Full text of The Relationship of Prices to Economic Stability and Growth : Compendium of Papers Submitted by Panelists Appearing Before the Joint Economic Committee

View original document

The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.

Session 88 }

JOINT COMMITTEE PEINT

THE RELATIONSHIP OF PRICES TO
ECONOMIC STABILITY AND GROWTH

COMPENDIUM
OP

PAPERS SUBMITTED BY PANELISTS APPEARING
BEFORE THE

JOINT ECONOMIC COMMITTEE

MARCH 31, 1958

Printed for the use of the Joint Economic Committee

UNITED STATES
GOVERNMENT PRINTING OFFICE
23734

WASHINGTON : 1958

For sale by the Superintendent of Documents, U. S. Government Printing Office
Washington 25, D. C. - Price $2.00

JOINT ECONOMIC COMMITTEE
(Created pursuant to sec. 5 (a) of Public Law 304, 79th Cong.)
WRIGHT PATMAN, Representative from Texas, Chairman
JOHN SPARKMAN, Senator from Alabama, Vice Chairman
HOUSE OF REPRESENTATIVES
SENATE
RICHARD BOLLING, Missouri
PAUL H. DOUGLAS, Illinois
HALE BOGGS, Louisiana
J. W. FULBRIGHT, Arkansas
HENRY 8. REUSS, Wisconsin
JOSEPH C. O'MAHONEY, Wyoming
HENRY O. TALLE, Iowa
RALPH E. FLANDERS, Vermont
THOMAS B. CURTIS, Missouri
ARTHUR V. WATKINS, Utah
CLARENCE E. KILBURN, New York
JOHN D. HOBLITZELL, JR., West Virginia

JOHN W. LEHMAN, Acting Executive Director
JAMES W. KNOWLES, Economist

LETTERS OF TRANSMITTAL
MARCH Si, 1958..
To Members of the Joint Economic Committee:
The papers transmitted with this letter were submitted by 47 leading economists invited to appear before the committee in panel discussions as part of the study: "The Relationship of Prices to
Economic Stability and Growth."
These papers are presented in advance of the committee's hearings, to be held May 12-22, to provide members of the committee, the
contributors, and the public an opportunity to examine the factsr
analyses, major issues, and conclusions in the various papers before
they are developed in oral statements and discussions at the hearings.

WRIGHT PATMAN,

Chairman, Joint Economic Committee.
in

MARCH 31,
Hon.

1958.

WRIGHT PATMAN,

Chairman, Joint Economic Committee,
House of Representatives, 'Washington, D. C.
DEAR MR. PATMAN : The papers transmitted with this letter were
submitted by 47 leading economists invited to appear before the Joint
Economic Committee in panel discussions May 12-22 as part of the
committee's study: "The ^Relationship of Prices to Economic Stability
and Growth." This is in accord with instructions to the staff approved
by the committee, October 7,1957.
The papers are presented as submitted by the contributors, without
additions or deletions. They are arranged by panel topics in the
order in which they are scheduled for discussion at the hearings.
In connection with these papers and the subsequent hearings, reference may be made to data included in the staff materials, "Productivity,
Prices, and Incomes," published by the committee last July. If th&
most recent data or revisions in data are not available readily in
current publications, supplementary tables are included in the appendix to this volume.

JOHN W. LEHMAN,

Acting Executive Director,

CONTENTS
Page

Introduction by Wright Patman, chairman, Joint Economic Committee.
Contributed papers:
I. Employment Act Objectives and the Stabilization of Prices
The Employment Act of 1946: The Dynamics of Public Economic Policy, Grover W. Ensley, National Association of
Mutual Savings Banks
,
Employment Act Objectives and the Stabilization of Prices,
Edwin G. Nourse, Economic Consultant
Price-Level Stability and Employment Act Objectives, Joseph
Aschheim, The Johns Hopkins University
How Important Is Price Stability in Stable Economic Growth?
G. L. Bach, Carnegie Institute of Technology
Price Behavior, Stability and Growth, William J. Baumol,
Princeton University
Economic Stability, Economic Growth and Price Stability, Betty
G. Fishman and Leo Fishman, West Virginia University. .
II. The Measurement of Price Changes and Price Relationships
The Measurement of Price Changes, Kenneth J. Arrow, Stanford
University
Administered Prices in the American Economy, Martin J.
Bailey, University of Chicago
The Price Indexes of the Bureau of Labor Statistics, H. E.
Riley, Bureau of Labor Statistics, United States Department
of Labor
Price and Income Measures for American Agriculture, Oris V.
Wells, Agricultural Marketing Service, United States Department of Agriculture
III. Past Price Behavior Viewed in the Context of Cyclical and Secular
Economic Changes
The Behavior of Prices, 1890-1940, Clarence H. Danhof, Tulane
University
An Interpretation of Price Movements Since the End of World
War II, Bert G. Hickman, The Brookings Institution
Cyclical Changes in Labor Cost, Thor Hultgren, National
Bureau of Economic Research, Inc
Trends in Product Prices, Factor Prices, and Productivity,
John W. Kendrick, The George Washington University
IV, Interrelationships Among Prices, Demands, and Costs
The Supply of Money and Changes in Prices and Output,
Milton Friedman, University of Chicago
Inflationary Depression and the Regulation of Administered
Prices, Abba P. Lerner, Roosevelt University and The Johns
Hopkins University
The Destabilizing Influence of Raw Materials Prices, Ruth P.
Mack, National Bureau of Economic Research, Inc
Relationships Between Foreign and American Prices, Gardner
Patterson, Princeton University
Prices, Costs, Demand, and Output in the United States, 194757, Richard Ruggles, Yale University, and Nancy D. Ruggles,
United Nations
Models of the Pricing Process, J. Fred Weston, University of
California, Los Angeles

VII

xi
xin
1
13
23
33
49
61
75
77
89
107
117
131
133
143
211
225
237
241
257
269
285
297
309

VIII

CONTENTS

V. Interrelationships Among Prices, Employment, Output, Incomes,
and Resources
The Composition of the Price Structure, Resource Allocation,
and Employment Levels, Clark C. Bloom, State University of
Iowa
Investment and the Price System, G. D. Bodenhorn, University
of Chicago
Resource Allocation, Economic Stabilization, and Public Policy
Toward Prices, Carl F. Christ, the University of Chicago 1
Inflation, the Wage-Price Spiral and Economic Growth, Otto
Eckstein, Harvard University
Pricing for Stability and Growth, John P. Lewis, Indiana
University
VI. Private Pricing Policies: Their Formulation and Effects
Business Price Policies and Economic Stability, Wroe Alderson,
Alderson & Sessions, Inc
Prices and Business Cycles, Lawrence E. Fouraker, Pennsylvania State University
Retail Price Policies, Stanley C. Hollander, Michigan State
University
Some Characteristics and Economic Effects of Pricing Objectives in Large Corporations, Robert K. Lanzillotti, State
College of Washington
Cyclical Implications of Private Pricing Policies, Alfred R»
Oxenfeldt, Columbia University
VII. Relationships Between Public Policies, Private Pricing Policies,
Price Changes, and Price Relationships
Price Effects of Tax Changes, George E. Lent, The Amos Tuck
School of Business Administration, Dartmouth College
Monetary Policy and the Structure of Markets, Warren L.
Smith, University of Michigan
Government Policy Toward Competition and Private Pricing,
Myron W. Watkins and Joel B. Dirlam, Boni, Watkins,
Jason & Co., Inc
The Effect of Government Spending Programs on Private Price
Formation, Murray L. Weidenbaum, Convair Division of
General Dynamics Corporation
The Influence of the Antitrust Laws and Related Governmental
Policies on Prices, Simon N. Whitney, Federal Trade Commission
VIII. Formulating Public Policies for Economic Stability and Growth
An Economic Policy for Economic Growth and Stability,
Oswald H. Brownlee, University of Minnesota
The Coordination of Price, Wage, Fiscal, and Monetary Policies
in Norway, Sweden, and the United States, Allan Gruchy,
University of Maryland
The Optimal Mix of Stabilization Policies, Richard A. Musgrave,
University of Michigan
Price Policy and Economic Growth and Stability, Arthur
Smithies, Harvard University
General summary
A Third Approach to the Analysis and Control of Inflation, Gardner
Ackley, University of Michigan
Maintaining Prosperity Without Inflation: Objectives, Problems, and
Policies, Neil H. Jacoby, University of California, Los Angeles. _ _
Price Level Stability and Economic Policy, Albert E. Rees, University
of Chicago
A General View of the Inflation Problem, Herbert Stein, Committee
for Economic Development
Relationship of Prices to Economic Stability and Growth: A Statement of the Problem, Robert C. Turner, Indiana University
App endix

319
323
335
345
361
375
397
399
417
425
441
461
477
479
493
513
529
555
573
575
583
597
611
617
619
637
651
665
671
685

CONTENTS;

CONTRIBUTORS
Ackley, Gardner, Chairman and Professor, Department of Economics,
University of Michigan, "A Third Approach to the Analysis and Control
of Inflation"
Alderson, Wroe, Alderson & Sessions Inc., Philadelphia, "Business Price
Policies and Economic Stability"
Arrow, Kenneth J., Professor of Economics and Statistics, Stanford University, "The Measurement of Price Changes"
Aschheim, Joseph, Assistant Professor of Political Economy, The Johns
Hopkins University, "Price-Level Stability and Employment Act Objectives"
Bach, George L., Dean, Graduate School of Industrial Administration,
Carnegie Institute of Technology, "How Important is Price Stability
in Stable Economic Growth?"
Bailey, Martin J., Professor of Economics, University of Chicago, "Administered Prices in the American Economy"
Baumol, William J., Professor of Economics and Sociology, Princeton University, "Price Behavior, Stability and Growth"
Bloom, Clark C , Professor of Economics and Assistant Director, Bureau
of Business and Economic Research, State University of Iowa, "The
Composition of the Price Structure, Resource Allocation, and Employment Levels"
Bodenhorn, G. Diran, Professor of Economics, University of Chicago,
"Investment and the Price System"
Brownlee, Oswald H., Professor of Economics, University of Minnesota,
"An Economic Policy for Economic Growth and Stability"
Christ, Carl F., Professor of Economics, University of Chicago, "Resource
Allocation, Economic Stabilization, and Public'Policy Toward Prices" __
Danhof, Clarence H., Professor of Economics, Tulane University, "The
Behavior of Prices, 1890-1940"
Dirlam, Joel B., Senior Consultant, with Myron W. Watkins, Vice President, Boni, Watkins, Jason & Co., Inc., Government Policy Toward
Competition and Private Pricing"
Eckstein, Otto, Professor of Economics, Harvard University, "Inflation,
the Wage-Price Spiral and Economic Growth"
Ensley, Grover W., Executive Vice President, National Association of
Mutual Savings Banks, "The Employment Act of 1946: The Dynamics
of Public Economic Policy"
Fishman, Leo, Professor of Economics and Finance, and Betty G. Fishman,
Lecturer in Economics, West Virginia University, "Economic Stability,
Economic Growth and Price Stability"
Fouraker, Lawrence E., Professor of Economics, Pennsylvania State
University, "Prices and Business Cycles"
Friedman, Milton, Professor of Economics, University of Chicago, National
Bureau of Economic Research, and Center for Advanced Study in the
Behavioral Sciences, "The Supply of Money and Changes in Prices and
Output"
Gruchy, Allan, Professor of Economics, University of Maryland, "The
Coordination of Price, Wage, Fiscal, and Monetary Policies in Norway,
Sweden, and the United States"
Hickman, Bert G., The Brookings Institution, "An Interpretation of Price
Movements Since the End of World War I I "
Hollander, Stanley C , Associate Professor of Business Administration,
Michigan State University, "Retail Price Policies"
Hultgren, Thor, Research Staff, National Bureau of Economic Research,
"Cyclical Changes in Labor Cost"
Jacoby, Neil H., Dean, Graduate School of Business Administration,
University of California, Los Angeles, "Maintaining Prosperity Without
Inflation: Objectives, Problems, and Policies"
Kendrick, John W., Associate Professor of Economics, The George Washington University, and Research Staff, National Bureau of Economic
Research, "Trends in Product Prices, Factor Prices, and Productivity" __
Lanzillotti, Robert K., Associate Professor of Economics, State College of
Washington, "Some Characteristics and Economic Effects of Pricing
Objectives in Large Corporations"
Lent, George E., Visiting Professor of Business Economics and Director
of Research, The Amos Tuck School of Business Administration, Dartmouth College, "Price Effects of Tax Changes"

Pag6
619
399
77
23
33
89
49

323
335
575
345
133
513
361
1
61
417

241
583
143
425
211
637
225
441
479

CONTENTS

Lemer, Abba P., Professor of Economics, Roosevelt University, and Visiting Professor of Political Economy, The Johns Hopkins University,
"Inflationary Depression and the Regulation of Administered Prices". _ 257
Lewis, John P., Professor of Business Economics and Public Policy,
Indiana University, "Pricing for Stability and Growth"
375
Mack, Ruth P., Research Staff, National Bureau of Economic Research,
"The Destabilizing Influence of Raw Materials Prices"
269
Musgrave, Richard A., Professor of Economics, University of Michigan,
"The Optimal Mix of Stabilization Policies"
597
Nourse, Edwin G., Economic Consultant, Washington, D. C, "Employment Objectives and the Stabilization of Prices"
13
Oxenfeldt, Alfred R., Professor of Marketing, Graduate School of Business,
Columbia University, "Cyclical Implications of Private Pricing Policies". 461
Patterson, Gardner, Professor of Economics, Princeton University, "Relationships Between Foreign and American Prices"
285
Rees, Albert E., Associate Professor of Economics, University of Chicago,
"Price Level Stability and Economic Policy"
651
Riley, H. E., Chief, Division of Prices and Cost of Living, Bureau of Labor
Statistics, United States Department of Labor, "The Price Indexes of
the Bureau of Labor Statistics"
107
Ruggles, Richard, Professor of Economics, Yale University, and Nancy
D. Ruggles, Bureau of Economic Affairs, United Nations, "Prices, Costs,
Demand, and Output in the United States, 1947-57"
297
Smith, Warren L., Associate Professor of Economics, University of Michigan, "Monetary Policy and the Structure of Markets"
493
Smithies, Arthur, Professor of Economics, Harvard University, "Price
Policy and Economic Growth and Stability"
611
Stein, Herbert, Director of Research, Committee for Economic Development, "A General View of the Inflation Problem"
665
Turner, Robert C, Professor of Economics, Indiana University, "Relationship of Prices to Economic Stability and Growth: A Statement of
the Problem"
671
Watkins, Myron W., Vice President, with Joel B. Dirlam, Senior Consultant, Boni, Watkins, Jason & Co., Inc., "Government Policy Toward
Competition and Private Pricing"
513
f
Weidenbaum, Murray L., Senior Operations Analyst, Convair Division,
General Dynamics Corp., "The Effect of Government Spending Programs on Private Price Formation"
529
Wells, Oris V., Administrator, Agricultural Marketing Service, United
States Department of Agriculture, "Price and Income Measures for
American Agriculture"
117
Weston, J. Fred, Professor of Finance, Graduate School of Business Administration, University of California, Los Angeles, "Models of the Pricing
Process"
309
Whitney, Simon N., Director, Bureau of Economics, Federal Trade Commission, "The Influence of the Antitrust Laws and Related Governmental Policies on Prices"
555

INTRODUCTION BY WRIGHT PATMAN, CHAIRMAN,
JOINT ECONOMIC COMMITTEE
The Employment Act of 1946 grew out of legislative debate of
public and private policies which would promote "full" or "maximum" employment of the Nation's productive resources—human and
material. Economic stability and growth were mainly thought of
in terms of minimizing fluctuations in employment and of maximizing the growth of real output of goods and services. From its initial organization, the Joint Economic Committee, however, has been
concerned also with problems of promoting stability of the general
price level and the need for avoiding inflation.
At a meeting of the Joint Economic Committee on July 24, 1957,
there was general agreement that the -committee should explore the
possibilities and problems of conducting a full-scale investigation of
prices and price-making processes in relation to economic stability
and growth. In accordance with my suggestion, the staff was directed to prepare a specific program for consideration by the full committee at a meeting on last October 7.
At the October 7 meeting, the committee reviewed the program
proposed by the staff, made several modifications and additions, and
then decided to proceed with the study.
PURPOSE OF PRESENT STUDY

In the present project, the committee's major goal is an objective
and authoritative exploration of general economic processes which involve prices, price relationships, costs, and price policies, in the expectation that this will reveal ways in which public and private
policies can contribute to attainment of the Employment Act
objectives:
SEC. 2. The Congress declares that it is the continuing policy and responsibility of the Federal Government to use all practicable means consistent with
its needs and obligations and other essential considerations of national policy,
with the assistance and cooperation of industry, agriculture, labor, and State
and local governments, to coordinate and utilize all its plans, functions, and
resources for the purpose of creating and maintaining, in a manner calculated
to foster and promote free competitive enterprise and the general welfare, conditions under which there will be afforded useful employment opportunities, including self-employment, for those able, willing, and seeking to work, and to
promote maximum employment, production, and purchasing power. (15 U. S. C.
1021.)
PREVIOUS AND CURRENT STUDIES

The Joint Economic Committee itself has, of course, already made
a number of studies in the field of prices. The first hearings ever held
by the committee were on current price developments and the problem of economic stabilization, in June and July 1947. In the fall of
1947, working through regional subcommittees, members visited many
parts of the country, holding hearings, gathering data, and report-

XII

INTRODUCTION

ing on consumer prices. An exhaustive study of profits in 1949, and
several earlier and later hearings on prices in the steel industry are
also cases in point. Prices, inflation, and controls have, moreover,
been recurrent topics for panel and committee discussions in connection with the committee's study of successive annual reports of the
President. More recently, acting upon instructions of the committee,
the staff compiled a factbook on Productivity, Prices, and Incomes,
leaving interpretations and conclusions to the user of the materials.
Other committees of the Congress are currently undertaking studies
of prices in varying contexts. The Senate Finance Committee, during the session last year, conducted extended hearings on price inflation, and on attempts to control it through monetary and fiscal policy.
The Senate Judiciary Committee's Subcommittee on Antitrust and
Monopoly has been investigating prices and pricing behavior, especially in the so-called "administered-price" area. A subcommittee
of the House Agricultural Committee has been studying the closely
related subject of marketing margins for farm products.
PLANS FOR THIS STUDY

The committee study of prices is being carried forward in four
stages. This compendium or collection of research papers, designed
to bring together analyses and findings of leading impartial experts
and the most authoritative information available, is the first part of
the program.
The second part will be a series of hearings in the form of panel
discussions scheduled for May 12-22. The panelists will be the contributors of the papers in this compendium in which the papers are
arranged by panel topics in the order in which they are scheduled
for discussion at the hearings.
As the third stage of the study, the committee is inviting economists
from labor and industrial organization to submit papers for a volume
of comments to be published after the May hearings. These papers
will concentrate on the analyses and issues raised by the experts contributing to the present compendium and participating in the May
panels.
The fourth stage in the program will be hearings in the form of
panel discussions to be held later in the year. The panelists at that
time will be the economists from labor and industry together with
some of the contributors to the present compendium.
^ When the experts were invited to contribute to the present compendium they were supplied with a preliminary outline of topics and
questions of interest to the committee. The outline was intended
only to give the contributors some guide to the scope of the investigation into which their individual papers would fit. The relevant part
of that outline is reproduced in each section of this volume directly
preceding the papers for that panel.
On behalf of the committee, I wish to commend the contributors
for the valuable materials they have presented in these papers. Their
generous gift of time and resources to the project is an outstanding
public service. I am confident their contributions will be reflected in
public deliberations on policies for economic stability and growth
for many years to come.

EMPLOYMENT ACT OBJECTIVES AND THE
STABILIZATION OF PEICES

/ . Employment Act objectives and the stabilization of prices
A. What price behavior would be consistent with the attainment of the other policy objectives of the Employment
Act "of creating and maintaining, in a manner calculated
to foster and promote free competitive enterprise and the
general welfare, conditions under which there will be afforded useful employment opportunities, including selfemployment, for those able, willing, and seeking to work,
and to promote maximum employment, production, and
purchasing power" in both the short run and the long
run?
Under what conditions would stabilization of prices be inconsistent with attainment of Employment Act objectives ?
B. What does "economic stability and growth" mean when used
in reference to a dynamic private enterprise economy?
C. What changes in the distribution of income and wealth
usually accompany changes in the general price level, and
what weight should these be given in assigning priorities
among the various objectives of the Employment Act ?
1. What are the effects of price level changes upon the
different income groups, especially upon the socalled "fixed-income" groups ?
2. How are holders of various types of assets affected
by price level changes ?
3. How do various types of debtors—financial and nonfinancial, corporate, individual, farmers, and
small-business men—fair under conditions of
changing price levels ?
4. What is the effect of changing price levels on the
relative position of Federal, State, and local governments, considering their traditional revenue
sources, their expenditure programs, and debtor
positions ?
XIV

THE EMPLOYMENT ACT OF 1946: THE DYNAMICS OF
PUBLIC ECONOMIC POLICY
Grover W. Ensley, Executive Vice President,1 National Association
of Mutual Savings Banks
INTRODUCTION

Twelve years ago, the enactment of the Employment Act introduced
into public policy the first specific statement of objectives for the national economy. The language expressing those objectives is broad:
* * * creating and maintaining, in a manner calculated to foster and promote
free competitive enterprise and the general welfare, conditions under which
there will be afforded useful employment opportunities, including self-employment, for those able, willing, and seeking to work, and to promote maximum
employment, production, and purchasing power.

This broadness of language reflects the necessity for compromise
among divergent viewpoints, which is the essence of our democratic
political institutions. The machinery set up by the act for implementing its objectives is simple. Both the generalized statement of objectives and the simple machinery of the act have proved to be truly
elegant.
In the 12 years since adoption of the Employment Act, the Nation's
economy has been subject to many and varying stresses and strains.
On the whole, it has performed very well, indeed, in providing an
increasingly prosperous life for a rapidly growing population.
Making due allowance for the recessions of 1949 and 1954, the stated
objectives of the act have to a very substantial extent been continuously achieved. One cannot, of course, assert that this success is
the unique result of the Employment Act and its machinery, but there
is little question that they have contributed very materially.
One aspect of our postwar economic history, however, has caused
mounting concern. This is the virtually continuous upward pressure
on prices, reflected in a 59 percent increase in consumer prices, a 73
percent increase in wholesale prices, and a 60 percent increase in the
implicit gross national product price deflators since 1945. Inflation,
even the so-called creeping variety, is hardly a new phenomenon. The
problems of maintaining a strong, stable economy associated therewith, however, appear to become increasingly trying as our Nation becomes increasingly industrialized. Stabilization of some widely accepted concept of a general price level has more and more commanded
attention of those responsible for carrying out the mandates of the
Employment Act.
At the same time, the persistent rise in consumer prices during the
past two and a half years, despite the increasing vigor with which
antiinflationary public policies were pursued and the seeming abrupt1
The views expressed are the author's and do not necessarily represent those of the
National Association of Mutual Savings Banks or the association's individual members.

ECONOMIC STABILITY AND GROWTH

ness with which the economy in the latter part of 1957 slid into sharp
recession, has raised widespread questions concerning price-level stabilization. Is price-level stabilization an appropriate objective of the
Employment Act ? Can general monetary and fiscal policies achieve
such stabilization without simultaneously precluding realization of
other ob j ecti ves of the act ?
These questions, I believe, may be answered affirmatively if there is
continuation of the alertness and flexibility in public economic policies
wiiich has been demonstrated repeatedly since the Employment Act
became law. In this connection the very broadness of the act's statement of objectives has been one of its major strengths. Greater specification of objectives would, in all likelihood, have unduly circumscribed economic policy, and diminished its effectiveness in contributing to achieving any of the more broadly stated objectives of the act.
CHANGING PRIORITIES I N ECONOMIC POLICY OBJECTIVES

In the rapidly changing condition of the postwar world, questions
concerning the standards the Nation's economy should meet and the
costs which may be incurred if these standards are to be met have
commanded ever increasing attention. Several criteria have assumed
primary importance. The objectives of the Employment Act have, as
a result, come more and more to be stated in terms of achieving growth
and stability of a dynamic, free, competitive, private-enterprise
economy.
The apparent paradox in the semantics of public economic policies
warrants a brief definition of terms before coming to grips with our
principal question: the interrelationship of objectives.
Economic growth should be defined as an increase in the Nation's
capacity to produce goods and services. A more conventional concept
of economic growth is expansion of real per capita gross national
product. While increases in capacity and in product are surely closely
related over time, in the short run the economy may grow without an
increase in total output or, conversely, total output may increase without expansion of productive capacity. I shall leave detailed discussion
of these matters in the capable hands of the other contributors.
Economic stability means that there is a limited range of fluctuation
in the rate at which the Nation's economic capacity, human and material, is used. While stability in a growing economy, therefore, implies
a constantly rising level of employment and of total production of
goods and services, it does not imply any specific rate of resource
utilization. All other things being equal, of course, stability at a high
rate is always to be preferred over a low rate of resource use. In the
present American economy, as a matter of fact, it is unlikely that the
incentives for growth and dynamics would be long maintained without stabilizing resource use around a relatively high rate.
"Dynamic" describes perhaps the most outstanding characteristic of
our economic life: the frequency and magnitude of the changes in our
tastes and preferences, in our effectively expressed demands for goods
and services, and in the ways in which resources can most efficiently
be combined and used to produce these goods and services. Although
we need not be detained by elaboration of the sense of the phrase,
"free, competitive, private enterprise," I would emphasize the importance of the word "free." Too often, restraints on freedom of enter-

ECONOMIC STABILITY AND GROWTH

prise are thought of primarily as those imposed by government,
through the exercise of its regulatory, fiscal, and monetary powers.
Such restraints are not to be taken lightly, but they must not be
permitted to obscure the frequently more significant limitations imposed from within the private sector of the economy.
While a dynamic economy need not be a growing economy nor a
stable one, these three characteristics are not mutually inconsistent.
Their simultaneous realization, in fact, has become the overall objective of public economic policy.
This was not always the case. The Employment Act was legislated
in 1945 and early 1946 in an atmosphere of uncertainty and apprehension. Economists—with few exceptions—were persuaded that the
postwar era probably would be marked by reemergence of persistent
unemployment and economic stagnation. These fears had their
origins in the great depression of the 1930's and the failure of public
and private programs to restore real prosperity. At the outbreak of
World War II, there were still nearly 10 million unemployed persons
in the United States.
Substantial idle industrial capacity still
haunted management after 10 years of virtually no expansion of plant
and equipment. Here was economic stability of a universally unacceptable variety.
In view of this pessimism at the close of World War II, it is hardly
surprising that the Employment Act emphasized the promotion of
maximum employment, production, and purchasing power, and the
creating and maintaining of conditions affording useful employment
opportunities for those able, willing, and seeking to work.
There is no occasion for deprecating this early postwar emphasis
in economic policy objectives. As of the end of the war, there was
little empirical evidence to suggest the strength of accumulated consumer desires nor the facility with which they could be translated into
effective demand for goods and services by virtue of the vast wartime
accumulations of liquid savings. By the same token, it is difficult to
see how economic analysts in 1945-46 could have determined the impact
of the rapid wartime exploitation of technological advances on postwar business capital expenditure plans, or have estimated the consequences of high business liquidity for financing the realization of
these plans. In short, nothing is harder for the economist to predict
than the dynamics of the economy. Prewar experience hardly
afforded a basis for such predictions. Emphasis in public policy
on maximizing employment and production, without explicit regard
for growth or dynamics in the early postwar days, therefore, is quite
understandable.
The great and generally unexpected burst of business and consumer
demand in the early postwar years almost immediately shifted the
focus of concern by those entrusted with public responsibility for
achieving the objectives of the Employment Act. Widespread upward
price pressures originating in excessive total demand made it clear
that the stabilization problem was not the "deflationary gap" anticipated by many at the end of the war, but rather a renewal of the
"inflationary gap." The first major inquiry of the Joint Economic
Committee in June and July of 1947, therefore, was devoted to price
developments.
23734—58

ECONOMIC STABILITY AND GROWTH

The recession of 1948-49 and the outbreak of hostilities in Korea in
mid-1950 provided a further test of the flexibility of the Employment
Act. Within this 3-year period:
Gross national product (in current prices) rose $13 billion—6%
percent in a year's time, fell $8.5 billion—3.2 percent in another year,
and then rose nearly $50 billion—19.4 percent in the third year. Unemployment increased by nearly 90 percent between the second quarter of 1948 and the fourth quarter of 1949 and fell by 44 percent by
the fourth quarter of 1950.
Public policy was called upon twice in 3 years to reverse its field—
and drastically so. There is widespread agreement that Federal fiscal
policy, which at that time carried the major burden of economic
stabilization, was on the whole well suited to meet these rapidly
changing challenges.
A concurrent development in determining the focus of economic
policy was the emerging emphasis on economic growth, proceeding
largely from increasing awareness of the responsibilities imposed
upon the United States as the leader of the free world in the cold war,
and of the economy's demonstrated growth tendencies. This emphasis has come to the fore in the post-Korean period and with
shocking abruptness since the fall of 1957. The work of the Joint
Economic Committee and its subcommittees in this area is a testimonial to their alertness to the importance of economic growth and
the diligence of their efforts to bring these considerations to the attention of the Congress, the Executive, and the Nation as a whole.
A NEW FRONTIER : STABILIZING THE PRICE LEVEL

Without belaboring the details or effectiveness of policy changes in
the postwar period, I think it is clear that experience under the Employment Act has demonstrated amply the dynamic quality of public economic policy. The recent past and our present situation suggest quite clearly the basic outlines of a major test to be faced:
achieving reasonable stability in the general level of prices without
sacrificing growth, stability, and the dynamic characteristics of the
economy nor the institutional freedom we prize.
Compatibility of public policy objectives
Are these objectives inconsistent? The answer is "Yes" only if
we choose to regard each objective in some absolute sense. Instead,
it must be recognized that the realization of additional gains with
respect to any one objective involves costs in the sense of some limitation on the extent to which gains with respect to any other can be
achieved. We must, therefore, continually effect compromises among
conflicting claims. This is the essence of the basic economic process
in any environment in which ends exceed means.
We look forward to a higher living standard tomorrow than we
enjoy today and recognize that to make this gain possible some sacrifice of today's consumption of goods and services or of leisure is required. We are continually required to choose between the satisfactions derived from variety in consumption and in the ways in which
we produce our incomes, on the one hand, and the losses in terms of
temporary unemployment, obsolescence, and the disruption of familiar
routines involved in shifting to new products and new methods of

ECONOMIC STABILITY AND GROWTH

production. We are continuously faced with the need for readjusting these priorities if our first approximations involve too great
changes in the rate of use of our resources. The fact that we are, all
of us, continually making these compromises demonstrates the compatibility of these objectives rather than their inconsistency.
We make these decisions as individuals in the setting of what we
hope is a substantially free market place. The aggregate of these
decisions reflects the order of priorities accorded by the economy as
a whole to various economic objectives. It is to be expected that
changing circumstances will lead to changing priorities.
Government, through the exercise of its monetary, fiscal, and regulatory powers, provides major conditioning factors which in turn influence our individual decisions. It is in fact virtually impossible to
conceive a complex of Government economic policies which would be
neutral in its impact on this system of choices. Public-policy makers,
therefore, are invested with tremendous responsibility to provide policies which will facilitate, rather than frustrate, the implementation
•of the choices registered in the market, without giving way to pressures from speculative excesses nor to irrational responses to minor
changes in economic conditions. It also imposes a substantial burden
on public instrumentalities to keep the market as free as possible from
restraints on the allocation of resources.
If these public responsibilities are appropriately discharged, the
Nation can be confident of realizing a high rate of economic growth
without unduly sacrificing stability or the dynamic characteristics of
our economic life. We must avoid, however, underestimating the
cost involved, lest we become too rigid in our thinking and our policy
attitudes.
For example, the more dynamic the economy, the greater will be
the problems of maintaining stability in the rate of resource use
while continuing to provide the real resources required for economic
growth. The more frequent and the greater the changes in technology and in resulting market demand and supply conditions, the
more frequent and the greater will be the pressure for changes in
the use of resources. Such changes cannot be achieved instantaneously, but in fact may require a considerable lapse of time. The
rate of frictional unemployment of both human and material resources, therefore, will tend to be higher in a relatively dynamic economy than in a static economy. An excessively rigid stabilization
policy may impair economic dynamics by increasing the real costs of
effecting necessary shifts in resources.
Achieving a, high rate of growth in a dynamic setting, therefore, requires a flexible approach to economic stabilization. Such an approach has been aptly visualized as a band representing a range of
fluctuation in unemployment through time, as contrasted with a line.
How wide or narrow this band should be cannot be precisely determined, since our attitudes assuredly will change through time. It
should be emphasized, however, that accepting some fluctuation in unemployment rates is perfectly consistent with the stabilization objectives implied in the Employment Act. Moreover, this acceptance
does not imply a relaxation of vigilance by those responsible for
economic policy. Rapid changes in the employment situation, even
though they do not at the moment exceed accepted tolerances, never-

ECONOMIC STABILITY AND GROWTH

theless demand alertness to the possibility that they shortly will, and
preparedness to deal promptly with an excessive fluctuation should it
in fact materialize.
The importance of relative price flexibility
How is stability of prices related to whatever composite objective
of growth, stability, and dynamics may be desired? At the outset,
the distinction between stabilizing the prices of particular goods and
services and stabilizing the level of prices must be emphasized. Stabilizing specific prices can hardly be justified as a public policy objective for a free-market economy except under the most extraordinary circumstances. Flexibility of prices of individual goods
and services is, of course, the specific mechanism a free, privateenterprise economy uses to induce shifts in resources and in economic activity. In the absence of relative price movements, the mechanism for such shifts must be public or private rationing. Either
will result in a loss in economic welfare in a private-enterprise setting.
Moreover, the more dynamic we wish our economy to be, the more
important is mobility of resources, and, consequently, the greater the
emphasis we must place on the flexibility of individual prices. If,
for whatever reason, the change in the price of a good or service for
which supply or demand conditions have changed is restrained relative to the prices of other things, the extent of the shift in resources
which otherwise would have occurred will also be limited.
Maintaining flexibility in prices of individual goods and services,,
therefore, is an important collateral objective of public policy, the
responsibility for which rests primarily in the antitrust field. Since
such price flexibility is, in part, a concomitant of the uncertainty inherent in a dynamic economy, the inclination of those controlling the
use of resources to insulate themselves from price flexibility is readily
understood. It must not be overlooked, however, that to the extent
this insulation is achieved, the economy will suffer, as a necessary
corollary, a loss in dynamics. Antitrust policy which fully recognizes
this basic drive toward relative price rigidity and its adverse implications for the character of the Nation's economic life, therefore, is a
necessary adjunct of public economic policy seeking the objectives of
the Employment Act.
Encouraging flexibility in individual prices is in no wise inconsistent with stability in some overall average of prices, at least conceptually. All that is required by the former is change relative to other
prices. Theoretically, relative price changes can take the form of
increases offset by decreases, leaving the average unchanged; by
smaller increases more than offset by larger decreases, with a corresponding declining average level; by larger increases only partially, if
at all, offset by smaller decreases, resulting in a rising average leveL
In practice, only the latter alternative appears to have been feasible
during the postwar period.
Distinguishing between relative price flexibility and general price
instability
It is generally agreed that while relative price flexibility is necessary and desirable for our economy, price-level instabilitv must be
minimized. Although conceptually the distinction here is fairly clearr
in practice it is considerably more difficult to determine. Other papers

ECONOMIC STABILITY AND GROWTH

in this volume will deal with the problems, often extremely technical,
of applying the concepts of price-level stability to public policy purposes. It will suffice at this point to suggest only one of the many
difficulties policymakers must face.
Suppose the price of some basic industrial raw material rises. The
factors underlying this type of price change are more often than not
extremely difficult to assess with any precision with information of
the accuracy and extent now available. Yet, if price level stability is
to be maintained, public policy must be based on judgments whether
these factors are fairly widespread and, therefore, whether the particular price rise is to be regarded as a harbinger of general upward
price pressures. Moreover, public policy must appraise the degree of
substitutability for the raw material in question to determine the
extent to which its price rise is likely to be reflected in other prices.
The greater the substitutability and the more promptly substitution
can be achieved, the less need be the concern over extension of the
price movement. Public policy, therefore, must depend also on evaluation of market structure, pricing practices, resource mobility, etc., in
the affected industries, since without this analysis, it cannot appraise
the likely aggregate demand and supply responses with which it must
deal. In practice, therefore distinguishing between desirable relative
price flexibility and undesirable general price instability is now—
and always has been—difficult. Granting this sort of practical difficulty does not, however, derogate price-level stability as a public policy
objective.
Creefing inflation and economic instability
There is no occasion in this discussion for dwelling on the manifest
evils of hyperinflation or hyperdeflation. The price-level movements
with which public policy in the United States must be concerned are of
the "inching" or "creeping" variety. This term, as generally used,
may include price movements which are so pronounced as to average several percent a year. Such changes are modest, as compared with
the hyperinflation of post-World War I I Hungary and China, for
example. It is precisely because of the relative slowness of "creeping"
inflation that it is so insidious and difficult to cope with.
Over a considerable part of our postwar experience with "creeping" inflation, this type of price movement has been widely interpreted
as a reflection of excessive total demand. So long as this assumption
appeared to accord closely with the facts revealed by other statistical
series, economic stabilization, in the sense of stabilizing the rate of
total resource use, and price-level stabilization could be and were embraced as a single objective. In 1946-48 and 1950-51, for example,
the need for stabilization efforts was signaled by general price movements, and public policies aimed at restricting total spending were
synonymous with restriction of the rate of price-level increases. In
other words, so far as anyone could see, the upward pressure on prices
originated in excessive demand which also was reflected in strong
tendencies toward "overemployment."
Since mid-1955, however, it has become increasingly apparent that
price-level stabilization and stabilization of the rate of resource use are
not necessarily synonymous. Experience during the latter part of
1957 and the early part of 1958 in particular has demonstrated
graphically the possibility of divergence of price-level changes from

ECONOMIC STABILITY AND GROWTH

changes in the rate of resource use. Timelags undoubtedly account
for some part of this divergence. Recently there has been considerable discussion, which will be extended and refined in other papers
in this compendium, of other factors which may contribute to it. My
purpose in alluding to this divergence is merely to show that economic
stability or instability is not necessarily the same as price-level stability or instability as was widely assumed in the past. Price-level
stabilization, as an objective of public economic policy, therefore,
should be examined, at least in part, on its own merits.
Equity aspects of inflation
The considerations which strongly support price-level stability as an
objective of economic policy can be placed in the rough, but by no
means mutually exclusive categories of equity and economics. The familiar equity argument is that instability in the general price level—
more specifically, in terms of postwar experience, inflation—diminishes the purchasing power of individuals and business entities with
fixed or "sticky" incomes and assets relative to the purchasing power of
those whose incomes and assets are responsive to general price movements. Insofar as those with fixed incomes and assets are concentrated
at the lower end of the income and wealth distribution or are the small
or new business entities, inflation represents a regressive tax. As
such, it effects changes in income and wealth distribution which are
antithetical to those sought by the explicit provisions of the Nation's
tax structure.
My fellow panelist, Dean Bach, has directed intensive studies concerning the impact of inflation on income and wealth distribution.
I shall not, therefore, attempt to cover this ground with any detailed
discussion.
Some economists have suggested that these equity consequences of
inflation can be and have been substantially ameliorated by appropriate private and public policies. It is argued, for example, that
persons retired under public retirement systems have suffered only
temporarily from persistent general price increases, since these retirement systems are liberalized periodically to take account of inflation.
Industrial retirement systems, it is argued, can turn more extensively
to variable annuity devises to hedge the retiree against price-level
fluctuations. Individuals and businesses, it is maintained, can change
the disposition of their personal savings toward more price-levelsensitive assets, particularly common stocks.
The answer to the first argument, of course, is that the losses suffered
by the public system retiree are never made good and are, therefore,
permanent. Adjusting public retirement system benefits frequently
enough to reduce these losses to insignificant proportions would hardly
be feasible. The answer to the other contentions is that real costs
are incurred by giving up liquidity and safety in seeking inflation
hedges. In the absence of such costs, the great savings institutions of
this country would never have reached their present stage of development, nor should we now see the great diversity in the character and
operations of financial intermediaries. It seems to me to be extremely
difficult to justify asking the economically weak to incur these costs to
protect themselves from inflation which is in any case regarded as
inequitable.

ECONOMIC STABILITY AND GROWTH

Moreover, it is at least questionable whether the personal thrift
patterns which would emerge from these suggestions would contribute
so well to providing for the economic growth which is one of the primary objectives of public policy. Personal savings habits dominated
by inflation hedging might well result in a perverse reinforcing of
cyclical movements and of inflation. Those who save to provide some
relatively fixed amount for retirement, for example, might well be
inclined to reduce their current savings under boom conditions if
their inflation hedging were, in fact, successful, and to increase their
savings during an economic decline if inflation-responsive assets were
also recession responsive. I do not know what proportion of total
personal savings are, in fact, motivated by some fixed total objective,
but whether the proportion is large or small, it is difficult to see why
we should encourage distortion of these savings patterns as a means
of justifying inflation.
Inflation and growth
According to a widely held view, the rate of economic growth can be
expected to respond favorably to mild inflation. The argument rests
primarily on the assumption that costs will lag behind prices, so that
profit margins rise. Rising profit margins imply a positive shift in
the schedule of the marginal productivity of capital which, in turn,
serves to increase investment demands. An expansion of capital outlays will be reflected in a multiple increase in total demand which adds
to upward price pressures, presumably still in advance of factor costs.
Thus, it is maintained, even mild inflation tends to be cumulative,
largely through its effects on stimulating growth-generating expenditures. The process presumably comes to a halt when the increase in
productive capacity exceeds the increase in effective demand or when
factor costs close the gap with product prices and pinch off profit
margins.
Dean Bach's studies, upon which he reported in this committee's
Tax Policy Subcommittee study of Federal tax policy for economic
growth and stability, cast some serious doubts about this delineation
of the relationship between inflation and growth. It seems to me that
an alternative thesis suggests that inflation may impede economic
growth more than stimulate it.
If product and resource markets were substantially free of restraints
or monopolistic elements, a system of monetary and fiscal policies
which limited increases in total demand in line with increases in capacity would also provide stability in the general price level. Failure
of such a system to prevent general price movements, therefore, indicates imperfections in market structures. If these imperfections are
accepted or tolerated by the Nation, the relative price changes necessary to effectuate dynamic impulses in the economy must be reflected
either in general price level changes, in instability in the rate of
resource use, or in an otherwise undesired shift in the division of
resources between growth and current consumption. Suppose that the
Nation's composite choices place, as a first approximation, a higher
priority on stability in the rate of resource use and on the desired
rates of growth and of current consumption than on price-level stability. On what grounds may one conclude that the resulting general
movement of prices is undesirable?

ECONOMIC STABILITY AND GROWTH

In the first place, inflation cannot finance the attainment of real
product objectives for which real resources are inadequate. I t can
serve only to change the total product "mix" by shifting effective
demand in favor of successfully hedged claimants and to the disadvantage of the relatively unsuccessful. Some sacrifice of the desired
rate of growth or of consumption, therefore, will necessarily result.
Moreover, such shifts are likely to be difficult to anticipate or to estimate in advance of their actual occurrence. Inflation, therefore, is
likely to produce what appear to be haphazard and capricious changes
in income positions and in asset and liability values. These effects
must serve to increase uncertainty. As a consequence they also must
add to the costs of growth by requiring more extensive use of resources
for hedging against the increased risks accompanying capital accumulations.
Second, inflation shifts resources along the lines of least resistance,
so to speak. The sacrifices of claims to resources, therefore, will fall
more heavily on economic units lacking the strength of the monopoly
elements which give rise to market imperfections. In other words,
the weakest economic units—the new or the small business, to take
familiar examples—will be required to bear a disproportionate share
of the burden of financing in real terms the shift in the use of currently available resources. Failure to control a little bit of inflation,
therefore, is quite likely to lead to increasing difficulty in curbing
subsequent inflationary pressures. It seems clear that a relaxed policy
toward inflation more greatly jeopardizes the economic position of
the State or local government or school district, of the new or small
business, of the prospective home buyer relying on a thin equity, to
use the favorite examples of critics of general monetary controls, than
does a, tight-money policy. This, I believe, is the principal basis for
asserting that moderate, "inching," or "creeping" inflation always
holds the promise of ever stronger inflationary movements. If, finally,
after long neglect, opposition to the adverse equity, competitive, and
resource-allocating consequences of inflation becomes strong enough,
efforts to curb its further extension then are likely to involve more
painful consequences in terms of arresting growth and sacrifices of
current production and employment than would have been required
if the initial modest inflation had been halted.
General price level stability, therefore, must be regarded as a necessary collateral objective of public policies aimed at promoting growth
and stability in a dynamic, free, private-enterprise economy.
FEASIBILITY OF PRICE-LEVEL STABILIZATION AS AN EMPLOYMENT ACT
OBJECTIVE

Can general monetary and fiscal policies achieve price-level stability without undue sacrifice of growth and stability in the rate of
resource use ? Much of the criticism of using general monetary and
fiscal restraints to curb inflation depends on the assertion that some
business and labor organizations can insulate their price and wage
decisions from changes in total demand conditions. General restraints on demand, it therefore is maintained, will impinge only on
economic entities lacking this power. Restraints sufficiently vigorous
to break through the insulation of big business and big labor, the argu-

ECONOMIC STABILITY AND GROWTH

ment runs, necessarily must hold total demand to less than fullemployment levels.
I must agree, as indicated above, that if, indeed, such power rests
in big business and labor (it should be noted that the validity of this
contention has not yet been demonstrated), the consequent limitations
on relative price flexibility and resource movement will impede the
successful use of general constraints against inflation. I cannot
agree, however, that the existence of such power calls for abandoning
the use of fiscal and monetary policy for price-level stabilization purposes, or for foregoing price-level stability as an objective of public
policy, or for relying on the good faith and intelligence of private
power groups to behave in a manner consistent with stability of
prices generally. The proposition that the value of the Nation's
money depends on the good intentions and behavior of a relatively
small private group in the population must surely be abhorrent to any
free society. If, in fact, the proposition is correct, it calls for constructive measures to reduce this power. Such measures also would
contribute to a more dynamic economy with greater private freedom
to be enterprising.
Price-level stabilization, therefore, can be a practicable objective
of public policy. Indeed, taking the steps required to make it so
will also strengthen policies directed toward achieving a high rate
of growth and stability in the rate of resource use.
As in the case of these latter objectives, precise specification of the
objective of price-level stability is not possible. Moreover, I doubt
that it would be desirable. After all, our economic stabilization objective does not rule out some fluctuation in the rate of resource use.
By the same token, as a practical matter we cannot pursue absolute
rigidity in the price level. What we seek is the best possible "mix n
of all of our major economic policy objectives. Our best hope for its
attainment, I believe, rests in the alertness and adaptability of those
charged with responsibility for public economic policy.
Our public-policy instrumentalities have shown flexibility and
alertness to desirable shifts in emphasis. They have not always, of
course, been completely successful. We have had recessions and interruptions of growth. We well may have experienced, without recognizing it, some inhibition of the economy's dynamism. There are
many expert observers of the American economy who assert that we
have suffered increasing private monopolistic restraints on the freedom
of resources and markets. And we assuredly have had inflation. But
while we recognize these shortcomings, we should not lose sight of
our accomplishments. We have not had a serious depression and we
have achieved remarkable progress in raising our living standards
since the war's end.
I think we can be confident, therefore, that the present language
of the Employment Act can continue to provide the framework for
the dynamic policies the Nation's economic development demands.
This study by the Joint Economic Committee, which carries forward
the high standards of those of the past, will, I am sure, contribute
to an extended appreciation of the significance of price-level stabilization policy in the context of the Employment Act's broad statement of
objectives.

EMPLOYMENT ACT OBJECTIVES AND THE
STABILIZATION OF PRICES
Edwin G. Nourse, Economic Consultant
I believe this committee is marking a new milestone in the interpretation and application of the Employment Act by conducting this
series of hearings on "The Relationship of Prices To Economic Stability and Growth." Such an inquiry is of utmost timeliness just
now, as the policy set forth in the Employment Act is facing its first
severe test. And great promise for the outcome of the investigation
may be found in the terms in which Chairman Patman's announcement of the hearing stated its purpose—it was to be—
an exploration of general economic processes which involve prices, price relationships, costs, and price policies * * * public and private [that] can contribute
to * * * maximum employment, production, and purchasing power.

The depth and breadth of this definition are in refreshing contrast
to some of the oversimplified and overmechanistic concepts of the
employment problem that are still current,
THE ACT^S STATED OBJECTIVES

This opening panel raises the question whether or how the stabilization of prices was included in, or related to, the objectives stated in
the Employment Act. Superficial evidence of such a relationship or
inclusion is lacking. Nowhere in the act can the expression "price
stabilization" or "price level" or even the word "price" be found.
This is probably to be explained by the fact that Public Law 304 of
the 79th Congress was simply a revised version of the Murray fullemployment bill of 1945. Its specific objective was still stated in terms
of jobs—"useful employment opportunities for those able, willing, and
seeking to work." There were, however, three other parts of the declaration of policy (sec. 2) that demand attention.
(a) The original full-employment objective was expanded to set
"maximum purchasing power" alongside "maximum employment and
production." Obviously, purchasing power is a price-oriented concept. (&) The declaration of policy also stated that the employment
and purchasing-power objective was to be pursued "in a manner calculated to foster and promote free competitive enterprise." Freeenterprise competition takes place in the market, and its outcome is
price. Here, again, price objectives are clearly embraced in the mandate of the act even though the words "price" or "price level" are not
used, (c) Section 2 of the act further declares that the Federal Government's objective of "maximum employment, production, and purchasing power" is to be pursued "with the assistance and cooperation
of industry, agriculture, labor, and State and local governments." I
would suggest that this stipulation of assistance from and cooperation
with industrial management, labor leadership, and agricultural organi13

ECONOMIC STABILITY AND GROWTH

zations has little meaning if it does not relate in the most positive
and comprehensive manner to prices, wages, and the market process
generally.
A trend toward greater emphasis on stability as an objective of
the Employment Act, due to fear of inflationary boom leading to
deflationary recession, led fo an attempt near the close of the last
session of the Congress to write a specific price-stabilization objective
into the act. The Bush amendment declared that the previously
stated objectives of maximum employment, production, and purchasing power "must be attained, if they are to be meaningful, in an economy in which the cost of living is relatively stable. To this end,
the agencies and instrumentalities of the Federal Government must
utilize all practicable and available means to combat inflationary
pressures as they develop within the economy." In the following
section, the President was instructed to include in his Economic Report "current and foreseeable trends in the price level prevailing
in the economy and the steps, if any, which have been taken to stabilize
the cost of living and to combat inflationary pressures existing within
the economy." Finally, the amendment expanded the description of
qualifications for membership on the Council of Economic Advisers.
They were to be persons competent "to formulate and recommend
national economic policy to promote employment, production, and
purchasing power under free competitive enterprise and [the new
language] in an economy of relatively stable prices."
This proposed amendment died in committee, so the question still
stands how the objectives as stated in the original act are being or
should be interpreted.
INTERPRETATION OF THE STATED OBJECTIVES

To understand the part that prices play in the interpretation and
application of the Employment Act, it is necessary to note a broad
analytical difference among those who try to interpret usefully the
objectives of the Employment Act or, more basically, the economic
philosophies and experimentation of which it is a particular legislative
expression. This divergence is one between rival but not mutually
exclusive values held by economists and laymen. One group vigorously proclaims itself the exponent of "high-pressure economics"
and ever-full (or over-full) employment (more jobs than applicants),
with "pressure" to keep it so exerted through positive governmental
policy and action, fiscal and monetary. Over against this interpretation of the objectives of the Employment Act are the sober but by
no means complacent economists, like myself, who would not care to
have the disparaging label "low-pressure economists" pinned on them
but might call themselves exponents of safe-pressure stabilization with
vigorous growth. We place primary emphasis on such fullness of
employment as can stand on its own bottom and, thus, reflect internal
stability in the market (where government has now become a major
supply-and-demand factor), and believe there is such a thing as
inflationary overemployment, a condition in which production is at a
destabilizing maximum of inventory surplus, excess plant building,
and wage-price "leapfrogging."
To take this position is not to espouse the heresy of "general overproduction" but to stress the fact that misallocation of resources de-

ECONOMIC STABILITY AND GROWTH

rives from faulty price, wage, and profit adjustments. As such, it
is to be attacked through specific market institutions, practices, and
policies, not through the blanket devices of interest rates and tax
levels, and only in emergencies through the processes of Federal spending. The "high-pressure" group stresses growth—though they are
not unmindful that market instability might retard growth. Safepressure economists stress the functional balance of prices and incomes
as the surest means to sustained growth in jobs, in production, and
in real consumer purchasing power. Advance would be at as fast
a rate as can be sustained within a competently administered market
and fiscal process.
Now this divergence in interpretation of the policy and responsibility of the Federal Government very evidently hinges on price
issues. I have suggested above that a mandate for the price-incomeadjustment line of attack on the employment problem is clearly evident
at three points in the Employment Acts' declaration of policy. Nor
have these phases of the objectives statement been ignored in the 13
Economic Reports of the President which have thus far been submitted, nor in the studies of this committee. In President Truman's
first report (January 1947), we read:
The Congress, by setting maximum purchasing power as an objective of
national policy in the Employment Act, pointed to the importance of purchasing
power in keeping our economy fully employed and fully productive. * * * The
rise in prices that occurred in the latter half of 1946 greatly reduced the
purchasing power of the current incomes received by the large majority of
people. * * * How to effect a mutual adjustment of income and prices which
will provide purchasing power adequate to sustain maximum production in the
years ahead thus becomes a central problem for private enterprise and Government (pp. 1, 2). [Italics added.]

In this report I do not anywhere find the word inflation, though the
first of several propositions in the closing summary began:
Chief among the unfavorable factors is the marked decline in real purchasing
power of consumers, resulting from the large price increases in the second half
of last year (p. 19).

Six months later, the midyear report stated:
Price and income adjustments stand foremost in need of attention. * * * There
is need to hold the price line in the face of recent developments which revive
some fear of another upswing of inflation (p. 2).

In January 1948 we read:
A year ago I warned against the danger of advancing prices, which would
undermine our structure of national prosperity. I strongly urged businessmen
to bring prices into line with the requirements of a stabilized economy. I called
upon workers to limit their demands for wage increases to those situations where
wages were substandard or where wage increases would not necessitate higher
prices. * * * The first objective for 1948 must be to halt the inflationary trend
(PP. 3, 5).

The midyear Economic Report (July 1948) clearly linked price
objectives to the mandate of the Employment Act, saying:
The policy proclaimed in the Employment Act requires us to devise and adopt
positive measures to stop this inflation and secure relative stabilization. * * * I
realize that the anti-inflation program I have offered will impede some business
plans, will curb some profit opportunities, and may limit some wage advances.
It is of the very essence of a plan to counteract inflation that this be done. All
groups will ultimately benefit when it is done (p. 2).

ECONOMIC STABILITY AND GROWTH

In January 1949 it was noted that the general rise in prices had
reached a crest in August 1948, but the Economic Report stressed the
point that a—
rising spiral [had] created more and more maladjustments among prices, wages,,
and other incomes * * * brought higher but uneasy profits to business firms,
squeezed the family budget of workers, who in turn sought to press wages upward
as the cost of living advanced (pp. i, ii, 4).
With a slight recession in 1949, the Economic Report of 1950 keynoted disinflationary price adjustments in recent months, and, with
premature complacency, suggested that this created "the relative
stability on which firm business and consumer plans can be based."
* * * The relatively safe passage from inflation to greater stability was noaccident. Businessmen, workers, and farmers demonstrated much greater judgment and restraint than in earlier similar periods. * * * The effective teamwork between free enterprise and Government confounded the enemies of freedom who waited eagerly, during 1949, for the collapse of the American
economy. * * * If we are to continue our economic growth, the major economic
groups must all pull together—businessmen, wage earners, and farmers must
work toward the same ends. Government, in turn, must carry out the aspirations of the whole people. * * * To promote an environment in which businessmen, labor, and farmers can act most effectively to achieve steady economic
growth is a major task of the Government. * * * It must keep open the channelsof competition, promote free collective bargaining, and encourage expanded opportunities for private initiative. The Council at midyear 1949 did not recommend increases in public spending for the puipose of stimulating the economy,
and our confidence in its internal recuperative forces has thus far proved justified (pp. 1,6, 7,103).
As the Korean war renewed inflationary pressure, the President
stressed the need for "increased production of the right kind of
goods" and "some sacrifice of domestic consumption." His midyear
(1950) Economic Report urged prompt imposition of war taxes, but
stressed large dependence—
upon business polices * * * fostering production along those lines which are most
needed under changing circumstances. It depends upon price and income practices which maintain a balance between full output and buying power, so as toavoid either inflation or deflation. * * * Labor should continue and enlarge its
contribution toward increasing productivity. Wage demands of a character
which might lead to another inflationary spiral should be avoided (pp. 12-13).
Stress on price and income aspects of the employment and production problem and of the importance of private business policies and
market processes continued throughout the Truman administration.
In the midyear 1952 Economic Report the President observed :
Although the longtime rise in prices since before the start of World War I I
has not prevented the great economic progress which has been made since then,,
we would now be even better off if the price level had been even better
held. * * * A further inflation of the price level or diminution in the valueof the dollar can and should be avoided (p. 11).
He cited the recent steel strike to highlight the heavy impact of wageprice relations in a basic industry on national stability and growth.
He urged that the Congress enact—
new legislation which would permit the Government to maintain essential production, to be fair to both sides in the dispute, and to retain the maximum degree of free collective bargaining (p. 14).
This, however, was to be emergency, not permanent, legislation. As
to the controls of materials, prices, wages, and credits invoked to meet
the Korean war pressure, the administration tended to rely on the

ECONOMIC STABILITY AND GROWTH

continuance of such procedures rather than to explore institutional
changes that might curtail rather than contribute to the built-in inflationary bias of a full employment economy and would permit early
removal of controls.
The final Economic Report of the Truman administration (January
14,1953) has deservedly been much praised for its stimulating analysis
of the economics of full employment. In this analysis it gave extended attention to prices processes and price policies. In an outstanding paragraph it said:
Private enterprise, under our free system, bears the major responsibility for
full employment. This report has already set forth the basic features of that
responsibility, and how its exercise is contributing to the well-being of the
American people. The role of responsible Government, while vital, is in a sensa
supplemental (p. 18).

In amplification of this theme, the report reasoned:
Expansion cannot continue smoothly unless it is based on a sound and fair
distribution of the increasing product. Our economy is built upon mass markets.
Unless each important sector receives a workable share of the expanding output,
the expansion will come to an end because the market demand will be lacking.
Growing capacity to produce requires growing ability to buy. * * * If business
and labor plan their price and wage policies to encourage the balanced expansion
of production and consumption, of jobs and markets, then our economic growth
can be steady. * * * We must learn more about the value of individual and
group self-restraints, about the general economy and its interrelationships, and
about the private price and wage policies which may contribute most to a stable
and growing economy (pp. 16,17,20).

Turning now to the present administration, the five Economic Reports which it has submitted have consistently sounded the theme of
"reasonably full employment with a reasonably stable price level."
Our economic goal—

said President Eisenhower in his letter of transmittal of his first Eco-nomic Eeport—
is an increasing national income, shared equitably among those who contribute.
to its growth, and achieved in dollars of stable buying power. * * * Government
must use its vast power to help maintain employment and purchasing power, as,
well as to maintain reasonably stable prices (pp. iii, iv).

More specific attention was paid to the objective of free competitive
enterprise.
The role of competitive markets is as basic to the proper functioning of
economic order as the secret ballot is to our political democracy. Government
has vital responsibility in this area, immensely complicated by large aggregations
of capital under single management and large organizations of labor. Govern-,
nient must, nevertheless, remain alert to the danger of monopoly, and it must
continue to challenge through the antitrust laws any outcropping of monopoly
power. It must practice vigilance constantly to preserve and strengthen competition (p. 5).

Competition, however, was stressed primarily as a means of promoting enterprise, improving the allocation of resources, and accelerating
growth, rather than as a means of regulating prices or preventing
inflation.
The 1955 Economic Keport repeated the belief that "Government
should persist in its efforts to maintain easy entry into trade and industry, to check monopoly, and to preserve a competitive environ^
ment" (p. v.), but there was no spotting of places where monopolistic
tendencies were apparent in the economy. There was a recommenda-.

ECONOMIC STABILITY AND GROWTH

tion to "strengthen the deterrent to violation of the Sherman Antitrust Act by raising substantially the maximum fine that may be imposed under the act" (p. 50).
The 1956 Economic Report again paid its respects to the general
principle of competitive enterprise, which the Government should
strengthen "through monetary, fiscal, and housekeeping policies to
promote high and rising levels of economic activity; by helping
small- and medium-sized businesses overcome impediments to their
expansion; and by vigorous measures for preventing monopolistic
practices and combination" (p. v.). In elaborating this last point,
the report pointed to the work of the National Committee To Study
the Antitrust Laws and the "vigorous enforcing of those laws" by the
Department of Justice and the Federal Trade Commission. The
President made six recommendations for strengthening the antitrust
laws governing industrial, commercial, and banking corporations, but
made no reference to price policies within the law or to the possible
impairment of free competitive enterprise through present institutions or practices of wage making.
The 1957 Economic Report was about equally reticent. Its review of economic developments during the preceding year referred
briefly to "the advance in industrial prices * * * especially after
steel prices were raised following the strike settlement * * * the
combination of heavy demands from the investment-goods sector of
the economy, rising labor costs, and renewed advances in prices of
many raw materials resulted in price increases for a broad range of
semimanufactured materials, components, and supplies. And these
price increases became cost increases to producers of finished goods,
many of whom were also experiencing rising labor costs. * * *
Wage and salary rates advanced during the year * * * average
hourly earnings of production workers in manufacturing and building construction rose 6 percent and in retail trade 4 percent. While
the increases in wage and salary rates were only slightly greater
than those in 1955, the improvement in productivity appears to have
been substantially less. * * * Total corporate profits before taxes
fell from an annual rate of $45 billion the second half of 1955 to * * *
$4.1 billion in the third quarter of 1956 (partly because of the steel
strike)" (pp. 32-34).
In his letter of transmittal, the President said:
Government must strive to strengthen competitive markets and to facilitate
the adjustments necessary in a dynamic economy. Even more exacting are the
responsibilities of individuals and economic groups. Business managements
should formulate and carry out their plans so as to contribute to steady economic growth. They must also recognize the broad public interest in the prices
set on their products and services. Both management and labor should remove
restrictions on the operation of competitive markets * * * and reach agreements on wages and other labor benefits that are consistent with labor-productivity prospects and with the maintenance of a stable dollar" (pp. iii-iv).

There was a repetition of recommendations for further legislation
to strengthen the antitrust laws but no specific comment on wagemaking institutions or practices.
Finally, the Economic Report submitted just a few weeks ago
"acknowledges that there is an unfavorable feature in recent economic
developments. * * * Four-fifths of the increase in gross national
product in 1957 was accounted for by rising prices. There are criti-

ECONOMIC STABILITY AND GROWTH

cal questions here for business and labor, as well as Government.
Business managements must recognize that price increases that are
unwarranted by costs, or that attempt to recapture investment outlays
too quickly, not only lower the buying power of the dollar, but also
may be self-defeating by causing a restriction of markets, lower output, and a narrowing of the return on capital investment. The leadership of labor must recognize that wage increases that go beyond
overall productivity gains are inconsistent with stable prices, and that
the resumption of economic growth can be slowed by wage increases
that involve either higher prices or a further narrowing of the margin
between prices and costs. Government, for its part, must use its
powers to keep our economy stable and to encourage sound economic
growth with reasonably stable prices" (p. v).
There is in this report the familiar pledge of allegiance to the dual
objectives of the Employment Act: Economic policy must "strive to
limit fluctuations in the rate of overall economic growth to a relatively narrow range around a rising trend" (p. 3). This is followed
by an impeccable section on "Free Competitive Enterprise." Here
is the emphasis made familiar in previous Economic Reports on the
efficacy of free enterprise in guiding the allocation of productive resources, but nothing about competitive adjustment of price and income structures in a day of giant corporations and unions. The
report rightly says that "the authors of the Employment Act made
it explicit that Government * * * should foster and promote free
competitive enterprise." But does this not mean that the Congress
should shape business institutions so as positively to induce price
competition or minimize monopolistic price controls? History suggests that more is needed than the mere official pronouncement that
"policies and practices of individuals and private groups must [sicj
contribute to, not hinder, the achievment of economic growth with
reasonably stable prices."
EFFECTUATING THE OBJECTIVE OF FREE COMPETITIVE ENTERPRISE

From examination of the language of the Employment Act (in
the perspective of its legislative history) and review of its administration (in the context of contemporary economic thinking), I am led
to five conclusions:
1. Both the framers of the act and those who have sought to forward its broad purpose of sustained high-level use of the Nation's
productive resources have been quite aware that they must deal with
a complex, interrelated process of prices and incomes.
2. There has been general recognition that there are three avenues
of constructive approach to the maximum production problem: Fiscal
policy, monetary policy, and market (or private price income) policy.
Because of preoccupation with the current economic fad of aggregate
demand as the antecedent rather than the concomitant of maximum
production, there has been undue faith placed in monetary and fiscal
policy as the means of attaining the ends of national growth and
stability.
3. There is progressive disenchantment with monetary policy as a
major means of achieving the objective of the Employment Act, both
because of inherent limitations and political back-seat driving. Fis23734—58

ECONOMIC STABILITY AND GROWTH

cal policy is still recognized as a powerful agency of growth and
stability, but one subject to similar political hazards in application.
It is a basic limitation of both these types of control that they are
aimed at aggregates, or the statistical artifact of a price level, rather
than specific functional price and income relationships. They may
aggravate rather than correct the specific and local situations where
increased costs, disbursed incomes, and price realizations attain or fail
to attain so good a balance as to clear the market at full capacity
operation. To the extent that such balance is not attained (or approximated), the compensatory or offsetting task thrown on public
action (fiscal and monetary) is increased—with resultant failure or
desperate resort to authoritarian controls.
4. In lieu of such a drift toward Government control ("creeping
socialism" or whatever) we have the alternative deeply rooted in our
traditions—"free competitive enterprises." This has two aspects.
One is the "economic statesmanship" of corporation pricemaking executives, and a like concern for the welfare of the economy on the part
of wage-negotiating officials of large and strategically placed labor
unions. Appeals to this important source of business efficiency and
economic stability are consistently emphasized through the Economic
Eeports of the President from 1947 to 1958. I believe that such appeals are not fatuous. With growing economic sophistication and
clearer sense of their responsibility on the part of executives of large
industrial, commercial, and labor aggregations, we may hope to move
closer toward the self-sustained balance which both President Truman and President Eisenhower have repeatedly stated should be the
contribution of private business. They have, I believe, been soundly
advised in taking this position.
5. The greatest service this committee and the Congress can render
at this juncture is to clarify the meaning of free competitive enterprise in this day of corporation and labor union giants. With the
degree of concentration of economic power that has grown up at these
centers and the institutional structures they now have it is quite possible for the free competitive enterprise of their leaders to work
against rather than for the stabilizing of the economy in a strong
growth trend. We need to reestablish conditions of price competition instead of power competition. A full employment economy needs
flexibility of its price and income structure to displace the built-in
rigidities and ever-widening institution of "escalation," whether of
farm-price supports, union contracts, cost-plus procurement, variable
annuity insurance, Government pay scales, and even proposed fixedincome bonds.
Ten years ago the Economic Eeport of the President said:
The policy proclaimed in the Employment Act requires us to devise and adopt
positive measures to stop this inflation and secure relative stabilization.

That responsibility has not been diligently carried out in the intervening years. Instead, we have followed the easier but dangerous
course of accepting what in 19511 called "inflation as a way of life."
The present administration, as noted above, has instituted two inquiries into the field of monopoly, and at least three committees of the
Congress have been conducting hearings on various phases of the
matter. But there has been no comprehensive attempt to relate these
specific inquiries to the central purpose of the Employment Act. They

ECONOMIC STABILITY AND GROWTH

have been lacking in comprehensiveness also in that none has faced
forthrightly the inseparability of price, wage, and profit structures
in a free-market economy. In my opening testimony before the
Kefauver Subcommittee on Antitrust and Monopoly, I was moved to
suggest that "the processes of pricemaking and of wage making are
so intertwined in the modern industrial world that neither can be
effectively analyzed in isolation from the other. I believe also that
the phenomenon known as monopolistic competition or as administered pricing manifests itself in essentially similar ways and with
essentially similar consequences in the two cases."
There are places in the program of the present hearings at which it
would seem that the pricing of labor and its relation to full employment would naturally enter the discussion, but the word "wages" is
nowhere used. I anticipate that this gap will in some measure be
filled by several of the contributors to the compendium and that the
panel discussions as they unfold will not only broaden and deepen our
understanding of the nature of free competitive enterprise as an objective of the Employment Act in an age of large-scale technology but
will also point to specific measures for buttressing the old faith and
giving it modern implementation.

PRICE-LEVEL STABILITY AND EMPLOYMENT ACT
OBJECTIVES
Joseph Aschheim, The Johns Hopkins University
INTRODUCTION

The postwar period has been marked by a curious asymmetry between widespread economic opinion on the one hand, and official
economic policy objectives on the other. Few issues have attracted
wider or more recurrent attention over the last dozen years than the
problem of full employment versus price-level stability. Yet our
basic guide to national economic policy, i. e.? the United States Employment Act of 1946, has remained devoid of any reference to pricelevel stability or to some other desired price-level behavior.
One does not have to look far afield for reasons underlying this
chasm between public discussion and congressional legislation. The
Employment Act came into existence at a time when recollections of
the severest depression in the history of the American economy were
all too vivid. No less important, the Employment Act was born under the bad omen of the famous, by now infamous, forecasts of substantial unemployment to follow the war. Thus the act bears the
unmistakable imprint of the predominant economic views in the context of which it was formulated. The Joint Economic Committee's
repeated concern with the relationship of prices to economic stability
and growth is but another illustration of the extent to which actual
developments over the last 12 years have belied the gloomy expectations reflected in the Employment Act.
Under certain conditions, the continued exclusion of a price-stability
objective from the Employment Act could be quite appropriate.
These conditions include (a) the presence of empirical evidence pointing to a serious conflict between price-level stability and present Employment Act objectives, and (5) the existence of a cogent basis for
the claim that it would be more desirable to forego price-level stability than drastically to compromise present Employment Act objectives. Accordingly, we shall first consider the extent of likely conflict
between price-level stability and the goals of "maximum employment,
production, and purchasing power." In the light of this discussion,
we shall subsequently appraise the desirability of adding to the Employment Act the objective of price-level stability.
ARE FULL EMPLOYMENT AND PRICE-LEVEL STABILITY COMPATIBLE?

There are two facets to the question of the extent of conflict between price-level stability and maximum or full employment in the
contemporary American economy: (1) The volume of unemployment
which would be required for price-level stability; and (2) the degree
of inflationary pressure which a state of maximum or full employment
will produce. Let us take up these two subjects in turn.

ECONOMIC STABILITY AND GROWTH

A. Volvme of unemployment
The postwar notion of a Hobson's choice between price-level stability and full employment is no more an accident than the fact that
the Employment Act itself reflects no trace of this notion. With
the money wage level rising faster than output per man-hour, with
the rapidity of adaptation of product prices to higher costs, with the
growing importance of wage leadership in the industrial sector of
the economy, and with the trade-union movement comprising approximately one-fourth of the civilian labor force, it is hardly surprising
to encounter the view that the market structure of the American
economy may involve substantial unemployment as a prerequisite for
price-level stability. Thus, only a few years separate the pessimistic
forecasts of mass unemployment in the immediate postwar period
from the almost equally disturbing predictions that future price-level
stability
would require from 10 to 15 percent of the civilian labor
force 1 to be unemployed. Have the latter estimates proven to be
more accurate than the former? Does postwar experience to date
bear out the expectation that price-level stability will necessitate heavy
unemployment ?
While the postwar economy has thus far been spared the test of
10 to 15 percent unemployed, we do not entirely lack other relevant
evidence as to whether such mass unemployment is prerequisite for
price-level stability. The experience of the first two postwar recessions, as indicated by the data in table I, is quite suggestive.
During 1949 unemployment averaged 5.5 percent of the civilian
labor force. Between December 1948 and December 1949, average
hourly earnings of production workers in manufacturing industries
rose by 0.6 percent when taken inclusive of overtime or by 0.7 percent
when taken exclusive of overtime. Between2 the same 2 dates, the
Consumer Price Index declined by 1.9 percent and the wholesale price
index fell by 5.0 percent. Manifestly, however one may characterize
the behavior of money wage rates during that period, they did not
prevent the price level from falling.
TABLE I.—Price and wage levels for selected months and years

December of year—

1948
1949
1953
1954

Consumer
Wholesale
price index
price index
(1947-49=100) (1947-49=100)

103.0
101.0
114.9
114.3

104.4
99.2
110.1
109.5

Average hourly earnings of
production workers in
manufacturing industries
Gross

$1,400
1.408
1.80
1.83

Excluding
overtime
$1,358
1.368
1.74
1.77

Source: for consumer price index: Survey of Current Business, April 1953, p. 23, and Economic Indicators,
March 1955, p . 3: for wholesale price index: Survey of Current Business, March 1952, p. 22, and Economic
Indicators, March 1955, p. 4; for hourly earnings: Monthly Labor Review, February 1950, p . 234, loc. cit.,
February 1951, p . 241, Employment and Earnings, May 1954, p . 38, and loc. cit., May 1955, p . 38.

Turning to the second postwar recession, during 1954 unemployment averaged 5.0 percent of the civilian labor force. Between
*Cf. A. P. Lerner, Economics of Employment (New York: McGraw-Hill, 1951), pp. 193194 ; and A. G. Hart, Money, Debt, and Economic Activity, 2d edition (New York : PrenticeHall, 1953), pp. 247-248.
a It is noteworthy that not only the food component of the Consumer Price Index declined
but also the apparel and the house furnishings components.

ECONOMIC STABILITY AND GROWTH

December 1953 and December 1954, average hourly earnings of production workers in manufacturing industries, whether taken inclusive
or exclusive of overtime, rose by 1.7 percent. Between the same two
dates both the consumer3 price index and the wholesale price index
declined by 0.5 percent. Here, too, however one may depict it, the
behavior of money wage rates did not prevent a decline in the price
level.
It seems reasonable to conclude that the minimal volume of unemployment necessary for price level stability is much closer to 5 percent
of the civilian labor force than to 10 percent, let alone anything
higher. In any event, there exists no empirical basis to date for the
contention that unemployment of massive proportions is a prerequisite for price level stability in the postwar market structure of the
American economy. This contention is an outcome of the failure
to distinguish between a state of high-level employment amid an excess of aggregate monetary demand, such as characterized the years
1945-48, and a state of high-level employment in the absence of an
excess of demand. Since the demand for productive factors derives
from the demand for output, an excess of aggregate monetary demand
is likely to imply an excess of demand for labor. And, indeed, an
excess of demand for labor was imminent in the general inflationary
conditions of 1945-48. To infer from the behavior of prices and
money wages in that context of inflationary demand to a future state
of high-level employment in the absence of an excess of aggregate
monetary demand was a procedure bound to lead to serious exaggerations. For under general inflationary-demand conditions, there
would be grounds for expecting the money wage level to rise in excess
of the average increase in output even in largely unorganized labor
markets; the substantial increase in money wage rates immediately
following World War I is a case in point. Thus to expect labor
organizations to restrict their demands for higher money wage rates
to the average increase in output per manhour when the pursuit
of such a course would imply constant or even reduced real wage
rates is to engage in economic fantasies. Amid inflationary demand
conditions, the bargaining power of wage earners is at its maximum.
In such a context, employers have both the incentive and the ability
to translate higher labor costs into higher product prices. In consequence, no exhortations to either labor or management or both to moderate their wage-price policies will stem the tide of inflation in the
face of a general excess of aggregate monetary demand. At best,
such exhortations are tantamount to locking the barn door after the
horse has been stolen; at worst, they deflect from the obligation of
government to avert or counteract inflationary demand conditions
by means of monetary-fiscal policy.
B. Extent of inflation
We may now turn to the second facet of the problem of pricelevel stability versus full employment, viz, the degree of inflationary
pressure likely to occur under full employment conditions. Obviously,
the prediction that the inflationary outcome implicit in
a full employment economy may go so far as to disrupt capitalism 4 must involve a
3
Again, there was a decline not only in the food component of the consumer price index
but also in the apparel, the transportation, the reading and recreation, and the other goods
and
services components.
4
Cf. C. E. Lindblom, Union and Capitalism (New Haven: Yale University Press, 1949),
p. 139.

ECONOMIC STABILITY AND GROWTH

high estimate of the inflationary bias envisioned. Is such an ominous
estimate corroborated by the postwar record ?
The answer to this question depends not only upon a scrutiny of
empirical data but also upon the meaning attributed to the concept
of full employment. We shall briefly consider only three out of a
spate of possible interpretations of "full employment."
First, let us take up the well-known definition provided by Lord
Beveridge that full employment is "having always more vacant jobb
than unemployed men." Clearly, this definition connotes an excess
of demand for labor. And since the demand for labor is derived from
the demand for output, the Beveridge definition implies an excess of
aggregate monetary demand. Now, as previously noted, we have no
reason to doubt that inflationary demand conditions will involve a
rise of the money-wage level in excess of the average increase in productivity. Thus, if full employment is taken to mean an excess of
demand for labor, full employment and price-level stability are by
definition incompatible. And the degre of upward bias in the price
level that is implied by such a definition of full employment will vary
with the magnitude of the inflationary pressures extant. For example, under the immediate postwar conditions of pent-up demand
and excess liquidity, the 1947 monthly average of the Consumers
Price Index rose by 14.5 percent above the monthly average for the
previous year. In contrast, amid the much milder inflationary demand conditions of more recent years, the 1956 monthly average of the
Consumer Price Index was 1.5 percent above the preceding year's
level.
Strict adherence to the Beveridge definition of full employment is
coterminous with the doctrine of "full employment at any price" or
more accurately, with "full employment at any price level." The
proponents of direct controls will be quick to suggest that this is not
necessarily so; with wage and price controls, they will aver, full employment and price-level stability could be rendered compatible. We
cannot digress here to discuss the dubious character of this assertion.
Suffice it to note that the Beveridge definition of full employment
implies a chronic state of open or repressed inflation.
Next we may turn to the definition according to which full employment is the maximum level of employment consistent with price-level
stability. In this instance, the constraint of a stable price level is
built into the definition of full employment; full employment and
price-level stability are by definition mutually consistent.
The finding that the volume of unemployment necessary for pricelevel stability is of the order of 5 percent of the civilian labor force
may suggest to some that the foregoing definition of full employment
is socially acceptable. On the other hand, it should not be overlooked that 1949 and 1954 were years of recession, as reflected by the
data in table II. Measured in current dollars, 1949 gross national
product remained at the level of 1948, and 1954 gross national product
was 0.7 percent below the level of 1953; measured in constant dollars
the decline for 1949 was 1.0 percent, and for 1954, 1.5 percent. We
shall presently revert to the significance of these declines.

ECONOMIC STABILITY AND GROWTH

TABLE II.—Gross national product and private domestic investment for selected
years
Gross naGross national prodtional prod- uct (billions
uct (millions of 1947 dolof dollars)
lars)

Year

1948
1949
1953
1954

257,325
257,301
3f3,218
360,654

Gross private domestic investment (millions of
dollars)

Capital consumption
allowances
(millions of
dollars)

41,176
32,549
50,325
48,032

16,494
18,431
26,486
28,760

243.9
241.5
305.3
300.8

Net private
domestic
investment
(millions of
dollars)
24,682
14,118
23,839
19,272

Source: Survey of Current Business, July 1956, pp. 10-13. 24-25.

Finally, full employment may be defined as the highest possible
level of employment without an excess of aggregate monetary demand.
In principle, adherence to this definition would be consistent with
averting or counteracting inflationary pressures from the demand side
whilst acquiescing in inflationary pressures from the cost side. In
practice, this definition would be favored by those who would tolerate
a gradually rising price level on the assumption that the upward
movement can be kept gradual indefinitely.
Thus our brief survey of some possible interpretations of full employment indicates that even a rough estimate of the extent of inflationary bias inherent in a condition of full employment will partly
depend on what this condition is taken to mean. At the same time
it must be noted that the difference in the employment levels implied
by the various definitions is, in reality, likely to be relatively small.
We must, in any case, expect seasonal unemployment plus "turnover"
unemployment (persons moving, for whatever reason, from one job to
another) to amount to at least 2.5 percent in the peacetime American
economy. Consequently, even the Beveridge definition of full employment encompasses a 2.5-percent unemployment margin. At the
other end, the definition of full employment which incorporates the
price-stability constraint does not appear likely to entail considerably
more than a 5-percent unemployment margin. Even the latter figure
could hardly be viewed as awesome by past historical standards. Yet,
it does imply some sacrifice of gross national product, and, more important, a relatively larger sacrifice of private investment. The
slight decline in gross national product during the 1949 and 1954 recessions has already been mentioned. Much more pronounced was the
drop in net private domestic investment (as shown in table II) : in
1949 it was a decline of 42.8 percent from the 1948 level, and in 1954
it was a decrease of 19.2 percent from the 1953 level.
Private investment is crucial to the expansion of the American
economy's productive capacity. In an international context in which
we should hardly want to compromise our economy's growth potential,
acceptance of a 5-percent unemployment margin for the sake of pricelevel stability is, to say the least, a questionable approach. This is
not to suggest that it would be desirable to insert a maximum-tolerableunemployment figure into the Employment Act. The Employment
Act sets forth general targets rather than detailed commitments; it
conveys broad directions rather than specific limits. In implementation, at any rate, "maximum employment, production, and purchasing
power" have been interpreted as referring to a zone rather than to a

ECONOMIC STABILITY AND GROWTH

level. And this is as it should be in an economy in which forecasting
tools are still imperfect and in which some flexibility in the use of
economic-policy instruments may be desired at all times. I am suggesting, however, that it would be an error to follow a path to pricelevel stability that would also lead to chronic underutilization of the
economy's potential for expansion and growth.
CAN WE AFFORD GOVERNMENT ACQUIESCENCE I N MILD INFLATION?

Where, then, are we left in the matter of price-level stability as a
possible additional objective to be specified in the Employment Act?
Are we to conclude that a gradually rising price level must be accepted as a deleterious but inevitable byproduct of a state of highlevel employment and uninterrupted economic growth? Is there no
practicable alternative to creeping or galloping inflation in the Age
of Outer Space?
A. Some common observations
While high levels of production and employment have obtained
since the end of World War I I , the purchasing power of the consumer's dollar has undergone a substantial decline. These concurrent
developments have given rise to the view that the Federal Government
is more reluctant to check inflation than to counteract the recession.
Admittedly, the possible imminence of a recession evokes calls for
governmental action from all sides. By contrast, the prospect of a
rise in prices is a much less potent factor of organized demands for
governmental intervention. Application of restrictive monetary and
fiscal measures is likely to elicit vigorous opposition from various
groups, whereas the pursuit of expansionary policies is agreeable to
most. In general, precipitating a recession appears a more hazardous
course than erring on the side of inflation. Furthermore, the existence of a large and widely distributed national debt and the necessity
of recurrent refunding operations induce the Treasury to favor low
interest rates. And the Federal Reserve, even with the "accord," can
hardly be oblivious to the implications of a significantly restrictive
monetary policy for Government and private security markets. The
upshot is that the Federal Government is commonly regarded as
inclined to acquiesce in an inflationary trend over the foreseeable
future just as it apparently has over the recent past.
B. Adequacy of Government services
I submit that the conclusion that a gradually rising price level is
the Government's "easy way out" is basically defective. I do so not
only on the grounds that it may eventually prove impossible to prevent a gradual rise in the price level from becoming rapid or that a
persistent erosion of the purchasing power of the consumer's dollar
has inequitable redistributive effects. Rather, I suggest that from
the viewpoint of government—National, State, and local—the most
adverse effect of a gradually rising price level is the concomitant
deterioration in the quality of Government services.
In times of high-level employment—when the Government labors
under severe budgetary constraints, under powerful pressures against
raising already high tax levels, and under pronounced unwillingness
among policymakers to add further to existing inflationary pressures—a gradually rising price level ceases to be "the easy way out"

ECONOMIC STABILITY AND GROWTH

and becomes "the dangerous way out." Being unable to raise salaries
in competition with the private sector without substantial timelags,
Government agencies are, in effect, compelled to lower their standards
in hiring additional and replacement personnel. In like mannerr
being unable to obtain increased appropriations for expansion of
essential physical facilities in the face of rising costs, Government
must of ttimes resort to elimination or curtailment of such expansion
programs, in this instance, too, at the expense of the quality of the
services being rendered. Thus, the stickiness of salary levels of
teachers, scientists, military personnel, Foreign Service personnel,
statisticians, accountants, and other Government employees, as well as
the sluggishness in the upward adjustment of expenditures for school
construction, new hospitals, improved civil-defense facilities, additional public libraries, more adequate statistical information, and a
variety of other projects result in reduced quality and curtailed efficiency of vitally important services which are wholly or partly the
responsibility of Government. In a period of acute international
challenge, no person or institution could have a greater stake in preserving price-level stability than Government itself.
The time has come to remove the existing asymmetry between the
Employment Act and the increasing awareness of the hazards of inflation by incorporating into the act the additional objective of pricelevel stability. This change will constitute explicit affirmation of the
Federal Government's heightened concern for the maintenance of
relatively stable price level amid the mounting needs for expansion of
Government activities in several directions.
C. Some possible objections
Various objections may be raised against the inclusion of a pricestability goal in the Employment Act. First, it has been argued that
the objective of price-level stability is, in any case, implicit in the
present formulation of the act, which includes mention of "essential
considerations of national policy" and "general welfare." But this
argument is rather specious; consistent adherence to it would obviate
the need for mention of the already stated objectives of "maximum
employment, production, and purchasing power," since these may be
considered as no less implicit in "general welfare" than price-level
stability. Furthermore, to state an important implication explicitly
would contribute to the clarity of the act by removing reasons for
doubting whether the particular implication was intended or not.
Second, and with more justification, it may be contended that the
concept of price-level stability is too vague and, therefore, open to
interpretations too different to constitute a meaningful objective for
national economic policy. It is undeniable that different types of
price indexes exist and that their behavior is far from uniform. Thus,
the wholesale price index is more sensitive to changes in business conditions than the consumer price index, whilst the latter is closer to an
average measure of the economy's final products. Moreover, there
are times of divergent movements of the two indexes. In addition, the
description "relative" or "reasonable" price-level stability would not
indicate the precise extent of upward or downward deviation of the
price level that is to be deemed tolerable. In response to the foregoing
points, it should be noted that a concept such as "maximum employment" is also subject to different interpretations and yet has not proven
to be a source of much confusion in policy formulation or execution.

ECONOMIC STABILITY AND GROWTH

Indeed, explicit reference to a relatively stable price level would elucidate the meaning of the act by specifying that the stated objectives of
"maximum employment, production and purchasing power" are not
to be interpreted as condoning a chronic inflation.
Third, it may be held that the Federal Government should not assume responsibility for reasonable price-level stability when its ability
to achieve this objective may depend on the activities of private groups
beyond its direct control in a free-enterprise system. This view is
based on the assumption that while Government can control a demandinduced inflation, it lacks the capacity for coping with a cost-induced
inflation. However, this assumption is erroneous. An important prerequisite for cost-induced inflation is a monetary policy which validates
upward price-level pressure from the cost side by permitting an expansion of the reserve base of the economy's money supply. The fulfillment of this prerequisite is not an act of God, but the expression of
discretionary decision making by the monetary authorities. It follows that the occurrence of a cost-induced inflation cannot properly
be viewed as a development with respect to which Government is
without influence in a free-enterprise economy.
Finally, it may be objected that inserting a price-stability goal into
the Employment Act would create a dilemma as to which of the
goals—maximum employment or price stability—is to be given preference in case it should be impossible to achieve both at the same time.
However, as we have pointed out, there is no basis in postwar experience to date for assuming the existence of a serious conflict between
full employment and price-level stability, unless full employment is
defined as coterminous with an excess of aggregate monetary demand.
Furthermore, to set aside a desirable policy objective only because its
attainment currently appears as a difficult challenge is a counsel of
despair. The present absence of a price-stability goal from the Employment Act is a gap attributable to the historical accident of erroneous forecasting toward the end of World War I I . Whether this
gap would be worth perpetuating under normal peacetime conditions
is a question which need not be settled here and now. Confronting us
are not normal peacetime conditions, but the exigencies of intense
international competition in the scientific, technological, military, and
economic fields. Meeting this competition is likely to require considerable increase in the extent and quality of Government services in
the years ahead. Under such circumstances, the addition of a pricestability objective to the Employment Act is both timely and imperative.
CONCLUDING REMARKS

(1) We have examined the extent of conflict between price-level
stability and present Employment Act objectives in the light of postwar evidence. We have found that price-level stability appears unlikely to require unemployment in excess of 5 percent of the civilian
labor force. While much more moderate than some other estimates,
a 5-percent unemployment margin is apt to entail a significant sacrifice in terms of net private domestic investment.
(2) We have suggested that perhaps the most serious drawback of
chronic gradual inflation is a concomitant deterioration in the quality
of Government services. Amid the need for improvement and expan-

ECONOMIC STABILITY AND GROWTH

sion of various governmental activities, future Government acquiescence in a slowly rising price level is undesirable.
(3) In view of the foregoing, we have recommended that the objective of a relatively stable price level be incorporated into the Employment Act. Failure to make this change in the act would be tantamount to impicit admission that the Federal Government is prepared to permit continual, if creeping, inflation.

HOW IMPORTANT IS PRICE STABILITY IN STABLE
ECONOMIC GROWTH?
G. L. Bach, Carnegie Institute of Technology
My task is to examine the impact of inflation on the functioning
of a private enterprise economy, and to consider the significance of
these findings for governmental policy aimed at promoting stable
economic growth. In particular, should the objective of maintaining
a roughly stable price level be specifically included in the Employment Act of 1946 ?
In the first major section below, I shall analyze the impact of
moderate (or "creeping") inflation on an economy like ours, drawing
both on recent experience and on economic theory. Then I shall
consider specifically the question of how important the goal of price
level stabilization should be, on the basis of these findings, within
the broad framework of our attempt to maintain a pattern of stable
economic growth.
Briefly, my conclusions are these. The effects of moderate inflation on a private enterprise economy, while much less disruptive and
alarming than often claimed, are appreciable and inequitable. Moreover, such inflation provides few, if any, of the stimulative benefits
often claimed for it. Persistent inflation in the United States (outside of war periods) will probably occur only as the result of combined income pressures of major economic groups and resulting
expansive monetary-fiscal policy by the Government. The resulting
inflation would probably nave appreciable disadvantages with few
advantages. It is, therefore, important for the Government to emphasize specifically that price level stability is one important goal
among the several specified by the Employment Act tor economic
policy directed toward maintaining stable economic growth in a free
society—and to recognize the correlative role of monetary stability
in its policy behavior.
This section analyzes the effects of moderate inflation on a modern
private enterprise economy, like the United States. By inflation I
shall mean simply a rise in the commodity price level, or (what is the
same thing) a fall1 in the purchasing power of the monetary unit
over commodities. This simple definition considers any rise in the
price index used as inflation; a little price rise is a little inflation,
and a big price rise is big inflation. It includes price rises when less
than full employment exists, because this may be part of the situation
we wish to consider. It does not look behind the inflation to see its
1
The Bureau of Labor Statistics indexes of consumer prices or wholesale prices can be
used to measure changes in the price level, though in principle a broader based commodity
price index would be better.
33

ECONOMIC STABILITY AND GROWTH

cause, though I shall argue later that some causes are much more
likely than others.
My analysis covers only relatively moderate ("creeping") inflation,
comparable to that experienced in the United States over the last
two decades. It specifically does not include massive runaway inflations like those suffered in central Europe and Asia after World Wars
I and II.
The economfc effects of inflation may be conveniently analyzed
here by asking what the effects of inflation are on (a) society's total
real output (real gross national product); (i) the distribution of
that real output among economic groups; and (c) the distribution
of ownership of society's wealth. My analysis, considering only
moderate inflation, suggests tentatively that:
1. There is little evidence that relatively mild inflation reduces the
current real output of society, though relatively sharp speculative
inflation probably does help induce an ensuing price collapse and
recession.
2. Neither is there much evidence that relatively mild inflation
significantly stimulates current real output, although strong total demand (of which rising prices may be one side effect) helps to stimulate output, especially in periods of substantial underemployment.
3. There is no clear evidence that moderate inflation either increases
or decreases the rate of economic growth, although again strong total
demand may speed economic growth in an economy that would otherwise be underemployed.
4. The impact of relatively mild inflation in redistributing current
income among major economic groups is apparently less than is often
claimed. To a considerable extent, major economic groups seem to
have adjusted their effective income claims upward at roughly comparable rates in the last two decades of inflation, although the share
of wages and salaries in the national income has increased substantially. The income shares of "passive" economic groups who do not
sell directly in the market (retired people, employees of government
and eleemosynary institutions, and so on) apparently lost most.
Within the major economic groups, individuals and subgroups were
affected very diversely.
5. Inflation's effect in redistributing control over wealth is substantial over the past two decades. More than half a trillion dollars
of purchasing power of creditors was wiped out by inflation.
6. In transferring purchasing power from net creditors to net
debtors, inflation in the modern American economy transfers purchasing power mainly from the "household" sector (which is a heavy
net creditor in spite of its substantial debts) to governments (which
are heavy net debtors). But governments are only intermediate
organizations. When they gain as debtors, part of this gain accrues
to taxpayers, but much of it (so far as the Federal Government is
concerned) is in effect dispersed throughout the economy to all
spenders, who gain at the expense of savers in government securities
and currency.
7. Nearly all major groups of households are net creditors, and
they therefore suffer on this account from inflation. Only very poor
families and young people just establishing households (25-34 age
group) are net debtors on balance. Conversely, older families and

ECONOMIC STABILITY AND GROWTH

high income, wealthy families are heavy net creditors, especially susceptible to loss from inflation.
8. Businesses (corporate plus unincorporated) have seen their relative share of the national income decline slightly under moderate
inflation, on the basis of reported profits. In spite of corporations'
position as moderate net debtors on balance, rapidly rising wage and
other costs have offset, or more than offset, this factor. Moreover,
real equity of stockholders in corporations has grown less than reported figures indicate; since depreciation and cost of goods sold are
generally underreported in inflation periods, reported profits are
larger than they would be if calculated on a replacement cost basis.
The evidence for these tentative conclusions is drawn primarily
from the experience of the American economy of the past two decades,
supplemented by experience abroad and in earlier periods. These historical observations have been supplemented by economic theory, since
information on the past never provides a really satisfactory basis for
prediction of the future. We can never be sure in the complex interplay
of forces just what observed effects in the past were actually caused
by inflation; even though some changes occurred with or just following
inflation, we cannot know that other third forces were not the causes.
Moreover, even if we could identify cause and effect precisely in the
past, we can never be sure that the future will be like the past. The
following paragraphs indicate generally
the kinds of evidence on
which the various conclusions rest.2
1, Does inflation reduce current real output?
During the past two decades of inflation in America, total real
output has risen persistently. This has also been generally true in
earlier periods here, and abroad. And there is little a priori reason
to suppose that moderate inflation reduces the size of current national
output.
The common belief that inflation disrupts the economy so as to
reduce total output traces back in America largely to the massive
hyperinflations of central Europe following World War I, when
currency became worthless and the diversion of energy from normal
productive work to speculation and rapid expending of funds became
a vast drag on the production of real goods and services. Even in
milder inflations, it is often argued that erratically rising prices
seriously disrupt economic planning and managerial decisionmaking
in particular. There is no convicincing evidence on the importance
of this effect in the American economy. Since the economy has grown
rapidly and relatively stably over the past two decades of intermittent inflation, the disruptive effect on managerial decisionmaking can
hardly have been overwhelming.
The fact that total real output has increased about fourfold in
the past 20 years while prices doubled does not, of course, prove
that inflation has not exerted a downward pressure on total output
that was persistently overcome by other expansive forces. But the
facts of history do indicate that if this was the case, the output-depressive effect of inflation was a relatively weak one.
2
A more complete analysis, underlying this summary, is presented in G. L. Bach, Inflation • A Study in Economics, Ethics, and Politics (Brown University Press, 1958), especially ch. I. Some earlier data and conclusions, on which the present findings partially
rest were presented to this committee, The Impact of Moderate Inflation on Incomes and
Assets of Economic Groups, in Federal Tax Policy for Economic Growth and Stability
(Joint Economic Committee, 84th Cong., 1st sess.).
23754—58
4

ECONOMIC STABILITY AND GROWTH

But a short-run depressive effect may result when sharp speculation-based inflation contributes to an ensuing price collapse and recession in employment and output. This can happen. The postWorld War I price inflation and collapse of 1920 is a clear case in
point. But even quite rapid persistent inflations have continued over
long periods without ensuing collapse and depression. This has, of
course, been true in the United States for the past two decades. It
has been true in Brazil, for example, which has had substantially continuous inflation averaging over 10 percent per annum for the past
two decades while total real output has grown persistently. In the
United Kingdom and most of western Europe, more or less continuous
inflation over the same period has been paralleled by a generally persistent growth in real national output. It is clear that inflation does
not necessarily presage collapse and unemployment just around the
corner—certainly not around a very nearby corner—unless the inflation reaches a rapid rate.
2. Does inflation increase current real output?
Moderately rising prices have generally been accompanied by rising total output. On the other hand, total output has also risen in
periods of stable prices—for example, the period of the 1920's in the
United States. Historical evidence alone, therefore, does not indicate satisfactorily whether inflation increases current real output.
There is little doubt, on theoretical and empirical grounds, that
rising total demand (spending) exerts an expansive force on total
output, especially in periods of widespread unemployment of resources.
Eising prices may be a side effect of this situation. Thus it is difficult
to dissociate what is due to increased total demand and what to rising
prices per se. The main arguments that inflation increases current
output have been: (a) that inflation induces lagging income groups
to work harder and longer; (&) that by pushing up prices faster than
costs (especially wages), it increases profits and thereby stimulates
investment and output; (c) by stimulating buying and output now,
expectation of continued inflation puts a premium on early purchases;
and (d) by easing the transfer of resources from declining to expanding industries, inflation helps increase total output. It is agreed that
inflation can increase current output greatly only when substantial
unemployment exists; only (a) is apt to be important in periods of
substantially full employment.
(a) Casual observation turns up numerous cases where inflation
has driven lagging income groups to work more and harder in order
to protect their real incomes—retired men driven back to part-time
work, wives of college professors working to supplement their husbands' lagging salaries, schoolteachers driving taxis or working in
industry during summer vacations—but it is doubtful that this effect
is a major one in creeping inflation. Two factors are critical: How
far behind prices do incomes lag, and how important are the lagging
income groups in the economy? Recent evidence suggests that in
America the major income groups have generally adjusted their income claims upward roughly in proportion to rising prices so money
incomes have not lagged far behind prices for most groups. Moreover,
those with the most significantly lagging incomes—retired persons
and employees of governments and eleemosynary institutions—are
neither large in the total potential productive power of the economy

ECONOMIC STABILITY AND GROWTH

nor in a position readily to increase greatly the total amount of work
in the national output. The labor force has shown no significant
tendency to grow during inflation as a percentage of the total population of labor force age, except during the war period when factors
other than inflation primarily explain this change. And while the
proportion of women holding jobs has risen steadily over the past
quarter century, outside the war period there is no significant relationship between the rate of increase and the rate of inflation On the other
hand, the recent increase in "moonlighting"—holding a second job—
may be partially a result of inflation.
(b) The main argument that inflation stimulates current output
has been that it increases profits as wages (and other costs) lag behind
rising prices, and this in turn induces more investment and output.
This wage lag has indeed apparently played a major role in many
past inflations. But, as the evidence presented below clearly indicates,
wages have not lagged behind prices in the American inflation of the
past two decades. During the demand-pull inflation spurts of 194648 and 1950-51, profits temporarily rose faster than wages, but this
situation was reversed immediately thereafter. And in the 1955-58
surge of prices, the wage share grew markedly relative to profits.
Thus overall wage costs have risen somewhat more rapidly than
selling prices with the result that profits have been squeezed. Indeed,
wages throughout the western industrialized world seem to be increasingly mobile upward, in many instances linked to rising prices
through built-in escalator clauses. Perhaps this situation will
change—and clearly a wage lag would be more likely in an underemployed economy—but potent economic and political pressures suggest that the wage lag is likely to be gone for a long time to come.
Some other costs may lag in inflation, even though wages do not.
Interest charges, rents, many salaries, and other costs are temporarily
fixed in dollar terms as selling prices of products rise. But these lags
can easily be overcome by only a modest wage lead. A special argument is advanced on the lag of interest costs, that this stimulates borrowing for real investment because the borrowed funds can be paid
back in cheaper dollars. This sounds reasonable, but few corporate
officials report this as a major consideration in capital-goods planning.
Inflation does lead to substantial overstatement of profits under prevailing accounting practices, because depreciation and inventory replacement costs are understated, and this overstatement of profits may
induce businesses to invest and produce more than they otherwise
would. Partial estimates suggest that this understatement of replacement costs may have approached one-third
of corporate reported
profits during the decade of the 1940's.3 While most businessmen
surely recognize the phantom nature of part of their profits in inflation, it may be that large accounting profits stimulate them to increase
output and investment spending beyond that which would be induced
by the "real" profit figures. If this force is important, it is surely
more so for investment than for total current output, however.
8
See Ralph C. Jones, Price Level Changes and Financial Statements: Case Studies of
Four Companies, and Effects of Price Level Changes on Business Income, Capital, and
Taxes (American Accounting Association, 1955 and 1956, respectively) ; and George Terborgh, Corporate Profits in the Decade 1947-56 (Washington, Machinery and Allied Products Institute, 1957).

ECONOMIC STABILITY AND GROWTH

(c) Expectation of continuing inflation may lead to increased current real output, especially in an underemployment situation, by stimulating buying ahead for inventories and for speculative resale. But
people cannot pile up inventories indefinitely on speculation that
prices tomorrow will be higher than today. Except as a "shot in the
arm," this can hardly be a major effect.
(d) Rising prices may make it easier for resources to be shifted
away from declining into growing industries, thereby helping to expand output. Without inflation, prices in declining
industries may
need to fall to force resources out, whereas wTith inflation resources
can be bid into growing industries by higher wages and prices. Inflation may thus deserve some credit as a social lubricant. But the historical evidence suggests mainly that resources shift easily when times
are prosperous and badly when depression prevails, regardless of
whether prices are rising or not. Agriculture is the major historical
case in point.
Overall, there is little real evidence that moderate inflation increases
current real output, though it is hard to dissociate the stimulative
effect of strong total demand from the effect of rising prices per se.
Certainly rising total demand may stimulate output and employment
in periods of underemployment, even though rising prices have little
contribution to make. In periods of substantially full employmenty
neither seems likely to increase current total output much, and the
argument that inflation is a necessary evil to increase output is
weakened accordingly.
3. Does inflation stimulate economic growth?
Historians have argued that over the centuries inflation has generally shifted income from the poor to the rich and from workers to
businessmen. This, they argue, has increased the volume of
saving
and investment and speeded the rate of economic growth.4 While
much evidence seems to support this view, the case is by no means
clear. Whatever the lesson of history, more recent experience throws
grave doubt on this hypothesis for, as will be shown below, at least in
the recent, American inflation the share of profits has not increased
relative to wages, nor has the share of the rich increased. The contrary has been true. Thus, the argument that inflation induces more
saving and investment via this channel cannot be accepted if recent
experience predicts the future.
More recently, many economists have argued precisely the contrary—that inflation discourages savings and thereby retards capital
accumulation, because inflation erodes the value of accumulated savings and encourages spending on current consumption. There is little
doubt that this effect prevails in very rapid, runaway inflation. But
for more moderate inflation, the case is not convincing. In almost all
the post-World War I I inflations of the Western World, capital
accumulation has proceeded rapidly. Inflation obviously militates
against saving in most forms, but the motives for saving are many and
mixed, and modern society provides some effective saving-investment
channels to escape the erosion of inflation, at least for sophisticated
savers.
* See Earl Hamilton, "Prices as a Factor in Business Growth," Journal of Economic History, Fall 1952. A counter argument is presented by David Felix, "Profit Inflation an<J
Industrial Growth," Quarterly Journal of Economics, August 1956.

ECONOMIC STABILITY AND GROWTH

Conclusion: A Scotch verdict. It is not clear that inflation of
modest proportions either increases or decreases substantially the rate
of capital accumulation.
A concluding note on the relationship between inflation, production,
and employment is provided by table 1, which summarizes the changes
on these 3 scores between 1952 and 1955 in the United States and 7
major European countries. This table shows no consistent relationship between inflation and changes in production. The largest increase in output and employment occurred iri West Germany, which
had no inflation at all; the smallest in Sweden which had an intermediate amount of inflation. The next largest increases in production were in Italy and the Netherlands with intermediate inflation;
the next smallest in the United States with very little inflation. Nor
can the apparent failure of inflation to explain differing rates of
growth and output be explained by introducing such other obvious
mtercountry differences as differential increases in the supply of
money, differing money wage ratio average
to prices, or differing
positions on international trade account.5
TABLE 1.—Inflation, production, and employment, 1952-55 *
[1952=100]
United
States
Cost of living:
1953
1954 .
1955
_- _
Industrial production:
1953
1954
1955
Employment:
1953 _
1954
1955

Belgium France Germany

Italy

Nether- Sweden United
lands
Kingdom

101
101
101

100
102
101

99
99
100

98
98
100

102
105
108

100
104
106

101
102
105

103
105
110

108
100
112

100
106
114

97
106
117

109
122
141

110
120
130

110
120
128

100
104
110

106
114
119

103
100
102

99
99
101

98
99
100

104
109
118

100
101
103

103
107
109

97
98
100

101
104
107

i Data from J. Herbert Furth, Indicators of Inflation in Western Europe, 1952-55, Review of Economics
and Statistics, August 1956, pp. 336-337. See that article for an analysis of the various countries' experiences.

4- Does inflation redistribute income ?
Mild inflation apparently redistributes current income among major
economic groups less than is often claimed, though the income redistribution may be appreciable. The redistribution that does take place
does not correspond very well to some of the common preconceptions
about inflation, at least in the recent American economy.
Over the past two decades, every broad functional economic group
in the United States has gained substantially in real income. Within
the rapidly growing total the wage share grew appreciably over the
period; wages and salaries, often considered laggards, considerably
outdistanced profits as a share of the national income. Farmers, who
are commonly supposed to gain most from inflation, saw their share
of the national income decline persistently—perhaps in spite of inflation, but decline nevertheless. Corporation profits declined slightly.
Unincorporated businesses, usually thought to be gainers from infla5
In the interest of brevity and simplicity, the international effects of inflation are completely omitted. For the United States as a whole, this probably does not distort greatly
the conclusions stated. For some sectors of the economy, however, the foreign trade
repercussions may be dominant. And for most other countries, more heavily dependent
on foreign trade, such omissions would not be tolerable.

ECONOMIC STABILITY AND GROWTH

tion, took a substantial cut in their share of the total. The interest
share fell drastically during World War II, as interest rates were
held down by Government policy, but it has grown moderately back
toward its earlier level since then. The rent share has changed little.
Nor is there evidence that inflation has shifted income from the poor
to the rich. If there was any effect, it apparently was the other way.
Figure 1 pictures these changes.
CONSUMER'S PRICE
INDEX
1947-1949 = 100
ID

% OF NATIONAL INCOME
O4

O
1

o
1

° \\

<O
Ol
<O -

3
-

3
Ol

5
Ol

—

V
\\
\\
\
I

x>
m

I
/

1
1
1

CORP"

NET

1947

{

]

to —.

-0
30

is s

EFOR

1951

I
\
\

z
z

. CO
(
-{

1 5

Pi
-4

//
•

Ol
Ol
1

195

I
\

o
z
o

•
*•

•

m
co

1
1

CO
Ol

•
#

o
-<

co
"

o
c
O
o m
m

){)1

h >

O)
X

/ 1
/ I

•

•
(0

c
m

•

ECONOMIC STABILITY AND GROWTH

Since inflation penalizes primarily those whose incomes rise more
slowly than the incomes of their fellows, it was those on relatively
fixed incomes who lost relatively, whether they were rich or poor,
young or old, farmer or city dweller. Above all, it appears to have
been the older people whose piece of the national income pie suffered
most. Their share of the national income dropped substantially as
inflation ate away at their largely fixed incomes. Moreover, here it
appears that inflation did discriminate against the middle incomes and
the poor. Well-to-do retired individuals could afford to diversify
the investments underlying their retirement incomes and to include
substantial amounts of such variable income assets as common stocks
and real estate whose dollar yields increased with inflation. To the
lower income families and widows this alternative was hardly open.
Social-security payments have been raised greatly, and many older
people have jobs. But this does not eliminate the basic fact that inflation is a major blow to this group, living considerably on past savings.
The other most important lagging income group appears to have
been employees of governments and eleemosynary institutions generally, especially teachers. In some cases, teachers' salaries have
lagged so greatly that their real incomes have actually declined during
the greatest boom in our history.
More generally, it was the "passive" economic groups—those who
sold no products on which they could raise the price or who worked
under arrangements where their output could not readily be raised
in price—who suffered at the expense of the "active" groups in society.
A broad generalization is suggested. We may live in a society where
the major economic groups are increasingly effective in protecting their
own income shares during slow inflation; and where private wage and
salary earners are especially successful, at the expense of more passive,
quasi-fixed income groups throughout the economy who have neither
prices to raise nor sufficient political power to push up their incomes
apace with rising prices.
5, 6, and 7. Does inflation redistribute control over wealth? 6
Inflation transfers control over wealth (as contrasted to current
income) from creditors to debtors. This is so because the debtor who
borrows $100 and repays the same $100 later when prices are twice as
high repays only half as much in real purchasing power. As best I can
estimate, the American inflation over the past two decades has wiped
out in this way well over half a trillion dollars of creditors' claims on
debtors (in 1957 prices). All fixed-dollar-value intangible assets (such
as bank deposits, currency, mortgages, Government and corporation
bonds, life-insurance reserves, and pension and retirement funds) are
debts owed to creditors that are susceptible to this erosion by inflation.
In 1938 all such assets totaled just over $300 billion. If we calculate
the loss of purchasing power on these debts up to the present, and make
a similar calculation for the additional net debts of each following
year, we obtain the very rough estimate of over $500 million inflationary erosion of real purchasing power of creditors over the period.
6
The impact of inflation on the distribution of wealth was analyzed in my earlier testimony before this committee, The Impact of Moderate Inflation on Incomes and Assets of
Economic Groups, in Federal Tax Policy for Economic Growth and Stability (1955), and
in G. L. Bach and A. Ando, The Redistributional Effects of Inflation, Review of Economics
and Statistics, February 1957. Only a brief summary, therefore, is included here.

ECONOMIC STABILITY AND GROWTH

Who gained this huge sum of purchasing power which creditors lost ?
On balance, households have consistently been heavy net creditors, and
governments (especially the Federal Government) consistently heavy
net debtors, with the two offsetting each other roughly at something
over a quarter of a trillion dollars each. Unincorporated businesses
and nonfinancial corporations roughly offset each other as net creditors
and net debtors, respectively, but the figures involved are small compared to households and governments.
In summary, inflation has caused a huge transfer of purchasing
power from households primarily to the Federal Government. But
this is clearly not the end of the matter, since the Government is not
some separate entity but, rather, an agency for all of us. We must
look through the Government to see who are the actual beneficiaries
of this inflation-induced levy on creditors.
At first blush, it would appear that taxpayers (that is, all of us in our
capacities as taxpayers) are the gainers. We now need to give up less
purchasing power in taxes to meet payments on interest and principal
on the Government debt. But it is highly unlikely that the Government debt will be paid off through taxation in the foreseeable future.
Who, then, is the gainer of the purchasing power confiscated from Government creditors by inflation ? The answer is, the buying public as
a whole, in proportion to its expenditures. Bondholders' real purchasing power is reduced, thereby increasing the share of the total current
output that can be commanded by the rest of the buying public as their
incomes rise with inflation. Put in commonsense language, Government bondholders and money holders are partially expropriated by
inflation, and the benefit is distributed over the whole population, with
the biggest benefits to those who buy the most.
What types of households lose most as creditors? All households
combined hold about 30 percent of their total wealth in the form of
fixed dollar value assets. By contrast, they are in debt up to only
a little over 10 percent of their total wealth. The difference is a
measure of their net creditor position. Every major group of households is a substantial net creditor by this measure, except for very
poor families and young families in the 25 to 34 age range which are
heavily in debt as they are setting up families and housekeeping.
But the extent to which different groups are net creditors varies
a good deal. The heaviest net creditors, relative to their incomes,
are older people, especially those who are retired. They hold a larger
proportion of their wealth in fixed dollar value assets than do any
other major group, largely because of the importance to them of insurance, and pensions and other retirement funds. Moreover, they
are least in debt, to reap offsetting benefits on that score.
At the other extreme, the very penurious or injudicious who are so
heavily in debt as to have a negative net worth and the younger
families mentioned above are least susceptible to "creditor loss" from
inflation.
Very well-to-do families appear to be in a mixed position. They
are slightly higher net creditors than the average (relative to their
total wealth), and their debts are small. Thus they appear vulnerable to inflation. But they hold an exceptionally large proportion
of their total assets in "variable price" form (common stock, real

ECONOMIC STABILITY AND GROWTH

estate, and so on),
which serves as a partial offset to their exposure
as net creditors.7
S. Hoto does inflation affect businesses?
Nonfinancial corporations are net debtors on balance, but only to
a modest extent. Moreover, on the average about one-third of all
individual corporations have been net creditors at any given time over
the past two inflationary decades. Thus, nonfinancial corporations
stand to gain somewhat on asset account from inflation, but only to
a moderate extent.
For most companies, other factors apparently play dominant roles
in determining the economic well-being of the corporation during
periods of inflation. Sales volume and increases in current cosTs
relative to selling prices appear to explain changes in the economic
position of corporations to a larger extent than do inflation-induced
transfers on asset account. As was indicated above, reported corporate profits in the aggregate have maintained a roughly stable proportion of total national income since the beginning of the two-decade
inflation period, though they have declined gradually since World
War II. "Keal" corporation profits declined appreciably as a share
of the national income, since reported profits over an inflationary
period are larger than they should be if depreciation charges were
sufficient to actually replace wearing out equipment. Moreover, inventory costs are generally undercharged in inflation. Eecent research suggests that such underreporting of costs and resultant
overreporting of profits in inflation was enough to eat up over onethird of reported profits in a small sample of diverse companies
studied. Thus, on balance nonmanuf acturing corporations as well as
unincorporated businesses received a declining relative share of the
real national income through the recent inflation, even though they
gained moderately as net debtors.
ii

The economic impact on America of moderate, creeping inflation
has been substantial, but hardly disastrous. This case against creeping inflation is clear, if modest. But the stronger case against permitting creeping inflation is that this governmental acquiescence
greatly increases the likelihood of more rapid inflation, whose results
will be more disruptive and inequitable. And acceptance of inflation
as a "price" for maintaining high-level employment will generally
be a mistaken choice, since it is unlikely that inflation will in fact
produce the desired full employment.
The history of America appears to provide little basis for expecting
major inflation outside the impact of war.8 If the peacetime inflation
danger is now great, there must be a change to account for it. I believe there has been such a change.
Bursts of inflation may come from many causes. But the danger
of serious continuing peacetime inflation arises fundamentally from
7
For statistical data on the position of different household groups, see Federal Tax Policy
for8 Economic Growth, pp. 78-80.
Major war would bring inflationary dangers of the first magnitude. But in that event
we can anticipate such widespread changes in the entire financial structure and economic
arrangements that discussion at the present level may have little relevance.

ECONOMIC STABILITY AND GROWTH

two major interacting factors: powerful "excess income claims,"
and governmental support of high-level employment through expansionary monetary fiscal policy. The danger is that this combination will gradually turn the creeping inflation to a walk and perhaps
even to a run. Major economic groups—in labor, business, and agriculture—have come increasingly to demand, through the market place
and through the governmental process, larger total income shares than
are consistent with maintenance of a stable price level. There is nothing new about the desire for larger income shares. But to this desire
has been added both the increased degree of organization of major
groups in the market places and, most important, the acceptance of
responsibility by the Federal Government for maintaining high-level
employment and production.
Government acceptance of this responsibility, which has been properly and widely acclaimed, means that if major wages and commodity
prices are pushed up faster than is consistent with high-level employment and production, the resulting unemployment and falling sales
will be bailed out by expansionary governmental monetary-fiscal policy, either directly or indirectly. A complete Government guaranty
of "full employment" would substantially remove the pressure on any
particular income group (seller) to moderate its own income claims,
by removing the fear of unemployment or lost sales. So long as any
group's demands did not get seriously ahead of other expanding
claims in the economy, it could count on continued employment and
sales at new higher cost and price levels. While inflation would
result for the economy as a whole, each income claimant would see
a chance of getting ahead of the parade. Moreover, with some relatively fixed income groups in society, there would always be a "rational" basis for excess income claims, since the active claimants
would always gain at the expense of passive participants in the economic process.
The apparent demonstration of the World War I I period that
massive Government spending can create substantially full employment—that the war and the huge deficit spending connected with it
did solve the vast unemployment problem of the 1930's—has consciously and subconsciously deeply affected the thinking of the present
generation. Since Government monetary and fiscal policy (it appears) can cure unemployment and depression, there is little need
to have unemployment and depression. Thus the normal restraint
against excess income claims, the fear of pricing oneself out of the
market, has been substantially weakened.
The danger of persistent inflation is further increased by the pervasive belief that the Government has a considerable responsibility to
take care of groups in trouble and that a rather high degree of social
security is a desirable public-policy goal. Even more important, the
high cost of defense, probably over many years ahead, increases the
tendency toward inflation. Taxes are always harder to raise than
expenditures, and higher Government expenditures even with a balanced budget probably exert some inflationary pressures.
Logically, this analysis leads to a prediction of cumulatively rising
inflation as income claimants try to outgrab each other. But in spite
of this reasoning, I foresee little danger of truly galloping inflation
of the postwar central European variet3^ in the United States short of
all-out war. This could only occur through a complete collapse of

ECONOMIC STABILITY AND GROWTH

responsible governmental monetary-fiscal policy. While there is little
doubt that the public fears unemployment more than it fears inflation,
should inflation reach anything approaching runaway levels in an
economy like ours, it seems to me overwhelmingly likely that the
public would demand and actively support governmental restraint on
the inflation both through more restrictive monetaty-fiscal policy and
through imposition of widespread direct controls, unpalatable as these
may be in less drastic situations. But the likelihood that inflation may
go beyond a creep to several percent a year seems by no means remote,
if governmental assurance of substantially full employment and maximum production becomes widely counted on as the dominant goal of
national economic policy. Under that circumstance, it is easy to see
how income claimants, both large and small, can persistently raise
their asking prices faster than is consistent with the full-employment
output at stable prices, and that these asking prices may be raised
cumulatively as time goes on.
Indeed, one of the greatest dangers of "moderate" inflation is that
it will grow fast enough to generate strong pressure for imposition of
widespread direct Government controls over individual wages and
prices, under the mistaken belief that this is the only or most effective
way to check rising prices. "Creeping" inflation has not recently
generated such strong pressure to interfere with private markets. But
I suspect that only a small further increase in the rate of inflation
would indeed bring a major danger to the preservation of substantially
free farkets. Imposing a new regulation always offers the illusion
of avoiding the painful restrictive impact of limiting total spending
power.
Many suggestions have been made to blunt the growth of excess income claims. But to date none of these seems very promising. Exhortation to moderation sounds good, but is a doubtful weapon against
self-interest. At least until a better suggestion comes along, the main
protection against an increasing "excess income claims inflation"
must be a recognition by economic groups (sellers generally) that they
do run a risk of pricing themselves temporarily at least out of the
market; in other words, that the Government is not committed solely
to maintaining full employment and high-level output.
The biggest objection raised against this analysis is the widely cited
"unemployment versus inflation" dilemma. What if sellers' wages and
prices are persistently pushed up and the Government refuses to augment the money supply, directly or indirectly ? Unemployment will
result, the argument runs. Only by providing increased purchasing
power can the Government prevent unemployment and falling sales.
Inflation will result, but this inflation is the necessary price for avoiding unemployment.
Such Government-supported inflation may temporarily ease the
excess-income-claims inflation. But if my analysis above is correct,
only very temporarily. For as assurance grows that Government
action will bail out excess-income-claims-induced unemployment and
falling sales, the restraining forces against inflationary wage-price
demands weaken proportionally. Wage and price demands will be
bigger next time around and the Government-supported inflation
required to avoid unemployment will be larger. What will stop the
inflationary demands and responses from cumulating upward?

ECONOMIC STABILITY AND GROWTH

Creeping Government-supported inflation does not solve the unemployment versus inflation dilemna which arises from excess-income
claims. At best it only postpones the dilemma, and even temporary
success becomes increasingly unlikely, even though the rate of inflation is repeatedly stepped upward. The problem of excess-incomeclaims-induced unemployment must ultimately be faced directly. We
cannot have full employment without inflation if excess-income claims
prevail. Indeed, it may become increasingly difficult to have full employment with inflation, if inflation becomes increasingly accepted
and expected. If we must choose between high-level employment and
inflation, most people will choose high-level employment. If we could
assure full employment by having a little inflation, few would hesitate
to incur the inflation. But this is not the meaningful way to state the
choice. To avoid inflation without full employment, we must generate
and preserve an economic climate where sellers expect Government
policy to emphasize both higH-level employment and relatively stable
prices. The solution to the unemployment or inflation dilemma is,
as with many other dilemmas, not to choose one or the other painful
alternative; rather, it is to take steps to avoid having the dilemma
arise.
In the long run, we will avoid an excess-income claims inflation
only if the public really wants to avoid it—only if economic groups
are willing, on the whole, to keep their income claims roughly within
the bounds of full employment output and stable prices. Ultimately
the public must understand the danger of serious inflation and must
consider this an important public policy objective if the long-run inflation danger is to be contained. "The Government" or some part of
it, such as the Federal Reserve Board, can provide leadership against
inflation, and can even restrain inflation counter to widespread public
pressures for some period of time. But in the last analysis, no Government agency can enforce a counter-inflation policy on a public
which does not want it.
ni

What are the implications of this argument for the Employment
Act of 1946? The case for inclusion of substantial stability of the
consumer price level as one objective of governmental policy in the act
seems to me a convincing one. The language of the act is general and
aimed at the promotion of multiple goals. Maximum employment,
production, and purchasing power are central objectives. Addition of
substantial stability of the price level would in no sense weaken the
importance of the goals already stated, but would recognize specifically
another important goal of public policy.
It is the responsibility of the Congress and administrative officers of
the Government to provide leadership on economic policy, as in other
areas. This is critically true in the case of counterinflation policy,
since only expansionary Government monetary-fiscal policy can support a continuing excess income claims inflation. Specific inclusion
of the price level stability objective would be such an act of leadership.
Constant visibility to Members of Congress, to the administrative
branch of the Government, and to the public of the price level stability
goal along with the presently stated goals of the Employment Act
would help to provide a more restraining framework for all in making
their public and private economic policies.

ECONOMIC STABILITY AND GROWTH

It has been argued that the price level stability goal is already
implicit in the act. This may be so, and undoubtedly many Government servants so interpret the present langauge. But one of the
main factors in the inflation since World War II, and one of the
main factors in any future inflation, will be the fact that consumers,
union leaders, and businessmen expect prices to rise. Their resulting behavior has contributed substantially to price rises in the past.
They may well do so again. A clear warning against such expectations can help at least some; it can hardly do any harm.
The major argument against inclusion of price stability in the act
is that this would lessen the promptness or vigor with which the
Government may act against future depressions. I cannot believe
that this is true. The pressures for strong, quick governmental action
against unemployment are powerful and pervasive. But by contrast I do believe that specific stable money language might help to
strengthen the hand of those in the Government who have the painful
and often superficially unpopular task of fighting inflations as they
develop. Temporarily our attention has been diverted from the
problem of creeping inflation to a serious recession. But if inflation
does again grow to more major proportions it will be basically because the Government provides the fuel for the inflationary fire
through monetary and fiscal policy, and because the public comes to
anticipate the provision of this fuel. Inclusion of an announced
goal of substantial price stability is the minimal responsible level of
protective action.
Treading the precarious path toward stable economic growth is at
best difficult, especially with the multiple objectives implied in that
general goal. Building a self-reinforcing set of stable money plus
high output expectations is at best a slow, hard process in our democratic society. We cannot expect success overnight. But current
policy needs to be directed toward this goal at the same time it is
necessarily focused also on the instabilities of each present situation.

PRICE BEHAVIOR, STABILITY, AND GROWTH1
William J. Baumol, Princeton University
When Johnson first arrived in London * * * " 'Thirty pounds a year was
enough to enable a man to live there without being contemptible * * *' It
may be estimated that double the money might now with difficulty be sufficient."—Boswell
I. AN OVERALL VIEW

A. The historical persistence of inflationary pressures
Among the few things which economists predict with any degree
of confidence is that, with some possible exceptional periods, prices
will, on the average, be higher in future decades than they are today.
This belief is based in part on a remarkably consistent historical
record. Scrappy evidence going back to the 13th century apparently
indicates that prices in England and the United States were never
stable for any protracted period, and that, considering periods of the
order of magnitude of 50 years, the price level has risen almost constantly.
However, its persistence is not the only reason we have for expecting
at least a mild but chronic inflation to continue in the future. The
fact is that there are many forces which make for inflation. For example, it is always easier, politically, for a government to increase its
expenditures than to increase its taxes and a counterinflationary
budget surplus is certainly likely to be unpopular.
B. The role of the American firm in economic growth
Perhaps more relevant for the present discussion are the relationships between economic growth and the behavior of the price level.
First I should like to say something about the nature of the forces
which make for growth in our economy. The centers of expansion
decisions in a free enterprise economy are its business firms. This is
not as obvious as it may seem—conservative business practice could
well call for firms to hold the line until rising population and consumption demand created the market for an expanded output. But
business enterprises simply do not operate in this way. Schumpeter
argued long ago that in order to increase their net worth firms must
keep innovating—introducing new products, new sources of supply,
new production methods, etc.
There are other related reasons for management's preoccupation
with growth. Growth can make a firm more attractive to the money
market and its products more attractive to customers. The morale
of its personnel is likely to be much higher than is the case in a declining enterprise. Only by growth can a small firm hope to enter
those lines of enterprise where a large amount of capital is required.
Other advantages of growth to the firm can easily be cited.
1
1 am extremely grateful to Lester V. Chandler, Kenneth Galbraith, and Gardner Patterson for their very helpful comments and suggestions.

ECONOMIC STABILITY AND GROWTH

There is yet another reason why American business enterprise has
worked for expansion of our economy—a reason which will play an
important role in the discussion later in this paper. So far I have
argued that businessmen want their firms to expand because growth
helps them to achieve their other objectives—it opens new profit opportunities, it helps with personnel management, etc. But growth is
important to the businessman not only as a means to an end. In large
segments of American business practice it has become a goal in and of
itself.
I have had occasion to consult with a number of firms in various
industries and of various sizes, and in almost every one this appeared
to be the case. In the so-called oligopolistic sectors of our economy,
the industries containing a relatively small number of large firms
which account for so large a proportion of our Nation's output, among
top management's prime yardsticks of success are sales revenue and
market share. When a businessman in such a firm is asked how things
are going one can be quite sure that he will reply cheerfully that sales
(meaning dollar volume) have gone up x percent (or sadly that they
have gone down y percent) and that only subsequently, if at all, will
he say anything about his profits.
This is not meant to deny that those businessmen are in business
at least partly to earn money. But so long as profits are at a level
which it considers reasonable—enough to keep stockholders satisfied
and to contribute to the financing of growth—management will bend
its efforts to the expansion of sales revenues rather than to further
increases in profits.
To summarize, we have in the nature of American business enterprise powerful dynamic forces which can be depended upon to produce
a remarkable rate of long-run expansion of our economy. Of course
there have been a number of severe and important short-run breakdowns in this mechanism but a glance at the broad outlines of our
economic history suggests that expansion and not depression is its
preponderant characteristic.
C. The effect of growth on price levels
Let us return now to the central subject—the relationship between
growth and the price level. This is a two-way relationship—economic
growth influences price behavior and price behavior influences growth.
Economic expansion clearly puts upward pressures on prices.
When employment is high and businessmen are bidding against one
another for resources which are in short supply it may naturally be
expected that the sellers of these resources will raise their prices.
Unions will demand higher wages. The owners of natural resources
will raise their prices, etc. Furthermore, businessmen will be inclined
not to resist these demands too strongly because their experience is
that during a period of expansion they can raise their prices accordingly without any serious effects on sales.
This is made possible, at least partly, by the fact that the price
raising occurs simultaneously with the rise in costs. For increasing
costs take the form of higher incomes and higher purchasing power
in the hands of the sellers of productive resources who can now afford
the higher prices asked by the businessman. Of course any individual
businessman might nevertheless do better for himself were he able
to resist wage and other cost raising demands so long as no one else

ECONOMIC STABILITY AND GROWTH

did so. For then he could benefit from the increased demand for his
products which results from the increased incomes of other firms'
employees without himself paying any higher costs. But observation
suggests that the corporate form, in which management is in a position of stewardship over the property of others, has made the executive highly reluctant to risk good will and public approbation in
major battles with the unions. So long as rising costs can easily be
passed on management will therefore put up no more than token
resistance to wage and other factor price demands.
Thus it is clear that either because he cannot help it or because he
does not wish to resist too hard, economic growth usually raises the
businessman's costs and hence tends to induce him to raise his prices.
Occasional exceptions are produced by cost-saving inventions which
help to expand sales through reduced prices or by economies of large
scale production made possible by growth of the firm. But, by and
large, expansion can be expected to create bottlenecks and competition
for factors of production whose effect is inflationary.
This is particularly so because growth is contagious. It is not likely
to occur in one or a few industries at a time. Growth in one sector of
the economy creates the income and hence the demand for the products
of other sectors. Moreover in any one industry management's concern
with market share means that the expansion of any one firm is likeley
to be followed by emulatory effort on the part of other companies.
Hence we may expect that everyone will be wanting labor, raw materials, and equipment at the same time and that the effect on prices
will therefore be as described.
D. The effect of price levels on growth
Up to a point slowly rising price levels probably act as a stimulant
to growth. On the one hand they produce an optimistic atmosphere
in which the businessman has confidence in the chances for success
of his expansion programs. Moreover, as we have seen, increasing
costs do not seem to serve as much of a deterrent to output during such
a period.
However when the price rises go on for too long or are too sharp they
can be expected to act as a drag on real output in a number of ways :
1. Unless the public becomes accustomed to uninterrupted inflation
as a way of life people are likely to begin eventually to think that it
must some day come to an end. This expectation effect can lead businessmen to liquidate inventories and consumers to postpone purchases,
particularly purchases of durable goods. A softening of the market
for any increased output (growth) is particularly likely to result if
consumers think prices are too high and decide therefore to postpone
any purchases which represent increases in their living standards.
2. Rising prices can also deter growth because they eat into the value
of an important class of savings. Holders of money and Government
securities will find that the value of their accumulations has evaporated and if there is any reason to make up the loss (e. g., saving
for retirement) they may be forced to reduce their demands accordingly. This is the familiar effect which has gone under various naries
such as the Pigou (and Keynes) effect, the real balance effect, the
wealth saving relationship, etc.
3. So far I have discussed the adverse effects of slow chronic inflation on economic growth when it acts as a deterrent to demand.
23734—58

ECONOMIC STABILITY AND GROWTH

frequently we hear nowadays of the effects on supply but in the long
run I think these effects are of at least equal importance.
The first of these is related to one that has already been discussed,
the effects of expectations on demand. I have already shown that if
people are led to expect an end to the inflationary process, demand may
be decreased and the steam may be taken out of the growth process.
However the alternative can be equally unattractive. If the public
gets to feel that the future holds only still higher prices many people
may cease saving altogether and so an important source of business
financing may disappear. The effect may be a squeeze on the construction of plant and equipment which, in the long run, can seriously
reduce national output. However here there may be an important offset in that increased consumer demand may greatly facilitate corporate saving.
4. A more universal effect of inflationary pressure on supply is an
extremely serious reduction in quality standards. This applies equally to the quality of labor, of raw materials and of product. The
absence of real economic pressures—the universal seller's market
which accompanies inflation—means that there is little motivation to
preserve standards of workmanship. Why produce a better product
if the old commodity can be sold anyhow? Why work hard if employers are short of men and one can always get another job ?
This danger is increased by the fact that reduced quality is a convenient way of disguising price rises. If a product has a traditional
price it is so much easier to put a little less into the packages, to use
shoddier materials, to reduce expenditure on quality control so that
the customer instead of paying much more for the same products pays
the same or only a little more for a formerly cheaper product.
Signs of this are all about us. The candy bar is smaller than it
used to be, food products which require it are not aged as long as
they used to be and so on. When this process is allowed to go on long
enough the economy is transformed into a world of Potemkin villages.
Two houses are produced where there was one before—but these two
houses are basically well disguised shacks. In other words chronically
creeping inflation is likely to turn economic growth into a sham and
to produce a blight on the economy and the landscape from which it
can take a long time to recover.
E. Policy implications
I believe strongly that there are right conclusions and wrong conclusions to be drawn from the foregoing analysis. One wrong conclusion is that price rises are going to occur anyhow and there is therefore very little that can be done about it.
Equally unacceptable to me is the view that the Government must
reconcile itself to some rise in prices partly because they act as a
stimulant to growth, but that the role of monetary and fiscal policy
in this connection should be to keep the rate of increase small—say
to some figure np higher than 4 or perhaps 2 percent per year.
I object to this view on two grounds. One is a matter of tactics
while the other follows directly from the foregoing analysis. The
tactical consideration is that whatever our objective there is some
danger that we will fall somewhat short of it. Experience suggests
that if we are prepared to do no more than to seek to keep the rate of
inflation down to 4 percent per year it is highly likely that prices will

ECONOMIC STABILITY AND GROWTH

rise faster than that. Only from a determination to take all measures
necessary to put a stop to inflation altogether can we with any confidence expect that there will be much of a contribution to its deceleration. In part, but only in part, this is a consequence of the structure
of our political arrangements. It is an inherent and in fact a desirable feature of a political system built about a multitude of checks
and balances that governmental plans are likely to come out much less
sharp and extreme than they were when first conceived.
But aside from this problem of tactics the adoption of a constant
slow rate of inflation as a goal can be objected to on the grounds that
the adverse effects of inflation on growth which were discussed above
can be produced by slow chronic price rises perhaps at least as effectively as they can by occasional sharp increases in the price level.
The attrition of savings can occur in either case and the destruction
of quality standards may be more complete for having occurred in
slow imperceptible steps.
To put the matter affirmatively, the promotion of economic stability and growth requires monetary and fiscal policies which can
and do prevent uninterrupted inflation. This conclusion is not based
on considerations of equity which clearly may by themselves also
make it imperative that inflation be brought to a halt. For the
question to which I have addressed myself is increased growth of
total national output rather than its distribution. Perhaps on either
ground it may be argued that we cannot tolerate a policy which passively accepts uninterrupted inflation as an immutable feature of the
national fate.
II. THE ROLE OF THE MARKET STRUCTURE

In an important recent contribution Professor Galbraith has called
our attention to the fact that the process by which price
changes
occur is dependent on the nature of the market structure.2 Prices are
set in a different way in those sectors of the economy which are composed of many small firms than they are in industries where there
are 5,10, or 20 major producers (the oligopolistic sectors).
Galbraith has argued that oligopolistic firms are particularly effective conductors of inflationary pressure (pp. 127-130) and that
they are relatively immune to the counterinflationary influences of
monetary and fiscal policy (pp. 130-132). From these results of his
analysis he concludes that the efficacy of monetary and fiscal policies
must depend largely upon their impact on the competitive sectors
of the economy which must bear the brunt of the burden of readjustment (p. 132) and points out that this raises serious questions
about the effectiveness, the equity, and the wisdom of such policies.
In particular this view suggests that much more extreme monetary
and fiscal measures may be required to achieve a given disinflationary
effect than we would otherwise have thought to be necessary.
I must differ with Professor Galbraith on a number of these points.
I think it can be shown that the difference in role of the competitive
and that of the oligopolistic firm in the inflationary process ip not
nearly as great as he would have us believe. To indicate my reasons
a J. K. Galbraith, Market Structure and Stabilization Policy, Review of Economics and

Statistics, May 1957.

ECONOMIC STABILITY AND GROWTH

for this view I must first digress into a discussion of the nature of
oligopolistic pricing.
A. Galbraith on the oligopolistic firm
To Professor Galbraith, as to many investigators, a prime characteristic of the oligopolistic firm is that it does not normally set a
price which maximizes profits. So far I am in agreement. But in
his analysis there seems to be no alternative explanation of the pricesetting process. Somehow, perhaps by historical accident, the price
of each item seems to have been set up, and then for fear of setting
off a price war through retaliatory efforts of competitors the oligopolist apparently never changes such a price unless universal cost
changes permit him to do so without being misunderstood by business
rivals. The executive is apparently taken to be less concerned with
the level at which his prices are set than he is in keeping his prices
at that level. We are given little indication how those prices happened to be set at those levels in the first place.
It would be a very odd coincidence indeed if those prices were to
happen to lie at their precise profit maximizing levels. In fact we
are told that they will generally lie below those which will maximize
profits apparently because inflation calls for rapid upward price
shifts as demand increases in monetary terms—price rises greater
than those which the oligopolist is willing to permit. Furthermore,
according to Professor Galbraith, "A commonplace feature of a firm
under inflationary demand is a backlog" (p. 128). This backlog can,
of course, be eliminated by a price sufficiently high to cut the quantity
demanded down to the quantity supplied, but such a price rise the
oligopolist presumably refuses to permit.
As a result of all of this, he argues, monetary and fiscal measures
may be rendered largely ineffective in the oligopolistic sector. Any
rise in costs incurred as a result of interest rate increases need not
reduce the oligopolist's profits. For, since he is not maximizing his
profits to begin with, he can raise his price and increase his profits sufficiently to make up for the higher interest cost. Any reduction in
demand produced by such policies may serve only to reduce the backlog
of demand and may actually result in no decrease in sales.
The oligopolist will then, according to Galbraith, be able to go on
as before, with no loss in profits and perhaps no less in sales. He
will have no reason to cut down his investment expenditure. His contribution to inflationary pressure will therefore not be diminished
unless monetary or fiscal actions are particularly severe.
Not such will be the fate of the competitive small-business man
where prices are fixed on the market and where rises in costs and cuts
in demand really hurt and hurt at once. Thus, according to Galbraith, besides being less effective than might have been expected,
counterinflationary monetary and fiscal policies can serve to increase
industrial concentration and help to contribute to the danger of decline of the small-business man.
B. Sales maximization, price setting, and the backlog
The short run purposes of the large firm can, I believe, be explained
somewhat further. There is some logic to the process whereby prices
are set, but it is not the logic of profit maximization. As I have already stated, I believe most oligopolistic firms aim to maximize not
profits, but sales volume (measured in money terms; sales ^tre what the

ECONOMIC STABILITY AND GROWTH

economist is his jargon calls "total revenue"). So long as profits
remain at a satisfactory rate, management will devote further effort
to increasing its sales rather than its profits.
It follows that Professor Galbraith is indeed right in asserting that
the oligopolist's profits will not normally be maximized. But the
reason is not because he is dominated by fear of making price adjustments, although such fears may also play their role. He will fail to
maximize profits also because he has another partially overriding purpose to which he is willing to sacrifice some profits.
If this is correct, it follows that we must expect only under extraordinary circumstances to encounter the backlog of orders which Galbraith considers to be a normal feature of oligopolistic operation during an inflationary period. A firm which tries to maximize sales will
not hold back on production unless either the expansion of its capacity
has been unable to keep up with demand or unless management is so
fearful of the future that it hesitates to undertake the required investment commitments.
Consider now the effects on such a firm of a rise in costs which might
occur, for example, as the result of an increase in wages or in interest
rates. As in Professor Galbraith's analysis, this may lead to an increase in the company's selling prices sufficient to prevent any reduction in profits. For otherwise the rise in costs can reduce the oligopolist's profits below that level which he considers to be the minimum uncler which he can operate. But several other consequences
follow from this step:
In the probably more usual case where there is no backlog of orders,
this price rise must mean some cut in the number of items sold, or at
least unit sales will not increase as quickly as they otherwise would
have. Moreover, if the initial price had been chosen to maximize
sales revenues, it is clear that the rise in price must result in a reduction in (or a reduction in the rate of increase of) dollar sales. It
follows that—
1. An interest rate or wage increase will also hit the oligopolist
where it hurts him—not in his profits but in his sales volume.
2; An interest rate rise need produce no increase in industrial concentration—the oligopolist's sales will decline right along with those
of the competitive firm.
3. By reducing the demands for his products or increasing his costs,
monetary or fiscal policy can effectively reduce the oligopolist's contribution to inflationary pressure. If his sales fall it will pay him to
reduce his demand for factors of production either for immediate use
in his manufacturing process or for investment purposes. But this is
precisely what a disinflationary monetary and fiscal policy seeks to
get him to do.3
Thus in each of these respects I am forced to disagree with Professor
Galbraith's policy conclusions—considering the oligopolistic sector
alone, fiscal and monetary policy need not on the face of it be inequitable, it need not promote monopolization, nor is it to be expected to be
ineffective.
8
Incidentally, I must take exception to Galbraith's assertion that "* * * firms in the
oligopolistic sector have the opportunity of offsetting any credit restrictions to which they
are subject by increasing their prices and their earned resources and devoting them to
investment" (p. 131). It can take several years of increased accumulation out of profits
to offsejb a large reduction in credit. Hence this means may well be insufficient to prevent
forced postponement of oligopolistic investment which is all the deflationary effect that is
usull
ht

ECONOMIC STABILITY AND GROWTH

Even in the case, which I have not yet discussed, in which there is a
large backlog of unfilled orders the Galbraith conclusions can be disputed. If the backlog occurs by design because the oligopolist is afraid
of the future and he wishes to retain a cushion of demand against
contingencies, the rise in price or the reduction in demand which results from anti-inflation policy will cut down on that cushion. The
oligopolist may then be expected to hold back even more in his investment plans, thereby contributing once again to a reduction in inflationary pressure. Similarly, if his backlog of demand piled up because he has, so to speak, been unable to get sufficiently quick delivery
on expanded plant and equipment a reduction in the backlog can mean
partial cancellation of investment orders.
C. The competitive sector
Yet there is an important observation which I think lies behind
Professor Galbraith's discussion. In times of inflation it is true that
the firm can frequently raise prices in response to cost changes and get
away with little or no observable ill effect. I have already had occasion
to refer to that process earlier in this paper. What occurs here is that
the cost rise, because it is universal, provides the purchasing power
which permits consumers to buy as much as they did before the price
rise. This is the mechanism of the familiar wage-price spiral.
But here the competitive firm is in no worse a strategic position than
is the oligopolist. The demand for the small-business man's products
will be raised by increases in wages, both his own and the oligopolist's.
Thus the competitive firm is able to participate wholeheartedly in the
inflationary process just as is the oligopolist. Indeed, it may almost be
suspected that the small firm will in this respect occupy the slightly
better position. For in competitive markets the translation of increased costs into higher prices is rapid, impersonal, and automatic.
To the extent that the large firm hesitates to change prices even in the
upward direction, it may therefore be placed at a relative disadvantage.
Of course rises in interest rates differ from wage rises in that they
do not all go into increases in money income in the hands of the public,
A large proportion of higher interest payments may represent a transfer to banks and insurance companies where they will not immediately
be used to swell the monetary demand for commodities. That is essentially why interest rate increases can act as an offset to inflationary
pressures while wage rises do not ordinarily do so. But if interest
rate rises do prevent increases in demand they will do so equally for
the oligopolist and for the small firm.
I am not to be interpreted to hold the position that different monetary and fiscal policies all and always fall with exactly equal weight
on the competitive and oligopolistic sectors of the economy. Clearly
this need not be so. It depends on costs structures, the relative elasticities of demands for the products of the two sectors and many other
considerations. My position is just that the relative effects can only
be determined after far more careful and detailed investigation of
their product lines and related matters, and that I will be very surprised should such an investigation show that all the tools of monetary
and fiscal policy consistently favor the one group against the other.

ECONOMIC STABILITY AND GROWTH

HI. PRICE POLICY FOR GROWTH

The preceding section was primarily negative in purpose, seeking
largely to deny that there were grounds for expecting a certain type
of relationship. Nothing was said affirmatively about the types of
policy which can promote price stability and economic growth. In
this section I hope to supply this deficiency.
A. Growth as a goal
As already stated, complete price stability and economic growth
are not totally compatible objectives. Growth tends to cause prices
to rise. It also has other costs. The most important of these is that
rapid growth requires a very high level of physical investment—of
construction of plant and equipment, etc. In a period when there is
little unemployment and unused capacity an increase in the production of investment goods must involve some reduction in the output of
consumers' goods. As is well known this is precisely the sort of decision to which the ruling clique in the Soviet Union has stuck these
many years. And, at least in part, this decision has been responsible
for their very noteworthy success in expanding Russian military and
industrial capacity at so phenomenal a rate.
Before effective policies for the promotion of economic growth can
be decided upon in this country we must therefore decide first how
much we are willing to give up for it. This is no idle question Our
overall economic growth has been proceeding by itself at a remarkably
high rate and there is reason to believe that, except for occasional interruptions, it is going to continue to do so. It is possible to give up
too much for the future like the miser who is starving because he is
putting aside every penny he earns.
Even the Russian threat must not force us to jump to hasty conclusions on this question. There is much to be said for the view that
to meet the competition of the iron curtain countries we need to speed
up not our overall rate of economic expansion but the growth of some
key outputs—scientific, military etc. Certainly this has been the
Soviet strategy.
B. Policy for overall growth
Suppose it is decided that somewhat more of our national effort
should be devoted to speeding up our economic expansion and less to
the production of current consumption goods. What can the Government do to promote these goals?
I have already argued, for one thing, that strong measures must be
taken to prevent uninterrupted inflation. When price rises threaten
to get out of hand or to last too long the usual array of weapons—•
budget surpluses, tight money, et cetera, must be employed to the extent necessary to put an end to the inflationary pressures.
But I do not believe this is by any means the most important measure to be taken to promote growth. In a free enterprise system where
the Government seeks to avoid the regulation of production by fiat
except in time of national emergency, businessmen must be induced
to take the appropriate measures voluntarily. Nor is this a matter
for exhortation and propaganda whose effects are doubtful and whose
use is usually inappropriate except in times calling for public heroism. Management's responsibilities are too great to run the business

ECONOMIC STABILITY AND GROWTH

enterprises with whose operations it is entrusted in accord with the
utterances of remote public figures.
There is only one dependable way to achieve the desired results—
that is to make them pay. And there is one instrument ideally suited
to achieve the Government's purposes in this respect—the tax laws.
Already the capital gains provisions served to make the retention of
corporate earnings attractive to large stockholders. This is an incentive for growth because retained earnings are readily employed to
finance plant expansion and other types of investment.
In the same way, if Congress wishes to promote economic growth
(and, I repeat, there is something to be said on either side on such a
decision) it must consider undertaking a revision of the tax laws in
a way which favors investment as against current production of consumers' goods. Businessmen, who are already anxious to promote
economic growth, will be quick enough to take advantage of any
opportunity to reduce their tax burden through increased investment
or any other expansionary activities which may be favored.
In the case of underdeveloped areas I have elsewhere gone so far
as to suggest that firms be offered a tax exemption which is larger the
greater is the percentage rate of increase in their production. I do
not believe so radical a proposal is appropriate for the United States
but it does serve to underline the nature of the policy approach which
I am urging.
I shall make no attempt to indicate the details of the tax-law revisions which would be in line with this view. That must be left to
persons more expert in the tax structure and its legal and economic
complications. I hope however that I have suggested one promising
avenue for exploration.
C. The long-run position of the United States
Let us return now to the question of the Soviet challenge although
this represents somewhat of a departure from the subject of the rest
of this paper. I think it is quite clear that the American industrial
capacity is still considerably greater than that of the Eussians. This
means that were it deemed crucial, there is little doubt that a crash
program could permit us to catch up on the production of missiles,
rockets, jet planes, or any other particular weapons on which we are
behind. But catching up on special items is not the problem. We
cannot plan on just barely keeping up by means of an interminable
series of emergency programs conceived in the haste which is engendered by panic.
Indeed, it is to be expected that in the normal course of events the
Eussians will from time to time come up with developments before we
do. The Soviet successes are frightening rather because they suggest we are falling behind in a more fundamental sense—in our supply of ideas and our ability to carry them out. Should we ever lose
our lead in the fundamental knowledge which we possess and the skills
to produce and use ever more knowledge we will really find ourselves
ready prey to the totalitarian world.
The difficulty appears to lie partly in our national propensity to
favor practicality. American industry, for example, is more than
willing to Sponsor researchers and thfcir research, provided that their
projects are of a variety which give promise of practical application.
From the point of view of the firm this may make good sense although,

ECONOMIC STABILITY AND GROWTH

increasingly, management, too, is showing signs of questioning this
position. But from the point of view of the Nation's future this
represents the most catastrophic sort of shortsightedness. Time and
time again it has been shown that some of the most fundamental and
revolutionary scientific and industrial developments would not have
been possible without the work of the basic researcher, the man who
does not know and, at least within limits, could not care less how his
results will be applied. Innovations from electricity to the hydrogen
bomb bear their debt to the completely impractical investigator.
The Soviets have certainly recognized this and their efforts to sponsor both research and researcher are apparently ahead of ours in a
number of respects. Their educational system is geared to supply
new scientific talent at a very high rate and it is already showing its
effects.
Though it is perhaps beyond the province of the Employment Act
this problem is clearly fundamental for economic growth. Here
again congressional action is probably called for. The design of a
comprehensive policy is beyond my intentions and my abilities. I
should, however, like to call attention to two suggestions which, I hope,
merit further exploration.
I believe the Government should consider setting up a system of
national scholarships on the order of those which were provided by the
G. I. bill. These scholarships should be available only for college and
university training. They should again permit the recipient to attend
and approved school of his own choosing subject to his acceptance for
admission by that institution. However, such scholarships should
be awarded on the basis of examinations which require of the applicant
evidence of training at least equivalent to that offered in a Soviet
gymnasium. This proposal is designed to achieve two purposes: to
increase the number of able college students and to force a drastic
upgrading of standards of our high schools which are notoriously so
far behind the equivalent Russian institutions.
However, this does not cope with another fundamental problem—
the relatively small proportion of able American students who go into
research rather than into industrial, legal, and other more lucrative
occupations. It is absolutely imperative for this country's future that
the research activities be made more attractive. One proposal that
has recently been made calls for the setting up in each field of research about 50 well publicized chairs bearing a minimum stipend of
$30,000 per annum plus adequate research funds. This would be
awarded to the most outstanding members of the research professions
and would not be attached to any particular academic institution.
Whatever the merits of this particular proposal it points up the
need for a dramatic move to increase the attractiveness of a career in
fundamental research. Such a conclusion is inescapable if we are prepared to accept two premises: (1) that America's long-run future is
at stake and depends on its ability to maintain and increase its supply
of knowledge and ideas, and (2) that in a free-enterprise economy
coercion is an inappropriate instrument and exhortation an undependable means to promote the goals of our society. Only by an
appeal to the profit motive which is so fundamental among the Nation's institutions can we feel confident that our goals will be achieved.

ECONOMIC STABILITY, ECONOMIC GROWTH, AND PRICE
STABILITY
Betty G. Fishman and Leo Fishman, West Virginia University

The terms "stability" and "growth" have been used so frequently
during the past few j^ears in discussions of the Employment Act of
1946, that a goodly number of intelligent citizens may very well believe that these words were used in the act, to set forth its purpose.
Actually, of course, this is not the case.
According to section 2 of the act—
it is the continuing policy and responsibility of the Federal Government to use
all practicable means * * * for the purpose of creating and maintaining, in
a manner calculated to foster and promote free competitive enterprise and
the general welfare, conditions under which there will be afforded useful employment opportunities, including self-employment, for those able, willing, and
seeking to work, and to promote maximum employment, production, and purchasing power.

Subsection (c) 4 of section 4 states that one of the duties and functions
of the Council of Economic Advisers shall be—
to develop and recommend to the President national economic policies to foster
and promote free competitive enterprise, to avoid economic fluctuations or to
diminish the effects thereof, and to maintain employment, production, and
purchasing power.

But the words "stability" and "growth" do not appear in the Employment Act either separately or together.
The number of years which has elapsed since the passage of the
Employment Act, together with the nature and diversity of the economic developments which have occurred during that time are sufficient to justify the conclusion that the language of the Employment
Act is remarkably well suited to the general purpose of the act. That
language is general enough to provide for flexibility in interpretation
and also for the flexibility in action necessary if the act is to furnish
a useful guide to economic policy in particular circumstances, the
exact nature of which could not be foreseen at the time the act was
passed. For example, there is no doubt among either economists or
policymaking officials that the Employment Act provides the basis
for Government action designed to restrain inflation when such action
appears necessary, as well as the basis for Government action designed
to mitigate or halt a recession when such action appears necessary.
Employment, production, and purchasing power are each highly
significant concepts in assessing the adequacy with which the economy
is functioning, and in determining what steps, if any, need be taken
to improve its functioning. They are also concepts which are generally understood and defined in a similar fashion by economists and
policymakers alike. In the case of employment and production, more-

ECONOMIC STABILITY AND GROWTH

over, policymakers need not confine their attention to the abstract
conceptions or to indirect methods of approximating the actual level
of these indicators of economic activity. Fortunately, there are available statistical series which provide reliable data on both employment
and production with very little time lag. In the case of purchasing
power, there is no equally satisfactory statistical series on which those
concerned with economic policy may rely. Even here, however, there
are data available which, when used with proper discrimination and
judgment, can be very helpful.
It may be pointed out that implicit in the language of the act is
the belief that maximum employment, maximum production^ and
maximum purchasing power either are or can be made to be consistent
with one another, and also with the objective of avoiding economic
fluctuations or mitigating their effects. Whether or not this is actually so under all circumstances may, of course, be questioned. But
it is evident that principal emphasis in the act is placed on employment, and this emphasis could furnish a clue to policymakers in what
otherwise might prove to be a perplexing situation.
II

Although the terms "stability" and "growth" have been used for
some years now in discussions of the Employment Act both in and
out 01 Government circles, those terms have not always been construed or applied in the same way, nor have they always been assigned the same degree of importance or priority.
Thus there have been times when stability has been emphasized and
growth has received little if any attention. At other times growth
has been emphasized and stability, at least overtly, has been substantially neglected. More recently it has been fairly common to link the
two together and to talk of "economic stability and growth" together.
In general in the years subsequent to 1946, concern about stability
and use of the term increased together with signs of actual or impending recession or actual or impending inflation. Under such circumstances, stability was generally not defined clearly by those who
used the term. It was, perhaps, sufficiently obvious that they applied
it in a negative rather than a positive fashion. In one set of circumstances it meant avoiding recession or at least avoiding a deepening
of recession. In another set of circumstances it meant avoiding inflation or at least avoiding a heightening of inflation. During some
periods when neither recession nor inflation were present or imminent,
attention shifted from stability, and the term "growth" was used
more frequently. This term, too, was rarely defined, but it was generally clear that it was held to be integrally related to increasing
employment opportunities, and sometimes to increasing productivity,
and/or a rising level of production as well.
Until quite recently, when both economic stability and growth were
discussed in the same paper or in the same public statement, "stability" was frequently referred to as a suitable goal for short-run
policy, and "growth" as a suitable goal for long-run policy. In such
discussions, also, while the term "stability" was never clearly defined,
it was generally applied in the negative sense described above.
More recently, however, there has been increasing recognition of
the fact that it is impossible to keep short-run and long-run consider-

ECONOMIC STABILITY AND GROWTH

ations neatly enclosed in separate compartments in the formulation
of public policy. For one thing it has become increasingly apparent
that policies which are adopted to cope with so-called short-run problems have a bearing on long-run developments as well, while policies
adopted to cope with long-run problems also have a bearing on shortrun developments. Even more fundamental, however, is the fact that
in an economy as highly developed and large and rapidly changing
as ours, it becomes increasingly difficult to draw and maintain a satisfactory distinction between long-run and short-run considerations.
In the not too far distant past many economists might agree that
the short run meant any period of time shorter than one complete
business cycle. They might agree also that in dealing with the short
run, the matter of economic growth could be ignored with impunity,
since the amount of growth in the short run was generally negligible.
Today, however, a large and increasing number of economists in the
United States realize that when an economy as large as ours is growing at an average annual rate of 3% percent, the average growth of a
single year accounts for a real increase of approximately $14 billion
in net national product and a correspondingly substantial increase in
employment. They realize too that with our population and labor
force increasing at their current rate, growth must be maintained
at all times, if "useful employment opportunities" are to be created7
and maintained for all "those able, willing, and seeking to work.'
And they are aware of the corollary of these facts; namely, that even
a comparatively short cessation or interruption of economic growth
will result in a large increase in unemployment.
In other words, we must run just to keep up with ourselves. But,
as Roy F. Harrod and Evsey D. Domar have both demonstrated, the
rate of speed at which we run is also important. If we proceed too
slowly, we will still be faced with increasing unemployment. On the
other hand, a sharp increase to a rate above that which can be maintained for a long period of time will result in an unhealthy inflationary situation which paves the way for future contraction or recession.
Thus within the past few years it has become increasingly common
for public officials to link the words "stability" and "growth" together
and to use them as one term in discussing the objectives of the Employment Act and the implementation of the act. The two are conceived
of as being closely interrelated, indeed, interdependent. For economic
stability is thought of not in the dictionary sense of permanence or
absence of change, but rather in the sense of minimum fluctuations
around the growth trend. Growth itself is generally equated with the
real increase in net national product.
It is fairly easy to understand why the 2 words are so often linked
together and used as 1 term. Each of them expresses one aspect of
the type of economic process which is consistent with the stated objectives of the Employment Act and indeed necessary for achieving
those objectives. Yet neither of them by itself is sufficient. The word
"stability" by itself is not satisfactory for this purpose. We must
know something of the level at which stability is to be maintained.
But the level which is satisfactory now will be unsatisfactory a year
from now and still less satisfactory a year after that.
On the other hand, the word "growth" by itself is not completely
satisfactory. Growth must proceed at a rate such that it can be maintained in the long run. Some fluctuations in the rate of growth, it is

ECONOMIC STABILITY AND GROWTH

generally realized, must occur and must be tolerated in a free society,
but those fluctuations should be relatively minor. In other words,
economic growth should proceed at a relatively stable rate.
Such usage of the words "stability" and "growth" by public officials is probably in accord with prevailing opinion among reputable
economists today. But is this usage really desirable?
Attention has already been called to the fact that the meaning generally ascribed to the word "stability" in current discussions of the
Employment Act is very different from the dictionary meaning. It is
different, too, from the meaning generally ascribed 'to the word only
a few decades ago by both economists and policymakers. And it is
certainly different from the meaning currently ascribed to "stability9'
by economists and policymakers when they discuss price stability.
In view of these facts, the current usage of the word "stability" in
discussions of the general objectives of the Employment Act is bound
to be confusing, at least occasionally. In addition, if the type of
growth which is desirable and consistent with the objectives of the act
is a stable rate of growth, the use of the word "stability" in the term
"stability and growth" is somewhat redundant. If the word "growth"
by itself is considered inadequate to convey so complex a meaning, or
if we wish to be more explicit on this score, might it not be preferable
simply to make use of the expression "a steady rate of economic
growth" or "a stable rate of economic growth" in our thinking and
talking and writing, rather than "economic stability and growth" ?
It would be desirable also for those who use the word "growth"
either in theoretical analysis or in discussions of economic policy, to
give more frequent explicit recognition to the fact that economic
growth is not uniquely determined by technological change and/or
changes in the size of the labor force. The rate of economic growth
is more properly thought of as resulting from the interaction of a
number of factors. Both technological change and an increase in the
size of the labor force, whether occasioned by population growth or
by change in social conventions and attitudes, are of course important
in this connection. But their importance should not cause us to overlook the role played by such factors as newly developed forms of business organization and new business methods, increased accumulation and application of capital even in the absence of technological
change, increased utilization of existing capital, an increase in hours
worked, or an increase in the availability of natural resources.
m
Both Domar and Harrod have demonstrated that while growth
does not insure stability unless the proper rate of growth is achieved
and maintained, it is possible to determine theoretically a stable rate
of growth for a free economy under certain specified conditions. But
is a stable rate of growth for the economy of the United States actuallv a realizable ideal ?
Practically all economists today recognize that there is some relationship between economic fluctuations and economic growth. It is
rather generally realized that economic growth does not proceed
smoothly or evenly. During some periods growth is retarded. At
times it may even stop or be reversed. During other periods growth
is accelerated, sometimes slightly, and sometimes considerably more
than slightly.

ECONOMIC STABILITY AND GROWTH

Not all economists agree on the precise nature of this relationship,
however. There are probably a large number who, even if they have
not followed or do not accept all the details of Harrod's and Domar's
demonstrations, nevertheless do accept their general conclusion that a
tendency to instability is inherent in the growth process; i. e., that
departures from a stable rate of growth will not automatically be
corrected, but will instead lead to further departures. But while
some economists believe that the fluctuations accompanying growth
are necessarily cyclical in character, others believe that this need not
be so.
The distinguished British economist, Harrod, although he does not
believe the existence of cyclical fluctuations in the real world is caused
solely by growth, has nevertheless demonstrated that an understanding of the growth process leads to the expectation that economic
fluctuations of a cyclical nature will occur. His analysis leads to the
conclusion that such fluctuations are not likely to be eliminated completely in a free society, where the actual rate of growth at any time
is the result of the combined effect of so many different individual
decisions. But Harrod does suggest the possibility of reducing the
magnitude of the fluctuations by appropriate policies.
The late Joseph A. Schumpeter believed that there is an inherent
causal relationship between economic growth and cyclical fluctuations, and that cyclical fluctuations are the mechanism through which
growth occurs. Devoted followers of Schumpeter may, therefore,
question whether it is possible to maintain growth without cyclical
fluctuations, and whether any attempts to eliminate or even sharply
reduce cyclical fluctuations may not result in economic stagnation,
or at least in a long-run rate of growth much lower than that which
might otherwise prevail.
Most economists in the United States today, however, do not subscribe to this point of view. A substantial number would probably
agree that complete elimination of fluctuations—both cyclical and
noncyclical—from the optimum rate of growth is impossible of
achievement, at least in a free society, and that maintenance of a
perfectly stable rate of economic growth in the United States for an
extended period of time is, therefore, equally impossible—especially
since the theoretical optimum itself may change from time to time.
They would probably also agree, however, that the concept of a
stable rate of economic growth is a useful one, not only as a theoretical tool of analysis, but as a guide for policymakers both in and out
of the Government. It keeps us moving in the right direction, and
while we may never quite reach it, we can certainly reduce the magnitude of fluctuations in economic activity and move closer to the
ideal of a stable rate of growth than we have yet come.
IV

Economists are generally agreed that prices are closely and significantly related to the economic process and the general level of economic activity. They are agreed, too, that the role of prices is complex,, since they are a cause with respect to some economic phenomena,
an effect with respect to others, and in many situations may be regarded as either cause or effect depending on the point of view or

ECONOMIC STABILITY AND GROWTH

point of departure of the analyst. Moreover, they are agreed that
price developments may, and frequently do, have an important influence on the level of economic activity. With respect to the precise nature of that influence, however, opinion is divided.
Within the past 2 years increasing attention has been given to the
concept of price stability in discussions pertaining to the Employment
Act of 1946. A number of economists and public officials have voiced
the belief that stable prices are essential to the promotion and maintenance of maximum employment, production, and purchasing power.
Some have even proposed that stable prices should formally be made
a goal of public policy and that an explicit declaration to that effect
should be incorporated in the Employment Act. Among those who
have urged that the act be amended in this fashion are Arthur F.
Burns and William McChesney Martin.
These economists, of course, and others who have voiced similar
opinions are concerned not with individual prices or price relationships, or even with prices and price relationships of groups of closely
related commodities, but rather with the general price level. Often,
however, for reasons which are wholly or largely practical, rather
than conceptual in nature, "the level of consumer prices" is accepted
either explicitly or implicitly as a substitute for "the general price
level."
Some economists who are in favor of amending the Employment
Act believe that the goal of policy should be "reasonable price stability" ; while others apparently lean toward a more rigid or less flexible
application of the concept of price stability by policymakers. In all
cases, however, those who propose that the declaration of policy in the
Employment Act be expanded to include price stability, use the word
"stability" not in the sense in which it is often used in current discussions of the general level of economic activity (i. e., minimum fluctuations around the growth trend), but rather in the dictionary sense
(i. e., steadiness, absence of change, permanence).
In view of the eminence and prestige of some of the economists and
public officials who have suggested that a stable price level be made an
explicit goal of public policy by amending the Employment Act, it
may seem bold to question the desirability of such action. Nevertheless, the present writers feel that it is important to raise this question.
It may be well at this point to call attention to the fact that within
the past few years there has been an increasing tendency to use the
word "inflation" without defining it, and to'apply the term to any rise
in the price level. Such loose usage can only result in a cer.tain degree
of confusion and inconclusiveness in discussions pertaining to inflation
and the price level, since not all participants are talking about the same
thing. In the following pages the term "inflation" is used by the
present writers only to describe a rise in the price level which is not
accompanied or immediately followed by an increase in total production of at least equivalent significance. In accordance with this usage
price rises which are accompanied or immediately followed by an
equally significant (or a more than equally significant) increase in
total production are not considered or labeled inflationary.
The proposal that the declaration of policy in section 2 of the Employment Act be amended to include price stability, is generally based
on two assumptions: (1) That in the absence of effective preventive
measures, a rising price level will be a basic characteristic of the

ECONOMIC STABILITY AND GROWTH

American economy for many years to come; and (2) that effective
implementation of the goal of price stability is and always will be consistent with the promotion and maintenance of maximum employment,
production and purchasing power. A third assumption, although
frequently omitted, or at least not clearly stated, should really also
be included here": namely, that from a technical and administrative
point of view stability of the price level constitutes a desirable goal
of public policy.
The assumption that a rising price level is virtually certain to be a
feature of the American economy for an indefinitely long time to
come is in turn generally based on one or more of the following
assumptions: (1) That the expansionary forces which have played so
important a role in shaping economic conditions in the United States
since the end of World War I I will continue to play an equally important role in the future; (2) that in the future, economic recessions
will be relatively infrequent, and also relatively short, and mild: No
more protracted or serious than the two relatively minor recessions
of 1949 and 1954, for example; and (3) that the Federal Government
will deal with future economic recessions swiftly and effectively. The
course of action pursued during the 1949 recession and that pursued
during the 1954 recession, are frequently cited in support of this
point.
The future is, as ever, uncertain and in the end it may turn out that
these assumptions were justified. Nevertheless, there are at present
sound reasons to question the validity of each of these assumptions.

If we analyze the forces making for economic expansion during
the postwar period, it is apparent that we cannot expect them to
operate in the same fashion throughout a future of indefinite duration. At the end of World War I I there was a huge backlog of
pent-up demand for many types of consumer goods and capital goods,
accompanied by a record volume of liquid savings held by both individuals and business firms. In addition, throughout the postwar
period the United States has had a rapid rate of population growth,
a rapid rate of technological advance, and a high level of public
expenditures. The increase in Federal expenditures is by now a
familiar story. It is perhaps less generally realized that from 1946
to 1957 total expenditures by State and local governments increased
more than 250 percent. A considerable part of this increase can be
ascribed to the backlog of unsatisfied needs which accumulated during World War II, and to the rapid rate of population growth.
The backlog of pent-up demand in the private sector of the economy
which existed at the end of World War I I persisted in making its
effects felt for longer than was originally anticipated. It seems probable that the outbreak and the duration of hostilities in Korea were at
least partly responsible for this prolongation. By now, however, the
backlog of pent-up demand is, of course, a thing of the past. The
rate of population growth continues high. But one of the lessons
of the past 15 to 20 years has been that the rate of population growth
is less stable than we formerly supposed. For many years the rate
of population growth was considered to be subject to such slow, gradual change that the level of population could be predicted with con23734—58

ECONOMIC STABILITY AND GROWTH

siderable accuracy for some time ahead. More recent developments
have routed this belief. As recently as 1940, many economists and
public officials were concerned about the economic effects of the declining rate of population growth. There was no hint then of the change
which was to occur within a few years. And when the change did
become evident, it was for some years considered to be a purely temporary phenomenon. Knowing these things, it would be imprudent
for us to base our cifrrent actions and plans on the expectation that
the current rate of population increase will be maintained or will rise
still further.
Technological advance, although certainly an important factor, is
notoriously unstable. Continuation and even expansion of the huge
research and development programs which have provided the basis
for most of the technological advance since World War I I will not
insure the continuation of an equally rapid rate of technological
advance in the years ahead.
Research and development programs may produce inventions. But
inventions are not synonymous with technological advance. New
inventions result in technological advance only when economic conditions favor their economic exploitation. Schumpeter's theory that
the rate of technological advance fluctuates in a cyclical pattern,
although inventions may appear in a steady stream, is widely subscribed to by economists.
Moreover, technological change does not necessarily result in a
net addition to employment opportunities or to income. In many
cases, the immediate effect of a labor-saving machine is to displace
labor. Only if the surrounding economic conditions are favorable
do the displaced workers find other employment opportunities. Similarly, a new product may simply displace an older product without
increasing the total level of production.
Government expenditures cannot be expected to rise as rapidly in
the future as they have in the past 12 years, and it is not even certain
that they will continue at their present level. National security expenditures, which accounted for 87.8 percent of total Federal purchases of goods and services from 1946 to 1957, are very largely responsible for the increase in Federal expenditures and for their current high level. Successful disarmament negotiations and/or a lessening of international tension, however, would undoubtedly result
in a reduction in national security expenditures. Such developments
do not appear probable in the immediate future. It is to be devoutly
hoped that these developments will occur before very many years
have passed, however. And in a world that is so full of surprises,
they may occur much sooner than any of us dare hope at the moment.
Even in the absence of successful disarmament negotiations and a
lessening of international tension, there is a possibility that national
security expenditures may decline in the years ahead. Modern implements of warfare may prove to be less expensive, in the aggregate,
than the older implements they are displacing, and they may not
become obsolete so quickly.
State and local government expenditures will probably level off in
the near future, and may even decline. Some States are already actively seeking methods of paring their budgets, and with a decline in
the general level of economic activity other State and local governments may be expected to follow suit.

ECONOMIC STABILITY AND GROWTH

The same arguments just marshaled to support the view that we
cannot expect the forces making for economic expansion during the
postwar period to continue operating in the same fashion indefinitely,
may also be cited in questioning the validity of the assumption that
future economic recessions will be relatively infrequent, and also relatively short and mild.
Our experience since the end of World War I I gives us some basis
ior hope on this score, but it does not provide a satisfactory basis for
firm belief. The mildness of the recessions of 1949 and 1954 may
very well be related to the strength of the forces we have just been
discussing. If that is so, as these forces lose their strength we may
once against experience more serious and protracted declines in economic activity.
Eecent experience and our hopes for the future should not blind
us to the fact that our basic economic institutions are still substantially the same as in the 1920's and the 1930's. The role of the capital
.goods sector of our economy is as important as ever, and no foolproof
method of eliminating the inherent instability of this sector of our
•economy, or neutralizing the effects of this instability has yet been
devised and tested.
The Federal Government acted wisely and well to help counter the
Recessions of 1949 and 1954. Those recessions, however, were essentially inventory recessions which are relatively simple to deal with
and which in any case tend to be relatively mild and brief. A sharp
decline in public expenditures or in investment outlays, might produce
an economic decline of a more serious character. We do not yet know
whether the Federal Government can act with the requisite speed and
.effectiveness in the face of such a situation.
Moreover, in any actual situation it is no longer a question of what
tthe political authorities can do but what they will do. Consider the
sharp differences of opinion being expressed about the nature and
scope of the measures which should be adopted to deal with the present
economic recession. When conflicting advice is offered a choice must
be made, and the choice may prove to be faulty. If this occurs, even
a decline which theoretically could be arrested will not be arrested.
J t may instead deteriorate into a genuine depression.
VI

The second assumption of those who propose amending the Employment Act to include price stability as a goal of public policy, is that
effective implementation of the goal of price stability is and always
-will be consistent with the promotion and maintenance of maximum
employment, production, and purchasing power. If ibis assumption
is evaluated on the basis of either history or theory, however, it appears to be of questionable validity. It is generally accepted that
economic growth is necessary for the promotion and maintenance of
maximum employment, production and purchasing power. And the
historical record reveals a remarkably close relation between economic
growth and a rising price level.,
In the past, economic growth has occurred largely during the expanding phase of the business cycle. The tendency of prices to rise

ECONOMIC STABILITY AND GROWTH

during a cyclical expansion has been recognized by virtually all economists. Arthur F. Burns himself writes:
One of the plainest teachings of history is that rising prices are a recurring
feature of the expanding phase of the business cycle.1

Moreover, the work of Simon Kuznets indicates that this relation between rising prices and economic growth is not purely
a short-run
phenomenon: it holds good in the long run as well.2
Theoretical analysis suggests that the association of rising prices
with economic expansion or growth in the past has not been fortuitous,
and that a similar association may be expected in the future. Economic expansion or growth occurs when aggregate demand is strong
in relation to current output, and gives evidence of increasing further. These are the same conditions which provide the basis for rising prices. When aggregate demand is not strong in relation to current output, and when there is no evidence that demand is increasing,
there is no sound basis for rising prices, and at the same time no reason for economic production and capacity to increase.
Moreover, a slowly or moderatively rising price level tends to encourage economic expansion and growth. Since interest rates and
wage rates tend to lag behind other price changes, a slowly or moderately rising price level is,generally associated with rising profits. In
addition, both consumers and business firms tend to satisfy their needs
more promptly and even to anticipate them to some extent when the
price level is rising, so that the increase in prices tends to enhance
demand. Businessmen thus have a double incentive to expand their
output and capacity.
Nothing herein should be construed to mean that a rising price level
is desirable under all circumstances, or that a continuously rising
price level is a necessary requisite for continuous economic growth.
The conclusion suggested is much more modest—namely, it does not
appear likely that price stability will, under all circumstances which
may arise in the future, be consistent with promotion and maintenance of maximum employment, production, and purchasing power.
VII

The assumption that price stability constitutes a desirable goal of
public policy from a technical and administrative point of view is
seldom, if ever, explicitly stated by those who propose that the declaration of policy of the Employment Act be amended to include maintenance of a stable price level as one of the objectives of public policy. But this assumption must be accepted by anyone who seriously
considers the proposed amendment necessary. Here again, however,
past experience and theoretical considerations both give rise to several questions.
Since it does not appear likely that price stability will, under all
circumstances which may arise in the future, be consistent with promotion and maintenance of maximum employment, production, and
purchasing power, if price stability is added to the Employment Act,
policymaking officials will sometimes be faced with the necessity of
choosing between price stability and the other objectives of the act.
1
2

Arthur F. Burns, Prosperity Without Inflation (Fordham University Press, 1957), p. 12.
Simon S. Kuznets, Secular Movements in Production and Prices (Houghton Mifflin Co
1930), pp. 200-2.

ECONOMIC STABILITY AND GROWTH

This will tend to cause confusion and dissension, and may result in
other undesirable consequences as well.
If price stability is always granted priority, this will greatly limit
the flexibility of action which it is desirable and presently possible
for policymaking officials to exercise. It may seriously reduce our
chances of realizing or approximating the optimum growth rate.
If price stability is not always to be granted priority, it does not
really constitute a very satisfactory guide to policymakers without
some indication of the circumstances under which it is to receive
priority or the level at which they should attempt to maintain stability. Would all price rises be regarded as inflationary ? Or would
some attempt be made to distinguish between a price rise accompanying an expansion in economic activity following an economic
recession, and a truly inflationary situation? If so, how would the
distinction be made ? In particular cases, it might well prove necessary to decide when healthy economic expansion ends and when inflation begins. And similarly, would all downward movements in the
price level be regarded as unhealthy? If not, how would the distinction be made between those which should be tolerated and permitted to run their course and those which should be halted or
reversed ?
If the goal of price stability is not to be granted priority except
at the discretion of policymaking officials, or if price movements are
merely to be one of a number of factors considered by policymaking
officials in reaching decisions on public policy, why is it necessary to
amend the Employment Act? Surely price movements are already
seriously studied and taken into account by those responsible for giving effect to the provisions of the Employment Act.
In this connection it is perhaps worth quoting from a paper submitted to the Joint Committee on the Economic Report in 1955 by
Alvin H. Hansen:
A high degree of stability ill the value of money must be an important consideration of public policy. Yet we are * * * in considerable danger of making
a fetish of rigid price stability. This fetish
could easily become a serious obstacle to optimum expansion and growth.3

I t seems pertinent also to raise the question of whether price stability constitutes a feasible goal of public policy. Quite aside from
any of th.6 finer technical points involved here, is our current knowledge of economic theory, techniques of public policy and the manner
in which our economy functions and responds to policy decisions,
sufficiently far advanced so that we can realize the goal of price stability ? If the answer be made that although we cannot achieve complete stability, we may nevertheless achieve reasonable stability, it
merely raises another question. What is reasonable for this purpose ?
If the act does not specify, it must be left to the discretion of policymaking officials. But is that not where the matter stands right now
without any amendment ?
Actually, current developments as well as those of the recent and
more distant past indicate that change or lack of change in the price
level does not always constitute a very useful guide to policymakers
who must decide what course of action to recommend or pursue.
Fedea Tax Policy for Economic Growth and. Stability (U. S. Government Printing
Office, 1955>, p. 17.

ECONOMIC STABILITY AND GROWTH

Prices tend to lag behind changes in underlying economic conditions. Price movements are, therefore, frequently a tardy indicator
of the need for action. They may give warning after the event has
alreacly occurred, or after the time for successful application of restraining or corrective measures has already passed. Considered by
themselves they may even suggest the need for action of a type which
is clearly inappropriate in the light of other economic indicators.
The continuing rise in the Bureau of Labor Statistics cost-of-living
index during the current economic recession is a case in point. The
decline in the BLS Consumer Price Index from 1926 to 1929 to
another.
Similarly, stability of a price index is not always an indication that
all is well, nor does it always augur well for the future. Consider,
for example, the remarkable stability of the BLS Wholesale Price Index from the beginning of 1927 to the autumn of 1929. A price index
may remain stable even if significant fluctuations occur among various components of the index. The relative stability of the BLS Consumer Price Index from 1952 to 1955 was actually the result of a
decrease in prices of many commodities and a simultaneous increase
in prices of services. The relative stability of the Wholesale Price
Index over the same period was actually the result of a decrease in
wholesale prices of farm products and processed foods and a simultaneous increase in prices of other commodities. At times, changes
in the components of a price index are more significant and more
useful as guides to policy makers than relative stability of the total
index.
Consider also the possibility of a general price decline which might
occur as the result of increasing productivity. This is no longer considered a very likely development by most economists. But if it did
occur, should the price decline be accepted as an indication that corrective or restraining measures are desirable?
Actually, of course, in a technical or statistical sense, there is no
such thing as a the general price level." It is well known that although we do have several good indexes which are useful for measuring price movements affecting specific aspects of our economy, we
do not have a satisfactory index of general price movements, nor are
we likely to have one in the foreseeable future. What, then, should
be used as an indicator of general price movements?
Since the different indexes useful for specific purposes sometimes
move in diverse fashion, they cannot all be given equal considerations
in deciding matters of general policy. It is necessary to choose among
them. At present there is considerable support for using the BLS
Consumer Price Index as an indicator of price movements for general policy purposes. Undoubtedly, this proposal has much to commend it from a political point of view. In the past, however, most
economists and statisticians have leaned toward the view that the
Consumer Price Index is less useful than the Wholesale Price Index
for this purpose. This attitude was based on their recognition that
consumer prices tend to lag behind wholesale prices, and that wholesale prices are most basically and intimately related to the economic
process than consumer prices. This jpoint of view is quite as valid now
as it ever was.
These technical and administrative considerations point to the same
conclusion indicated by the arguments presented in the preceding two

ECONOMIC STABILITY AND GROWTH

sections of this paper; namely, that it would be unwise to amend the
Employment Act by adding price stability to the objectives presently
included in the declaration of policy.
vm
It is perhaps worth noting that several times in the past when
congressional action has been proposed to make price stability the
sole or chief goal of the policies and actions of the Federal Reserve
Board, Federal Reserve officials have presented eloquent testimony
in opposition to these proposals, and have made use of some—although not all—of the same arguments cited above.
Even now, despite Chairman William McChesney Martin's public
statements in favor of the proposal that the goal of price stability
be incorporated in the Employment Act, the members of the Board
of Governors apparently are not unanimously agreed on the desirability of this proposal. Not very long ago, Charles N. Shepardson, a
member of the Board of Governors, observed:
As is evident, we have not been completely successful in our efforts to contain
inflationary pressures. But perhaps we should not be too severe on this lack
of perfection. I doubt that perfect price stability can ever be achieved in a
free-enterprise system, or any other system for that matter. Furthermore, I
am not at all sure that it would be wholly desirable. Some upward drift in
prices during periods when demands are pressing against our resources and
some decline following these unusual periods of hyperactivity are not only unavoidable but perform a useful function in helping
to bring about adjustment
of spending and saving decisions in the economy.4

One of the arguments most frequently cited by those who favor
amending the Employment Act by adding price stability to the declaration of policy, is that even a slowly or moderately rising price level
has uneven economic effects and that it has particularly undesirable
effects on retired persons living on past savings, on recipients of
pensions or annuities, and other individuals with fixed incomes.
In arguing against the proposed amendment to the Employment
Act, the writers do not wish to appear unmindful of the validity of
this argument. But the validity of this argument does not weaken
our position. The point of view expressed in this paper is, first of
all, based on the belief that policymaking officials already have ample
basis for adopting measures to restrain undesirable price rises. More
rigid adherence to the goal of price stability might have undesirable
effects on the general level of economic activity and the general level
of economic well-being which would far outweigh the possible beneficial effect on the groups referred to in the preceding paragraph. It
might even have undesirable effects on some of the individuals in these
groups. In a severe economic contraction, individuals sometimes lose
all or a substantial part of their savings very quickly. Business firms
necessarily reduce or suspend dividend payments and sometimes default on contractual interest payments as well. Many incomes which
are customarily regarded as fixed nevertheless do decline during a
protracted depression.
4
See his talk before the National Agricultural Credit Conference in November 21, 1957,.
entitled, "Monetary Policy in Our Economic Climate," p. 11.

ECONOMIC STABILITY AND GROWTH

Rigid adherence to the goal of price stability might thus do more
harm than good even to the group it was intended to help. I t would
seem preferable, therefore, not to rely on price stability for this purpose, but rather to devise new techniques more specifically directed
toward the special problems of this group, which would not involve
the possibility of such widespread undesirable side effects.

II
THE MEASUREMENT OF PRICE CHANGES AND
PRICE RELATIONSHIPS

/ / . The measurement of price changes and price relationships
A. How do changing technology, changing physical characteristics, changing uses of products and services, affect the significance and usefulness of price comparisons between different time periods ?
B. What is the distinction between relative price movements and
changes in "the price level" ?
C. Given today's markets and institutions, what are the identifying characteristics of "administered" compared with "competitive" prices ?
D. What would be the characteristics of a general price index
adequate for economic policy purposes ? Would more than
one index be needed ? If so, why ?
E. When "the price level" enters into decisions about policies to
promote economic stability and growth, which of the available indexes would, in theory, be best as a measure of general
price movements; e. g., the Consumer Price Index, the
Wholesale Price Index, the gross national product deflators,
or other ? How could existing indexes be improved to come
closer to the ideal ?
76

THE MEASUREMENT OF PRICE CHANGES
Kenneth J. Arrow,1 Stanford University
In this paper I will be concerned almost exclusively with measurement of the cost of living, or the Consumer Price Index, as it is now
called. It is this field of price indexes to which economic theory has
made the most important contributions directly. However, I am quite
convinced that there are no great differences in principle as far as
other price indexes, such as the Wholesale Price Index, are concerned.
All commodities in a sense are intended ultimately for consumption
and therefore all prices reflect indirectly the valuations placed on them
by ultimate consumers (including, in many cases, investors).
Stress will be laid on the theoretical contributions. Recommendations as to sources of data and representativeness of sampling can
hardly be done effectively by one outside the Government agencies
responsible.
I. BASIC CONSIDERATIONS

A consumers' price index is a measure for one period of time, say
1, with respect to another period, say 0. It is generally defined as the
ratio of the expenditures needed in time 1 to maintain a given standard of living, to the expenditures needed in time 0 for the same purpose. The standard of living for this purpose clearly cannot be identified with a fixed basket of commodities. Suppose, for example, that
in the base period 0, the price of beef is high, while that of lamb is
low, while the reverse is true in period 1 due to, say, a change in supply conditions. We would expect the consumer will purchase a great
deal of lamb and relatively little beef in period 0, and the reverse in
period 1. If we use as a fixed basket of goods that consumed in the
base period, then we will find a rise in the cost of living because the
lamb, which has appeared so strongly in period 0, has undergone a
price rise. On the other hand, the reverse would be true if we started
with period 1 as the base period.
There seems no recourse but to recognize frankly that a standard
of living is not any fixed basket of goods, but a subjective level of satisfaction. In the example just given, it may well be that the consumer
feels about as well off in one period as another since he has had a
chance to compensate for the price changes by changes in the proportions in which different commodities are consumed. The Government
statistician, for obvious reasons, shies away from the notion of tying
the objective-seeming price index to subjective concepts of utility or
satisfaction, but in fact there is no escape in this proposition. Any
attempt to explain the meaning of the cost-of-living index must eventually come to the notions just described unless we are to stop at banal
lf
This paper was prepared during the tenure of a Ford Foundation faculty research
fellowship. I wish to thank also the Office of Naval Research, whose grant permitted
research and secretarial assistance.

ECONOMIC STABILITY AND GROWTH

tautologies analogous to "national income is that which is measured
by national income statisticians."
It is, of course, perfectly true that we are not in the position now,
and we may never be, to measure the subjective satisfactions of consumers directly. However, as the previous example already hints,
economic theory argues that something, at least, can be learned by
studying the overt behavior of the individual on the market. The
reason is that the individual is motivated to secure for himself the
highest utility consistent with his monetary resources. His utilities
or satisfactions then are already reflected in market behavior and it is
not implausible, therefore, that we may use this behavior to cast some
light upon his satisfactions and in particular upon the cost-of-livingindex as defined previously. Unfortunately, as is well known, the inference from quantities and prices observed to the true cost-of-living
index is by no means a simple one.
Let us consider the simplest inference, a very well-known one.
Suppose we wish to know the cost-of-living index defined as the minimum expenditure in period 1 needed to obtain the level of satisfaction
of period 0 as a ratio to the expenditure in period 0. One way to make
sure to achieve the same level of satisfaction in period 1 as in period 0
is to purchase the same goods. This will, in general, not be the cheapest
way of achieving that level of satisfaction. Let p 0 and pi denote
prices in periods 0 and 1, respectively, and let q0 represent quantities
in period 0. Then the above reasoning shows that—

will be at least as great and probably greater than the true cost-of Iiving index.
We thus see that price-quantity figures permit some inference as to
the true cost-of-living index. In this case, they set an upper bound,
but we also see that a simple set of figures does not give a complete
determination. In particular, the formula used is a base-year weighted
price index (frequently referred to as Laspeyres' formula); that is,
it is precisely the formula which is used in most ordinary statistical
work. We see, then, that this formula tends to overstate the rise in the
cost of living from any fixed base. It is indirectly the realization of
this fact which causes the base to be revised as frequently as it is.
In the following sections we will suggest the use of additional
information other than that used in the formula just given to narrow
down the limits which we can place on the true cost-of-living index
by inference from price-quantity observations.
n . THE USE OF ENGEL CURVES FOR PRICE INDEXES

The definition of the cost-of-living index number suggests immediately that it really should not be regarded as a single number. The
cost-of-living index number corresponding to one level of satisfaction
in a base period may be very different from that in another. This
shows up conspicuously when the differences are large, as between poor
and rich. The consumption pattern of the rich is quite different from
that of the poor and a, shift in prices which increases the cost of living
to one may decrease it to another. Thus, if servants' wages rise while

ECONOMIC STABILITY AND GROWTH

the prices of manufactured goods fall, it may well be that the expenditure needed by a wealthy person to maintain his standard of living;
i. e.? his level of satisfaction, will be going up, while that of the poor
person is going down. This fact is implicitly recognized in index
numbers which restrict their announced coverage to "middle-income
families," but it argues that quite a bit more can be done along these
lines. There should be a separate cost-of-living index number for each
income level.
Empirical data on which any such index numbers are stressed are
found in budget studies. For those, we have 2 years, 0 and 1, in
which the budgets of individuals over a wide range of incomes have
been made available. For each commodity, we can find the average
consumption in each income class. The relation between consumption
of a particular commodity and income is known as an Engel curve,
after the German statistician of the last century, who pioneered in this
area. We assume, then, that we have an Engel curve for each commodity in each of the 2 years under consideration. We now construct
two new curves which represent, at each income level, the purchases
of commodities in 1 year evaluated at the prices of the other. Thus,
if Ei represents an expenditure level in year 1, take the average
amounts consumed of each commodity by those with expenditure level
Ei, and calculate what that bundle would have cost in the prices of
year 0. In symbols, if qi(Ei) represents quantities purchased in
year 1, when total expenditures are E a , then in the notation previously
used, we calculate
which will be designated as E 1 0 (Ei). A similar calculation can be
made interchanging the roles of the years 0 and 1. Thus, we have
a second curve E O i(E o ). For any expenditure level E o in year 0, let
q o (E o ) be the commodities purchased by individuals whose total
expenditure was E o . Then
EOi (Eo) =

We will plot the two curves on the following diagram, figure 1 (p. 87).
The horizontal axis represents expenditures in year 0, the vertical axis
those in year 1. The curve E O i(E o ) will be plotted against the horizontal axis. The curve E 10 (Ei) will be plotted against the vertical
axis.
The problem of finding the true cost of living is that of matching
tip expenditures E o and E x so that for each value Ep we find the
corresponding minimum expenditure E x which will yield the same
level of satisfaction. We want then a curve relating the two variables.
To read a cost of living index number for any given level of satisfaction on this curve, we may take a particular value of E o and find
the corresponding value of E x ; the ratio of these two values is the
cost of living index corresponding to an expenditure level of E o in
the base period. The fact that the curve will not usually be a straight
line to the origin shows that the cost of living index number will in
general vary by income levels.
The question is the determination, or at least the approximation,
of this expenditure-equivalence curve- The reasoning of the preceding section shows that for any E o , E 0 i(E o ) is an overstatement of the

ECONOMIC STABILITY AND GROWTH

equivalent expenditure in year 1. Reversal of the argument shows
that for any expenditure E x in year 1, the number Eio(E;i) is an overstatement of the equivalent expenditure in year 0. Therefore we
can say without any further discussion that the expenditureequivalence curve must lie between the two curves E O i(E o ) and
E K ^ E X ) . For a verbal reference, we will refer to the first of these
as the current-year weighted base-year expenditure curve and the
base-year weighted current-year expenditure curve.
It is already clear from this discussion how much more information is contained about the cost of living index by the knowledge of
the two sets of Engel curves than from simply knowing some kind
of national average. The latter at best amounts to knowing one point
on each of the two expenditure curves. While there is some information contained in this statement since the expenditure-equivalence
curve would have to lie between the two points, it is clear that the
restrictions on the location of the curve are not very severe.2
In many cases the two expenditure curves will be very close together
so that the expenditure-equivalence curve will be determined quite
accurately. An approximation formula has been developed for the
interpolation
of expenditure-equivalence curve by the late Abraham
Wald.3 If we refer to figure 1, we can take any base year expenditure
E o and find from the curve E K ^ E J ) that expenditure in year 1 for
which the base-year weighted expenditures are equal to E o . In the diagram this value is represented by Ei O *(E o ). It is clear from the diagram that this value is an underestimate of the equivalent expenditure
m year 1. We have already noted that E O i(E o ) is an overestimate
of the equivalent expenditures. It is natural then to seek for a method
of averaging the two. The particular average suggested by Wald is
the following: Let aOi be the slope of the upper curve at point B,
and let a* (Oi) be the slope of the lower curve at A. In both cases the
slope is found with respect to the E0-axis, that is, the way the curves
are drawn now. Then Wald's formula is

In the present state of theory this seems about as satisfactory a formula as can be found.
I t will, of course, be objected that obtaining Engel curves from annual budget studies is an expensive operation. Here, as elsewhere,
the rule that you don't get something for nothing is applicable. It
seems clear that higher accuracy in the index numbers require greater
expenditure by the Government. The value received in terms of ability to make better economic plans will unquestionably repay additional expenditures many fold over.
This is especially true in the present context because annual budget
studies have many other uses than the improvement of index numbers.
(Changing consumption patterns will be detected at a much earlier
a
The importance of Engel curves for a more accurate determination of price index numr
bers was first pointed out by Ragnar Frisch. For a very clear exposition at greater length
than the above, see his paper, Some Basic Principles of Price of Living Measurements,
Econometrica, vol. 22, No. 4, October 1954, pp. 407-421.
9 Wald's results were first published in A New Formula for the Index of Cost of Living,
Econometrica, vol. 7, No. 4, October 1939, pp. 319-331. For a simpler exposition, see
K. S. Banerjee, Simplification of the Derivation of Wald's Formula for the Cost of Living
Index, Econometrica, vol. 24, No. 4, July 195-6, pp. 296-298.

ECONOMIC STABILITY AND GROWTH

stage and will find a great deal of use both in private industry and in
many Government policies, such as the prediction of revenues from
excise taxes; the consumption weights used in national income calculations can be greatly improved.
I would further urge that if budget studies were put on an annual
basis provisions be made that a major proportion of the families be
reinterviewed annually. The value of panel studies—that is, repeated studies on the same individuals has been amply demonstrated
in recent work on consumption patterns as well as on sociological
phenomena. The effect of income changes and changing age and
family composition on consumption can be studied in this way as in
no other. The variability of income and expenditures over time will
form an important tool in the study of income distributions. There
are thus many side benefits to annual budget studies and their analysis
in terms of Engel curves, so that one can easily justify the additional
expenditure.
m . MULTIPLE TIME COMPARISONS

The preceding section has dealt with comparisons between 2 points
of time, say 2 successive years. However, one of the main functions
of an index number is to provide a time series for the cost of living.
We would like to be able, in fact, to make all sorts of comparisons
including, ideally, comparisons of the cost of living in periods quite
remote in time. To begin with, we will consider only the single comparisons based on national averages which are currently used, and
abandon the Engel-curve approach of the last section.
It has been seen to begin with that the use of quantity weights as of
a fixed time period tends to overstate the increase in the cost of living
in a subsequent time period. Thus, if 1947 rates are used, the price
index of 1955 will be higher with respect to 1947 than it should be.
The same will be true of the price index of 1956. The indexes are
also used to compare 1955 with 1956. Here it is not clear what the
bias will be but it is clear that the 1947 quantity rates may or may not
be very useful ones for the 1955-56 comparison.
There is a considerable divergence of viewpoint here between the
practice of statisticians of almost all governments and the views of
index number theorists. In comparisons involving distant points of
time, index number theorists generally take 1 of 2 positions: Either
they argue the comparison between any 2 years should be based solely
on the data for those 2 years and not on any others, or they advocate
a chain index. In a chain index comparisons are made only for adjacent years directly. The resulting index numbers for pairs of
years are then multiplied together in an obvious way to get index
number comparisons between years which are not consecutive. Thus,
an index number for 1949 with respect to 1947 would be obtained by
multiplying the index for 1948 with respect to 1947 by the index for
1949 with respect to 1948.
It is not possible to give a definitive argument on the relative merits
of chain index numbers and the more conventional fixed weight index
numbers, primarily because neither is a thoroughly accurate solution. Nevertheless, there are some considerations that suggest that
as between the two, the chain index number is preferable.

ECONOMIC STABILITY AND GROWTH

First, it is clear from the preceding discussion that the difficulty
with index number comparisons arises because of the inadequate
knowledge of the want structure of the individual. An increase in
the number of observations made on market behavior would be expected to increase the possible accuracy of index numbers.4 The upward bias in base-year-weighted index numbers illustrates the inevitable errors due to the use of a limited number of observations—
in this case, a pair. A chain index number between 2 points of time
separated, say, by 5 or 6 years, makes use of all the intermediate
observations.
Second, the difficulties and possible errors of index number comparisons are increased when the consumption patterns in the two
periods being compared are farther apart. Economic magnitudes
tend to change continuously. Comparisons of adjacent years, then,
are comparisons between more nearly homogeneous universes. It
is more legitimate to use various approximations such as averaging
of quantity weights. Indeed, the weights based on quantities in 2
adjacent years will differ less than they will in years taken further
apart, as a rule. Therefore, each link in the chain index number will
be more accurate than direct comparisons at some distance apart.
That these considerations play some role in practice is shown by the
relatively frequently changes in base years employed in practice. If
the base of an index number is changed relatively frequently, there is
in effect a chain index number, calculated perhaps on a quinquennial
rather than annual basis. Once the need for changing bases is recognized, it becomes merely a question of discussing the optimal frequency. An attempt at a logical foundation
for chain
indexes has been given by Frangois Divisia.5 If we write
down the statement that total expenditures equals the sum of
expenditures on individual items, where expenditures on any
one item is equal to the product in price and quantity, E=Spq,
then during a small period of time the change in the rate of expenditures can be expressed approximately in the following way:
AE=2pAq+SqAp. Here the symbol A means "change in." This
formula assumes that in a small period of time prices and quantities
can only change by small amounts. The first term represents that
part of the change in expenditures attributable to a change in quantities evaluated at the original prices and can therefore be thought of as
the change in a certain quantity index. Similarly, the second term can
be thought of as attributable to the change in prices using quantities
as weights. If we imagine that we have both a price index and a
quantity index, it is natural to demand that their product be equal
to total expenditures E = PQ and therefore a small change in expenditures will be approximately expressed by A E = P A Q + Q A P . Comparison of the two formulas for AE suggests that the right-hand sides
can be identified term for term. Carrying out in detail the reasoning
suggested here shows that the price index is a chain index with quantity
weights. In each length of the chain, if the links are sufficiently short,
* For a more extended discussion of the possibilities and limits of improving index number comparisons by increased observations, see Paul A. Samuelson, Foundations of Economic
Analysis, Cambridge, Mass.: Harvard University Press, 1947, pp. 146-163.
5
For a description of Divisia's viewpoint, see Erland v. Hofsten, Price Indexes and
Quality Changes, Stockholm: Bokforlaget Forum, and London: George Allen and Unwin.
1952, pp. 21-23.

ECONOMIC STABILITY AND GROWTH

the base-year and current-year quantities will be very similar so that
it makes little difference which is used. Ideally, actually, the links
should be very short in time indeed but there are some difficulties with
seasonal fluctuations which suggest that it is unwise to use periods
shorter than 1 year.
Divisia's reasoning is plausible but is not closely tied to the definition
of a cost-of-living index in terms of expenditures needed to maintain a
given level of satisfaction. The connection between the 6two concepts
has been investigated to a certain extent by Jean Ville. In general
he shows that the chain index does not give exactly the true cost of
living index except in the special case where at any fixed set of prices
an individual will divide his income among different commodities in
the same proportions, regardless of the level of his income. That is,
the Engel curves are straight lines through the origin. This case is, of
course, unrealistic although it may be approximately valid for relatively small changes in real income.
The reasons why the chain index falls short of perfection can be put
in different ways. One thing to note is if additional observations are
designed to get a better knowledge of the want structure of the individual, then there is no reason to confine oneself to observations in
temporal order. All information within a time period in which
wants can be regarded as homogeneous should be equally relevant to
a comparison between any two time periods. Another point is that
if the level of satisfaction of the population is rising as is usual, then
the comparison between successive years referred to different levels of
satisfactions. After a period of time, then, the chaining involves
price changes which are irrelevant to different standards of living.
The special case, which Ville discussed, is one in which the consumption pattern is the same at all levels of satisfaction and so no ambiguity
arises. One form in which these imperfections may show up is the
following paradox: It can happen that prices vary over time in such
a way that after a few years, let us say, they return to their initial
values^ Obviously in that case, any properly defined price index
should be equal to 100 as between the. beginning and end years. Yet
it is possible for a chain index number to differ from 100, either above
or balow. If, for example, as is usual in seasonal fluctuations, prices
tend to be high when quantities are low, and vice 7versa, then in general
the chain index will tend to give too high a value.
Too much should not be made, however, of paradoxes like this.
Virtually any index number known can lead to similar difficulties in
suitably unfavorable circumstances.
On balance, the case for chain index numbers as opposed to fixed
weight aggregates seems strong in spite of the drawbacks just noted.8
However, the case is considerably strengthened if we combine the
chain index of this section with the Engel curve methods of the last.
Since each pairwise Engel curve comparison gives us a complete set
of equivalences between expenditures in one year and expenditures in
6
See his paper, "Sur les conditions d'existence (Tune ophelimite totale et d'un indice
dn niveau des prix, Annales de l'l'niversite de Lyon, sec. A (3), vol. 0 (194fi), pp. 32-39,
particularly pp. 36-38: translated into English as The Existence-Conditions of a Total
Utility
tility Function, Review of Economic Studies, vol. XIX (1051-52), pp. 123-128, particularly pp.
p. 125-127.
»TMss was pointed out by Ladislaus von Bortkiewicz; see v. Hofsten, op. cit., pp. 1415,8 27.
This position is also taken by Frisch, op. cit., p. 417, and Bruce D. Mudgett, Index
Numbers, New York: John Wiley & Sons, and London: Chapman and Hall, 1951, pp. 70-79,
23734—58
7

ECONOMIC STABILITY AND GROWTH

the next, the chaining is quite straightforward. Given any expenditure level in, say, 1947, we can find expenditure in 1948 which gives
approximately the same level of satisfaction. Starting with expenditure levels for 1948, we can find the corresponding expenditure levels
for 1949. By linking the two pieces of information we can find an
equivalence between expenditure levels in 1947 and 1949. This is still
a chain comparison subject to the disabilities noted above. It finds
its primary justification in the assumption that neighboring comparisons are apt to be more accurate than those at a distance. In terms
of figure 1 the two bounding curves are apt to be much closer together
than they are for comparisons of several years apart.
A more experimental approach to index numbers based on multiple
comparisons of time has been
originated in a paper by Lawrence
R. Klein and Herman Rubin.9 If we have observations for a number
of years, we have some possibility of estimating the effect of price
changes on the consumption of different commodities. One can hope,
at least, to obtain demand functions which express the consumption
of any commodity in terms of the prices of that commodity and competitive ones, and the income of the country or of the individual if
we have observations on separate individuals. If the demand functions are accurately obtained, it is possible to infer the underlying
want structure and therewith to make all desired cost of living comparisons. The problem resides in the difficulty in accurately determining statistical demand functions; nevertheless a great deal of progress
has been made on this in recent years. It would seem worth while to
experiment with this approach in different w^ays and see how it compares with other, more traditional methods. In principle, it is certainly the most satisfactory since, for example, all observations are
treated equally thereby using more information.10
IV. THE PROBLEM OF QUALITY CHANGES

One of the greatest problems in any index number comparisons over
time is that of changes in the quality of existing commodities or the
introduction of new ones. A thorough theoretical analysis of the subject is found in the book of Erland von11Hof sten who has been in charge
of the Swedish cost of living index.
We will discuss the problem
here mainly from the Engel curve viewpoint of section II.
It should be made clear that we are discussing this problem at a
theoretical level where it is assumed that we have a price for every
conceivable commodity, including every variety. We ignore here the
9
A Constant-Utility Index of the Cost of Living, Review of Economic Studies, vol. 15,
1948,
pp. 84-87.
10
Klein and Rubin analyzed the particular case where the demand functions are such
that the expenditure for any commodity is a linear function of prices and income. They
obtain, therefore, a specific formula for the cost of living index number. Paul Samuelson
has objected that the linear demand functions are not likely to be found in practice; see
Some Implications of Linearity, Review of Economic Studies, vol. 15, 1948, pp. 88-90.
However, linear functions have been fitted to British data by J. R. N. Stone; see Linear
Expenditure Systems and Demand Analysis : An Application to the Pattern of British
Demand, Economic Journal, vol. LXIV, No. 255>, September 1954, pp. 511—527. For an
excellent exposition, see Ragnar Frisch, Linear Expenditure Functions, Econometrica,
vol. 22, No. 4, October 1954, pp. 505-510.
Klein and Rubin assumed that the demand functions needed for their index would be
obtained statistically by fitting to aggregate data. If Engel curves were available for each
year, the chances of reliable fits would be much improved. In any case, it should be made
clear that any statistically derived demand functions which satisfy certain consistency
conditions can be used to derive index numbers ; they need not be of the linear form
assumed
by Klein and Rubin.
11
For reference, see footnote 5.

ECONOMIC STABILITY AND GROWTH

question of sampling in groups of related goods at a representative
price. The splicing method frequently used in discussing quality
changes depends in part on a sampling concept as well as a purely
theoretical one. The difficulties attached12 to the sampling process
have been ably pointed out by von Hof sten.
At this level of abstraction there is no logical difference between a different variety of the
same commodity and a different commodity. We will assume our
classification as fine as called for by the circumstances so that, for
example, two different models of Cadillacs are to be regarded as separate commodities.
Suppose then we observe, in comparing year 0 with year 1, that a
commodity appears in year 1 which was not consumed at all in year 0.
Actually, from the Engel curve point of view, the problem might arise
at each expenditure level. We may find a commodity purchased only
by upper income people in year 0 which is now purchased in
year 1 by lower income individuals. We will postulate that the want
structure of individuals is the same in the two situations in spite of
the introduction of a new commodity. The absence of a commodity in
the former period will be explained by the hypothesis that the price
at which it could have been produced was so high that the demand for
it would be 0. The problem comes in evaluating the lower boundary
curve of figure 1, Ei O (Ei). Some of the current year expenditures
will be of the new commodity. When reevaluated at base-year prices,
what price shall be assigned to it? The theoretically correct answer
will be the lowest price which will keep every individual from purchasing the commodity. Unfortunately, this price, unlike the others
that have been used to this point, is a hypothetical one, not an actual
one, and its use introduces a hypothetical element into the calculations.
However, I believe that any rule which will accomplish the end of
accounting for quality changes in a price index must involve judgment
somewhere and it is deceptive to state an objective-sounding rule
which is not based on a logical theoretical foundation.
It will be useful to distinguish between several situations. One is
that in which a new commodity rather different from any now existing
is being introduced-—e. g., automobiles around the turn of the century,
or television sets more recently. Here there is apt to be a continuous
rise in consumption. The first year in which the commodity is introduced the consumption is probably rather small. Therefore any error
committed in attributing to the previous year a hypothetical price
will not have a great effect on the base year weighted current year
expenditure curve. Here again we have an example of the value of
a chain index. There is no point at which the introduction of a new
commodity will produce great problems providing the introduction is
gradual. Once it has appeared as an item of expenditure, the successive future steps of its growth are accounted for in the price index,
just as any other commodity.
An even more favorable situation is that in which the newly introduced commodity is a close substitute for one previously existing, say
an improved model. If both are available in the second period, then
it is reasonable to postulate that the consumer would reject completely
1 or the other of the 2 goods if the price ratios differed very much
from that which actually prevailed in year 1. Therefore, a hypo» Op. cit., pp. 53-58, 71-72.

ECONOMIC STABILITY AND GROWTH

thetical price for a newly introduced variety in year 0 is so chosen
that the ratio of its price to the price of variety in existence in both
periods is slightly higher in year 0 than in year 1. This approach
to the pricing of new varieties is very similar to the splicing method.
However, it meets von Hofsten's strictures so long as all the varieties
are counted each time. The difficulty encountered in the splicing
method arises because the particular varieties selected may be
unrepresentative.
A more difficult case is that in which one variety disappears and is
in effect replaced by another. The most common instance of this, is
the change of models in many consumers' durable goods, notably autoifiobiles. The principle is not altered. However, it is now necessary
to have a hypothetical price in both situations. If model A was produced in year 0 and replaced by a more or less equivalent model B in
year 1, the index maker must hypothesize a price for model B in year
0 just high enough so that none of it would have been consumed even
if it were available. Similarly, he must hypothesize a price for model
A in year 1 just high enough, again, to insure that there would have
been no consumption even if available. At this point, the index
maker will probably start to resort to objective measures of quality of
gome kind, such as performance or durability characteristics of the
object. He will, in effect, be postulating that the consumer would
choose between two varieties if both were available, according as the
price ratio exceeds the quality ratio or not. Of course, the only true
measure of quality is the satisfaction yielded to the consumer, and the
quality ratio used by the index maker must be related to his guess as
to the consumer's tastes.
The entire argument to this point has been based on the assumption
of an unchanged want structure between one period and the next.
We have gone so far as to impute tastes for commodities not available
in one time period or the other in order to preserve this theoretical
foundation. However, as von Hofsten stresses, this point of view
cannot be maintained for goods subject to style changes. As we are
currently observing with regard to women's dresses, it is possible for
styles to change from year to year and return to their starting point.
If all prices had remained unchanged during this period, then the
change in varieties <^an only be explained by the assumption of a
change in tastes. There seems no simple way out of this problem
except a judgment by the statistician that the new style performs,an
equivalent function in terms of satisfaction to the old one.
V. CONCLUDING REMARKS

Index numbers are, of course, desired for other purposes than to
measure the cost of living. One obvious possibility is to considersome subset of cost-of-living items, such as food. The logic of the
preceding argument goes through precisely provided that we assume
that the distribution of food expenditures in any period among different foods depends only on the total volume of food expenditures and
is independent of the prices of other goods, for any given total volume
of foorl expendihirps. This does not deny substitution between foods
and other commodities, but we assume that the total effect of this
substitution is already reflected in the choice of a volume-of food
expenditures. In a broad way, similar considerations apply to the

ECONOMIC STABILITY AND

GROWTH

pricing of producers' goods which should be interpreted as reflecting
indirectly consumers' preferences. However, there is undoubtedly alot more in the detailed working out of the theory that has never been
developed.
This leads to the final suggestion that considerable effort be put
into pure research on the theoretical problems of index number construction. This has to be done, of course, in close context with practical problems and, therefore, through the existing statistical agencies.
Either there should be provision for a research unit within existing
statistical agencies, or arrangements should be made for contract
research by universities under the supervision of the Bureau of Labor
Statistics and similar agencies. For real progress, a good deal of
freedom must be granted. The possibility of experimental construction of index numbers must be allowed a wide scope. In this research,
issues of comparability with the past should not be allowed to dominate too strongly. The most important thing is the collection of the
data necessary for price measurements. Even though such data were
not available in the past, we should at least now plan for the future an
adequate amount of information.

.0<E0>

Figure t

(

oi<Eo}

ExpenditureEquivalence

E|O(E.)
l0
'

ADMINISTERED PRICES IN THE AMERICAN ECONOMY
Martin J. Bailey, University of Chicago
The principal aim of this paper is to examine afresh the issues,
both theoretical and practical, involved in the concept of administered
prices. Because of the limitation of time and resources, I have had
to stick to two limited objectives in preparing the paper: To survey
and summarize the conclusions that can be drawn from other people's
work in this field, and to make a very modest additional contribution to the facts and theory that have been brought to bear on it.
Some of the relevant contributions in this field, including all the
most important ones, are listed in the bibliography at the end of this
paper.
The factual research carried out by the authors listed provides us
with a very considerable amount of information about price behavior
and the factors affecting it; and the theoretical analysis given by
them says about everything that can be said in this area.
The factual evidence which has accumulated has not been sufficient,
however, to produce general agreement about what the facts are
about administered prices, nor has this combined with the analysis
presented been able to produce agreement about their practical importance. In general, opinion is divided sharply into two groups:
Those who think administered prices or monopoly prices are widespread and extremely important, and those who think such prices are
nonexistent or of no great importance, except in public utilities and
other enterprises regulated by government. My survey of the opinions and evidencec has led me to find myself more and more firmly
set in the second group. It is my opinion that the subject of administered prices in the free or unregulated part of the economy is not of
itself a (proper concern of public policy nor a subject worthy of the
attention of the Congress. That monopoly as such is a proper concern
of public policy, and that the impact of existing public policies on
monopoly and on related matters is of great importance to everyone, I
am firmly convinced; I do not want anything I say in this paper to
convey any impression to the contrary. However, the proper focus
of attention in these matters should be on measures to insure and
promote competitive marketing of goods and services and the efficient
use of resources. Some existing public policies which do not promote
these ends are in need of reappraisal and reform.
One point on which nearly all observers agree is that there are
wide areas in the economy in which prices are administered, but where
this is no cause for concern. For example, virtually all retail trade
is of this character. It is agreed that retailers' margins are essentially competitive and flexible, although from day to day the retailer
has a wide range of discretion within which he could set his prices
without immediate drastic shifts in sales. As many consumers are
very price conscious, however, very large shifts in a retailer's sales
89

ECONOMIC STABILITY AND GROWTH

would definitely take place after a moderate delay if his prices were
substantially out of line either way. From a long-run standpoint,
therefore, he has practically no discretion about the prices he can
charge. Hence, it is generally agreed, this area of administered prices
should not b3 a source of public concern.
Another important aspect of administered prices is that a misleading or superficial appraisal is painfully easy to give, even by the most
careful and well-intentioned observer, because of the shortage of reliable and unambiguous information. Most of the work that has been
done in this field has had to use 1 of 2 sources of price and related
data: the wholesale price quotations published by the Bureau of Labor
Statistics, and the unit value or realized price data obtained from
the census of manufactures. Results from both these sources have
quite justly been criticized as inconclusive (and the critics have therefore felt free to write their own tickets). The BLS quotations usually
relate to narrowly defined specific commodities, which is as it should
be; but they refer to quoted prices, without cognizance of discounts,
freight absorption, alteration of quality, and other devices for changing the price actually charged for a standard (unchanging) commodity. The census of manufactures data, which avoid this defect by
giving information about net unit values received by sellers, unfortunately have such broad commodity and industry categories that
changes in the product mix of a serious magnitude may occur unnoticed within these categories and entrap the investigator in false
inferences.
Although these defects in the available data on past years are difficult to correct, the situation for the future is more favorable. Problems of this nature in the construction of the Consumer Price Index
have already to some extent been dealt with effectively. For example,
in the pricing of new cars, where trade-in allowances make a mockery
of the list-price schedule, the BLS has worked out a successful surveying technique for dealing with the problem. A modest appropriation by the Congress could make possible the extension of this good
work, including the development of a corresponding set of techniques
for getting realistic wholesale price data. At the same time, research
by university people, and others outside the Government, into enterprise sales data and other sources can bring great improvements in
our information about the past. Some examples in respect to this latter nos^bility will be presented later in this paper.
Keeping firmly in mind the weaknesses of the available factual data,
which carry with them subtle problems of interpretation and analysis,
we may proceed to a summary of what is known and certain about
administered prices in relation to competition, monopoly, and the concentration of industry. At certain points we shall have to conclude
that the only thins: that is known and certain is that no conclusion enn
properly be reached, but even that is better than nothing, especially
in an area where unsupported assertions are frenuently encountered.
Insofar as possible, I shall try to make it clear in every ea^e why a
definUp, conclusion may or may not be drawn from the information
available.
1. Tf a seller actually sells at an administered price; i. e., a price
which he may keep unchanged for weeks or even months at a time in
the face of changing market conditions, he possesses a degree of mo-

ECONOMIC STABILITY AND GROWTH

nopoly power. What economists mean by a degree of monopoly power
is that a seller possessing it may raise his price noticeably above the
prices at which he does a satisfactory volume of business without a
complete or disastrous loss of sales, and may lower it noticeably without thereby making possible an unlimited or enormous increase in sales.
If changing market conditions do not force a seller to change his price
to avoid intolerable fluctuations in the volume of his sales, this by
itself means that he possesses a degree of monopoly power.
Gardiner Means, the best known writer on this subject, likes to
stress that administered prices may occur in markets that are essentially competitive, that is in which the degree of monopoly possessed
by the sellers is not sufficient to imply excessive profits or an injury
to the public due to continuous overpricing. In this he helps to emphasize the fact that what economists generally refer to as a degree
of monopoly power does not correspond to the notion of monopoly
that implies gouging of the public. Nevertheless, as Means also emphasizes, even a very limited and apparently harmless degree of
monopoly power, if widespread throughout some sector or sectors of
the economy, could conceivably involve a disastrous contraction of
output at unchanging prices, rather than the maintenance of output
at sharply falling prices, in the face of a contraction in aggregate
demand. This might seem to imply that public policy could and
should be directed at making even essentially competitive prices less
administered and more competitive; Means leans somewhat to this
view himself, but qualifies it by saying that administered prices are
an inevitable part of the modern industrial economy. Others, such
as Galbraith, say instead that because of this prices should be made
more administered, by subjecting price changes by big business to
prior review by congressional hearings.
Two points remain extremely unclear from what has been said so
far. In the first place, no one has demonstrated that "administered
prices" are in fact widespread or sticky enough to be a matter for
public concern. In the second place, it is not clear or generally
agreed whether administered prices in fact involve overpricing of
an economically significant quantity of goods, to the detriment of the
public, in a way that cannot be corrected by vigorous application of
the antitrust laws and related measures. Further, in connection with
the first of these points, it remains to be answered whether a tendency
for outputs instead of prices to fall in the face of falling demand has
any special implication for a general national policy directed at maintaining output and employment.
2. Before these major points are considered, one thing must be emphasized : Quoted prices, or list prices, are frequently not the prices
actually charged. As was already mentioned, the price that matters
to both the buyer and the seller is the net price after discounts, allowances, and so on. These discounts and so on may fluctuate from
day to day, or even from hour to hour, although the list quotation
remains unchanged for weeks or months; and as they change, so
changes the net realized price.
Suppose it were true that every observable case of an "administered," or sticky price, was of this kind, where sharp price competition between firms caused net realized prices to fluctuate from day
to day without the outside observer's being able to see it because of the

ECONOMIC STABILITY AND GROWTH

unchanging list price. Then we would in fact have all the beneficial
and desirable characteristics of competition, and any reasoning based
on the apparently harmful sticky prices would be wrong and misleading. In particular, the whole structure of Means' treatment of this
subject, which has as its foundation the relative infrequency with
which many of the prices quoted by the BLS change, would fall to the
ground if it could be shown that the actual prices charged, masked
by these quotations, change from day to day and from week to week
according to the state of the market.
As a matter of fact, it cannot be shown conclusively that all or most
of these quotations are nominal or that they generally fail to reflect
the true frequency of price change. At this point I can only say that
a very large question mark has to be placed alongside all the work
on so-called administered prices that explicitly or implicitly uses the
frequency of changes in quoted prices as a criterion of the sensitivity
of these prices to market forces. This question mark will get larger
and larger as I go on to consider some important examples of such
prices on which definite and conclusive information is available.
3. The theory of oligopoly, industrial concentration, and related
ideas are seriously defective, and provide us with no basis for judging
whether prices are as senstive as they should be. Oligopoly theory,
and related ideas about enterprises occupying a large share of the
market for some product, have gained a wide degree of popular acceptance because their logic looks good and their conclusions accord
with our natural and to some extent justifiable suspicions of big business. This, however, does not tell us anything about the facts of the
case, nor does it by itself even assure us that the logic of the theory
itself is adequate or air tight. It is not.
The prevailing theory of oligopoly states that a large seller will
hesitate to cut his price when he knows that if he does other sellers
will follow suit, because he is so large that his action cannot go unnoticed. If all the output of the product they sell is produced by a
few large firms, this would seem to imply that no one will cut the
price unless demand has fallen to such disastrously low levels that
price cutting is obviously preferable even when everyone does it together. The gap in this reasoning is that the circle is not complete
unless it can be asserted that every firm in the industry is very confident that if it lets customers get out the door because it refused a
price concession, not a single other firm will make the concession and
make off with the trade. If a firm has serious doubts on this score,
as well it might, and if the current sales and price situation is such
that the firm is very anxious to secure the additional trade, it will be
sorely tempted to offer the concession. The effect of "oligopoly" will
then be that the firm will request that the customer not shout from
the housetops about the concession given.
The temptation to cut prices covertly in this way will occur any
time that the extra costs incurred to produce additional output are
substantially less than the going price. Even if the difference between the two does not yield as much as the producer would like toward the overhead, it will come closer to doing so if he can expand
volume at the expense of his competitors by a small price concession;
and the operating profit will most certainly fall much further short
of covering overhead if his competitors expand their volume at his
expense. In a situation of this kind, it is obvious to everyone con-

ECONOMIC STABILITY AND GROWTH

cerned that a price well above marginal cost is unnatural and difficult
to maintain.1 All firms are not of equal size, and smaller ones are less
likely to try to keep the price up than are the big ones when this
situation affords them the opportunity to make what for them is a
big increase in sales. Hidden price cutting is likely to break out any
time the natural level is below the going level, and is unlikely to stop
until equality between the two is reached. Any attempt by the bigger
firms to hold the line, as they are the ones more likely to try it, simply
results in all the business going to the smaller firms in the industry.
The big firms get no opportunity to recover this lost business in times
of booming demand, and so run the risk of gradually being replaced
as the largest firms of the industry if they do not yield and follow
the market closely.
This sort of pattern of behavior is clearly visible in the nonferrous
metals industry: In a weak market the smaller firms press sales while
the larger firms follow along dragging their heels and complaining
all the while. The big firms have not been getting smaller, but only
because they meet market prices, however grudgingly. Here as elsewhere the structure of quoted prices, though volatile, understates
somewhat the fluctuation of actual prices; discounting is common in
a normal market, rife in a weak one, and negative when supplies are
tight. In a nutshell, in this market generally recognized to be competitive, we can see a certain amount of oligopolistic protestation
alongside truly competitive pricing.
That such concessions, discounts, and so on are given in a great
many industries in which quoted prices change infrequently is a matter of common knowledge, admitted by everyone except the large
firms in those industries themselves. A casual reading of trade publications and business periodicals will convince one that this sort of
behavior is extremely widespread, although data on the subject are
extremely hard to come by. The problem is simply that the one
element of unquestionable truth in the theory of oligopoly is that large
firms do not like to talk about sales at prices below the quoted price.
If one were to read only the statements of big steel executives, one
would be led to believe that no one in the steel industry would ever
dream of cut-price sales, as such sales would obviously be detrimental
to the profits of steel firms as a group. Certainly they leave no doubt
that they wish no such sales were made.
This leads to the ironic situation that big firms in truly competitive
industries bring floods of criticism and epithets of "monopoly" upon
their own heads by their stiff-necked secrecy about their pricing policies. They are so convinced that oligopoly theory ought to be true
and applicable to their own situations, and they so wish they could
price like oligopolists, that they talk like oligopolists while they price
like competitors. This is regrettable as much for the public as for
big business, because it distracts public attention from true monopolies,
price-fixing arrangements, and misuse of resources. It also makes it
difficult for the trained investigator, even if he knows just what he
1
This, of course, does not mean that it is "unnatural" for a firm to make an operating
profit sufficient to cover overhead in times of average or high demand. When operating
rates are high the extra costs incurred to produce additional output are higher than at low
operating rates (because of greater pressure on maintainance staffs, and other factors).
A price that covers these extra costs at high outputs will more than cover variable costs
for the entire output, therefore, and generally will yield a net profit.

ECONOMIC STABILITY AND GROWTH

Is looking for, to find where the monopolies and other abuses really
are.
For the time being we must conclude from this that oligopoly pricing, or any other type of essentially monopolistic pricing by big business, is less common than it would appear by considering quoted price
behavior and the public statements of businessmen themselves. Further discussion of this point, from a long-run standpoint, is given in
section 6, below.
4. The evidence is conclusive that the prices of steel and petroleum
are sensitive to month-to-month changes in market conditions. Steel
and petroleum are the two major industries widely believed to be of
the "administered price" variety on which I have been able to find
some hard factual evidence on prices actually charged. Unfortunately, the evidence on steel dates from 1940 and before, but there is
no reason to suppose that the situation has changed in any essential
respect. Studies by the TNEC on the sales data of the major steel
producers and by the BLS on the purchase data of a large sample of
steel users, covering various years up through 1939, show that price
concessions on the major steel products averaged some 6 percent off
the base price in 1939 and some 50-60 percent off the standard extras,
which meant that the base price plus extras was reduced by some 17
percent by these concessions. That most of the concessions were in
the extras rather than the base price may be due to the greater ease
of concealment in the category of extras. That these concessions
fluctuated from month to month was also made clear in these studies.
Chart I shows the average monthly mill net in relation to the quoted
base price of the United States Steel Co. for the years 1912-39, representing a composite of all their products sold. In most years the
net price received was actually less than the base price, which meant
that concessions were greater than the extras. (These figures are only
broadly representative of what was happening, since small month-tomonth fluctuations may be partly due to changes in the product mix.
Individual product studies give the same picture, however, so that
there is no reason to suppose that this chart is misleading.) Note that
generally on a rising market the quoted or "hoped for" base price
shot up far above and ahead of the prices actually received.

CHART I

REPORTED COMPOSITE PRICE AND COMPOSITE MILL NET YIELD
1926 -100
250

250
225

%
\ n
i

200
175
150

7
1

J
75

175
150

125

100

22*
200

REPORTED
COMPOSITE PRICE
<IRON AGO

COMPOSITE
MILL NET YIELD
(TO U.S.S.C.
SUBSIDIARIES)

«)
ce

125 I

\
\

-f\

17
*\
y \

100 x
o

z
75

C
O

§P

s
t>
d

50 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 50

Ordinarily the general level of the reported steel prices, as indicated by the Iron Age composite price
of steel, reflects the relative level of the mill net yields, i. e., the amounts received per ton by the U. S.
Steel Corporation subsidiaries on the various products after deduction of cost of delivery. However,
at times the level of the mill net yields has been slightly above or slightly below the relative level of the
reported prices, except that during the periods of intense demand during the World War and in 1920
when the prices charged by the Corporation's subsididiarles were beneath the level of the going prices.
Factors tending to lower mill net yields with respect to reported base prices are principally (a) reductions from base price, (b) excess of actual cost of delivery over freight added to base price in computing
the delivered price, (c) quantity discounts and (d) deductions for quality, size, eic. Factors tending to
raise mill net yields with respect to reported base prices are principally (a) extras for special finish,
quality, size, heat treatment, etc., and (b) extras for small quantity.

ECONOMIC STABILITY AND GROWTH

I strongly suspect, therefore, that the "price rise" of July 1957 was
largely fictitious; although a recovery in economic activity will very
likely save the steelmakers from the indignity of openly rescinding
the price increase.
The data on the petroleum industry are current, fortunately, and
with a little work can be made available for a considerable past period.
This industry is frequently alleged to have administered prices for
crude oil and for refined products at the refinery and in wholesale
markets. It is a straightforward matter to check up on the latter
prices by observing daily quotations in newspapers (such as the New
York Journal of Commerce) and the monthly figures in the Oil and
Gas Journal. The daily quotations posted by regular suppliers in
wholesale markets, less reported discounts, change on the average
every 3 or 4 weeks; but a more accurate picture is obtained by reading
the journalistic reports in the newspaper columns alongside these
postings. They report actual market prices, which change every few
days, according to the state of the market. When the postings change
they appear to change largely to get in line with prices actually
charged.
Price figures at the refinery level are obtainable from the Oil and
Gas Journal and the Independent Petroleum Association. Chart I I
and the accompanying table give an average of these figures by months
for 1957, along with the BLS series on petroleum products at wholesale for comparison. Both these series understate somewhat the
downtrend in these prices from March 1957 onwards, because of the
slight lag of movements in posted prices behind prices actually
charged; both series are based on postings less published discounts
only.
TABLE 1.—Monthly prices of petroleum products, 1957
Average
BLS quotation for
BLS index
value refined
crude oil
(1947-49=100)
products 1
(dollars per (dollars per
barrel)
barrel)
January...
February..
Marco
April
May
June..
July
August
September.
October....
November.
December.

124.9
131.0
130.7
130.4
129.8
128.4
126.4
125.5
125.6
124.6
123.5
123.5

4.13
4.25
4.21
4.17
4.16
4.10
3.99
3.97
3.99
3.97
3.92
3.94

2.82
3.07
3.07
3.07
3.07
3.07
3.07
3.07
3.07
3.07
3.07
3.07

1
Average of figures given by the Oil and Gas Journal and by the Independent Petroleum Association of
America (for east of California prices).

ECONOMIC STABILITY AND GROWTH

CHART II - MONTHLY PRICES OF PETROLEUM PRODUCTS, 1957
::::{;•'-1": T^ rT^t i:*r ^z z~. ~ r -rrr :rr: " " *." ~ ~-~ rnr !*~! \J~ " r-~ n ~ " : : —r' ..: — f~—[""F

The price of crude oil presents a more delicate problem, because
no regularly published data can be found on the prices actually being
charged. It is clear that the realized value of crude oil for the' big integrated producers went down, because of the fall in the prices of
final products; and since the 22 largest oil firms are net purchasers
of crude from smaller producers to the tune of 25 percent of United
States crude oil output, it does not seem at all probable that they
actually paid the high posted price for crude when the crude wasn't
worth it and when a multitude of suppliers were piling up inventories.
Late last year an oil company executive was reported in the Oil and
Gas Journal as having remarked in a speech that large quantities of
crude were moving at cutrate prices. Private inquiry of an oil
economist on my part uncovered the information that for some
months a major importer of crude has been selling it at prices as

ECONOMIC STABILITY AND GROWTH

much as 30 percent below the posted prices, and that price-cutting
is quite normal in the trade by various covert devices, such as shipping
a higher quality of crude than that for which the purchaser is invoiced. Within the last few months posted prices of crude have at
last been falling more into line with actuality, but they are still misleading.
5. Differing degrees of price flexibility in the different sectors of the
economy are apparently almost entirely explainable in terms of differences in the behavior of costs, without regard to concentration or
monopoly. The work of other researchers, combined with the direct
evidence on steel and petroleum which I have presented and with
the impressions one gains by reading trade journals and the like for
many industries, definitely point in this direction. Means found that
a small selected sample of quoted prices did show a weak but definite
relation between degree of concentration and list price rigidity, although Tucker and others have questioned whether this can be generalized for the whole economy. I would not be surprised if it can,
since maintaining a fiction of a quoted price different from actual
prices is behavior, pointless though it may appear, more to be expected
in a concentrated industry than in an atomistic one.
However, when one turns to realized prices, the prices actually
charged, various researchers, especially Neal, Thorp & Crowder, and
Tucker, found no relation at all bstw^een concentration and price
rigidity or at most a negligible relation. These results tended to show
instead that prices vary closely with costs, which explain (even when
only crudely measured) some 80 percent to 90 percent of the
variation
in prices between 1929 and the bottom of the depression.2 However,
these studies have to be supplemented with the additional material
and impressions which I have put forward, because the product classifications used are so unavoidably crude. Even all this together is
not absolutely conclusive, however, and many economists who are
familiar with all or most of this material find it possible to continue
to give full play to their suspicions about the harmful pricing practices of big business. I believe that a continuing accumulation of
evidence may one day convince them that they are overdoing a good
thing.
2
Neal (item 21 in the bibliography) found that, between 1920 and both 1033 and 1035,
the coefficient of determination (r2) between costs and realized prices was consistently of
the order of 0.80 to 0.90, depending on the sample used. The partial coefficient of determination, after allowing for the effect of co«ts. between prices and the concentration ratio
was consistently around 0.09 : that is, of the totnl variation of prices, only from 1 to 2
percent was attributable to variation in concentration ratios.
Thorp and Crowder (29) found that in their data there was no relation at all between
changes in realized prices and concentration, nor between changes in realized prices and
changes in concentration. Correlation methods were not applied, but it was evident that,
if thoy had >^een, a zero result would have been obtained.
Tucker (30) concentrated ^is attention on Means' data, and pointed out that if a lareer
sample h^d been used the relation Menns found between list-price rigidity and concentration would 'ar^ely disappear: it would disappear entirely, he said, if allowance were made
for the irisleading character of list pricing and its correlation with the degree of concentration. The other two studies te^d to bear him out in this contention.
One ot^ev question mark that hns to be put alongside these studies, besides that arising
from possible changes in the product mix. is that nationnl concentration ratios do not
equally reject degrees of concentration in local markets for all products. Tt is not possible
to say how systematic consideration of this factor would affect the result, if nt nV, but such
attention as has been paid to it has led to the conclusion that the effect would not, in any
case, be large.
All this does not denv that a monopolist may have sensitive prices like everyone else,
and th^t he mav, nevertheless, suffer a smaller drop in profits during bad times tban do
other sellers. This aspect of the matter is best viewed as a part of the problem of longrun overpricing, discussed in the next section.

ECONOMIC STABILITY AND GROWTH

One other serious qualification must be entered at this point, however, and that concerns the rigidity of the wage structure. There is
little doubt that here is a price that is not very sensitive to changing
economic conditions. Nourse, among others, has repeatedly cited this
factor as a source of excessive price rigidity in the nonagricultural
sectors of the economy, and in particular he has tended to indict the
labor unions for the existence of this factor. Now, it is true that since
the New Deal era wages have shown no tendency to drop in recessions,
which they had generally done previously. However, careful studies
of wage movements have not tended to find much relation between
these movements and the degree of unionization, but rather seem to
show differing degrees of reluctance on the part of employers to cut
wages. Other aspects of presentday policy than the labor laws, such
as unemployment insurance, may have reduced the downward flexibility of wages; but unions do not seem to have made much difference.
6. Long-run overpricing of commodities is not a widespread problem in our economy. In this connection, the histories of the steel and
petroleum industries are very suggestive. Both had monopolistic
pricing and profits in the first decade of this century, and neither
made any secret of the fact that this was the way they wanted it to
be. Both monopoly situations fell apart, but for rather different
reasons. The United States Steel Co., which on its formation in 1901
controlled the production of some two-thirds of the steel ingots produced in the United States, priced monopolistically for 10 years or so,
providing an attraction and an umbrella for the expansion of its
competitors. The United States Steel Co., now controls only one-third
of the steel ingots produced in the United States. Its control of steel
pricing broke down by the advent of the First World War, and never
returned, although this has never been admitted. During the depression, the company joined the international steel cartel, accepting
a quota for exports by the entire industry with penalties for exceeding
the quota; as a result, the company regularly paid penalties as its
competitors regularly exported in excess of the quota, and the scheme
had no real effect as far as this country was concerned. This incident
occurred during a period of especially intense j)rice competition among
steelmakers within this country. The only time that United States
Steel has had its way on prices since the First World War was during
the unfortunate interlude of the NRA, which, mercifully, died quickly
except for its labor provisions. Yet to this day the company's executives talk like oligopolists, apparently looking wistfully back over the
decades to the time when this kind of talk meant whfit it said.
The Standard Oil cartel was a more tough and durable phenomenon,
having lasted from around 1880 to 1911, and it required the strong arm
of the Government to break it up. Throughout the period of its success, the cartel controlled 80-85 percent of refinery capacity in the
United States, and by one dubious means or another kept the remainder in line on prices and output quotas. After the cartel was
dissolved by court order and it became obvious that the successor
companies could no longer maintain discipline as before, competition
expanded rapidly. At the present time the successor companies control only 25-30 percent of the industry's output, and in many markets
are in competition with one another.
These two histories point up the fact that the public has two major
forms of protection from long-run overpricing of products: competi23734—58

100

ECONOMIC STABILITY AND GROWTH

tion, potential and actual, on the one hand, and the antitrust laws on
the other. By and large, these two forms of long-run protection are
adequate over that area of the economy where they are applicable.
The paths of history are strewn with the bones of cartel and monopoly
schemes designed to gouge the public, and nowhere have they failed
more regularly than in the United States. The chances for success
in such enterprises have always been slim, even without the antitrust
laws, and the failures have been spectacular and costly to their
promoters.
The main reason that cartel and monopoly schemes tend to fail is
indicated by the analysis presented in section 3. An overpriced commodity, one whose price is substantially above the cost of production,
presents an overwhelming temptation to existing and potential producers to add to existing output. Hence, a monopoly or cartel which
raises the price above its natural level will, in so doing, invite its own
destruction unless it can effectively prevent such additions to output.
Effective prevention of additions to the monopoly's chosen output
requires strict policing of outputs and prices, something which can be
done usually only with the assistance of Government (or with the tacit
sufferance of Government if it uses strong-arm methods, taking the
law into its own hands). Rockefeller failed twice before finally succeeding in attempts to form his oil-refining cartel, because of this
difficulty.
The notorious methods used by the diamond monopoly in the South
African mining areas, employing armored cars, halftracks, and the
like, and an extensive secret service, have been sufficient to maintain
its control of the diamond market for generations. However, this
control is now breaking down because new producing areas nominally
cooperating with the monopoly do not police so well. Schemes to
raise prices of tin, copper, rubber, and coffee have repeatedly broken
down and brought considerable losses on their promoters because of
insufficient control over total production. Examples of this kind are
legion.
It is also true that the price-raising power of any monopoly, even
if successful, is likely to be limited by competition from other products
satisfying the same basic need or use. For instance, a monopolist in
one type of fuel would to some extent have to hold his hand for fear
of causing people to change over to other types of fuel. This consideration may explain the recent trend toward moderation in the
demands of the coal miners' union, who seem to have seen the handwriting on the wall.
Nevertheless, even in the United States, where collusive agreements
and the like are especially difficult to police and maintain, there have
been some successes, costly to the public. The occasional successes
point up the need for the existence and enforcement of the antitrust
laws. A similar conclusion is reached if one examines the consequences of widespread monopoly in other countries, where, generally
speaking, no such laws exist.
By and large, available evidence on corporate assets and profits
tends to suggest that the extent of monopoly in the United States

ECONOMIC STABILITY AOT) GROWTH

101

economy is much smaller than one would expect if he accepted concentration ratios as a measure of monopoly.3
The main areas in which the public is not fully protected from
monopoly in this country are ones in which the Government, usually
actively but sometimes passively, has prevented or allowed the prevention of competition and has caused or permitted wrong and inefficient pricing to prevail. Licensing restrictions, franchises, patents,
rate regulation, and so on have effectively curtailed competition in
many areas of the economy and have permitted gouging of consumers
and misuse of resources. The most important by far of these areas is
transportation and public utilities, which are subject to rate regulation. Although these rates are usually kept down to a level which
only allows a "reasonable return" to capital, the structures of rates
set could scarcely be more viciously inefficient and harmful to the
public interest if that were their explicit object. Further, the "reasonable return" criterion is not always followed, as in many cases
the sole object of regulation is to set minimum rates so as to prevent
effective price competition between alternative forms of supply of a
basic service. Basic reforms in all the areas where the Government
actively or tacitly restricts competition are badly needed. In this
particular respect, one may quite properly say that maladministered
prices are a real and serious phenomenon, both from a short-run and
a long-run point of view.
7. History has shown repeatedly that the only cure for a drastic
fall in aggregate demand is for aggregate demand to rise again; the
appropriate public policy for stability and full employment, therefore,
consists of countercyclical monetary and fiscal policies, combined with
a structure of policy that encourages the private economy to maximum
growth and stability. Price flexibility has never been sufficient to
avoid sharp and undesirable drops in aggregate output and unemployment, and there is no reason to suppose that such flexibility ever will
be sufficient to do so. An adequate countercyclical monetary and fiscal
policy has never been put to the test, although the performance of
such policy in the recessions of 1937,1949, and 1954 was not bad. Most
economists are agreed that well-formulated monetary and fiscal policy
can be sufficient to maintain output and employment.
These remarks are equally applicable whether "administered prices"
are a widespread phenomenon in the private, unregulated economy or
not. Even if Means' ideas could be accepted at face value, these would
still point in the direction of maintaining adequate monetary and fiscal
policies as the main defense against undesirable fluctuations in output, employment, and the price level. It would be very convenient if
nearly all prices, including wages, were sufficiently flexible that no
Government action would be necessary to maintain output and employ3
Harberger's study of this question (10) based on 1925-29 data Indicates that monopoly
profits before income taxes are at most of the order of 1% percent of the national income.
This would imply, on my own calculation from his data, that from 6 to 8 percent of the
national product is under unregulated monopoly control. 'This estimate is made by assuming that all differences in profits are due to differing degrees of monopoly power, i. e., that
none are due to the rise and decline of industries with growth and change in the economy.
Many of these profit differences must really be due to the latter factor, so that this estimate undoubtedly overstates the amount of monopoly.
This estimate roughly agrees with the more conservative of Nutter's estimates (25) based
on concentration ratios and on Wilcox study (35). As not all concentrated industries
are high-profit industries, nor vice versa, however, these estimates do not agree on which
industries are monopolistic. Further, even where they do agree, some of the high profits
will be due to the dynamic factors just mentioned.

102

ECONOMIC STABILITY AND GROWTH

ment regardless of fluctuations in aggregate demand. They are not
that flexible, they never have been so far as we know, and it is not
likely that they ever will be. Government therefore has a useful function to perform in this matter, by meeting such fluctuations with carefully weighed counteraction.
The encouragement of the private sector to maximum growth and
stability in such a manner that will require the minimum use of
countercyclical devices does involve questions of prices and price
policies, although these are of a second order of importance in this
context. Where Government is involved in the regulation and restriction of private enterprise, this involvement should be such as to promote the highest possible real income through the best possible use of
resources. Preliminary study indicates that the reforms I have previously mentioned with respect to such involvements could produce an
increase in the real national
income of at least 5 percent and perhaps
as much as 10 percent.4
Concerning the unregulated private sector of the economy, the role
of public policy is more limited but is also important. In my view
the Government, apart from preventing monopolization, price-fixing
agreements, and blatant price discrmination, should adopt a position
of benign neutrality toward the private economy. Where monopoly
and artificial restraint of trade are not present, it may virtually be
taken for granted that particular products for which the demand falls
will suffer an appropriate fall in price. It may almost be taken for
granted that products for which the demand rises will enjoy an appropriate rise in price; this may not quite be taken for granted in every
case, however, because firms may fear an inappropriate adverse public
reaction or even prosecution under the antitrust laws. This is inappropriate because if the scarcity of the commodity relative to demand is real and not artificial, a high price is the only efficient way
to ration the scarce supply among its most important and valuable
uses. In peacetime the efficiency of the price mechanism is not questioned if the commodity in question is wheat, copper, or textiles; and
it is equally efficient if the commodity in question is steel or any other
commodity that happens to be produced by big business. As long as
the industry is competitive, so that the scarcity is not artificial, the
bigness of the firms which will make windfall gains in times of scarcity of their product (and windfall losses in times of plenty) is not
relevant to how the price should behave. It would be most helpful
to this end if the Congress made it clear to businessmen exactly where
they stand—what the limits are to the application of the antitrust
laws, and how much freedom they have to price according to the
State of the market. Corporate enterprises, both big and little, are
an integral part of the highly productive American economic system,
and will remain so for a long time to come. Pointless harassment of
these enterprises cannot possibly serve any useful purpose, and could
be very prejudicial to the objectives of maximum growth and stability
of the economy.
I also believe it to be true that the measures I endorse, namely appropriate monetary and fiscal policies, reform in the areas of the economy
where Government has a direct restrictive or regulatory influence,
and benign neutrality in the rest of the economy, are sufficient for
4

These are rough estimates based on calculations by Harberger and myself.

ECONOMIC STABILITY AOT) GROWTH

103

maintaining long-run stability of the price level in the face of upward
as well as downward pressures. The correctness or not of this belief
is of course of considerable importance to long run growth and stability, since many students of the subject have come to the conclusion
that stability of the price level and full employment are inconsistent
because of the pressure of wage.demands (or of the upward movement
of "administered prices"). Certainly one cannot deny that this could
conceivably be the case; but in those instances I know of where it
demonstrably has happened (i. e., where unions have demanded and
gotten wages that were inconsistent with full employment at existing
prices) wages have been escalated so effectively that full
employment
was not possible even with mild or unlimited inflation.5
It is easy to jump to the conclusion, whenever the consumer price
index goes up a few points, that someone has been "pushing" it up;
and of course big labor and big business are natural targets in this
respect. Showing that it has been happening in the American economy
is another matter, and I do not believe that it can be shown. The significant inflationary episodes of the past have almost certainly been
monetary phenomena, at least in the sense that sufficient monetary
restraint would have stopped them effectively without serious unemployment. There is little reason to suppose that the episode of 1955-57
was any different.
As Buggies pointed out last year to the Senate Judiciary Committee, the consumer price rises in 1955-57 occurred almost entirely
in services and utilities. Prices in the industrial sector could not by
any stretch of the imagination be credited with responsibility for
the upward movement of the Consumer Price Index. But this is
where the rise would have had to have occurred to be consistent with
giving big labor or big business the blame. On the other hand, I
do not find it difficult to interpret the events of 1955-57 in terms of
growing aggregate demand insufficiently restrained by monetary and
fiscal policy. These years saw an unprecedented boom in domestic
investment, a boom which ought to be expected to place pressure on
resources in a fully employed economy. At the same time, expanding consumer incomes were increasingly directed to services, in keeping with a long-run trend. Services, in order to keep employees and
attract them away from the booming investment sector, where jobs
were plentiful, had to continue even more sharply their long-run
tendency to bring their wages (and therefore costs and prices) up
toward a level consistent with this. Similarly, so far as one can tell,
the rises in prices in the utilities and other regulated areas made sense
in terms of rising costs and ample demand.
The price movements of the 1955-57 episode are highlighted in
table 2, which shows the changes in the general price level (gross
national product deflator) and its components. Consumer goods rose
the least (3.8 percent) of any of these components, while the biggest
rises were in compensation of general government employees (17.9
percent) and in producers' durable equipment (15.0 percent) which
had dropped slightly from 1953 to 1954. These movements make a
pattern very much in keeping with the interpretation of the episode
as a demand inflation generated from the active investment and government sectors.
8
I have in mind the cases of Germany's inflation of the 1920's and Italy's inflation
after World War II. In both these cases there was severe unemployment during inflation.

104

ECONOMIC STABILITY AND GROWTH

A more restrictive monetary and fiscal policy would have dampened
down somewhat the demands of the investment sector, which in turn
would have implied less pressure elsewhere in the economy. I have
seen no evidence to suggest that this would have necessitated serious
unemployment had the additional restraint occurred while the pressure in domestic investment was still high. After all, we had 5 years
of virtual price stability since the end of 1950, and without serious
unemployment. I think this can continue.
TABLE 2.—Components of the general price level
[1947=100]

1954
Gross national product deflator _
Consumer prices
Goods
Services
Construction
Producers' durable equipment
Government purchases of goods and services
Compensation of general government employees

119.9
116.8
112.6
128.1
129.7
128.1
125.2
134.1

1957
129.6
124.0
116.9
137.5
142.9
147.3
142.4
158.1

Percent
Increase
1954-57
8.1
6.0
3.8
7.3
10.2
15.0
13.7
117.9

i This item represents the wage and salary level in general government, and is not strictly comparable
with the other indexes because of the possibility that there has been technological advance in government
as elsewhere.
Source: Survey of Current Business, February 1958.

BIBLIOGRAPHY
1. Adams, Walter and Horace M. Gray, Monopoly in America (New York:
Macmillan, 1955).
2. Backman, Jules, Price Practices and Price Policies (New York: The Ronald
Press, 1953).
3. Blair, John M., and Arthur Reeside, Price Discrimination in Steel, TNEC
Monograph No. 41 (Washington: Temporary National Economic Committee, 1941).
4. Blair, John M., Economic Concentration and Depression Price Rigidity,
American Economic Review, May 1955, p. 566.
5. Daugherty, Carroll R., Melvin G. de Chazeau, and Samuel S. Stratton, The
Economics of the Iron and Steel Industry (New York: McGraw-Hill,
1937).
6. Dennison, H. S., and J. K. Galbraith, Modern Competition and Business
Policy (New York: Oxford University Press, 1938).
7. Elliott, William Y., et al., International Control in the Non-Ferrous Metals
(New York: Macmillan, 1937).
8. Federal Trade Commission, Control of Iron Ore, Report to the Antitrust
Subcommittee of the Judiciary Committee of the House of Representatives (Washington: Government Printing Office, 1952).
9. Galbraith, John K., Market Structure and Stabilization Policy, Harvard
Business Review, 1957.
*10. Harberger, Arnold C, Monopoly and Resource Allocation, American Economic Review, May 1954.
11. Jones, Eliot, The Trust Problem in the United States (New York: Macmillan, 1921).
12. Laidler, Harry W., Concentration of Control in American Industry (New
York: Thomas Y. Crowell, 1931).
13. Livermore, Shaw, Concentration of Control Now as Compared With 1890,
Journal of Marketing, April 1940.
14. Mason, Edward S., Economic Concentration and the Monopoly Problem
(Cambridge, Mass.: Harvard University, 1957).
15.
, Price Inflexibility, Review of Economic Statistics, May 1938
16. Means, Gardiner C, Big Business, Administered Prices, and the Problem
of Full Employment, Journal of Marketing, April 1940.

ECONOMIC (STABILITY AND GROWTH
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.

105

"industrial Prices and Their Relative Inflexibility," Senate Document
No. 13, 74th Congress, 1st session (Washington: U. S. Government
Printing Office, 1935).
, The Structure of the American Economy, Part I (Washington: National Resources Committee, 1939)
Mills, Frederick C, Prices in Recession and Recovery (New York: National
Bureau of Economic Research, 1936).
National Bureau of Economic Research, Business Concentration and Price
Policy (New York, 1954).
Neal, Alfred C, Industrial Concentration and Price Inflexibility (Washington: Public Affairs Press, 1942).
Nelson, Saul, and Walter G. Keim, Price Behavior and Business Policy,
TNEC Monograph No. 1 (Washington: Temporary National Economic
Committee, 1940).
Nourse, E. G., Industrial Price Policies and Economic Progress (Washington: Brookings Institution, 1938).
, Price Making in a Democracy (Washington: Brookings Institution,
1944).
Nutter, G. Warren, The Extent of Enterprise Monopoly in the United States,
1899-1939 (Chicago: University of Chicago, 1951).
Purdy, Harry L., Martin L. Lindahl, and William A. Carter, Corporate
Concentration and Public Policy (New York: Princeton-Hall, 1950).
Stocking, George W., The Oil Industry and the Competitive System (Boston:
Houghton Mifflin, 1925).
Tarbell, Ida M., The History of the Standard Oil Co. (New York: McClure,
Phillips & Co., 1904).
Thorp, Willard L., and Walter F. Crowder, The Structure of Industry, TNEC
Monograph No. 27 (Washington: Temporary National Economic Committee, 1941).
Tucker, Rufus, Concentration and Competition, Journal of Marketing, April,
1940.
, The Essential Historical Facts about "Sensitive" and "Administered"
Prices, The Annalist, February 4,1938.
, The Reasons for Price Rigidity, American Economic Review, March
1938.
United States Steel Corp., TNEC Papers (New York: United States Steel
Corp., 1940), volume II.
Weston, J. Fred, The Role of Mergers in the Growth of Large Firms
(Berkeley and Los Angeles: University of California, 1953).
Wilcox, Clair, Competition and Monopoly in American Industry, TNEC
Monograph No. 21 (Washington: Temporary National Economic Committee, 1940).
f

THE PRICE INDEXES OF THE BUREAU OF LABOR
STATISTICS
H. E. Riley, Chief, Division of Prices and Cost of Living, Bureau of
Labor Statistics, United States Department of Labor
In this paper I shall be concerned with some of the practical problems affecting the development and use of two widely Known indexes
of price movement: The Consumer Price Index and the Wholesale
Price Index of the Bureau of Labor Statistics. Both of these price
measures serve two broad areas of need: For general economic
analysis and to guide policy decisions by government, business, and
the public. Neither index can be completely satisfactory for both
theoretical analyses and practical day-to-dajr decision making. In
fact, price measures of any kind have certain inherent limitations
which must be recognized in any application.
Index numbers provide a convenient way of studying price behavior
quantitatively. Simply defined, an index number is an average of
price changes between two points of time. This implies that there is
a universe of items for which we wish to measure the average price
change. Since an average is involved, it applies in addition that
we know how important each item is in the universe so that we can
weight together the measures for individual items. If we know all
about the universe of items at both points of time for which we are
computing the index of price change, there is no problem—one could
define and compute literally an infinity of measures of price change.
But no one of these indexes would be right for all purposes. A more
important point is that complete quantity and price information is
not available at any given moment for any index of significant scope.
The census of manufactures, for example, identifies something like
8,000 individual materials and product items, not counting all variations in size, quality, or marketed form.
These commodities are changing hands continuously, passing from
producer to wholesale distributor to retailer to the ultimate consumer.
Each transaction involves a price determination. Any workable plan
for a price index, designed to reflect the movement of prices at any
level in this complex? must rest on a sampling structure. It requires
a sample of commodities, of sellers, and a sample of price quotations
or transactions. Thus, in the Consumer Price Index we have a sample
of about 300 commodities and services. We have a sample of retail
stores and service establishments clustered in 46 cities. From each
store we obtain a sample of prices—the price in effect at a point in
time for each sample commodity offered for sale. Accuracy in each
of the samples is essential for a valid index number.
Price can be defined as the amount of money "paid" for a unit of
commodity or service, but this formal definition provides no more
than a starting point. There are various kinds of transactions and
the price, or money payment, may enter into the transaction at dif107

108

ECONOMIC STABILITY AND GROWTH

ferent stages. In some cases different transactors pay different prices
for equivalent volumes of goods and services. Therefore, for a price
to be useful in statistical measurement, its exact meaning must be
understood.
For the BLS indexes, the commodities to be priced are defined in
terms of more or less exact physical specifications, depending on the
item involved. When necessary—i. e., where the price might be affected by these factors—the classes of customer and/or seller involved
in the transaction are likewise specified. There is a wide variety of
prices that can serve as the basis for statistical measurement; not all
are equally good or appropriate. There is what can be called a list,
or posted, price. In most stores of this country that is the price at
which transactions are made. For some items, including some sold
at retail, the list or posted price is very different from the actual
transaction price. Witness the haggling which commonly accompanies the sale of a new car. In this case the determination of price
is usually further complicated by the fact that the transaction involves
cash plus a "trade-in." It is obvious that a realistic measure of
price change should be based on actual prices paid rather than list
prices, if the two differ.
The distinction between list price and transaction price is extremely
significant in the case of nonretail prices. The practice in many industries is to maintain a list price for a long time and change the price
as market conditions require by offering or withdrawing discounts or
"extra" charges. Variations from the list price are made not only
over time but also at any given time for different classes of customers.
Obviously, in comparing prices for supposedly the same item over
time, one of the essentials is to insure that the prices cover equivalent
terms. Exactly the same item, for example, may be quoted on an
f. o. b. basis or on a delivered basis, or inclusive of installation costs
in some cases. Changes in the terms of sale may be used as a device
for adjusting real prices without changing published list prices.
Generally, the commercial transaction unit is the one used for purposes of price measurement. Thus, the cost for a monthly telephone
service with unlimited calls is an appropriate measurement consistent
with this concept. A different price, if it could be computed, would
be secured if the price per actual phone call made or received was to
be considered. Similarly, the price per thousand cubic feet of natural
gas is an appropriate price measure, or the price of a ton of coal.
Either one of these prices could be converted into price per equivalent
quantity of heat, measured by the British thermal unit, which might
be a useful concept for some purposes. For some of our series we do,
in fact, refer to the B. t. u. content in the specifications of these two
items. In most instances, however, it would be impossible to define
units of performance in this manner.
A basic problem in the measurement of price behavior is that few
products remain constant in their physical characteristics over any
length of time. In some cases it is possible to adjust the prices available for the two periods being compared to take account of technical
differences in individual products, which, it is assumed, are reflected
in market valuations. In other cases the products have changed so
much that comparisons over time cannot be made except with explicit
statement of the basic assumptions involved. Thus, we hear, for example, how the car of today differs from the car of 10 years ago.

ECONOMIC STABILITY AXD GROWTH

109

Some of the differences can be accounted for and prices adjusted; for
example, automatic shift and power steering. Other differences, however, cannot be evaluated in dollar terms. Some manufacturing
changes may have reduced costs and at the same time increased the
utility of the product. In very few cases are the ideal conditions for
computing the true market evaluation of the differences between an
old product and its current modification satisfied. To do this would
require that both products be available on the market in significant
quantities for an overlap period of suitable length.
Of course, new products are continually coming on the market,
and others drop out. This constant change in the commodities flowing through the market presents especially difficult problems in measuring average price changes over a long span of time.
Computation of an index to represent the average price change
between two periods of time requires both price data for individual
items and quantity data for the weights by which the prices are
combined. Determination of the quantities presents many technical
and theoretical problems. In the early days of index number research a great deal of effort went into finding some intrinsically
"true" method of weighting which would be best in a general sort
of way. We now recognize that the choice of weights depends (1)
upon the objective which the index number is designed to serve and
(2) the availability of data from which the weights can be derived.
Regardless of the objective, however, the availability of data often
becomes the dominant criterion in the selection of weights.
The interpretation of a particular index number depends not only
on £he way in which the price data are combined but on the logic
of the weighting structure for the particular purpose to be served
by the index number, and the firmness of the statistical foundation
oi the weights. In some widely used index numbers, the weights
are hypothetical or arbitrary, or are based on very limited data. The
weights of our daily spot market index are arbitrary. By averaging
the prices without varying the weights we give each quotation equal
weight in the computation. This may not impair the usefulness of
the index numbers provided users are aware of what the weights
mean. For other index numbers, the weights rest upon very broad
statistical foundations as well as carefully thought out logical structures.
The major BLS price indexes are good examples of indexes with
weights derived from comprehensive statistics for a base period.
Thus the weights of the Consumer Price Index represent the quantities
of the various commodities and services consumed by city worker
families as measured in an extensive survey of family spending in
1950. The weights were adjusted to 1952 conditions before being
introduced into the Consumer Price Index as issued in January 1953.
In effect the Consumer Price Index formula is a variation on the
Laspeyres base weighted index, named for the 19th century statistician who developed the original equation.
We use the term "fixed basket" in connection with the Consumer
Price Index because each monthly set of price quotations is combined
by the quantities represented in the basket to derive the current value
aggregate. A comparison of this current aggregate value with a
similar aggregate value based upon 1947-49 prices gives the index
in relation to the standard base period for Government statistics.

ECONOMIC STABILITY AND GROWTH

A comparison with a similar aggregate for the previous month gives
the percent change shown by the Consumer Price Index since the
preceding month.
Many statisticians argue that the indexes should be constructed
with current rather than base period weights. This is known as
the Paasche formula. An index of this type has value for many purposes but, as a practical matter, the necessary data for weights cannot
be obtained in time for current computation of price indexes. Moreover, a current weighted index presents a comparison of prices as
against a base reference period. Comparison between any other two
periods are invalid because the differences in index numbers involves
not only changes in price, but also changes in weights. Thus current
weighted formulas do not satisfy a maior requirement of BLS price
indexes that the indexes be comparable not only as against a base
period, but also as against previous months and previous years.
Prices indexes derived from careful specification pricing of the
tvpe that the BLS builds into its Consumer Price and Wholesale
Price Indexes are often compared, with confusing results, with indexes
of realized unit value changes. For example, using data from consecutive censuses or annual surveys of manufacures, it would be possible to compute the dollar change in unit value per ton of all steel
produced, or of all gasoline. Such comparisons are affected, however,
not only by price change of carefully defined types of steel or gasoline
sold through specific channels, but also by changes in the product
mix. There might be more or less high-grade steel or high-octane
gasoline in one period than another, for example. Moreover, the
change in unit values may reflect different sources of supply or different customers. An index of unit values derived from purchasers
would give a still different result as compared with unit values derived
from producers' shipments since changes in transportation and distribution costs, as well as price change, would be involved.
Some confusion exists regarding the meaning of the term "base
period" as applied to the BLS price indexes. Although the weights
for the CPI are based on the year 1950, adjusted to 1952, the index
as published describes the change in prices from the average of 194749. The latter reference point might be termed the index base, or base
reference period. Ideally, the weight and reference bases should
coincide, but in practice this is not always feasible. Basic data suitable for weights are not always available for the standard reference
period established by the Bureau of the Budget for all Government
statistical series. Moreover, it may be desirable for some index series
to change the weighting structure more often than the reference
period is changed.
For specific purposes it may be convenient to shift the reference
base to a period other than the standard one for a given series. Thus
a chart of the CPI and its major components on a 1939 base affords a
graphic picture of the movements in consumer prices since immediately before World War TT. Of course, as Weslev C. Mitchell demonstrated in the famous
BLS bulletin on The Making and Using of
Index Numbers,1 shifting the reference base may lead to sizable errors,
unless the translation is computed separately for each item in the
index. This arises from the fact that as the index is carried further
1

U. S. Department of Labor, Bureau of Labor Statistics, bulletin No. 656.

ECONOMIC STABILITY AND GROWTH

111

away from its weight base the relative movements of the individual
prices become widely scattered. The short method of shifting in
effect applies a different set of weights to the individual prices than
that which was carefully built into the original index series.
I have described in general terms the nature of our Wholesale and
Consumer Price Indexes, indicating that each embraces a wide variety
of commodities involved in market transactions in a defined stratum
of economic activity. The question often arises, How well do these
indexes indicate changes in the general price level ? This is difficult
to answer, because no one has succeeded in defining the universe of
transactions included in a general price level in terms capable of
measurement. In its broadest context the general price level embraces all transactions involving transfers of goods, services, money,
and financial obligations however expressed. Many of these elements
cannot be defined in terms such that a price or value could be determined for purposes of index construction or that weights for combining different elements could be established.
The nearest approach to an index of all prices is the implicit deflator
which is a byproduct of the translation of current dollar estimates of
the gross national product into constant dollars of a base reference
period. The GNP implicit deflator is derived by first deflating each
component of the GNP by the most appropriate price index from
whatever source available. Each component is thus expressed in terms
of constant dollars of the desired base period. A comparison of the
aggregate of deflated components with the current dollar aggregate
provides the implicit price index. If the deflation is carried out in
maximum detail the resulting price index has the characteristics of a
Paasche, or current weighted index. To the extent, however, that
the original deflation is accomplished for larger groups only, using
already available composite price indexes, the resulting implicit deflator has a larger component of weights from earlier periods, and
loses some of its Paasche characteristics.
For most purposes measures of general price movements have more
usefulness if limited to a definable transaction level or economic
sector. It is in this respect that the BLS indexes provide effective
measures of changes in price levels. The Consumer Price Index is
a reasonably good measure of retail price trends, representing household purchases of goods and services. By definition the CPI does
not measure changes in prices paid by high income or very low income
families, farmers, or the self-employed. To a certain degree, however,
the index covers the modal range in the entire retail market distribution of household goods.
The Wholesale Price Index may be considered a measure of general
price movements at the production level of the economy. Although
its commodity content is much larger than that of the CPI, it is a less
satisfactory general price measure in that it does not include business
services, construction, real estate, transportation, and securities.
There is a tendency to think of the CPI and the WPI as two comparable measures of price movements at different levels in the economy.
In a sense, of course, this is true. But the assumption of similarity
leads to erroneous conclusions regarding the relationship between the
two. We are often asked to explain why the CPI is rising while the
W P I is falling, or vice versa. When the Consumer iMce Index rises

112

ECONOMIC STABILITY AND GROWTH

more rapidly than the Wholesale Price Index, many users leap to the
hasty conclusion that the housewife is being victimized by the rapacious
middleman. Such a conclusion cannot be supported by comparing
the two indexes. In the first place the indexes do not measure prices
of similar groups of commodities at two transaction levels. The
Wholesale Price Index includes raw materials. It also includes semifabricated parts and components made from those raw materials and
the final products incorporating the parts and components. The
manufacturer's price of the consumer good may parallel the retailer's
price, but in the W P I that price change may be offset by movements
m prices of things that never enter the consumer market. This means
that comparison of the two indexes will no.t provide a valid measure
of trends in price markups or margins.
Thus far, I have referred to the Consumer Price Index and the
Wholesale Price Index in discussing price index problems and objectives. A more detailed examination of the two indexes will serve to
indicate their uses and limitations and provide a background for suggesting needed improvements in and additions to these basic measures
of price movement.
The Consumer Price Index, inaugurated in essentially its present
form in 1918, is defined as "a measure of changes in prices of the
goods and 2services bought by families of city wage earners and clerical
workers." I t is a price index, and not a measure of changes in family
living costs. This distinction is worth emphasizing, because it has
significant implications with respect to the use of the index in wage
escalation.
The CPI is computed by comparing prices from period to period for
a fixed "market basket" of goods and services. A cost-of-living index,
on the other hand, would be analogous to an expenditures index. The
usual concept of cost of living includes the changes in family expenditures which occur when its living habits change, its income rises or
falls, and its requirements for food, shelter and clothing change with
increases or decreases in the family size.
Although the index is not constructed to reflect changes in living
costs as thus defined, the market basket items are representative of the
things that wage-earner families actually buy. The content and importance of the items in the basket are determined by periodic surveys
of family expenditures. The latest of these studies was carried out
during the period 1949 through 1952, and resulted in the revision of
the index first released in January 1953. The basic index structure
has not been altered since that time, although we have changed commodity specifications or substituted new items in the market basket
as items originally included were replaced in the market by others.
The market basket contains about 300 specific items of goods and
services, including foods, clothing, house furnishings, rent, home
maintenance, personal and medical care, recreational goods and services, and other items in the consumption pattern of the population
group specifically covered by the index. The current prices of these
items are weighted to represent all items consumed by the index
families.
The prices for the index are obtained from representative stores
and service establishments, or other appropriate sources, in 46 cities.
2
See BLS Bulletin No. 1168, Techniques of Preparing Major BLS Statistical Series, ch. 9,
fojr full technical description of the Consumer Price Index.

ECONOMIC STABILITY AND GROWTH

113

The sample of cities includes the 12 largest urbanized areas in the
country. An additional 9 large cities, 9 medium-size cities and 16
small cities, each representing a sampling cell in the universe of about
3,000 urban places in the United States, makes up the balance of the
city sample.
In addition to the national figure, an index is published for each of
the 20 largest cities in the sample. For five major cities monthly indexes are provided. Indexes for the other 15 are issued quarterly.
This schedule of releases is governed by the frequency of price collection. We cannot collect prices each month in each city for all items
in the market basket. Food prices are obtained monthly in all cities,
and all items are priced each month in the five large cities. Most of
the nonfood items are priced quarterly in the remaining 41 sample
cities, on a rotating basis, so that prices are available each month from
a subsample of cities of all sizes.
The most obvious and important use of the Consumer Price Index
today is in wage escalation provisions of long-term labor-management
contracts. It is estimated that approximately 4 million workers are
employed under such contracts by which their wages are adjusted on
a quarterly or semiannual basis for changes in the purchasing power
of the wage dollar as measured by the index. In fact, of course, the
index is a consideration in virtually every wage determination made
in this country, whether by management alone or through collective
bargaining.
Although the escalation provisions hinge on a single figure, usually
the national index, other important uses involve not only the summary
figure but the individual item and group series underlying the total.
All of these uses have focused widespread attention on the index,
raising questions about its meaning, accuracy, and appropriateness in
various situations. Many questions arise from misunderstanding,
but others point to the need for expansion and improvement in the
underlying current price data and additions to the number and types
of price measures made available. The increasing need for retail
price data in deflating components of the national accounts can be
satisfied only by expanding the scope of the current retail price reporting. The appropriate vehicle for this purpose is the current price
program of the Bureau of Labor Statistics.
There is gi*eat demand for additional individual city consumer
price indexes and for comparisons of living-cost differences among
cities. For intercity comparisons, the most effective device is the
standard budget. Such a budget, described as the content of a
"modest but adequate standard of living" for a 4-person city wageearner's family, was developed by the Bureau in 1947 and its cost
computed for each of 34 large cities. That budget base is now out of
date, but the Bureau is engaged in developing a new budget structure
reflecting current standards. If funds can be provided for the necessary additional price collection, it is hoped that new budget estimates,
in terms of annual dollar totals, can be issued by mid-1959 for each
of the 46 cities in the Consumer Price Index sample. These figures
will furnish a direct and easily understood means of comparing living
costs among those cities. Standard budgets are also needed for other
family types and for various levels of satisfaction. The present
Bureau program includes the preparation of a budget for an elderly

114

ECONOMIC STABILITY AND GROWTH

couple, to provide a basis for evaluating the adequacy of retirement
benefits and for estimating changes in living costs for retired couples
when they move away from the place of former employment.
Reference has been made to the fact that the basic structure of the
Consumer Price Index has not been brought up to date since 1952.
While maintenance of a fixed base for price comparison is essential if
the index is to serve as a wage-income deflator, it is recognized that
the sample of items and their weights must be reviewed at periodic
intervals. While there are some dissenters, it is generally agreed that
such review and revision should be undertaken at intervals of not
more than 10 years. In view of the time required for the completion
of a new consumer expenditure survey, it is essential that a revision
program be initiated in the very near future if the index is to be
revised on that schedule.
As the earlier discussion has indicated, the Wholesale Price Index
differs fundamentally in many respects from the Consumer Price
Index. In fact, it specifically excludes from its scope sales to household consumers. The term "wholesale" in this instance refers to sales
in large lots, not to prices paid or received by wholesalers, distributors, or jobbers. 3 The index does not include Government, services,
or constructions. Whenever possible, the prices used in constructing
the index are those applying to the first important commercial
transaction for each commodity. The index is therefore frequently
and more properly referred to as a measure of primary market prices.
The Wholesale Price Index is based on monthly price data for nearly 2,000 commodities, ranging from raw materials, such as grains,
fibers, and iron ore, to finished products, such as canned foods, clothing, automobiles, and machine tools. For most of the raw materials
the prices used are those established in the organized exchanges and
quoted in the trade journals. Prices of semifabricated and finished
products are generally obtained directly from the producers by mail
questionnaire.
The weighting structure of the index is designed to account for
the value of all commodities sold in the domestic market, including
imports. The weighting data are obtained from value of shipments
as reported in the census of manufactures, the value of agricultural
and extractive industry products are reported by the Department of
Agriculture and Interior, and imports as reported by the Department
of Commerce. The general policy of the Bureau is to review and
revise the weighting structure each 5 years, as the results of the
quinquennial Census of Manufacturers becomes available.
Although, as I have indicated, the Wholesale Price Index is used as
a measure of general price trends, it is the detail underlying the total
index that has the widest usefulness. Individual item and group indexes are used in deflating components of the gross national product
estimates. Segments of the index are used in escalation provisions
of long-term production contracts, commercial leases, and supply
contracts. For example, virtually all of the heavy power generating
equipment produced is made under an arrangement by which the
contract sum is adjusted for changes in the prices of selected materials and components between the initiation and completion of the
job. Federal shipbuilding contracts contain similar provisions.
s
Rf»e Bureau of Labor Statistics Bulletin No. 1168, ch. 10, for full technical description
of the Wholesale Price Index.

ECONOMIC STABILITY AJSTD GROWTH

115

The wholesale price series are nationwide in scope, with no local
detail except for individual item prices as quoted on organized exchanges. Thus, the monthly detailed report provides prices and indexes for potatoes at Boston, Chicago, New York, and Portland, Oreg.
Flour, butter, poultry, milk, and other basic foods are reported for
several market centers. Central and eastern regional prices are
quoted for paperboard. For virtually all other commodities in
the index, local detail is not available, chiefly because of sampling
problems and the necessity of avoiding disclosure of confidential
data. As prices are reported on an f. o. b. point of production or
freight equalized basis, the combination of prices originating at different shipping points provides a consistent series. To the individual consumer of industrial products the index may, however, appear incorrect. Thus, the purchaser of, say, finished steel products,
may find that the index movement doesn't follow the trends of his
own costs. His costs may rise because of freight-rate increases, or
because in a period when the terms of sale are changing he is unable
to take advantage of better quantity or cash discounts offered by the
producers.
It has been suggested that a valuable adjunct to the present wholesale price series could be developed through a series based on reports
of prices paid by the distributor or industrial consumer. Such an
index would reflect fully the effects of changes in freight rates and
might also provide a useful check on the ultimate effect of the discounts
and allowances offered by producers. The maintenance of a consistent
and continuous index of prices paid presents several knotty problems.
One of these is the sample of purchasers. Most industrial users of
raw materials, components, and capital equipment do not buy all of
these items every month. A very large reporting sample might be
required? therefore, to furnish sufficient data to establish the average
price paid at a point in time. Moreover, the composition of the reporting sample would vary from month to month, which would introduce problems of equating the variations from buyer to buyer in freight
costs, discounts, and allowances.
The wholesale price series suffers from several deficiencies and inaccuracies, some of which are almost impossible to correct within the
present frame of operations. While these deficiencies have little effect
on the overall index, they present serious obstacles to the use of many
subgroups and special product combinations that are desired for analytical, deflation, or escalation purposes. An important problem area
includes heavy industrial equipment, ships, locomotives, and aircraft.
Most of these products are manufactured to order, and no "market
price" is established for them. Each ship differs from every other
ship. Each purchaser of aircraft specifies particular features to be
included in his order. The value of these unique products is included
in the weight base, but their price in the current index computation is
imputed to the price movement of their components or other items of
a generally similar type.
The use of mail-price reporting has imposed some limitations on
our ability to obtain accurate data. There is evidence that some
manufacturers may have allowed special discounts or sales rebates to
their dealers and have not reported the fact in filling out the monthly
price schedule. The solution to this problem appears to rest in regular
personal contacts with the firms, to explain the objectives of the price23734—58

116

ECONOMIC STABILITY AND GROWTH

index program, to determine the best method of obtaining an actual
realized price, and to encourage cooperation in accurate reporting. For
some items it appears that direct collection, by personal interview,
rather than use of the mail questionnaire, may be necessary.
Mention should be made of the need for price indexes on an industry
basis. Some types of economic data, such as employment, capitalization, and dividend payments, are available on an industry but not a
product basis. Comparable price indexes are frequently needed in
the analysis of these series. Industry price indexes are also essential
for the development of productivity measures, and for studies of
interindustry transfers.
The derivation of adequate industry price indexes involves more
than a mere regrouping of the present commodity series. Aside from
the fact that the output of some major industries is poorly represented
in the present series, there are difficult weighting problems. Ideally,
the weights should be constructed on a value added basis, so that each
industry series would be influenced by the actual value contribution
made by that industry to the total output. The present item weights
in many instances represent values of groups of items for which price
movements are generally similar. This imputation pattern may
ignore industry of origin, which means that the value of output for one
industry may be included in the weight attached to the product of
another. Among the other problems involved are the derivation of
adequate weights for secondary products, and the inclusion of interplant transfers, now excluded from the weighting structure. Despite
these problems, however, this is one of the gaps in the Bureau's price
program which should receive early attention.

PRICE AND INCOME MEASURES FOR AMERICAN
AGRICULTURE
Oris V. Wells, Administrator, Agricultural Marketing Service,
United States Department of Agriculture
The invitation to prepare this statement was generous: I was to concern myself with the general price level and nonagricultural phenomena only to the extent I so desired, I was not to feel mortgaged
to any existing statistical measures simply because they were now provided for, and finally I was not to be overly concerned with theory;
rather, I should so far as possible keep the discussion in terms of
what appeared to be reasonably practical.
There are three comments I should like to make with respect to these
terms of reference:
First, I am of course aware of and agree with the argument in
favor of a stable "general price level." However, I do not know how
to measure, at least in any precise way, the general, overall price level
nor do I think that any single measure, assuming we were to agree on
one, would wholly serve the purpose which the committee has in
mind.
That is, it never seemed to me in the twenties and early thirties that
the Wholesale Price Index measured the general price level even
though we often talked as if it did. Nor do I today accept the Consumer Price Index, useful as it is, as a single sufficient measure of the
overall concept. Further, except under pressure of great economic
strain, and usually then also, it seems to me that many, often the most,
of our price problems have to do with differences in price and related
economic developments as between different sectors of the economy.
Very simply, this means that we are as much interested in sector or
partial price level measures as in the direction and magnitude of the
total price movement.
Second: In discussing price and related economic measures for
agriculture, it is difficult for me to start anywhere except with the
measures now being used, along with our current recommendations
for improvement.
Our farm statistics have gradually developed over quite a long
time, they have been and are being used to evaluate programs and
arrive at decisions, and the problem as I see it has to do with strengthening and improving our current measures, not with designing an entirely new set.
Third: Analysis of why prices behave as they do and the effects
or implications of such behavior are as important as any set of price
statistics. Prices or price indexes by themselves are purely neutraL
They only tell in a predetermined way what has happened—not why
nor whether it is good or bad, nor what should be done. So I conclude that we should never spend all our funds simply to collect, com-

117

118

ECONOMIC STABILITY AND GROWTH

bine, and release statistics. Appropriate analyses must also be arranged for.
There is a good argument that the only really good index or
aggregative measure for any practical purpose is one which has
been designed to fit precisely and particularly the problem at
hand. In fact, I often find myself arguing along the lines of
Kenneth Arrow's recent (Econometrica, October 1957) conclusion: "In view of the magnitude of an economic system, it would
take only a very small percentage of improvement in economic stability or growth to make almost any conceivable data collection
worthwhile" (i. e., the cost of collecting the relevant data would be
quite small as compared with the possible gain). But there are
practical limits: We must often analyze actual situations and arrive
at decisions on the basis of statistics collected in accordance with
some earlier estimate of what might be useful while at the same time
fiscal administrators and congressional committees are rarely as enthusiastic about statistics as are the economists and the statisticians.
So we must design our statistics to supply our analysts and policymakers with as much data as possible bearing on what we believe to be
the main questions ahead.
There has long been a strong, continuing interest in comparing
prices received by farmers with either prices paid by farmers or prices
of nonfarm commodities. As a result, we have now been calculating
and publishing monthly indexes of (1) prices received and (2) prices
and cost rates paid by farmers since the mid-1920's. I should now like
to consider these two indexes, omitting, however, any discussion of
the parity price calculations which also involve both indexes. There
is a double reason for this omission: These are the two basic farmprice measures irrespective of their use in the parity calculations,
while the parity problem itself has been recently discussed in a report
to the Senate as well as in the statement which I submitted to the Joint
Economic Committee only last November.1
PRICES RECEIVED BY FARMERS

The index of prices received by farmers as now published by the
US DA in our monthly report, Agricultural Prices, is a straightforward aggregative price index, calculated from the base 1910-14=100.
The weighting pattern for the period 1910 to 1935 is based on the
average volume of farm sales of about 50 commodities during 1925-29,
while from 1935 to date the weighting pattern is related to the average
sales of 52 commodities accounting for 92 percent of all farm cash
sales for the years 1937-41.
There are no really difficult problems associated with this index.
The commodity mix has not changed so radically over the years as to
raise serious questions as to comparability as between even widely
separated periods and except for truck crops, some fruits, and tobacco,
the accurate measurement of month-to-month changes is not too
difficult.
1
See : Possible Methods of Improving the Parity Formula, Report of the Secretary of
Agriculture Pursunnt to Section 602 of the Agricultural Act of 1956. S. Doc. No. 18, 85th
Consr., 1st sess., February 1, 1957 : and Orvis V. Wells, Parity Prices and Parity Income
Formulas, 1933-57. included in Policy for Commercial Acrriculture—Tts Relation to Economic Growth and Rtabilitv, papers submitted by panelists appearing before the Subcommittee on Agricultural Policy, joint committee print, November 22, 1957.

ECONOMIC STABILITY ANT> GROWTH

119

There are nevertheless some improvements that would be desirable.
We are now primarily dependent upon mail questionnaires for most of
our basic data. This leads to certain limitations—limitations which
can be removed only by providing for the direct enumeration of a set
of dealers selected by modern sampling methods. This would go far
to eliminate variations resulting from intermittent nonresponse of
mail reporters—i. e., "holes" in the basic sources of price information—and to insure improved representativeness and stability. The
data collection procedure for the Consumers' Price Index may be
looked upon as something of the pattern toward which we should
move. With this development should go the collection of prices for
more commodity breakdowns by main classes and main methods of
sale. Farm forestry products, which account for several hundred
millions of dollars of farm sales, might also be priced. Finally, it
would be desirable to shift both the official reference base and weighting pattern for the index to a recent, post-World War I I period.
Since American farmers exhibit a strong tendency toward continuing full production, although shifting substantially as between various
crops and classes of livestock depending upon relative returns, fluctuations in the prices received index have a relatively high correlation
with changes in cash farm sales. Further, since prices farmers receive
for most commodities are still flexible, changes in this index, when
measured relative to changes in the far more stable jndex of prices
and cost rates paid by farmers, give an approximate indication as to
the prosperity or well-being of commercial farmers (roughly, the 44
percent of farmers accounting for about 91 percent of all farm
products sold).
COSTS AND MARGINS

Changes in prices received by farmers are also one of the main
factors affecting the price of food at retail. However, what the housewife buys at retail is not the raw farm product but rather a combination of the farm commodity along with all of the associated assembling,
processing, transporting, and selling services. Since food costs are
one of the principal components of consumer expenditures, this means
that it is desirable to measure the costs and margins which intervene
between farmers and consumers in order that food costs can be broken
down into the farm and nonfarm portions. Such calculations are
facilitated to the extent that retail prices are collected on such food
items as to allow a good composite price to be calculated for comparison with the prices of the raw commodity at the farm level—this is
especially true for meat animals, for example.
We now estimate the value or retail cost of an average food basket
of products grown on American farms. That is, the average quantities of food purchased by an average-sized family, using the same
quantity food weighting pattern as used by the Bureau of Labor Statistics in calculating the Eetail Food Price Index, is compared with
the estimated farm value of an equivalent amount of the various commodities concerned. Such measurements are essential to understanding the forces affecting both short-run changes and longer term trends
in prices of food at retail.

120

ECONOMIC STABILITY AND GROWTH
PRICES AND COST RATES PAID BY FARMERS

Commercial farming is a business operation with the farmer's net
income depending not only upon prices received, the efficiency of his
own labor and management, and the effect of weather upon yields, but
equally upon the level of prices and cost rates paid. As a result,
efforts to measure prices paid by farmers, or comparisons between
farm and nonfarm prices, go about as far back as efforts to measure
changes in the average level of prices received.
Since the late 1920's the United States Department of Agriculture
has been calculating and publishing an index of prices paid by farmers
which has gradually expanded into an index of prices and cost rates
paid. Compared with the index of prices received by farmers, where
the statistical problems are relatively simple, the maintenance and
improvement of this index presents some difficult, complex problems.
As in the case of the prices received index, the prices paid indexes
are still calculated from the official base 1910-14=100. But whereas
the general mix of farm commodities sold has remained relatively
stable, this index now covers three considerably different periods as
it extends back to 1910-14.
The first period was characterized by the farming methods which
prevailed prior to World War I when horses and mules were almost
the sole means of farm power, and is represented in the current calculations by price series representing farmers' purchases for living
and production, excluding automobiles, trucks, and tractors, but including allowances for taxes on farm real estate, interest on farm realestate mortgages, and wages to hired farm labor.
Then there followed a transition period during which farmers
were shifting rather rapidly from horses and mules to automobiles,
trucks, and tractors. Accordingly, price series for automobiles, trucks,
and tractors, and various automotive supply items were incorporated
into the index as of March 1924. The weighting pattern prior to
January 1935 was based upon estimated average purchases and outlays
of farmers for the years 1924-29.
The third period starts in the late 1930's and so far as the index as
now calculated is concerned, extends from March 1935 to the present.
This period is characterized by the increased use of fertilizer and
spray materials, increasing wages to hired farm labor, and the substantial completion of the shift from horses and mules to automobiles,
trucks, and tractors. The weighting pattern for this period is the estimated average purchases and outlays by farmers for the years 1937-41.
Obviously, the measurement of price changes over periods in which
such wide differences have occurred in the actual mix of goods and
cost elements entering into farm production and farm family living as
have occurred during the last 50 years raises some very real statistical
problems. The direct and by all means the most satisfactory solution
to these problems is to shift the base and weighting pattern for the
index of prices and cost rates paid by farmers to a post-World War
I I period. This was one of the chief reasons for the Secretary of
Agriculture recommending to the Senate in January 1957 that the
official reference base of this index be shifted to the 10-year period,
1947-56.
Meanwhile, the Congress did make funds available for a comprehensive survey in 1956 of farmers' expenditures for both production
and farm family living purposes covering the calendar year 1955.

ECONOMIC STABILITY AND GROWTH

121

We are now analyzing these survey data with a view to calculating a
current weighting pattern for the prices-paid index. This should
be done whether or not the reference base is changed. But the statistical work would be simplified and the index strengthened if we could
dispense with having to trace back through three sets of linkage factors to the original base, 1910-14.
In addition to adopting a new weighting pattern and base period
for the index of prices and cost rates paid by farmers, it would also
be desirable (a) to expand the coverage of the index to include important service or commodity areas not now covered and to strengthen
the data for groups now inadequately covered, and (6) to adopt an
objective probability sampling procedure combined with enumerative
collection of data as already mentioned with respect to prices
received.
The index of prices and cost rates paid by farmers as now calculated
covers (a) prices paid by farmers for 191 items used in farm family
living, (&) prices paid by farmers for 199 items used in farm production, and (c) allowances for taxes on farm real estate and farm mortgage interest paid, and wage rates paid to hired farm labor. Actually, prices are now covered for 350 items or commodities since 40 of
them are common to both the family living and the farm production
subindexes. The prices-paid index is calculated and released on $
monthly basis but the wage-rate data are quarterly and annual rates
are used for taxes and interest.
The 1955 expenditure survey indicates that our current information
covers only about 81 percent of farmers' expenditures for commodities
used for both farm family living and farm production.
The chief farm family living items not covered are those for medical, dental, and hospital purposes which amount to over $1 billion, or
7 percent of total farm family living expenditures; expenditures for
personal insurance of about $400 million, or about 2.5 percent of total
living expenditures; and expenditures for recreation amounting to
$300 million, or about 2 percent.
In the field of farm production, classes of items not now covered
include machine hire and work paid for on a custom-rate basis, cash
rent, irrigation charges, and marketing expenses for crops and livestock. Representative prices or charges should be collected covering
these expenditures. Marketing expenses, for example, account for
about $500 million, or 2.5 percent of all production expenditures. In
addition, the coverage of some commodity groups already represented
in the index should be substantially expanded; e. g., containers for the
marketing of farm products and pesticides for the control of insects
and fungi.
Meanwhile, for many items used in the farm family living field, we
ask local merchants to report the average price of the kind of item—e. g., workshirts, overalls, shoes, flour, etc.—most commonly sold. For
the more costly items, we do ask for prices according to specifications.
We should increase the use of specification pricing although it seems
to me that the "most commonly sold" method has some advantages
which also need to be considered, especially when dealing with essential, everyday items. These problems of coverage and the use of specification pricing relate in considerable part to the problem of available
funds. We are now mostly using mail questionnaires. Widening the

122

ECONOMIC STABILITY AND GROWTH

coverage and enumerative pricing on an objective sampling basis
would considerably increase the costs of our price work.
INDEX OF FARMLAND VALUES

One of the price indexes to which I would like to call the committee's attention is the index of average value per acre of farm real
estate. Farm real estate accounts for about 70 percent of the value of
farmers' nonfinancial assets, and changes in market values and rates of
transfer serve as indicators of the general economic position of agriculture. Although, technically speaking, the index measures only
changes in market prices of one productive factor, land, the nature of
this resource is such that its price also reflects the confidence or judgment of farmers and others who invest in farmland as to the longer
run prospects of agriculture. Consequently, the broad movements that
occur in land prices often have significant implications with respect to
agricultural credit, rural welfare, and the attainment of f armownership.
The regular crop reporters of the Department are the primary source
of basic data used in constructing the index of farm real-estate values
by the Farm Economics Research Division, Agricultural Research
Service. They provide estimates of prevailing market values in their
localities as of March 1, July 1, and November 1. Index numbers are
computed for States, type-of-farming areas, and special groupings of
States as well as for the United States as a whole. Special mail surveys are also directed twice each year to about 10,000 farm real-estate
dealers, bankers, lawyers, and others in close contact with local market
developments. Data from these surveys provide a check against the
crop-reporter estimates as well as a general appraisal of local supply
and demand conditions, availability of credit, types of sellers and
buyers, and related information.
Although the National and State indexes are sufficiently reliable for
most purposes, a larger number of reports, particularly by class or type
of land, would improve the accuracy of the index in certain States.
Steps have been taken in recent years to improve the reporting procedures in New England, Florida, Nebraska, and California with this
end in mind. However, a need exists for information as to the level
and trend in market prices for areas smaller than States which cannot
be met by present reporting procedures. By necessity, a State index
often averages significant variations in price movements within the
State. A substantial increase in the number of reports for relatively
homogeneous areas within States would be necessary to permit the
calculation of indexes below the State level.
n
Price policy is chiefly a means to an end and price indexes themselves
rarely measure the final results in which we are chiefly interested. For
this reason, it seems necessary to call attention to the need for improving our estimates of farm income and farm expenditures as well as
the need for breaking down farm income estimates as between different
classes of farms. At the same time, attention is called to the fact that
all of the data which we use in deriving average prices and price indexes are also useful and necessary in estimating farm income.

ECONOMIC STABILITY AND GROWTH

123

Formerly, farm expenditure and net farm income estimates were
made for the United States as a whole with no breakdown either as
between different classes of farms or as between the various States and
geographic divisions. However, the census of agriculture now classifies farms and tabulates value of production, farm expenses, and related data by farm class. In addition, funds are now available for
estimating farm expenditures and farm operators' net income on a
State basis. The committee will also recall that during the recent
hearings on Policy for Commercial Agriculture, Messrs. Nathan Koffsky and Ernest Grove of the Agricultural Marketing Service were
asked to prepare a paper breaking down the national farm income
estimates so far as possible between low-production and high-production farms (i. e., farms with annual farm sales of $2,500 or more) for
the years 1946 through 1957.
We need (1) to strengthen our current farm income estimates
wherever possible, especially the estimated income of farm people
from nonfarm sources, (2) to substantially improve our estimates of
farm production expenditures both by States and for the United States
as a whole, and (3) to find ways of breaking down our annual farm
income estimates as between classes of farms, especially commercial
versus noncommercial farms. Good farm income estimates properly
broken down as between the various States and as between commercial
and noncommercial farms would, along with a revised index of prices
paid by farmers, yield a much better judgment than is now available
as to the relative well-being of farm people.
Since the committee is chiefly interested in price policy, I shall not
endeavor to go into all of the various problems having to do with the
estimating of farm income and farm expenditures but they are considerable, and we also need to find some way of getting more adequate
annual estimates of both the numbers of farms and of farm
population.
So far as the aggregative farm income estimates are concerned, the
accompanying table does set forth the main aggregates in which we
are interested and to which our attention should be directed. Since
farm income may be looked at in several different ways, it may be
worthwhile to briefly define what each of these aggregates measures:
Cash receipts from farm sales or marketings measure the value
of farm products sold by farmers during the calendar year.
Gross income from agriculture represents the total value of
commodities and services produced by farms in the United
States, without any deduction for production expenses. In addition to cash receipts from farm marketings, gross farm income
also includes direct Government payments to farmers, the value
of the net change in inventories of farm products, and noncash
income—i. e., estimated allowances for food and fuel consumed
directly in farm households and the rental value of the farm
dwelling.
Production expenses comprise the aggregate costs paid out or
incurred by farm operators for production purposes. Current
farm operating expenses include wages paid hired farm labor,
both in cash and in kind, purchases of feed, livestock, seed, fertilizer, outlays for repair and operation of farm buldings, motor
vehicles and other machinery, etc. Charges are also included
for depreciation of motor vehicles and machinery, taxes levied

124

ECONOMIC STABILITY AND GROWTH

on farm property, interest paid on farm mortgage loans, and
net rents paid to nonfarm landlords. Production expenses do
not include an imputed value for labor of the farm operator and
his family or for land and other capital owned by the operator.
Eealized net income of farm operators from iarming is the
balance which remains when production expenses are deducted
from realized gross farm income. That is, it represents what
farm operator families have available during the year for farm
family living and savings on the basis of actual farm sales or
family use of farm-produced commodities or services, including
use of the farm dwelling. It is not adjusted for the value of the
net change in farm crop inventories or numbers of livestock.
Total net income of farm operators from farming differs from
realized income by the value of the net change in farm inventories. The changes in the physical volume in the various crop
and livestock items are valued at the average price of the commodity for the year. This adjustment is added to or subtracted
from realized gross or net income of farm operators to convert
it to total gross or net income. Both the realized and the total
estimates measure returns for the labor and management of the
operator and his family and the return on land and other capital
owned by the farmer.
Net income to all persons on farms from farming adds wages
paid for farm work to hired farmworkers living on farms to total
net income of farm operators from farming. Income to persons
on farms from all sources adds income received by persons on
farms from sources other than agriculture to income from agriculture. Farm families on the average receive a fourth or more
of their total net income in the form of earnings from employment in nonagricultural occupations or as returns from investment in nonfarm property. Hence the measurement of these
items is obviously important as a supplement to the regular
measures of farm income.
Each of these aggregative or total measures has its appropriate
place or use, and the net income estimates when divided by the estimated number of farms or farm population give a much more accurate measure of changes in funds available to farm families than
do any of the price indexes. The price indexes do, however, influence and help forecast changes in income, while the index, or subindex, of prices paid by farmers for goods and services used in farm
family living is also our best measure of changes in the purchasing
power of the farmer's net income dollar.
INCOME BY TYPES AND SIZES OF FARMS

In addition to strengthening our aggregative measures of farm income and findings ways of breaking them down as between parttime, small-scale, and commercial farms, we also need supplementary
analyses indicating the changes that are taking place with respect to
different types of commercial farms.
The best approach to this that I am so far acquainted with is the
costs and returns series for specific types of farms (e. g., central
Illinois cash grain farms, Corn Belt hog-beef fattening farms, Cen-

ECONOMIC STABILITY ANTD GROWTH

125

tral Northeast dairy farms, central Kentucky tobacco farms, Southwest cattle ranches, Delta cotton farms, etc.) which are now maintained in the Farm Economics Research Division, Agricultural Research Service.
On the basis of census, special survey, and other data, representative commercial farming systems are derived for specified types and
sizes of farms within given areas. Estimates for these representative
farms are then made from year to year as to changes in organization,
crop and livestock production, sales, expenses, net income, and related
information. These series provide information on year-to-year
changes as well as an appraisal as to differences among various types
and sizes of commercial farms in major producing areas. So far,
such series are available for about 30 types of farms m 16 of the more
important type-of-farming areas. To be most useful, the number of
farm types or areas for which such series are calculated needs to be
sufficiently increased to be representative of broad types of farming
(cotton, dairy, etc.) for the United States as a whole. Also, the collection of data from current surveys should be strengthened with
respect to such items as hired labor, fertilizer, pesticides, farm machinery, and building repair and improvement.
# * *
In discussing the various price and income measures in sections I
and I I above, I have tried to indicate what the main measures are and
discuss briefly the improvements that should be made. The views
expressed are personal rather than official. For more detailed information, the committee is referred to volume 1, Agricultural Prices
and Parity; volume 3, Gross and Net Farm Income; volume 4, Agricultural Marketing Costs and Charges; and volume 6, Land Values
and Farm Finance, of Agriculture Handbook No. 118, Major Statistical Series of the United States Department of Agriculture; How
They Are Constructed and Used. United States Department of
Agriculture, 1957.

126

ECONOMIC STABILITY AND

GROWTH

Number and percentage of farms and proportion of market sales, by economic
class, United States, 1954
Economic class

Commercial farms:
With sales of $2,500 and over:
Class I
Class II
. .
Class III
Class IV

Value of sales

$25,000 and over..
$10,000 to $24,999
$5,000 to $9,999
$2,500 to $4,999

Total
With sales of less than $2,500:
Class V._
Class VI
_
_

$1,200 to $2,499
$250 to $1,999 i

Total
_

$250 to $1,199 i
Under $250

Percentage
of all farms

Percentage
of market
sales

Thousands

Percent

134
449
707
812

2.8
9.4
14.8
17.0

31.3
26.9
20.5
12.1

2,102

44.0

90.8

763
462

16.0
9.7

5.7
1.4

1,225

25.7

7.1

575
878
3

12.0
18.3
.1

1.5
coco

Other farms:
Part-time
Residential
Abnormal 2

Number of
farms

Total

1,455

30.4

2.0

All census farms

4,782

100.0

» Farms with sales of $250 to $1,199 were classified as part-time if the operator worked off the farm as much
as 100 days, or if other income of the operator family exceeded farm sales.
* Public and private institutional farms, experiment stations, and S3 on.
Based on Bureau of t h e Census data. Adapted from table 2 of Family Farms in a Changing Economy,
Agriculture Information Bulletin No. 171, March 1957, USDA, ARS.

Selected data relating to farm prices and incomes, United States, 1939 and 19J/.6-58
Food market basket 3

Prices received and paid
(1910-14=100)

Year

1939
1946
— _
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957 5
19f6: 4th quarter
1957:s
1st quarter
2d quarter
3d quarter
___
4th quarter
1958: 1st quarter

Prices
Prices paid
received
b y farmers' or parityindex i
index

Parity
ratio
(percent) 2

95
236
276
287
250
258
302
288
258
249
236
235
242
234

123
208
240
260
251
256
282
287
279
28:1
281
285
296
289

77
113
115
110
100
101
J07
100
92
89
84
82
82
81

237
243
247
241
254

294
296
295
298
302

81
82
84
81
84

Farm
value

a $172
(I)

467
497
435
432
497
482
445
421
395,
390
400
393
385
395
414
407

Marketing
margin

Measures of income from farming or to farm people (billion dollars)

Cash
receipts
from farm
sales <

Realized

485
493
488
527
552
558
565
574
582
607
590

7.9
24.8
29.7
30.3
27.9
28.4
32.9
32.6
31.2
29.9
29.5
30.4
30.0
30.9

10.6
29.3
34.0
34.6
31.6
32.1
37.1
36.7
35.1
33.7
33.2
34.4
34.4
35.2

6.2
14.3
16,8
18.6
17.9
19.2
22.3
22.5
21.2
21.5
21.6
22.3
22.9
22.6

4.4
15.0
17.2
15.9
13.7
12.9
14.8
14.3
13.9
12.2
11.6
12.1
11.5
12.6

597
607
616
607

30.3
30.0
29.8
30.0

34.5
34.4
34.3
34.5

22.8
23.0
22.8
23.0

11.7
11.4
11.5
11.5

a $279

1 Index of prices and cost rates paid by farmers for goods used in faim family living and
production including allowances for hired labor and taxes and mortgage interest paid per
acre of farm real estate
2 Index of prices received expressed as a percentage of the index of prices paid.
3 The market basket incluies estimated quantities of United States grown foods purchased per urban worker family in 1952. 1939 estimates based en same market basket as
for 1947 to date. Comparable data for 1946 not available.

Gross farm
Production
expenses *
income 4

Net income of farm
operators from
farming*

Total
4.5
14.9
15,5
17.7
12.9
13.7
16.1
15.1
13.3
12.7
11.9

n.6
12.1
12.0
12.0
12.1
12.2
12.2

Index of farm
Income to persons land values
on farms
(1912-14=100)

From
farming

From ail

5.2
16.7
17.4
19.7
14.7
15.5
18.0
17.0
15.1
14.4
13.6
13.4
13.9
fl
()

7.7
21.0
22.3
24.8
19.9
20.8
23.6
23.1
21.1
20.2
19.9
20.1
20.2

(«)
(6)

sources

(«>

82
141
157
170
177
174
200
221
221
216
224
232
247
?241

(6
(6

8 247

•253
7 259

I
3

* Quarterly data are seasonally adjusted annual rates.
« Preliminary.
* Quarterly data not available.
* Nov. 1.
s Mar. 1.
»July 1.
Compiled from Agricultural Marketing Service and Agricultural Research Service data.

128

ECONOMIC STABILITY AND GROWTH

FARMERS9 PRICES
% OF 1910-14
300

200

100
% OF PARITY

150

Prices received

100" •••••••••••••- ••v
50
1910

1920

1930

1940

1950

* MONTHLY DATA
A

INCLUDES

INTEREST, TAXES. AND WAGE RATES. ANNUAL AV DATA, 1910 23,
BY QUARTERS. 1924-36, BY MONTHS, 1937 TO DATE

U. S. DEPARTMENT OF AGRICULTURE

The indexes of prices received and prices and cost rates paid by farmers are
shown in the first section of the above chart. Note that while the index of
prices received by farmers remains relatively flexible, the index of prices and
cost rates paid by farmers has become increasingly inflexible, showing no real
decline from 1952 to date. The relationship between prices farmers receive and
the prices and cost rates they must pay—commonly referred to as the "parity
ratio"—is shown in the second section of the chart.

FARM FOOD MARKET BASKET0
Farm Value and Marketing Margin

DOLLARS

Marketing margin
Farm value

600

1939*'47'48'49'50'51 '52'53'54'55'56'57'58'59
° BASED ON ESTIMATED QUANTITIES OF FARM FOOD PRODUCTS BOUGHT BY URBAN WORKER FAMILIES, 1952
* ESTIMATES BASED ON SAME MARKET BASKET AS 1947 TO DATE

U.S. DEPARTMENT OF AGRICULTURE

NEG. 6008-58(3)

AGRICULTURAL MARKETING SERVICE

ECONOMIC STABILITY AND GROWTH

129

The above chart breaks the retail value of an average American family's
annual purchases of foods derived from American farms into its two basic components—that is, the farm value of the equivalent raw commodities and the total
costs and margins which intervene between farmers and consumers. Note that
from 1951 into 1956 marketing costs rose relative to farm prices so that the
farmer's share of the retail food dollar declined from 49 percent to 40 percent.
In 1956 and again in 1957, 60 percent of the cost of the food market basket was
accounted for by marketing charges. Falling farm prices tended to offset the
increase in marketing costs and margins from 1951 through 1956, but with some
increase in farm prices in 1957, along with another increase in marketing costs,
prices of food at retail also increased.

INCOME
OF FARM
OPERATORS

$ BIL.

Realized
gross*

40
30
PRODUCTION
„ EXPENSES!!

20
10

1930

1935

1940

1945

1950

1955

1960

• EXCLUDING INVENTORY CHANGE, INCLUDING GOVERNMENT PAYMENTS
U . S . DEPARTMENT OF AGRICULTURE

NEC 443-58(3)

AGRICULTL

The above chart traces the realized gross and net income of farm operators
from farming from 1930 through 1957 and indicates the increasing importance
of production expenses as a determinant of net farm income. The rise in production expenses reflects two main influences: First, technological developments
have substituted machines and other industrial products for human labor which
has brought a high, relatively inflexible cash cost structure to modern American
agriculture. Second, persistent inflation during the last decade has had a
more basic effect on farmers' costs than on prices of products sold by farmers.
In 1957, production expenses accounted for 2 out of every 3 dollars received by
farmers from farming operations, as compared with a ratio of only about 1 out of
every 2 dollars in 1947-49.

Ill
PAST PEICE BEHAVIOR VIEWED IN THE CONTEXT OF
CYCLICAL AND SECULAE ECONOMIC CHANGES

131

23784—58

/ / / . Past price behavior viewed in the context of cyclical and secular
economic chmiges
A. What have been the general price movements based on
the various available indexes? What cyclical and
secular economic changes have been associated with
these price movements ?
B. In studying trends, particularly those of recent years,
what has been the relationship between price changes
and changes in the cost of the various factors of
production—
1. To what extent have price changes preceded or
lagged behind changes in labor costs ?
2. To what extent have price changes exceeded
changes in labor costs?
3. What has been the effect of changes in capital
costs—i. e.? interest rates and other costs—on
prices?
132

THE BEHAVIOR OF PRICES, 1890-1940
Clarence H. Danhof, Tulane University
Information on the movements of prices over time plays an indispensable role in the analysis of economic fluctuations. When employed in conjunction with other indexes of business activity, price
series contribute to the appraisal of business conditions. In short-run
situations, price trends provide guides in suggesting probable economic developments. Prices play an essential role in the interpretation of data on production and income values and are similarly indispensable in appraising the effectiveness of the economy in supplying
real goods and services. They are useful in measuring changes in the
relative positions of economic groups. The internal structure of
prices—of factors and end products—also invites attention. The
possibility exists that analysis of the changing pattern of prices within
the total structure may uncover recurring relationships that will contribute and assist in the development of control programs.
This paper reviews briefly some of the more significant aspects of
price level movements and certain relationships within the structure of
prices from the point of view of their relevance to the promotion of an
economy operating at maximum effectiveness. Many aspects of the
subject will be treated lightly since they will be dealt with more fully
elsewhere in this compendium. The period covered is the half century, 1890-1940. The paper is primarily a summary of observations
and conclusions of recent research.
ROLE OF PRICES

A price originates as the money side of a single exchange action involving a buyer, a seller, and a specific good or service. The fundamental function of prices is to make possible the appraisal by buyers of
the goods and services available, and the appraisal by sellers of opportunities to offer goods for sale. The system of prices acts to give
producers minute information as to the terms of availability of production factors as well as to measure the effective desires of purchasers.
Prices are thus the mediums by which the resources allocating and
product determining functions are carried out in a free enterprise
system. In performance of this service, specific prices are constantly
changing, displaying varying degrees of sensitivity, shifting relationships one to another with varying frequencies and differing
amplitudes.
To effectively analyze prices, whether at a point in time or over a
period, they must be merged or consolidated. Thus, we may consider
the prices of groups of commodities, classified in some logical relationships as to origin or use; prices of factors of production as contrasted
with end-use goods; and prices as a totality, which when viewed as a
historical series, constitutes the concept of the price level. Such

133

134

ECONOMIC STABILITY AND GROWTH

groupings may reveal significant trends, obscured in the great mass
of historical data.
Such groupings are, however, no longer true prices. Each item in
such a group of price measures as, for example, the BLS wholesale price
series shown in the table, must be envisioned as an abstraction. It is
an artificial summarization of a multitude of prices which over the
period covered have moved in opposite directions and with varying
amplitudes. It is obvious that much of what has actually happened
is lost in reducing such activity to a single figure; on the other hand,
the essential meaning in the complex of changing prices is approached
in a sufficiently simplified form to provide a tool of wide applicability.
The changes found in such measures of the general price level may
be usefully classified according to the period of time involved, or to
conform to some change which is to be interpreted. For our purposes the importance of price movements relates to changes in economic activity and our primary interest is in the referral of price
movements to the phases of the business cycle. There exist also
movements over time periods longer than the cycle.
SECULAR TRENDS

The 30 years from the close of the Civil War to 1896 were characterized by a persistent decline in the price level. The wholesale
index which stood at 116.3 in 1866 (1926=100) fell to 46.6 in 1896.
This downward movement was remarkably continuous, interrupted
only by a few short upturns and a few years of stability.
From the 1896 low, prices turned upward, moving sharply until
1900, and then more slowly until 1915, at which time prices were
50 percent above the low. War conditions stimulated further sharp
increases, prices reaching in 1920 the highest levels of this half
century.
The peak was short lived. Precipitous declines occurred in late
1920 and in 1921, the fall in the wholesale index from 154 to 98 being
by far the sharpest short-term decline in the period. The movement
was remarkable for its brevity. Stability was achieved in 1924 at
levels well above prewar.
There was relatively little movement in prices in the 5 years following. This period of stability was interrupted by a renewed downward movement; the severe declines which characterized the early
thirties. The low point came in 1933. Since that time the price level
has moved consistently upward, interrupted only by mild reaction
in 1937-38.
The causes of such long-term secular movements are not likely to
be found in analysis of the internal structure of prices. Long-term
trends in the prices of individual commodities usually reflect technological developments, changes in raw material supplies, or in consumption patterns and institutional relationships. Long-term swings
in the general price level are generally associated with changes in
external factors such as wars, technological developments, and institutional changes, particularly those relating to the currency supply.
Though the relationship of secular to cyclical changes is the subject
matter of a considerable literature, no significant generalization seems
possible at this time.

ECONOMIC STABILITY AND GROWTH

135

PRICES IN CYCLICAL CHANGES

During this half century 14 cyclical movements in economic activity
may be identified following the criteria developed by the National
Bureau of Economic Kesearch. Cycle peaks and troughs are detailed as shown on the table, and the data grouped accordingly.
Some of these cycles were periods of mild adjustment, movements
of output and prices ranging narrowly. The major disturbances were
those of 1893-97 (two cycles), 1902-4,1907-8,1913-14,1920-21,192933, and 1937-38.
Significant declines in production and employment occurred in some
of the pre-World War I contractions but price changes were in most
instances relatively modest, declines being moderate and quickly recovered. During the contractions of 1899 and 1904 the prevailing
rising trend of prices persisted although dampened in 1904 by declines concentrated in the metals and chemicals groups. From 1904
to 1907 metals were much the most volatile price group; suffering the
largest decline in 1908. The contraction of 1908 was accompanied
by a severe financial crisis, heavy unemployment, and reduced production. The decline of prices was, however, moderate and of brief
duration.
The contraction of 1911 was accompanied by a general price decline
of 6 or 7 percent and a recovery within a year to 1910 levels. The
briefer but more severe contraction of i914 made little impression on
the general price level though declines in metals and fuels were substantial.
Manufacturing activity rose rapidly after 1914, reaching a peak in
1918. Prices followed behind, at first slowly, and then more rapidly,
continuing to rise long after manufacturing activity was contracting.
The peak was reached in 1920 at a level more than double that of
1910-14, representing an expansion proportionately far greater than
occurred in physical production. The rise in prices was shared by
all categories though there were marked differences in the timing of
peaks and course of the subsequent declines. Chemicals reached a peak
in 1918, farm products in 1919, and metals and building materials in
1920.
The 1921 contraction, though brief, was one of the most severe up to
its date, measured in declines in production or employment. It is
noteworthy for the precipitous decline in prices, a large part of the
increase of the preceding 5 years being wiped out in an 18-month
period. Metals experienced the smallest group decline in 1921 but
continued to fall in 1922. All other price groups achieved stability in
1922 at a level about 30 percent above prewar.
Some recovery in prices was evident in 1922, the upward movement
continuing in 1923. A mild contraction in 1924 proved a minor and
temporary setback; prices continued to press upward, reaching a postwar high the following year, the all commodity index standing at 103.
A mild contraction in 1927 was accompanied by widespread price
drops which reduced the all commodity index by 5 percent. Stability
was characteristic of 1928 and 1929, but a decline of about 10 percent
occurred in 1930. Further declines in 1931 and 1932 brought the allcommodity index down to 65 in 1932 and early 1933, a drop of roughly
one-third in 6 years. Farm products, textiles, fuel, and lighting ma-

136

ECONOMIC STABILITY AND GROWTH

terial led the decline in time and in magnitude; metals and building
materials evidenced the greatest resistance to the trend. Prices of raw
materials in 1932 were 45 percent of 1926 levels while manufactured
commodities were off 35 percent.
Marked recovery of prices occurred in 1935 and 1936, leveled off in
1936 and resumed in 1937. Increases were most marked in farm products. A renewed decline then pushed prices back to 1934-35 levels,
with farm products and raw materials once again suffering the greatest reduction. Metals and building products successfully resisted the
reversal. Prices in 1938-40 remained stable at these levels until 1940,
although farm products and raw materials suffered further declines.
A much more detailed analysis of price history of this period is
readily possible. This brief review serves here as background for
analysis which utilizes far more detailed information. Such analysis
suggests the operation of some simple but fundamental relationships.
It seems reasonable to conclude, for example, that there is no such
thing as a "normal" price level; at least none seems to have existed
in this half century. There have, however, been glimpses in this
period of a desirable structural complex of prices: that is, a pattern of
interrelationships which operate in such manner as to permit or induce
a rate of maximum output, that maximum being established by
resource availability.
It also seems clear that increases in the general price level occur
when output is moving to higher levels. As the volume of output
approaches the capacity of the system, strains develop, and very rapid
price increases of considerable magnitude occur. On the other hand,
declines in prices are general when productive activity declines. It
would seem to be true that many prices can resist declines with sufficient strength so that very sharp reductions in activity may occur
with only modest general price reactions. Perhaps of greatest significance is the fact that a low level of output can apparently be
maintained alongside of prices stabilized at what seem to be disproportionately high levels.
A variety of relationships between prices and output exists within
the price level. In agriculture, for example, production is more responsive to physical than to market forces. In such commodity groups
as metals or building materials, there appears to exist, on the other
hand, a substantial ability to resist price declines despite disproportionate drops in production.
SOME CHARACTERISTICS OF PRICE MOVEMENT

Economic output consists of goods and services valued at prevailing
prices, measurable as an aggregate because of prices. Over time the
aggregate value varies as a result of changes in both the prices and!
quantities involved. There is, of course, very great diversity in the
magnitudes of price and quantity changes but the role of prices in
such changes can be roughly identified.1
In the earliest phases of an upturn, increases in quantities seem
completely responsible for expanding activity in nondurables, farm
products, and consumer products. Prices rise but lag and quantity
increases dominate the continuing expansion in all commodity groups.
1
These paragraphs are based upon Frederick C. Mills, Price-Quantity Interactions in
Business Cycles, National Bureau of Economic Research, New York, 1946.

ECONOMIC STABILITY AND GROWTH

137

Thereafter price rises become a weightier factor in the increasing
aggregate of business activity. Increased capacity is being brought
into production and quantities increased correspondingly, but price
rises continue dominant. In the final stages of business expansion,
quantity factors become increasingly important, with accompanying
weakness in prices. As contraction develops, prices weaken more
slowly than quantities decline. Changes in physical quantities tend
to reverse the direction of business movement more frequently than
do prices.
Price movements differ as to the frequency of change, in the duration of a change in one direction, the magnitude or amplitude of the
change, and in the timing of the movement with reference to other
prices.
Some price series show daily changes as do many common stocks.
Others change less frequently: wheat prices characteristically changed
about once per month in the period 1890-1925. Still other prices
remain fixed for much longer intervals. Steel prices in the same
period changed once every 9 months. More recently automobile prices
have characteristically changed only with annual model changes.
Moreover the magnitude of such changes vary. For example, the
annual mean deviation of the price of steel rails was only 2.4 percent
in the 1890-1925 period; in contrast, coke prices over the same period
deviated 14.5 percent from the annual mean.
The amplitudes of price movements are widest for goods at the
raw material level and at the consumption levels, less m the intermediate categories. Correspondingly, variations in quantities are
greatest in such groups as durable goods. The sensitivity of prices
has been intensively investigated by F. C. Mills. Mills concluded
that there was a "definite tendency toward a decline in 2the variability of commodity prices during the period 1890-1912." With reference to the years 1922-25, Mills believed that price variability was
greater than in the earlier period due to the operation of war induced
forces. The nature and significance of variability or flexibility of
prices aroused a great deal of interest in the 1930's, the typical point
of view being that increased
sensitivity of prices would produce greater stability in production.3 The subject is discussed elsewhere in this
compendium.
From the point of view of the problems of identifying and measuring cyclical movements, these differences in variability are significant.
The movements of the BLS Wholesale Price Index of Basic Commodities have been determined by the National Bureau of Economic
Research as leading cyclical
turns by 2.6 months at the peaks, and
3.2 months at the troughs.4 The BLS Index of all wholesale prices
other than farm products or foods, moves in rough coincidence with
the cycle. Other price groups lag severely, as in the case of metals.
On the whole, price series are useful in cycle prediction only in support of other types of indicators.
2
Frederick C. Mills, The Behavior of Prices. National Bureau of Economic Research,
New
York, 1927, p. 156, of 1957-58.
8
A review of the literature is conveniently available in Kenneth D. Roose, The Economics
of Recession and Revival. An Interpretation of 1937-38. Yale University Press, New
Haven, Conn., 1954, ch. 9.
* Geoffrey Moore, Statistical Indicators of Cyclical Revivals andi Recession, National
Bureau of Economic Research, New York, 1956.

138

ECONOMIC STABILITY AND GROWTH
PRICES AS COSTS

Up to this point our concern has been with certain characteristics
of the movement of the price level. Specific product prices reflect
the conditions which bear upon the supplying of the goods, and particularly the prices that prevail for labor and raw materials. In the
long run the relationship between the costs of the factors of production and the prices at which products are sold, must be such as to
leave to the producer a profit sufficient to permit and induce him
to continue or make indicated adjustments in his activities.
The intensity of resource use, and hence the volume of production of the individual firm, is influenced by its profit position, particularly by the direction of change of profits. Such profit experiences
and anticipations are a function of the relationship between the quantities which can be sold at prevailing or anticipated prices and the
trend of the cost of providing such quantities. The shifting relationship between product and factor prices invites attention as offering
an approach to the problems posed by cyclical fluctuations.
There is evidence that the preeminent student of the business
cycle, Wesley Mitchell, entertained some such general hypothesis.
Profits of some sectors of the business world declined more quickly
than those of others, responding to rising costs. Such reduced profits—though profits in general might be rising—would produce declines in capital expenditures, inventory liquidations—and in other
ways might induce a retrenchment. Localized, such developments
might have little, if any, effect, but if such
adjustments became sufficiently widespread, a cycle is established.5
RAW MATERIALS

Movements of raw materials prices can be judged only generally for
the years 1890-1912. Thereafter BLS wholesale price groupings of
raw materials, semimanufactured goods and manufactured products
provide more suitable aggregate data.
Raw material prices since 1913 to 1940 have characteristically moved
in advance of prices of fabricated products in both advancing and
declining phases of the cycle. They have also varied more widely.
Wholesale prices of semifabricated articles, on the other hand, have
tended to rise higher and fall more slowly than manufactured goods.
On the basis of a detailed examination, F. C. Mills concluded that
changes in materials costs played a neutral part in price movements
of 1914-29. Though raw material costs rose somewhat relative to
prices of fabricated products, increasing efficiency in operations6
served to maintain material prices in relation to finished products.
There is no reason to believe that any significant and persisting
change in relations between raw material and finished prices operated
during this period.
WAGES

Wages are a very large part of aggregate production costs. In
specific industries, wages may play a smaller role in costs but, never5
Wesley C. Mitchell, Business Cycles, University of California Press, Berkeley, 1913,
pp.8 457-514, 562-569.
F. C. Mills. Economic Tendencies in the United States, National Bureau of Economic
Research, New York, 1932, pp. 104, 216, 388-389, 414.

ECONOMIC STABILITY AND GROWTH

139

theless, typically a critical one. As changes in the wage element in
production costs differ from product prices, they may constitute a
factor responsible for the amplitude of the aggregate cycle.
Many influences bear, of course, on the labor costs of production.
Aside from hourly costs, there are such factors as labor utilization
under or over the standard weekly hours, fringe costs, quality, and
efficiency of the labor force. The wage rate is, however, probably
the most important factor, particularly when viewing economic activity in the aggregate.
Data on average hourly earnings as shown in the table are available,
although no single series covers this half century. Dealing with the
period 1919-1938, Creamer and Bernstein conclude that the data on
average hourly earnings probably reflect
wage costs with sufficient
accuracy for general analytical purposes.7
Average hourly earnings fluctuate much less, relatively, than manufacturing activity, wholesale prices, and with some exceptions, than
prices of semifinished goods. The movements in wage rates lag well
behind changes in business activity and in employment. In the
period 1919-38, the lag in wage-rate movements, whether for manufacturing as a whole or for component industries, was generally in
excess of 6 months and an average 9 months behind business activity.
If major turning points are considered the lag was 7 months. Wage
rates turned on an average 10 months later than factory employment.
Moreover, wage rates do not typically decline to the levels of other
measures such as wholesale prices; rather, prices appear to move up
to meet the weakening wage rates. In the case of prices of finished
manufactured products, the experience is more fixed. However, in
general, prices for finished products fluctuate more widely than earnings, particularly on the downswing. Earnings show some tendency
to rise more than prices on the upswing.
CAPITAL

Consideration of capital costs involves a wide range of intimately
interrelated factors. To mention a few, these include the policies of
the banking system with regard to currency supply, the public's propensity to save, business profit expectations as contrasted with the
desirability of liquidity, the relationship of price movements to debt
burdens, the relationship of product prices to prices of producer durables, and many more.
Interest rates on short-term commercial paper are shown in the
table, and data on other types of capital are available. The influences
reflected in such prices are, however, exceedingly complex. Probably
no relationship between commodity prices and capital costs—our
principal concern—can be identified
in this period in such a way as to
be useful for analytical purposes.8 The costs of capital are influenced
on both the supply and demand side by numerous forces unrelated
to specific prices though reflecting broad price movements.
Fortunately, it can be argued that the price of capital is rarely as
important a consideration in business calculations as the costs of other
7
Daniel Creamer and Martin Bernstein, Behavior of Wage Rates During Business Cycles,
Occasional Paner 84, National Bureau of Economic Research, New York, 1950.
s Frederick R. Mncanlny, The Movements of Interest Rates, Bond Yields, and Stock Prices
in t*>» United States Since 1856, National Bureau of Economic Research, New York, 1938,
ch. VI.

140

ECONOMIC STABILITY AND GROWTH

productive factors, and that it is frequently sufficiently small to be
of no determinate significance.
PROFITS

In a free-enterprise economy, the entrepreneurial function of committing resources to production hinges upon the profits, which may
be anticipated in the light of interpretations of recent experience.
Profits are, hence, a social cost of production, and a brief resume of
profit experience in recent cycle history is relevant. Quarterly data are
necessary to meaningful analysis, and are available for the period
1920-38 for a sample of corporations principally engaged in manufacturing and mining.9 Aggregate profits of the corporate sample
move with changes in industrial production. Such profits have the
significant distinction of rising and falling more dramatically than
any other of the comprehensive indicators of the National Bureau of
Economic Research. The amplitude of movement contrasts as follows:
Average amplitude of movement—Percentage

Corporate profits
Industrial production
Wholesale commodity prices...

of mean

values durin g a cycle1
Rising stage

Falling stage

168.8
35.2
8.7

174.6
32.5
8.9

i Arthur F . Burns, New Facts on Business Cycles, 30th Annual Report, National Bureau of Economic
Research, New York, 1950, p. 20.

Changes in profits are the resultant of movements of product prices,
factor costs, quantities sold, or some combination. The average amplitude of prices is substantially less than for quantities, as the latter
is reflected in such an indicator as industrial production. At the level
of the firm, it is typically true that changes in volume of sales, rather
than in prices, are the major factor in a change in profits. The firm
has no control over sales except insofar as it can achieve a superior
price-cost relationship. The possibilities of achieving a superior cost
structure is limited, since the firm bids for factors in a competitive
market. Faced with a changing volume of sales, the firm may adjust
by changes in price, within a limited range, or by varying its production rates.
A downturn in business activity usually reveals itself to a business
firm as resistance to purchase at prevailing rates; i. e., as a decline
in the volume of goods that are taken. The consequence is an unplanned buildup of inventories of finished products and, probably, of
semifinished goods and raw materials as well. The firm reacts by
reducing its purchases of materials, labor, and, most of all, of producers' equipment. Selling prices may or may not be reduced, depending upon inventories, and the likelihood of reestablishing volume
above its break-even point.
Available data suggest that the reactions of the firm also describe developments in the aggregate economy. There are, however, important
exceptions. The aggregate data obscure the fact that some firms will
almost invariably be experiencing trends opposite to those which
9
Thor Hultgren, Cyclical Diversities in the Fortunes of Industrial Corporations, Occasional Paper 32, National Bureau of Economic Research, New York, 1950.

ECONOMIC STABILITY A3STD GROWTH

141

characterize the whole. At any point in time, there will be a group
of firms within a given industry enjoying rising profits, another
suffering from declining returns, if not losses, and a third enjoying
stable returns. From the point of view of the cycle, declines in profits
will occur for an increasing number of firms well in advance of the
peak of aggregate profits. The opposite is similarly true; an increasing number of firms will have expanding profits well before the low
point is reached in the profits of all firms.
There exist in the total economy at all times significant cross- and
counter-currents. In any period of general stability, it would appear
that these currents are of equal strength and their forces are canceled
out; that is, the effects of the reactions of firms with poor profits upon
the materials, capital, and labor markets tend to be wholly offset by
the expanding activities of those firms with attractive returns. When,
however, the currents run more vigorously in one direction than another, the result must be a shift in the level of aggregative activity.
When the number of firms experiencing shrinking demand, and hence
profits, is very numerous, and if their size in terms of employment
and material purchases is substantial, the forces which produce an
aggregative contraction may overcome the opposing forces, thereby
transferring the contraction movement to the whole economy.
Kelatively little is known of the characteristics of business firms
grouped, as suggested, according to their profit trends. The inadequacy of available data and the complexity of the problem make
analysis difficult. Nevertheless, further exploration in this area offers
substantial possibilities of contributing critical information to our
knowledge of the nature of business fluctuations.
CONCLUSION

This brief paper has attempted to address itself to the question:
Is there in the history of the price structure of the half century, 18901940, any evidence that suggests that forces within the price system
were responsible for initiating the excessive cyclical fluctuations to
which the economy has been subjected ?
Tentatively the answer appears to be "no". There has been a substantial increase in our knowledge of the price structure but no pattern
of relationship has emerged, either in the data or in abstract generalization, which clearly suggests the presence within the system of prices
of a cycle-producing force of an initiating nature and adaptable to
controls.
It is sometimes held that the business cycle is peculiar to our type
of society and, if not a direct product of the increasingly complex
nature of our economic order, then closely associated with that fact.
To the extent to which such a hypothesis may be true it becomes important to develop a price structure of sufficient sensitivity so as to
maintain those relationships which will assure a level of activity close
to full employment.
The problems of the flexibility of price responses offer some promise.
Even in that area, however, nonprice determinants seem too important
to suggest that anything more than minor improvements in the price
mechanism might be anticipated. By and large the search for controls
over excessive economic fluctuations would seem to lie in other directions.

142

ECONOMIC STABILITY AND GROWTH
Business activity and prices, 1890-1940

Cycle

PJuly..
T May

P January
T June
P December _
T June.
P June
T December.
P September.
T August
P May..
T June
P January
T January
P January _ _
T December .

P August

T April
P January
TJuly

Percent
gainful
Year workers
unemployed

1890
1891
1892
1893
1894
1895
1896
1897
1808
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918

1919
1920
1921
1922
1923
P May
T July.
1924
1925
P October... 1926
T November. 1927
1928
1929
P June
1930
1931
1932
1933
T March
1934
1935
1936
1937
P May
1938
TJune .
1939
1940

5.67
5.74
1.62
4.78
4.38
1.85

2.12
6.27
1.92
1.45
4.06
2.35
2.58
5.56
5.87
.46

1.33
11.2
6.79
1.71
4.56
1.81
1.01
3.45
3.88
.89

5.91
14.19
22.71
23.36
19.04
17. 56
15.00
12.15
18.40
16.32
13.08

Average
Wholesale prices, 1926=100
Physfullical
time
output, weekly
manu- earn- Interest
Fuel Metals
Chemfactur- ings,
All
and
Build- icals
rates,
Farm and
ing
ing,
manu- prime comprod- lighting metal
and
1899= factur- com- modiucts mate- prod- mate- allied
100
ing,
mercial ities
rials
rials
products
1890- paper
ucts
99=100

100
102
115
129
132
124
148
159
161
133
158
168
161
185
198
186
218
259
257
254

222
242
194
249
280
266
298
316
317
332
364
311
262
197
228
252
301
3C3
376
295

374

101.0
100.8
101.3
101.2
97.7
98.4
99.5
99.2
99.9
101.2
104.1
105.9
109.2
112.3
112.2
114.0
118.5
122.4

Index
of payrolls.
1939=
mo
1UU

103.2
123.5
79.7
85.5
108.4
101 2
106.6
109.9
107.9
109.1
116.4
94.1
71.2
49.2
.52.8
67.8
78.0
90.5
108.2
84.2
100.0
114.5

6.91
6.48
5.40
7.64
5.22
5.80
7.02
4.72
5.34
5.50
5.71
5.40
5.81
6.16
5.14
5.18
6.25
6.66
5.00
4.67
5.72
4.75
5.41
6.20
5 47
4.01
3.84
5.07
6.02

5.37
7.50
6.62
4.52
5.07
3.98
4.02
4.34
4.11
4.85
5.85
3.59
2.64
2.73
1.73
1.02
.76
.75

3.94
.81

..59
.56

56.2
55.8
52.2
53.4
479

50.4
54.2
49.5
51.3
416

48.8
46.5
46.6
48.5
52.2
56.1
55.3
58.9
59.6
59.7
60.1
61.8
65.2
62.9
67.6
70.4
64.9
69.1
69.8
68.1
69.5
85.5
117.5
131.3

43.9
39.6
42.5
44.9
45 8
50.5
52.8
58.4
55.6
58.5
56.4
57.3
62.2
62.2
69.6
74.3
66.8
72.6
71.5
71.2
71.5
84.4
129^0
148.0

13S. 6
154.4
97.6
96.7
100.6
98 1
103 5
100.0
95.4
96.7
95.3
86.4
73.0
64.8
65.9
74.9
80.0
80.8
86.3
78 6
77.1
78.6

157.6
150. 7
88.4
93.8
98.6
100 0
109.8
100.0
99.5
105.9
104.9
88.3
74.8
48.2
51.4
65.3
78.8
80.9
86.4
68.5
6\3

67.7

38.1
37.0
34.8
35.3
34.3
40.3
39.5
33.9
34.5
41.2
46.3
44.6
51.8
60.3
53.3
49.6
52.0
54.4
53.7
51.6
47.6
46.7
51.4
61.3
56.6
51.8
74.3
105.4
109.2

105.3
92.2
84.0
76.8
65.7
70.4
71.2
65.0
65.3
100.0
98.0
93.1
91.0
90.2
79.9
89.1
102.4
109.8
86.3
84.5
85.2
80.8
89.5
90.8
80.2
86.3
116.5
150.6
136.5

46.5
44.2
41.7
41.6
39.8
38 8
38.9
37.4
39.6
43 6
46.2
44.3
45.3
46.7
45.0
48.1
54.0
56.8
52.0
53.7
55.3
55.3
55.9
56.7
52.7
53.5
67.6
88.2
98.6

73.2
74.0
74.6
72.7
65.5
64.7
65.0
70.9
77.4
81.1
82.1
84.2
86.5
84.1
84.1
82.3
76. S
78.5
79.6
79.9
82.0
81.6
80.7
80.2
81.4
112.0
160.7
165.0
182.3

104.3
163.7
96.8
107.3
97.3
92 0
96.5
100.0
88.3
84.3
83.0
78.5
67.5
70.3
66.3
73.3
37.5
76.2
77.6
76.5
73.1
71.7

130.9
149.4
117.5
102.9
109.3
105 3
103.2
100.0
96.3
97.0
100.5
92.1
84.5
80.2
79.8
86.9
86.4
87.0
95.7
9\ 7
94.4
95.8

115.6
150.1
97.4
97.3
108.7
102.3
101.7
100.0
94.7
94.1
95.4
89.9
79.2
71.4
77.0
86.2
85.3
86.7
95.2
90.3
90.5
94.8

157.0
164.7
115.0
100.3
101.1
98.9
101.8
100.0
96.1
95.0
94.0
88.7
79.3
73.9
72.1
75.3
79.0
78.7
82.6
77.0
76.0
77.0

Sources: Cycle identifications: Arthur F. Burns and Wesley C. Mitchell, Measuring Business Cycles.
National Bureau of Economic Research. N. Y. 1947, p. 78.
J
Percent gainf'il workers employed: Comnuted from data in Historical Statistics of the United States,
1789-1945, Government Printing Office, Washington, 1949, p. 65. 0 indicates that estimated employment
equaled or in some years (1903, 1917, 1918, 1919) exceeded the estimated total gainful workers.
Physical oitDit: Manufacturing, National Bureau of Economic Research in Historical Statistics, p. 179.
Weekly earning: Ibid., n. 66.
Interest rate: Ibid., p. 278.
Wholesale nrices: Hdnd'iook of Labor Statistics, Bureau of Labor Statistics, Government Printing
Office, Washington, 1950, p. 118.

AN INTERPRETATION OF PRICE MOVEMENTS SINCE
THE END OF WORLD WAR II
Bert G. Hickman, The Brookings Institution 1
This paper attempts to identify and analyze the principal factors
which have shaped the course of prices during the (past 12 years. The
emphasis is upon the interpretation of price movements rather than
the influence of price relationships on income distribution, resource
allocation, or, for that matter, the level of aggregate business activity,
although the grosser sort of expectational effects are considered in the
latter connection. For the most part, then, the economic consequences
of changes in relative prices are neglected except insofar as they are
relevant to the behavior of the overall average of prices. I will not be
interested here in questions of equity or responsibility or whether
unregulated economic forces should be permitted to establish prices
within the going institutional framework, but only in the analysis of
why prices behaved as they did. All these other topics are bypassed
not because they are unimportant, but because a documented interpretation of price tendencies during a span of more than a decade is a
sufficiently lengthy undertaking. A good part of the justification for
the undertaking, on the other hand, lies in whatever contribution it
may make toward better understanding of the broad issues of cause
and effect, equity, and responsibility.
The paper is organized chronologically, since the principal task is to
interpret price behavior in terms of its several causes, and since this
job of synthesis is best handled by introducing particular factors
as they become important over time. It will not, however, be necessary to give equal weight to the occurrences of every year—we will be
selective on that score—and there is at least one topic of an analytical
nature which will interest us continuously: the interaction of prices
and wages. A major advantage of the predominantly synthetic treatment of the paper is that it will enable us to see how wage-price interactions differ according to the surrounding economic circumstances.
In this way needed perspective will be provided on a topic which is
often subject to oversimplification and misplaced emphasis in the
form of either-or propositions.
A PRELIMINARY LOOK AT PRICES I N THE POSTWAR PERIOD

Let us take a moment to decide which aspects of postwar price
experience we should most like to understand. A glance at chart 1
reveals a number of intriguing features which merit attention. Foremost is the fact that prices have moved substantially higher over the
12 years spanned by the chart. Thus the Consumer Price Index rose
56 percent between January 1946 and December 1957. The pace of
advance was markedly uneven, however, and this is the second major
fact that we notice. It suggests at the outset that it would be unwise to
ignore developments during the period in an explanation of the
increase over the period, and that a search for a single cause of the
overall advance would be unrewarding.
1
The views expressed in this paper are those of the author. They do not necessarily
reflect the views of other memhers of the Brookings staff or of the administrative officers
of the institution. My thanks are due to William R. Belmont, who drew the charts and
assisted in the preparation of the paper in countless other ways.

143

144

ECONOMIC STABILITY AND GROWTH
Chart 1
Indicators of Prices, Production, and Employment
(Monthly, '1946 -1957)

€5

150

140

130

1946

Sources" Bureau of tabor Statistics, Board of Governors of the Federal Reserve System.
Vertical lines indicate business cycle peaks(P) and troughs(T) as dated by the National Bureau of
Economic Research, except for the peakof August 1957 which was dated by the author.

ECONOMIC STABILITY AND GROWTH

145

These initial reflections are strengthened when we observe that the
unevenness is of several kinds. First, there are variations associated
with the ebb and flow of physical activity during business cycles.
Prices increased during the 3 business expansions of the postwar period, but stood firm or declined during the 2 complete contractions
included in the chart. There is a hint here of the operation of systematic cyclical factors affecting prices, production, and employment.
A goodly amount of irregularity is to be found even during like
cyclical phases, however. Thus each successive business expansion
saw a smaller rise of average consumer prices, with increases respectively of 33, 13, and 5 percent. The first expansion was clearly inflationary, whether one means by "inflation" a sustained and rapid
advance of prices, or a rise which is both substantial, and substantially
larger than the concomitant increase of production. The second
upswing contains a subperiod which is also definitely inflationary;
namely, the 8 or 9 months following June 1950 during which consumer prices increased nearly 1 percent per month. No comparable
advance in the cost of living occurred during the most recent expansion. The index began to rise midway in the upswing and increased
thereafter at the not inconsiderable rate of 0.3 percent per month,
however, and it is both interesting and relevant to note that prices
rose, faster than production and employment at that time.
Apart from these differences in the speed and magnitude of price
increases, there are striking distinctions with regard to the internal
patterns of prices and production in the several expansions. Both
prices and output rose fairly steadily during 1946-48, and when the
one quickened or slowed, so also did the other, and roughly in proportion. A similar correlation is to be observed during the early inflationary phase of the second expansion, but at that point the resemblance ends. Wholesale prices actually fell during the remainder
of the expansion, and consumer prices first rose more slowly and then
leveled off. Production and employment paralleled these price tendencies from mid-1951 to mid-1952, but rose strongly in the face of
stable prices from then until mid-1953. The pattern of the third
expansion differed in still another way, with production and employment rising rapidly during the first 18 months and prices changing
scarcely at all, only to be followed by an equally lengthy period in
which these tendencies were reversed. Again, these variations suggest that different sets of forces, or different combinations of them,
shaped the course of prices during one or another phase of the postwar
experience.
The indexes of chart 1 serve admirably to map the time path of the
overall average of prices. They are averages, however, and they
sometimes conceal wide divergencies; after all, there would be little
need to describe a central tendency by an average were it not for
dispersion of the individual observations. Many hundreds of prices
enter these indexes, and for present purposes it would be both impractical and uniformative to study them all. We will find it useful
to give separate attention to a few major subgroups, however, for the
light that they cast on the determinants of the overall price level and
for what they reveal about the operation of specific forces. These
subgroups are plotted in charts 2 and 3. Notice that the first of these
charts begins in 1947 instead of 1946.

146

ECONOMIC STABILITY AND GROWTH
Chart 2
Indexes of Wholesale Prices
(Monthly, 1947-1957)

100 -,

145

135

125

Finished Durable
Goods
115

Other Nondurable

Goods
105

•"*" •

yiy

a a*
•

a*"' !
'-•••.•.•••.

.••.•"••.••"
••...••

*•" 'Foods
85

•

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the National
Bureau of Economic Research, except for the peak of August 1957 which was dated by
the author.
Source: Bureau of Labor Statistics.

147

ECONOMIC STABILITY A3STD GROWTH
Chart 3
Indexes of Consumer Prices
(Quarterly, 1946-1957)
Index
1947-49-100
145

Services

135

^ r

125

115
Food
105
Durable Goods
^fr

Other Nondurable
Goods

. /

75
/
t

0
1946

l i i

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the National Bureau
x>f Economic Research, except for the peak of August 1957 which was dated by the author.
Source: Bureau of Labor Statistics.

The dominant feature of chart 2, of course, is the gap that opened
between the indexes beginning in 1951. I t will be seen from the upper
panel that prices of farm products and processed foods declined irregularly from 1951 through 1955, all the while that "industrial"
prices were stable or rising. Since agricultural products are the basic
raw materials for many manufactured nondurable goods, it is reasonable to expect the weakness of the former to be reflected in the prices
of the latter, though not necessarily to the same degree because of
changes in other components of cost or by reason of differential movements of demand at the various market levels. The expectation is
confirmed by the breakdown of wholesale prices of finished goods
shown in the lower panel.
Like all changes in relative prices, these disparate movements had
their effect on the distribution of national income. The question
which presently concerns us is whether they also influenced the price
level. It is surely pertinent to observe that the gap opened and wid23734—58

148

ECONOMIC STABILITY AND GROWTH

ened most noticeably during the only extended interval of overall
price stability in the postwar period—that is, from about mid-1951
to late 1955 or early 1956. The observation alone is not enough to
establish causation, of course. An unsupported claim that prices
were stable on the average because agricultural prices fell could with
equal arithmetic justice be countered with the assertion that the rise
of nonagricultural prices caused the average to remain stable. The
error in this kind of statement is obvious, but it is perhaps less apparent that it is also insufficient to show that the movement of, say,
agricultural prices is due to factors peculiar to them alone; for this
leaves open the possibility that the height of the price level is determined by general causes and that the change in agricultural prices
simply induced a counterbalancing adjustment in other prices without
affecting the overall average. These matters must be carefully considered at the appropriate points.
The classification of prices by durability of product is of interest
in yet another connection. It will be seen that the curve of durable
goods prices more nearly resembles an ascending staircase than an
inclined plane. The steps do not appear in every year, but when
they do it is usually after midyear. Traces 6f this tendency for price
increases to concentrate in certain months can also be distinguished
in the index of prices of nondurable goods excluding foods, but it
occurs less regularly there, is centered on the months before and after
the turn of the year, and is less pronounced, except possibly in 1955
and 1956. Food prices do not exhibit the tendency, although they
seem to be subject to seasonal peaks of their own in the late summer
and autumn. An investigation of the causes of the step pattern and
its comparative strength among classes of products promises to reveal something about the nature and timing of inflationary processes.
The wholesale prices included in chart 2 refer only to goods. A
breakdown of the Consumer Price Index includes prices of services as
well, and shows that they are much less subject to short-term fluctuations than prices of goods (chart 3). Services reveal in purest form
the underlying inflationary cast of the postwar period. Since they
are produced as used and cannot be stockpiled, they are unaffected
by temporary shifts of inventory demand by households or businesses.
At the same time, their persistent upward course is symptomatic of
constant pressure from increasing demand, increasing cost, or both.
While it is true that their rise relative to prices of goods during the
past 7 years must reflect a tendency toward lagging productivity or
growing relative demand, their absolute increase need not have occurred if gains in real national income due to advancing productivity
had been distributed in the form of general price reductions at a constant level of money income.
BASIC CONCEPTS

The theoretical apparatus which underlies the subsequent discussion rests on one or two key concepts which can readily be illustrated
by examples.
First, there is the notion of aggregate supply. If we think of an
individual business firm, it is apparent that its costs of production can
vary in several ways. If it has unutilized plant capacity, it may be able
to increase production considerably without forcing costs up, but as

ECONOMIC STABILITY AND GROWTH

149

it approaches capacity operations, unit costs are bound to rise even
when there is no increase in the prices which it pays for labor, materials, or other resources. This is because there is an optimal balance
between materials, manpower and equipment, and when that balance
is exceeded, the output per additional worker or batch of materials
must decrease—it will do little good to put two workers on a machine
which can be operated efficiently by one. Costs may also be forced up
at higher rates of utilization because prices of resources rise—premium pay for overtime is an important example—or because standby
units of inefficient equipment may be pressed into service or it may
become necessary to hire persons who are poorly trained or otherwise
inherently less efficient than the normal complement of workers.
Analogous considerations apply when we think of the economy as
a w^hole. When coming out of a period of recession, most firms will
be operating well within capacity, and increases of demand will go
primarily to stimulate production rather than prices. As national
output continues to expand, however, bottlenecks will increase prices
of some resources, capacity will be reached in some businesses, overtime pay will increase, and so forth. If demand continues to rise, firms
will produce more, but they will also charge higher prices to cover the
resultant increase of cost. Now, in actual practice it is always possible to get a little more output with a sufficiently lavish expenditure
of resources, but it is a useful simplification to think of national output
as having a fixed limit at any given time, at a point corresponding to
full man-hour employment of labor. Once the limit is reached, further
increases of money demand can only raise prices without inducing
more output. That is about what happens during periods of pure
inflation.
It must now be emphasized that two important cost influences were
left out of account in the preceding paragraphs. This was deliberately done, for they are quite different from those discussed thus far.
The variations in costs already considered were the result of changes
in output which were induced by changes in demand—costs and prices
would not have risen had demand not increased. The type of cost
change to be considered now can occur at any given output and independently of any change of demand.
Physical limitations on output can be gradually overcome, given
sufficient time, by construction of additional productive facilities and
the growth of the labor force. Technological improvements, moreover, may make possible new methods of production which lower costs
no matter what the rate of capacity utilization. In these cases, the
same total output as before can be produced at lower prices, or alternatively, a larger output will be forthcoming at the same price level.
The second main type of independent cost change is an "autonomous" variation in the price paid for a resource. Wage increases
negotiated in collective bargaining may usually be regarded as autonomous, since they are not necessarily induced by an expansion of
output and will raise money costs of producing each alternative output. Independent changes may commonly occur also in the prices
of raw materials imported from abroad, traded in world markets, or
subject to supply interruptions due to the vagaries of weather. Provided demand remains unchanged, autonomous increases in resource
prices will raise the prices of goods and reduce their output, whereas
decreases will have opposite effects.

150

ECONOMIC STABILITY AND GROWTH

Demand will not remain unchanged in the ordinary course of
events, however, and this brings us to the next major point—that on
an aggregative or economy wide basis supply and demand are not independent, but interact with one another in several ways. Thus in the
present instance, an autonomous wage boost may increase aggregate
demand in addition to raising costs. This is because wages affect incomes as well as costs, and income is one determinant of demand.
Suppose that a general wage increase goes into effect at a time of full
employment and when businessmen have no independent reason to be
pessimistic about the near-term future of sales. Under those circumstances, businessmen are likely to maintain output, increase prices, and
wait to see whether the increase diminishes the physical volume of
sales.
With higher wages and stable employment, consumer demand will
be augmented, though not enough to match the price increase. The
increase of wage-earner consumption demand will be smaller than the
price increase partly because prices will usually be raised enough not
only to cover the wage increase but also to increase profits, and partly
because workers will save some fraction of their added incomes. If
no increase of demand is forthcoming from other sources, the same
physical volume of sales as before cannot be maintained at the new
price level, and either prices or production or both must be cut.
It should not be assumed, however, that other demands will remain
unchanged. Large components of final demand are partly or entirely
independent of the current level of income. Under the full employment conditions assumed in our illustration, it is probable that businessmen and governmental units will pay the higher prices now necessary to carry out their previously planned physical investments.
There is likely also to be sufficient flexibility in the financial arrangements of nonwage personal income receivers to permit them to increase
their consumption outlays and defend their living standards even if
their incomes lag somewhat in time or amount. Thus aggregate demand may rise enough to sustain the price increase without any reduction of output.
This last result is quite possible and even probable under the
assumed conditions, but it is by no means inevitable. To cite a single
important exception, financial constraints might prevent the necessary
expansion of money outlays for plant and equipment, business inventories, or durable consumer goods. Thus "it all depends"—but that is
precisely the point.
It will be helpful to distinguish two other types of supply-demand
interaction in these introductory remarks. The first occurs when
the initial change on one side is conditioned by the actual or expected
tstate of the other. Thus unions will ask for larger wage increases
If they expect demand to be high or rising, and management will resist a given increase less strenuously if they hold the same expectation.
Secondly, there is the interaction occasioned by the fact that an initial
adjustment of price or output in response to a primary disturbance
may induce secondary repercussions, as when a price rise caused by
increased demand induces a wage increase which puts further pressure
on prices.
In summary, the following points should be kept in mind as we
interpret postwar price experience. What happens to prices and production depends on both supply and demand. Under given condi-

ECONOMIC STABILITY A3STD GROWTH

151

tions of supply, prices and output will rise and fall together along
with demand, and their relative movements will be correlated with
the level of economywide resource utilization. Under given conditions of demand, autonomous cost increases will raise prices and reduce
output, and vice versa for autonomous decreases. An initial change
on either side is unlikely to leave the other unaffected, however, and
this means that one must be alert to several forms of interaction between the two. In view of these complexities, inferences about the
causes of observed behavior should be drawn with care and assertions
that single factors are responsible received with skepticism.
POSTWAR INFLATION", 194 6 - 4 8

The economy was still operating under the wartime "disequilibrium
system" as 1946 opened. The essential ingredients of the system were
simple. Resources were diverted to war use only partly by taxation
and the remainder of the transfer was deficit-financed. Since taxes
were not used to reduce private incomes proportionately to the supply
of civilian goods, it was necessary to control prices and allocate quantities by rationing. Since money incomes far exceeded permissible expenditures at controlled prices, and did so for several years, households and businesses perforce accumulated large holdings of money,
Government bonds, and other liquid assets.
It was called a disequilibrium system, of course, because the demand latent in the prevailing level of money income and wealth was
not allowed to determine prices. So long as prices were controlled,
the economy was prevented from seeking a market equilibrium. The
removal of controls during the last half of 1946 freed the economy to
react to the accumulated demand pressures which had been contained
previously by administrative devices. The result was a powerful
inflationary shock (chart 1).
Effective demand was extraordinarily high at the time of decontrol
partly because of the wartime heritage of financial liquidity and deferred demands which characterized alike the household, business, and
foreign sectors. It was augmented also by the "first round" of wage
increases of early 1946. Higher wage incomes increased the postdecontrol money demand for consumer goods, and since it is unlikely
that other income receivers diminished their consumption or investment demands correspondingly, if at all, this meant that prices rose
more than they would have at the old wage rates.
Thus, although the first round of wage increases was by no means
the sole cause of the price inflation of 1946-47, it did contribute to itTo argue that it was the sole cause would be to assert that autonomous
wage increases disturbed an initial equilibrium in which there was no
excess demand, which is absurd. Similarly, to cite the lag of prices
behind wages as evidence of a cost-price spiral is to ignore the fact
that the lag was really determined by the timing of price decontrol
relative to wage decontrol.
The pace of the inflation slowed noticeably in the second quarter
of 1947 (chart 1). The index of wholesale prices actually declined
somewhat and the average of consumer prices was stable. A corresponding deceleration occurred in physical activity. The period of
tranquility was short-lived, however, and with the coming of summer,
production recovered moderately and prices again shot upward. If

152

ECONOMIC STABILITY AND GROWTH

the respite is regarded as a temporary equilibrium, two questions
come naturally to mind: What conditions produced the equilibrium ?
Why was it unstable ? These questions will be answered in turn.
The first possibility to consider is that financial stringency may have
retarded the expansion of demand in one or more parts of the economy.
The essence of the sort of pure inflation now under discussion—that
is, a situation where real output is virtually fixed and additional
spending goes primarly to raise prices—is, of course, that it results
from competition among the various claimants on the national product
for the limited supply of goods and services. Gradual increases of
supply aside, the inflation will continue as long as some spending units
are displeased with their level of real expenditures—their share of
total output—and are able to increase money expenditure in an attempt to bid goods and services away from rival spending units.
Conversely, it will be halted if all spending units become satisfied
(perhaps better, reconciled) with their current real expenditure in the
sense that they cannot, or will not, find the means to finance a larger
money outlay.
Any spending unit, be it in the household, business, Government, or
foreign sector, has two sources from which to finance current purchases of goods and services. It can undertake expenditure out of
current income receipts; or it can alter its wealth, either by borrowing
or by drawing upon its assets. Thus financial stringency could take
the form either of an income limitation, or a constraint on the desire
or ability to diminish net worth.
Let us examine the incomes and expenditures of the major sectors
for evidence of monetary constraints on spending (chart 4). The
reader will note the continuous rise of personal consumption expenditure from 1945 to mid-1947 and beyond. He will notice also that
throughout 1946 and early 1947 expenditure climbed more rapidly
than income, a symptom of the excess demand for consumer goods
which was released by decontrol and of the determination of consumers to maintain their real purchases after decontrol. Even the
diminished increase of disposable income in the first quarter of 1947
and its actual dip in the second quarter failed to deter the rapid expansion of expenditure.
The fall in the rate of personal saving during the first postwar wave
of inflation was accompanied by some deterioration in the financial
position of consumers as a group, of course. The moderate reduction
of consumer liquidity during 1946 may have inhibited the growth of
spending in theiirst half of 1947, just as the retardation of disposable
income may have done, but the fact remains that neither force was
powerful enough to prevent further substantial increases of consumer
expenditure in those months. We must look elsewhere for signs of
flagging demand.

ECONOMIC STABILITY AND GROWTH

153

Nor do we fare much better when we turn to the Government and
foreign sectors. Government purchases of goods and services did diminish during the first half of 1947, but relatively little in comparison with the increases in other categories of final expenditure. The
reduction came entirely in the Federal sector and reflected primarily
the further progress of war demobilization rather than any financial
constraint. Net tax receipts (receipts less transfer payments) were
substantially in excess of Government purchases during the last half
of 1946 and in 1947. Incidentally, the Government surplus served
as a partial restraint on the progress of the inflation, but it was not,
of course, enough. The sharp rise of receipts during 1946 was largely
induced by the expansion of money incomes and was therefore, itself,
a reflection of the inflation which automatic increases of tax receipts
could mitigate but not prevent.
No assistance was forthcoming from foreign sources. On the
contrary, net foreign investment rose substantially during 1947 Under
the spur of postwar needs for relief and rehabilitation in war-ravaged
countries.
We come finally to domestic investment and its finance. It is apparent from chart 4 that gross private domestic investment fell in the
early months of 1947 and remained relatively depressed until late in the
year. Since gross retained earnings continued to rise, private deficit
financing for purposes of investment was reduced. This last development was effect rather than cause, however, for there was no shortage
of external funds to restrict investment during this period.
A breakdown of gross private domestic investment reveals that its
decline after the fourth quarter of 1946 was due entirely to the inventory component (chart 5). Business fixed investment and residential construction continued to rise along with consumption and net
foreign investment. Taken altogether, these facts demonstrate that
the lull during the spring primarily reflected diminished demands for
inventory and not a deficiency of final demands. There is strong
evidence, moreover, that the decline of inventory demand was caused
by nonmonetary factors.

154

ECONOMIC STABILITY AND GROWTH
Chart 4
The Nation's Income, Expenditure, and Saving By Major Economic Sectors
Seasonally Adjusted Quarterly Totals at Annual Rates, 1945-1950
(Billions of Dollars)

Consumers

190

260

130

Personal Consumption Expenditures

Business

105

Excess of Investment Jf
Gross Private Domestic Investment

^^^•V^

Excess of

TT1T
U l l \JjJ

ore

1 Gross Retained Earnings 1/ I

^l^^\ 1
\ 4 \ \ III

Government

105

Purchases of Goods and Services
75

Receipts
(Less Transfer Payments)

25
0

Net Foreign Investment

25
1

1 IIM"*1

^—i—|—

T 1 1 1 1ITI

15
1947

1948

1949

1950

1/ Includes undistributed corporate profits, inventory valuation adjustment, and capital
consumption allowances.
Source: Department of Commerce.

ECONOMIC STABILITY AND

155

GROWTH

Chart 5
Gross Private Domestic Investment and Its Major Components
Seasonally Adjusted Quarterly Totals at Annual Rates, 1946-1950
(Billions of Dollars)
P
T

Gross Private Domestic Investment

30 -

20 -

10 -

-10

Components of Business Fixed Investment
P

40 -

30 -

Producers* Duralble Equipment
—«

^
0

Nonresidential
i

1946

Construction
i

1947

1948

1949

1950

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
national Bureau of Economic Research.
Source: Department of Commerce.

156

ECONOMIC STABILITY AND GROWTH

The reason for believing that monetary factors had little to do with
the behavior of either fixed or inventory investment during 1946-47 is,
of course, that the entire economy was quite liquid and bank credit was
readily available at generous terms. Since the monetary authorities
accepted the goal of stable interest rates, they were unable to act to
prevent an expansion of bank credit and the money supply. Any increase in the demand for loanable funds which could not be satisfied from current saving or by dishoarding must needs be met by the
banking system if interest rates were to be kept from rising.
As it happened, the contribution of additional money supplies to
increased spending was comparatively small during the period under
review. Thus gross national expenditure increased 12 percent and
the private portion of it 15 percent between the second quarters of
1946 and 1947. The money supply, consisting of currency outside
banks and adjusted demand deposits, rose less than 3 percent over the
same interval. A rise in the income velocity of circulation of money
therefore accounts for three-fourths of the increased spending on
national output and four-fifths of the rise of private spending. This
implies a sharp reduction in the proportion of the money supply
held idle. The dishoarding occurred, moreover, at low and stable rates
of interest. Here is ample testimony of the involuntary nature of
much of the liquid asset accumulation of the war years, which resulted
by wars' end in holdings which were considerably larger than desired, given current prices for goods and going interest rates.
What, then, does account for the reduced inventory demand of early
1947? It was primarily a reflection of pessimistic short-term expectations. As the President's Economic Report of January 8, 1947,
put it (p. 16) : "Threatening the continuation and expansion of business investment is the fear that a drop in general consumer demand may
be in the offing." This fear rested upon the knowledge that inflation
had diminished the real value of disposable personal income during
1946 and that consumers had maintained real expenditure only by
increasing money outlay sharply relative to income. This process obviously could not go on indefinitely and might be near its limit, especiallv since real demands for specific commodities could be expected to
fall as the goods became available to rebuild household stocks.
The close parallels between the events of 1919 and 1946 had not
escaped attention. Each were years of postwar transition during
which abnormal domestic and foreign demands had driven prices
rap^dlv upward. The deflation which had followed the earlier experience during 1920 was among the sharpest on record. History
might repeat itself, and in the circumstances it was wise to pursue
conservative inventory policies.
Chart 6 indicates how conservative those policies were. The ratio
of inventories to sales in manufacturing: was actually reduced during
1946. Retail inventories, in contrast, did rise strongly relative to sales
between April and December. It is significant, however, that the ratio
remained far below the prewar relationship of 1939-41. This last was
true also of the wholesale and manufacturing sectors. Businessmen
generally were making do with relatively smaller stocks than in prewar days.

157

ECONOMIC STABILITY AND GROWTH
Chart 6
Inventory Sales Ratios
(Monthly, 1945-1950)

1.60

1939-41 Average

1.30
\*^/~""^

RetaU

1.00
)

1939-41 Average

1.30

1.00
Wholesale
0.70
t

2.20
1939-41 Average
1.90

1.60
1

Manufacturing

1.30

1945

1946

1947

1948

1949

1950

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the National Bureau
of Economic Research.
Source: Department of Commerce.

Diminished inventory demands were felt largely by producers of
nondurable goods. Some decline of postwar inventory investment was
inevitable once stocks were rebuilt sufficiently to provide appropriate
selections and acceptable delivery schedules. To judge from the
levels maintained in later years, the appropriate "technical" ratio of
stocks to sales was approximated by nondurable goods retailers late in
1946. Technical considerations would have been overridden, nonetheless, had retailers expected prices to advance rapidly or sales to spurt
strongly. For reasons already discussed they expected neither to
occur, or at least they were uncertain enough to play it safe.

158

ECONOMIC STABILITY AND GROWTH

Prices of durable goods were not exposed to the same degree of
downward pressure from diminished inventory demand as nondurables. Average wholesale prices of finished durable goods rose more
during the second quarter of 1947 than in the first, although prices of
intermediate durable materials advanced more slowly than before.
The latter retardation may reflect some easing in the intensity of demand, since unfilled orders of durable goods manufacturers leveled off
in the first quarter and dropped in the second. Perhaps the most interesting development in this sector, however, was the stability of
steel prices during the first 6 months of the year.
What makes it particularly interesting is that steel prices were not
raised at the time of the April increase in wages of steelworkers, and
did not rise until 3 months had passed. On no other occasion from
then until now have steel prices lagged a general increase of steel
wages. Because of their key position in the industrial price structure,
the stability of steel prices probably contributed importantly to the
shift toward expectations of stable or falling prices which characterized the period. It appears likely that steel and other administered
prices were, in turn, strongly influenced by the public attention which
was focused on the desirability of voluntary restraint in setting
prices in the State of the Union Message and Economic Report of
January 1947, and again in speeches made by the President in April.
Be that as it may, price stability was short lived in these cases and
most others, for it could not persist in the face of the resurgence of
demand which soon developed. The next task is to identify the factors back of that resurgence.
Reduced inventory demand consequent upon pessimistic expectations about the short-term course of sales and prices was primarily
responsible for the precarious balance of aggregate supply and demand in the spring of 1947. The balance was upset and inflationary
pressuries revived during the summer partly because a recasting of
expectations took place. The reversal of expectations did not, however, occur for independent psychological reasons. Quite the contrary, it was due to a sequence of concrete economic events which first
removed the drag exerted by uncertainty about short-term developments and a bit later contributed directly to an acceleration of final
expenditure.

ECONOMIC STABILITY AND GROWTH

159

First, there was the fact that the pessimism about final demand
proved to be unjustified. Consumer expenditures, business investment
m plant and equipment, residential construction, purchases of State
and local governments, and net foreign investment all increased during the winter and spring. Fears that new homes would be priced out
of a market by rising building costs were lessened when new housing
starts rose more than seasonally during the spring. It had been anticipated that the drain on gold and dollar assets of foreign countries
would soon force a substantial reduction in net exports, but the prospects in this area were improved by the announcement of Secretary
of State Marshall early in June that the United States would support
a joint program for European recovery.
The acceleration of prices did not await actual increases of final
demand. Specific factors boosted prices in two prominent sectors
during the early summer. Food prices rose during July and August
owing to a combination of short domestic and foreign crops and
normal seasonal influences. A widely publicized wage increase in coal
mining was accompanied by a simultaneous advance in coal prices in
July and was followed by a steel price increase in August. These specific increases doubtless fostered expectations of a new wave of generalized price advances, the more so because it now appeared that
earlier wage gains would be reflected in prices, as indeed they were.
Thus, speculative purchases were partly responsible for the steep
rise of manufacturers' new orders and sales which commenced in
September (chart 7).
Retail sales also spurted strongly in September, however, providing
still another stimulus to orders and prices. Legislation passed late in
July permitted redemption of Armed Forces leave bonds on or after
September 2. Quick advantage was taken of the opportunity to
supplement current incomes, as transfer payments leaped more than
$10 billion at an annual rate in September and produced corresponding bulges in personal income and retail sales which persisted for
several months. The spurt of final demand was augmented by substantial inventory accumulation at the retail level, and although the
latter was partly offset by inventory reductions of wholesalers, factory sales and production nonetheless moved strongly upward. As
production mounted, so also did earned incomes, further feeding the
expansion of retail sales.

160

ECONOMIC STABILITY AND GROWTH
Chart 7
Indexes of Personal Income, Retail, Wholesale, and Manufacturers' Sales
and Inventories and Manufacturers' New Orders
Seasonally Adjusted, Monthly, 1946-1950
Index
1947-49 = 100
1947-49=100
135

125

105

•
160

150

130

120

110

130
120

100
Manufacturers* Inventories

1946

1948

1949

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
National Bureau of Economic Research.
1/ Not seasonally adjusted.
Source: Department of Commerce.

ECONOMIC STABILITY AND GROWTH

161

The expansion was given new vigor by the events just reviewed.
The underlying sources of strength which stemmed from backlogs of
consumer and investment demand and which had been present all
along now asserted themselves for a timx3 without the debilitating
offset of inventory disinvestment. When these forces began to weaken,
moreover, fresh stimuli came forward to prolong the movement.
The year 1948 had scarcely begun before the economy experienced
a jolt which raised anew the possibility of imminent deflation. Farm
and food prices broke sharply downward in February (chart 2).
The reductions were apparently the result of favorable crop prospects
at home and abroad, prospects which were disproportionately influential because of expectations that the extraordinary height to which
agricultural prices had risen by the end of 1947 could not be sustained
under normal conditions. This episode could have touched off a wave
of deflationary inventory disinvestment had it created expectations of
a generalized price decline, but it did not do so, and industrial prices
remained largely unaffected. Expectations remained unshaken partly
because the potential decline of agricultural prices was limited by
support programs and partly because important shortages of durable
goods and materials persisted and the response of prices in those sectors to augmented supplies or diminished demands could be expected
to take place slowly in any event.
Although final demand was sufficiently high to support prices and
outputs of most goods in the face of the agricultural price declines
of February and March, and indeed for some months thereafter, the
first quarter nevertheless marks the emergence of important deflationary factors. These factors were of more than transitory importance
because they reflected a change in the basic conditions which had
fostered the inflation. I refer to the weakening of real investment
demand in the business and housing sectors and to the acceleration
of the decline of net exports which had set in earlier.
A survey of anticipated expenditures taken in the opening weeks
of 1948 foreshadowed a gradual downslide of real expenditures for
plant and equipment by nonagricultural business during the year.
The actual peak in physical volume occurred in the first quarter, but
when account is taken also of investment by farmers, professional men,
and institutions, the peak of total business fixed investment is delayed
until the third quarter in physical terms and until the fourth in dollar
amount (chart 5).
The early decline of nonagricultural business investment was concentrated in manufacturing industries, where it probably reflected
diminished needs for deferred replacements and for capacity expansion to meet postwar levels of demand; modernization and expansion
programs had progressed more rapidly in manufacturing than in
railroading or electrical utilities, to name two important sectors in
which investment continued upward during 1948. Some reduction
of investment demand was to be expected once firms had attained
satisfactory postwar relationships between capacity and output, even
if output continued to increase as rapidly as before. Actually, manufacturing production increased rather slowly during 1947, and this
fact may have influenced the formation of investment plans at the
turn of the year. In some industries the retardation during 1947
was due to inadequate plant capacity or shortages of materials, but
in many product lines it must be traced to a diminution of consumer

162

ECONOMIC STABILITY AND GROWTH

demand as households rebuilt their "capacity" to derive current services from durable or semidurable goods.
The foregoing reflections serve as a reminder that the inflation was
driven throughout by demand elements which were to an important
degree independent of the level or rate of change of real national income. Satisfaction of backlog investment or consumption demands
would in itself tend to reduce inflationary pressures at any given level
of real income and independently of financial considerations.
This is not to say that the slowdown of business investment was
completely unrelated to financial developments, but the latter influences were secondary at most. Although corporate liquidity had
decreased as the inflation progressed, investment funds from current
operations improved as the ratio of gross retained earnings to corporate investment increased from 48 percent in 1946 to 69 percent
in 1947 and 81 percent in the following year. External funds, moreover, were easily obtainable. Long-term interest rates had been allowed to rise by the Federal Reserve in the closing months of 1947.
The increase was moderate, however, and the monetary authorities
took the necessary steps then and later to supply reserves to the banking system whenever the stability of interest rates was threatened.
The money supply declined fractionally after the first quarter of 1948,
but bank loans rose substantially and income velocity and total spending continued to increase.
Financial constraints were more important in the areas of home
building and foreign trade. The rise of gilt-edge interest rates had
the effect of making GI and FHA mortgage loans at fixed rates less
attractive to lenders, many of whom tightened up on loan applications.
From April 30 to August 10 there was no statutory authority for
FHA action on the most liberal type of mortgage for lower priced
houses or on mortgage insurance for multifamily rental housing
units. Housing starts fell sharply beginning in July, and it appears
that tightened purchase terms and reduced availability of lower
quality mortgage credit should receive a good deal of weight in an
explanation of the decline.
Net exports form the third important category of final demand
which weakened well before the downturn of aggregate activity. Net
foreign investment plummeted between the third quarters of 1947 and
1948 because foreign gold and dollar holdings had been depleted by

ECONOMIC STABILITY AND GROWTH

163

the extraordinary purchases of 1946-47 and because loan-financed
aid by the United States Government was slashed (chart 4). Government expenditures in the form of gifts under interim aid and the
Marshall plan rose during 1948 and offset a part of the reduction of
foreign investment, but the export surplus nonetheless fell substantially, and on an. annual basis accounted for seven-tenths of the total
drop of $7 billion in foreign investment. Fortunately, the bulk of the
decline in the export surplus came before mid-1948 while other elements of demand were still favorable to the maintenance of physical
activity, so that its major effect was to relieve inflationary pressures
rather than to initiate a contraction.
After this lengthy recital of early developments that sapped the
strength of the underlying forces which had fostered the inflation in
previous years, the reader who does not have the details of 1948 fixed
freshly in mind may be surprised to learn that the indexes of wholesale and consumer prices spurted yet again after March, and by September had risen respectively another 3.5 and 4.6 percent. This last
fillip was given to the inflation by the actual and anticipated effects
of Government fiscal actions. These included final approval of a new
foreign-aid program under the Marshall plan, enactment of a steppedup defense program, and a reduction of personal income taxes effective April 2.
The bulk of the substantial increase of Federal expenditures between the first and third quarters of 1948 (chart 8) resulted from
foreign aid and the agricultural price-support program. Domestic
defense expenditures rose only in the final quarter of the year. The
economic stimulus provided by the defense program did not await
the increase of expenditures, however, but was partly felt during the
spring when its discussion and enactment affected business expectations and when substantial orders for aircraft were placed. The tax
reduction could also be expected to have an early effect on retail sales
because withholding from wages and salaries would be decreased at
an annual rate of $3 billion beginning in May.

23734—58

164

ECONOMIC STABILITY AND GROWTH
Chart 8
Government Purchases and Personal Consumption Expenditures
Seasonally Adjusted Quarterly Totals at Annual Rates ,1946-1950
(Billions of Dollars)
P
T

Log Scale

Government Purchases of Goods
and Services

60
40

v
30

Total

V^^~^

\ Federal

/""

>
..

State and Local

^^^^^

10
9
i

Personal Consumption Expenditures
200
150

Total

100
80

Nondurable Goods

60
50
Services
40

20
Durable Goods

1946

1947

1948

1949

1950

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the National
Bureau of Economic Research.
Source: Department of Commerce.

ECONOMIC STABILITY AND GROWTH

165

It was in this expectational context that prices of agricultural products turned upward again in April and that springtime wage negotiations were conducted in the large durable goods industries. Heavy
industry had resisted wage increases following the agricultural price
break of February, since it appeared that the cost
of living would
stabilize and that demand pressures were easing.2 When farm and
food prices recovered and the defense program and tax reduction were
announced, the determination of labor to win wage increases was augmented and that of management to resist them diminished. The
result was a series of wage and price increases in major heavy industries between May and July. Thus prices advanced on a broad
front from April through July or August, with participation by
agricultural and industrial commodities alike.
During the second half of the year the forces of deflation gained
final ascendancy. The downturn developed gradually from an accumulation of depressing factors, each mild in itself, but in sum sufficient finally to cause a decline of aggregate activity. For reasons
already discussed, residential construction eased downward in the
third quarter and business fixed investment in the fourth. These developments had earlier origins, but a new depressant was added when
retail sales leveled off during the third quarter (chart 7). The retardation affected particularly sales of nondurables, and since stocksales ratios were being closely controlled by retailers, new orders were
immediately trimmed to prevent further accumulation of stocks,
wholesale trade in nondurables declined after July and factory sales
after September, and manufacturing production of such goods eased
during the third quarter and again in the fourth.
The retardation of retail sales was the more disappointing because
it followed the April tax cut. Price and wage increases augmented
personal incomes during the second and third quarters, and the tax
reduction boosted disposable personal income even more, but these
developments had the effect mainly of raising saving instead of consumption expenditure (chart 4). Personal saving increased from an
annual rate of $4.2 billion in the first quarter to $10.6 billion in the
second and $12.6 billion in the third. The corresponding saving percentages were 2.4, 5.7, and 6.6. Thus the retardation of consumer
spending occurred because spending did not keep pace with income
and not because income failed to increase as rapidly as before; on
the contrary, disposable income increased faster during the spring and
summer than during the preceding half year.
The decline of consumption expenditures relative to income was a
consequence of the diminution of pent-up consumer demands. To
the extent that it was an expression of a general easing of demand,
it was probably influenced by the progressive decline of liquidity during the preceding years of inflation. Individuals' holdings of liquid
assets had increased more slowly than disposable income, inflation had
reduced their real value 15 percent between the end of 1945 and 1947,
and there may have occurred a shift of assets to firmer hands. Certainly the intense desire to increase expenditures in any and all directions which had characterized the first year or more of the inflation
was no longer manifest, and this argues for the presence of some
general constraint such as diminished liquidity.
2

Midyear Economic Report of the President, July 1948, p. 33.

166

ECONOMIC STABILITY AND GROWTH

Since household stocks of durable and semidurable goods had been
replenished by 1948, however, the satisfaction of deferred demands
proper may have provided the stronger general constraint. That this
is so is suggested indirectly by the behavior of automobile purchases,
for this was the principal example of a consumer good still in short
supply. The rate of real expenditure depended upon production and
showed no signs of flagging during the spring and summer of 1948
or indeed for many quarters thereafter. Consumer credit controls
had been dropped in November 1947 and were not reimposed until
September 1948. It is probable that had more cars been produced in
mid-1948, more would have been purchased, and at higher prices if
desired by sellers. Other expenditures would not have decreased
correspondingly, moreover, because any current retrenchment of consumption outlays which might be made by families purchasing new
automobiles would be smaller than the total outlay for the automobile,
the bulk of which would be financed by asset liquidation or new credit.
The emerging inventory problems of retailers were not confined to
nondurable goods. Expenditures for furniture and household equipment also leveled off in the third quarter and actually dropped in the
fourth. Inventories of durable goods were not quite so closely controlled by distributors, however, and factory production of major
durable household goods was maintained a few months longer than
was true of nondurables, before joining the general decline after October. Thus by the final quarter aggregate activity was being depressed
by reductions in business fixed investment and residential construction, by stable or falling retail sales, and by diminished inventory
demand consequent upon the latter. About the only important category of demand still expanding was Government expenditure, and
this was not enough to prevent a downturn of aggregate physical
activity in November or December.
Prices fell along with physical activity as inflationary pressures
ebbed. The indexes of wholesale and consumer prices led the October
peak of industrial production by 1 month. Agricultural prices had
started downhill for certain in July because increased production at
home and abroad augmented domestic supplies at the same time that
retail sales of food and other nondurables stabilized. Retail food
prices dropped in concert with farm prices. Wholesale prices of
products other than farm commodities and foods did not fall until
December, however, and the Consumer Price Index of all items less
food also held firm until year end. Reductions of industrial finished
goods prices at wholesale and retail awaited the further easing of
demand which came after the turn of the year.
MORE ABOUT THE INFLATION

The many references to demand and expected demand in the preceding pages will not have escaped the attention of the reader. The
inflation was started by the discharge of excess demand after decontrol
and it was ended when final demand weakened even as real income
continued to rise and largely because backlogs were worked off in
many lines. This does not end the matter, however, for the progress
of the inflation—its duration and magnitude and the timing of the
successive waves—might have been quite different in another economic

ECONOMIC STABILITY AND GROWTH

167

•environment. As a bald statement, this is a self-evident proposition,
but it is one which is often forgotten in relation to specific historical
episodes, where the role of disturbing forces is likely to be emphasized
to the neglect of factors which condition the response of the economy
to disturbances and which are equally to be regarded as causes of the
observed behavior.
There can be no doubt, for example, that the inflation could have
been terminated earlier and with a smaller overall price rise by sufficiently vigorous monetary or fiscal actions. As it was, external
monetary intervention did comparatively little to bring the inflation
to a close and fiscal actions actually prolonged its final phase. Not
much is to be gained by speculating about what might have been,
however, especially since it cannot be known how far alternative
policies could have been pursued without producing consequences less
attractive than the gradual diminution of liquidity through price
inflation and the concomitant gradual elimination of the more pressing physical shortages.
What can be done with profit is to inquire more closely into the time
path of the inflation, given the fact that external intervention was
minimal. We ask why it progressed in 3 waves instead of 1 uninterrupted advance, and why each wave was smaller than its predecessor (chart 1). Let us test first the hypothesis that the successive waves were recurrent phases in a wage-price spiral which was
started and kept in motion by a series of wage increases. In that
event, the interval between waves would be determined by systematic
lags in the process of wage-price-wage adjustment, and the magnitude of the waves would depend upon the size of the successive adjustments.
The hypothesis is leant a surface plausibility by the fact that wages
led prices during the first two waves (chart 9). Thus average hourly
earnings of manufacturing workers increased more rapidly than consumer prices during the first half of 1946. Prices moved ahead of
wages after decontrol, however, so that by early 1947 the purchasing
power of wages was considerably reduced. Wages spurted again during the spring of 1947, and after a lapse of 2 or 3 months, prices
resumed their climb. Up to this point, then, there were alternate
spurts of wages and prices. This neat phasing broke down in the
third round of inflation, however, when prices and wages rose simultaneously in the spring and early summer of 1948. Apparently
there was nothing inherent in the organization of the economy to make
prices lag wages. But this reminds us that special circumstances may
account for the two earlier instances of lagging prices. That is certainly true of the first round, when prices were decontrolled later
than wages. Eecall also the unusual conditions that prevailed at the
time of the second lag, when public attitudes and short-term expectations were unfavorable to price increases and when springtime wage
increases were absorbed for a few months in steel and other heavy
industries.

Chart 9
Indexes of Average Hourly Earnings and Prices
(Monthly, 1946-1949)
Index
1947-49 "100

Index
1947-49'100

7 in

Wholesale

105

Prices^^

100

95.

jfj
/Average Hourly
JF
/
Earnings
90 ~
I
jr (Manufacturing)
Consumer M s
Prices ?M /

35
0

85
/
/

^r
Average Hourly Earnings
(Nondurable Goods Manufacturers)

75
70

Average Hourly Earnings
(Durable Goods Manufacturers)

105

100

110

J.J.U

y
-

70
65

•

1946

1947

1948

1949

Vertical lines indicate business cycle peaks (P) and troughs (T) as
dated by the National Bureau of Economic Research.
Source: Bureau of Labor Statistics.

1946

1947

1948

1949

ECONOMIC STABILITY AND GROWTH

169

Apart from timing, there is the question of amplitude. May not
the progressively smaller size of the price waves be due to the fact that
each round of wage increases was smaller than the last, beginning
with a "typical" raise of 181^ cents in 1946 and moving on to 15 cents
in 1947 and 10 to 13 cents in 1948 ? The answer must again be "No".
The reason may be seen by asking under what conditions the statement
might be true. It implies first of all that the role of demand is passive ; that the amount of a price increase is determined by the size of
an autonomous wage increase and that demand merely rises enough
subsequently to sustain the higher price. Applied to a sequence of
damped wage-price movements, it implies secondly that some mechanism exists to make each wage increase smaller than the last, unless,
of course, this important fact is left unexplained. A plausible hypothesis for the sort of situation now under consideration is that wage
demands are equal to the increase in the cost-of-living since the last
wage increase, and that each wage increase induces a less-than-equal
rise in consumer prices because of productivity gains, partial absorption of cost increments by manufacturers or distributors, or the sluggish response of prices of consumer services.
Now either or both of these assumptions may at times approach
realism, but neither can be accepted in the present instance. With
regard to the first, it is obvious that the margin of excess demand
which existed at the time of decontrol for reasons independent of the
wage advance was an important determinant of the size of the initial
price wave. Nor was the wage increase of the first half of 1947 as
large as the price increase of the last half of 1946. Wages caught
up with consumer prices momentarily in mid-1947, but this was partly
because they had gained relatively in the 4 or 5 months preceding
decontrol and not because they advanced as much as prices in later
months. All the same, there is empirical substance in the notion that
the amplitudes of the price waves of 1947 and 1948 were approximately
determined by the amplitudes of the corresponding wage movements.
This does not mean, however, that we have here an example of "costpush" inflation.
It is important to recognize the full implications of these last remarks. The crux of the matter is that wages and prices may spiral under conditions of either "cost-push" or "demand-pull." The "costpush" terminology implies that the spiral originates and is kept going
by autonomous wage increases because workers are trying to increase their share of real national income, whereas "demand-pull"
carries the connotation that prices rise because of demand pressures
and induce subsequent wage increases as workers strive to maintain
their income position. A spiral may occur in either case, however,
and for the same underlying reason—the attempts of contending parties to defend or increase their shares of the national product. These
attempts may embrace investment goods as well as consumer goods,
and the contenders are not limited to those facing one another across
the bargaining table.
Since a spiral may occur for either reason, it may be difficult to
decide whether a particular experience is caused by one or the other
or some combination of the two. The task is complicated by the fact
that lags in the process may be short or even nonexistent when current
changes on either side are influenced by expected changes on the

170

ECONOMIC STABILITY AND GROWTH

other, and the further fact that independent disturbances may enter
the picture as well. Let us nevertheless examine the movements of
wages and prices during 1946-48 against the background of the earlier
analysis of the period, which should help us around these difficulties.
It is easy to detect changes in basic wage rates in the statistics of
average hourly earnings for a given industry, since abrupt vertical
steps appear in the data whenever such changes occur. Unfortunately,
these vertical movements tend to be smoothed out when hourly earnings are averaged for large numbers of industries, because the individual changes are spread over time. The smoothing is not complete,
however, for some bunching does occur, and at those junctures the
curve accelerates (chart 9). If a close positive relationship exists
between changes in wages and in prices, it should be possible to observe corresponding accelerations of prices. For reasons already discussed, we take the timing and amplitudes of prices and wages during
1946 as given, and inquire into the subsequent reactions to this first
wave of inflation.
The first important retardation of prices came during the second
quarter of 1947. Was this the result of a prior deceleration of wages
in manufacturing industries, which because of lags in the processes
of production and distribution did not affect prices of goods to final
users for several months? That is not the explanation, since the
retardation was in prices of nondurable goods and especially foods,
and since wages had accelerated in the nondurable goods industries
in the preceding quarter and had been accompanied at that time by
correspondingly rapid increases in prices of finished nondurables at
wholesale and retail (compare charts 2, 3, and 9). The cause of the
lull in prices was the reduced inventory demand analyzed earlier,
coupled with the temporary absorption of springtime wage increases
in the durable goods industries at their source.
Nor should the resumption of the inflation in the second half of
1947 be interpreted merely as a lagged response to previous wage increases. Prices moved ahead partly under the stimulus of rising demand. This is not to say that wage increases were unrelated to the
movement, however. In the first place, they affected its timing. The
coal mine wage-price increase of July and the August rise in steel
prices helped touch off the new round of price increases in durable
goods. Together with the simultaneous increases in agricultural
prices, these specific advances contributed to the upward revision of
price expectations which stimulated inventory demand. Secondly,
the size of the price adjustment was doubtless influenced by the
amount of the increase of labor costs under the 1947 wage agreements.
Thus average hourly earnings rose 7 percent between April and November. Allowing for the price lag, we find that wholesale prices
of finished goods increased 8.5 percent and the Consumer Price Index
7.5 percent between June and January.
Expected demand permitting, it is natural to suppose that business
firms will increase prices at least proportionately to unit labor costs.
The increases may not hold if the necessary demand fails to materialize, but under conditions of high liquidity, easy credit, and
strong "real" needs for producer and consumer goods, money outlays
will usually rise sufficiently to sustain the increase, and perhaps even
more if speculative buying develops during the process. Because

ECONOMIC STABILITY AND GROWTH

171

profit margins may vary firm conclusions are impossible, but it appears likely that had the average wage increase been smaller during
1947, the average price increase would also have been smaller. Notice,
however, that this would have meant that wage earners must needs
resign themselves to a lower real wage than they had received prior
to price decontrol. As it was, their attempt to compensate for the
rise in the cost of living was successful for only a month or two during
the second quarter of 1947, and once the new wave of price increases
got underway wages again fell behind.
The wage pattern of 1948 was quite similar to that of 1947, but
prices behaved rather differently because independent demand forces
assumed greater importance. It will be recalled that one effect of
the agricultural price break of February-March was to foster a holdthe-line attitude on wages and prices in the heavy industries. When
the tax cut and defense and foreign aid programs were approved,
however, it appeared that demand would prove adequate to sustain
new price increases, and the result was a simultaneous advance of
wages and prices during the late spring and summer. Average
hourly earnings in durable goods manufacturing plants and wholesale prices of finished durable goods both increased 6 percent between
May and September. Wages rose about as much in the soft goods
industries, but in that sector demand was already weakening and
prices of farm-produced raw materials were beginning to slide, so
that prices of finished soft goods merely held their own instead of
rising (chart 2). The same combination of easing demand and enlarged supplies forced food prices downward beginning in September,
so that the consumer price index declined and real wages improved
correspondingly during the closing months of the year.
This third wave is an interesting example of joint and simultaneous
participation by demand and supply in an inflationary episode. On
the one hand, it appears that substantial wage boosts would not have
occurred without the announcement of the new Government programs,
and on the other, the wage-price increases did not await an actual increase of final demand, and average prices of durable goods rose just
enough to compensate for the average wage increase. Nondurable
goods were subjected to approximately the same wage pressure, but
present demands were clearly less intense and since the unconcentrated market structure of most of the nondurable goods industries
makes prices quickly responsive to market forces, prices did not move
up commensurately with wages.
We may sum up the principal analytical lessons from the wageprice behavior of 1946-48 as follows. The existence of a wageprice spiral does not in itself show whether it is due "basically" to
cost or demand forces. Lags between wages and prices are not necessarily indicative of causal sequences and special factors may sometimes account for their occurrence. When wages rise autonomously
and prices are marked up correspondingly, demand must play at
least a passive role in sustaining the price increase if output is not
to fall. Expected increases of demand may induce wage-price reactions before final demand actually rises, in which case wages and
prices are likely to advance simultaneously. An observed sequence
of price waves may be the result of the interplay of wages and prices
in response to a single initial disturbance, but one must be alert to the

172

ECONOMIC STABILITY AND GROWTH

fact that autonomous forces operating outside of the mechanism of
the spiral may exert a decisive influence over the course of events.
With regard to the pattern of the inflation itself, it was initiated
by the excess demand released by decontrol, revitalized during 1947
by a combination of induced wage increases and largely permissive
demand increments (including anticipatory inventory buying which
was itself prompted by rising prices), and prolonged in its third
phase by an autonomous increase of expected demand which induced
a simultaneous wage-price advance. The duration of the waves and
the intervals between them were modified by wage behavior, but cannot be explained by systematic lags between wages and prices. The
inflation would doubtless have moved faster and with fewer pauses
had it been customary to negotiate wage contracts twice a year rather
than annually, or had cost-of-living adjustments been written into
the contracts. Whether it would have been greater or smaller in
overall magnitude depends upon a number of unknown factors, such
as the extent to which frequent wage-price revisions would affect
price expectations, the vigor of anti-inflationary public policy under
the alternative conditions, and the relative importance of real and
monetary stimuli to demand under circumstances in which liquidity
was reduced more rapidly by price advances, giving less time for the
reduction of physical backlogs of demand.
PRICES DURING CONTRACTION AND RECOVERY, 1 9 4 9 - 5 0

The truly remarkable feature of the 1949 contraction of business
activity was that it was so mild. What factors were responsible for
the "* * * unique and fortunate experience of liquidating
a major
inflation without falling into a severe recession"3 and why was the
price decline so moderate in comparison with the increases which
came before and after? (See chart 1.) These are interrelated questions, for the moderate price decline was both consequence and cause
of the mildness of the contraction in physical activity. What one
would like to disentangle if he were sufficiently wise to do so, is the
relative importance in the contraction of transient factors specific to
it and of more lasting attributes of the postwar economy. It will be
helpful to keep this ideal in mind during the subsequent discussion.
The contraction was mild enough to be called an inventory recession, and it has often been described by that teim. Insofar as the
term is taken to mean that final demand declined little and that most
of the moderate fall in gross national product was accounted for by
inventory change, it is quite accurate. This leaves unsettled the
question, however, of whether inventory investment exerted more
than a passive influence over the course of events. The amount of
decline of inventory demand depends in the first instance, after all,
on the behavior of final expenditures. If these drop only moderately,
current production will soon fall below sales, attempts to adjust
stocks to the smaller volume of sales will meet with early success,
and new orders and production will quickly revert to equality or more
with sales. Minor contractions are kept minor by those factors—not
3
The quotation is from The Economic Situation at Midyear 1949, A Report to the
President by the Council of Economic Advisers, p. 3, included in Midyear Economic Report
of the President; July 1949.

ECONOMIC STABILITY AND GROWTH

173

always the same in every contraction—which foster stability of final
demand. Let us first inquire what they were in 1949, and then go on
to consider the role of inventory investment and whether it was partly
an active one.
The behavior of the major components of final expenditure is summarized in charts 5 and 8. The following facts stand out: State and
local government expenditures rose strongly throughout the contraction, just as they have done every year of the postwar era under the
impetus of population growth and migration and because of the stubborn persistence of backlogs of needs for community services. Federal expenditure also increased during the first 6 or 7 months of the
contraction, and the subsequent decline of defense spending was
small and did not begin until other sectors were assuming active
importance in the recovery. Thus autonomous demands on the part
of governmental units moderated the decline and favored a prompt
upturn.
The early revival of residential construction and personal consumption expenditure is also noteworthy. In both cases backlogs of
real demand remained large—automobiles were at the forefront in the
improvement in retail sales—and easier credit facilitated renewed
expansion. Not only was credit eased generally by the cyclical fall
of demand for external funds on the part of business and the expansionary monetary policy of the Federal Eeserve System, but selective
devices were used to improve terms of home mortgages and consumer
installment credit. Indicative of the strength of combined real and
monetary demand stimuli is the substantial rise of consumption expenditure which commenced after the first quarter of 1949 and continued even as disposable personal income dropped through the remainder of the year (chart 4).
Consumption expenditure would doubtless have increased as a percentage of disposable income in any event, given the special stimuli
just discussed and the regular tendency for consumers to defend their
living standards against falling incomes, but disposable income
actually declined comparatively little, and this meant that expenditures rose rather substantially in absolute amount and became a positive force for expansion. Between the fourth quarters of 1948 and
1949 disposable income decreased about $6 billion and consumption
increased $3.5 billion. Automatic stabilizers cushioned the contraction of income and helped foster the expansion of consumption. National income dropped $16 billion, but induced reductions of undistributed corporate profits and corporate profit tax liabilities absorbed
about $6.5 billion of the decline, while nearly $2 billion was offset by
smaller personal taxes, and as much more by increased transfer payments which supplemented the flow of production income.
Business fixed investment led on the downturn, lagged on the upturn, and fell about 12 percent in physical volume between the third
quarter of 1948 and the first quarter of 1950. Thus it was by all
odds the major depressant of final expenditure at work in the contraction. The point which deserves greatest stress, however, is not that it
declined, but that it did not decline much more. This is a good part
of the explanation of how the inflation was liquidated without a severe
contraction. The inflationary waves of 1946-48 had not engendered
overly optimistic long-term investment expectations and had not caused
an unwarranted expansion of productive facilities. The available

174

ECONOMIC STABILITY AND GROWTH

data on capital stock in relation to production and on capacity utilization by industry indicate that capital facilities were smaller in relation to output before the contraction began than they had been in
1929 or 1937. This favorable start might have been overcome had
other elements of final demand decreased sharply during the contraction, of course, but the independent strength of Government and
consumer spending prevented the development of any widespread condition of excess capacity.
The situation was much the same with regard to inventory investment. Inventories were not excessive before the downturn began—
partly, of course, because they had been watched so very carefully
during 1948, with mildly deflationary results during the latter months
of that year in the nondurable goods industries. Even by the end of
1948 they were below the prewar relationship to sales in manufacturing and distributing alike (chart 6). Here is rather strong evidence that the restocking of 1946-48 was needed to reestablish normal,
working relationships between stocks and sales, and that speculative
inventory purchasing in anticipation of advancing prices had not
gotten out of hand. This meant that inventory disinvestment during
1949 was confined largely to that required to bring stocks down to the
reduced level of sales and production and was uncomplicated by any
need to liquidate speculative holdings. Even the feedback effect of
inventory disinvestment on production income and hence upon final
demand was limited by the automatic changes in profits, taxes, and
transfers mentioned earlier.
Inventory disinvestment could still have become a powerful independent force in the contraction had adverse price expectations and
a psychology of liquidation-at-any-cost developed as it progressed.
The initial absence of speculative holdings in itself inhibited such a
development, however, and other factors were at work as well, including some significant changes in the institutional structure. I refer to the agricultural price-support programs, the Government commitment to promote full employment contained in the Employment
Act of 1946, and the augmented importance and strength of labor
unions in the mass-production industries dating from the midthirties.
Taken altogether, these factors limited actual and potential price
declines, thus undercutting such nascent fears as may have existed
about the stability of prices.
The price declines which did occur were orderly and conformed
to the usual pattern of changes during periods of economic contraction. That is, prices of finished goods fell least and those of materials experienced greater reductions. This is typical of contractions because deflationary impulses are propagated backward with
increasing intensity from final sellers to manufacturers and thence
to suppliers of raw materials; since, in their desire to trim inventories,
firms at each stage of production or distribution curtail purchase orders by more than their own sales decline. The pressure on prices
resulting from the magnification of downshifts of demand at each
stage is especially powerful in markets where large numbers of sellers
are actively engaged in price competition and where output changes
slowly in response to demand. And, of course, the more prices of
purchased materials weaken at each market level, the greater is the
scope for reductions in prices of finished goods. It is in the light of

ECONOMIC STABILITY AND GROWTH

175

these facts that the importance of agricultural supports as a cushion
to the price structure should be viewed, since farms supply the basic
materials for a large number of nondurable goods. It is relevant also
that a good part of the potential decline of farm prices during 194849 occurred before the contraction of aggregate business activity was
actually underway (chart 2).
Additional support to prices was forthcoming from the downward
inflexibility of wage rates (chart 9). Since the same factors which
act to depress prices of purchased materials and components during
contractions operate in the reverse direction during upswings, the
margin between prices of unfinished and finished goods will usually
have narrowed during an expansion and will provide room for some
reduction of material costs relative to finished-goods prices during the
ensuing decline. Under given circumstances, there is only so much
scope for cost reductions from this source, however, and, once that
limit is reached, unless wage rates are cut, businessmen will prefer
to curtail production rather than reduce prices. This fact has several
ramifications.
First, wage inflexibility retards the decline of finished-goods prices
for many months during a contraction. The wage lag provides time
for a moderate realinement of prices and inventories before deflationary pressures become intense—and a moderate realinement may prove
sufficient if basic growth factors remain favorable to expansion. If
wage rates were to fall from the beginning of the contraction, the
chances of widespread postponement of purchases in anticipation of
future price declines would be greatly enhanced.
Wage rigidity during minor contractions is not a new phenomenon
confined to the postwar economy. Wages also lagged by long intervals
during the business cycles of the interwar period, and responded
scarcely at all to the mild downswings of 1923-24 and 1926-27 (chart
10). Hence, it cannot be asserted with absolute confidence that wages
were stable during 1948-49 and again in 1953-54 because organized
labor was stronger than in prewar years. Wages may have remained
firm because the contractions themselves were mild and short lived, so
that management determination to seek lower wage schedules was
never put to the test and might not have been even in the absence of
labor unions.

176

ECONOMIC STABILITY AND GROWTH
Average hourly earnings (cents)

S 8 S 8

£ ^ 2 9
( 0 0 1 : £261 !!-•<*V)

( 0 0 1 : S2-E26I^) luauiAotduw AJOJDOJ

That is not the end of the matter, however. If businessmen become
convinced for any reason that wage rates will not fall, they are, in
the same measure, convinced that substantial price declines are not
in the offing. These dual convictions must have been cultivated by the
growth in the power of organized labor. Indeed, there is some evidence
of this in the behavior of wages and prices during the severe contraction of 1937-38, which occurred after the mass-production industries
were organized. It is inconclusive evidence, because the contraction,

ECONOMIC STABILITY AND GROWTH

177

though severe, was short lived, and fell within the range of the normal
timelag of wages behind business downturns. All the same, the fact
that the wage and price reductions associated with that contraction4
were comparable to those of 1923-24 and 1926-27 rather than 1920-21
is consistent with the a priori prediction of greater rigidity. Granted
that businessmen expected greater rigidity, this factor would be particularly important in a contraction such as 1948-49, which followed
a sharp inflation. This is because it would modify price expectations
which might otherwise become quite pessimistic, and, by so doing,
help to prevent the contraction from becoming severe enough ever to
press heavily on wages.
The final aspect of wage-price rigidity to be discussed is, perhaps,
the most important of all in connection with the contraction of 194849. It is doubtful that there existed then or exists now any sizable
segment of opinion in business or Government circles that widespread
wage reductions are beneficial as a means of combating unemployment,
or that a deflation of wages and prices is a necessary though painful
penalty for permitting an inflation to run its course. To quote the
Council of Economic Advisers in mid-1949:
There would be no purpose, and much potential damage, in an attempt to get
back to some drastically lower price level by wage cutting, since incomes are now
geared to prices substantially higher than before the war or immediately after
the war. Such an effort would involve a deep and vastly unsettling decline
in wages as the accompaniment to a prolonged period of severe depression. From
that, practically no one would gain; the economy would lose tragically.5

This attitude was reflected in the fiscal and monetary policies of the
day. It contrasts strikingly with the situation of 1920-21, when deflation was regarded as a necessary evil, and when Federal Reserve
actions were taken to tighten rather than ease credit. The differing
attitudes toward the inevitability and desirability of wage-price deflation, and the associated differences in public and private actions, go
far toward explaining the greater severity of the earlier experience
despite the presence of underlying elements of strength similar to
those following World War II.
In summary, prices declined comparatively little during 1949 because speculative excesses had not Ibeen a feature of the inflation;
because institutional factors and public policies minimized actual and
expected declines and prevented cumulative deterioration of shortterm price and sales expectations; and because, in those favorable
circumstances, the forces for expansion inherent in the postwar population upsurge and remaining backlogs of public and private demand
could quickly assert their influence. The deflationary force of inventory disinvestment was largely spent by midyear, and a recovery led
by Government spending, homebuilding, and automobiles actually
began at that time. The upswing of industrial production was interrupted during the autumn by coal and steel strikes, but expansion was
quickly resumed and developed rapidly during the first half of 1950.
Prices responded, as they normally do in a period of recovery, with
gentle increases for materials but little change for finished goods.
* Creamer, Daniel, Behavior of Wage Rates During Business Cycles, Occasional Paper 34,
National
Bureau of Economic Research, 1950, New York, N. Y., p. 37, table 5.
8
Op. cit, p. 8.

178

ECONOMIC STABILITY AND GROWTH
THE KOREAN INFLATION AND ITS AFTERMATH, 195 0 - 5 4

The normal processes of recovery had carried the economy most of
the distance toward full employment by mid-1950, at which time the
economic forces set in motion by the Korean conflict intervened to alter
radically the character of the expansion. Prices were forced up violently in 2 consecutive waves of forward buying which swept over the
country during the first 9 months of hostilities, and, although wholesale
prices subsequently receded, the average of consumer prices never did
(chart 1).
It will not be necessary to follow the events of the inflationary phase
of the Korean expansion in detail, since the main facts are easily read
from the behavior of expenditures, incomes, and prices during the
period. First, we note that military spending was not the sole or even
the principal cause of the buying waves, although the placement of
defense orders doubtless stimulated business activity during the latter
months of 1950. Rather, it was the forward buying of consumers and
businessmen in anticipation of potential shortages which created acute
inflationary pressures. The speculative nature of the surges of consumer demand is readily apparent from the wide swings of expenditure relative to income, timed as they were with the onset of hostilities
and the later entry of the Chinese Communists into the war (charts 11
and 12). The fluctuations of retail trade were duplicated at wholesale and in manufacturers' new orders and sales, so that inventories
rose all along the line from factory to retailer (charts 14 and 15 ).
A general price and wage freeze was announced on January 26,
1951. From then until early in 1953, prices were under varying degrees of wartime administrative control, but it was a fascinating and
unexpected feature of those years that average wholesale prices actually declined substantially and, for many commodities, fell well below
legal ceilings. This occurred, moreover, during a period of continuous full employment of labor and while the Government share of real
gross national product was increasing from 7.7 percent in 1950 to 15.5
percent in 1952. What happened was that a decline of consumer
demand in reaction from the early buying sprees neatly offset the rise
of defense spending. This is not to say that price and allocation controls were unnecessary or ineffective, however. By insulating the several sectors of the economy from one another, they gave maximum
effect to the relaxation of consumer demand.
As 1951 opened, the production of consumer goods was geared to
high and rising retail sales. Household stocks had been augmented
and personal liquidity depleted during the buying waves, however,
and civilian goods appeared to be in plentiful supply and could no
longer rise much in price now that controls were in effect. Eetrenchment was indicated under the circumstances, and, beginning in February, retail sales declined suddenly and sharply, causing an equally
abrupt and unwanted increase of distributors' stocks during the next
few months (charts 14 and 15). The necessity subsequently to work
off these excessive inventories magnified and prolonged the deflationary impact of the cutback of consumption expenditure well into 1952,
despite a gradual recovery of retail sales after mid-1951.

179

ECONOMIC STABILITY AND GROWTH
Chart 11
The Nation's Income, Expenditure, and Saving By Major Economic Sectors
Seasonally Adjusted Quarterly Totals at Annual Rates, '1950-1954
(Billions of Dollars)

Consumers

.260
Disposable Income

•230

Personal Consumption Expenditures
170

Business

Gross Private Domestic Investment

vffn

-- L111 ' '

l^^ft^^

Excess of Investment

7fnTnml

'1 1 I [
Gross Retained Earnings £

i
Government

Receipts
(Less Transfer Payments)
60

Purchases of Goods and Service,
30

Surplus

Net Foreign Investment

+ 15

-15

1950

1951

1952

1953

1954

1/ Includes undistributed corporate profits, inventory valuation adjustment, and
capital consumption allowances.
Source: Department of Commerce.

23734—58

180

ECONOMIC STABILITY AND GROWTH
Chart 12
Government Purchases and Personal Consumption Expenditures
Seasonally Adjusted Quarterly Totals at Annual Rates, 1950-1954
(Billions of Dollars)

hog scale

90
80
70

Government Purchases of Goods
and Services

yS*^Total

" - — ^

—*»^

^ ^

40
/Federal
30

———
^

-0~~0~~"^

State and Local

300
Personal Consumption Expenditures
200
Total
150

100
90
80

m*^

nondurable Goods

70
~~

Services

50
40

——-

—S

^^S
Durable Goods

10
1950

1951

1952

1953

1954

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the National
Bureau of Economic Research.
Source: Department of Commerce.

181

ECONOMIC STABILITY AND GROWTH
Chart 13
Gross Private Domestic Investment and Its Major Components
Seasonally Adjusted Quarterly Totals at Annual Rates, 1950-1954
(Billions of Dollars)

-10

Components of Business Fixed Investment
j

Producers' Durable Equipment

\-X
20

"-*

wmmmmmm wmmm

Nonresidential

1950

1951

Construction

1952

1953

1954

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
National Bureau of Economic Research.
Source: Department of Commerce.

182

ECONOMIC STABILITY AND GROWTH
Chart 14
Indexes of Personal Income, Retail, Wholesale, and Manufacturers1 Sales
and Inventories and Manufacturers' New Orders
Seasonally Adjusted, Monthly, 1950-1954
Index
June 1950 s 100
135

150

100

1950

1951

1952

1953

Index
June 1950 -100
T

1954

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
National Bureau pf Economic Research.
Source: Department of Commerce.

183

ECONOMIC STABILITY AND GROWTH
Chart 15
Inventory-Sales Ratios
(Monthly, 1950-1954)
Per cent

170

\^v

A Retail

150

130

120

Wholesale
100

80
0

\
if

195

175

\—y\

Manufacturing

155

135
0

1950

1951

1952

1953

1954

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
National Bureau of Economic Research.
Source: Department of Commerce.

184

ECONOMIC STABILITY AND GROWTH

For a time it was as though two economies existed side by side without touching. The grand division was between consumption and
nonconsumption goods. Reduced consumption and inventory demands
depressed especially factory production and wholesale prices of textile, apparel, rubber, and leather products from early 1951 to mid-1952.
^Production of consumer durable goods also declined substantially, but
their prices remained at ceiling levels. The production cutbacks were
required by defense restrictions on materials, but retail demand had
eased for durables as well as nondurables, and distributors' stocks remained above the 1949-50 levels, so that supplies were adequate at
existing prices. Defense contracts and defense-oriented programs of
facilities expansion replaced the business lost in the consumer sector,
;and average prices and production of durable goods were generally
stable.
The downdrift was ended by mid-1952 insofar as nonagricultural
business is concerned. The gap which had opened between income
and expenditure with the retrenchment of consumer demand in 1951,
narrowed again in 1952 as consumer purchases rose somewhat faster
than income (chart 11). Inventory liquidation had substantially
improved the position of distributors by early summer (chart 15).
Allotments of critical metals to nondef ense purposes were increased as
defense requirements began to taper off toward a plateau planned for
1953. The incipient recovery of the consumer sector was interrupted
l>y the steel strike of the summer of 1952, but the work stoppage ended
late in July, and production, employment, and incomes expanded
swiftly thereafter as factories and distributors rebuilt inventories and
final sales of consumer goods accelerated. Thus the defense expansion was capped by a vigorous boom in civilian goods as it entered
its closing phases. No comparable rise occurred in prices, however.
Indeed, the indexes of wholesale and consumer prices decreased somewhat during late 1952 and early 1953—surely peculiar behavior for an
economy in which production was high and rising and unemployment
was less than 3 percent of the labor force. Nor is it adequate to observe
that sharp reductions in foods lowered the average of wholesale prices
and that prices of other finished goods actually increased a bit (chart
2), for the recovery of consumer demand might have increased nonfood prices even more were it not for price controls. Let us follow up
this thought and inquire into the causes and effects of the price divergencies which developed during the period of price and wage control.
Changes in relative prices may reflect a variety of possible forces
acting singly or in combination. They may be initiated either by
shifts of demand or supply among products; either type of shift may
occur for several different reasons; and the timing and amount of the
resulting price changes will depend partly on the structure of the
affected markets—for example, whether there are many buyers and
sellers, or so few as to confer some degree of individual control over
price. Now, one obvious fact about the shifts of consumption and
nonconsumption demand during the Korean period is that they occurred under conditions in which some prices were not free to move
upward in response to market forces whereas others could fall below
ceiling prices if demand weakened or supply increased enough. The
probable result is that the differential between prices of durable and
nondurable goods did not become as large as it otherwise would have

ECONOMIC STABILITY A3STD GROWTH

185

done, or alternatively, that the differential would have remained about
the same but that prices of durables would otherwise have risen more
and those of nondurables fallen less. In either event, the overall
average of prices would have been correspondingly higher.
Suppose by way of illustration, that no administrative controls had
been imposed on prices, wages, or critical materials and that no fiscal
or monetary curbs to private expenditure had been in effect when
consumer demand declined in 1951. Is it not likely that bottlenecks
in metals and metal products would have forced prices higher in those
sectors—including prices of durable consumer goods? Consumers
might have reacted to this by purchasing fewer cars or refrigerators
until prices came down again, or they might have substituted more
spending on durables for less on nondurables and depressed the prices
of the latter, but it is also quite possible that they would have increased total expenditure even at the same level of money income, thus
sustaining a higher price level. More important, of course, is the fact
that money income would not have been the same. Money wages
would have risen along with prices in the durable goods industries,
and the wage increases would have spread to other industries as well,
raising the whole structure of prices and money incomes at the given
level of real national income. Thus the relaxation of private demand
contributed so much to overall price stability partly because it occurred in an institutional complex that constricted the normal channels for transmission of inflationary impulses between sectors and
permitted deflationary forces full scope in the affected areas. By the
same token, of course, the task of containing inflationary pressures
by direct and indirect controls was greatly simplified by the voluntary
reduction of consumer demand and the associated inventory adjustments.
Perhaps even the preceding discussion has left the reader inadequately prepared for the amplitude of the price declines in nondurables during 1951-52 (chart 16). The reaction was especially severe
for prices of intermediate materials—intimately related to basic agricultural commodities—just as the preceding rise had been under the
impact of speculative inventory demand at home and abroad. Prices
of finished nondurable goods did not fall nearly as much, partly because the demand for finished goods always holds up better than that
for materials during periods of declining sales and inventory liquidation, and partly because wages in nondurable goods factories continued to rise along with those in durable goods industries (chart 17).
Average hourly earnings of manufacturing workers kept closely in
step with the cost of living until the summer of 1952. Consumer
prices moved ahead during the first Korean buying wave, but wages
responded quickly and by the time of the general freeze in January
1951, had achieved approximate parity with the Consumer Price Index
not only for factory workers but for other large groups of employees
as well. Wages were under direct control after the freeze, of course,
but gradual increases occurred as adjustments were granted to groups
whose real earnings had lagged in 1950, and to workers in general to
compensate for increases in consumer prices during the period of price
control. In this way higher wages became part of the cost-price-income structure of the economy, thereby limiting the price declines in
the weaker sectors. With the exception of foods, average wholesale
prices of nondurable finished goods never did fall back to the pre

186

ECONOMIC STABILITY AND GROWTH

Korean peak of 1948, and not even foods did so at retail (charts 3 and!
surged forward during the last phase of direct controls in
the latter half of 1952 and early 1953. The rise reflected a breakdown
of the hold-the-line policy on real wages, and was part of a general
movement to relax direct and indirect controls as it became apparent
that this could be done without undue danger to the mobilization program or to economic stability. Substantial advances for steel, copper,
and aluminum were permitted under a corresponding relaxation of
the criteria for price increases, and metal users were allowed to pass
through the resulting absolute cost increases but not to pyramid them
by proportional markups. Were it not for the ceilings, prices of
finished durable goods would doubtless have risen substantially after
the steel strike, under the combined pressure of heavy consumer and
inventory demand and the concomitant rise of resource prices. As it
was, they increased only fractionally before the final elimination of
price controls in April 1953.
Increases were small also among nondurables, however, where ceilings had been suspended for a large proportion of commodities, and
despite a substantial upswing of wages, production and sales. The
principal depressant at work—the emergence of agricultural surpluses—has ueen an important determinant of prices of nondurable^
ever since. In late 1952 and early 1953, increased domestic production
and a sharp drop of exports augmented domestic supplies sufficiently
to depress prices of food at wholesale and retail and to offset increases
of demand for materials important in the manufacture of nondurable
goods. During the last phase of the Korean expansion, then, the
spread which had developed between prices of durables and nondurables initially because of shifts in relative demands for final products, was kept open with the assistance of shifts in export demand
and domestic supply which were in good part independent of current
final demands.

ECONOMIC STABILITY AND GROWTH
Chart 16

Index
1947-49'10$

150

140

130

120

110

Indexes of Wholesale Prices of Intermediate
and Finished Goods
(Monthly, 1947-19&7)
P
T
P

Materials
T

A00

150

140

130

120

110

100

Consumer Other Nondurable Goods

1947
48
49
50
51
52
53
54
55
SB
57
Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the National
Bureau of Economic Research, except for the peak of August 1957 which was dated by
the author.
Source: Bureau of Labor Statistics.

187

Indexes of Average Hourly Earnings and Prices
(Monthly, 1950-1954)
Index June 1950-100

Average Hourly Earnings
(Manufacturing)

^ /

120

Wholesale Prices

Average Hourly Earnings
(Durable Goods Manufacturers)

[**<

115

Consu mer

Prices

110

105

r^f^
]f

f
I

_^i

Average Hourly Earnings
(Nondurable Goods Manufacturers)

/¥

"
, /

J f

115

100

125

100

1950

1951

1952

1953

1954

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
National Bureau of Economic Research.
Source: Bureau of Labor Statistics.

•

1950

1951

1952

1953

1954

ECONOMIC STABILITY AND GROWTH

189

The subtitle of the present section—"The Korean Inflation and Its
Aftermath, 1950-54"—was chosen advisedly from two points of view.
The contraction of 1953-54 was caused partly by the post-Korean cutback of defense expenditures and defense-oriented investment and
was in that sense an aftermath of Korea. More directly relevant is
the fact that price behavior during the contraction was heavily influenced by the preceding reaction to the inflationary waves of 1950-51.
These points will be discussed in turn.
The sharp rebound of business activity which came after the steel
strike of mid-1952 and under the spur of mounting private demands
has already received attention. The rate of recovery of sales and production was rapid indeed during the latter months of 1952, but the
pace of expansion slowed materially during the winter and spring of
1953 (charts 1 and 14). Retail sales had increased more rapidly than
personal income as consumers stepped up their purchases during the
closing months of 1952. Retail trade leveled off early in 1953, however, partly because total consumption expenditure increased only in
line with income instead of by more, and partly because this meant a
still more pronounced retardation of sales of goods as distinct from
services, just as the preceding acceleration of spending relative to income had been for goods rather than services. With sales flagging
and inventories still increasing, distributors cut back on their orders,
with the result that production of nondurables and of major household durable goods dropped after May, anticipating hy 2 months the
downturn of industrial production. Similar retardations appeared
prior to the actual 1953 declines of defense expenditure and fixed
investment, and probably contributed also to the downslide of new
orders which preceded the July peak of aggregate manufacturing
sales and production by many months (chart 14).
If the foregoing events sound a bit familiar, it is because they are
not unlike those of 1948. Retail sales also slackened at that time,
although for somewhat different reasons, and helped precipitate the
downturn. Again, fixed investment flattened out before actually declining in the earlier period. Other parallels are to be found in the
further development of the two contractions. Thus most of the decline of gross national product during 1953-54 was accounted for by
inventory change. Just as in 1949, the deflationary impact of inventory adjustment was largely spent after 6 months of contraction
(chart 13). Just as then, moreover, a substantial portion of the inventory decline was induced by falling retail sales in the early months
of contraction; and as before, inventory disinvestment was quickly
arrested because final expenditure held firm after its brief initial
lapse. Even the proximate cause of the stability of final expenditure
during the last half of either contraction was the same—an early
recovery of consumption demand and residential construction which
served to offset the further deterioration of final demand for business plant and equipment in both contractions and for Federal purchases in the later one.
I have emphasized these parallels as a reminder that "inventory
recessions" remain just that only if final demand has independent
strength and if short-term sales and price expectations do not deteriorate in cumulative fashion to undermine that strength. This is not
to say that final demand held firm for identical reasons in 1949 and

190

ECONOMIC STABILITY AND GROWTH

1954, although there were important common influences even there in
the operation of automatic stabilizers, the stimulus imparted to home
building and consumer spending by easier credit, and the favorable
state of long-term investment opportunities as evidenced by the comparatively minor declines of business fixed investment in both years.
A principal difference, of course, was that Federal spending rose in
1949 and fell in 1954, whereas tax reductions stimulated private income and demand in 1954 but not in 1949.
Another difference between the two contractions, and one which is
a little unexpected because the fluctuations of physical activity were
approximately equal in duration and amplitude, is that average prices
declined considerably during 1948-49 but edged fractionally higher
in 1953-54 (chart 1). Is this evidence that, in the short space of 5
years, prices had become increasingly resistant to business contractions because of structural changes in the economy? Or was the
greater stability during the second contraction due primarily to factors peculiar to it ?
One possibility is that wages were more rigid than before. This
is not borne out by the behavior of average hourly earnings in manufacturing industries, however, which leveled off as usual during contractions (chart 17). There is a point worth noting about the timing
of industrial wages and prices in relation to business activity, nonetheless. Although demand was weakening before the business-cycle
peak of July 1953, substantial cutbacks in production and sales did
not occur until after that date. This meant that wages and prices
were decontrolled a few months prior to the actual contraction. The
increases which came with decontrol were moderate, but they were
firm enough to hold during the ensuing recession, with the result that
industrial finished-goods prices averaged higher than in the months
preceding the downturn.
But price increases during the very months when the economy is in
the process of transition from expansion to contraction are not especially surprising if allowance is made for the gradual and gentle form
which the transition often assumes. The increases which occurred
in prices of finished goods during the latter months of 1948 come immediately to mind (chart 2). It is not the fact that prices went up
just prior to the downturn that distinguishes 1953-54 from 1948-^9,
but the fact that they stayed up throughout the contraction.
Since the decline of final demand was small and wages were firm,
the scope for industrial price reductions was limited largely by the
potential fall in prices of materials. As we have seen, substantial declines may occur from this source during mild contractions because
the demand for materials weakens considerably more than that for
final goods. Even this avenue was closed in 1953-54, however, principally because prices of nondurable materials had been forced down
previously—the adjustment which ordinarily would come with a
general contraction had already occurred because of the isolated contraction in that sector during 1951-52. Producers of intermediate
nondurable materials were caught in a wage-price squeeze because their
wage scales reflected the general boosts in the economy which came
after 1950, but their prices were no higher than in 1948 (charts 16
and 17). Still, these prices did decline a little in 1954, and would
have decreased more were it not that prices of crude materials were

ECONOMIC STABIIilTY AND GROWTH

191

lent independent support by the agricultural program. Again, agricultural price supports were already providing effective floors by
1953 because of the declines which had gone before, whereas in 1948-49
farm prices could fall much further before they were arrested by
supports.
It appears, then, that the price stability of 1953-54 was in good part
the result of factors peculiar to that contraction, and does not necessarily mean that prices have become permanently more rigid. To the
extent that agricultural supports remain a feature of the economic
structure, however, and especially when prices have been under support before a contraction begins, they will minimize cyclical declines
in prices of important classes of goods. Since any diversion of purchasing power which might arise, if prices of foods and other nondurables were lower in the absence of supports, would during a period
of contraction tend to stimulate larger physical sales of durables rather
than higher prices for them, it is likely that the overall average of
prices would be kept higher by the supports than it otherwise would
be. And, of course, the net injection of income from Government
sources via the support operation helps to firm demands generally.
If the cost of living were, indeed, prevented from falling as much as
otherwise, this would reinforce the tendency toward firmness by reducing wage adjustments under escalator clauses. This effect would
be comparatively unimportant, however, since downward wage escalation is usually limited bjr agreement during any given contract period,
and since small changes in wage rates up or down probably have little
influence over administered prices.
EXPANSION AND INFLATION, 1954-57

"Expansion and then inflation" would be a more descriptive heading for this section. The business expansion which began in August
1954 lasted exactly 3 years. During the first half of this span, the
average of consumer prices kept within a range of 1 index point, but
it began rising in March 1956, and has continued upward ever since,
with the advance persisting even into the present phase of cyclical
contraction. While not as rapid as the vigorous inflations of 1946-48
and 1950-51, the consumer price increases of the past 2 years were at
a substantial rate and were more than double the increase of disposable
personal income in constant prices.
Price stability for finished goods is to be expected during the initial
phase of expansion from a cyclical trough. Prices of materials may
recover rapidly from the very beginning, but these elements of cobt
will have fallen so low during the contraction that increases can be
absorbed for a time without encroaching on profit margins. Because
excess capacity will have developed during the downswing, owing to
the continued growth and modernization of plant and equipment and
the decline of production, moreover, real cost savings will accrue as
activity expands to make more efficient use of productive facilities
and to make use of newly more efficient productive facilities. These
real savings will make it possible to pay higher prices for both labor
and materials without raising production costs.
Purely cyclical factors are npt enough to account for the continued
stability of consumer prices into early 1956, however. Industrial

192

ECONOMIC STABILITY AND GROWTH

wholesale prices were fairly steady until mid-1955, but they increased
sharply thereafter (chart 2). Leaving this rise unexplained for the
moment, we notice that its effects on average consumer prices were
minimized by two factors of noncyclical origin. Prices of farm products and foods fell off sharply during the last half of 1955, as agricultural commodities continued in excess supply because of a combination of high carryover stocks and record or near record 1955 farm
outputs. Retail food prices declined in sympathy, though by smaller
amounts, tending directly to reduce the Consumer Price Index. Consumers probably took advantage of lower food prices partly by increasing their food consumption and partly by increasing other expenditures. If so, the induced shift of demand was insufficient to raise other
prices enough to compensate for the decline of food prices, since average prices were stable despite higher personal incomes.
One reason for the lack of fully compensating increases was that
persistent agricultural surpluses had also weakened prices of basic
materials used in other nondurable-goods industries, serving to restrain increases for finished goods in that sector. Another important
factor was a fall of retail prices of durable consumer goods in the face
of rising wholesale prices (charts 3 and 16). This was not a new
development, however, for retail margins on durable goods had been
narrowing for several years under the competitive pressures associated
with the growth of the discount house and emergence of a buyers' market in automobiles. The potential for retail price reductions from
this source appears to have become exhausted about mid-1956. Prices
of household appliances did fall slightly over the next 18 months, but
furniture prices rose moderately, and those of automobiles more substantially. In future, aside from seasonal changes in dealer margins,
significant reductions of retail price tags will probably await or induce
corresponding drops in wholesale prices.
Let us revert now to mid-1955 and the rise of industrial prices which
began at that time. We need to go back, in fact, just a bit further
in order to recall 2 important features of the recovery of 1954-55.
One was the spectacular reception accorded the 1955 automobile
models and the acceleration of consumer purchases of durable goods
for which it was primarily responsible (chart 18). The other was
the onrush of business plans for capacity expansion during the winter
and spring of 1955. These plans were a lagged response to the business
upturn, so that business fixed investment did not rise strongly until
the latter half of the year (chart 19). The upsurge was foreshadowed
during the spring, however, in surveys of investment plans and, even
more concretely, in new orders for producer goods. Thus, prospects
for 1955 sales and profits were excellent at the time of the wage settlements in major durable-goods industries during the late spring and
summer.

193

ECONOMIC STABUJITY AND GROWTH

Chart 18
Government Purchases and Personal Consumption Expenditures
Seasonally Adjusted Quarterly Totals at Annual Rates, 1954-1957
(Billions of Dollars)
Government Purchases of Goods and Services

Los Scale
9b
80
70
60
50

Federal
40
30

State and Local
Personal Consumption

Expenditures

300

'•

250

— —_

200

150

—

Total

Nondurable Goods

—
100
90
80

—

••

-—

—

Services

70
60

50
40
—

—•———

—

Durable Goods
1954

1955

1956

1957

Vertical lines indicate business cycle peaks (P) and troughs (T) as
dated by the National Bureau of Economic Research, except for the
peak of August 1957 which was dated by the author.
Source: Department of Commerce,

194

ECONOMIC STABILITY AND GROWTH

Gross Private Domestic Investment and Its Mayor Components
Seasonally Adjusted Quarterly Totals at Annual Rates, 1954-1957
(Billions of Dollars)
T

Gross Private Domestic Investment

^ ^ ^ ^ ~

^ ^

Business Fixed Investment

~mmmmmmm

Residential
—ii

•—'

Construction

^ _—

Inventory Investment
i

-10

Components of Business Fixed Investment

1954

1955

1956

1957

Vertical lines indicate bttsiness cycle peaks (P) and troughs (T) as dated by the National
Bureau
of Economic Research,
except for the peak of August 1957 which was dated by the author.
Source: Department of
Commerce.

1947-49'100

120 -

Chart 20
Indexes of Average Hourly Earnings, and Prices of Intermediate Materials
and Finished Goods
(Monthly, 1954-195®
Index
1947-49*100
j
Nondurable Goods Manufacturing
(Excluding Foods)
120
Average
Hourly Earnings

r
Durable Goods Manufacturing

115 -

Average Hourly Earnings

105

Prices of Finished Goods
mjj

/ Prices of Finished Goods

295

Prices of Intermediate Materials

100

•^Prices of Intermediate Materials
95

o
w
o

50 "

90
0 >

a 2-

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the National Bureau
of Economic Research, except for the peak of August 1957 which was dated by the author,
$ource: Bureau of Labor Statistics.

Chart 21
Indexes of Average Hourly Earnings and Prices
(Monthly, 1954-1958)
Index_
Vug.l954=100 2

-J.UU

125 -

125

120

Average Hourly Earnings
(Manufacturing)

115 -

110

105

j f

Wholesale

^ *
^—m^

Prices^/"^"""

100

115

110

105

Average Hourly EarningsJTT^
(Durable Goods) J r V

rjf
^/
-[Average Hourly Earnings
jV~ir
(Nondurable Goods)

100

0 >

95
i

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
National Bureau of Economic Research, except for the peak of August 1957 which
was dated by the author.
Source: Bureau of Labor Statistics.

1954

1955

1958

ECONOMIC STABILITY AND GROWTH

197

With product demands high and expected to rise further in many
industries, sizable wage increases were demanded and received by
employees in heavy industry. Parallel changes occurred in prices of
intermediate materials for durable goods manufacturing—comprising
principally primary metals, lumber, and plastics—and of durable finished goods (chart 20). Between June and November of
1955, average hourly earnings of production workers in durable-goods
factories increased 3.7 percent, and prices of intermediate materials
5.1 percent. Average wholesale prices of finished durable goods increased 3.4 percent between July and December, with consumer
durables up 2.6 percent and producer goods 4.3 percent.
Were these price raises caused by increased demands or higher
costs? The answer can only be: by both. Current demands for
durable goods had recovered strongly before the wage advances, and
further increases of demand were in the offing and the orders were
on the books. These facts certainly augmented the size of the wage
increases—contrast the increases of 1954 and 1955—but the timing
and magnitude of the price increases were just as definitely influenced by the behavior of wages.
Prices of nondurable manufactures also rose during the last half
of 1955. These increases were probably occasioned less by wage increases than by current demands. Eetail sales of nondurables moved
ahead strongly during 1955 and continued to rise as sales of durables
fell during the last few months of the year. For reasons to be discussed later, prices of nondurables as a class are quicker to respond
to demand shifts than are those of durables. Thus prices of nondurables dropped in the spring of 1955 despite substantial wage increases, and when average hourly earnings in nondurable manufacturing caught up with those in durable goods factories with a
spurt during the spring of 1956, prices were steady and did not rise
until late in the year (charts 20 and 21). Nor is the lag of price increases behind wage increases in 1955 and again in 1956 evidence that
the latter caused the former. Technical lags in the processes of production and sale cannot explain a wage-price lag on the order of 6
months; and in point of fact, prices of intermediate materials and
finished goods moved closely together from month to month, as
the chart shows. Wage and cost increases do influence prices, but so
also does demand, and in the instance of nondurables both factors
operate quickly and may sometimes be offsetting, sometimes reinforcing, so that no simple relationship exists between prices and either
factor alone.

198

ECONOMIC STABILITY AND GROWTH

A final feature of 1955 should be mentioned before we move on.
The year marks a revival of labor-management interest in the longterm contract which specifies deferred annual wage increases and
provides for escalation to protect the real value of wages during the
life of the agreement. Such contracts were negotiated in the automobile industry and with General Electric in 1955. They spread to
steel, meatpacking, and railroads in 1956. By the end of 1957, the
Bureau of Labor Statistics had on file long-term agreements involving
fixed wage increases of a deferred nature which covered 4.4 million
workers. The deferred increases specified by "annual improvement
factors" or "annual productivity increases" are generally in line with
the long-term average rate of growth of output per man-hour for
the economy as a whole, but that fact does not necessarily render
them noninflationary. Even apart from other determinants of price
and cost, the annual productivity gains in the industries directly
affected by the contracts may change unevenly over time and may
deviate widely from the average for the economy, both in any given
year and over the long term. And, of course, since the deferred
increases set a standard for other wages, they may lift the general
level of money wages at a faster rate than overall productivity rises
in any particular year, especially since productivity growth is markedly uneven. When this happens the upward pressure on prices from
the cost side is increased everywhere, although average prices will
not rise commensurately or at all unless demand permits.

199

ECONOMIC STABILITY AND GROWTH
Chart 22
Indexes of Personal Income, Retail, Wholesale, and Manufacturers* Sales
and Inventories and Manufacturers' New Orders
Seasonally Adjusted, Monthly, 1954-1958
Index
Aug. 1954 -100

Index
Aug. 1954 = 100 ,
125
Wholesale Sales
(Left Scale)

115 -

105 -

140

Manufacturers* Sales
110

(Right Scale)

100
120

Wholesale Inventories /

,'Manufacturers
Inventories

110

Retail Inventories
(Left Scale)

Ao
1954

1955

1956

1957

1958

Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
National Bureau of Economic Research, except for the peak of August 1957 which
was dated by the author.
Source: Department of Commerce.

200

ECONOMIC STABILITY AND GROWTH
Chart 23
Inventory-Sales Ratios
(Monthly, 1954-1958)

Percent
170

160

150

Retail

140
i

120

110

Wholesale

100
0

195

185

175

165
0

V
>

Manufacturing

1954

1955

1956

1957

1958

Source: Department of Commerce.
Vertical lines indicate business cycle peaks (P) and troughs (T) as dated by the
National Bureau of Economic Research, except for the peak of August 1957 which
was dated by the author.

ECONOMIC STABILITY AND GROWTH

201

The pace of business expansion slowed notably during the first 6
or 8 months of 1956. Industrial production actually drifted downward, whereas more comprehensive measures of physical activity such
as real gross national product merely flattened out. Industrial wholesale prices rose at a much diminished rate, although farm and food
prices rebounded from the 1955 lows and the average of all wholesale prices moved ahead faster than in the preceding year (charts
land 2).
The slowdown occurred because demand pressures eased generally.
In two major sectors of the economy final expenditures declined by
substantial amounts: consumer purchases of automobiles, which had
turned down after the third quarter of 1955, and residential construction, which had reached its peak slightly earlier (charts 18 and 19).
Retail sales of nondurable goods did not actually decline during the
early months of 1956, but neither did they increase much until May.
Hence, automobile dealers took steps to reduce inventories and other
retailers to avoid increases, with the ultimate result that inventory
investment declined in sympathy with stable or falling retail sales,
and production fell off among durable and nondurable factories alike
(charts 22 and 23). About the only major component of private
demand which was still rising strongly was business fixed investment.
The structure of industrial prices mirrors the pattern of relative
demands just described. Prices of finished producer goods continued
to rise during the first half of 1956, although at a slower rate than
during the preceding or subsequent spates of wage-price increase.
Average wholesale prices of consumer durable goods were stable as
retail prices of automobiles declined moderately. Steady also was
the average of consumer finished nondurable goods other than foods
in primary markets (chart 16).
Granted that easier demands were instrumental in the retardation
of price increases during this period, the next question is obvious. The
slowdown of demand was in response to events which had occurred
during 1955. In particular, the decline of homebuilding and automobile sales during the closing months of 1955 caused production to
drop not only in those sectors but in supplying industries as well.
About the turn of the year the declines became substantial enough to
depress hours and employment of factory workers and slow the growth
of aggregate personal income. Increased personal taxes restrained
the increase of disposable personal income still further. Nonautomotive consumption expenditures continued to increase, but at a
diminished rate and at the expense of a considerable reduction of
personal saving. Apparently the retardation of disposable income
contributed to easier demand conditions not only in automobiles but
in consumer goods generally during the first 3 or 4 months of 1956.
Taking the analysis back another step, the decline of homebuilding
after mid-1955 was caused primarily by measures employed at that
time to stiffen terms on mortgages insured or guaranteed by the Federal Government, and by the concomitant restrictive actions of the
Federal Reserve authorities which increased interest rates generally
and made residential mortgages less attractive to lenders. Automobile
sales, on the other hand, were probably not materially influenced by
general credit constraints. The spectacular sales of 1955 models had
satisfied the needs of many new-car buyers for at least 2 or 3 years

202

ECONOMIC STABILITY AND GROWTH

ahead, producing a temporary saturation of segments of the new-car
market and a corresponding decline of sales of 1956 models. I conclude that the easing of demand in 1956 was partly a lagged response
to credit policies which had accomplished the intended purpose and
partly the result of an independent reaction from the prior surge of
automobile sales.
The downward movement in automobiles and housing was prolonged into the summer months, although nonresidential construction
offset much of the latter decline. Eetail sales moved ahead during
the late spring and early summer, but the increase reflected primarily
higher prices for foods, and the physical volume of retail sales continued below the level reached late in 1955. Consequently, production
of durables and nondurables alike remained depressed until July, and
in that month a sharp though brief curtailment of steel production
occurred because of a strike.
Industrial production and prices surged forward on a broad front
beginning in August. The prime mover appears to have been increased activity in the automobile and ancillary industries as the new
models went into production. Federal defense expenditure turned
upward as well, however, and an added thrust was lent by restocking
demand on the part of metal fabricators whose inventories of materials had been drawn down during the steel strike.
The entire fourth quarter was colored by the rebound of automobile
production. As it turned out, sales increased less than expected, but
in the meantime production was augmented by the inventory accumulations of automotive manufacturers and dealers. The increase of
automobile production was superimposed, moreover, on the gentler but
steadier rise of machinery production as business expenditures for
equipment continued their advance. Since total industrial production advanced, so also did employment and earnings, the latter with
an assist from wage increases. In this way, increased production
consequent on expanded automobile output served to bolster personal
incomes and helped to induce a part of the increased volume of retail
sales which had been expected and the expectation of which had started
the process.
Demand, expected demand, and independent or semi-independent
cost increases are again intertwined in the price-output movements of
this period. Again there is a close correspondence between the increases of average hourly earnings and prices of intermediate ma^
terials and finished products in the durable goods sector (chart 20).
The increases were timed with the wage-price boost which followed
the steel settlement. With wages and prices rising all along the line
in metal fabricating, and with business demand for equipment still
high, prices of finished producer goods rose 3.4 percent between July
and October. Increased price tags on new model automobiles were
primarily responsible for the abrupt 2.1 percent rise of average wholesale prices of consumer durable goods between September and November. Prices of nondurables, excluding foods, also increased after
August, and by year end had risen 1.1 percent. These increases in
rimary markets were reflected in retail prices, and the Consumer
'rice Index gained 1.0 percent between August and December despite
a slight decline in prices of food over those months (charts 1 and 3).
As 1957 opened the expansion was already more than 2 years old.
I t gradually lost its momentum as the year progressed, and by autumn

ECONOMIC STABILITY AND GROWTH

203

a downturn was underway. Interestingly enough, production, employment, and prices behaved much as they had done in 1956 until the
period of actual contraction came during the last 4 months of the year
(chart 1). For that matter, the contrast after August is found primarily in physical activity, because prices of finished goods rose after
the contraction had begun and by amounts that while smaller than in
1956, were still significant (charts 20 and 21).
The early resemblance to 1956 is partly due to a repetition of the
same deflationary forces that developed at the outset of the former
year. Automobile sales increased moderately in the first quarter but
fell thereafter. Eesidential construction continued downhill until
July.
These declines were smaller than in 1956, but a new and ominous
depressant was added when business fixed investment tapered off
(chart 19). Money outlays for producers' durable equipment diminished fractionally after the first quarter and, since prices kept rising,
the decline was somewhat larger in physical terms. Signs of the
imminent slackening of fixed investment had appeared a few months
earlier, of course, when new and unfilled orders of producers of
investment goods entered a decline.
Thus, the year had scarcely begun before three important elements
of final demand—business investment, residential construction, and
automobile purchases—were exerting concerted downward pressure on
production. Government purchases and net exports rose, however,
with especially favorable results for the aircraft and petroleum industries—the latter because of the foreign demand stimulated by the
Suez crisis. Nonetheless, the balance was on the side of contraction in
durable goods. Employment, hours, and weekly earnings of factory
workers in the affected industries declined along with production, but
gains among salaried workers and elsewhere in the economy kept
personal income rising until August. Purchases of consumer durables
failed to rise along with personal income, but retail sales of foods and
soft goods moved ahead strongly during the spring and summer as consumers augmented their total expenditure and shifted its composition
in favor of nondurables. Since distributors were watching inventories
carefully, however, the new business of nondurable manufacturers did
not rise correspondingly and the volume of production was merely
maintained instead of increasing.
The incipient contraction became actuality after August. What
happened was that an economy already balanced on the edge of contraction because of internal developments in the private sector, was
subjected to additional, autonomous deflationary shocks from reductions of defense spending and export demands. Neither decline was
especially large, but they occurred when the economy was vulnerable
to disturbance and hastened a downturn that would doubtless have
come soon anyway. Now industrial production dropped more swiftly,
nonagricultural employment began to decline, and labor income
turned downward and carried aggregate personal income with it.
Retail sales decreased as personal income dropped, spreading the
contraction to the soft-goods industries.
The shifting fortunes of the economy and its several parts during
1957 left their mark on prices as well as production. Since demand
had eased generally after the turn of the year, industrial wholesale
prices were stable on the average until summer arrived (chart 1).

204

ECONOMIC STABILITY AND GROWTH

Even prices of finished producer goods rose comparatively little
(chart 16). It is of course true that these prices tend regularly to
rise more slowly after the annual wage-price adjustment is completed
near the end of 1 year and before it begins in the next, but the diminished intensity of demand for producers' durable equipment is reflected in the small rate of price increase in comparison with the first
half of 1956. Among important classes of finished goods, only foods
rose appreciably in price during the spring.
It is apparent from the indicators of investment preparations—
new orders, investment plans, and the like—that demand for investment goods stopped rising in 1957. It is quite another and more
complicated issue, however, as to why this came about. Financial
constraints were doubtless partly responsible. The Federal Eeserve
kept pressure on bank reserves through most of the year, and bank
loans increased only half as much as in 1956. Since the demand for
loanable funds exceeded the supply from banks and other sources,
interest rates rose through October. Overall credit tightness probably limited investments in certain directions, especially when it came
time to finance projects already planned. And, of course, financial
stringency during 1956 may have affected 1957 investment. Perhaps
the early decline of plant and equipment expenditure in the commercial sector reflects restricted access to credit sources on the part of
small business. The continued decline of homebuilding until midyear was certainly owing in part to difficulties in obtaining mortgages on favorable terms.
Another factor was at work to retard business investment at least
in the manufacturing sector, however, and that was the sharp expansion of capacity implied in the high rate of current investment outlays during 1955 and 1956. Capacity utilization rates were still satisfactory at the end of 1956, but the growth of production had slowed
in 1956 and could be expected to be gradual in 1957 since the economy
was already at full employment. Merely maintaining current capital
installations at the level already attained would insure future increases of capacity at a steady, high rate. Such considerations as
these probably damped the planned increases in the rate of current
investment, and by so doing helped to make capacity more redundant
than otherwise, for the slowdown in the rate of investment implied a
cutback of orders for producer goods and worked in combination with
weakness in automobiles to reduce production throughout the metal
and metal using sector of the economy. By September 1957, McGrawHill could report: "For the first time since we began asking the question in 1955 every manufacturing industry is now operating at a lower
rate of capacity than it generally prefers," although noting also
that no industries reported really depressed levels of business at that
time.6
The industrial price stability which had been a feature of the first
half of the year was disrupted by new increases for durable goods
during the summer and fall. Prices of producers' equipment accelerated after June, whereas the index for consumer durable goods took
its usual vertical step later in the year when the 1958 automobiles were
introduced at higher prices. The price raises were occasioned, of
course, by wage increases. This joint movement is reminiscent of 1948,
•McGraw-Hill's Fall Survey of Preliminary Business Plans for Capital Spending in
1958-59, prepared by McGraw-Hill department of economics, p. 3.

ECONOMIC STABILITY AND GROWTH

205

when wages and prices advanced together in the durable goods sector
just prior to another economic downturn. Actually, the timing itself
is not particularly relevant, since demand for investment goods had
eased but was still high before both downturns and markets could
hardly be described as distressed. The principal difference between
the two experiences is that in 1948 management resisted wage increases
until expectations were altered by the tax cut and defense program,
whereas in 1957 the increases were largely an automatic result of bargaining decisions made in 1955 and 1956. Thus about two-thirds of
the 7.5 million workers who received increases under major collective
bargaining contracts in 1957, did so because of deferred wage increases
under long-term agreements.
The present contraction has been underway for about 6 months as
this is written. Thus far the prices indexes—available now through
February—are still rising. Retail food prices are moving irregularly
upward as they have done since early 1956, primarily because of smaller
supplies of livestock. Consumer services continue to rise gradually
but steadily in price. Apparel has dropped a bit and retail prices
of new automobiles were shaved after November but remain several
percent higher than before the model changes. In primary markets,
durable finished goods have held their gains of late 1957, but average
prices of nondurables excluding foods dropped in February and may
have entered a cyclical decline in that month.
SOME REFLECTIONS ON THE INFLATIONARY PROCESS I N THE POSTWAR
AMERICAN ECONOMY

The reader must have been struck by the similarity of behavior of
durable goods prices during the past 3 years and in 1947-48. (During
the Korean war, of course, prices were under control.) In both periods
these prices rose swiftly, and in both the advance accelerated in those
months of the year when major wage increases went into effect, or
shortly thereafter in 1947.
This timing relationship is certainly not surprising. Major wage
increases are likely to raise unit labor cost in the short run, although
productivity increases may offset part of the effect then or later.
Prices of many durable goods are administered, moreover. Where
industries are sufficiently concentrated so that individual firms recognize mutual interdependence, but there is no collusion, it is advantageous to all if prices are changed infrequently and in response to clearly
identifiable factors known by each firm to affect the others in about
the same way. When the individual firms are also large and in the
public eye, a further advantage accrues if price increases can be
attributed to cost increases, even though profit margins may be maintained or augmented in the process.
The magnitude of the increases cannot be explained simply by the
fact that prices are administered, however. Even administered prices
have to be set at some level—and at a level which is profitable to the
company. This means that product demand cannot be ignored; it
sets limits within which the firm must price if it is to attain profits
which are satisfactory or better. By the same token, it sets limits
within which costs must be held. These facts apparently are recognized by both labor and management, since the largest wage-price
increases occur during periods of rising or high demand for durable

206

ECONOMIC STABILITY AND GROWTH

goods. (Deferred wage increases written into long-term contracts
may invalidate this statement in 1958, at least with regard to wages.)
The high demand for durable goods during the past 3 years cannot
be dismissed from an explanation either of changes in relative prices
or of inflation of the price level. Given that demand, sizable wage
and price increases appeared feasible and satisfactory at expected
levels of production and employment for at least one, and in some
cases several years ahead. But the wage increases spread to other
industries
as well, although often with a lag of several months (chart
21).T Labor costs of production were thereby raised generally, but
prices did not rise by the same amount everywhere (chart 16). Why
not?
First, although wage rates rose about equally through the economy,
prices of materials did not. Many raw materials of agricultural origin
were in chronic excess supply, with the result that prices were stabilized at levels governed by Federal programs and the entire cost
structure of products fabricated from the materials or their substitutes
was anchored.
Second, relative demands differed among the several product classes.
Among durables, this shows up in the fact that prices of producer goods
rose more than those of consumer items, and within the latter group,,
those of automobiles more than household appliances (these actually
dropped at retail), despite the similarities in prices of materials and
labor. As between durables and nondurables, the comparative
strength of market demands is difficult to gage. The changes in wage
rates and unit labor costs—whether fringe benefits and the earnings of
nonproduction workers are included or excluded—have been about
the same in the two divisions since the end of Korea, but the changes
in prices of materials were not, and this may account for most of the
contrast in the amounts of average price increase. Largely because
many important industries in the division are unconcentrated and
prices not administered, average prices of nondurables are rather
sensitive to current market demands, so that prices do not always go
up when wages do, and vice versa. This distinction rests on characteristics of market structure, however, and not on the relative strength
of final demands among products.
I do not wish to imply by my emphasis on demand factors that there
are no significant differences between 1947-48 and 1950, on the one
hand and 1955-57 on the other. For one thing, money wages of factory
workers lagged the cost-of-living during the earlier intervals, whereas
they moved ahead of it in the last one until 1957. In the former inflations organized labor was reacting to previous price increases and
attempting to restore the real value of labor income, whereas in the
latter one it was seeking to augment real income. It is this last f actt
coupled with the observation that money wages have outrun belowaverage productivity increments and raised unit labor costs, that has
led many observers to speak of cost-push inflation during 1956-57.
There are two kinds of inflation, it is asserted: demand-pull and costpush—and it is the latter which we have recently experienced. This
is misplaced emphasis.
7
Average hourly earnings rose as much or more in most of the nonmanufacturing indus~
tries for which data are available, including building construction, railroads, and trade.

ECONOMIC STABILITY AND GROWTH

207

There are not so much two kinds of inflation as two degrees of inflation. The earlier inflationary episodes of the postwar period were
more vigorous, and more widely diffused among the various sectors of
the economy than the last. Prices tended everywhere to move upward in common surges. One reason for this, of course, was that
market demands generally were more intensive and less easily discouraged by price increases than latterly, and this in turn was due to
the powerful inflationary potential supplied by deferred demands
and postwar liquidity in 1946-48 and by generalized, war-inspired
expectations of physical shortages and price advances during 1950.
This is not to say, however, that cost pressures and, more specifically,
wage increases were unimportant in 1947-48, for they did affect the
timing and size of the price waves in those years. They did so not
only in the durable-goods industries, and not alone because of the fact
that they raised unit labor costs, but also because key wage-price settlements affected price expectations and hence product demands
generally.
Demand pressures were clearly less intense in 1956-57, but that
fact does not render them unimportant in the gentler inflation of those
years. If one insists on a distinguishing categorization for this inflation, "bottleneck" may be more suitable than "cost-push." The
former term at least carries the connotation of increased demand as
well as increased cost in the sectors where prices rise strongly. It also
has the virtue of emphasizing the fact that inflationary pressures may
originate in particular sectors and spread to others, rather than appearing simultaneously everywhere.
That, in fact, is what one must expect under normal peacetime conditions. Widely diffused, powerful surges of excess demand are
easily recognizable precisely because they are abnormal. Such abnormal conditions aside, inflationary forces will tend to fan out from
initial areas of disturbance.
Demand may figure in two ways in the process. First, specific demands may foster individual price and wage advances and serve as
inflation starters. Secondly, aggregate money demand will have to
rise if prices are to increase in other sectors. If the postwar American
economy does indeed have an inflationary bias—and I think that it
does—it is because its institutional framework favors the initiation
and propagation of inflationary impulses, and guards against their
liquidation.
With regard to the initiation of inflationary impulses, there is the
fact that organized labor groups will press for money wage increases
during periods of business expansion, since this is the variable affecting real income over which individual unions have some degree of
direct control. Their success in winning wage increases will depend
in part upon management estimates of the extent to which wage increases may profitably be passed on in product prices. This will mean
that wage levels will tend to be determined by the increases which occur in the industries whose profit prospects are most favorable, and it
is at this point that high demands for specific products become crucial
in helping to set standards for wage increases. The standard-setting
wage increases may or may not exceed the long-term average rate of
overall productivity increase, but they are quite likely to do so for any
given year and especially during years of full employment expansion.

208

ECONOMIC STABILITY AND GROWTH

Apart from acute inflationary disturbances like price decontrol or
Korea, a problem of adjustment is posed for the economy each time
wages and prices go up in key industries. Confronted with the fact
that over much of our basic industry wages and prices are determined
at discrete intervals and set for a year or more ahead, the question
is whether aggregate money demand will rise sufficiently in response
to the specific increases to sustain a higher level of prices and money
incomes. This question, be it noted, is the same no matter what the
causes of the specific increases themselves: whether, for example, they
are heavily influenced by expected demand as in 1948, or are the lagged
result of bargaining decisions made 1 or more years previously, as in
1957.
The additional money demand will be readily forthcoming when real
demands are strong and financial constraints weak, as in 1947. Under
those circumstances, increases of current money demand—including
speculative inventory demand—will be large enough to raise prices
in all markets and hence wage and nonwage incomes in all industries
about equally. When current (as contrasted with expected) real
demands are weakening, however, as in 1948 or 1957, prices in other
sectors, and hence nonlabor incomes per unit of output, will not rise
correspondingly. And, of course, sales may become depressed everywhere, so that total wage earnings, profits, and other variable incomes
may decline even in those industries where prices went up substantially.
The inflationary bias of the economy is apparent also when it comes
to this question of the propagation of inflationary impulses. Our
money supply is managed, and it is managed with regard to domestic
stabilization objectives. This means that monetary controls will be
used to curb an expansion of money expenditure under full employment conditions, and this whether aggregate demand is surging forward on a wide frontier because of powerful autonomous forces, or
rising unevenly in response to the gentler prodding of demand or cost
increases in specific sectors. It means also, however, that monetary
(or fiscal) curbs will tend to err on the side of too little restraint, since
the goal is not stable prices at any cost, but stable prices accompanied
by full employment and economic growth.
A distinction is sometimes drawn between demand inflation and
cost inflation on the grounds that the former can be stopped at a
given level of real income by eliminating excess demand, whereas
even if that is done, autonomous cost increases will renew the latter
type of inflation and force either a relaxation of the demand constraint or a reduction of output. I suspect that this contrast is more
a property of static equilibrium models than of the dynamic economy.
In the first place, one should remember that autonomous demand
shifts mi£;ht also disturb a stable equilibrium if that were achieved
through fiscal or monetary controls—and autonomous demand shifts
occur frequently in the real w^orld. Credit controls
would have to be
tight indeed to prevent a price advance fosfered by new autonomous
demands and financed by the activation of idle money balances. Secondly, financial constraints powerful enough to keep prices from
rising under demand pressures, would almost certainly prompt a
contraction of physical activity. If they did not lead directly to a
downturn, they would do so indirectly by retarding or stopping the

ECONOMIC STABILITY AND GROWTH

209

expansion of physical activity, with adverse consequences for real
inventory demand and perhaps for fixed investment as well. Such
considerations argue for that cautious application of inflationary controls which is observable in practice, no matter what the origin of
the inflationary pressures.
Finally, our institutions and policies guard against the liquidation
of inflationary pressures. Deflation brings not only lower prices but
unemployment and lost production, and these are adjudged the worser
evils. Instead of forcing credit deflation, the monetary authorities
pursue easy money policies during contractions. Expansionary fiscal
actions—increased Government spending and tax reductions—are
more likely than not. Automatic stabilizers cushion the drop of
income and demand. Agricultural supports slow or prevent price
declines, and administered prices are preserved by company policy.
General wage reductions are neither recommended nor anticipated.
In short, the preponderance of public and private economic forces
work directly, and in many instances deliberately, against price
reductions during business contractions.
Since my subject has been price behavior, I have written of inflationary bias in these paragraphs. It is readily to be seen, however,
that the bias is largely a byproduct of properties of the postwar economy which most persons would agree were desirable ones. This fact
should be kept in mind when judging the performance of the economy during these past years, and it should come to the forefront
whenever the benefits of alternative goals and the risks of alternative
policies are to be weighed.

CYCLICAL CHANGES IN LABOR COST
Thor Hultgren, National Bureau of Economic Research, Inc.
The cost of producing a commodity or service puts a floor under its
price. If cost rises enough to wipe out the margin of profit at a
former price, suppliers will have no motive for furnishing the product
to their customers at that price, and either the price will rise or the
suppliers will drop out of the market. At least over short periods,
however, the relation between cost and price is often very loose. If
cost rises moderately, business enterprises sometimes accept a smaller
profit margin, at least temporarily, instead of increasing their prices
m proportion. If cost falls, they sometimes collect a larger profit
margin instead of reducing their prices proportionately. The study
of cost is, nevertheless, an essential component in the analysis of prices.
At the National Bureau of Economic Research we have been trying
to learn something about the relation between cost and the level of
production in various industries, and also between cost and the general
level of business activity. Costs may conveniently be divided into
three classes: materials, labor, and general overhead. Labor cost is
the kind on which there is most information and most of our findings
pertain to it. These findings have not jet passed through the National Bureau's review procedures, and I am presenting them here on
my own responsibility and not as official conclusions of the Bureau.
AVAILABLE INFORMATION ON LABOR COST

For the analysis of an industry's prices, the aggregate labor cost
incurred in the industry is of little significance except as a starting
point for calculations. The important figure is cost per unit of
product. The word "cost" will be used in this latter sense in this
paper. There are few if any industries for which such a figure is
published as a routine statistic. Consequently it is necessary to compute such figures. To make that computation we must have figures
on the aggregate labor cost as a point of departure. But we must
also have figures on the physical volume of production, otherwise we
cannot divide aggregate cost by product. Moreover, we need monthly
figures on both, since in this particular investigation we are interested
in the relations between cost and fluctuations in production or business
rather than in long-run trends, and monthly figures are usually far
more instructive in the study of such fluctuations. There are many
industries for which we cannot get the necessary information. Either
labor data, or production data, or both are lacking. We have, however, been able to assemble figures for anthracite and bituminous coal
mining, quite a few although by no means all branches of manufacturing, and railroads. The several industries and the periods covered
by our monthly data are shown in table 1. For railroads the figures
go as far back as the middle of 1921, for some important industries
23734—58

211

212

ECONOMIC STABILITY AND GROWTH

they begin in 1932, for others there is no usable information until
1947.
Labor cost per unit has two components: the number of hours paid
for per unit of product turned out, and the average amount of wages
paid per hour. It is instructive to consider hours per unit and labor
cost separately. We have therefore computed figures on man-hours
per unit of product as well as labor cost per unit of product.
In general we use Bureau of Labor statistics data on labor. For
many industries that Bureau publishes for a week near the middle of
each month an estimate of the number of production workers, the
average hours worked per week, and the average hourly earnings.
Multiplying the first by the second, we get the aggregate hours worked
during the middle week.
TABLE 1.—Periods covered by monthly data on man-hours and labor cost per
unit of product for various industries
Industry
Anthracite mining
_
Bituminous mining
Manufacturing:
Meatpacking
Confectionery
Cigars
Cotton
Wool
Textiles
Hosiery
Men's and boys' suits and coats.
Women's outerwear
__
Lumber
_
Millwork and plywood
Paper and pulp
Petroleum refining
Tires and tubes
Cement
_
Steel
Iron and steel foundries
Aluminum and copper mill shapes.
Nonferrous foundries
Railroads

1st month
January 1932
January 1934
January 1932
January 1939
do.
January 1932.
do
January 1947.
January 1934January 1947.
do
do...
do
January 1932..
January 1947.,
January 1932..
January 1947..
January 1933.
January 1932-.
January 1947..
do..
do
July 1921

Last month
September 1955.
December 1955.
December 1949.
May 1949.
June 1955.
May 1949.
Do.
December 1955.
May 1949.
December 1955.
Do.
Do.
Do.
June 1955.
Do.
December 1940.
July 1955.
June 1955.
Do.
December 1955.
Do.
Do.
August 1955.

From this point onward, our procedure depends on what kind of a
measure of production we use. In some industries, for example anthracite, production can be measured tolerably well in terms of some
simple natural unit, in this instance tons of anthracite mined. In
such a case, we multiply the aggregate hours in the middle week by
4.33 to get an estimate of aggregate hours for the month, and divide
that aggregate by the total production. In other industries, the diversity of products is so great that we have to use a composite index
number as a measure. In that case, we construct an index number
of the aggregate hours and divide it by an index number of production.
For the latter, in general, we use one or another of the indexes computed by the Federal Reserve Board for various industries.
Where we use a production measure in natural units, we multiply
man-hours per unit by average hourly earnings to get labor cost per
unit. Where we use indexes, we multiply the index of man-hours per
unit by an index of average hourly earnings to get an index of labor
cost per unit.
In the case of railroads our labor as well as our production data
come from the Interstate Commerce Commission. We measure pro-

ECONOMIC STABILITY AND GROWTH

213

duction in terms of traffic units, which we define as revenue ton-miles
plus 2.4 times passenger-miles. The 2.4 factor is based on the average
relative revenue per ton-mile and per passenger-mile over a long
period of years.
Finally, we adjust both the man-hours per unit figures and the labor
cost per unit figures for seasonal variation. If ,man-hours per unit
in a particular industry average higher in December than in the year
as a whole, for example, we reduce the figures for the various
Decembers.
IDENTIFYING CYCLES IN PRODUCTION

We likewise adjust the figures on production in each industry for
seasonal variation. This enables us to determine what cycles, if any,
have occurred in its production. By a cycle we mean merely an upswing followed by a downswing, or vice versa. The process can be
illustrated with the help of chart 1, which depicts production in one
of our industries, aluminum and copper mill shapes, for 1947 to 1955.
As often happens, there are numerous minor month-to-month ups
and downs in the curve. We disregard these, but try to pick out the
larger swings. A helpful rule in this connection, and one that we
follow, is that nothing shall be recognized as a cycle unless the up and
down swing taken together last for at least 15 months. On chart 1,
we find that production of mill shapes reached a peak in July 1948.
After that there was a downswing, reaching a trough in May 1949.
This was followed by an upswing, which reached a peak in August
1950, and so on. Asterisks indicate the dates of peaks and troughs.
CHABT 1.

ALUMINUM AND COPPER SHAPES : FEDERAL RESERVE INDEX OF PRODUCTION, JANUARY 1947-DECEMBER

1955

Index (19*»7-*»9«1OO)

19*»7

19*8 19*9

1950 1951 1952

1953 195*»

1955

214

ECONOMIC STABILITY AND GROWTH

LABOR PER UNIT OF PRODUCT OFTEN VARIES INVERSELY WITH PRODUCTION

We are now in a position to learn what relation there is, if any, between man-hours per unit of product in this industry and the level of
its production. Our first question is, how does man-hours per unit
at each peak in production compare with man-hours per unit in the
following trough, or vice versa ? In answering such a question, however, we do not like to depend on the figure for a single month. We
therefore take an average of man-hours per unit in the month preceding each peak, the actual peak month, and the following month.
Likewise, we take an average for each group of 3 months that has a
trough month in the middle.
Mill shapes is one of the industries for which we use index numbers. Computing 3-month averages, we find that the index of manhours per unit (1947-49=100) was 91.3 in the vicinity of the 1948
production peak and 116.0 in the vicinity of the 1949 trough. We
conclude that in this contraction of production man-hours per unit
of product increased. At the 1950 peak the figure is 94.8. In the
1949-50 upswing, therefore, man-hours per unit of product declined.
So far, it looks as though man-hours per unit were inversely related to production in this industry. But we have data on only three
completed contractions and only two completed expansions. This
perhaps is not enough to justify a generalization about this particular industry. We do feel, however, that by pooling our data for
this and other industries we obtain something more significant. In
one industry or another, at one time or another, we have data on 69
expansions of production and 77 contractions. In 62, or 90 percent,
of the 69 expansions, we find a net decline in man-hours per unit.
In 58, or 75 percent, of the 77 contractions, there was a net rise in
man-hours per unit. The pooled data suggest a strong tendency
toward an inverse relation between man-hours per unit and total output. When one goes up the other goes down and vice versa. This
means that man-hours in the aggregate usually do not rise and do
not fall by as great a percentage as output.
These results, incidentally, show the superiority of monthly over
annual figures. For the same industries and the same periods of time
we have, of course, annual figures on production, and we have computed annual figures on man-hours per unit. When we base our
conclusions entirely on annual figures, we get a similar result in
expansions but not in contractions of output. Man-hours per unit
falls in 86 percent of the expansions, but also in 55 percent of the
contractions instead of rising in most of them. Note, however, that
the majority of declines is much smaller in contraction than in expansions. This reduced majority of declines masks an actual preponderance of rises that is disclosed by the monthly data.

ECONOMIC STABILITY AND GROWTH

215

This comparison enables us to extend the statistical basis for our
generalization about an inverse relation. There are an additional
197 expansions and 220 contractions in production for which we have
nothing but annual data. Man-hours j>er unit fell, according to the
annual data, in 84 percent of the expansions and, again, in 55 percent,
or a much smaller majority, of the contractions. It seems likely that
the small majority of declines in contractions here also masks a preponderance of rises that monthly data, if we had them for these industries and periods, would disclose.
Annual figures minimize or conceal the cyclical rises in man-hours
per unit because they understate the declines in production and consequently the frequently adverse influence of diminished volume,
while they do not similarly understate the effect of technological improvements, which operates to reduce man-hours per unit both in
expansions and in contractions. It is true that annual figures also
understate the influence of expansions in volume but in that case their
influence is reinforced rather than opposed by that of technology.
Although such figures obscure the inverse relation between input
per unit and production, there are some industries in which the relation is so strong that even annual data convincingly disclose it. We
hesitate to generalize about an industry unless we have data for eight
or more phases of production. (By a phase, we mean either an expansion or a contraction.) Among industries for which we have that
much data (in either annual or monthly form), we find 22 with a
strong inverse relation, 9 without any marked relation one way or
the other, and none with a strong positive relation; i. e., none in which
man-hours per unit rises and falls with volume.
LABOR COST ALSO VARIES INVERSELY WITH PRODUCTION, BUT RISES
MORE OFTEN

We can analyze the direction of change in labor cost per
unit of product in the same way that we analyze man-hours per
unit. We can average the labor cost figures for the 3 months in the
vicinity of each peak in an industry's output and likewise those for
the 3 months in the vicinity of each trough. We can then note the
direction of change during each expansion or contraction. For example, we find that the index of labor cost for mill shapes rose from
93.5 at the 1948 peak to 122.5 at the 1949 trough, then fell to 106.5 at
the 1950 peak. When this is done for the 69 expansions, we find that
labor cost declined from the trough of production to its peak in 41 of
them, or 59 percent. There was a net rise from the peak to the
trough in 70 of the 77 contractions, or 91 percent. Growing production was accompanied in most instances by falling labor cost, and
diminishing production by rising labor cost. Like man-hours per
unit, labor cost was inversely related to volume.

216

ECONOMIC STABILITY AND GROWTH

But rises in labor cost were more common than rises in man-hours
per unit in both kinds of phases. In expansions, the percentage with
rises is only 10 for man-hours per unit, but 41 for labor cost. In contractions the percentage with rising man-hours per unit is 75, the percentage with rising labor cost is 91. The reason for the difference is
to be found in average hourly earnings. In expansions these have
usually risen, and the rise has not infrequently been large enough to
more than offset the fall, if any, in man-hours per unit. In the contractions for which we have monthly data, average hourly earnings
often rose, although not as fast as in expansions. In some cases these
rises were enough to offset the decline in man-hours per unit that did
occur in a minority of contractions.
The greater frequency of rises in cost than in hours per unit may
not have been characteristic of earlier contractions for which we have
no monthly data. During contractions of production in the neighborhood of 1920-21 or 1929-32, we know that average hourly earnings
fell in many industries, and in these periods falls in labor cost may
have been more common than in more recent periods.
INVERSE RELATION MORE PRONOUNCED AT THE BEGINNING THAN AT THE
END OF SWINGS I N PRODUCTION

So far we have confined our attention to the net change in hours per
unit of labor cost between the beginning and end of an upswing,
or the beginning and end of a downswing in production. But their
direction of change is often not continuous during the course of such
an upswing or downswing. Chart 2, again for the mill shapes industry, illustrates how the course may alter. The asterisks on this
chart no longer represent turning points in production itself; the
latter are indicated by broken vertical lines for peaks and solid vertical
lines for troughs. The asterisks here indicate peaks and troughs, and
mark off cycles, in man-hours per unit. Obviously the direction of
change in man-hours per unit is not always consistent throughout a
phase of production. During the 1950-51 contraction in production,
hours per unit at first rose, afterward fell. (As before, we ignore
minor fluctuations and take note only of the longer movements marked
by the asterisks.) The decline continued all through the following
expansion. On the other hand, both the beginning and the end of the
]ast completed upswing in hours per unit coincide neatly with the beginning and the end of the last complete downswing in output.

ECONOMIC STABILITY AND GROWTH

217

CHART 2. A L U M I N U M AND COPPER S H A P E S : INDEX OP MAN-HOURS P A I D FOR PER
U N I T OF OUTPUT, JANUARY 1947 TO DECEMBER 1955

Index (19V7-U9-1OO)
130

1955
Broken vertical lines indicate peaks, solid vertical lines troughs in production of aluminum and copper shapes.

We have graphs like chart 2 for all our industries. They enable us
to classify every production phase according to the sequence of change
in man-hours per unit. In the present instance we classify the 1950-51
contraction in output as a case of rise, fall, the 1951-52 expansion as
a case of (continuous) fall, and the 1952-53 contraction as one of
(continuous) rise. We have classified all our production phases in
this manner (table 2).
Man-hours per unit declined during the earliest months in a large
majority of the output expansions, 58 of 69. In 23 of the 58 instances,
however, the initial fall turned into a rise that persisted to the end
of the expansion. Still other sequences occurred, but just before the
end declines were far less common than at the beginning. Even so
they somewhat outnumbered rises, by 36 to 33.

218

ECONOMIC STABILITY AND GROWTH

Most contractions, 45 of 77, began with hours per unit rising. Of
the rises, 24 were consistently maintained, 20 turned into falls that
persisted. Continuous declines also were fairly common. Whatever
the sequence, hours per unit were falling at the end in a majority
of instances, about as large as the majority of rises in the beginning.
In labor cost there was a similar contrast between the earliest and
latest months of production phases. Declines were in a majority
at the beginning, rises in a heavier majority at the end of expansions.
Although rises preponderated at both the beginning and at the end of
contractions, the preponderance was much heavier at the beginning.
In both kinds of phases, however, and both at the beginning and the
end, rises in cost were more frequent than rises in hours per unit.
TABLE 2.—Expansions and contractions of production in various industries,
classified according to sequence of change in man-hours per uint or in labor
cost
Number of expansions or contractions with indicated
sequence of change
Sequence of change

In labor cost per unit

In man-hours per unit

Rise
Rise, fall.
...
Rise, fall, rise
Rise, fall, rise, fall
Fall.. . .
.
Fall, rise
Fall, rise, fall
_ .
Fall, rise, fall, rise
Fall, rise, fall, rise, fall

Expansions

Contractions

Expansions

8
3

24
20

18
3
2

29
23
3
2
1

1
23
7
2

12
31
1
2

Contractions
38
26
1
1
8
3

All sequences

Rising at beginning _ _
Falling at beginning
Rising at end
Falling at end

11
58
33
36

45
32
31
46

23
46
53
16

66
11
42
35

The contrary movement of production on the one hand, and hours
per unit and labor cost on the other, was therefore much more conspicuous at the beginning than at the end of upswings or downswings
in production.
Changes can be studied in a more systematic and progressive manner by dividing each upswing or downswing into a standard number
of stages regardless of its length. We define the initial, previously
described, 3-month trough period of an expansion as stage I, the
next peak period as V, and the terminal trough period of a contraction as stage IX. We divide the months between the actual
trough month and the next peak and likewise those between each
peak and the next trough into three periods, equally long if possible;
if necessary we make the middle stage a month longer or shorter than
the others. This procedure gives us stages II, I I I , and IV of an expansion or VI, VII, and V I I I of a contraction. We strike an average of, say, man-hours per unit for all the groups included in a stage.
In effect, we break each expansion into four successive segments, the
first running from I to II, the second from I I to I I I , and so forth;
and we likewise break each contraction into four segments. The
method is illustrated for one cycle of cement production in table 3.

219

ECONOMIC STABILITY AND GROWTH

The figures for man-hours per barrel on each line of the last column
is an average for the months indicated on the same line in preceding
columns. (This is one of the industries in which we use "natural"
physical units rather than index numbers.)
Sometimes an expansion is too short for this procedure; the first and
last stage would include almost all the months. In such cases, we restrict stages I and V or V and I X to a single month, the actual peak
or trough month. Even so, 1 of our 69 expansions and 2 of our
77 contractions are too short for the application of this method.
3.—Division of 1988-H cycle in production of cement into stages (initial
trough in production, February 1988; peak, April 191$; terminal trough, May
19U)

TABLE

Months included
1st

I
III
IV
V
VI .
VII.._
VIII
IX

. .

January 1938
March 1938 —
July 1939
December 1940
_ March 1942
May 1942
January 1943
September 1943.
April 1944

Last
March 1938
June 1939
November 1940.
March 1942
May 1942
__
December 1942__
August 1943
April 1944
June 1944

Number of Man-hours
per barrel
months
3
16
17
16
3
8

00 00 00

Stage

0.447
.396
.375
.363
.365
.370
.427
.526
.538

In the remaining 68 expansions and 75 contractions, we can now find
out how frequent rises and falls in hours per unit were in the various
segments. In table 3, for example, man-hours per barrel was 0.447 in
I but only 0.396 in I I ; we therefore count the 1938-42 expansion in
cement as one of the expansions in which hours per unit fell from I to
II. All in all, this happened in 60 of our 68 expansions, or 88 percent
of them. Similar percentages for all pairs of stages appear in table 4.
We have computed stage averages and percentages for labor cost in
exactly the same way; the percentages also are shown in the table.
Declines in hours per unit became less and less frequent in expansions. The percentage fell from 88 percent in the first segment to 85
in the second, 74 in the third, and 66 in the fourth. In contractions,
the characteristic initial rises likewise became less frequent; the successive percentages are 72, 64, 60, 49. In the case of hours per unit,
however, this does not mean that the characteristic change for the
phase as a whole was actually reversed in the last segment. True, the
percentage of rises is a trifle below 50; but the percentage of declines
is only 45; in addition there were a few instances of no change.
Declines in labor cost likewise became less and less frequent from
segment to segment of expansions. In fact, rises outnumbered declines in the third, and even more strongly in the fourth segment. In
contractions, the progression was less regular, as the percentage of
rises did not turn down until the third segment; but declines were
considerably more common in the fourth than in the first.

220

ECONOMIC STABILITY AND GROWTH

TABLE 4.—Man-hours and labor cost per unit of product frequency of rises and
falls from stage to stage of production cycles
Percentage of expansions and contractions in production
in which—
From stage

To stage

Man-hours per unitRose

III .

Ill
IV

IV
V

VII

V .
VII .
VIII._

VIII..
IX

Did not
change
9

25
34
72
64
60
49

3
2
1
0
0

Cost per u n i t -

Fell

Rose
88
85
74
66
28
36
37
45

Did not
change

16
38
54
71
79
84
81
63

0
0
0
1
0
1
2
1

Fell
84
62
46
28
21
15
17
36

PRODUCTION CYCLES VERSUS BUSINESS CYCLES

Everyone recognizes nowadays that there are broad upswings and
downswings, that is, cycles, in economic activity at large. To go
back no farther, there was a general contraction in 1953-54, followed
by an upswing that now appears to have ended in 1957. At the
moment of writing, we are in another downswing, although just how
far along in it we do not yet know.
But production in a single industry often follows a different course
from that of general business activity. A particular kind of output
frequently declines during at least part of a business expansion and
rises during at least part of a business contraction. From some points
of view, it may be more interesting to compare the changes in hours
per unit or labor cost between turning points in aggregate business
rather than between turning points in each industry's own production. The National Bureau has worked out a chronology of turning
points, and therefore cycles, in business at large, which we may use
for this purpose. This means that from here on we no longer compare man-hours per barrel of cement around February 1938 with manhours per barrel around April 1942 (see table 3); instead, for example,
we compare man-hours per barrel around May 1938 with man-hours
per barrel around February1 1945; these dates are the trough and
peak of a business expansion.
We have hours per unit and labor cost data for these 2 dates in
10 industries, giving us 10 observations in that business phase. In
one business expansion or another and one industry or another, we have
51 observations altogether. We have 61 observations for business
contractions. The expansions represented are 1933-37, 1938-45,
1945-48, and 1949-53. The contractions are 1937-38,1945-45,1948-49,
and 1953-54.
FALLING MAN-HOURS PER UNIT MORE COMMON I N BUSINESS EXPANSIONS
THAN I N BUSINESS CONTRACTIONS

In 39 of the 51 observations for business expansions, or 76 percent,
there was a net decline in hours per unit from the business trough
1
The 1938 date has been revised to June; but to keep our work consistent with other
studies we use the old date. The effect on our conclusions is negligible.

ECONOMIC STABILITY AND GROWTH

221

to the business peak. In this respect, business expansions resembled
production expansions. But in 41 percent of the 61 observations for
business contractions or 67 percent, we also find a net decline in hours
per unit. In production contractions, on the contrary, there was a
majority of rises. The only distinction we can make between business
expansions and business contractions is that declines in hours per
unit are even more frequent in the former than in the latter.
The explanation of the difference between downswings in business
and downswings in production is similar to one we have already
discussed in other connection. When an industry's contractions fail
to coincide with those in business, declines in production measured
between business dates are smaller percentagewise than declines measured between production dates. Any adverse effect of diminishing
volume is correspondingly minimized, while the contrary effect of
any technological improvements is not correspondingly minimized.
Consequently, an industry may have a less efficient labor-to-product
ratio at the trough than at the preceding peak in its production, and
still have a more efficient ratio at a business trough than at the preceding business peak.
LABOR COST RISES AND FALLS WITH BUSINESS

In the case of labor cost there is a much more striking difference
than in the case of hours per unit. In 41, or 80 percent, of our 51
observations for business upswings, labor cost increased. In 35, or
57 percent, of our 61 observations for business downswings, labor
cost diminished. In other words, cost appears to be positively related
to the level of business activity, whereas it is inversely related to
the level of production in the several industries.
In the case of upswings, the explanation is connected with the fact
that the production expansions were on the average a good deal shorter
than the business expansions. At the beginning of the 1933-37 business expansion, the recovery from the banking holiday, and the imminence of the blue eagle and NBA, seem to have stimulated a vigorous
wave of buying and production. For many industries the rise proved
to be temporary, although a good number had a second upswing after
an intervening decline in production. In the 1938-45 business upswing, the construction phase of the war effort reached a comparatively
early climax. This meant an early downswing for industries heavily
dependent on construction. Manpower shortages apparently caused
such a decline in others. In the 1949-53 business expansion, the
Korean crisis caused a tremendous upswing in demand from consumers
and from business itself; but as inventories accumulated this receded
and presently restrictions on materials also limited production in some
industries. Consequently there were a number of contractions in
production, more or less in the middle of the business upswing, often
followed by renewed rises in output during its later stretches. The
difference in length is important because the cumulative effect of rising
average hourly earnings on cost is greater over the longer periods; in
many cases it preponderated over whatever net rise in the scale of
production occurred over the same periods.
In business contractions, as already noted, technological progress
tends to predominate over diminishing volume, bringing about a net
fall in hours per unit, while average hourly earnings do not rise as

222

ECONOMIC STABILITY AND GROWTH

vigorously as in business expansions, if they rise at all. The net result
was the rather narrow majority of declines in labor cost.
FALLING MAN-HOURS PER UNIT MOST FREQUENT I N EARLY BUSINESS
EXPANSION AND LATE BUSINESS CONTRACTION

Business cycles, like those in production, can be divided into stages
on the plan previously described. The 3-month periods in the vicinity
of business troughs and peaks become stages I, V, or IX. Intermediate months in a business expansion are grouped into stages II, I I I ,
and IV, those in a business contraction into VI, VII, and VIII. We
can compute averages of hours per unit or cost for each of these for
any industries for which we have data covering a business phase. In
that phase the stages will refer to the same time periods for all industries. In each phase, we get a number of observations of change
between each pair of stages equal to the number of industries. In all
four of the business expansions covered, in one industry or another,
we get the same number of observations for each pair of stages as for
net change over the phase, namely 51. Similarly, we have 61 observations for each pair of contraction stages.
As in production expansions, declines in man-hours per unit are
very common in the first segment of business expansion and become
less numerous thereafter (table 5). There is a slight increase in frequency, however, in the last segment. In contractions of business, as
in production contractions, declines in hours per unit are comparatively infrequent in the first segment but become more and more frequent thereafter. In our data, they outnumber rises in the last three
segments.
DECLINES I N LABOR COST SOMEWHAT SIMILARLY DISTRIBUTED

The frequency of rises in labor cost does not increase at all consistently from segment to segment (table 5). They were most numerous
in the second segment, next most common in the fourth. The pattern
is largely influenced by the peculiar history of average hourly earnings in the 1933-37 expansion. The blue eagle and NRA brought
widespread increases in hourly pay; these tended to raise labor cost
between I I and I I I . Another round of sharp increases in wage rates
tended to boost cost between IV and V. Taking all the observations
for all four expansions together, however, the first segment was the
only one in which declines in cost predominated over rises. The net
declines in labor cost during business contractions were concentrated
in the last two segments; it was only in these that declines exceeded
increases in number.

223

ECONOMIC STABILITY AND GROWTH

TABLE 5.—Man-hours and labor cost per unit of product frequency of rises and
falls from stage to stage of business cycles
Percentage of observations in which—
From stage

To stage

Man-hours per u n i t Rose

III

IV
V
VI

vVI

III
VII
VIII

IV.._
VII .
VIII-.

22
33
39
37
54
48
30
30

Did not
change
2
0
0
0
2
1
4
0

Labor cost—

Fell

Rose
76
67
61
63
44
51
66
70

43
86
69
76
62
56
28
36

Did not
change
0
0
0
0
2
0
2
0

Fell
57
14
31
24
36
44
70
64

T E C H N O L O G I C A L PROGRESS M I N I M I Z E S C Y C L I C A L R I S E S I N COST

In the foregoing discussion a complicating influence, that of technological progress, has been mentioned incidentally. I t deserves a
more emphatic recognition. Technological improvements tend to reduce man-hours per unit not only in expansions but in contractions.
If it were not for such improvements, declines in hours per unit would
not be as frequent as they are in expansions of output, while rises
would be even more frequent than they are in contractions. Even in
an industry that clearly has an inverse relation between hours per unit
and output, technological improvements enable the industry to reduce
hours per unit in some contractions. Technical advances likewise
maximize the frequency of decline in cost and minimize the frequency
of rises in cost. The influence of technology on labor cost reenforces
the usual influence of expanding volume and, mostly, to be sure, at
other times, the influence of declining hourly earnings; it opposes the
usual influence of contracting volume and the influence of rising
hourly earnings.
MEANING AND LIMITATIONS OF THE COST DATA

If we were dealing in each industry with the production of one
homogeneous commodity, changes in total cost (labor, materials, and
overhead) per unit would necessarily be reflected either in proportionate changes in the average price received by the industry or in
altered margins of profit. With a single product, if total cost per
unit rises, and the price does not rise by as great a percentage, the
margin narrows; if price rises by a greater percentage, the margin
widens in spite of the rise in cost. If cost falls, and the price does not
fall in proportion, the margin widens; if price declines by an even
greater percentage, the margin narrows in spite of the fall in cost.
We have presented data for labor cost only. In fact they do not
include all of that, for administrative, general supervisory, and central
office labor is not included. Total cost no doubt often fluctuates by
different percentages than labor cost and may at times move in the
opposite direction.
Even if the data represented total cost, they would still not pertain
m any instance, except perhaps cement, to an industry with a vir-

224

ECONOMIC STABILITY AND GROWTH

tually homogeneous product. Where the product is complex, changes
in total cost, based on the measures of production we are obliged to
use, might understate or overstate the percentage by which prices
would have to be raised or lowered to preserve an initial margin of
profit. If the production of high quality coal rises faster than that
of all coal, for example, the mining companies, without changing the
price of any quality, will collect more revenue per ton of all coal, and
hence if cost per ton rises, prices will not have to rise in proportion,
and yet the original margin could be preserved. In more serious
cases we try to minimize the difficulty by using composite indexes of
output. But while those indexes are based in principle on value at
constant prices, they do not pretend to reflect the full complexity of
pricing. An extremely detailed measure of production, completely
reflecting that complexity, might rise or fall from time to time by a
greater percentage than the available index. Use of the latter would
at such times overstate or understate the size of the adjustment in
prices needed to preserve the initial margin.

TRENDS IN PRODUCT PRICES, FACTOR PRICES, AND
PRODUCTIVITY
John W. Kendrick, The George Washington University and National
Bureau of Economic Research, Inc.
This paper is based on estimates of product prices, factor prices,
and productivity in the private domestic economy for selected years
of high-level economic activity since 1919. The estimates make possible an analytic description of the trends and interrelationships of
these variables and of related variables such as the real income per
unit of factor input in comparison with productivity, and the changing factor shares of the national income as a result of divergent price
and quantity movements of labor and capital inputs.
The interrelationships described by the time series are based on
identities, and by themselves do not reveal the mechanism by which
prices change. They do provide part of the background required
for analysis of the mechanism, and some interpretation is given in the
paper, but this study is intended primarily to provide one of the
key pieces of the mosaic to be built up by all the contributions to
these hearings.
The estimates on which the analysis is based were prepared by the
author in connection with a study for the National Bureau of Economic Research of productivity trends in the United States since
the late 19th century. The sources of the estimates will be described
in detail in the National Bureau study; a very brief description is
appended to this paper. The bulk of the paper is drawn from a
preliminary draft of one of the chapters in the Bureau study, but
the
draft has
not
been
reviewed
byn*/vtheVXXXD
board
of directors
of the
I/IAO \JLM.a>±v
iiao
n u yet
t j
"-f^xi
IOVIOVVDU
KJ\J<XI.\JL
K/JL KXXI. CA;iuia
" assumes full
" responsibility for the
National Bureau, and"" the author
here.
estimates and his interpretation of their meaning presented he
C U l

THE TOTAL, FACTOR PRODUCTIVITY CONCEPT AND MEASURE

Since the measures of real product and of real factor inputs are
basic to the subsequent analysis, and since the concept of "total factor
productivity" differs from the more usual "output per man-hour" partial productivity measure, a few words as to underlying concepts and
methods are in order.
Total factor productivity is defined as the ratio of the physical
volume of final output to the physical volume of labor and capital
inputs, the units of the various types of output and of input weighted
together by their base period unit values. Labor input is measured
in terms of the man-hours worked in the various industry groupings,
weighted together by base period average hourly earnings. Capital
input is assumed to move proportionately with the constant dollar
value of the net stock of real capital available for use (land, inventories, and plant and equipment net of depreciation); indexes of real

225

226

ECONOMIC STABILITY AND GROWTH

net capital stock in various industry groupings are combined by baseperiod capital compensation in each. Total factor productivity differs from the conventional "output per man-hour" measures in two
respects: (1) Since man-hours employed in each of about 40 industry
groups are weighted by base-period average hourly labor compensation, shifts in relative man-hours employed from lower-to-higherpaying industry groups result in a greater relative increase in the
weighted "labor input" series than in unweighted man-hours, but do
not influence the aggregate productivity measure which is in effect
a weighted average of productivity indexes of the component industries; (2) since capital input is counted in the input denominator
along with labor input, increases in output per unit of labor input that
result from substitution of capital for labor do not show up as changes
in productive efficiency; only if final output increases in relation to
both corresponding factor inputs can we say that there has been a net1
saving of inputs or cost elements, and thus an increase in productivity.
To illustrate the difference in movement of the several conceptually
different productivity ratios, as well as to provide background for the
later analysis, consider the following average annual percentage rates
of change for the period 1919-53. This period is used since the trend
rate of increase in total factor productivity in the private domestic
economy has been significantly higher than it was in the decades preceding World War I. The average annual rate of increase in real
product per man-hour was approximately 2.5 percent, compared with
a 2.3 percent rate of increase in real product per unit of labor input
(weighted man-hours) ; the 0.2 percentage point difference represents
the effect on labor input of shifts from lower-paying to higher-paying
industries. The difference between the 2.3 and 2.1 percent average
annual rates of increase in real product per unit of labor input and in
total factor productivity, respectively, reflects the net effect of substitution of capital for labor over the period.
PRODUCTIVITY AND THE PRICES OF PRODUCTS AND FACTORS

The general level of prices of the final goods and services produced
by the economy must increase to a lesser extent than the average prices
of the productive factors (or decline more) in proportion to advances
in factor productivity. Increasing productivity means that the quantity of inputs used per unit of output declines, and this decline in unit
factor requirements provides an offset against rising factor prices in
like degree. Productivity gains thus act as a cushion whereby rising
factor prices are reflected to a lesser extent in final product prices; or if
factor prices decline, productivity advances are associated with a
proportionally greater decline in product prices.
The relationship among the three variables may be stated precisely
if the terms are appropriately defined. Factor price is the weighted
unit compensation of the factor services, obtained by dividing national income (which is the sum of factor cost or compensation) by
the sum of real labor and capital inputs (see table 1). The average
unit compensation of labor is the compensation per man-hour in the
various industries combined by variable man-hour weights. The
1
For further discussion of concept, see John W. Kendrick, Productivity Trends: Capital
and Labor, Occasional Paper No. 53 (National Bureau of Economic Research, 1956).

227

ECONOMIC STABILITY AND GROWTH

average unit compensation of capital is, in effect, the product of prices
of capital goods and the rate of return on capital (including profit)
in the several industries, combined by variable capital input weights.
Operationally, the average price of each factor class is obtained as a
quotient of the total compensation of each divided by the corresponding aggregate real input measure (see table 2a).
TABLE 1.—Private domestic economy—Factor prices, product prices, and
productivity
Net domestic product at factor cost
Current
dollars

1929
dollars

(1)

(2)

Factor
Factor producinput
tivity
1929
index
dollars *
))
(3)

(4)

Average price
Factors
index

Products
index

(5)

(6)

D+()

A. Millions of dollars and index numbers, 1929=100
60,848
82,669
66,433
203,191
268,946

1919.
1929.
1937.
1948.
1953.

56,711
82,669
84,157
135,991
168,562

70,207
82,669
73,720
92,827
102,739

80.8
100.0
114.2
146.5
164.1

86.7
100.0
90.1
218.9
261.8

107.3
100.0
78.9
149.4
159.5

115.3
90.1
243.0
119.6
302.0

93.2
78.9
187.4
106.8
148.6

B. Link relatives
135.9
80.4
305.9
132.4
442.0

1929/19.
1937/29.
1948/37.
1953/48.
1953/19.

145.8
101.8
161.6
124.0
297.2

117.8
89.2
125.9
110.7
146.3

123.8
114.2
128.3
112.0
203.1

O. Average annual rates of change (percentages)
3.8
.2
4.5
4.4
3.3

3.1
-2.7
10.7
5.6
4.5

1929/19.
1937/29.
1948/37.
1953/48.
1953/19.

2.2
1.7
2.3
2.3
2.1

1.7
-1.4
2.1
2.1
1.1

1.4
-1.3
8.4
3.6
3.3

-0.7
-2.9
6.0
1.3
1.2

i Factor input here is derived as the sum of labor and capital inputs in absolute terms. It differs
slightly from an index which represents a variable weighted average of indexes of labor and capital inputs.
NOTE.—Table,may not be internally consistent due to rounding, Indicated derivations apply to sec. C
only if 100 percent is added to the percentage rates of change.
TABLE 2a.—Private domestic economy—Average factor prices: Labor and capital
[Millions of dollars and index numbers, 1929=100]
Labor cost

1919
1929
1937
1948
1953

_
__

Current
dollars

1929
dollars

Average
price of
labor
index
(D + (2)

(1)

(2)

(3)

43,814
59, 749
52,400
154, 769
213,145

51,802
59, 749
52, 221
66,859
70, 743

84.6
100.0
100.4
231.5
301.3

Capital cost
Current
dollars

1929
dollars

Average
price of
capital
index
(4)+ (5)

(4)

(5)

(6)

17,034
22,920
14,033
48, 422
55, 801

18,405
22,920
21,499
25,968
31,996

92.6
100.0
65.3
186.5
174.4

In order to obtain an index of final product prices consistent with
the factor price measure, it is necessary to compute the quotient of net
23734—58

228

ECONOMIC STABILITY AND GROWTH

private domestic product at factor cost in current prices and in constant prices. As a "net" measure, the implicit price deflator accords
a smaller weight to the prices of capital goods than would price deflators of gross product, since capital outlays required to offset capital
consumption are excluded. As a measure of the average prices of
national product "at factor cost," the effect of indirect business taxes
on market price is eliminated. In practice, it is obtained by dividing
national income by real net product at factor cost. The latter variable
is gotten by extrapolating base-period income by the index of real
net product, since the deflated net product at factor cost should show2
virtually the same movement as deflated net product at market price.
Now, net national product at factor cost equals national income
(Y); when divided by real product as defined (O), a measure of
average product price is obtained ( P o ) ; when divided by an index
of real factor input (I), a measure of average factor price is obtained
( P ^ . The following equation demonstrates that average product
price is the quotient of average factor price and productivity
(T=O/I);

In table 1, the values of these variables have been entered for key
years beginning with 1919. Over the period 1919-53, factor price has
risen at an average rate of 3.3 percent a year, productivity by 2.1
percent, and the product price level by 1.2 percent. Thus, productivity advance has mitigated the effect of increasing money demand on
factor price by approximately two-thirds.
The importance of the cushion provided by productivity against
the impact of inflation may be realized more vividly in terms of the
aggregate percent changes over the 34 years. If productivity had not
grown at all, and all other things had been the same, prices would have
tripled—which is what happened to factor price. But actually, productivity doubled, which reduced the price increase to 50 percent.
These changes were the net result of divergent tendencies over the
several subperiods. In the first decade shown in the table, 1919-29,
product prices actually dropped somewhat, despite a 15-percent rise in
factor price, as a consequence of the accelerated rate of productivity
advance that set in around the end of World War I. Between 1929
and 1937, in the absence of full recovery from the great depression,
factor price was 10 percent lower in 1937 than in 1929, but product
price was more than 20 percent lower due to the continued, although
2 Cf. John W. Kendrick, The Estimation of Real National Product, Studies in Income and Wealth,
vol. 22, N. B. E. R. The implicit price deflators for the net product at factor cost and at market prices
do not diverge substantially over the period, as shown below:
Implicit price deflators for net private domestic product
[Index numbers, 1929=100]
At market
price
1919
1929
1937
1948
1953

_. __
_..

108.3
100.0
82.2
144.3
159.1

At factor
cost
107.3
100.0
78.9
149.4
159.5

Ratio
(l)+(2)
100.9
100.0
104.2
96.6
99.7

ECONOMIC STABILITY AND GROWTH

229

somewhat retarded, advance in productivity. With the wartime and
postwar inflationary pressures, unit factor price in 1953 stood almost
3 times higher than the 1937 level, while the product price level had
doubled. Continued strong productivity gains had cushioned the
effect on prices of monetary inflation, which was accentuated by continuous rounds of wage increases in the postwar period. But it is
clear that the magic of technological advance alone cannot prevent
significant price rises in the face of major inflationary pressures.
At least the 1.3 percent average annual rate of increase in product
prices 1948-53 is far lower than the 6 percent rate over the preceding
decade. The 1.3 percent rate is close to the long-term rate, and more
recent figures indicate that approximately the same rate prevailed
from 1953 to 1956. The rate would have been lower in the last 3 years
except for an apparent slowing of productivity advance. On the other
hand, the rate of advance in factor prices also slowed—a condition not
likely to prevail indefinitely for reasons we shall go into in the next
section which deals with the individual factor prices.
RELATIVE CHANGES I N FACTOR PRICES

The index of total factor prices is a composite of the prices of the
various component types of factors. Each individual factor price
may have changed by more or less than the weighted average of all.
Average hourly labor compensation has changed in somewhat different proportions in the various occupational or industry groupings;
and unit capital compensation has varied among the several industries.
But the interindustry structures of wage rates and of unit capital
compensation have been relatively stable over time in contrast to the
marked difference in movement between the prices of the two major
factor classes, labor and capital.
Between 1919 and 1953, average hourly labor compensation increased at an average annual rate of 3.8 percent a year—almost double
the 1.9 percent average increase in the price of capital. Over the 34
years, these average rates of increase were associated with a 256 percent increase in labor rates compared with an 88 percent rise in unit
capital compensation. Reflecting the heavier weight of unit labor
compensation, total factor compensation per unit rose by 200 percent
over the period, which reduces to the average annual gain of 3.3
percent.
It will be noted that the 1.9 percent average annual increase in the
price of capital was somewhat higher than the rate of increase in average product prices generally. Since there was little trend in the rate
of return on capital, the explanation lies primarily in a somewhat
faster rise in the prices of capital goods, as measured, than in the prices
of other final products. Insofar as the quality of capital goods increased more than the quality of other goods, the differential price
movement is overstated but cannot be quantitatively adjusted for.
During the first decade 1919-29, total factor price rose by 1.4 percent a year—less than half its rate of increase over the whole period.
But the average increase in wage rates of 1.7 percent was twice the
average increase in the price of capital, as was true for the longer
period. Between 1929 and the submerged peak of 1937, wage rates
were barely able to hold their own, while the price of capital declined

230

ECONOMIC STABILITY AND GROWTH

substantially as a result of the incomplete recovery from the great
depression. The next subperiod, 1937-48, was the only one in which
the rate of increase in the price of capital exceeded that in the price
of labor—10 as compared with 8 percent a year—due both to the low
1937 base and to the postwar shortage of capital still prevailing at the
high 1948 peak. The relationship during this period also reflects the
high degree of monetary inflation accompanying the war and continuing well into the postwar period.
The relationship between factor prices reversed completely in the
following period. Between 1948 and 1953, while the rate of increase
in wage rates slowed somewhat to a 5.5 percent annual average, the
price of capital actually declined. Here, it would seem inflationary
pressure tended to originate on the labor cost side, with mild restraint on the part of the monetary authorities in the latter half of
the period resulting in some squeeze on profits. Rough estimates for
the period 1953-56 point to a continuation of the same basic situation. Wage rates rose by more than 4 percent a year on average,
while a further squeezing of the rate of return on capital more than
offset small increases in the price of capital goods and produced a 2
percent average annual decline in the price of capital.
Despite the decline in the rate of return on capital in the postwar
period, the return was still sufficient to induce a volume of new investment consistent with relatively full employment. But it is obvious
that the rate of return cannot continuously decline and still produce
sufficient investment demand. Assuming an eventual leveling or
rising rate of return on capital, and the same rates of increase in wage
rates and productivity as have prevailed in the postwar period, the
rate of increase in final product prices would accelerate as unit factor
costs, following the current contractions, rise in greater degree than
in the period 1948-57.
Forces behind relative factor price changes
Whatever the rise in unit factor costs, it may be expected that wage
rates will rise faster than the price of capital in the future as in the
past. In the past, this relative movement has resulted in the labor
share of productivity gains and of the national income rising. So it
is worth pausing to consider some of the reasons for a relative increase in the price of labor.
The variables determining relative price movements of the factors
are complex, but two major influences stand out—one relating to the
rate of return on capital and the other to the prices of capital goods,
the product of which is the price of capital as we define and measure
it.
With respect to the first influence, the statistics show that net capital
formation has been high enough secularly in this country to result in
a significantly greater increase in real capital stocks and services than
in the labor force and man-hours worked. The law of diminishing
marginal productivity tells us that under these circumstances, and in
the absence of technological advance, the rate of return on capital
would decline both absolutely and in relation to the wage rate. Actually, technological advance has shifted the factor demand curves
upward so that there has been no pronounced trend in the rate of return to capital, while real wage rates have risen.

231

ECONOMIC STABILITY AND GROWTH

The second influence relates to the prices of capital goods. Increasing productivity in the capital goods industries as in the economy generally means that capital goods prices fall in relation to wage
rates (increase less), assuming relatively full employment and competitive conditions which tend to cause prices to approximate the cost
of production per unit, and result in labor being paid in accordance
with its (rising) marginal productivity. Since there has been no
corresponding offset in a rising rate of return on capital over the
long run, the decline in capital goods prices relative to wage rateg
is a built-in factor in dynamic economies that promotes the substitution of capital for labor.
TABLE 2b.—Private domestic economy—Relative factor prices

Price per unit of factor input
Year
Labor
(1)

R e l a t i v e factor
prices, reciprocal
ratios

Capital

Total

Labor to
capital
(D-K2)

Capital
to labor
(2)-KD

(2)

(3)

(4)

(5)

A. Index numbers, 1929=100
1919
1929
1937
1948
1953

84.6
100.0
100.4
231.5
301.3

92.6
100.0
65.3
186.5
174.4

86.7
100.0
90.1
218.9
261.8

91.4
100.0
153.8
124.1
172.8

109.4
100.0
65.0
80.6
57.9

109.4
153.8
80.7
139.2
189.1

91.4
65.0
124.0
71.8
52.9

B. Link relatives
1929/19...
1937/29...
1948/37....
1953/48—
1953/19—

118.2
100.4
230.6
130.2
356.1

108.0
65.3
285.6
93.5
188.3

115.3
90.1
243.0
119.6
302.0

C. Average annual rates of change (percentages)
1929/19...
1937/29__
1948/37...
1953/48. _.
1953/19...

1.7
.1
7.9
5.4
3.8

0.8
-5.2
10.0
-1.3
1.9

1.4
-1.3
8.4
3.6
3.3

0.9
5.5
-1.9
6.8
1.9

-0.9
-5.2
2.0
-6.4
-1.9

NOTE.—Table may not be internally consistent due to rounding. The quotients in cols. (4) and (5) of
part O are derived as indicated only when 100.0 percent is added to the average annual rates of change..1

232

ECONOMIC STABILITY AND GROWTH

TABLE 3.—Private domestic economy—Factor shares of national income in
current and constant dollars
Percent distribution of realfactor
cost in 1929 prices

Relative factor
prices i (1929=100)

Year

1919 1929
1937
1948
1953

Percent distribution of national
income in current
prices

Labor

Capital

Labor

Capital

Labor
d)X(3)

Capital
(2)X(4)

(1)

(2)

(3)

(4)

(5)

(6)

97.6
100.0
111 4
105.8
115.1

106.8
100.0
72.5
85.2
66.6

73.8
72.3
70.8
72.0
68.9

26.2
27.7
29.2
28.0
31.1

72.0
72.3
78.9
76.2
79.3

28.0
27.7
21.1
23.8
20.7

i Index of ratios of individual factor prices to total factor price (see table 2b).
NOTE.—Table may not be internally consistent due to rounding.

On the demand side, it is conceivable that inventions might be sufficiently labor saving (that is, require increasing quantities of capital
relative to labor, given constant relative factor prices) to cause the
demand for capital to increase more rapidly than the demand for labor.
But if this has been the case (as distinguished from the substitution
of capital for labor as a result of changing relative factor prices),
it has not been a strong enough tendency to offset the effect on relative
price of the greater increases in the supply of capital than of labor—
since the estimates show that wage rates have consistently risen relative to the price of capital in all periods when capital per unit of labor
was rising.
Factor shares in national income
The national income accruing to each factor is the product of the
quantity employed and its price (unit cost). Aggregate national
income is the sum of the compensations of all the factors. Thus, the
share of each factor in total national income will vary in accordance
with the net effect of changes in the quantity of the factor employed
relative to total input, and the price of the factor relative to average
factor price.
It was apparent in table 2b that the input of capital rose substantially relative to labor input between 1919 and 1953, and in all subperiods except 1937-48. The ratio of capital to labor input went up
by 27 percent over the whole period. As a ratio to total factor input,
the increase was only 7 percent—since the marginal rate of substitution of capital for labor was more than 3:1 based on the average
weights accorded the two factors over the period. The ratios of the
input of each factor to total factor input is shown in the first two columns of table 3. From 1919 to 1953, the ratio of labor to total factor
input declined from 74 to 69 percent.
But the decline in relative labor input was associated with a more
than proportional increase in the price of labor services relative to total
factor price. The 18 percent increase in this ratio (derived from the
data of table 2 and shown in the third and fourth columns of table
3), is less than the 89 percent increase in the ratio of the price of labor
to that of capital alone, again because of the much higher weight of
labor.

ECONOMIC STABILITY AND GROWTH

233

Only in the subperiod 1937-48 did the relative price of labor decline.
But in all subperiods, the relative prices and relative inputs of the
two factors moved inversely. It is clear that it was through relative
price movements that the different rates of supply of the two factors
were absorbed into the productive system. That is, cost economies were
achieved by producers in substituting the factor that was becoming
relatively cheaper for the one that was growing dearer as a result of
changing relative supplies. Over the period as a whole, the ratio
of the percent change in relative factor prices to the percentage
change in relative factor inputs was — 0.24. The coefficient of substitution varied considerably among the subperiods, however.
The last two columns of table 3 show the net effect on income shares
of the inverse movement of relative factor inputs and prices. Since
the decline in the relative input of labor was significantly smaller than
the increase in relative labor price, the share of labor increased from
71 percent in 1919 to 79 percent in 1953. The same percentages may
be calculated directly from table 2. Only in subperiod 1937-48 did
labor's share in the national income temporarily decline due to the
peculiar circumstances described earlier. This implies that in all
subperiods except 1937-48, and over the long period, labor received a
significantly larger share of the productivity increment to real income
(product) than its share of total income at the beginning of each
period.
RELATIVE CHANGES I N REAL FACTOR COMPENSATION PER UNIT

Once the prices of the factors have been calculated, it is easy to
compute the real earnings per unit of each of the factor inputs. This
involves dividing the factor prices (i. e., the current dollar compensation per unit) by an index of the prices of products for which factor
incomes are spent, directly or indirectly. For that index, we use the
implicit price deflator for the net domestic product at factor cost.
This index is composed of the prices of new capital goods and goods
purchased by Government, as well as consumer goods, although consumer goods have by far the largest weight.
It could be argued that labor income is distributed in a somewhat
different way among these types of goods (i. e., as among spending,
saving, and taxes) than is the income accruing to capital, and that
to measure the purchasing power of the different types of unit compensation different price indexes should be employed with weights
based on the patterns of spending out of each type of income. But
both types of income are used for all the major types of final product,
and it is statistically impossible to relate patterns of spending to type
of factor income since most spending units do not receive a pure form
of either. In any case, the results using a different deflator would
not differ substantially. Over the long period 1919-53, the consumers'
price index increased by only a few percent less than our deflator.3
The results of deflating current dollar unit factor compensation by
product price is shown in table 4a in index number form. Since
average hourly earnings had increased substantially more than average compensation per unit of capital, it follows that the real increase
in the former would also be greater. But price deflation accentuates
3
Conceptually, a market price index would be preferable to our index at factor cost, but
the differences between the two are minor and our index has the advantage of permitting
precise definition of the relationships between productivity, prices, and unit factor costs.

234

ECONOMIC STABILITY AND GROWTH

the differential movement: between the years 1919 and 1953, real
average labor earnings increased by almost 150 percent, more than
4 times as much as the 35-percent increase in real compensation per
unit of capital.
The gams in real unit compensation of each factor may be compared with the gain in productivity, which is also the gain ill real
income per composite unit of factor input. (See table 4b.) The
proportionate gain in real average hourly earnings of labor was onefourth greater than the proportionate increase in total factor productivity over the period (although about equal to the rate of gain in
real product per man-hour). The proportional gain in real unit compensation of capital was only about one-fourth of the total productivity increase.
The marked difference between the increases in real unit earnings
of the two factors and the productivity increases is explained by the
differential movement in the prices of the two factors. (See table 4b.)
The price of capital fell by around 38 percent relative to composite
unit factor price, and it was this relative decline that caused the real
earnings per unit of capital to rise less than productivity and, conversely, made it possible for the real average earnings of labor to rise
substantially more than the proportionate increase in productivity.
TABLE 4a.—Private domestic economy—Real factor income per unit
[Index numbers, 1929=100]
Current income per unit
(factor price)

1919
1929
1937
1948 . 1953

Labor

Capital

(1)

(2)

84.6
100.0
100.4
231.5
301.3

92.6
100.0
65.3
186.5
174.4

Average
product
price
(3)
107.3
100.0
78.9
149.4
159.5

Real income per unit
Labor

Capital
(2)-f(3)

(4)

(5)

78.8
100.0
127.2
155.0
188.9

86.3
100.0
82.8
124.8
109.3

235

ECONOMIC STABILITY AND GROWTH

TABLE 4&.—Private domestic economy—Productivity and real factor income
per unit

Year

Relative factor
prices

Productivity

Real income per
unit

Labor

Capital

Labor
(DX(2)

Capital
(DX(3)

(2)

(3)

(4)

(5)

(1)

A. Index number, 1929=100
1919.-1929.—
1937.-1948....
1953-...

80.8
100.0
114.2
146.5
164.1

97.6
100.0
111.4
105.8
115.1

106.8
100.0
72.5
85.2
66.6

78.8
100.0
127.2
155.0
188.9

86.3
100.0
82.8
124.8
109.3

126.9
127.2
121.9
121.9
239.7

115.9
82.8
150.7
87.6
126.7

B. Link relatives
1929/19.
1937/29.
1948/371953/48.
1953/19.

123.8
114.2
128.3
112.0
203.1

102.5
111.4
95.0
108.8
117.9

93.6
72.5
117.5
78.2
62.4

C. Average annual rates of change (percentages)
1929/191937/291948/371953/48.
1953/19-

2.2
1.7
2.3
2.3
2.1

0.2
1.4
-.5
1.7
.5

-0.7
-3.9
1.5
-4.8
-1.4

2.4
3.1
1.8
4.0
2.6

1.5
-2.3
3.8
-2.6
.7

NOTE.—Table may not be internally consistent due to rounding.

So long as capital increases more rapidly than labor, input and the
price of capital rises less rapidly than wage rates, it is apparent that
wage rates can rise somewhat faster than total factor productivity and
still be consistent with a stable product price level. Beyond a point,
however, increases in wage rates are associated with price inflation,
if the monetary authorities accommodate the rise of unit costs. This
historical survey indicates, at least roughly, where that point lies.
TECHNICAL NOTE

The estimates of real private product since 1929 are those published
by the United States Department of Commerce (National Income,
1954 edition, a supplement to the Survey of Current Business, as revised) carried back to 1919 by estimates of the common components by
Prof. Simon Kuznets supplemented by our own estimates of Government purchases of goods and services from private industry, based on
Treasury Department and Census Bureau data for Federal and for
State and local purchases, respectively.

236

ECONOMIC STABILITY AND GROWTH

The nonfarm employment estimates from 1929 forward are also
those published by the Commerce Department inclusive or proprietors,
supplemented to include unpaid family workers as estimated by the
Census Bureau. The estimates of farm employment, including unpaid
family workers, are those of the Department of Agriculture back to
1919. Nonfarm employment was extrapolated from 1929 to 1919 by
estimates based largely on Census Bureau sources, and contained for
the most part in various studies of output, employment, and labor
force published by the National Bureau of Economic Eesearch. The
estimates of average hours worked per week in private nonfarm industries needed in conjunction with employment to estimate manhours worked, are largely those of the Department of Labor and the
Census Bureau. Farm man-hour estimates of the Department of
Agriculture were used with a level adjustment. In establishing levels
for some of the nonfarm industries in the early part of the period, use
was also made of estimates for the years 1920-22 based on sample
surveys by W. I. King for the National Bureau of Economic Eesearch.
The estimates of real net capital stocks in the private nonfarm, nonresidential sector are largely those prepared by Raymond Goldsmith
and published in A Study of Saving in the United States, volume 3
(Princeton University Press, 1955), subsequently revised and extended
since 1945 by Mr. Goldsmith. The estimates of real farm capital are
chiefly those prepared by Alvin S. Tostlebe, Capital in Agriculture:
Its Formation and Financing Since 1870 (Princeton University Press
for the National Bureau of Economic Research, 1957). Estimates of
the real net stock of nonfarm residential real estate are based on those
contained in the volume by Leo Grebler, David M. Blank and Louis
Winnick, Capital Formation in Residential Real Estate (Princeton
University Press for the NBER, 1956).
The national income estimates, used both as a basis for weighting
the factor inputs by type and by industry, and for computation of
the compensation per unit of factor input by type, are those published by the Department of Commerce since 1929. They were extended back to 1919 largely by the estimates of Professor Kuznets,
National Income and Its Composition, 1919-38, adjusted for consistency with our employment estimates. The net income of proprietors was separated into labor and capital components by imputing the
average annual compensation of employees in each industry group
to the proprietors. The sum of the imputed labor income of proprietors plus wages, salaries, and supplements make up total labor
compensation; the rest of national income equals capital compensation and comprises interest, net rents, and royalties, and profits.

IV
INTERRELATIONSHIPS AMONG PRICES, DEMANDS,
AND COSTS

237

IV. Interrelationships among prices, demands, and costs
A. General price movements.
1. Under what circumstances can we expect general
price movements—inflation or deflation—to
originate in an excess of demand over the supply
forthcoming at constant prices? To originate
in changes in unit costs? Can a cost-push price
movement continue to operate in the absence of
an "excess demand situation" ? If so, then for
how long, and under what conditions ?
2. Accepting relative price movements as proper and
necessary under a dynamic economy, will a
change in relative prices induce more general
price movements—i. e., are there any individual
products or services so important to the economy
as a whole that changes in their prices are necessarily followed by widespread changes in prices
of other products and services ? In other words,
should policies for the control of inflation or deflation be substantially concerned with influencing prices of certain particular goods or services ?
3. To what extent do general price-level changes tend
to feed upon themselves, with accelerating or
cumulative movements away from a stabilized
price level ?
B. In short-run situations, do movements of prices of some
products and services tend to be determined mainly by
changes in demand while others reflect mainly changes
in unit costs? If this distinction seems useful, what
products and services would you put in each classification, and why ?
C. Relationships between prices and—
1. Aggregate demand.
(a) What are the relationships between the
level of, and changes in, the supply of
money, and the level of, and changes in,
the general average of prices of goods
and services? Of productive factors?
(&) How and to what extent are prices in the
United States affected by developments
in other countries, especially changes
in international prices? Under what
circumstances does this relationship
run in the reverse direction from
changes in the United States to changes
in international prices ?
2. Consumer demand.
3. Investment demand.
4. Government demand.
238

ECONOMIC STABILITY AND GROWTH

239

D. Relationships between prices and costs.
1. The determinants of costs:
(a) How are unit costs affected by the rate of
utilization of plants and equipment!
Of labor force f Of other resources ?
(&) How are unit costs affected by changes in
the technical efficiency of productive
factors or of the way in which they are
combined ?
(c) How are unit costs affected by the size and
scale of enterprises ?
(d) How are unit costs affected by changes
in the prices of productive factors?
2. Factors affecting prices of productive resources:
(a) What influences the price of productive
factors: Wages? Profits? Interest?
Eents? etc.
(6) How are changes in demand for goods
and services related to changes in prices
of productive resources ?
3, How are unit costs related to prices in the long
run? In the short run? What classes of costs
are relevant to the analysis of such cost-price
relationships?

THE SUPPLY OF MONEY AND CHANGES IN PRICES
AND OUTPUT
Milton Friedman, University of Chicago, National Bureau of
Economic Research, and Center for Advanced Study in the
Behavioral Sciences
This paper deals with two broad issues that have arisen again and
again in connection witli movements in the general level of prices.
One issue is the connection between such price movements and
changes in the supply of money. The other is the relation between
price changes and changes in output.
The course of economic history is replete with substantial price
disturbances. Whenever such disturbances have occurred, two different explanations have been offered. One, common to all disturbances, is that the price movements reflect changes in the quantity of
money, though the source of the monetary changes has varied
widely—from clipping of currency to gold discoveries to changes in
the monetary standard to the printing of paper money to the creation
or destruction of deposit money by central banks and commercial
banks. The other explanation has been in terms of some special circumstances of the particular occasion: good or bad harvests; disruptions in international trade; lack of confidence; the activities of
"profiteers" or "monopolists" selling goods or of employers seeking
to hold down wages; the activities of workers or unions pushing
wages up; and so on in great variety. Perhaps the one common core
of such explanations is that they generally attribute the price movements to the (socially) misguided behavior of particular individuals
or groups. My own view is that these alternative explanations play
little or no role in either long range or large movements in prices,
though they may in short and minor movements, except indirectly
as they affect the supply of money. It is clearly impossible to argue
this view in detail within the compass of this paper. My reason for
stating it is to make clear that I am putting such explanations to
one side and concentrating instead on the monetary forces at work.
The relation between the supply of money and prices has been explored so frequently and thoroughly that I can hardly hope to add
much that is new on an analytical level. My reason for dealing with
it nonetheless is twofold: on the one hand, though it is the essence
of the problem of long run and large price movements, it tends to be
pushed to one side and neglected—partly, perhaps, because of the
desire to be novel; on the other hand, extensive empirical work that
is currently underway puts flesh on the analytical skeleton to an extent
that has not heretofore been possible. One of the major aims and
justifications 1of this paper is to summarize some of the broad findings
on this work. I shall do so in section 1 for the longer term changes
in money and prices, in section 2, for the shorter term changes.
1
These are based partly on the preliminary results of an extensive study by Anna J.
Schwartz and myself under the auspices of the National Bureau of Economic Research on
the secular and cyclical behavior of the stock of money in the United States, partly on a
series of studies done in the workshop in money and banking at the University of Chicago.
The views expressed in this paper are of course my own and are not necessarily those
of the organizations sponsoring these studies or of the other participants in them.

241

242

ECONOMIC STABILITY AND GROWTH

Discussion of public policy with respect to prices necessarily involves the issue what kind of movements are socially desirable. One
major problem is the relation of price movements to economic growth.
Is a rising price level favorable or unfavorable to rapid growth in
output? No conclusive answer can be given to this question in the
present state of our knowledge. Some analysis and evidence to
justify this assertion are given in section 3.
The final section of this paper presents some implications for
policy that are suggested by the relation between monetary and price
change and between price change and output change.
1. Relation of stock of money to prices over longer periods
There is perhaps no empirical regularity among economic phenomena that is based on so much evidence for so wide a range of circumstances as the connection between substantial changes in the stock
of money and in the level of prices.2 To the best of my knowledge
there is no instance in which a substantial change in the stock of money
per unit of output has occurred without
a substantial change in the
level of prices in the same direction.3 Conversely, I know of no instance in which there has been a substantial change in the level of
prices without a substantial change in the stock of money per unit of
output in the same direction. And instances in which prices and the
stock of money have moved together are recorded for many centuries
of history, for countries in every part of the globe, and for a wide
diversity of monetary arrangements.
There can be little doubt about this statistical connection. The
statistical connection itself, however, tells nothing about direction of
influence, and it is on this question that there has been the most controversy. It could be that a rise or fall in prices, occurring for whatever
reasons, produces a corresponding rise or fall in the stock of money,
so that the monetary changes are a passive consequence. Alternatively, it could be that changes in the stock of money produce changes in
prices in the same direction, so that control of the stock of money
would imply control of prices. The variety of monetary arrangements for which a connection between monetary and price movements
has been observed supports strongly the second interpretation, namely,
that substantial changes in the stock of money are both a necessary
and a sufficient condition for substantial changes in the general level
of prices. But of course this does not exclude a reflex influence of
changes in prices on the stock of money. This reflex influence is often
2
"The stock of money" is not of course an unambiguous concept. There is a wide range
of assets possessing to a greater or lesser degree the qualities of general acceptability and
fixity in nominal value that are the main characteristics of "money." It is somewhat
arbitrary just where the line is drawn which separates "money" from "near-money" or
"securities" or "other financial claims." For most of what follows, the precise line drawn
will not affect the analysis. For the United States at present, I shall treat as "money in
the hands of the public" the sum of "currency outside banks," "demand deposits adjusted,"
and "adjusted time deposits in commercial banks," as these terms are defined in Federal
Reserve monetary statistics. I shall note explicitly any point at which the precise definition
3 adopted affects the statements made.
The nearest thing to an exception I know of is German experience from the midthirties
to 1944. See John J. Klein, German Money and Prices, 1932-44, in Milton Friedman (Ed.),
Studies in the Quantity Theory of Money (University of Chicago Press, 1956), pp. 121-159.
The qualification, "per unit of output" is needed only to cover movements spanning
long periods of time, like the long term decline in prices in the late 19th century. For
moderately short periods, even the qualification is unnecessary.

ECONOMIC STABILITY AND GROWTH

243

important, almost always complex, and,4 depending on the monetary
arrangements, may be in either direction.
This general evidence is reinforced by much historical evidence of
a more specific character demonstrating that changes in the stock of
money, at least when they are fairly large, can exert an independent
influence on prices. One dramatic example is from the experience of
the Confederacy during the Civil War. In 1864, "after 3 years of
war, after widespread destruction and military reverses, in the face
of impending defeat, a monetary reform that succeeded in reducing
the stock of money halted and reversed for some months a rise in prices
that had been going on at the rate of 10 percent a month most of the
war. It would be hard to construct a better controlled experiment to
demonstrate the critical importance of the supply of money."5 The
effect of discoveries of precious metals in the New World in the 16th
century and of gold in California and Australia in the 1840's, of the
development of the cyanide process for extracting ore plus gold discoveries in South Africa in the 1890's, and of the printing of money
in various hyperinflations, including our own Revolutionary War
experience and the experience of many countries after World War I
and World War II, are other striking examples of increases in thq
stock of money producing increases in prices. The long price decline
in the second half of the 19th century in many parts of the world is a
less dramatic example of a decline 6in the stock of money per unit of
output producing a decline in prices.
The relationship between changes in the stock of money and changes
in prices, while close, is not of course precise or mechanically rigid.
Two major factors produce discrepancies: changes in output, and
changes in the amount of money that the public desires to hold relative
to its income.
For the moment, we shall treat output as if it were determined
independently of monetary and price changes, postponing to section 3
the relation between them. This is clearly a simplification that is to
some extent contrary to fact, but certainly for the longer periods and
larger changes that are discussed in this section, the simplification
neither does serious violence to the facts nor leads to any significant
errors in conclusions.
Suppose the stock of money were to remain unchanged for a
period of years but total output over the same period were to double.
Clearly, one would expect prices to fall—other things the same—to
something like half their initial level. The total amount of "work"
for the money stock to do, as it were, is doubled, and the same nominal quantity of money could perform the "work" only at lower levels
of prices. Eoughly speaking, this is what happened in the United
States in the period from the end of the Civil War in 1865 to the
resumption of specie payments in 1879: The stock of money was
4
For example, under a gold standard, a rising level of prices discourages gold production and so, after a lag tends to produce a decline in the stock of money. On the other
hand, under a fractional reserve banking system, if rising prices lead banks to reduce the
ratio
of cash to liabilities, rising prices may tend to produce a rise in the stock of money.
5
Mlton Friedman, The Quantity Theory of Money—a Restatement, in Studies in the
Quantity Theory of Money, p. 17. The quotation summarizes one item from a s-tudy
by Eugene M. Lerner, summarized in his article, Inflation in the Confederacy, 1861-65, in
the6 same volume, pp. 163-175.
The decline in the stock of money per unit of output occurred as a result of (1) exhaustion of then-known gold mines; (2) the shift of many countries from a silver to a gold
standard ; (3) the rapid increase in output.

23734—58

244

ECONOMIC STABILITY AND GROWTH

roughly the same in 1879 as in 1865—if anything, some 10 percent
higher; output grew very rapidly over the period, probably more7
than doubling; and wholesale prices were half their initial level.
Thus, for price movements, the relevant variable is the stock of money
per unit of output, not simply the global stock of money.
The second major factor that can introduce a discrepancy between
movements in money and in prices is a change in the ratio that 8the
public desires to maintain between its cash balances and its income —
the public including individuals, business enterprises other than banks,
nonprofit institutions, and the like. The number of dollars an individual wants to keep in cash depends of course on the price level—at
twice the price level he will want to hold something like twice the
number of dollars—and on his income—the higher his income presumably the larger cash balances he will want to hold. But the price level
is what we are trying to explain, and we have already taken account of
the effect of changes in output. This is why we express this factor
in terms of the ratio that the public desires to maintain between its
cash balances and its income, rather than in terms of the number of
dollars it desires to hold.
Broadly speaking, the public as a whole cannot by itself affect the
total number of dollars available to be held—this is determined primarily by the monetary institutions. To each individual separately,
it appears that he can do so; in fact an individual can reduce or increase his cash balance in general only through another individual's
increasing or reducing his. If individuals as a whole, for example,
try to reduce the number of dollars they hold, they cannot as an aggregate do so. In trying to do so, however, they will raise the flow
of expenditures and hence of money income and in this way will reduce the ratio of their cash balances to their income; since prices will
tend to rise in the process, they will thereby reduce the real value of
their cash balances, that is, the quantity of goods and services that
the cash balances will command; and this process will continue until
this ratio or this real value is in accord with their desires.
A wide range of empirical evidence suggests that the ratio which
people desire to maintain between their cash balances and their income is relatively stable over fairly long periods of time aside from
the effect of two major factors: (1) The level of real income per capita,9
or perhaps of real wealth per capita; (2) the cost of holding money.
(1) Apparently, the holding of cash balances is regarded as a
"luxury," like education and recreation. The amount of money the
public desires to hold not only goes up as its real income rises but goes
up more than in proportion. Judged by evidence for the last 75 years
in the United States, a 1 percent rise in real income per capita tends
to be accompanied by nearly a 2 percent increase in the real amount
of money held and thus by nearly a 1 percent increase in the ratio of
7
The basic data underlying this statement are from the National Bureau study mentioned in footnote 1 above. They will appear in a monograph by Anna J. Schwartz and
myself that is now in preparation.
s9 The reciprocal of this ratio is termed "the income velocity of circulation."
On this subiect, see Phillip Casan, The Monetary Dynamics of Hyperinflation, and
Richard T. Selden, Monetary Velocity in the United States, in Studies in the Quantity
Theory of Money. The statements that follow are based also on additionoal work done
in connection with the National Bureau study referred to in footnote 1.
For shorter periods, an additional factor enters. Cash balances are apparently adjusted
to longer term income expectations ("permanent income") rather than to current income
as measured on a monthly or annual basis. This introduces additional changes in the
ratio of cash balances to current measured income. (See sec. 2 below.)

ECONOMIC STABILITY AND GROWTH

245

cash balances to income. This tendency is highly regular over the
long sweep of time from 1875 to World War I I ; it has not been operative since the end of World War I I but it is yet too soon to judge
whether this a fundamental change or simply a reaction to the abnormally high ratio of cash balances that was reached during the war.
(2) The cost of holding cash balances depends mainly on the rate
of interest that can be earned on alternative assets—thus if a bond
yields 4 percent while cash yields no return, this means that an individual gives up $4 a year if he holds $100 of cash instead of a bond—
and on the rate of change of prices—if prices rise at 5 percent per
year, for example, $100 in cash will buy at the end of the year only
as much as $95 at the beginning so that it has cost the individual $5
to hold $100 of cash instead of goods. The empirical evidence suggests
that while the first factor—the interest rate—has a systematic effect
on the amount of money held, the effect is rather small. The second
factor, the rate of change of prices, has no discernible effect in ordinary
times when price changes are small—of the order of a few percent a
year. On the other hand, it has a clearly discernible and major effect
when price change is rapid and long continued, as during extreme
inflations or deflations.10 A rapid inflation produces a sizable decline
in the desired ratio of cash balances to income; a rapid deflation, a
sizable rise.
Of course even after allowance is made for changes in real income
per capita and in the cost of holding money, the ratio of cash balances
to income is not perfectly steady. But the remaining fluctuations in
it are minor, certainly far smaller than those that occur in the stock of
money itself.
Some idea of the quantitative magnitude of the changes in the
United States over long periods of time can be obtained by comparing
average values of various items over the most recent complete business
cycle—that running from a trough in 1949 to a peak in 1953 to a
trough in 1954—with those over the earliest for which we have the
relevant data—that running from a trough in 1878 to a peak in 1882 to
a trough in 1885. The money stock multiplied 67-fold over these seven
decades, and real income ninefold, so the money stock per unit of
output rose about 7.5-fold. Prices something less than tripled, so the
ratio of the money stock to money income roughly tripled. In the
initial cycle, the stock of money averaged about 24 percent of 1 year's
money income—that is, cash balances were equal to the income of
about 3 months; in the terminal cycle, the stock of money averaged
about 67 percent of 1 year's income—that is, cash balances were equal
to the income of about 8 months. Over the period as a whole, the
money stock rose at an average rate of 6 percent per year, money income at nearly 5 percent per year, prices at nearly IV2 percent per
year, total output at about 3 percent per year, and population at
about iy2 percent per year.
Of course, these changes did not occur smoothly. Figure 1 shows
the more detailed behavior based on average values for each of the 19
business cycles that we have experienced since 1879. It is clear that
there is an exceedingly close connection between movements in the
stock of money per unit of output and in prices. The only major dif10

Evidence for this is presented in Cagan, op. cit., and is available also from work by
John Deaver on monetary changes in Chile.

246

ECONOMIC STABILITY AND GROWTH

ference is the more rapid long-term growth in the stock of money
which in turn reflects the effect of the long-term growth in per capita
real income and the associated rise in the desired ratio of money stock
to money income.
8. Relation of stock of money to prices over shorter periods
Over the longer periods considered in the preceding section, changes
in the stock of money per unit of output tend to dominate price
changes, allowance being made for the effect of the growth of real
income per head. This is less so over the shorter periods involved in
the fluctuations we term business cycles, though the general and average relationship is very similar. The reason for the looser connection
in such periods presumably is that movements in both the stock of
money and in prices are smaller. Over longer periods, these movements cumulate and tend to swamp any disturbance in the relation
between desired cash balances, real income, and the cost of holding
money; in the ordinary business cycle, the disturbances, though perhaps no more important in an absolute sense, are much more important relative to the movements in money and prices.

•ill ;(|t( I l l (1

III II'III III
tillIIIllll ll|l
1' PI lill! r

Iuiiiiii
I llit iiiillli
l l i l ll
'

mill 'IImil ii!
|| i11
Illl'l
I

'•
j>

|
•
Tn.
t
•hi
||
T

|l|

!•' T I

iii
[j!|

f r f -\\ f-in "~~"rr
•

l: r' 1 1ILi
j • f (j j

II•1 I It1 i||l'III illi"P^>
'
IJijjiTi |U It i i I
ill'
Ijljr
inn

mil
mil

' (i

T - - ^ ' -1

i
' *'
ujKp if . , .-L,
L u L ' J. . . . .

1 !

•

' i l l

1 j !'

' | || » J 1 '

' 1 ' i

[

i III! (I

/I llil

I I I I II
1
II
U || | ||

—1 »ullil i l l

L. 1 friilll tit
\ llliHl III

\\M\\\\ '

11,

TsJUIllT'lll
1 %

_i|M ,
\

,
py

inl l

—ifmTryft ' " ' i f i — '
I'llill /III
If 1
lill/ill
i/l
I I 1 • j •• -m---

/mil i i

: :

-S;

Tf
B1

JI
in

"1M\\ \ " m""
I

HI | |

II PHI

ljj i

/|ll
| ii|| II||
II Illlllili

sj imil

3$|f|I
pllirli
'

>3||| |;l l l j l l j > |

A\ i

'
r

'-- m
1 F' ::::
•
.:.,
i
' tt
1\l1

— ' 1l l1 l i
_i
333:
3jn in
*31
oLi i' I1'i |i1
II
|1
1
I i

a
J

•

___
"

__ _

::::-!-::—-i:

;

....:

...
:

...,

!.. . _

247

3•i i1nitilll 1lilli i Eit1
1

I'
1
1*

HI |||| IIII
II1 'iff lilt
lllHiirmT

UIIIIII Illl

IINUIIIIII
IIIUJJIU
I Illl

III Illlllili
llllllll
I
1 Illll
I 'DM 1
III lufi '

III
iiiiii

iiiiiy

ifflilf

II
liiitf' ' V
IIIIII
—Jillli
1 Pfltt •
lilil Hi
II II II
ll
III rlFii'
III l u l l

IIITIIJ 1

II
IIIIrPi
11 f
\(MM
\f
1 i'l'll'' i Ii
ll 1 iiln i
I I ||)||I|| j
imTlrl

1 llllll
III

11 inII 1||II

1 || UJ

1 UMJ
ll Iffflfipl I'll
|| l|| I'l | | j | 1,

II illlll ll III Illll
IIII llllll llllll II
II mill |||j

lioni
III II

l l H l r j •<

I'l II[iiIII 1 ,
li'iliiiilli i
II 'uiul'i '

t 'r rH "i
Eli| li|ijj' |I j
l i'|' 'J •' 11

1
T i i r r * !'
1 llllllll i| '
_ i__
.:::::....)..... -ft(!|iP ! h

^ --

I* 1 • 1
I
:::::::;a._._:.[
«J iji.. | in '[ 1
J!
_ _ 3 11 i|'|| || <

I t
,
D :"":":::::::s_

:i ...I
t
\

\ \' i i

1 l''l' 1
HM ' | 1 | ,
ia l' ' '1
"sl 'ii

n I
::::
,, . ,...

-^fl- — f|

, : : . . : ! . . ! :!

|1L:[I

u "n:— ^

- fUff H
~~ i]Lii:

- ! f r f l jj-±-is

Output

ECONOMIC STABILITY AND GROWTH

rf| |i • 11 i i ~f~
' i i ii 1 ' I I l_u-

;1
nx ill
11 H1'' 1 tJ j_
LL
^ nlf rf, J 7ijT~
rt""

-.•:::L1:JLFF^

S+OCK of r\<x*>ey />er U-nit of

¥::::::

•

fft _ __ 1 1 Ilill Illl
1 1 mil I I
t
t.:
_ * lI illil
Ml itll
f
4

:.:._ t;:__: r
t
I
5

i , i
: j I

:::::::::::::::

'
I 1

4
i :

ntj:.--

i_i i j_ 1

Jfjj nj;"""T
1i'1

>. * t t r I " + ' " "
' i 11 i
j_i i t i i
'' | f
'
i p| !

_ iJ-LTitt

a' i l l " " " " " I : : "

** 4 ' 1
111

i yl J 1

4'
|J i

O ' •'
i\ i'[[

1 [M ' ^"n'T'M
'1

ml ul ' i '
ft " ii i ! <:
Ilili ' i

tj M

(

Hill '!' i ] \\ i L '
IJJil ii'j lii j i h
IjTT

III 1 1 ' 1
til 1 'L
p j L j i j i [ U L U . * --<J|j|}

l l i i . ' L ^ ' ^ ' i i i i [ti
II- : f '
|p [Ii
I ji M

|||
l|

1 |j

' !t

1
'
i
1
*s !
•'
!

i]rf t

Ij1 i j

*-: [jf • £T" T ""

I I > JI

'
I 0 :-(-$ '
n rrt• i j i , i M
' FT'
• i f j" T

mil

! j1 j

.i ij

1'./ til: ' |
f ] n" 4 L

;, !jjj St.i > j : . .

Kill i '
llllll 1 l l

Ijtfjj r i! m ~h^ "77j—

mil LJ '
' • i' i
rrpjj $
, j ,
•*•
H i • i' f' i• • i l l ' i i

IIP i !H • • ' '
Hill ii i ''I ' ! i ' i '

||fei;::i:

l'!' Ill i_J
II t IfIl j l j i'Il j j' '[[ | !
l''i i i
• LXJ.1' I'.._
rx _ _
ilj||[' : ; ; !! 1 ' !

Vertical scale logarithmic.
Source: All series are from unpublished data developed in connection with studies a t
the National Bureau of Economic Research. Money figures from study by Milton Friedman
and Anna J. Schwartz; output (net national product in constant prices) and price
(deflator implicit in computing net national product in constant prices) figures preliminary
estimates by Simon Kuznets.

248

ECONOMIC STABILITY AND GROWTH

On the average, prices rise during an expansion phase of a business
cycle, fall during the contraction phase. In the usual fairly mild
cycle of peacetime since 1879, wholesale prices have on the average
risen about 10 percent from trough to peak, and have fallen by somewhat less than half that amount from peak to trough. The general
pattern has not changed much except for the relation of the rise to the
fall. During the period of generally declining prices from the 1880's
to the mid-1890's, prices tended to fall more during the contraction
than they rose during expansion; during the subsequent period of
generally rising prices, the reverse was the case and in some instances
prices continued to rise during part of the contraction; in the 1920's,
the rise and fall were roughly the same; in the two postwar cycles the
rise was decidedly larger than the fall, as in the pre-1914 period.
Taken as a whole, these mild cycles would have imparted a generally upward drift to prices. The failure of such a drift to develop
during peacetime was a consequence of the more severe depressions
that occurred from time to time. In the five business cycles for
which the contractions were most serious and can be designated deep
depressions (1891-94,1904-08,1919-21,1927-33, and 1933-38), wholesale prices on the average rose about 10 percent during expansions,
about the same as in the mild cycles, but then fell during the contractions over twice as much, ending up on the average some 12 percent
below their level at the start of the cycle. It was the price declines
during these deep depressions that, as a matter of experience, offset
the upward tendency during mild cycles—"creeping inflation" in this
sense is by no means a unique post-World War I I phenomenon.
The stock of money shows the same relation to these cyclical price
movements as that depicted in figure 1 for longer periods. During
the mild cycles, the stock of money almost invariably rose during both
expansion and contraction, but at a faster rate during expansions than
during contractions. On the other hand, during the deep depression
cycles listed above, the stock of money invariably fell during the
course of the contraction, and there is only one other cycle during
which there is an appreciable absolute decline during any part of
the contraction (1894-97). This resemblance between the cyclical
movement in the stock of money and in prices holds not only on the
average but also from cycle to cycle, though of course with more variability for the individual cycles.11
11
One difference between the comparison made here and in the preceding section is that
the money series used is the stock of money, not the stock of money per unit of output.
The reason for this is the problem referred to in footnote 9 above. Over the longer periods,
the stock of money rises more rapidly than money income; an increase in real income per
capita leads to a more than proportional increase in real money balances—income velocity
falls with a rise in real income. Over the cycle, the reverse relation holds, if money income is measured by a figure like the regularly published national income or net national
product estimates. Money stock falls relative to measured money income during expansion and rises during contraction—income velocity rises during expansion and falls during
contraction. It turns out that this apparent contradiction can be accounted for, both
qualitatively and quantitatively, by distinguishing between measured income and a longer
term concept that I have called permanent income and also between measured prices and
permanent prices. One implication of this interpretation of the behavior of velocity is
that division of the money stock by measured national income in constant prices would
yield estimates of the stock of money per unit of output that were formally comparable
to those plotted in figure 1 but did not have the same significance and meaning; the latter
use an average output figure that is closer to permanent output or income than to annual
measured income. Unfortunately, full analysis of this issue is impossible within the confines of the present paper. The forthcoming annual report for 1957 of the National Bureau
of Economic Research will contain a somewhat fuller summary ; and the monograph referred
to in footnote 7 above, a full analysis.

ECONOMIC STABILITY AND GROWTH

249

There can be little doubt on the basis of this evidence that there is
a close link between monetary changes and price changes over the
shorter periods within which business cycles run their course as well
as over longer periods and during major wartime episodes. But
three important considerations must be borne in mind if this fact is
not to be a misleading guide to policy.
The first is that the direction of influence between the money stock
and income and prices is less clear-cut and more complex for the
business cycle than for the longer movements. The character of our
monetary and banking system means that an expansion of income contributes to expansion in the money stock, partly through inducing
banks to trim more closely their cash reserve position, partly through
a tendency for currency in public hands to decline relative to deposits;
similarly, a contraction of income contributes to a reduction or a slower
rate of rise in the money stock by having the opposite effects on bank
reserve ratios and the public's currency ratio. Thus changes in the
money stock are a consequence as well as an independent cause of
changes in income and prices, though once they occur they will in their
turn produce still further effects on income and prices. This consideration blurs the relation between money and prices but does not
reverse it. For there is much evidence—one important piece on timing will be presented in the next paragraph—that even during business cycles the money stock plays a largely independent role. This
evidence is particularly direct and clear for the deep depression periods. There can be little doubt, for example, that Federal Eeserve
action in sharply raising discount rates in January 1920 and again in
June 1920 (5 months after the onset of the contraction in January 1920) played an important role in the subsequent decline in the
money supply and unprecedentedly rapid fall in prices or that Federal
Reserve policy in the early 1930's played an important role in producing a decline of a third in the stock of money from 1929 to 1933—by
far the largest decline in the whole period covered by our data.12
A second, and perhaps more important consideration, has to do with
the timing of the changes in the money supply and in income and
prices. The generally upward trend in the money supply which
accounts for its continuing to rise, though at a slower rate, during
most contractions in economic activity as well as during expansions
makes it difficult to judge timing relations from ups and downs in
the money supply itself. For this and other reasons, we have found
it most useful to examine instead the ups and downs in the rate at
which the money supply is changing. The rate of change of the
money supply shows well-marked cycles that match closely those in
economic activity in general and precede the latter by a long interval.
On the average, the rate of change of the money supply has reached
its peak nearly 16 months before the peak in general business and
13
The other deep depression episodes are a bit more complex. The decline in the stock
of money from 1893 to 1894 seems connected with the uncertainty about silver; in 1907,
quite clearly with the banking panic which was of course in part a consequence of a prior
decline in economic activity but not through the particular channels described above and
which once begun very likely served as an important factor in making the contraction as
deep as it was; in 1937-38, with the doubling of reserve requirements by the Federal
Reserve System in two steps in 1936 and in 1937—the first step coincides with a sharp
reduction in the rate of growth of the money stock, the second with the beginning of
decline.

250

ECONOMIC STABILITY AND GROWTH

has reached 18
its trough over the 12 months before the trough in general business.
This is strong though not conclusive evidence for the independent
influence of monetary change. But it also has a very different significance. It means that it must take a long time for the influence of
monetary changes to make themselves felt—apparently what happens
now to the rate of change of the money supply may not be reflected
in prices or economic activity for 12 to 16 months, on the average.
Moreover, the timing varies considerably from cycle to cycle—since
1907, the shortest time span by which the money peak preceded the
business cycle peak was 13 months, the longest, 24 months;
the corresponding range at troughs is 5 months to 21 months.14 From the
point of view of scientific analysis directed at establishing economic
regularities on the basis of the historical record—the purpose for
which the measures were computed—this is highly consistent behavior;
it justifies considerable confidence in the reliability of the averages
cited and means that they cannot easily be attributed simply to the
accident of chance variation. But from the point of view of policy
directed at controlling a particular movement such as the current
recession, the timing differences are disturbingly large—they mean
that monetary action taken today may, on the basis of past experience,
affect economic activity within 6 months or again perhaps not for
over a year and 6 months; and of course past experience is not exhaustive; the particular episode may establish a new limit in either
direction.
The long time lag has another important effect. It leads to misinterpretation and misconception about the effects of monetary policy,
as well as to consequent mistakes in monetary policy. Because the
effects of monetary change do not occur instantaneously, monetary
policy is regarded as ineffective. The most recent example is the
tight money policy of 1956 and 1957 which coexisted with rising
prices but whose delayed effects are with us in the current recession.
A similar and even more dramatic example is the tight money policy
from early 1928 on and the associated lack of growth in the money
supply which coexisted with economic expansion but contributed to
both the occurrence and the severity of the 1920 downturn. The fact
that these policies had a delayed effect in turn misled the monetary
authorities; on those occasions, and even more clearly in 1920, they
were induced to believe that still stronger measures were required
and so tended to overdo a repressive policy. On other occasions, notably in 1932 as well as earlier in that major catastrophe, the failure
of tentative movements toward easy money to have an immediate effect
led them to regard their actions as ineffective and to permit and contribute to the sharp decline in the stock of money which occurred and
which played so crucial a role in that episode.
The third consideration is in some ways a different aspect of the
one just discussed. The variation in timing means that there is considerable leeway in the precise relation between changes in the stock
13
The average at peaks is based on 18 observations, on troughs on 19. Of course, instead of interpreting the cycles in the rate of change as conforming positively with a lead,
they could be interpreted as conforming inversely with a lag. A number of pieces of statistical
evidence, however, argue strongly for the former interpretation.
14
These are for the period since 1907 because our money data prior to that date are
annual or semiannual. While the annual and semiannual observations give the same
average timing as the monthly, individual observations are not comparable.

ECONOMIC STABILITY AND GROWTH

251

of money and in prices over short periods of time—there are other
factors at work that lead to these variations and mean that even if
the stock of money were to change in a highly regular and consistent
fashion, economic activity and prices would nonetheless fluctuate.
When the money changes are large, they tend to dominate these other
factors—or perhaps one might better say, they will force these factors to work in a particular direction. Thus there seems little doubt
that a large change in the money supply within a relatively short
period will force a change in the same direction in income and prices
and, conversely, that a large change in income and prices in short
periods—a substantial short-period inflation or deflation—is most
unlikely to occur without a large change in money supply. This is
certainly the conclusion suggested by the evidence for the deep depression cycles and for sizable inflations. But when the money changes
are moderate, the other factors come into their own. If we knew
enough about them and about the detailed effects of monetary changes,
we might be able to counter these other effects by monetary measures.
But this is Utopian given our present level of knowledge. There are
thus definite limits to the possibility of any fine control of the general
level of prices by afineadjustment of monetary change.
3. Changes in prices and changes in output over longer periods
Over the cycle, prices and output tend to move together—both tend
to rise during expansions and to fall during contractions. Both are
part of the cyclical process and anything, including a monetary
change, that promotes a vigorous expansion is likely to promote a
vigorous rise in both and conversely. The preceding section implicity assumes this connection.
Over the longer period, the relation between price changes and output changes is much less clear and in the first section we took the behavior of output for granted. Now this seems clearly valid, not only
as an expository device but also as a first approximation to reality.
What happens to a nation's output over long periods of time depends
in the first instance on such basic factors as resources available, the
industrial organization of the society, the growth of knowledge and
technical skills, the growth of population, the accumulation of capital
and so on. This is the stage on which money and price changes play
their parts as the supporting cast.
One proposition about the effect of changes in the stock of money
and in prices that is widely accepted and hardly controversial is that
large and unexpected changes in prices are adverse to the growth of
output—whether these changes are up or down. At one extreme, the
kind of price rise that occurs during hyperinflation seriously distorts the effective use of resources.15 At the other extreme, sharp
price declines such as occurred from 1920 to 1921 and again from
1929 to 1933 certainly produce a widespread and tragic waste of
resources.
So much is agreed. The more controversial issue is the effect
of moderate change in prices. One view that is widely held is that
35
However, even open hyperinflations are less damaging to output than suppressed inflations in which a wide range of prices are held well below the levels that would clear the
market. The German hyperinflation after World War I never caused anything like the
reduction of production that was produced in Germany from 1945 to the monetary reform
of 1948 by the suppression of inflation. And the inflationary pressure suppressed in the
second case was a small fraction of that manifested in the first.

252

ECONOMIC STABILITY AND GROWTH

slowly rising prices stimulate economic output and produce a more
rapid rate of growth than would otherwise occur. A number of reasons have been offered in support of this view. (1) Prices, and particularly wages, are, it is said, sticky. In a market economy, the
reallocation of resources necessitated by economic growth and development requires changes in relative prices and relative w^ages. It
is much easier, it is argued, for these to come about without friction
and resistance if they can occur through rises in some prices and wages
without declines in others. If prices were stable, some changes in
relative wages could still come about in this way, since economic
growth means that wages tend to rise relative to prices, but no
changes in relative prices, and, of course, there would not be as much
scope even for relative wage changes. (2) Costs, and in particular,
wages, are, it is argued, stickier than selling prices. Hence generally rising prices will tend to raise profit margins, giving enterprises both a bigger incentive to raise output and to add to capital and
the means to finance the capital needed. (3) The most recently popular variant of the preceding point is that costs are not only sticky
against declines but in addition have a tendency to be pushed up with
little reference to the state of demand as a result of strong trade
unions. If the money stock is kept from rising, the result, it is
claimed, will be unemployment as profit margins are cut, and also a
higher level of prices, though not necessarily a rising level of prices.
Gently rising prices, it is argued, will tend to offset this upward pressure by permitting money wages to rise without real wages doing so.
(4) Interest rates are particularly slow to adapt to price rises. If
prices are rising at, say, 3 percent a year, a 6 percent interest rate on
a money loan is equivalent to a 3 percent rate when prices are stable.
If lenders adjusted fully to the price rise, this would simply mean that
interest rates would be 3 percentage points higher in the first case than
in the second. But in fact this does not happen, so that productive
enterprises find the cost of borrowing to be relatively low, and again
have a greater incentive than otherwise to invest, and the associated
transfer from creditors to debtors gives them greater means to do so.
In opposition to this view, it has been argued that generally rising
prices reduce the pressure on enterprises to be efficient, stimulate speculative relative to industrial activity, reduce the incentives for individuals to save, and make it more difficult to maintain the appropriate structure of relative prices, since individual prices have to
change in order to stay the same relative to others. Furthermore, it
is argued that once it becomes widely recognized that prices are rising, the advantages cited in the preceding paragraph will disappear:
escalator clauses or their economic equivalent will eliminate the stickiness of prices and wages and the greater stickiness of wages than of
prices; strong unions will increase still further their wage demands
to allow for price increases; and interest rates will rise to allow for
the price rise. If the advantages are to be obtained, the rate of price
rise will have to be accelerated and there is no stopping place short
of runaway inflation. From this point of view, there may clearly be
a major difference between the effects of a superficially similar price
rise, according as it is an undesigned and largely unforeseen effect of
such impersonal events as the discovers of gold, or a designed result
of deliberative policy action by a public body.

ECONOMIC STABILITY AND GROWTH

253

Some who believe that slowly rising prices are adverse to economic
growth regard stable product prices with slowly rising wage rates
as most favorable, combining the advantages of stable price expectations with some easing of frictions involved in relative wage
adjustments. Others view gently falling prices and stable wages as
most favorable, arguing that additional problems in wage adjustments would be balanced by the stimulus to thrift and accumulation.
Historical evidence on the relation between price changes and output changes is mixed and gives no clear support to any one of these
positions. (1) In the United States, the period from 1865 to 1879
was a period of exceedingly rapid progress; and during the same
period, prices were cut in half. True, neither price changes nor
output changes proceeded regularly within the period. Output apparently grew most rapidly during the cyclical expansions in the
period when prices rose mildly or were roughly stable; most of the
price declines occurred during cyclical contractions. Yet the problem at issue is less the cyclical relation than the longer period relation and there can be no doubt that during the period as a whole
prices fell sharply and output rose sharply. (2) The period from
1880 to 1897 was a period of generally declining prices, from 1897
to 1913, of generally rising prices; taken as a whole, the second period
has generally been regarded as displaying more rapid growth than
the first. But it is not clear that this is a satisfactory interpretation.
The period of great monetary uncertainty in the early 1890's was
associated with generally depressed conditions and was followed by
a rapid rebound. If both are excluded, the remaining periods show
about the same rates of growth in real output per head, although
prices were generally falling during the 1880's and rising after the
turn of the century. Moreover, the period from 1908-14 is one of
relatively slow growth despite rising prices. (3) The decade of
the 1920's, after the recovery from the deep depression of 1920-21,
was a decade of rapid growth and prices were relatively stable. (4)
In Great Britain, output per head apparently grew at a definitely
higher rate during the period of generally falling prices before the
mid-1890's than16 during the subsequent period of rising prices up to
World War I.
(5) On the other hand, the attempt to achieve
mildly falling prices in Britain in the 1920's was associated with
considerable economic difficulties and something close to stagnation.
AH in all, perhaps the only conclusion that is justified is that either
rising prices or falling prices are consistent with rapid economic
growth, provided that the price changes are fairly steady, moderate in
size, and reasonably predictable. The mainsprings of growth are
presumably to be sought elsewhere. But unpredictable and erratic
changes of direction in prices are apparently as disturbing to economic
growth as to economic stability.
^. Policy implications
The preceding account of the relation of money to prices over long
and short periods and of price changes to output changes has some
fairly direct and immediate implications for public policy with respect
both to growth and stability.
16
See James B. Jefferys and Dorothy Walters, National Income and Expenditure of the
United Kingdom, 1870-1952, Income and Wealth Series V, table III.

254

ECONOMIC STABILITY AND GROWTH

(1) In order for the price level to be reasonably stable over the
decades ahead, the total stock of money will have to grow to accommodate itself to the growth in output and in population. In addition, if
past patterns continue, it will have to grow to satisfy the desire of the
public to increase the ratio of cash balances to income as their real
income rises. Past experience suggests that something like a 3 to 5
percent per year increase
in the stock of money is required for longterm price stability.17
(2) An essential requirement for the avoidance of either substantial
inflation or substantial deflation over the coming decades is the avoidance of a substantially more rapid or a substantially less rapid increase
in the stock of money than the 3 to 5 percent per year required for price
stability. A substantially more rapid rate of growth in the money
supply will inevitably mean inflation; conversely, continued inflation
of substantial magnitude cannot occur without such a large rate of
growth in the money supply. A substantially slower rate of growth
m the money supply, let alone an absolute decline, will inevitably mean
deflation; conversely, continued deflation of substantial magnitude
cannot occur without such a small or negative rate of growth in the
money supply.
(3) A highly fluctuating price level is as disturbing to economic
growth as to economic stability. Given that this is avoided, it is not
clear what pattern of long-term price behavior is optimum for economic stability—whether a roughly stable price level, a gently rising
price level, or a gently falling price level. It does seem clear that any
of these is consistent with rapid economic growth. If it is necessary
to state objectives in terms of a price level goal, then a stable price
level has the very great advantages of (a) ease of public understanding, (b) definiteness rendering successive alterations in the precise
goal less likely, and (c) probably the closest approach to equitable
treatment of the various members of the community. However, the
difficulty of assuring the close attainment of any price level goal suggests that it might be better to express the immediate policy goal in
terms'of some variable other than the price level, for example as being
the attainment of a steady 4 percent per year rise in the stock of
money, and then to let the price level be whatever would be consistent
with this money goal. The resulting price level behavior could hardly
depart much from relative stability and would certainly not be
violently unstable.
(4) For cyclical movements, a major problem is to prevent monetary changes from being a source of disturbance. If the stock of
money can be kept growing at a relatively steady rate, without erratic
fluctuations in short periods, it is highly unlikely if not impossible
that we would experience either a sharp price rise—like that during
World Wars I and I I and after World War I—or a substantial price
or output decline—like those experienced from 1920-21, 1929-33,
1937-28.
17
This range is for the stock of money as defined in footnote 2, namely, currency outside
banks plus adjusted deposits, demand and time, of commercial banks. For a narrower
definition, currency outside banks plus adjusted demand deposits, the required rate of
growth is less; for a broader definition, the preceding plus all time deposits, in mutual
savings banks and the postal savings system as well as commercial banks, the required
rate of growth is greater. The reason is that time deposits have been growing relative
to demand deposits and currency, and, until 1957, mutual savings deposits relative to other
time deposits.

ECONOMIC STABILITY AND GROWTH

255

(5) A steady rate of growth in the money supply will not mean perfect stability even though it would prevent the kind of wide fluctuations that we have experienced from time to time in the past. It is
tempting to try to go farther and to use monetary changes to offset
other factors making for expansion and contraction. Though the
available evidence demonstrates a close connection between monetary
change and price and income change in the course of business cycles
as over larger periods, it also casts grave doubts on the possibility of
producing any fine adjustments in economic activity by fine adjustments in monetary policy—at least in the present state of knowledge.
The evidence suggests that monetary changes take a fairly long time
to exert their influence and that the time taken varies considerably.
In terms of past experience, for example, action taken now to offset the current recession may affect economic activity in some 6 months
or not again for over a year and 6 months. The tight-money policy
of late 1956 and most of 1957, which was taken to offset the then existing inflationary pressure, almost surely had little effect on that situation and is only now exerting its influence and contributing to the
current recessionary tendencies; the inflationary pressures in 195(>
may well themselves have been in part a delayed consequence of the
expansionary monetary policy taken to offset the 1953-54 recession.
There are thus serious limitations to the possibility of a discretionary
monetary policy and much danger that such a policy may make matters worse rather than better. Federal Reserve policy since 1951
has been distinctly superior to that followed during any earlier period
since the establishment of the System, mainly because it has avoided
wide fluctuations in the rate or growth of the money supply. At the
same time, I am myself inclined to believe that in our present state of
knowledge and with our present institutions, even this policy has
been decidedly inferior to the much simpler policy of keeping the
money supply growing at a predesignated rate month in and month
out with allowance only for seasonal influences and
with no attempt
to adjust the rate of growth to monetary conditions.18
(6) To avoid misunderstanding, it should be emphasized that the
problems just discussed are in no way peculiar to monetary policy.
Fiscal action also involves lags. Indeed the lag between the recognition of need for action and the taking of action is undoubtedly longer
for discretionary fiscal than for discretionary monetary action: The
monetary authorities can act promptly, fiscal action inevitably involves serious delays for congressional consideration. It has been
argued that this defect of fiscal action is counterbalanced by a shorter
lag between the action and its effects. This may well be, though there
is little concrete empirical evidence that I know of; the belief is based
18
This is not intended to be a full statement of the optimum monetary structure. I
would prefer automatic arrangements that would reduce the area of discretion. One
particular set of such arrangements is suggested in my A Monetary Fiscal Framework
for Economic Stability, reprinted in my Essays in Positive Economics (University of
Chicago Press, 1953), pp. 133-156.
The extensive empirical work that I have done since that article was written has given
me no reason to doubt that the arrangements there suggested would produce a higher
degree of stability ; it has, however, led me to believe that much simpler arrangements
would do so also ; that something like the simple policy suggested above would produce a
very tolerable amount of stability. This evidence has persuaded me that the major
problem is to prevent monetary changes from themselves contributing to instability rather
than to use monetary changes to offset other forces.
On the issues in question, see also my 'The Effects of a Full Employment Policy on
Economic Stability: A Formal Analysis, reprinted in the same book, pp. 117-132.

256

ECONOMIC STABILITY AND GROWTH

on general considerations of plausibility, which can be a misleading
guide. And there are certainly no reasons for believing and no
empirical evidence to show that the lag, whatever its average length,
is any less variable for fiscal than for monetary action. Hence the
basic difficulties and limitations of monetary policy apply with equal
force to fiscal policy.
(7) Political pressures to "do something" in the face of either relatively mild price rises or relatively mild price and employment declines are clearly very strong indeed in the existing state of public
attitudes. The main moral to be drawn from the two preceding
points is that yielding to these pressures may frequently do more
harm than good. There is a saying that the best is often the enemy
of the good, which seems highly relevant. The goal of an extremely
high degree of economic stability is certainly a splendid one; our
ability to attain it, however, is limited; we can surely avoid extreme
fluctuations; we do not know enough to avoid minor fluctuations; the
attempt to do more than we can will itself be a disturbance that may
increase rather than reduce instability. But like all such injunctions,
this one too must be taken in moderation. It is a plea for a sense of
perspective and balance, not for irresponsibility in the face of major
problems or for failure to correct past mistakes.

INFLATIONARY DEPRESSION AND THE REGULATION
OF ADMINISTERED PRICES
By Abba P. Lerner, The Johns Hopkins University and Eoosevelt
University
Inflation, by which I mean a condition of rising prices, may be
the result of action either by buyers or by sellers. We are much
more familiar with inflation caused by buyers trying to buy more
goods than are available, that is, spending more money than can
buy (at current prices) the available supply of goods. When this
happens, prices are bid up to the level at which the buyers are no
longer trying to buy more than is available. The market is then
cleared with every buyer able to buy as much as he wants to buy.
If, as a result of this development, there arises a still further increase
in the amount of money spent by the buyers, perhaps because they
have received more money as sellers of something else, we have a
continuing inflationary process.
Such a process cannot go very far unless there is an increase in
the supply of money. Otherwise, with the rising prices, the public
finds that the stock of money is too small for the greater volume of
transactions, in monetary terms, that is going on. Many people then
reduce their buying or increase their selling (so as to hold on to or
to get hold of more money) and this, tends to stop the inflationary
process. But if the monetary authorities increase the supply of
money, or permit the supply of money to be increased, then the inflationary process can continue.
Because we are much more familiar with this particular type of
inflation, we have tended to assume that it is the only kind. This
has led to the habit of considering an increase in the supply of money
not as merely one of the necessary conditions for an inflationary
process to be able to continue, but as the cause of the inflation, which
it need not be. Our overoccupation with this particular type of
inflation has also led many economists, including myself, to use the
word "inflation" not only to stand for the condition of rising prices,
but also to stand for "excess demand," the attempt to buy more goods
than are available at the current prices, which is the cause of this
type of inflation.
This extension of the meaning of the word inflation would be quite
harmless if it were true, as it apparently was assumed to be true, that
rising prices could come about only as a result of excess demand by
buyers. This usage furthermore had the advantage of permitting
the condemnatory word inflation to be used for attacking a condition in which prices were prevented from rising, as by price controls,
when the economy would be better served if they were permitted to
rise. Such price control under conditions of excess demand could
then be called a kind of inflation—repressed inflation—which can be
257

258

ECONOMIC STABILITY AND GROWTH

even more harmful to the economy, and to society in general, than
an open inflation with rising prices. So it seemed like a good idea
to identify inflation with a condition of excess demand, whether the
resulting tendency for prices to rise was permitted to express itself
or not. Repressed inflation could therefore be called a certain kind
of inflation and given a blacker name, and this seemed harmless even
though it was something like calling an anti-Communist a certain
kind of Communist.
But excess demand by buyers is not the only possible cause of a
condition of rising prices. Prices may rise not because of the pressure of buyers who are finding it difficult to buy all they want to buy
at the current prices. Prices may rise because of pressures by sellers who insist on raising their prices even though they may find it not
especially easy to sell. We would then have not a buyer-induced
inflation but a, seller-induced inflation. To distinguish this from the
kind of inflation we have discussed above, and which we may call a
buyers' inflation (or demand inflation), we may call this kind of inflation a sellers' inflation.
If sellers' inflation is possible as well as buyers' inflation, it is not
such a good idea to use the word "inflation" to stand for excess demand. That use of language tends to suggest that if there is no "inflation" in the sense of excess demand, there can be no inflation in
the sense of rising prices. It leaves us with no way of describing the
kind of situation in which we find ourselves when prices are rising
because of upward pressure by sellers, and the authorities, in endeavoring to stop the rise in prices, have taken steps which have been very
effective in removing excess demand, but which have not removed the
upward pressure on prices from the sellers' side. Indeed such measures as budgetary restraint and tight money can be so effective in removing excess demand that they can overdo this and remove some
demand that is not in excess. They would bring about a condition of
deficient demand, or not enough demand to enable us to make full use
of our productive potential. Nevertheless, prices may keep on rising.
The net result would be both inflation and depression at the same
time—prices rising—even though we are not fully utilizing our available labor force and productive potential.
This appears paradoxical only because of our habit of using 1 word,
"inflation," to represent 2 different things, rising prices and excess
demand, that do not necessarily have to go together in the actual world.
The distinction between buyers' inflation and sellers' inflation is
related to but is not exactly the same as the distinction between demand
inflation and cost inflation. While demand inflation seems to be synonymous with buyers' inflation, cost inflation suggests that there is a difference between cost, on the one hand, and profits, on the other, in their
operation on price. This is especially true when the phrases "cost-push
inflation" or "wage-cost inflation" are used as synonymous with "cost
inflation." The impression is given that the whole of the blame falls
on labor or on trade unions. When trade unions raise wages bv more
than can be absorbed by increasing productivity, costs rise. The employer then seems to be completely innocent of "profit inflation" in passing on the increase in costs as long as he does not increase his rate of
markup, i. e., as long as he does not increase the prices he charges for
the product in a greater proportion than his costs have increased.

ECONOMIC STABILITY AND GROWTH

259

There is, however, no essential asymmetry between the wage element
and the profit element in the price asked for the product. A sellers'
inflation could just as well be started by an increase not in the wage
asked, but in the percentage of markup of price above cost. Prices
would rise and wages would then be raised by workers in attempts to
maintain (or restore) their original buying power. Business would
then "innocently" raise their prices again only in proportion to the
increase in their costs, and we would have the inflation upon us as well
as boring discussions about who started it first and the famous chicken
and egg.
The "who started it first" debate is a complete waste of time because
there is no original situation in which there was a "just" or "normal"
distribution of the product between wages and profits. Any increase
can be seen either as the disturbance which bears the full responsibility
for the inflation, or as nothing but the correction of an inequity perpetrated in previous history—all depending on the point of view. The
term "sellers' inflation," by treating wages and profits on exactly the
same footing, avoids the fruitless game of mutual recrimination. Sellers' inflation takes place whenever wage earners and profit takers together attempt to get shares that amount to more than 100 percent of
the selling price. When the sum of what they try to get comes to more
than 100 percent of the selling price it is futile to ask whether this is
because the wages demanded are too high or whether it is because the
profits insisted on are too great. No matter where justice may lie between the 2 claims, the only significant thing for our problem is that
the sum of the claims is more than 100 percent. That is what causes
the inflation.
It is, of course, impossible for the two parties to succeed in getting
more than 100 percent of the proceeds between them, but it is precisely
on an impossibility such as this that any continuing process depends.
Buyers' inflation is similarly built on an attempt to reach the impossible. In that case, it is the attempt of buyers to buy more than 100
percent of the goods than can be made available. Their attempt bids
up prices, but since that does not (and cannot) succeed in enabling
them to obtain more than 100 percent of the goods that there are
available to be got, they continue the attempt and we have the continuing process of buyers' inflation. In our case, the impossibility
that generates the process is the attempt of wage earners and profit
takers between them to get more than 100 percent of the money proceeds from the sale of the product. Each increases the part he tries
to take, by increasing wages or by increasing prices. Since they cannot succeed, they keep on raising wages and prices and so we have
the continuing process of sellers' inflation.
There is great resistance to recognizing the possibility of sellers'
inflation. Sometimes, this takes the form of saying that there must
have been some excess buyers' demand or prices could not have risen.
This begs the whole question. Since it assumes, without apparently
thinking it necessary to provide any support for the assumption, that
the only possible cause of rising prices is excess buyers' demand, the
argument assumes what it wants to prove.
A more sophisticated version of this argument points out that if
output shrinks by less than the increase in prices, and this is usually
the case during a sellers' inflation, there must have been an increase
2373 1—58

260

ECONOMIC STABILITY AND GROWTH

in the total amount spent in buying the output. The arithmetically
irrefutable increase in expenditure is then triumphantly exhibited
as the excess buyers' demand that is responsible for the inflation. Expenditure is the same thing as buyers' demand, but an increase in
expenditure is not the same thing as excess buyers' demand. An increase is not the same thing as an excess. An excess of demand by
buyers induces the price increases—it is the cause of the price increases. An increase in expenditure could be induced by—it could
be the result of the increases in prices brought about by the pressure
of sellers. If there is no increase in expenditure the number of units
of goods bought must fall in the same proportion as the price per unit
is raised by the sellers. A 10 percent increase in prices would thus
result in a fall in output of about 10 percent. This involves depression and unemployment that the authorities naturally seek to remedy
by monetary and fiscal measures. Such remedies all involve increases
in money expenditure, so that even if only a part of the unemployment
is corrected (and this is usually the case because of the authorities'
reluctance to undertake inflationary measures when prices are rising),
we would observe an increase in total expenditure. Buyers' demand,
however, instead of being excessive, could still be deficient, i. e., it
could still be insufficient to enable the potential output of the economy
to be sold (at the prices demanded by the sellers). An observed increase in total expenditure is therefore no proof that the price rise
is due to excess buyers' demand. The increase in expenditure could
have been induced by attempts by the authorities to keep down unemployment induced by price increases imposed by the sellers. In a
sellers' inflation, an increase in expenditure is perfectly compatible
with deficient buyers' demand.
A still more sophisticated argument along the same lines goes on
to claim that even if prices are being raised by the insistence of sellers
rather than by the pressure of buyers, the orthodox measures of reducing total demand would still check the inflation. By reducing total
expenditure, or perhaps by merely refusing to permit the increase in
total expenditure needed to accommodate the increased prices, the
authorities would bring about depression and unemployment. This
would stop the sellers from increasing prices. The question then resolves itself into how much unemployment would be necessary to
stop the sellers' inflation, and whether it is morally desirable or
politically possible for the authorities to induce or permit unemployment of the required volume and duration.
I t has been suggested that even if the authorities are not really
prepared to bring about the degree of depression necessary to negate
the pressure of sellers' inflation, they could still do the trick by solemnly announcing a policy of refusing to provide the increase in
expenditure called for by a continuing sellers' inflation. The threatened unemployment would then sober the sellers into calling off their
inflationary wage and price increases.
I t seems pretty certain first that such declarations would not be
believed and that the bluff would quickly be called. But, even if it
were believed as regards the economy as a whole, that would not
prevent any specific wages or prices from being raised while the local
conditions still permitted this. It would perhaps even aggravate the
wage and price increases as each tried to get his increase quickly,
while the local going was still good.

ECONOMIC STABILITY AND GROWTH

261

All this brings us to the perhaps only too obvious conclusion that
sellers' inflation cannot be cured or prevented by measures directed
against excess demand by buyers. It can be successfully treated only
by attacking the pressure on prices by sellers.1
Before we can consider just how one can attack the pressure on
prices by sellers, it would be desirable to get a perspective on the
whole problem by a quick look at the general theory of inflation and
deflation.
A somewhat schematic formulation of the development of thought
on this subject shows four theoretical models of the operation of the
economy.
Model A assumes perfectly flexible prices and wages, so that any
excess of buyers' demands makes prices and wages rise, and any deficiency of buyers' demand (through the unemployment that results)
makes prices and wages fall, until price stability and full employment are restored. Both monetary policy and fiscal policy are unimportant, or even unnecessary. As long as the volume of money is
kept fairly stable by some automatic device such as the gold standard,
the price level will automatically adjust itself so as to yield full
employment with price stability and without inflation.
Model B embodies the recognition that we do not have the degree
of price flexibility in the downward direction to make complete laissez
faire a satisfactory monetary and fiscal policy. Unemployment
(caused by deficient buyers' demand) does not reduce the wage and
price level quickly enough to the level needed to restore full employment. The process is rather complex. To achieve the task, unemployment must reduce the wage level, and thereby the price level, to
the degree necessary to increase the value of the existing stock of
money (as each dollar becomes more valuable) to the extent necessary
to increase expenditure in real terms (as each dollar spent constitutes
more real purchasing power) to the volume necessary to give a satisfactory level of employment. This process can last for years, during
which time prices and wages are falling as different resistances to the
reductions are gradually overcome. Meanwhile, there continues an
expectation of price and wage reductions still to come. This expectation induces investors as well as consumers to postpone their expenditures as long as prices are still falling, so that buyers' demand
is reduced still further and the depression can get very much worse
before it gets better.
The recognition of the nature of such a process leads to the abandonment of laissez faire in monetary and fiscal policy. Instead of
1
In an outstanding article which concentrates on showing the inadequacy and superficiality of proposals to prevent inflation by monetary and fiscal policies and declarations,
Prof. Sumner H. Slichter seems to suggest that the distinction between buyers' inflation
and sellers' in^ation is a futile fantasy. Thus he says (using a somewhat different
terminology), "Much time has been wasted in recent years in discussing whether inflation
is demand inspired or cost inspired. (Some 70 or 80 years ago, the Austrian theory of
value produced a similar debate as to whether demand or cost determines the value; the
argument ended suddenly when it dawned on the economists that each blade in a pair of
scissors cuts)" (On the Side of Inflation. Harvard Business Review, September/October
11)57, p. 32).
However, the inapplicability of this analogy jumps to the eye in his very next sentence
which shows that it can make sense to distinguish between the blades, since he goes on to
say. "Thus changes in the price level may originate either with shifts in the demand
schedules or with shifts in the supply schedules," and in another article. Professor
Slichter definitely alines himself with the sellers' inflation blade in declaring t h a t :
"There is no evidence that prices are rising ahead of costs and are pulling costs up.
The evidence is all the other way: that prices are being sluggishly adjusted to slowly
rising costs" (Government Spending Can Reduce Taxes. Harvard Business Review,
July/August 1957, p. 106).

262

ECONOMIC STABILITY AND GROWTH

waiting for the price level to fall until it has adjusted itself to the
volume of money expenditure, a policy is developed of adjusting the
volume of money expenditure to the existing price level, so as to reach
and maintain a satisfactory level of employment at the current prices.
This switch from laissez faire to an active monetary and fiscal
policy is also applied in the opposite direction to deal with excess
buyers' demand. Although there is not the same resistance to price
and wage increases as there is to price and wage decreases, the necessary adjustment to excess buyers' demand by rising prices still takes
time. It is no instantaneous adjustment (if only because of the existence of long-term contracts, and because of attempts to stop profiteering by preventing the necessary price increases) and so it causes disturbances that are unjust and reduce the efficiency of the economy.
The policy is therefore applied in both directions, providing for increasing the volume of money expenditure
whenever necessary to prevent or correct an insufficiency of buyers7 demand; and for decreasing
the volume of money expenditure whenever that is necessary to prevent or correct an excess of buyers' demand.
The volume of expenditure may be adjusted either by working on
the stock of money (by monetary policy) or by working on the
velocity of circulation of money (by fiscal policy), or by some combination of the two.
Model B, which is, of course, the Keynesian general theory of employment policy, differs from model A primarily in incorporating a
policy of increasing or decreasing demand, if it should become too
little or too great. (It has a steering wheel to keep the car on the
road.) Because of this difference, a secondary distinction arises.
With policy coming into the picture, it becomes important which of
two instruments of policy is to be used, monetary policy or fiscal
policy. Model B makes use of both instruments. (The car can use
either kerosene or gasoline.)
Model C is not really a new model. It rather consists of a series of
publicity releases of model B dolled up to emphasize one or another
of its qualities as if this were a new invention that made model B
obsolete. One very crude pamphlet of this series emphasizes the
ability of model C to cut down on demand, if it becomes excessive or
threatens to become excessive, seeming to imply that model B was a
depression model, which could work only in the direction of increasing demand, if it became deficient or threatened to become deficient.
(Model C has a steering wheel that can be turned to the right.)
A more refined variant of model C, let us call it model C*, is concerned with the relative effectiveness of monetary policy and of
fiscal policy in different circumstances. An economy may be so saturated with money so that further increases in the stock of money
would not be effective in increasing expenditure^ and reductions in the
stock would have no significant effect in reducing total expenditure.
(This is sometimes expressed, though not explained, by saying that
changes in the money supply would be offset by opposing changes in
the velocity of circulation.) Monetary policy is then useless and expenditure can be increased or decreased only by fiscal policy—by the
Government increasing or decreasing its own expenditure, e. g. on
public works, or permitting others to spend more by reducing taxes or
forcing them to spend less by increasing taxes.

ECONOMIC STABILITY AND GROWTH

263

It is then suggested that model B works only in this case which is
called the Keynesian case. It should more properly be called Keynesian special case (of the Keynesian general theory) when it is
appropriate to concentrate entirely on fiscal measures to increase or
decrease expenditure on consumption and investment. (Only gasoline can be used.)
In this kind of situation, even extreme price flexibility is unable to
restore or maintain the desired level of real demand, because it operates, after all, as nothing but a roundabout way of increasing or
decreasing the real volume of money in terms of buying power. It
is a kind of automatic monetary policy which is useless for the same
reasons as other monetary policy is useless, when the economy is so
saturated with money that changing the quantity has no appreciable
effect.
When the economy is at the other extreme from being saturated
with money, and money is very tight, the situation is naturally reversed. Fiscal measures for increasing expenditure on consumption
or investment are ineffective, because an increase in expenditure anywhere in the economy, say in Government expenditure on public
works, results in an increase in demand for money to hold in connection with the increased volume of transactions. In the very tight
money situation, this raises the rate of interest, or in some other way
reduces expenditure somewhere else. Similarly, a decrease in expenditure anywhere releases holdings of money which permit an increase of expenditure somewhere else. Fiscal policy then is helpless,
and what is called for is monetary policy to increase or decrease the
money supply. (Only kerosene can be used.) This case is then called
the Classical Case, as if it were one in which the Keynesian theory
does not apply and where model B should be replaced by model C*
(which can burn kerosene). This case should more properly be
called the Classical Special Case (of the Keynesian general theory).
The Keynesian theory (model B) covers both situations in which
fiscal policy is strategic (when model B uses gasoline), and situations
in which monetary policy is strategic (when model B uses kerosene),
as well as the more normal situations when both policies are effective
{when model B can make use of both fuels, mixing the proportions
to suit the terrain).
Model D is a genuinely different model, in which unemployment
not only fails to make prices and wages fall quickly enough to serve
as a cure for the unemployment, but is even unable to prevent prices
and wages from continuing to rise. When we have strong trade
unions with the power to raise wages, strong corporations with the
power to set prices administratively, and a general atmosphere in
which it is considered normal, natural and only fair for wages to be
increased regularly, and by amounts greater than the average increase
in productivity or in the share of the product that labor can obtain,
prices increase, and the economy is subject to sellers' inflation. It is
now no longer a question of whether fiscal policy or monetary policy
is more effective m regulating the volume of buyers' demand or expenditure, since the inflation is caused not by excess buyers' demand,
but by the existence of powerful institutions and mores that enable
sellers to insist on and obtain continually higher prices. The widespread and generous feeling that workers are entitled to the increases

264

ECONOMIC STABILITY AND GROWTH

in wages that they get is made much easier by a, recognition that any
raise need not be taken out of profits, since it is possible, as well as
proper, to "pass it along" to the ultimate purchaser in higher prices.
Indeed, it is usually considered only right that profits, in dollars,
should be increased so as to protect real profits from the declining
value of the dollar.
We have already mentioned the argument that a really firm refusal
on the part of the monetary authorities to prevent the volume of
money from increasing, no matter what happened, would bring the
sellers to their senses. Eealizing, or discovering, that they will not
be able to sell so much if they raise their prices, they will refrain from
raising prices, and they will not grant, or ask for, wage increases that
raise costs by more than can be squeezed out of profits.
There are several reasons why this is not practical. In the first
place, the policy of firmly or obstinately holding the money supply
constant does not prevent excess buyers' demand from coming about.
It does not even prevent an increase in total expenditure. This is
because the policy of holding the money supply constant is essentially
a kind of monetary policy, and we may be in the Keynesian special
case where monetary policy is not effective. That we are at the
present time in such a situation is suggested by the fact that, while the
supply of money has been held fairly stable in recent years, the
volume of expenditure has continually increased. (Another way of
expressing this, which is more common perhaps because it sounds like
an explanation, is to say that the velocity of circulation has increased
and that this has frustrated the restrictive monetary policy.)
There is, of course, a limit to the degree to which expenditure can
increase without an accompanying increase in the money stock, and
if the inflation were a buyers' inflation it would come to an end when
this limit was reached (i. e., when the velocity of circulation could not
increase any more). But where the inflation is a sellers' inflation, it
does not stop at that point. After the increase in prices has absorbed
all the increase in expenditure that is compatible with a constant
money stock (i. e., that can be attributed to an increase in the velocity
of circulation) it continues to increase. The increase in prices goes
on further until it has reduced real expenditure and employment sufficiently to overcome the institutional forces that enable sellers to demand higher and higher prices. The question is how strong are these
institutions ? or, in other words, how severe a state of depression and
unemployment would have to be maintained in order to destroy these
institutions or to induce sellers not to use their power to raise prices;
and how able and willing the authorities would be to bring about and
maintain this degree of depression and unemployment ?
The continuing increase in wages and prices in the present depression would be some indication that it would require quite a severe and
prolonged depression to change people's notions of what is the proper
development of wage rates (and of the corresponding prices, since the
right of wages to increase goes together with the right of profits at
least not to fall). It would take perhaps an even more severe level of
unemployment to destroy the power of labor to force the wage increases on more reluctant employers who grant wage increases only
when they feel they are forced to—i. e., that they would lose more from
the strikes and other weapons of the trade unions than they would lose
by agreeing to the higher wages (and passing them on).

ECONOMIC STABILITY AND GROWTH

265

At the same time, a policy of full employment seems to have won
a firm place in the country's economic policy (even though its application may be rather shaky), not only because of the general acceptance
of the desirability of prosperity, for human as well as for international
political reasons, but because neither political party can afford the
blame for even a mild depression. With such a setup, there is no need
to worry whether the cure is worse than the disease—whether the depression would be more harmful than the inflation that it would prevent. This cure is not one that any government would carry out or
even seriously attempt to carry out.
None of the problems of sellers' inflation or of inflationary depression could arise in a perfectly competitive economy, because in a perfectly competitive economy, we cannot have the institutions and mores
that give sellers the power to push prices uj). In a perfectly competitive economy, all that is needed for stability of the price level is a
monetary and fiscal policy to keep buyers' demand from becoming
either excessive or deficient. No one holds back any product from the
market—or can establish a price which results in some of the potential
product or the available labor not being taken off the market, so that
unless there is excess buyers' demand, prices cannot rise, and if there
is a deficient buyers' demand, prices must fall. Unless there is full
utilization of resources, we cannot have inflation, and if there is a
depression (or recession), we will have deflation (i. e., falling prices).
In a perfectly competitive economy, we cannot have inflation and depression at the same time.
But where prices are administered by decrees of large firms, and
wages are administered by joint decrees of powerful unions, together
with powerful employers or employer groups, the situation is different. Sellers' inflation is a byproduct of the process; and, together
with sellers' inflation, we can also have depression—indeed we will
have depression with our sellers' inflation, just to the degree that the
authorities try to cure the inflation by reducing ("excess") demand.
In an economy where there are both administered and competitive sectors, the phenomenon of sellers' inflation can spill over from
the administered to the competitive part. It can even happen that the
contagion of sellers' inflation in the competitive sector is more pronounced than in the administered sector. There is then a tendency to
assume that the sellers' inflation thesis has thereby been disproved.
Actually, this does not prove anything either way in the debate that
can rage as to whether the inflation is a sellers' inflation or a buyers'
inflation.
The contagion can be explained as follows. Prices and wages being
raised in the administered sector but not in the competitive sector,
there will be a switch in demand from the products of the administered
sector to the products of the competitive sector. There is then a deficiency of demand in the administered sector and an excess of demand
in the competitive sector. With factors of production immobile, there
is unemployment in the administered sector, but there it does not cause
either prices or wages to fall so that the unemployment persists. Attempts to reduce total spending, so as to check the rise in the overall
price level, would increase still further the unemployment in the administered sector (while removing some, or all, of the excess demand
in the competitive sector). Pressure is then put on the Government

266

ECONOMIC STABILITY AND GROWTH

to alleviate the depression; and, in doing so, it must create enough demand to maintain the higher price level in both sectors.
As the economy gets used to such a process, in which wages and
prices are rising all the time, an increase in strategic or key prices or
wage rates in the administered sector come to be recognized all over
the economy as presaging a general rise in prices. The competitive
sector then does not wait for the excess demand to appeal. Its workers
demand higher wages, its employers expect to be able to get the higher
prices out of which to be able to pay the higher wages, and they grant
the increases and raise the prices. They do not have to go into the
calculations of what output and elaborate price maximize profit. They
have the businessman's rough rule of thumb of a more or less traditional markup on their cost. This brings them straight away to the
position that would be reached after the excess demand has materialized and has been validated and adjusted by the monetary policy undertaken by the authorities to cure or prevent the unemployment
threatened by the increase in wages and prices in the administered
sector.
The economist is tempted to draw diagrams showing the point of
maximum profits of a firm, competitive or monopolist, and to demonstrate, in classical vein, that an increase in wages will move that point
to the left, reducing the optimum output of the firm and causing the
firm to restrict output and to raise the price by less than the increase
in cost. This should cause unemployment which, in the competitive
sector, would restore the previous price and wage levels. The sellers'
inflation has disappeared into thin air.
The answer to this, in classical vein, is that the demand will not
remain the same, because the phenomenon is not happening only to
an individual firm (in which case it would be proper to assume the
conditions of demand to be unchanged), but that the monetary and
fiscal authorities, in providing additional overall demand to cure or
prevent the unemployment in the administered sector, will raise every
demand curve so that the firm will be able to sell as much as before
and provide the same employment as before, even though the price is
sufficiently higher to enable the higher wage to be paid. Profits, or
the gap between cost and price, will also be higher, of course, although
in real terms, allowing for the fall in the value of the dollar, everything will be just the same as in the beginning.
The answer in the businessman's language is that he has to increase
his price in proportion to the increase in costs, in applying his regular markup; and his experience is that since this is happening to everybody, including his competitors, and employment in the country is
more or less being maintained, he will be able to pass it on to the consumer.
Although the infection starts with the administered sector of the
economy, there is no reason why the epidemic should not hit the
competitive (and nonadministered monopolistic) sectors of the economy sometimes more severely and sometimes less severely than it hits
the administered sector. This is why th$ observation that prices rise
more or rise less in the unadministered sector than in the administered
sector proves nothing at all either way as to whether the inflation is a
buyers' inflation or a sellers' inflation But only a sellers' inflation is
compatible with a depression.

ECONOMIC STABILITY AND GROWTH

267

The inflation and the inflationary depression that result from administered wages and prices have important similarities to, and are
no less socially harmful than, the monopolistic exploitation that would
result from the administration of excessive prices by public utilities.
We have gone a long way toward eliminating the latter evil by the
regulation of prices that may be set by public utilities for the services
they supply. The same kind of device can be used to eliminate the
former evil. Just as the public utility prices can be and are being
regulated so as to prevent monopolistic exploitation, so administered
prices and wages can and should be regulated, so as to prevent sellers'
inflation and the depression it may bring with it.
The regulation of administered prices and wages so as to prevent
sellers' inflation would have to follow somewhat different lines. It
would not be directly concerned as to whether there is more than or
less than a fair rate of return on investments. That would be left to
the strong competitive forces that still prevail in our economy. Nor
would any other regulations whatsoever be involved other than price
regulation. The function of the regulation here proposed would be
only to prevent restrictive prices or wages from being administered.
A restrictive price is one that results in the demand for a product
falling below capacity output. A restrictive wage is one that results
in less than full employment in the specific labor market to which it
applies. With a monetary and fiscal policy concentrating on the maintenance of adequate buyers' demand for full employment at a constant price level, while preventing buyers' inflation, it would be
possible for wages per hour to rise on the average at the same rate
as productivity per hour, with aggregate profits rising too at the
same pace as aggregate wages and aggregate output, (except that increases in the degree of competition, which might be induced, could
reduce the share going to profits and increase the share going to labor).
The regulatory body would therefore have to follow a set of rules
which would do the following things:
(1) They would permit an administered price increase only
when production and sales are at capacity. Such price increases
should not be withheld on account of profits being high.
(2) They would enforce decreases in administered prices
whenever production and sales are significantly below capacity.
A price decrease should not be waived on account of profits being
low, or even negative on this item in the firm's output, as long
as the price more than covers current operating costs (more
strictly short period marginal costs).
(3) They would permit increases in administered wages in
general at a rate equal to the average trend of increase in national
productivity.
(4) They would permit increases in administered wages
greater than this wherever the labor market is tight—with say
less than half the national average rate of unemployment.
(5) They would permit only smaller increases in administered
wages, or no increases at all, where the labor market is slack—
with say more than twice the national average rate of unemployment. (The expected continuing increase m product per head
makes it possible to avoid reductions in money wages although it
is unavoidable, for price stability, that some prices must fall if
others rise.)

268

ECONOMIC STABILITY AND GROWTH

This is, of course, not a fully worked out solution ready for immediate application. There remains much to be developed—such as
generally acceptable criteria of capacity of different firms and industries and generally acceptable measures of slackness or tightness in
particular labor markets, or measures for dealing with possible attempts by monopolistic industries to restrict the installation of
capacity, if they are prevented from restricting the utilization of
existing capacity. (This would bring out the existence of a specific
monopoly situation that calls for treatment quite apart from the
problem of inflation.) The intensification of competition which the
regulation would enforce would also, in some instances, lead to the
elimination of high cost competitors. While the public would benefit
from the increased efficiency of the economy—in higher wages and
lower prices—such elimination of competition would conflict with
certain existing so-called antitrust policies that have become in effect
anticompetition policies and need to be reconsidered.
There remain also important problems of organization and administration of the regulatory body, as well as the need for widespread
and intensive public discussion to bring about the familiarity with,
and the understanding of, the nature of the proposed regulation which
is essential for its effective operation in a democracy. And in the
course of such examination and debate, important developments,
changes and improvements are to be expected. Nevertheless,
the general lines indicated seem to be inevitable if sellers5 inflation is to be
attacked directly and if we are not to depend on irrelevant nostrums
or pious exhortation because we do not dare to attack the problem at
its roots.

THE DESTABILIZING INFLUENCE OF RAW MATERIALS
PRICES
Ruth P. Mack, National Bureau of Economic Research, Inc.
Spot prices and other prices of many raw materials fluctuate in a
sensitive and extreme fashion. They are at the tail end of the price
structure and swing widely. The question that I want to explore
is the extent to which the tail wags the dog and, particularly, thereby
helps to drive him uphill. To put the matter in another way, is it
possible that the response of all prices to swings in sensitive prices
tends, under conditions that presently characterize our economy, to
foster an upward trend in the price level ? It seems highly probable
that the answer is "Yes."
SUMMARY OF THE ARGUMENT

Over the past decade and in earlier years, as well, the prices of
manufactured goods have tended to show an upward trend relative to
those of the raw materials of which they are constituted. Increases
in labor costs do not account for the growing spread. Doubtless,
more extensive processing, packaging, and increased distributing
costs, and certain social costs borne by the manufacturer, have filled
the gap.
But in spite of their diminishing importance over considerable
stretches of years, it seems probable that the gyrations of raw materials have contributed materially both to the fluctuation and to the
upward trend in the price level. The reason is that, on the one
hand, their extreme fluctuations are transmitted to later stages of
production. On the other hand, the price of fabricated goods responds more agilely to a rise in crude materials prices than to a fall.
Sensitive materials prices, in other words, operate on the price level
like a rachet jack.
Change in the price of crude materials, like any cost, affects the
price of the fabricated product. It also affects other prices by influencing expectations about prospects for prices and delivery conditions. The resultant change in short term demand-supply relationships affects prices at the later as well as at the earlier stages of
production through which the basic product passes to the final consumer.
It is the contention of this paper that these two avenues whereby
the price of crude demand-oriented materials influence other prices—
the avenues of cost and of their implication in buying or inventory
waves—join to cause rises in crude prices to be promptly and liberally
reflected in rises in other prices. However, the rigidities of other
aspects of the cost structure, along with common interpretations of
the influence of price reductions on short-term demand, cause the
response to falling sensitive prices to be far less wholehearted.
269

270

ECONOMIC STABILITY AND GROWTH

In view of the built-in resistance to downward pressures, reduction
of upward movements is in the public interest. The analysis of process
has implications about how such reduction might be achieved.
THREE CHARACTERISTICS OF THE TEMPORAL ASSOCIATION OF RAW- AND
PROCESSED-GOODS PRICES

Chart I shows an index of spot-market prices of 13 industrial commodities. They include scrap prices for copper, lead and steel scrap,
tin, zinc, print cloth, cotton, wool tops, burlap, hides, tallow, rosin,
and rubber. Just below it is the Bureau of Labor Statistics' price
index for wholesale prices of crude materials other than foods and
fuels. This index covers the many such items priced by the Bureau
and weights them according to their sales to manufacturers. The
third line is the Bureau of Labor Statistics' index of wholesale prices
of intermediate materials. The fourth covers all manufactured commodities other than farm and food products.
Correspondence in cyclical fluctuation
The chart shows that many of the fluctuations in any one series are
also found at about the same time in the others; these common movements are marked by an X placed at their peak and trough months.
(Other not generally shared reversals in direction are marked by O,
and are discussed later.) The correspondence lessens as the eye moves
down the chart, but even where the locations of maximums and minimums are not found at the same time in raw and in processed goods
there seems to be a good bit of almost immediate response at the
intermediate stage to price changes of crude materials. The association is clearer in rate of change. For individual commodities, of
course, the association is considerably closer in spite of the fact that
price indexes of partly finished or finished goods are notoriously imprecise images of the prices (including allowances for changes in
terms and quality) at which sales are actually made.

CHART I. INDEXES OF WHOLESALE PRICES OF COMMODITIES AND MANUFACTURED GOODS OTHER THAN FOODS (1947-49=100)

13 Spot Market Conmodities

I
a

to
19W

19^9

1950 1951

1952 1953

195^ 1955 1956 1957

272

ECONOMIC STABILITY AND GROWTH

Greater amplitude of fluctuation in crude-materials prices
The chart also shows that spot prices move more widely than all
crude non-food-materials prices and a great deal more widely than
wholesale prices of goods that have undergone manufacturing processes. If, for each series, we add the falls from the peak to the subsequent trough to the rise from the trough to the subsequent peak
(turns marked with X on the chart), the average rise or fall per
month for spot prices is 2.4 percent of the level in 1947-49. The
corresponding figure (excluding foods throughout) for crude materials is 1.2; for intermediate goods, 0.6; for all manufactured goods,
0.4; and for finished manufactured goods, 0.3.
Crude materials other than foods constituted in 1947 about 6.9
percent of the value of the sales of all industries other than food
industries. If we exclude the duplication of sales among manufacturing industries, crude materials constitute about 18.5 percent
of shipments other than foods to the final-demand sectors such as
export, Government, gross private capital formation and consumers.
The relatively heavy fluctuation of crude prices can account for a
very large proportion—an average of 57 percent—of the cyclical
fluctuation of the price of finished
manufactured goods other than
foods during the 11 years, 1947-57.1
Upward trend in processed—relative to crude-materials prices
Over stretches of years there has been a tendency for prices of fabricated products to rise more (or fall less) than those of the crude
materials of which they are made. Information for two periods may
be reviewed.
Continuous indexes, those shown in chart 1, are available beginning
January 1947. They show price trends after the immediate disruptions of wartime controls and decontrols have spent their force. From
January 1947 to December 1957 the index of spot-market prices
showed a net fall of 15 percent of the 1947-49 level. (Taken to its
peak at the end of 1955, there was a net fall of 10 percent.) All
crude material other than food and fuel rose 16 percent of their 194749 level, labor cost per unit of output rose about 15 percent and, even
if calculated along a trend line that covers 25 years (on the assumption that employers tried to adjust for wartime changes), rose only
about 27 percent of the 1947-49 level during the last 11 years. The
price index of all manufactured goods other than foods, on the other
hand, rose 34 percent of the 1947-49 average, and of finished manufactured goods other than foods, 33 percent.
1
1 am indebted to Beatrice Vaccara for providing the estimates of the proportion that
crude materials constitute of gross and net manufacturers' shipments, as explained in the
text. To estimate the percentage of cyclical fluctuation in finished-goods prices that can
be due to fluctuation in crude materials, I add the cyclical rises to the falls ("rises" and
"falls" are marked by X in chart 1) in crude materials prices and apply it to the crude
sector—18.5 cents on 100 cents' worth of finished goods. This accounts for 57 percent
of the rises and falls (analogously marked) in the price of finished manufactured goods
other than foods. This figure for the whole period averages very different relations for
each expansion and contraction. For contractions, the figures are a great deal higher
than for expansions.

ECONOMIC STABILITY AND GROWTH

273

The differential trend of raw and manufactured goods prices after
World War I I was not the result of 1 or 2 freak movements, nor
of differences in the date at which turns occurred, nor of the behavior of labor costs. Comparing movements that are shown as
matched on chart I, the ratio of rises in spot-market prices to corresponding ones in margin over labor costs 2 is always larger during contractions than during previous and following expansions in
spite of the fact that expansions in manufactured prices were usually
longer than in raw materials and contractions shorter. This was also
true when crude materials, rather than spot prices, were compared
with margins. Another way of describing these relations is to say
that when spot prices or crude-materials prices rose, manufacturers
raised prices more, even after allowing for the probable increase in
labor costs, than they lowered prices when materials prices fell, assuming that here, too, they would wish to allow for the movement of labor
costs. Failure to allow for labor costs increases the difference.
It would be useful to test whether the same tendency appeared for
matched pairs of prices for particular industries. But I have not
tried to assemble these comparisons for the postwar decade. There
is, however, an interesting group of well-matched prices for raw
products and finished-goods prices for 18 industries, 1913 to 1935.
These annual figures were collected by Solomon Fabricant many years
ago. They, like the aggregate data for the recent decade, exhibit the
tendency for finished prices to rise more over the years than the price
of the major constituent, crude materials.
Figures are given in table I. They make rough estimates of trends
by using trough-to-trough measurements. Ratios over 1.00 in columns 2, 4, and 6 indicate the presence of the lesser trend change of
crude materials. The second column shows that this differential trend
occurred in 15 of the 18 industries over the 20 years between about 1913
and 1933. Labor costs could not have been responsible, since cost per
unit seems to have been about the same at the two dates. Columns 4
and 6 show that the same differential trend was present in 15 or 16 of
the 18 industries when additional trough-to-trough measures are
used to break the period into change during and after World War I.
2
The allowance for labor costs was based on changes in a 12month moving average
of production-worker payrolls per unit of output as estimated in Productivity, Prices, and
Incomes, materials prepared for the Joint Economic Committee by the committee staff,
table 52. In determining margin, it was assumed that labor cost constituted 40 percent of
the total cost of manufacturing.

274
TABLE

ECONOMIC STABILITY AND GROWTH

I.—Trend in prices of manufactured goods compared to that in prices of
crude materials, 18 industries, 1913-38
Prewar trough to
1932-33 trough i

Prewar trough to
postwar trough i

Postwar trough to
1932-33 trough i

Ratio,
Ratio,
prices of
prices of
Percent manufac- Percent manufac- Percent
change tured goods change tured goods change
to crude
to crude
materials
materials
(1)
Bread and products
Flour and grain mill products
Butter, cheese, and processed milkMilk, raw at farm
Meat packing products
Livestock and poultry
Sugar, granulated
Sugar, raw
Tobacco pr- ducts
Tobacco leaf
Cotton goods
Cotton, raw
__.
Woolen and worsted goods
Wool, raw
Silk hose
Silk, raw
Boots and shoes
Hides and skins
Auto tires and tubes
Rubber, crude
Paper
Wood pulp
Fertilizers, mixed
Fertilizer materials
Petroleum products
Petroleum, crude
Linseed oil and meal
Flaxseed
Simple processed iron and steel
Pig iron
Copper products
Copper ingots.
Lead pipe
Pig lead
Zinc sheet
Zinc slab

140.0
55.0
77.5
70.9
89.6
63.1
93.7
83.5
128.0
51.7
118.9
64.2
113.9
60.6
64.4
45.6
158.0
39.4
26.4
5.4
154.0
82.8
76.5
84.2
66.3
121.8
103.2
87.6
128.1
131.7
64.8
41.6
115.3
82.0
117.1
61.7

(2)

(3)

2.55
1.09
1.28
1.12
2.48
1.85
1.88
1.41
4.01
4.89
1.86
.91
.54
1.18
.97
1.56
1.41
1.90

171.0
119.5
137.4
137.0
123.8
113.6
139.2
133.3
176.8
169.7
198.9
148.2
178.5
107.1
157.1
177.2
180.0
83.8
74.4
25.4
180.0
129.0
141.2
130.6
158.7
281.8
151.7
137.1
155.1
177.6
104.3
94.2
125.6
117.6
107.3
97.5

(5)

(4)

1.43
1.00
1.09
1.04
1.04
1.34
1.67
.89
2.15
2.93
1.40
1.08
.56
1.11
.87
1.11
1.07
1.10

81.9
46.0
56.4
51.8
65.1
55.5
67.3
62.6
72.4
30.5
59.8
43.3
63.8
56.6
41.0
25.7
87.8
47.0
35.5
21.3
85.6
64.2
54.2
64.5
41.8
43.2
68.0
63.9
82.6
74.2
62.2
44.1
91.8
69.7
109.1
63.3

Ratio,
prices of
manufactured goods
to crude
materials
(6)

1.78
1.09
1.17
1.08
2.37
1.38
1.13
1.60
1.87
1.67
1.33
.84
.97
1.06
1.11
1.41
1.32
1.72

1
The prewar trough was located at the lowest figure reached by each index, 1913 to 1915, inclusive. The
postwar trough wasjocated at the lowest figure reached by each index, 1921 to 1922.
Source: National Bureau of Economic Research; compilations of annual price indexes, 1913 to 1935.

T H E CAUSAL ASSOCIATION B E T W E E N PRICES OF CRUDE, D E M A N D - O R I E N T E D
M A T E R I A L S A N D PRICES OF PROCESSED GOODS

The fact that raw materials have tended to rise less over the years
than the prices of fabricated goods may seem to absolve them from
responsibility as a potential inflationary force. But to do so is, I
believe, to misunderstand the process whereby raw, particularly spot
market, prices affect other prices. We live in a world in which there
are few free competitive markets in the sense defined in Economics IA.
Most prices, at least for processed goods, respond to other forces as
well as to those of free competition. Businessmen make explicit
decisions about them; product prices are in part "administered."
Changes in crude materials prices bear on the character of decisions
about product prices. In the context of this investigation, the point
of chief interest is that they are likely to bear in a somewhat different
way when prices are going up and when they are going down. To
describe the character of this difference, it is necessary to concentrate

ECONOMIC STABILITY AND GROWTH

275

on sensitive prices for which changes in demand rather than extreme
changes in supply are the primary source of usual market variation.
A large proportion of basic industrial materials are of this sort.
Characteristically, prices tend to rise and fall with rises and falls in
orders and output. To see how this fact, and some others that I shall
mention, bear on pricing strategy, it is necessary to consider two chief
ways in which materials prices can affect product prices.
Crude materials as a cost
Raw materials are first and foremost a cost of production. When
raw materials prices rise, the price of the goods in which they are incorporated must also rise if profits are not to be reduced. The amount
that they need to rise depends on the relative importance of raw materials in the cost structure and what is happening to other costs.
Say materials are 30 percent of the total cost of an article that sells
for a dollar. Suppose other costs do not change and materials rise by4
10 percent. Then profits can be shielded by a rise of only 3 percent.
I want to underscore the character of this calculation; it is the absolute change in a cost that must be recovered, not its percentage change.
The change in finished prices that is likely to accompany a given
rise or fall in materials prices depends in part on whether other costs
are likely to be rising or falling at the same time. In recent years,
labor cost seems to have risen, at least over a good part of the time,
when materials prices rose, though its relation to cyclical fluctuation is
not very systematic. But for goods whose basic materials prices parallel changes in demand, one type of cost does bear a systematic relation
to the price of materials—the overhead or burden type of expense.
Since these expenses are not a function of the months' or often even
the years' output, they tend to represent a larger cost per unit when
output is low than when it is high. Industry has also tended to add
to other costs when opportunity arose—those associated with further
processing of product, packaging, research, marketing and the like.
These facts suggest that it might be wise strategy to raise finished
prices, when raw materials prices and demand rise, by more than
the sum of the increase in materials cost plus any increase in actual
labor costs and minus the decrease that had been experienced in actual
per unit burden and overhead expense. The extra recovered margin
would then provide for expected further increase in labor cost and
changes in product, in distribution and the like—changes believed to
strengthen the competitive position of the company. When materials
prices and output decline, the fact that overhead expense per unit is
increasing is a reason for prices not to decline by the absolute amount
of the decline in materials.
I have analyzed how sensible business pricing policy might generate a stipulated association of changes in the price of crude and of
processed materials. But why should the asymmetrical association
to rises and to falls in materials costs be attributed to fluctuation in
materials prices any more than to the character of the other costs
and the drive to improve product and strengthen competitive
advantage?
Of course, it is foolish to isolate any one factor in a total situation
and say it governs the rest. I claim merely that sensitive demand4

Ten percent of 30 cents is 3 cents.
23734—58
19

Three cents is 3 percent of a dollar.

276

ECONOMIC STABILITY AND GROWTH

oriented raw materials prices play a more important role in the causal
nexus than their position as cost justifies. One reason for this is
their publicity value. Since, these days, pricing policy is defended
at the labor bargaining table, at the lecturn and in the press, publicly
reported costs—those of materials prices—that are rising swiftly
provide a valuable talking point.
But a more important reason for assigning materials prices a strategic role in this process is a second avenue of impact of these prices
on those of processed goods.
Materials prices and buying waves
The prices of demand-oriented sensitive raw materials, particularly
prices quoted daily on open markets, provide a basis for expectations
about market conditions. The expectation involves delivery conditions, selections, and price, and, typically, all three factors change
together. A change in the price of a sensitive crude material often
bespeaks changes in supply conditions in markets in which the crude
material, semifinished goods made from it, and often even finished
goods are bought by manufacturer or distributor. The appropriate
response to an expectation that prices will rise, deliveries lengthen,
or selections thin is that of extending the number of weeks' supply
of materials held on hand and on order. A fall in the price of crude
materials can have the opposite implications at any or all stages of
a vertical sequence of production operations.
Needless to say, the buying associated with optimistic expectations
influences the price of the raw product just as changes in its price,
via the expectations based upon it, influence buying. Moreover the
process as a whole is affected by conditions of final demand. These
processes, like most social processes, are partly circular. The point
is simply that this circular process is one that links sensitive prices,
via buying waves for intermediate goods, to the prices of manufactured goods. Price fluctuation in raw commodities is increased by
the buying waves and fluctuations of finished prices are also thereby
increased. I shall come next to why the upward fluctuations may
well ba more augmented than the downward one, but, first, can one
demonstrate the association of sensitive prices and buying waves?
As evidence on this point I submit three charts. The first applies
to hide prices in the interwar years. The monthly course of hide
prices was "predicted" by two variables: (1) Shipments of leather
to leather-goods manufacturers and (2) a ratio depicting what may
be thought of as a tug-of-war between buyer and seller to acquire inventories of hides and of leather. I call it the "stock-location ratio";
purchased and in-process inventories (leather in the hands of leathergoods manufacturers and raw or in-process hides in the hands of
tanners) are divided by finished stocks (leather in the hands of tanners and hides in the hands of packers and dealers). The ratio depicts the efforts—stronger at some times than at others—of customers
to buy stocks away from sellers. The calculations indicate that these
efforts have an important impact on hide prices, just as expectations
about hide prices must influence buying behavior.

CHART II. CONTRIBUTION OF EACH TWO VARIABLES TO THE ESTIMATION OP HIDE PRICES, 1922-39 1
Dollars ptr hide

Dollars per htdt

I
8

1922 1923 1924 1925 1926 1927 1928 1929 1930
Estimated prices are computed by the following formulas developed by
the method of least squares :
1922-31 : P=— 3.98 + 6.089 i2+1.657#
1932-39 : P=— 2.46 + 2.534 12+2.4435
where P is price per hide in dollars
R is the stock-location ratio
S is leather shipments by tanners in millions of equivalent hides.

1932 1933 1934 1935 1936 1937 1938 1939
Coefficients of multiple correlation a r e : 1922-31, 0.86 ; 1932-39, 0.81.
Regression coefficients exceed their standard errors of estimate by between
6 and 18 times for the two values of the two variables.
Source: Consumption and Business Fluctuations: A Case Study of the
Shoe, Leather Hide Sequence, by Ruth P. Mack, National Bureau of Economic Research, 1956, pp. 225, 226.

1931

278

ECONOMIC STABILITY AND GROWTH

Similarity in the course erf stocks, orders, and prices is also indicated in charts I I I and IV. Chart I I I makes comparisons for a
group of semidurable industries. In the first 2 lines, an index of 9
spot prices may be compared with orders received by the manufacturers but not yet shipped to their customers. Unfilled orders, of
course, increase as buyers try to provide further ahead; they decline
as buyers return to a hand-to-mouth position. The third line is the
stock-location ratio—purchased materials and in-process stocks of
semidurable-goods manufacturers divided by their finished-goods
stocks. Chart IV shows just the same three series for manufacturers
of primary metals. In general, the three series seem to show a strong
similarity within the industry sequence and. sharp differences when
the two sets of prices are compared.5 Similar differences apply to
the two sets of unfilled orders and to the stock-location ratios. Note
how the semidurable series all decline throughout 1948 whereas metals
do not decline sharply until the beginning of 1949. The 1949 trough
is earlier and sharper in the semidurable group than in two of the
metals series. The post-Korean peak is sharp and occurs early in 1951
in semidurable manufacture. In metals it spreads out through much
of 1952 as the huge defense contracts stimulate the metals industries,
though price ceilings affected the course of prices in 1951. The
trough in both sets of data seems to locate in 1954, but the nondurable
group rises to a peak in 1955 and starts to decline, whereas metals continue to rise well into 1956. It seems proper to conclude that the
charts suggest an intimacy between buying waves and the most sensitive prices in the vertical sequences of industries concerned. I have
little doubt that finer analytic tools and deeper understanding of the
process would make it possible to sharpen the demonstration.
5
To some extent the price series is an element in the other two, but its influence could
not be responsible for the basic similarities. The spot prices included in the price index
apply at most to only a small part of the purchased stock since these are purchases at all
stages of manufacture and of all materials that are used. The raw materials deflators
calculated by the Department of Commerce have far gentler contours than the spot price
indexes. In the stock-location ratio, the presence of a price element in both numerator
and denominator further reduces the impact of prices on the course of the ratio.

CHART III.

PRICES AND INDICATORS OF BUYING WAVES IN SEMI-DURABLE-GOODS

INDUSTRIES

9 Spot Market Prices
(scale at left)
if t o -

ffrr-ft

1
M

Unfilled Orders (scale at right)

3000
MIL

— 2000

o
Stock-Location Ratio (scale at left)

\
19^6

19^9

1950

1951

1952

1953

195^

1955

1956

to
1957

280
ECONOMIC STABILITY AND GROWTH

ECONOMIC STABILITY AND GROWTH

281

Asymmetry in the response to rises and to falls of crude prices
I come now to why sensitive materials prices seem to have a special
responsibility in furthering an upward trend in price. The reason
involves why their upward fluctuations are more willingly reflected in
selling prices than their downward ones.
The causal association between prices of crude, demand-oriented
materials and prices of processed goods involves materials as an element in cost, that is, the relation of rises and falls in materials costs to
the total cost structure. It concerns also the association between sensitive prices and buying waves for those commodities for which supply
factors are not subject to abrupt autonomous change—demandoriented sensitive prices. Changes in prices arouse expectation
about further change; buying intended to lengthen or shorten market
positions results at any or all of the vertically related stages; buying
affects prices. Thus the buying wave, more popularly called an inventory cycle, is inextricably involved with sensitive materials prices.
It is involved in several ways.
The first way involves the effect on demand of a change in price.
We know from Economics A l that the maximization of utility implies
that a rise in price alienates demand and a fall stimulates it, other
things the same. The previous section indicated how "other things"
undergo systematic shifts because of price-associated expectations. A
man will buy materials to lengthen his market position when he expects markets to tighten; he will buy less when he expects them to
slacken. My point now is a further one: he is more willing to pass on
the rise than the fall in materials prices to his own customers because
he expects the demand of his customers to be, like his, stimulated by an
expected rise and discouraged by an expected fall in price (providing,
of course, he believes that his competitors, having experienced the
same buying problems as he has, will handle their selling problems in
the same way). Thus he has two reasons for raising prices—the fact
that costs have increased; the fact that he expects that short-term
demand will, if anything, be stimulated rather than repelled.
The impetus to pass on a fall in costs ordinarily comes more from
the pressure of the buyer than from the spontaneous wish of the
seller. This is true particularly since the argument of the previous
paragraph applies in reverse—a price reduction in and of itself can
be expected to discourage expectation-related demand. But the sellers' resistance to reducing prices may be paralleled by less than insistent pressure to do so on the part of his customers. When tight
markets have continued long enough for stocks of purchased materials to have built up, the owner of the stock is wary of setting a
downward spiral in motion by forcing price concessions from his supplier. The accounting law of cost or market, whichever is lower,
means that loss on inventories will follow a reduction in materials
prices. It may be better strategy to decrease buying and use up stocks
before pushing for lower prices.
The argument suggests that prices of manufactured goods would
lag sensitive materials prices more persistently at peaks than at
troughs. The scanty number of observations covered by appropriate,
readily available price statistics shows manuf actured-goods prices lagging sensitive prices at each of the four major peaks that the data

282

ECONOMIC STABILITY A1SID GROWTH

cover6 whereas at troughs, they lead once, synchronize once, and lag
once.
The line of thought that emphasizes the link between sensitive
prices and buying waves and the asymmetrical influence on pricing
policy for rises and falls should be subject to test. One implication is
that were it possible to hold other relevant factors constant, commodities having a high degree of cyclical variability of raw prices
should tend to have finished-goods prices that have a trend rise far
greater (or fall far less) than that of the price of the major raw
materials of which they are composed. Small differential trends
would be associated with low cyclical variability. The only data that
could readily be used for the test were the 18 industries for which
Fabricant had prepared price indexes for two or more stages of fabrication.
These figures seem to exhibit the sort of correlation between variability and differential trend that one would expect, though the te$t
can only be thought
of as highly tentative and requiring further
substantiation.7
PROBLEM AND THERAPY

In industries in which short-run changes in the prices of basic raw
materials respond primarily to changes in demand rather than to
sharp autonomous changes in supply, there is a triple causal association between materials prices and product prices. First, materials
prices are a business cost which, like other costs, must be recovered in
selling price. Second, the fact that many crude materials prices are
promptly and publicly reported gives them an instrumental value in
recovering actual and expected increases in labor costs or increases in
other costs which are expected to improve the competitive position
of the company. Provision for increased complexity of product, packaging, marketing, research, are cases in point. Third, materials prices
are linked with buying waves, and the associated inventory cycles, as
cause and effect.
These several causal links between prices of crude materials and of
products are relevant to the subject of this inquiry in two ways. In
the first place, materials prices impart cyclical instability via all three
causal links—enough to account for a large part of the cyclical fluctuation in the price of finished goods. The three causal links contribute to this result in various ways. As cost alone, the high per« Comparisons are for 1937, 1939, 1948, 1949, 1951, 1953, and 1957. The data used
a r e : Bureau of Labor Statistics price indexes: all manufactured goods prices (excluding
crude materials) since 1947 and combined semimanufactured and finished goods prices,
1935 to 1946. For this earlier period, sensitive prices were represented by an index of
spot-market prices of 28 commodities and beginning in 1947 by spot-market prices of
industrial commodities (shown in chart I ) . The lag ( + ) or lead (—) of manufactured
prices relative to sensitive prices i s : at peaks + 5 , + 8 , + 1, +28, and troughs 0, + 5 ,
or 7—10.
For the 15 industries in which raw-materials prices rose less or fell more 1913 to 1932
than did finished prices, average year-to-year differences in price indexes of raw materials,
adjusted for the trend component, give a rough measure of cyclical price variability. We
rank these figures from greatest to lowest variability and compare the rank standings for
each industry with a second set of rank standings that are based on the spread between
raw and finished goods prices between the trough in 1913, 1914, or 1915 to the trough in
1932 or 1933. Of course, other variables that ought to be controlled include differences
among industries with respect to the trends in productivity, importance of materials as
a cost, and peculiarities of supply conditions; institutional characteristics that influence
the capacity to administer prices and the objectives to be thereby served. None of these
matters could be controlled; but two industries, sugar refining and linseed oil and meal
manufacture, had very few firms and low materials costs. Omitting these, the rank
correlation for the 13 industries was 0.60 which is significant at the 5-percent level.

ECONOMIC STABILITY AND GROWTH

283

centage fluctuation of crude prices is watered down because they constitute only a fraction of all costs. Also the slow leaching of costs
through the productive process delays the impact of raw on finished
prices. Their involvement with buying waves, on the other hand,
increases the promptness and extent to which changes in crude materials prices reappear in other prices, since the short-term shifts in
demand-supply conditions that accompany buying waves have a direct
impact on prices not only of crude materials but also on prices of
intermediate and finished products.
Second, materials prices act as agent in facilitating an upward
trend in the price level. Here again all three causal links play a part,
but that of instrumental value and, especially, buying wave is particularly important. The impact of buying waves on the advantage
of buyer and seller during periods of tightening demand is such as
to further a rise in product prices relative to materials prices which
is proportionately greater than the relative fall dictated by the advantage of buyer and seller during periods of hesitant or slackening
demand.
The line of thought that I have outlined suggests that crude materials prices is an advantageous point at which to take steps to reduce
cyclical fluctuations in prices and to curb their upward trend. What
steps may be advisable ?
The direction of the cure is implicit in the analysis of the process.
It involves the damping of buying waves. This would reduce fluctuations in the price of demand-oriented crude materials. It would
reduce the impact of fluctuation in materials on later stages of manufacture. It would reduce the ratchet-jack aspect of the process.
Though we know far less than we need to know about the alternating over and under buying of stock in trade, it seems clear that
one element that generates this alternation is fear of scarcity. In
connection with certain materials, it may be feasible to diminish
fear of scarcity by the presence of reserve stocks handled in the interests of price stability.
But perhaps the most promising attack on actual scarcity is to
lessen imagined scarcity which tends to validate imaginings. It is
therefore reasonable to suppose that better and more complete information about stocks and orders—new and outstanding orders—would
reduce surges and cessation in buying.
Specifically, we need information on orders received and orders
placed by the same firms; we need information on unfilled orders and
orders outstanding of the same firms. We need this information for
vertical sequences that are continuous from the first to last processor.
Information in physical units is more useful than in dollar units
but is often too difficult to obtain. This committee has shown active
interest in businessmen's expectations about sales and inventories.
Yet businessmen are not clairvoyant and the the value of orders they
receive and of unfilled orders on their books must certainly be a major
determinant of expectation about sales and stocks. Order information is often much easier to get and much more reliable than information on expectations. The several inquiries about statistics instigated
by this committee and executed by the Federal Reserve Board 2 years
ago were unable to cover data on orders; unfilled orders were not covered by the inventory committee and new orders were not systematically covered by the Gainsbrugh committee on expectation. It

284

ECONOMIC STABILITY AND GROWTH

seems distinctly in the public interest to improve such information
both because of its value in forecasting general economic conditions
and because it would help industry in its own interest to lessen fluctuation.

RELATIONSHIPS BETWEEN FOREIGN AND AMERICAN
PRICES
Gardner Patterson, Princeton University
The task assigned me was briefly to survey the problem of the
interrelationships among American and foreign price levels from the
long-term point of view. I was specifically asked not to deal with
the immediate international problems arising from the present United
States recession. My concern is with the problems arising from a
persistent uneven pace of inflation among nations, and both "persistent" and "uneven" are important qualifications.
To properly assess the relationships between foreign prices and the
American economy one must have some notion as to the importance
of our international economic transactions. There is no simple measure of this. On the most general level, it can be said that as compared with many domestic economic activities the numbers are relatively small but that they are not so insignificant as to be safely ignored
by those formulating policies to cope with the problems facing this
committee.
I . RELATIVE IMPORTANCE OF INTERNATIONAL ECONOMIC TRANSACTIONS

Excluding military aid, our exports of goods and services have in
recent years been running in the neighborhood of 5 to 6 percent of
our gross national product, or some 9 percent of our movable goods
production. Our imports (f. o. b.) have been equal to from 3 to 4
percent of our gross national product, or equal to some 6 percent
of the estimated value of movable goods produced in the United
States. Looked at from another angle, it has been estimated that
about 7 percent of our labor force gain their livelihood, in one way
or another, from foreign trade. For many sectors of the economy,
foreign trade—and so price changes abroad compared to those at
home—is much more important than these overall averages indicate.
For example, foreign sales in recent years have, on the average, accounted for about 20 percent of our production of trucks, 25 percent
of our output of construction and mining equipment, 10 percent of
our machine tools, and from 25 to 40 percent of such agricultural
products as cotton, wheat, rice, fats and oils, and tobacco. For some
industries, on the other hand, exports are of very little importance:
printing and publishing, clothing, manufacturing, food processing,
and so forth. Similarly, in many areas the importance of imports is
much greater to our economy than the above overall averages suggest.
We are now depending upon foreign sources for about a quarter of
1
For more details on most of the points in this section, see U. S. Congress, Committee
on Ways and Means, Foreign Trade Policy, Compendium of Papers on United States Foreign
Trade Policy (1957), pp. 15 ff. See also U. S. Department of Commerce, Bureau of
Foreign Commerce, annual bulletins on Exports in Relation to United States Production,
in its World Trade Information Service, and, in the same Service, Contribution of Imports
to United States Raw Material Supplies.

285

286

ECONOMIC STABILITY AND GROWTH

our consumption of iron ore, a third of our copper and fish, half of
our raw wool and sugar, and most or all of our tin, nickel, lead, zinc,
bauxite, chrome, manganese, cobalt, and so forth, to say nothing of our
coffee and bananas.
From the aggregative point of view, it may be said that, by and
large, the greater the excess of exports over imports the greater net
financial boost or push (not always desired) our economy gets in the
first instance from its foreign trade. The size of our current account
export surplus—the excess of exports of goods and services over imports, or the excess of our foreign sales over our purchases from
others—varies from year to year. Since 1950 it has averaged, excluding military aid, less than 1 percent of gross national product and
was almost exactly 1 percent during 1957. This is a small percentage.
But in absolute amounts, again excluding military aid, it averaged
$2.5 billion for the period 1950-57, and was over $4 billion in 1957.
That is to say, last year this was only a little less than most of the
recent (March 1958) estimates of the anticipated decline in 1958, as
compared with 1957, in the value of business investment in plant and
equipment about which there has, quite properly, been much serious
concern of late.
Perhaps this is enough to indicate that there are important structural and financial relationships between the American and foreign
economies. Moreover, it seems likely that these interrelationships and
interdependencies will increase in the years ahead. As our income
goes up, the variety and quantity of goods which we demand grows
and so our interest in foreign products; the evidence seems quite conclusive that foreign sources of raw materials will become increasingly
attractive as our production grows and as the more cheaply exploited
domestic sources are used up. Corollarily, the volume and value of
our exports, especially of manufactured goods, also seem likely to
increase as foreign incomes improve. These forces making for growth
of trade seem likely to be more powerful than the drives toward protectionism which have been such a prominent part of our landscape
in recent years.
I I . TRANSMITTING PRICE CHANGES ACROSS FRONTIERS

The effects on one nation of price movements in another, and the
ways in which these effects are transmitted, are so complex that they
are not known with precision. At the outset, it should be noted that
price changes in some goods or services in one country may have virtually no discernible impact on prices elsewhere. This is the case for
price changes in those many items which do not move across frontiers—haircuts, house rents, and so on. There may, of course, even
here be indirect repercussions, especially through wage rates, but they
are likely to be small. Probably more significant are the cases where
price changes in one country serve to offset or counteract other economic changes and so are not transmitted in any clearcut way. Thus,
an increase in the prices abroad of goods we import may have little
impact on our economy if it is matched by a decrease in the cost of
transportation. Or a decrease in prices abroad may largely be offset
by a rise in our tariffs. Moreover, it may often be that the markets
for certain goods are so poorly organized, or the goods so differentiated, that great divergences in prices can take place for quite some

ECONOMIC STABILITY AND GROWTH

287

time as between nations before consumers shift their sources of supply
or before the new prices make themselves felt abroad in other ways.
Our chief concern here is with general price levels, not just selected
individual prices, and it can be said that any substantial change in the
price level in one country will have a noticeable impact on other
countries. For purposes of illustration, assume that prices of most
goods—and so the general price level—rise abroad faster than they do
in the United States; as we shall see later, this seems to have been the
common pattern in the past decade or so. To some degree these price
rises will be transmitted directly to Americans by the simple fact that
we are likely to continue to import at least some of these goods and the
cost per unit, or price, goes up. If the price rise abroad is great
enough we may reduce the quantity of our imports and rely more on
domestic producers. But under these circumstances the domestic
producers can usually obtain a higher price than before and so, here
too, prices have increased in the United States. Another aspect of
this is that since rising prices tend to lessen the danger or force of
import competition, one of the moderating influences on price increases
at home is removed. But this is not all. If foreign prices go up
faster than ours, there is a tendency for buyers all around the world to
shift to lower-price sources; this will mean an increase in demand for
many United States goods and so more pressure is applied to our
prices.
A bit more complicated, but fully as important, is the transmission
of inflation, by which term I mean a general rise in prices, via what
is called the "income effects." As inflation mounts in one country,
money incomes also tend to rise. As money incomes rise, the money
demand there for all sorts of goods increases. Included as the object
of this enlarged demand are likely to be at least some of the domestic
goods formerly exported plus all kinds of imports. More specifically,
still assuming the pace of inflation abroad is greater than ours, this
means an increase in our exports and a decrease in our imports. Now
if we have lots of unemployed resources, these developments will give
a welcome boost to our economy and may well increase total production and employment. But if we are already fully employed—and
are ourselves already suffering from rising prices, though more moderate than other areas—these changes will aggravate our inflation.
The actual ability of others to buy more and sell less depends on the
previous balance of trade, on the state of their reserves, on the availability of unilateral transfers and loans, etc. If the first were unfavorable, the second low, and the last very limited, then the situation
cannot long continue. But this state of affairs would raise some important policy issues for us which will be noted later. It was suggested
earlier that changes in the overall inflationary pressure (wanted and
unwanted) put on one economy by inflation in another via these income
effects can roughly be measured in the first instance by the changes in
the size of the export surplus; the absolute inflationary impact on the
more stable economy over time, however, will be some multiple of
this surplus, inasmuch as a large part of the additions to its national
money income represented by the surplus of foreign sales over purchases will be spent again and again inside the country and only
a part will "leak out" through domestic savings, higher taxes, and
larger imports.

288

ECONOMIC STABILITY AND GROWTH

Inflation also is sometimes quickly spread among nations via the
speculative channel. This is most likely to affect countries for which
raw materials bulk large in imports (as the United States) or in exports (as many of the underdeveloped countries). Once the prices
of raw materials start rapidly upward, for whatever reason, the fact
of increase tends to encourage users to accumulate inventories, these
goods typically being storable. The supply of most raw materials
tends to be price inelastic, which means that any such increase in demand tends to have a very great effect upon prices. Thus both the
importing and the exporting countries tend to inherit inflationary
pressures from one another. Similarly, of course, once raw material
prices start to fall they tend to fall far and fast.
The international transferral of inflations also takes place via its
effects on a nation's terms of trade. An increase in the price of imports as compared with exports—a deterioration in a nation's terms
of trade—means that the nation suffering such an adverse movement
must increase the physical quantity of her exports to maintain the
previous quantity of imports. This means (unless the worsening
terms of trade are the result of at least a proportionate increase in
productivity in the production of the exports) that, as compared
with the previous situation, there are fewer goods available for purchase inside the country and this, of course, is inflationary. This
phenomenon has had great quantitative significance for England in
recent years, especially after Korea. It has been less important,
though not negligible, for the United States inasmuch as our relative
dependence on imports is much smaller.
I I I . RECENT PATTERNS OF RELATIVE PRICE MOVEMENTS

If inflations do spread across frontiers, as has just been argued,
then it is important to know whether, in fact, in recent years inflation has proceeded at different rates in different countries, and, if so,
where the United States has stood on the ladder.
Price data are not available for some countries and international
comparisons of price movements are fraught with possibilities of
error, stemming from the wide differences of items covered, gross
inaccuracies in observation, incomparability of many items whose
prices are compared, the irrelevance of differences in prices for certain
comparable goods and services, and so forth. Nonetheless, readily
available data for most of the countries which now supply 1 percent
or more of United States imports do show such great differences in
the relative price movements that one can conclude, in the light of
what has been said above, that a source of problems for the United
States exists here.
The actual comparison varies much, depending on the year chosen
for the base of the price indexes used. In the early postwar years,
1945-48, as the United States abandoned its controls more quickly
than other nations, prices here very often rose faster than in many
foreign countries where price controls were retained. For the decade
since then, however, a substantial majority of countries appear to
have more inflation than the United States, as the table below shows.
That is, for the years 1948 through 1953 these data indicate the
United States was subjected to serious inflationary pressures from

289

ECONOMIC STABILITY ANI> GROWTH

abroad. Since 1953 the rate of inflation has tapered off in most countries and the degree of difference has declined, but inflation still
continues and it continues imevenly.
TABLE I.—Comparison of domestic inflations, autumn 1951
1953=100

1948=100

Wholesale
prices
Chile
Argentina
Indonesia
Brazil
Japan
Peru
Colombia
Mexico
Australia
Prance

1,880
372
291
274

_ _

TTnftp.fi K"ingr|nrn

Sweden
Netherlands
West Germany
Canada
Belgium
United States
India _. _
Switzerland l
Philippines
Italy
Venezuela
Cuba

205
204
183
159
150
143
117
115
113
113
108
103
103
99
97

Cost
of
living
1,650
590
347
342
195
209
206
200
201
185
153
147
152
117
127
115
118
115
111
104
132
120
90

Wholesale
prices
Chile
Brazil
Argentina
Indonesia _..
_
Colombia
Peru
_ ._
Mexico
France
Sweden
_ __
_ _
Australia
United Kingdom
Netherlands _ _ _
United States
Belgium
Denmark
Japan
West Germany.
_ _. __
Switzerland
Italy
Canada-.
_.__._ _
India
Philippines
Cuba -_
Venezuela

696
195
160
136
135
119
108
108
107
107
107
106
105
105
105
105
103
101
101
101
100
99

Cost
of
living
636
212
182
163
140
127
143
111
113
113
118
117
106
109
113
111
108
107
114
107
106
106
97
102

i 1950=100.

Source: Calculated from data published in United Nations, Monthly Bulletin of Statistics, various 1957
issues.

To us in the United States foreign prices have meaning only when
translated into dollars and this requires a foreign exchange rate. A
good many foreign countries have devalued their currencies during
the past decade and so the above prices do not accurately reflect the
prices to us in terms of dollars. Indeed, often devaluation was necessary in large part precisely because other countries had more inflation than the United States. Thus, another very rough indication
that most other countries have suffered more inflation than we have
is the fact that of the countries included in table I above all save Canada, Cuba, the Philippines, Switzerland, and Venezuela have officially
reduced the value of their currency in terms of the dollar at some time
or other since 1948. But it is also noteworthy that only the following
have officially cut the dollar value of their moneys since 1953: Bolivia,
Brazil, Chile, Colombia, France, and Mexico. Table I I takes account
of these official devaluations and expresses foreign wholesale price
indexes in terms of United States dollars.
According to this calculation, the extent of inflationary pressure
which was probably applied to the United States from abroad is much
less than table I indicated. However, it must be remembered that
most countries today follow a so-called "adjustable-peg" exchange rate
policy; that is, their exchange rates are not allowed to fluctuate from
day to day or even month to month in response to market forces, but
are held fixed over long periods of time and then, if found inappropriate, undergo major adjustment in one fell swoop. This means that
the impact of relative differences in degree of inflation tends to be
felt for a relatively long period before the rate is adjusted.

290

ECONOMIC STABILITY AND

GROWTH

TABLE II.—Comparison of wholesale prices, expressed in United States dollars,
autumn 1957
[1948 = 100]

Japan 1
Brazil
Australia
Peru 2
West Germany 2
Canada
France
United States
United Kingdom 2
Philippines
Switzerland
Sweden
Colombia3
Netherlands
Belgium
Venezuela
Chile 4
Mexico
Cuba 5
Italy
India

[1953=100]

290
185
142
134
123
123
121
113
104
104
103
102
99
99
99
98
97
95
90
89
78

Brazil *
Peru
Chile
Sweden
France
Australia
United States
Netherlands
United Kingdom
Belgium
Canada
West Germany
Switzerland
Japan
India
Philippines
Italy
Cuba
Venezuela
Mexico 3
Colombia

123
120
117
109
108
108
107
107
106
106
106
105
105
104
103
102
100
100
100
95
94

1
"Coffee export rate" used in converting domestic prices to dollar basis.
2
1950 = 100.
3
Index of "Home and Import Goods" prices. Exchange rate used was "Other principal
export rate."
*5 Average exchange rate of 40 pesos per dollar used for 1948 and 110 for 1953.
Cost of living index.
Source : International Monetary Fund, International Financial Statistics, December 1957,
various tables.

Another common indication of relative price movements is provided by the so-called terms-of-trade index, the ratio of the prices of
the commodities a nation exports to the prices of the commodities
it imports. A commonly used measure of this is to divide the unit
value of exports of United States merchandise by the unit value of
United States imports for consumption, the unit values being those
calculated by the United States Department of Commerce. These
calculations are based on actual prices paid and received, rather than
on general price indexes, and take into account changes in the exchange
rate. Using 1936-38 as equal to 100, the index shows a fairly steady
deterioration from 1945 through 1951 and then a slight improvement,
with a fairly stable relationship since 1954.
Again, the possibilities of errors are great and the recorded decline
could be the result of a series of factors other than more domestically
generated inflation abroad than in th'e United States. For example,
important changes in the kinds of goods imported or exported could
cause changes in the index even if prices for each good remained constant. Or, the higher prices for imports may have been the result of
an increase in American demand for the goods. Still, when considered
along with the other evidence this does lend support not only to the
thesis that prices have moved unevenly among nations but also to the
hypothesis that during much of the postwar period the United States
has imported some inflation.

ECONOMIC STABILITY AND GROWTH

291

TABLE III.—United States terms of trade index—Unit value of exports of United
States merchandise divided oy unit value of United States imports for consumption
[1936-38 = 100]
100 1951
97 1952
107 1953
91 1954
88 1955
85 1956
83 1957 (January-September)
74

1936-1938
1940 .
1945.
1946
1947
1948
1949
1950

71
74
71
72
72

Source : Calculated from data in U. S. Department of Commerce, Business Statistics, 1957
edition, and Survey of Current Business, January 1958.
IV. WHAT ARE THE PROSPECTS FOR THE FUTURE ?

What is likely to happen is always more interesting than what has
happened and it is to that we now turn. In my judgment, most other
countries are likely to witness steady and serious inflation in the foreseeable future. Whether this will be at a faster rate than in the
United States depends on our internal policy, and I leave consideration of this to others. My point is only that if our objective is to
have stable prices our policies must make allowance for some, probably moderate but unremitting, inflationary pressures from abroad.
For some of the industrialized countries with whom we have contact, these inflationary pressures will arise in part from the defense
burdens they have assumed. Much more important, I anticipate, will
be the conjunction of (1) a vigorous pursuit by these governments of
full employment policies, (2) the existence of strong and aggressive
labor unions, and (3) the fact that much of the public now generally
believes that by energetic use of monetary and fiscal measures any unwanted slack in demand or employment can be successfully combated.
If, as seems to be the case in many countries, the politically and socially acceptable levels of unemployment are now very small, then
tremendous upward pressure will be exerted on prices for the many
well known reasons that apply to our internal situation. One way to
insure full employment is for the Government, in one way or another,
to stand ready to increase monetary demand whenever any slack develops, thus putting a floor under wages and leaving them no place to
go but up. A full employment policy by the Government serves to
increase the strength of unions, one of whose jobs is to increase the
wages of members. Moreover, trade unions are becoming more
powerful in most nations and their attempts to raise wages pari passu
with productivity increases tend, as many have pointed out, to be inflationary, particularly (a) if the increase in wage rates tends to follow that of the more progressive industries, leaving those whose productivity is rising less no alternative but to meet the higher labor
costs through higher prices, and/or (b) if the increase in productivity
has been made possible by increased capital costs per unit of output,
leaving management no choice but to raise prices if it is to cover
costs. Another aspect of this is that full employment decreases the
resistance of employers to wage increases, inasmuch as in an economy
with full employment and rising wages management finds it easy to
pass on its rising costs in higher prices to its customers. This may
be even easier to do in economies where the form of business organiza23734—58

292

ECONOMIC STABILITY AND GROWTH

tion, tradition, and the laws encourage monopolistic-like actions by
businessmen. Furthermore, anticipation of full employment and
rising prices encourages businessmen to invest and this can be inflationary in a full employment situation.
Of course, all industrialized countries do not, and will not, act in
the same way on this or any other problem. Evidence of the past decade indicates, however, that the governments of many of the other
highly industrialized nations, including for example the United Kingdom, Austria, and Norway, have felt compelled to push employment
to levels well in excess of those determined as politically necessary to
the United States Thus, for example, in England, even under the
Conservative Government, unemployment greater than iy2 to 2 percent
of the labor force has become a matter of grave concern calling for
remedial action. In the United States unemployment of these dimensions would mean less than 1.5 million jobless. Italy, on the other hand,
has tolerated year after year a level of unemployment greater than we
have now.
In many of the economically advanced countries governments have
been assuming growing responsibility for providing social and welfare services at steadily increasing levels. Most indications are that
this trend will continue. Such activities are, to some extent, an additional demand on the available resources, thus tending to push up
prices. There is an accompanying tendency, frequently reported, of
individuals saving less, inasmuch as some of their future needs are
otherwise being provided for. This, too, is inflationary.
Many observers believe that another fundamental factor making for
inflation in both the richer and the poorer foreign countries is the
so-called "demonstration effect" of the higher and growing United
States levels of living, increasingly made known to others as communications of all types improve between ourselves and others. The essence
of the argument here is that as others watch us constantly increase the
quality, the variety, and the amount of things we consume, their appetites are whetted and they are tempted to try to "live beyond their
means." That is, the peoples of other nations attempt to consume more
than there is for sale at the present prices. The result is inflation.
Economists are not in complete agreement on this; some argue that
the"demonstration effect" of Americans' higher living standards may
encourage others to work harder, to save more, and to produce more,
with the consequence of disinflation. The latter certainly could happen.
But it is my hunch that in more cases than not the effect is inflationary.
Another factor making for a persistent rise in prices in both developed
and underdeveloped countries is that once inflation has gone on for any
length of time, as it now has in most foreign countries, to say nothing
of the United States, nearly all persons come to expect further rises,
and their attempts to hedge against this—by getting out of fixedincome assets, asking for escalator clauses in wage contracts, and so
on—act as a forceful propellent to further inflation.
There are special grounds for fearing that in the years just ahead
inflation is even more likely to occur in the so-called underdeveloped
areas than it is in the highly industrialized countries. In addition
to the factors just cited, the governments of most of these underdeveloped areas are firmly committed to programs of economic development—that is, to rapidly and drastically altering the structure
of their economies and to attempting to industrialize them. The

ECONOMIC STABILITY AND GROWTH

293

reasons for the adoption of this policy are well known to Congress
and need not be explored here.
Let it be acknowledged at once that economic development does not
of necessity lead to serious inflation. But it has done so in most cases
since World War I I ; and it seems more likely than not to do so in
the foreseeable future. The reasons are simple: changing the pattern of production of a country and increasing total production and
productivity are likely to be relatively slow processes, not only because of technical problems but also because they usually must be
accompanied by a great many fundamental social and cultural
changes. Moreover, most of these areas are poor and so their ability
quickly to raise large amounts of additional real capital internally is
low. And, as is well known, the amount of available foreign capital
also tends to be low compared to the demands or wishes of these countries. At the same time, the governments of these countries often find
it to their immediate political advantage to hold out great promise to
their peoples on these matters and so to increase the public's expectations of rapid improvement.
This complex of considerations has, in most instances, created for the
authorities an irresistible temptation to attempt a level of capital
formation and economic development well in excess of the physically
available goods and resources. The important aspect of this in the
present connection is that the favorite devices used have been "monetary" ones—budget deficits, easy construction credits, expanded agricultural-aid programs, etc. Often such programs are very successful
at first. There are quite a few unemployed resources, not only labor
but also land and, sometimes, capital, that can be put to work; the
public is often willing substantially to increase its holdings of cash
for a while; until labor is better organized—such organization is often
one of the desired components of an economic development program—•
wages can be held down in spite of serious rising costs of living; and,
for a time, past accumulations of foreign exchange holdings can be
spent, which action soaks up some of the new domestic cash. The
very fact of early success tends to be heady and all too often encourages an ever more ambitious program. But once these early "advantages" are exhausted, serious inflation begins. At this danger signal
most economists advise the governments to trim their sails and reduce
their economic development efforts. But this is asking more than can
be expected of the governments of most poor countries—often nations
with rather unstable political institutions. To them, and to many
other thoughtful persons, inflation, especially if it can be kept fairly
moderate by local standards—say, no more than 10 percent per year—
appears a far less serious problem than having a rate of economic
development less than might be the case. In practice, the inflation
often gets completely out of hand, but that is beside the point for our
present purpose, which is merely to note some of the reasons why we
may expect prices to rise significantly—say, something more than 7
percent per year—in the years ahead in the poorer countries of the
world. These countries supply many of our imports and the markets
for a good bit of our exports.

294

ECONOMIC STABILITY AND GROWTH
V. IMPLICATIONS FOR UNITED STATES POLICY

My thesis up to now has been that price changes can be expected
to move among nations; that in the period since World War I I inflation has characterized the economies of virtually every country
with which we have important economic relations; that the pace of
these inflations has varied a great deal; and that for most of this
period the United States, on balance, has probably imported a substantial amount of inflation, but this has decreased in recent years.
Finally, it has been argued that there are strong forces at work making for continuing inflation in both the other industrial economies of
the world and in the so-called underdeveloped ones.
I would like now, briefly, to turn to some of the policy implications of all this for the United States. First of all, it seems clear that,
in order to achieve internal price stability, the United States will have
to overcompensate for the purely domestically generated inflationary
pressures. That is, we will have to devise and pursue more rigorous—
and probably more unpopular—antiinflationary policies than our own
excesses would dictate. This is important, for it is always tempting
to pursue policies on the assumption that some of our domestic problems will be solved by the actions of others. In this field our problems will be bigger than they appear when looked at only from the inside.
Perhaps the next most important implication is that these considerations argue for the United States pursuing a more liberal import
policy than it now does. The fact of rising prices abroad will itself
tend to curtail our imports and those that do come in, it was asserted
earlier, will tend to raise our prices. But for them not to enter our
market would make our inflation even more severe. One of the most
efficient ways of restraining inflation is to increase the amounts of
goods and services in a market while reducing the amount of money
looking for goods to buy. Increasing imports does precisely that.
By the same token, any United States Government action designed to
keep out imports has direct inflationary effects. The price-raising
effects of restricting imports may be in some small part offset if other
countries retaliate and reduce their markets for our exports, but it
is clear I think that a worldwide policy of curtailing trade all around
makes for lower real incomes for everyone in the longer run. There
are of course a host of other considerations relevant to the formulation of our international trade policy; most of those also argue for
a reduction of American import barriers but a consideration of them
is outside the present assignment.
Third, if there continues to be inflation abroad and the United
States is able to stop inflation at home, or if the pace of inflation
abroad is faster than here, the attractiveness to foreigners of American goods will increase and the attractiveness of their goods to us
will decline. As a consequence, one may anticipate the emergence
in some form or other of what has been called the "dollar shortage,"
that is, severe pressure of the balances of payments of other nations.
As a consequence of this, one must expect an increase in barriers and
in discrimination against United States exports. Although a raising
of such barriers against our exports is to be regretted, we must refrain
from retaliating in kind, for this would only enlarge the problem
and add to inflationary pressures here in t\e United States.

ECONOMIC STABILITY AND GROWTH

295

The inflationary impact on us of rising prices abroad will tend to be
less as the price of foreign currencies goes down in terms of dollars.
From the point of view of maintaining stable prices in the United
States, the above analysis indicates that our policy should be to encourage other nations to make more frequent adjustments in their
foreign exchange rates than has been their usual practice, or at least
to widen considerably the margins about parity within which the exchange rates can fluctuate. Indeed, these considerations indicate
that the United States Government should reconsider its past policy
of supporting an adjustable-peg system of exchange rates with very
narrow margins which, in practice, means that the rates are virtually
fixed, i. e., the peg is adjusted only after long intervals during which
time others undervalue the dollar, and should seriously consider encouraging other countries to adopt a system of more flexible exchange rates. A policy along this line would seem to require an important change in the policies of the International Monetary Fund,
in whose deliberations the United States exercises great influence.
What can the United States properly do to help other nations avoid
inflation ? Probably not much. The most important thing would be
to prevent inflation at home. If we can do this, at least we will not
add to their price level problems, which, in turn, add to ours. We
also can refrain from such speculative stockpiling orgies as characterized 1950 and 1951, although those foreign countries providing
strategic and critical raw materials found our eagerness to buy at
almost any price welcome enough at the time. A continuation of
our policy of helping to share the military burdens of our allies may
in some cases give relief to inflation, but in others it is likely that
our help "encourages" them to make greater domestic military efforts
than they otherwise would and so, while strengthening the free world's
defenses, it also adds to inflation. In any event, our military-aid
policy should not be importantly changed because of this consideration. Probably the most important thing we can do to prevent inflation abroad is to resist the temptation to try to win friends in the
poorer countries of the world by building up their expectations that,
by undertaking elaborate economic development programs, they may
achieve a rapid and large-scale improvement in their economic wellbeing. To do so is indirectly to encourage irresponsible inflationary
policies. At the same time, if their expectations are already aroused,
either by us or by someone else, then the amount of inflation they
suffer, and so in part at least pass on to us, will be less as (1) we help
these foreign governments resist resort to excessive money creation,
and (2) as we provide, through both public and private channels,
some oi the capital they need to carry out their economic development efforts. It must be acknowledged, however, that, as respects
(1), we can expect to have virtually no influence and, as respects (2),
while enough aid will stop inflation abroad, not only does aid add
directly to the inflationary pressures in the granting country (if it
is operating under fairly high levels of employment) but it sometimes
actually encourages the recipient to attempt to live even more beyond
his means. Still, on balance, a strong case can be made for providing
some help to the underdeveloped countries in their investment programs. And there are a couple of clear-cut general guidelines as to
the sort of investment that should be encouraged by us to the extent

296

ECONOMIC STABILITY AND GROWTH

that our aim is to reduce inflation abroad: the favored projects should
be ones promising fairly quick returns—and so leading to fewer frustrated expectations; and they should make use of resources with which
these countries are most favorably endowed rather than those that
they only wish they had. A policy of discouraging unduly ambitious
development efforts and of facilitating those where there is good
prospect for genuine economic advantage is admittedly most delicate
and difficult. But the problem of coping with inflation is difficult—
and important—too.
My conclusions from all this are that the problem of stabilizing
prices in the United States will be made more difficult in the years
ahead, as it has for many of the past several years, by the fact that
serious inflation is likely to characterize for the indefinite future manyr
if not most, of the other economies with whom we deal. Although
the impact on our domestic economy of foreign trade and of disparate
international price movements is of much less importance for us than
such matters as our internal wage rates, productivity changes, defense
costs, Government fiscal policy, etc., it is much too important to be
ignored. The more important implications of this for our domestic
policies are that we must take even sterner internal anti-inflationary
measures than our own excesses would require; we should adopt more
liberal foreign trade policies; we should not retaliate when others
discriminate against us in trade matters; we should discourage overambitious development efforts by the poorer countries of the world;
and, while we should not try to solve their inflationary problems
by aid, we should continue to help them in certain of their efforts
to expand output. Fortunately, these policies can be defended on
the basis of our national interest even if inflation were not one of the
most serious economic problems facing us.

PRICES, COSTS, DEMAND, AND OUTPUT IN THE UNITED
STATES, 1947-57
Richard Ruggles, Yale University, and Nancy D. Ruggles, United
Nations
In the postwar American economy, the view has been widely held
that it is excessive demand that has forced prices up. Yet a review
of the period reveals that the only periods when demand was really
excessive were the 3 years immediately following World War I I (1946,
1947, and 1948) and the Korean war year of 1951. The peacetime
American economy has, if anything, been suffering from a lack, not
an excess, of effective demand. In the last 10 years there have been
3 years in which output has actually declined below the previous year:
1949, 1954, and again in 1958. The periods we tend to think of as
unparalleled prosperity are those which have followed immediately
upon the years of decline, when the increase in output was mainly due
to regaining a fuller use of available equipment and manpower.
Why, then, is it that the sense of excess demand persists ? There are
a number of reasons. The level of employment in the postwar period
has been continuously high; and the great depression of the 1930's
created in people's minds a lasting impression that employment could
be used as a barometer of the economic health of the economy. Furthermore—and this is probably more important—there has been a
general feeling that prices have been rising relatively rapidly throughout the period, and that this coupled with full employment meant
pressure on available resources. But this last premise cannot be
accepted at face value without exploring in some detail exactly what
has been happening to prices. It is, of course, possible to find price
indexes that will show almost anything one wants to show, and it is
difficult to obtain measures of price behavior that are unequivocal and
do reflect honestly what is taking place in the economy. For the present purpose, the most general price indicators are those which result
from the computation of gross national product in constant prices.
These are called the implicit price deflators of the gross national
product. They represent the difference between gross national product and its various components in constant prices, or real terms, and
in current prices, or money terms. They are given in table 1 below.
297

298

ECONOMIC STABILITY AND GROWTH

TABLE 1.—Implicit

price deflators for gross national
year percentage changes)
1946

1947

Consumer goods
7.8
Durable
1.8
Nondurable
9.7
5.0
Services
Producer goods:
Residential construction
8.3
Other construction.
8.6
10.2
Equipment
Government:
Federal
20.7
State and local
8.6
Total
10.2
Consumer price index
8.5
Gross Government product price index
18.0
Gross national product
in constant 1947 prices. _ -12.5

10.7
8.5
12.9
7.6

(year-to-

1949

1950

1951

1952

1953

1954

1955

1956

5.7 - 1 . 0
4.3
5.9 -3.4
5.9
2.8

1.3
0
1.0
2.3

6.9
6.6
8.6
4.2

1.6
-.6
1.1
3.4

1.3
.4
-.4
4.2

0.8
-2.3
.4
2.6

0.3

-.7
1.7

1.7
.7
1.4
6.9

3.2
3.2
3.2
3.2

-2.5
1.0
4.8

4.2
1.7
2.1

6.9
8.8
8.6

2.7
2.4

2.0
4.3
1.1

2.4
1.6
1.6

5.0
4.4
4.4

2.0
4.9
5.5

3.8
1.5
1.0

3.2
1.5
1.0
1.0

11.0
7.8
7.7
8.0

5.1
1.4
2.3

-1.3
.8
.3
2.4
2.0

3.9
2.0
1.1

3.2
5.6
3.1
1.5

4.0
4.5
3.8
3.3

3.6

2.9

5.4

3.4

4.0

6.1

5.6

5.3

9.6

6.9

3.8

3.9

-1.5

7.2

2.4

1948

20.0
20.0
14.4

12.0
10.9
8.1

8.5
12.6
11.7
14.5

10.8
5.5
9.6
4.5

-.5

product,

4.8

-1.0
6.2
-1.0

-2.3
3.1
1.2

1957

Source: Economic Report of the President, January 1958, table F-5, p. 122, table F-39, p. 161, and table
F-2, p. 118.
THE MEASUREMENT OF PRICES

The implicit price deflators shown above are based primarily upon
price information collected by Government agencies. This price information cannot always be a valid reflection of the true behavior or
prices in the economy. Prices are obtained for commodities of fixed
specifications. In fact, however, new items continually appear on the
market, and old items change in quality. When improvements in
quality cannot be measured, or new products appear which are more
desirable than those they replace, price indexes which must leave these
factors out of account will show too much price increase. In pricing
consumer goods, the Bureau of Labor Statistics does attempt to take
into account those changes in the nature of the product that result in
an upward change in price. But improvements in quality may occur
at no increase in cost, so that the consumer gets more value for his
money. In such cases, no adjustment is made in prices, and the price
indexes do not reflect the quality improvements. I t would indeed
be interesting if price comparisons taking quality improvements into
account could be made between periods. This question might be asked
another way. Suppose an individual were given $1,000 and a choice
of ordering goods either from an early postwar Sears, Koebuck catalog
(say 1948) or a current (1957) catalog. If he were permitted to spend
the money in terms of only one catalog, which catalog would he
choose ? The 1948 catalog has substantially lower prices, but also less
advanced products. If the 1957 catalog were chosen, it could not be
said that prices rose from 1948 to 1957, despite the evidence of the
price indexes. Different people would undoubtedly answer this kind
of question differently, but it is by no means certain that an overwhelming majority would choose to spend their money under 1948
price and product conditions rather than under 1957 conditions despite
the fact that the implicit price indexes for consumers' goods have risen
about 7 percent since 1948.
The measurement of consumer services is even more difficult. Generally speaking, the compensation of the person performing the service is taken as the major indicator of price. Thus the cost of domestic

ECONOMIC STABILITY AND GROWTH

299

servants, of haircuts, and of professional services all tend to be measured by this criterion. It is obvious, however, that the quality of any
of these services may improve. Thus, for example, medical service is
better than it used to be, but this is not taken into account. Similarlyy
quality changes in such services as housing and education cannot be
measured. It is a serious error, as a moment's reflection will indicate,
to assume that if teachers do not improve, education as a product does
not improve over the years. We have only to ask ourselves whether
we would be content to give children today exactly the same education as was given 50 years ago, using the same books and the same
fund of knowledge . Education, like other products, is a combination
of factors of production, and should not be assessed only in terms of an
intuitive judgment about the contribution of any one of these factors
of production.
Nor are consumers' goods and services the only area that presents
pricing problems. For producers' durable goods, it is well recognized
that equipment produced today is far more productive than that produced even 5 years ago, but such increases in efficiency are extremely
difficult to take account of in price indexes. In terms of the ability
of capital goods to yield productive services, there can be little doubt
that the increase from year to year is substantial. But price indexes
of producers' durable equipment generally reflect changes in costs of
production, rather than changes in the performance of the equipment
itself. Residential and commercial construction also poses problems.
Again, price indexes are constructed by determining what a structure
of standard specifications would cost. Here again, improvements in
design and cost-reducing changes in specifications are not taken into
account in the price comparison, so that the price index tends to have
an upward bias. Finally, the pricing of goods and services purchased
by Government presents considerable problems. For military equipment, it is often impossible to determine what happens to prices when
design changes radically. For the services of Government employees,
like services in general, it is assumed that there is no change in the
efficiency or output per man, so that all increases in salary are, in
effect, increases in the price of Government gross product. By this
measure, the price of Government services as measured by the pay of
Government employees has risen by an average of over 5 percent a
year since 1946. There is good reason to believe, however, that the
productivity of Government workers has increased substantially in
this period. For one thing, the introduction of data handling machines and computers speeds up the operations of many stages of
Government work. Statistics in the Government are now in large
part handled mechanically rather than by clerks. The mechanization
which is so characteristic of current developments in business is also
occurring in Government.
There is thus an upward bias in the price indexes for almost every
category of expenditure. For commodities it exists because quality
changes and new products cannot be integrated adequately into the
price data. For services it exists because by and large the value of
services is assumed not to increase, although there is strong evidence
that it does.

300

ECONOMIC STABILITY AND GROWTH
THE BEHAVIOR OF PRICES 19 4 7 - 5 7

If this upward bias in the price indexes is taken into account, a
rather interesting picture emerges. The large price increases of the
three post-World War I I years and the Korean war period stand out.
However, up through 1955, the other years exhibit overall price
changes which probably are smaller than the overall quality improvement, so that, in fact, from 1948 through 1955 the American economy
did not show an overall upward price movement in any period except
the Korean boom. For the 6 years involved, the average annual price
increase was only 1.1 percent; for consumer goods alone it averaged
0.6 percent.
For 1956 and 1957 the picture is quite different. Price increases
were considerably more pronounced, and at the same time the growth
in output was slowing down. In fact, even now, with a decline in output, prices are still rising.
Given this pattern of price development in recent years, how has it
been related to cost and demand, and what kind of price behavior can
we expect in the future ?
In the early postwar period and during the Korean war, the existence of excess demand is sufficient to explain the price movements
that occurred. Immediately after the war, the combination of long
postponed expenditures and accumulated liquid funds resulted in a
rapid increase in demand for consumer goods which were still in short
supply in an economy that had not fully converted from war production. During the Korean war, expenditures by the Federal Government on national security increased from $18.5 billion to $37.3 billion in
1 year alone, thus pumping into the economy almost $19 billion of
additional expenditures. At the same time the increase in the Armed
Forces reduced the civilian labor force, so that the normal increment
of manpower from population growth did not occur. In a period
of 2 years the real output of the economy rose by over 17 percent, and
employment rose only 4 percent. Under such conditions, it is not at
all surprising that the increase in real output could take place only
with rising prices.
But the absence of a significant price rise on average in the other
years prior to 1956-57 did not mean that there were no rising prices
anywhere in the economy. During this period agricultural prices
were generally falling. These falling agricultural prices were offset
in most of these years by rising wages, so that prices on average were
quite stable. From 1951 to 1955, there was a decline of almost 19 percent in wholesale prices of farm products, while average hourly earnings in manufacturing rose by 18 percent. In the commodityproducing industries, furthermore, wage cost rose more slowly than
hourly earnings because productivity increased. In manufacturing
as a whole, wage cost rose only 2 percent. Thus the pattern of price
behavior in these years can largely be explained in terms of the behavior of agricultural prices, wages, and productivity. The movement of agricultural prices and productivity growth tended to hold
down the increase in product prices, by keeping materials costs and
labor costs below what they otherwise would have been. Wages, on
the other hand, exerted an upward influence, increasing somewhat
more than in proportion to the productivity gain. The net result
overall was comparative price stability. The evidence of the forces

301

ECONOMIC STABILITY AND GROWTH

at work can be seen, however, in the changing price structure, as revealed by the components of the cost-of-living index or by the implicit price deflators of the gross national product. These indexes
show that prices of durable goods, where productivity increases were
greatest, have actually declined since 1951. Prices of nondurable
goods using agricultural materials—e. g., food processing and clothing—have been relatively stable; although productivity gains in these
industries probably were not as large as in the durable-goods industries, agricultural raw-material costs fell. In such areas as construction, productivity gains were less pronounced and material costs,
being mainly nonagricultural, did not fall. Prices in this sector
therefore rose since 1951 by 10 to 15 percent, a considerably greater
price rise than that exhibited by the other commodity-producing
sectors. The largest price increases occurred in consumer services
and Government services; here price increases since 1951 have ranged
from 15 to 27 percent. Services, in fact, have accounted for most of
the upward price movement that has occurred in this period. Prices
rose faster in 1956 and 1957 for 2 major reasons. First, the decline
in farm prices stopped, for all practical purposes. Second, the general softness of the economy meant that the rise in output per manhour was considerably less than in previous years. The increase in
output per man-hour for the private sector of the economy is estimated to have been in the range of 1.3 to 1.7 percent per year for
1956 and 1957, as against an average of 3.1 to 3.5 percent for the
previous 4 years. Thus the two important elements that had been
offsetting the increase in wages disappeared, and the result was increasing costs, which were passed along as increasing prices.
PRICES AND COST

It has frequently been charged that rising prices in these years
were the consequence of monopolistically administered prices, that
producers have been taking advantage of wage increases and other
increases in costs to raise prices even more. If this were the case,
profits should be increasing faster than wages. The evidence shows,
however, that the reverse is true. Profits have been decreasing in
relation to wages. This is shown in table 2 below.
TABLE 2.—Relation of corporate wages and profits
Corporate
wages and
salaries

1948
1949
1950
1951
1952
1953
1954
1955
1956

85.0
83.1
91.3
105.0
112.5
121.9
119.1
128.9
139.7

Corporate
profits and
inventoryvaluation
adjustment
30.6
28.1
35.1
39.9
36.9
36.0
33.1
40.7
40.4

Profits as a
percentage of
wages and
salaries

36.0
QO Q

38.4
37.8
32.8
29.5
27.8
31.6
28.9

Source: Survey of Current Business, July 1952, table 12, p. 15 .and National Income Supplement, 1954
edition, table 12, p. 75.

302

ECONOMIC STABILITY AND GROWTH

The decline in profits relative to wages is a further confirmation
that demand has not been excessive. When excess demand does exist,
it tends to pull prices up. Costs are raised, in such a situation, by
producers bidding against each other for existing resources. But the
increase in costs will lag behind prices, since the demand for resources
is essentially a derived demand and exists only because the prices of
final goods are higher than costs. Nor does it follow that an increase
in wages will always lead to excess demand that will result in a
wage-price spiral. A given increase in wages, unaccompanied by any
change in the rate of productivity increase, must be passed on as a
price increase by the producer if he is to maintain his profits per
unit at the same level. This means, however, that prices must rise
in exact correspondence with the increase in wages. In order to sustain a wage-price spiral, therefore, an increase in spending would be
required equivalent to the increase in wages. In a great many situations, such an increase in spending would not occur. In the first
place, wage earners do not receive as spendable income all that producers pay them; taxes are paid to the Government, and additional
funds may be siphoned off by contributory pension schemes and other
fringe benefits. Furthermore, wage earners may decide to save a
portion of their increase in wages. There is no reason to expect nonwage-earning groups to increase their spending to make up for the
lack of spending of the wage earners, so that the volume of expenditure in real terms may actually decline. Similarly, the increase in
prices of plant and equipment may mean that the real volume of
investment will be reduced, since many producers have a fixed amount
of money to spend for plant and equipment. This point is in a
sense the reverse of the well-known Pigou effect, which states that a
price decline will lead to a rise in real expenditures because of the
increased real value of cash holdings. The rise in prices means that
individuals' and businesses' cash holdings have a smaller value in real
terms and this will tend to decrease the amount of real expenditures
which they can and will make. It is quite possible, therefore, for
wage increases to have a dampening effect on output. Such a situation is not the familiar wage-price spiral postulated by orthodox
economics, but it seems quite likely that this is what has in fact occurred quite often in the United States economy. The increase in
wages in 1937 (over 12 percent in manufacturing), may have contributed to the 1938 recession, which occurred even though the economy
never reached full employment. Large wage increases also occurred
from 1947 to 1948, from 1952 to 1953, and from 1956 to 1957. The
recessions of 1949 and 1954 were not very severe, but nevertheless the
rise in wages in the previous period may have had a dampening effect
upon output. In 1956 and 1957, the rise in prices has been much
sharper, and it also appears that the current decline in output may
be more serious.
THE BEHAVIOR OF WAGES

From the foregoing discussion it might seem logical to conclude
that the culprit in the present situation is wages, and that everything
would be all right if the rate of increase in wages were slowed down.
Such reasoning really assumes that wages throughout the economy
follow key wage rates which in turn result from specific union bargains. A look at the data, however, reveals that the increase in

E€ONiOMIC STABILITY AND GROWTH

303

wages has not come solely in the industrialized or unionized portions
of the economy. The implicit deflator for Government services, for
instance, rose 58 percent from 1947 to 1957, while prices in general
rose only 30 percent. In private services also, the rise in wages has
been very considerable. In certain occupations the supply of labor
is not increasing as fast as the demand for it. In industries where
output has been expanding and productivity increasing, furthermore,
there is a natural tendency for wages to rise faster than in the lagging
industries, both because the increase in wages is necessary to attract
more labor and also because producers can pay higher wages and at
the same time enjoy increased profits, owing to falling unit costs and
increased output. If we are to permit adjustments of this sort to take
place and if we do not want to reduce wages somewhere else, it is
obvious that on average wages will increase. If the growth of productivity is small, the average increase in wages may exceed that in
productivity. But stopping the increase in wages would necessarily
involve either suppressing the natural readjustments in wages among
various groups in the economy, or bringing about actual declines in
wages in sectors of the economy where productivity increases were
slower than the average or where the supply of labor was greater.
This would imply, for instance, that Government employees, instead
of enjoying a 58 percent increase in wages since 1947, might have
suffered a decline at least in real terms, since Government service is an
area where the supply of labor is not limited, and where, according
to our statistical treatment, productivity does not increase. Such a
procedure as this, however, would run counter to the considerations
of equity which influence private producers and Government alike.
The concepts of fair wages and fair prices do exist even though they
may not be fully justifiable in the strictest economic sense.
There is no assurance, furthermore, that tampering with wages
would automatically correct the situation. Over the long run, the
rate of increase of productivity might be affected, since it is at least
in part the increase in wages and the expectation of future increases
that induce producers to plan for and install the labor-saving devices
that increase productivity. Moreover, even if the overall price level
were lower, there is no guaranty that real purchasing power would
be sufficient to keep the economy operating at full capacity. The
maintenance of a close correspondence between growing purchasing
power and growing capacity is a much more delicate problem than
we have believed heretofore; there is no necessary reason why a reduction in. wages should assure equilibrium between expenditures and
productive capacity over any extended period of time. Tampering
with wages would attack only the surface phenomena, and in the
process our normal price mechanism would be seriously disturbed.
The roots of the present difficulty are much deeper, and we must seek
a solution that will reach the heart of the problem without constituting a serious interference with the market mechanism of price
determination.
THE ROLE OF DEMAND

The level of demand in the economy has, obviously, an important
role in reaching such a solution, because of its repercussions on price,
cost, and output behavior. A high level of demand which absorbs
most of the capacity of the economy is a necessary condition for a

304

ECONOMIC STABILITY AND GROWTH

high level of investment. An economy operating considerably below
full capacity does not provide the incentive for producers to expand.
They do not expect that an increase in output would press upon their
capacity or require additional plant and equipment, and uncertainty
regarding demand makes even long-range planning difficult. The
high rate of technological change (and consequently obsolescence)
generally makes producers wary of long-range commitments except in
circumstances where even obsolete capacity could be utilized profitably.
It takes time, of course, before the capacity created by investment
is available for productive purposes. But if, when the increased
capacity does become available, demand does not or cannot increase
to keep pace with it, investment may be curtailed, or at least it will
increase at a lower rate. This will intensify the insufficiency of
overall demand, and a cumulative decline may result. Thus, the seeds
of a cumulative decline are potentially present in every period of high
demand. In Schumpeterian terms, growth comes about in cycles, and
a period of underutilization of capacity follows inevitably from a
boom. In more recent years, this same phenomenon has been discussed in terms of the growth rates and the levels of investment that
would be necessary to sustain capacity •operation. Given the increased absolute amount of-savings-that accompanies a growing economy, it is obvious that the absolute amount of investment will have
to increase if demand is to keep pace with capacity. If the economy
is geared to an increase in the relative level of saving as income rises,
the relative amount of investment must also increase. A failure of
investment to absorb all the income that people would like to save
will result in the underutilization of capacity. In the 1930's, much
the same subject was discussed in terms of secular stagnation. It was
argued that profitable investment opportunities were drying up, so
that the aggregate demand was bound to fall. Today one no longer
hears much about this secular stagnation thesis, but if we had a severe
depression, it might well emerge again as an explanation of the
behavior of the economy.
This relationship between the level of demand and investment has
important implications for productivity change over time. In periods of rising demand, the apparent productivity gain will be substantial as underutilized capacity is employed more efficiently. Many
sectors of the economy operate with decreasing costs. The distributive trades, for instance, can ordinarily sell a larger volume of goods
with the same amount of resources. Similarly, mass production industries, such as consumer durable goods, generally have decreasing
costs. A relative expansion of output in these sectors of the economy
will therefore result in an increase in output per man-hour for the
economy as a whole. On the other hand, true technological productivity increase will be relatively low in such periods, because of the
relatively low level of investment in the immediately preceding
periods when excess capacity existed. In periods of downswing, the
technological productivity gains resulting from the relatively high
level of investment in the preceding boom begin to appear, but they
are somewhat obscured by the inefficiency of less-than-capacity operation of industry and the necessity for spreading fixed costs over a
smaller output. Much of the real investment made during the boom
goes unutilized, and the productivity gain resulting from it is wasted.

ECONOMIC 'STABILITY AND GROWTH

305

In other words, because investment fluctuates the average level of
productivity gain is very much reduced. This means that the influence of productivity increases in offsetting increasing wages and thus
lowering costs is reduced, and price rises tend to be greater. A more
stably maintained high rate of investment and a fuller use of capacity
might well result in a considerably greater rate of increase in productivity, and so in a lowering of costs. This in turn would have a
beneficial effect on prices as well as on output.
It is often argued that wages, and therefore prices, will rise faster
under full employment conditions, and that the only way to keep the
increase in wages and prices under control is through the maintenance
of some unemployment. This argument leaves out of consideration
the impact of slackening demand upon productivity. There is good
reason to believe that investment is more sensitive to the level of operation of the economy than are wage rates. It is true that a decline in
profits tends to make producers resist wage demands more strenuously,
and a decline in employment makes labor less adamant in its demands.
But, as has been pointed out above, the same falling profits and the
same expectation of excess capacity have important repercussions on
investment. The slower increase in productivity through failure to
utilize capacity and low investment may over time lead to a greater
increase in wage costs than would have occurred had demand and investment been higher. To consider the effect of changes in demand
on wage rates alone therefore leaves out half the problem. For the
producer, the important consideration is not wage rates, but wage
costs.
THE CURRENT SITUATION

There can be little doubt that the American economy is now operating considerably below full capacity. The Board of Governors of the
Federal Reserve System has prepared a combined index of the degree
of capacity utilization in industries producing such major materials
as iron and steel, aluminum, copper, cotton yarn, synthetic fibers,,
cement, wood pulp, paper, petroleum products, coke and industrial
chemicals. At the beginning of this year, combined capacity in these
industries exceeded output by about 35 percent. It is true that these
are some of the industries that are hardest hit, but they represent a
substantial portion of the private economy. The Federal Reserve
Board index of industrial production indicates that the first quarter
of 1958 industrial output was at a level which had previously been
attained 5y2 years before (the last quarter of 1952). Furthermore,
when fewer goods are produced, the output of retail and wholesale
distribution will also fall. By now (April) the degree of underutilization of capacity in the materials-producing industries would
be greater than 35 percent. For the economy as a whole, it is not
unrealistic to suggest that output is a good 20 to 25 percent below
physical capacity. At current prices, this means that we could be
producing $100 billion in goods and services more than we are.
Given an estimate of this magnitude, the question naturally arises
whether we would have the capacity in terms of manpower, as well
as plant and equipment, to produce such an additional volume of goods.
Unemployment is now in the neighborhood of 7 or 8 percent of the
labor force. The length of the average workweek, however, has
dropped by another 4 or 5 percent, so that the total level of unem-

306

ECONOMIC STABILITY AND GROWTH

ployment, taking part-time work in account, might be as high as 10 or
12 percent. Moreover, the labor force has grown more slowly in the
last year, owing to lack of employment opportunities, than would be
normal for good times. During 1957, the labor force increased by only
0.6 percent. With full employment, an additional 2 percent increase
in the labor force might be expected. Taking all of these figures together, man-hours could probably increase at least 10 or 12 percent
without exceeding the full employment level; this would still leave
2 or 3 percent frictional unemployment.
Could such a 10 to 12 percent increase in man-hours result in a 25
percent increase in output if, as already suggested, the physical capacity to produce such output exists ? An answer to this question may
be sought by looking at the relationship between the changes in manhours and the changes in output that have actually occurred in years
of recovery from a recession or depression. Table 3 below shows the
data for five such periods.
TABLE 3.—Relation of changes in man-hours to changes in output in post-recession
periods
[Percentage change from preceding year]

Year

1922
1935
1939
1950
1955

Man-hours
employed
in the
private
sector

Real private
gross
product

8.6
5.5
5.3
2.4
3.5

16.3
13.2
9.4
10.1
7.9

Ratio of
col. 2
to col. 1
1.9
2.4
18
4.2
2.3

Source: Prices, Productivity, and Incomes, Joint Economic Committee Print, 85th Cong., 1st sess.,
table 5, p. 91.

In the most striking instance, a 2.4 percent increase in man-hours
in 1950 achieved a 10 percent increase in output. In the least favorable case, a 5.3 percent increase in man-hours in 1939 resulted in a 9.4
percent increase in output. The increase in output generated per manhour exceeded that which would be needed to produce an increase of
$100 billion in output given our present unemployment in 3 of the
5 periods, and in both of the postwar periods.
It is true, of course, that such an increase in output could not be
achieved in a short space of time; increases in output of the magnitude
of 25 percent cannot take place quickly. Even the most rapid rates of
increase seldom go beyond 15 percent a year. But it would not be
unreasonable to expect a 10-percent rate of increase to be maintained
for several years. In this connection it is interesting to note that the
increase in private gross product in constant prices for the 2-year
period from 1922 to 1923, following the recession of 1921, was 31 percent. If some of the increase in output were plowed into productivityincreasing investment it is reasonable to suppose that after 2 years of
10 percent growth sufficient additional capacity and productivity
would have been generated to make a continuing 5 or 6 percent growth
rate possible, instead of the 3 percent growth rate that the economy
has been averaging over good times and bad.

ECONOMIC STABILITY AND GROWTH

307

If it is true that the economy is operating at a level of $100 billion
below capacity at the present time, it is obvious that our most important
task is to get the economy back to full capacity operation. The amount
we are losing by operating at our current level is more than the total
expenditure of the Federal Government. Even more serious than the
waste of resources by idleness is the effect that underutilization of
capacity has on investment, and therefore on long-term growth. As
other countries continue to grow, we are likely to slip behind. The
impact of the underutilization of capacity is not so much in the immediate present as it is upon future increases in productivity. Future
levels of consumption and investment will be correspondingly reduced.
The cost of idleness is much greater than, for instance, the cost of as
much as $50 billion a year in foreign aid.
It has already been noted that the price-cost-output situation in the
United States at the present time is not unique in the history of our
economy. It is also true that many of the other industrial countries
are currently having much the same problem of rising wage costs and
prices without excess demand. In Sweden this price-cost problem is
most severe; Norway, the Netherlands, Canada, and the United Kingdom have had difficulties of this nature but in somewhat less degree.
Even in Germany, where unemployment has been a major problem
throughout the whole postwar period, wage rates in the last year or
two have risen more than productivity. France is the only one of the
western European countries that has recently been experiencing demand inflation. Belgium and Italy are still suffering from some
unemployment, and wage costs in these two countries have been falling.
In terms of growth of output since 1950, three countries stand at the
bottom of the list: Denmark, the United Kingdom, and the United
States.
CONCLUSION

There is no automatic mechanism that will keep the economy
operating at full capacity. Left to chance, there will be fluctuations,
resulting in considerable underutilization of capacity over time. The
fluctuations will not occur around the full capacity level. Excessive
demand will normally be of short duration and concentrated in periods
of large defense expenditures or other abnormal demand. The periods
of underutilization of capacity will be long, and although they may
not be reflected in unemployment they will result in general softness,
slow growth, and slow productivity increase.
This chronic underutilization of capacity results not from the drying up of investment opportunities or from the maturing of the
economy but from a failure of purchasing power to keep pace with
potential production. The problem is that at the levels of output
that would utilize capacity fully, the saving that consumers and business will want to do will exceed the investment that would normally
occur at this level of output. There are two alternative approaches
to this problem. First, a stimulation of consumption (either private
or public) would reduce, at least in real terms, the amount of saving
people want to do; and thus restore equality between saving and investment at capacity output. Alternatively, it would be possible to
stimulate investment and thus bring it into equality with the saving
individuals and business would want to do at capacity output. A free
2,3734—58

308

ECONOMIC STABILITY AND GROWTH

choice solution to the problem of the long-run full utilization of the
economy does not exist. Either consumption or investment, or both,
must be increased to take up the slack m the economy. This means
that the decisions of either consumers or those purchasing producers'
durable goods, or both, must be altered by tax cuts or other incentive
measures. It would always be possible to achieve full capacity utilization by providing more purchasing power to consumers, but this
entails the danger that the level of investment may be lower, and
growth and productivity increase smaller, than we as a nation would
choose given the opportunity to evaluate the true alternatives. In
such a situation accumulated liquid assets in the hands of individuals,
and the national debt, would tend to be large, but capital in real terms
would be small. Stimulation of real investment, on the other hand,
would yield greater increases in productivity and growth, and thus
would result in a larger amount of real capital.
The price-cost behavior in the economy is not the primary cause of
the underutilization of capacity. The main cause is that at full capacity the economy does not operate in such a way that it continues to
generate sufficient expenditures to purchase the output that can be
produced. There is no automatic mechanism that will assure stability
and full utilization of capacity. In the last analysis, the satisfactory
operation of the economy is the responsibility of the Government.
To the extent that the economy operates at less than full capacity, to
this extent the Government has failed in its obligation to maintain the
health of the economy. Monetary policy is not a sufficiently powerful
tool to accomplish this task.
There is considerable danger that in an attempt to prevent rising
wages and prices there will be agitation to atomize both labor and
business. The solution is not to be found in this direction. What we
need are positive steps toward solving the problem, not negative steps.
Nor is it an adequate solution for the Government simply to return
purchasing power to individuals to be spent on consumer goods. By
such a device, the Government would merely be thwarting the decision
of the private economy to save. It is true that providing stronger
incentives to invest will not necessarily insure full capacity utilization; it will create more capacity which, in turn, will demand either
further increases in the purchasing power of consumers or further
increases in investment. This, however, is not a cause for discouragement; it is merely the hallmark of increasing productive potential.
Instead of despairing at the increase in our productive capacity, we
should utilize it for an increasing standard of living and our future
growth.

MODELS OF THE PRICING PROCESS 1
J. Fred Weston, University of California, Los Angeles
One set of papers in this compendium deals with the relative importance of demand versus cost factors in explaining recent price
behavior in the United States. Another delineates the effects of prices
on employment, output, and factor returns (incomes to factors of
production).
The basic function assigned to this paper is to provide a connecting
link between these two sets of forces uy showing the interrelationships between them.
This paper seeks to achieve its objective by three stages. First, it
will illustrate the interrelationships by reference to the controversy
over dichotemization in the pricing process. Second, it will review
recent developments in the theory of macroeconomic factor returns
which illumine the issues under consideration. Third, it will draw
on empirical materials to test some of the ideas set out.
I. DICHOTEMIZATION OF THE PRICING PROCESS

One of the complaints lodged by the new economics was the indeterminacy of the absolute price level in classical economic theory. The
indeterminacy was said to have resulted from a dichotemy in the
analysis of the economic process of monetary economy. The dichotemy said that the commodity markets are dependent on and determine relative prices, but that the monetary sector is dependent upon
and determines absolute (money) prices.
Following the pattern set by the scholarly literature on this subject, the present paper analyzes these issues by an 11-equation model
of the economic system. This model has simple numerical parameters.
It is not strictly empirically based, for such systems are too complex
for easy exposition. The use of a simple model with illustrative
numbers accomplishes several things. Since many factors are involved, a literary discussion typically ends in the same disagreements
which provoked the initial debate, because the interactions between
the many elements operating in an economic process cannot be handled in an orderly fashion.
On the other hand, econometric analysis must limit the number of
variables to be manageable. For some situations, the key factors may
have been omitted. Both literary and econometric discussions often
have many unspecified, but implicit, assumptions which are crucial
for determining the essential properties of the operation of the
economy.
1
Support of studies of this and related topics by the committee on research, University
of California, Los Angeles, is gratefully acknowledged.

309

310

ECONOMIC STABILITY AND

GROWTH

The parameters of the equations in the present model are illustrative. But they make explicit the assumptions of the nature of behavior relationships. Different assumptions are readily made by
changing the number and selection of variables and the values of the
parameters of the equations. The system is set out completely in
table I.
TABLE I.—A model for investigating dichotemization of the pricing process
I. THE FUNCTIONS

(la) M»=Ji»
(16) M»=M* (r,Y)
(2a) M*=kY
(26) Md=L (r,Y)
(3) Af*=M<*

(4) I=I(r,Y)
(5) 8=8 (r,Y)
(6)

£=/

(7) Y=PR
(8) R=R (Nd)
(9) N*=F ^
(10) W=W+pf(N>) .P
(tOa) a=0,/3=l
(106) a = l , 0 = 0

(11) #<*=#•
M,rf=»supply of money
M =»demand for money
F= national money income
r representative interest rate level
/=»investment (private)
5=saving
jj-an index of physical output

II. THE EQUATIONS

(la) M«=200

(16) M'=2f
2f

(2a) Md=2/5Y
(26) M<*=.5F-12.5r
(3)
(4) /=.15y-4.5r+20

(5) £ = . 2 F + . 5 r - 2 5
(6)
(7) Y=PR
(8) 72=8.5 N*-.025N**
(9) N*= 170-20.
(10)
(10a) A['=8O-5O/t0

(106) TF=5.29
(ID
SYMBOLS
P=index of the price level
iV"d=labor demand
w=real wage rate=^—
iV«=labor supply
W= money wage rate=wP
11 equations and 11 unknowns

The system contains 11 equations and 11 unknowns. The values
of the solutions are close to current national income accounts figures.
The nature of the equations will be briefly described.
l he system begins with two alternative money supply functions.
Equation (la) is money supply given at $200 billions. This reflects the assumption that the money supply is determined by "banking policy" unrelated to economic variables in a clear or systematic
fashion. Equation (lb) states that deposit creation through commercial bank lending policies is influenced positively by the level of
business activity. Furthermore, at higher rates of interest more
loanable funds are offered. The interest rate is meant to be representative of the complex of interest rates. The level of interest rates
is most significant as an index of the kinds of credit standards and
credit terms currently employed by lenders—in short, an index of the
availability of loanable funds.
Equation (2a) is an expression of the quantity theory of the demand for money. The quantity theory in its most familiar form is
M V = P T . V, transactions velocity, can be expressed as 1/k which is
the income velocity of money. The equation becomes M=kPT. If
real output, R, is substituted for T, the total volume of transactions,
the equation becomes M=kPR. PR is the same as money income, Y.
The equation becomes M = k Y where k is the proportion of income
command over which people desire to hold in the form of money.
In equation (2a) the value of k is set at 2/5.

ECONOMIC STABILITY AND GROWTH

311

An alternative formulation of the demand for money is the liquidity
preference theory. This states that the demand for money is influenced by 2 sets of factors instead of 1. The demand for money
is influenced by the level of economic activity through the transactions
motive for holding cash balances. In addition, cash balances are
held for precautionary and speculative motives. The amount of
liquidity people will buy is influenced by its cost—interest rates.
The cost of liquidity is the interest which otherwise could have been
earned by the cash balances by investing them.
The fourth equation is the investment function. It has a segment
which indicates that investment depends upon output (sales), a form
of the acceleration principle. Since profits before taxes are correlated
with the level of economic activity, this segment of the investment
function may also be said to reflect the profit variable as well. The
sign and value of the interest term states that investment is negatively
interest elastic. Autonomous investment of $20 billion per year is
the final segment of the equation.
The savings equation contains the same variables as the investment
equation. The amount saved will reflect the level of business activity.
Savings may also be stimulated by the payment of higher interest
rates. Since savings is equal to investment by equation (6) and investment is income less consumption, consumption is also income less
savings. Thus the implied consumption function is C=25+.8Y—.5r.
The negative 25 in the savings equation reflects the positive intercept
of the consumption function.
The foregoing six equations complete the monetary subset of the
system. Alternative forms of the money supply and money demand
equations will be used in observing the solution process of the complete
system.
The last five equations represent the commodity or "real" subsector
of the system. Equation (7) states that the price level is the ratio
between money income and real output.
Equation (8) is an aggregate production function. It expresses
real output as a function of the quantity of labor input. The constants in the equation reflect the stock of capital in the economy and
indicate that this is a statical model. The parameters of this equation,
reflecting the quality and quantity of the capital stock, determine the
productivity of labor.
Equation (9) is the demand for labor function. It is obtained by
taking the first derivative dof the production function, equation (8).
The result is w=8.5 —05N . When this expression is rearranged in
the form of a demand for labor, it becomes equation (9).
Equation (10) is the labor supply function. We have two forms.
Equation (10a) is the classical labor supply function which states
that the supply of labor is an increasing function of the real wage.
Equation (10b) is a Keynesian labor supply function. It depicts a
situation in which labor union policy establishes what the level of
money wages will be.
We have now the complete equation system of the economy. It
seeks to represent practical and realistic elements of the economy
and to show how they interact. Specific assumptions about the behavior characteristics of the economic process are made unambiguous.

312

ECONOMIC STABILITY AND GROWTH

Situation I: Classical forms
The operation of this system can first be illustrated by the selection
of the classical forms of the equations where alternative forms are
presented. This is not a complete classical system because to conserve
space, the classical formulations of the investment and savings functions are omitted. Accordingly, equations (la), (2a), and (10a) are
used with the remaining 8.
Equation (3) becomes:
(3) 200=2/5Y
This equation is readily solved, Y is equal to $500 billions.
Equation (6) is:
(6) .15Y-4.5r+20=.2Y+.5r-25
Since Y is equal to $500 billions, r can be evaluated at 4 percent.
Thus investment and savings are equal at $77 billions.
We observe therefore that the monetary subsector can be solved to
obtain the money values of income, investment, savings, and the interest rate. Income is obtained from the two monetary equations. This
level of income, along with the interest rate, determines the level of
investment and savings.
The solution to the commodity market equations proceeds independently. Equation (11) is:
(11) 170 - 20w=80 - 50/w
The value of the real wage is $5,000, a rough estimate of the average
level of wages. With the real wage established, the demand for labor
and supply of labor are 70 million people.
The value of the demand for labor, 70, is used in equation (8), the
aggregate production function, to obtain real output of $472.5 billion
in constant dollars. This amount is divided into money income of
$500 to obtain a price level of 105.8.
We observe that employment, real wages, and output were determined in the commodity subsector without reference to the monetary
subsector. Money income from the money sector was used to determine the absolute price level. Thus, in classical economics, money
was indeed a veil and had no effect on the real factors in the economy.
To increase the money supply would be to increase money income, but
not real output or real wages and employment. Since the demand and
supply of labor determined real wages, there was no scope for the
establishment of a wage policy outside the boundaries defined by these
conditions.
The Keynesian formulation finds an interrelationship between the
monetary and real sectors. What characteristics of the Keynesian
system bring this about ?
Situation II: Keynesian monetary equations, classical labor functions
We begin with a complete money-supply equation (lb). This is
taken in conjunction with the liquidity preference form of the demand
for money equation (2b). Equation (3) now becomes—
(3) % 5 Y+20r=.5Y-12.5r
(6) .15Y-4.5r+20=.2Y+.5r-25

ECONiOMIiC STABILITY AND- GROWTH

313

Since two unknowns are involved, we need equation (6) to obtain
the values of income and the interest rate. These are, respectively,
$500 billion and 4 percent. The numerical solutions under all the
alternative assumptions will be identical so that attention will be
focused on the solution process. In this set of relationships, it is
necessary to have the money supply-and-demand equations to obtain
the money value of savings and investment. Also, the savings and
investment functions are needed to determine income as well as
interest-rate levels.
In the real subsector, nothing has been changed. Hence, the solution would proceed in the same fashion as under the completely
Classical case. Thus the liquidity-preference function is neither a
necessary nor sufficient condition for the monetary subset to have an
influence on real output, wages, or employment.
Situation III: Keynesian labor-supply fwwtions
We now combine the Keynesian labor-supply functions with either
form of the monetary subset. Since the monetary subset is solved
independently under either set of assumptions, we may take the solutions from the monetary subset as obtained under either situation I
or II. We then solve the real sector. To use equations (9) and
(10b) in equation (11), we change equation (9) into the form:
(9) ]ST d =170-^
We now solve for W.
Equation (11) becomes—
(11) 5.29=
Solve for N in terms of P.
(11) Nd=170d

This value of Nd is substituted in equation (8), which reduces to—
279.84
(8) #=722.52

In order to obtain solutions for the commodity sector, it is necessary
to have the value of money income, Y, from the monetary subsector.
Equation (7) then reads:
(7) 500=
We can solve for P, which is 105.8 as before.
The solution process with a Keynesian labor-supply function demonstrates how the monetary and real sectors may become interdependent.
Other relationships would bring about the same result. For example,
the "Pigou effect"—consumption will increase at enhanced real values
of cash balances—could be incorporated into the savings function to
read:
(5) 5=.25Y+.4r-.13M-20
P

314

ECONOMIC STABILITY AND GROWTH

Or the investment function may include a negative response to
rises in real wages—
(4)

I=.2Y-5w-4.5r+20

Other assumptions about economic behavior functions could be reflected in the set of equations to determine their consequences. Or
problems, other than the dichotemization of the pricing process, could
be investigated. In this connection, the Keynesian position on the
"flaw" in the price system—its inability to bring about adjustments
to restore the economy to a full-employment equilibrium could be
studied. In general terms, the flaw argument can be briefly stated.
Planned savings exceed planned investment, so income and output
fall. The declines in output give rise to wage cuts; reduced income
leads to less consumption, so the price level falls. Thus, the real
quantity of money falls, its value rises. People bid for other assets,
so that the rate of interest falls. Investment might rise toward the
full-employment equilibrium level, but may fail to do so for one or
both of two reasons. One, investment may not respond to declines
in interest levels—a negative rate of interest may be necessary to call
forth the necessary increase in investment. Two, at some lowered
level of interest rates, the cost of holding cash balances has become so
small that people may be willing to hold indefinitely large balances—
the liquidity trap. As a consequence of either situation, equilibrium
would settle at less than full-employment levels. This description
of the flaw in the price mechanism does not depend on rigid wages, as
the interdependence analysis did.
The rejoinder to the "flaw in the price mechanism" argument is
the Pigou effect. Pigou argued that, as the price level falls, the value
of cash balances rises and, as a consequence, consumption increases to
bring the economy back to full employment equilibrium. This
process does not depend on lowered interest rates to call forth more
investment.
Thus, the effect of wage changes or monetary policy on prices, real
output, and employment depends upon which of the alternative
assumptions about the nature of the behavior equations is the most
realistic approximation to actual conditions. This is an area requiring further empirical materials.
The system of equations presented above for analyzing the pricing
process illustrates the operations of the economic system. It provides a framework for analyzing the relationship between prices, output, employment, and real wages. However, more recent literature
has used a different approach to study wage policy.
n . A MODEL FOR STUDYING THE MACROECONOMIO THEORY OF WAGES

Recent literature has centered attention on the income and spending
aspects of wage increases. Gross national product from the spending
side or demand side is related to gross national product representing
the summation of factor returns. These relationships are shown in
table 2. Equations 1 through 5 are spending or demand functions.
Equations 6 through 15 represent aspects of factor incomes.

ECONOMIC STABILITY AND GROWTH

315

TABLE 2.—A model for studying the macroeconomic theory of wages
D equals GNP spending
1. Y=C+I+G
2. C=.65 + .G(Y-Tx)
3. Tx=.2Y
4. 7=20+.15F
5. G=$100B
6. P = - 8 - K l l F
7. CTX=.5P
8. D=.6Pa
9. CS=APa
10. Pa=P-CTx
11. *+r+p = $73
12. de=.OSY
13.
tix=-10+.1Y
14. TT=

15. « , - f = ^ = M > 2 8 6
F equals ^AT'P factor incomes
B. Symbols

C= Sales to Individuals (consumption)
/ = Gross Private Investment
(r=Spending by Government
Tx=Taxes
P— Corporate Profit before taxes
Pa= Corporate Profit after taxes
D=Dividends
CS= Corporate Saving
CTX= Corporate Taxes
i—Interest Income
r=Rental Income
W— Wages and Salaries
L=Labor Force
U= Unemployment
N= Employment
w=Average wage per worker
de=depreciation
P=unincorp. business profits
Us=indirect business taxes

With a wage increase, factor incomes rise. Potentially, the aggregate
demand function increases as well. Figure I illustrates the forces set
out in table 2.

FIGURE I

The increase in wages raises factor returns from F to F 2 . Whether
employment increases or decreases depends upon changes in effective
demand. Employment will either decrease, increase, or remain the
same depending upon the relative shifts of the F and D functions.
Unfortunately, we have little empirical material on the spending
propensities of different categories of income receivers in relationship
to income changes of the kind suggested by the above analysis. The
movement of the F and D functions will result from the operation of
forces which the recent literature on the subject has only cataloged,
but not adequately identified.
i n . PRODUCTIVITY, PRICES, AND INCOMES

Numerical models of the economic system are assuredly less powerful and informative than recent empirical, dynamic models of the

316

ECONOMIC STABILITY AND GROWTH

economy. Yet they possess the advantage of relative simplicity of
exposition and force a clear statement of assumptions. These simple
models demonstrate how complex the economic process must be and
illustrate how numerous variables interact. The models could be used
to make specific assumptions about behavior relationships and then
reason to a conclusion about the consequences of wage changes, changes
in monetary policy, changes in tax policy. Recently, studies have
been made of the time j)aths of response of selected variables to
changes in tax policy.2 Similar studies are needed for analyzing the
response of the system to changes in wages and prices.
Still another approach has argued that an analysis of market structure is necessary to provide "the first wholly satisfactory integration
of the wage-price
spiral with aggregative demand and price
analysis." 3 My emphasis would be the reverse, that most analyses
of postwar wage and price behavior have neglected macroeconomic
considerations of the kind suggested by the models presented in the
previous two sections. For the macroeconomic approach provides
a framework for encompassing the effects of economic change on
factor returns and their consequences for the rate and pattern of
spending.
Partial equilibrium studies of wages, profits, productivity, and
prices are characteristically incomplete. Much emphasis has been
placed on the following relationship:
["Percent ~|
["Percent ~|
|~Percent
"1
"Percent
change in I equals change in I less I change in
I plus change
Lpriees J
Lwages J
LproductivityJ
in profit
margin on
sales

However, the causative mechanism is not portrayed by this set of
relationships. In the short run, the absolute level of productivity
is a function mainly of the capital stock employed in an industry.
Changes in productivity overtime are highly correlated with changes
in output.4 But increases in output have differential price consequences. The demand pull which calls forth output increases is a
price rising influence. Increased output spreads overhead over a
larger number of units and is a cost reducing factor. The net result
depends upon the elasticities of supplies of products, as well as elasticities of demand, including market structure effects.
2
Arthur Goldberger, Properties of an Econometric Model of the United States, Econometric
Society, Philadelphia, December 30, 1957.
3
J. K. Galbraith, Market Structure and Stabilization Policy, The Review of Economics
and
4 Statistics, XXXIX (May 1957), p. 129.
Trends in Output Per Man-Hour and Man-Hours Per Unit of Output—Manufacturing,
1939-53. BLS report No. 100, p. 314.

317

ECONOMIC STABILITY AND GROWTH
TABLE

3.—Behavior of selected economic variables for manufacturing industries,
1949-36
(l)

Industry

Lumber and wood products
Furniture and fixtures
Stone, clay, and glass
Primary metals
Fabricated metals
Machinery (not electrical)
Electi ical machinery
Transportation equipment
Miscellaneous manufacturing. _
Food and kindred products
Tobacco manufactures
Textile mill products
Apparel
Paper and allied products
Chemical and allied
Products of petroleum and coal.
Rubber products
Leather and leather products..

(2)

(3)

Percent Percent
increase increase
in pro- in hourly
duction, earnings,
1949-56
1949-56

35.2
28.4
56.4
25.5
39 2
5L5
111.2
89.5
46.9
15.3
4.9
6.1
12.0
62.2
73.5
35.6
38.5

41.0
35.5
43.4
47.2
40.4
43.1
36.8
38 0
38.1
44.2
49.5
21.2
26 5
433
44.8
41.9
39.9
31.6

(4)

(5)

Percent inPercent
creases net
increases profits to sales,
1949-56
in prices
(1947-49),
1956
Before After
taxes taxes
25.4
24.8
26.9
53.4
32.7
41.7
37.6
29.1
-9.0
6.1
15.5
-12.1
-.3
27.2
7.2
17.3
45.8
-.7

-27.2
11.9
12.2
28 8
-10.3
2.8
-16.1
-20.2
8.1
-9.1
50.7
-24.3
43.2
13.2
15.4
17.5
50.9
10.3

-33.9
3.0
—4.7
13.6
-21.6
-15.6
-33.3
-35.7
0
-27.3
16.3
-36 6
23.8
-6.2
-2.4
20.2
15.8
-4.5

(6)
Average
level of
concentration of
employment,
1950

27.7
25.3
50.2
53.8
39.9
41.6
59.6
60.4
38.4
42.3
57.3
31.5
-.3
33.5
51.6
40.3
75.0
24.2

Sources: Cols. 2-4, Federal Reserve Bulletins. Col. 5, Federal Trade Commission, Securities and
Exchange Commission, Quarterly Financial Report for Manufacturing Corporations. Col. 6, Federal
Trade Commission, Report on Changes in Concentration in Manufacturing, 1935-47 and 1950, pp. 132136. Calculated by taking unweighted arithmetic average of concentration ratios of the 4-digit products
included under the industries listed above.

Table 3 illustrates some of the problems of the microeconomic
approach. With the data, matrices may be made of the influence on
prices, wages, or profits or changes in production at different levels of
concentration. These matrices show that for increases in production,
price rises would be larger, but not uniformly so. For the same increases in production, larger increases in hourly earnings result in
larger price rises, but not in a consistent fashion. Profit changes are
most directly connected with increases in output, but numerous exceptions are found. Increases in production haw a greater influence
on price increases than the average level of concentration in manufacturing industries. Small production increases are associated with
lower levels of concentration and large production increases are associated with higher levels of concentration. This explains why the
incautious or determined will find with selected data an association
between levels of concentration and price level changes.
Also impressive is the clustering of wage increases during the period
at the 40- to 45-percent level, except for industries where production
(sales) increases have been small. Where production increases have
been lower, wage rises have also been smaller.
These data show that the many simple generalizations explaining
recent price changes are subject to considerable reservation. Cost
push does not explain price increases, because hourly wage increases
and price increases have been smaller where production increases have
been smaller. Administered prices do not explain price increases,
since prices increases are small even among concentrated industries
where production increases have been small, e. g., tobacco manufacturers.

318

ECONOMIC STABILITY AND GROWTH

Little relationship can be found between profit increases and price
increases when these are stratified by groups of percent increases in
production or in hourly earnings.
Thus the macroeconomic models depicted in the early part of the
paper will not perform the complete task of explaining price changes;
neither will the usual kind of partial analysis. Both are needed. But
we cannot be content with the easy aphorisms suggested by the need
for both approaches and the need for separating the influence of different kinds of forces. Individual institutional developments as well
will have a substantial impact. Two illustrations will demonstrate
this.
The Revenue Act of 1954 which liberalized the rate at which depreciation could be charged against income for tax purposes reduced the
cost of capital equipment, i. e., increased its probable return. This
change was doubtless a significant factor in the capital equipment
boom of 1955 and 1956 and related price effects.
The lengthening of terms in auto installment contracts to 36 months
during 1955 was equivalent to a substantial reduction in automobile
prices as expressed by required monthly payments. This effective
price decrease had an important influence on the 7.9 million production achieved in 1955 with its consequent influence on the projected
levels of productivity in the automobile industry, the large wage increase and its subsequent spread through American industry.
IV. CONCLUSIONS

This paper has sought to show the interrelationships between the
cost-demand influences on price on the one hand, as well as the impact
of price changes on economic aggregates and factor returns. To do
this it sketched the framework of macroeconomic systems which exhibited these relationships.
To show the operation of the relevant forces in clearer focus, empirical data on production, hourly earnings, prices, profits and concentration were brought together. The data controvert the easy generalizations which have been used to explain recent price behavior in
the United States. The crucial role of institutional changes further
limits the validity of monistic explanations of causative factors.

INTERRELATIONSHIPS AMONG PRICES, EMPLOYMENT,
OUTPUT, INCOMES, AND RESOURCES

319

V. Interrelationships among prices, employment, output, incomes, and
resources
A. Price changes and the allocation of resources.
1. What may impede shifts in the proportions in which
factors are used when changes in their relative
prices call for such a shift ? How can the mobility of resources be increased ?
2. It is often said that in a dynamic economy highprofits opportunities are needed in order to attract additional resources to industries with rapidly expanding demand and thus bring down
prices and profits in the long run.
Does existing evidence show that industries and corporations making "high" profits tend to expand
production and productive capacity more rapidly
than industries and corporations making "lower"
profits ?
3. Do prices behave differently where capacity is being
expanded rapidly from industries where capacity
is more stable or declining ? If so, why ?
4. What are likely to be the effects of price level
changes upon patterns of real investment and the
allocation of resources?
5. To what extent are past and prospective price
changes likely to affect personal consumption and
savings patterns ?
6. To what extent and under what circumstances is the
choice of personal investment patterns of individuals affected by past and prospective changes in
the general price level ?
B. Relationship of prices to aggregate economic activity.
1. How, and to what extent, are aggregate employment
and output affected by the direction and rate of
change in the general price level? In relative
prices?
2. Does inflation tend to generate an ensuing collapse
of employment, production, and purchasing
power, as well as of prices? If so, by what process ? Does the rate of increase in prices influence
the outcome ?
3. What are the effects of price changes on profit margins, on rates of return, on investments, and on
the stimulus for technical progress and expansion?
4. What effects do general price level changes have
upon the size and composition of the labor force,
and upon labor and managerial incentives ?
320

ECONOMIC STABILITY AND GROWTH

321

5. What are the effects of price level changes on business financial structures—on depreciation of fixed
assets, on requirements for working capital versus
those for fixed assets, etc. ? What are the effects
on ability to finance technical improvements
which make possible greater productivity and
improved products ? On ability to finance expansion ? What are the characteristics of an "ideal"
financial structure for a business enterprise—
internal versus external funds, equity versus debt,
etc?
From existing data, what can be stated about the
relative amount of investment by firms of various
sizes in each industrial segment and about the
principal sources of capital utilized by each size of
firm?

THE COMPOSITION OF THE PRICE STRUCTURE, RESOURCE ALLOCATION, AND EMPLOYMENT LEVELS
Clark C. Bloom, State University of Iowa
THE SIMPLE VIEWS

Flows—Of output and of expenditures
The price-market-organized economy of the United States today
produces both a flow of expenditures and a flow of output. The flow
of expenditures is largely fed by the flow of income but varies therefrom by (1) decisions to hold more or less money, or (2) decisions to
borrow from, or to repay loans to, a fractional reserve banking system
and thus to expand or to contract the supply of money. The flow of
output is motivated by the flow of expenditures and thus represents
the response of producers, usually private businesses, thereto. This
flow of output is functionally related to the employment of inputs and
the generation of income. The two flows are thus obviously
interrelated.
Flows to maximize output
At any point in time, a price is determined for each unit type of each
factor of production, human or material. At this price, some units
will not seek employment either because of the greater psychic value
of leisure that would have to be given up or because of money costs
associated with that employment which reduce net compensation to zero
or below that level at which leisure seems more attractive. To force
the employment of these voluntarily idle resources obviously involves
either a waste of resources or a dictation of a real consumption rather
than a leisure choice for the individual. Both are usually seen as
undesirable. On the other hand, a society which attaches a positive
value to expanded real output will certainly wish to provide employment at the existing factor price to each factor unit seeking employment
at that price. The maximum flow of output is that made possible by
the full utilization of resource units seeking employment at existing
factor prices.
If this maximum output flow is held desirable—or, alternatively, if
the unemployment of resources is held undesirable in its own right because of the unsatisfactory redistribution of income which it implies,
or because of its social consequences, then desirable also is that flow of
expenditures which calls forth thisflowof output.
Policy for maximum output flow
In the simple view now being discussed, the existence of unemployment which signals a failure to maximize output requires action. Required is either an expansion of the flow of expenditures (via (1) reduced money holdings or (2) expanded borrowing from the banks)
or an expanded physical volume response by producers to a given
23734—58

323

324

ECONOMIC STABILITY AND GROWTH

expenditures flow (via a changed functional relationship between the
flow of expenditures and the employment of inputs).
Assuming prior maximizing behavior and its continuance under
static conditions by private individuals and businesses, only government can generate an expanded expenditures flow. It can do so by
either reducing its cash balances or—and more importantly—by bank
borrowing to cover deficits resulting from either (1) tax cuts, which
increase private incomes, or (2) increased purchases by governmental
units themselves. Since tax cuts may result in either (1) increased
money holdings, or (2) the repayment of bank loans by individuals in
lieu of increased expenditures, a given deficit is usually thought more
stimulating of employment and output when it results from spending
on goods and services by government rather than from a tax cut.
Prices and maximum output flow
Prices are implicit in this simple view. Implied is the orthodox
model of micro economics in which (1) outputs are rationed among
customers via price, (2) inputs are rationed among producers via price,
(3) output volume changes are signaled by changes in the relationship
between revenues (related to output prices) and costs (related to input
prices), and (4) input volume changes are signaled by changes in the
price thereof. But, in this simple view, these prices, once determined
by market forces, are seen as unchanging, so that the real value of the
expenditures flow and the real quantity of the output flow remain
unchanged so long as the dollar values thereof remain unchanged.
Indeed, the only conceivable explicit treatment of a price in this
simple view is with regard to the rate of interest. This rate can be
seen, sometimes is seen, as varying with (1) the desire to hold money,
or (2) changes in the supply of money, actual or potential, by the
banking authorities. Thus, an "easy money" policy (via (1) lower
rediscount rates, (2) open market purchases, or (3) reduced reserve
requirements) increases supply and lowers the rate of interest. The
decline in this price (relative to all others seen as unchanging) is
seen as increasing borrowing and as expanding the flow of expenditures by borrowers (usually ignoring its impact on the desire to hold
money or an asset values). However, this increased borrowing and its
consequent expansion of the flow of expenditures is frequently seen as
not sufficient to overcome any substantial failure to maximize output
because of unemployment. Action via Government deficits is still
seen as the more important policy position.
Certainly, the simple view does not explicitly take account of such
price changes as might be expected to occur in the face of a substantial
unemployment or a failure to maximize output. Neither does it explicitly consider whether or not these changes contribute to a growth
in output (and employment) or to a further decline in output (and
employment).
The causes of recession (depression)
The simple view, focusing on flows of expenditures and output but
ignoring (or minimizing) the impact of price movements, must find
the causes of recession in either (1) a decline in flow of expenditures,
or (2) an increase in the supply of the factors.
A decline in the flow of expenditures will follow upon the substitution of money holdings or bank-debt retirement for goods as the result

ECONOMIC STABILITY AND GROWTH

325

of (1) expectations with respect to such price movements as are anticipated, particularly with respect to the rate of interest and related
securities' prices, (2) tastes or preferences for goods, particularly important for durable goods when timing of purchases can change
drastically, and (3) bank policy which may be pointed toward a reduction in the money supply, and hence a reduction in bank loans outstanding.
An increase in the supply of factors obviously does not mean a
decline in output. An increasing supply of factors in the face of
constant employment means instead (1) a growing margin between
actual and potential output, and (2) increasing problems relating to
unemployment directly. A failure to expand a flow of expenditures
and output thus spells recession. A recession is not only a decline, but
a failure to grow. The avoidance of recession, therefore, requires a
broadening of the flow of expenditures with an increase in factor supplies which, in the simple view, can come only with a substitution of
goods for money accompanied by an expansion of bank loans.
In this view, it is worth emphasizing that recessions do not develop
as the consequence of, and are not accompanied by, general or relative
price movements.
PRICE MOVEMENTS—THE SIMPLE VIEW

Introduction
I t is, of course, perfectly possible to deal with price movements,
their causes and their consequences, within the simple situation of
the preceding section. This is now to be done.
This discussion is to be divided into three parts as follows: (1) An
indication of the kind of price movements which can be expected
either to accompany, or to cause, a movement away from a full-employment equilibrium situation; (2) an analysis of subsequent price
movements which accompany, or cause, either a further movement
away from equilibrium or a return to it; and (3) a look at the kind of
policies suggested by the specific consideration of the role of prices.
Price movements and the onset of the recession
Accompanying output price movements.—A narrowing of the flow
of expenditures (via an increased holding of money, repayment of
bank loans outstanding, or reduced borrowing from banks) implies
a diminution in demand for most outputs. Under competitive conditions, this means a short-run decline in price and in volume as
existing firms reduce output. This will be followed, in the long run,
by a rise in price toward original levels but with a further decline
in output accompanying a withdrawal of firms. Under monopoly
conditions (including conditions of monopolistic competition) for
firms a,ware of their demand and cost functions and motivated to
maximize profits, the narrowing of the flow of expenditures usually
results in both lower prices and reduced output. Under monopoly
conditions (including conditions of monopolistic competition) for
firms pricing on a "full cost" basis (i. e., setting price at average cost
plus a conventional markup), the narrowed flow means sharply reduced volumes and possible price increases as average costs rise with
the movement of outputs back from "capacity" (i. e., minimum average cost) levels. Under oligopoly conditions, a tendency to set prices
on a "full cost" basis plus tacit or informal agreements which cause

326

ECONOMIC STABILITY A1STD GROWTH

prices to be "sticky" are likely to yield price stability and sharply
reduced outputs. In each case, however, a narrowing expenditures
flow means reduced volume.
Accompanying factor price movements.—When output diminishes
with a narrowing of the flow of expenditures, the demand for factors
diminishes. Under competitive conditions in the factor market, this
means both lower factor prices and reduced employment thereof.
Under conditions of noncompetitive factor supply wherein factor
prices are given and unchanged, employment in response to a given
expenditures flow declines more sharply. In either case, however, the
result is a diminished volume of factor employment.
Similarly, if factor supplies increase while expenditures flows and
factor demand held steady, competitive conditions in the factor market
will mean lower factor prices and expanded employment while noncompetitive conditions mean factor-price stability and stability of
employment in the face of mounting unemployment.
Causal output price movements.—Given a flow of expenditures,
autonomous output price increases (via such developments as (1) a
change in market structure which replaces a competitive with a higher
monopoly price, (2) a reduction in supply consequent upon a technological change which moves marginal cost curves of firms to the left,
or (3) an increase in conventional markups for "full cost" prices)
reduce the real value of such flows. Physical output flows are diminished with a consequent diminished use of factors.
Incidentally, autonomous price increases of the kind noted above do
not seem likely of frequent occurrence. They do not seem likely to be
a usual trigger for a recession.
Closely related to such autonomous output price movements are
those movements which stem from autonomous shifts in demand.
Suppose for example an autonomous change in preferences in favor
of a specific good. In most market situations, the price of this goods
will rise and additional resources will be required for its expanded
production. At the same time, the demand for substitute goods will
fall, their price will fall, and fewer resources will be required for their
diminished production. In this case, two relevant possibilities exist.
First, there is the possibility that the changed situation may lead to a
diminished flow of expenditures. This is likely if the preferred goods
are conventionally purchased out of income while the disadvantaged
oods are conventionally purchased with the aid of borrowed money,
uch a set of facts may well lead to the retirement of debt without an
offsetting decline in money balances and thus yield a reduced expenditures flow. A reduced flow of expenditures is also likely if the supply
curve for the preferred goods is quite inelastic (so that expenditures
thereon rise relatively little) while the similar curves for disadvantaged goods are quite elastic (so that expenditures on them fall
rapidly). In either case, a reduced-expenditures flow will pinch
volume. Second, a given flow of expenditures now motivates a different pattern of output and the use of a different "bundle" of factors.
This flow may motivate an output which does not fully utilize the
existing supplies of any factor and may certainly leave supplies of
certain specific factors unemployed. It is thus possible that price
changes accompanying autonomous shifts in demand may lead to a diminished expenditures and output flow; although, of course, the
opposite effect is also possible.

ECONOMIC STABILITY AND GROWTH

327

Causal factor price movements.—Given a flow of expenditures,
autonomous increases in factor prices will motivate higher product
prices, reduced outputs, and diminished factor use. Such autonomous
increases are most likely via a change in market structure which
replaces a competitive with a higher monopoly price.
Price movements as recession correctives
The preceding section notes price movements which accompany or
trigger a narrower real flow of expenditures. More interesting and
more important than these price movements already noted are those
which accompany or motivate processes which (1) return the economy
to full employment, (2) push the economy to lower levels of employment and output, or (3) leave the economy in equilibrium at less than
full-employment levels.
A basic situation.—Assume a previously presented situation in
which the expenditures flow narrowed for reasons unrelated to a price
change. Price consequences were lower prices for some outputs (i. e.,
those outputs sold competitively or by noncompetitive firms which
know demand and cost functions and move to maximize profits) and
lower prices for some factors (i. e., those sold in competitive markets). Consequences also included a reduced volume of output and
reduced factor employment.
Viewing this situation further, the lower factor prices mean lower
costs. These lower costs will mean lower output price, greater output
volume, and diminished declines in factor employment provided that
the expenditures flow does not narrow further. Indeed, if the factorsupply schedule is sufficiently inelastic, or if it shifts sufficiently to
the right to take account of a growing real income provided by given
wage rates as outpmt prices decline; factor prices (and hence firm
costs) will fall rapidly enough so that the economy's real output and
employment will have declined but little. On the other hand, if the
factor-supply schedule is elastic, factor prices (and hence firm costs)
will fall so slowly that the economy's real output and employment
will remain at low levels. The key conclusion, however, is that output
and employment do not regain levels attained prior to the initial
narrowing of the expenditures flow.
The foregoing assumes that the new, narrowed flow of expenditures
can be maintained. If it can, the economy will stabilize at the new,
lower levels of output and employment. On the other hand, if the
flow of expenditures narrows further, additional declines in output
and employment are in prospect. Or, if the flow of expenditures does
now rise, then growth in output and employment toward full-employment levels can be anticipated. The important question has to do
with the relative likelihood of either a decline or an increase in the
expenditures flow.
Pressures toward decline are strong. They include (1) the immediately preceding decline in money income (by definition as the expenditures flow narrowed) and in real income (as output diminished)
which may well lead to protective action designed to increase money
holdings or to retire debt, both reducing the expenditures flow, (2)
the immediately preceding decline in factor employment which may
well suggest still further efforts at protective action, (3) the immediately preceding decline in output prices which may well lead to the
anticipation of further declines and a consequent postponement of

328

ECONOMIC STABILITY AND GROWTH

purchases, and (4) the immediately preceding likely unfavorable
bank experience in the face of declining incomes which will probably
motivate decline, rather than increase, in bank lending and the resultant money supply. If these strong pressures do, in fact, further
narrow expenditures flows, market adjustments—of which price
changes are a part—are inadequate to stem the recessionary tide.
However, an increase in the expenditures flow is not impossible.
If (1) the immediately preceding price declines are deemed only temporary and the anticipation of future increases stimulate purchases,
(2) efforts are made to protect living standards in the face of lower
incomes by reducing money balances or by bank borrowing, (3) the
increased real value of money balances or fixed dollar value asset
holdings when output prices decline so meet liquidity requirements
that buyers are willing to spend a larger portion of current income or
to borrow more, and (4) if bank policy looks toward maximum lending, the flow of expenditures will expand. If this expansion does,
in fact, occur, then market adjustments—of which price changes are
a part—are adequate to stem the recessionary tide and to return the
economy to full employment.
Market adjustments, price adjustments, are thus not always adequate to restore a maximum flow of output. Indeed, on balance, it
seems unlikely that they will usually be so. If they are not, policy
suggested by the simple view in which prices were taken as given still
seems desirable.
Other situations.—A flow of output at less than full-employment
levels has previously been seen to appear with a constant expenditures
flow but (1) growing factor supplies, (2) autonomous output price
rises, or (3) autonomous factor price rises. Each of these situations
can be analyzed similarly to the basic situation. In each case, stability
at new, lower output and employment levels will accompany the maintenance of the constant expenditures flow. A decline in the flow will
mean a further decline in output and employment and the inability
of market adjustments to stem a downturn. An increase in the flow
will mean an upturn in output and employment to full-employment
levels and an ability of market adjustments to restore a position of
maximum output.
In the first case (growing factor supplies), a further view of the
situation discloses lower factor prices, lower costs, lower output prices,
greater output volume, and at least some increases—although not to
full employment levels—in factor employment if the expenditures
flow is just maintained. In this case, pressures toward a decline in
the expenditures flow seem less strong and an increase more likely.
While total money and real income both declined in the basic situation, money income now holds constant and total real income rises.
While employment actually declined in the basic situation, in the
present case it is growing. While bank experience was unsatisfactory
in the basic situation, it is unlikely to have been generally so with the
better-income positions here viewed. Only the immediately preceding decline in output prices, which may well lead to the anticipation
of future declines and a postponement of purchases, really point to a
reduced flow. Countering this is the possibility that increased factor supplies mean increased labor supplies, a growing population, a
more rapid rate of family formation, and an increased implied stimulus to borrow for consumer and civic durable items which this more

ECONOMIC STABILITY AND GROWTH

329

rapid population growth and rate of home formation implies. On
balance, the chances are good in this case that market adjustments
will lead to a widened expenditures flow and the achievement of
maximum output levels.
In the second case (autonomous output price rises), reduced output and reduced factor use will again mean lower factor prices, lower
costs, a decline of output prices toward—but not to—levels prior to
the exogenous increase, an increase in output volume toward—but
not to—the prior level, and at least some recovery toward—but not
to—full-employment levels if the expenditures flow is just maintained. In this case, pressure toward a decline in the expenditures
flow seem less strong than in the basic situation but stronger than
in the case of growing factor supplies. As compared to the basic
situation, money income has remained constant rather than declined
and output prices have risen rather than fallen. Both of these developments argue for reduced pressures for a decline in the expenditures flow and make a possible increase easier of visualization.
In the third case (autonomous factor price rises), higher factor
prices are followed by higher costs, higher output prices, reduced volume, and reduced factor employment if the expenditures flow is just
maintained. As in the second case, and for the same reasons, pressures toward a decline in the expenditures flow seem less strong than
in the basic situation and make a possible increase therein easier of
visualization.
Price movements and policy
Market adjustments, in which price movements are*an important
factor, are capable of restoring maximum output and full employment only when they permit, and are accompanied by, a widening
flow of expenditures. Any policy which encourages this broadening
of flow is thus to be desired.
Such policies certainly include (1) the inculcation of attitudes favorable to the viewing of price declines as temporary and likely to
be followed by increases, (2) the maintenance of the broadest possible
market for, and maximum liquidity of, securities holdings to reduce
to the minimum the desire to add to money stocks or fixed-income
securities or to encourage a partial liquidation thereof, and (3) the
encouragement of banks to pursue a vigorous policy of loan expansion, perhaps with the cooperation of Government via loan guaranties.
It should also be pointed out that at any given level of expenditures
flow, the greater the downward flexibility of output and factor prices,
or conversely the less the autonomous upward movement of output
and factor prices, the better is volume of output and employment
maintained. Price flexibility does serve to maintain output and employment except as it encourages a narrowing of the expenditures
flow (as has been shown likely in the basic situation). Autonomous
upward price movements do serve to reduce output and employment
except as they encourage a widening of the expenditures flow (as
has been seen to be quite possible, but not necessary, in a preceding
section).

330

ECONOMIC STABILITY AtfD GROWTH
PRICE MOVEMENTS—A MORE COMPLEX VIEW

Flows are not homogenous
Thus far, the flow of expendituies and the flow of output have each
been seen as broad homogenous flows. This view is, of course, unrealistic.
Expenditures flows are for thousands upon thousands of different
products. These flows are not easily interchangeable. For examples,
a flow of expenditures for commercial aircraft is not easily rechanneled into aflowof expenditures for guided missiles, a flow of expenditures for machine tools is not easily rechanneled into a flow of expenditures for automobiles, or a flow of expenditures for television
sets is not easily rechanneled into a flow of expenditures for television
repair services.
Output flows are similar of thousands and thousands of different
products. These flows too are not always easily interchangeable. For
examples, the same men and machines in the same locations may not
shift easily from the production of commercial aircraft to the production of guided missiles, from the production of machine tools to the
production of automobiles, or from the production of television sets
to the provision of television repair services.
Flows must be mutually consistent
In terms of output.—The many expenditures flows must call forth
many output flows which are mutually consistent therewith if the flows
are to persist at a given volume. Thus, aflowof expenditures on, and a
flow of output of, new machines must not so expand capacity of the
industry using the new machines that the flow of expenditures on its
product is incapable of motivating a full employment of the expanded
resources available to it. Similarly, a flow of expenditures on, and a
flow of output of, product to be added to stocks (inventories) must
not become excessive relative to the flow of expenditures on, and the
flow of output to, final users. Continuing, a flow of expenditures on,
and a flow of output of, any product cannot be inconsistent with the
flow of expenditures on, and the flow of output of, any complementary
product.
Any disarticulation of output flows like those pointed to above will
lead to the narrowing of an expenditures flow and—even with price
flexibility—a reduced volume of output and employment, reduced
money and real incomes, and reduced expenditures flows on other products as a consequence of these reduced incomes (except as the depressing effect of the decline in income is offset by reduced money holdings,
reduced bank-debt retirement, or expanded bank lending).
In terms of factors.—The many expenditures flows must call forth
many output flows which are consistent with the available supplies
of inputs. It is perfectly possible that a particular intended (ex ante)
expenditures flow may be frustrated by a shortage of a necessary
input, this frustration following either rationing of limited output by
producers or rising prices coupled with elastic demand. Or this particular intended flow may be widened by the shortage of a necessary
input and rising prices coupled with an inelastic demand. The shortage of a particular input, if it results in a narrowed actual (ex post)
expenditures flow, spells unemployment of cooperating factors not in
short supply, reduced incomes, and, again, a tendency for reduced

ECONOMIC STABILITY AND GROWTH

331

expenditures on other products because of these reduced incomes. The
shortage of a particular input, even if it results in widened expenditures on the product in which it is included, spells decline since (1)
cooperating factors not in short supply will be unemployed, and (2)
the widened expenditures here must be in the face of narrowed expenditures for other outputs (unless such expenditures are maintained by, again, reduced money holdings, reduced bank-debt retirement, or expanded bank lending).
Such disarticulations in factor requirements thus also create a tendency toward the narrowing of expenditures flows and recession.
In terms of income distribution.—Even if the many expenditures
flows mesh well together, even if they call forth a pattern of resource
use consistent with patterns of supply, disarticulation is possible if
income is distributed according to a pattern which is inconsistent with
the maintenance of the original expenditures flow. For example, if
income yielded by a specific pattern of expenditures and output flow
emphasizing expenditures on, and output of, durable consumers goods,
is accumulating to income receivers not likely to emphasize expenditures on such items, then expenditures thereon are likely to narrow
with the unfortunate consequences already frequently cataloged. Of
course, the income receivers with expanding incomes may widen expenditures flows elsewhere with stimulating consequences to the economy as a whole. It should be noted, however, that a full offset requires
that expenditures elsewhere increase by the amount of the decline in
durable consumers goods purchases, a requirement which means that
persons with growing incomes save no part thereof.
Disarticulations of income distribution with respect to flows of
expenditures and output may, therefore, also lead to declines in output
and employment.
Price changes and disarticulations
The arguments of the section entitled "Price Movements—the
Simple View" indicate that price changes and market adjustments (1)
accompany changes in a flow of expenditures or (2) cause changes in a
flow of output in response to a given flow of expenditures. These
arguments are equally valid here. Price changes and market adjustments both accompany the changed expenditures flows stemming from
disarticulations and cause the changed output flows which result in
disarticulations.
Disarticulations thus give rise to narrowed real expenditures flows
via either (1) narrowed money flows, or (2) constant money flows at
higher output prices. Conventional price changes and market adjustments do not restore real flows (outputs) no matter how flexible
prices may be.
Any narrowed real expenditures flow is likely to spread via the pressure of reduced incomes and reduced expenditures to other flows and
through the entire economy unless, at some point, money holdings
are reduced, bank-debt retirement is slowed, or there is net new borrowing from banks.
Income distribution and disarticulations
Introduction.—The view that there are a good many flows of output
and expenditures also allows explicit consideration of the impact of
changes in the distribution of income. Thus, a disarticulation which

332

ECONOMIC STABILITY AND GRiOWTH

leaves a flow of income without an immediately desired flow of output
and, hence, adding to money balances or reducing bank indebtedness
is depressing (via a reduced expenditures flow) and points to the
possible desirability of redistributing the income flow in directions
where there is an immediately desired flow of output and which will
encourage the activation of idle balances and new borrowing rather
than debt retirement to widen the flow of expenditures.
A specific and important instance of the kind described generally
above occurs when a flow of income to businesses and individuals
which normally goes to purchase capital goods (or inventories) is not
so spent (perhaps because of output disarticulation). Indeed, in this
area, it is likely that not only does current income go unspent, but new
borrowing is replaced by debt retirement. This situation is frequently held to point to the desirability of immediately and consciously changing the flow of income away from these recipients who
frustrate its translation into an expenditures, an output, and an employment flow, to those whose flow of expenditures would be stimulated thereby. The latter are usually seen as lower income bracket
consumers or as small-business men with a great expansion potential
but little current income or borrowing potential.
It has frequently been pointed out that the transfer of income is
inevitable in any case; that only the level of the economy's output
and employment at which the transfer will take place is in question.
This view holds that, if the transfer does not take place early, income
and activity in the entire economy will decline and that that income
to businesses and individuals which originally financed capital expansion will, along with the income of the unemployed, be that which
declines most sharply. Indeed, this view argues that the basis for
a restoration of this income and of a flow of expenditures on capital
goods (inventories) is best laid by an immediate transfer which will
remove the disarticulation via absorbing excess capacity and excess
inventories by way of expanded consumption.
Price changes and income transfers.—Disarticulations lead to decline via reduced real expenditures flows which are diffused through
the economy with general declines in income which conventional price
movements, by themselves, are powerless to halt. This has given rise
to proposals for specific price changes not called forth by orthodox
economic forces acting on individual decision makers.
Generally, these proposals call for an increase in prices for those
items sold by those whose incomes it is desired to increase because
they will, therefore, expand their expenditures flows, and a decrease
in prices for those items sold by those whose incomes it is desired to
decrease because they have already narrowed their expenditures flows.
These prices are to be maintained whether or not, atomistically
viewed, they maximize the individual positions of the market
participants.
Specifically, with a decline in expenditures flows for capital goods
(or inventories), wage rates should be raised while volume is consciously maintained and retained profits and dividends are pinched.
Or, conversely, wage rates might be maintained while output prices
are consciously reduced and retained profits and dividends are
pinched. The latter alternative, however, does create the added
danger of reduced consumption if further price cuts thereby come to
be anticipated.

ECONOMIC STABILITY AND GROWTH

333

Logically, this income redistribution via conscious changes in prices
(of labor and of output) and conscious maintenance of output is
a powerful antirecessionary device. It also requires either (1) substantial Government control over prices and outputs or (2) a very
high level of industrial statesmanship. Conventional market forces
do not call forth the actions required. Furthermore, under reversed
conditions wherein the output distortion requires, and stimulates, an
expansion of investment expenditures, a redistribution of income
toward savers (i. e., businesses and upper-bracket income persons) is
logically required. This shift, also not called forth by conventional
market forces, would necessitate (1) substantial Government control
over wage rates or (2) a very high level of labor statesmanship.

INVESTMENT AND THE PRICE SYSTEM
G. D. Bodenhorn, University of Chicago
In a capitalist, free enterprise economy, most decisions about output, investment, and consumption are made by individuals, families,
or corporations who are seeking to further their own self-interest.
It has long been the wonder of philosophers, economists, and political
scientists that such a system should, in most instances, lead to socially
desirable results. This happens, to put it simply, because the way for
people to serve their best interest is to make money, and the way to do
this is to make something which other people want and are willing
to buy. In such an economic system, prices are the measure of the
value of a product, and the prospect of selling a product at a profit
constitutes the inducement to produce.
The function of profit in a free enterprise economy is to draw
productive resources into the correct industries, so that these industries can increase their output. The resource should, from the point
of view of the society, be used where they will best satisfy consumer
wants and needs. Producers must pay for the resources which they
use, and then sell the product at a price which consumers are willing
to pay, when they compare this price with the prices of other products
which could be used to satisfy their wants. If consumers find a
product so valuable that they are willing to pay the producer more
than enough to pay the productive resources, and leave a large profit
besides, then more resources should be devoted to the production of
this product. In this way the resources will be used to produce that
product which is most valuable to the consumer.
Investment is undertaken by individuals, or groups of individuals,
who seek thereby to make profits. In general, if there is an industry
which presents opportunities for unusual profits, as compared with
the necessary investment, entrepreneurs invest and attract resources
to the industry. This increases the output of the industry, and profits
are reduced to a more normal level.
Usually this flow of investment funds and of resources into high
profit uses is in the public interest. This is because the industries and
firms which make high profits use resources more efficiently and satisfy
consumer wants better. This procedure also requires that low profit
firms and industries contract, and sometimes are even forced out of
business. The system then involves the survival and growth of the
fittest, with fitness measured by the ability to serve the consumer and
earn profits.
As a Nation, however, we are interested in investment not only
from the point of view of the expansion of particular industries, but
also because investment employs people and creates income when it is
undertaken, and permits growth of national product and income at a
later date. We are interested, therefore, in appraising the performance of the investment sector of our economy from three points of

335

336

ECONOMIC STABILITY AND GROWTH

view: that of the economy as a whole, that of the firm, and that of the
industry. I am going to discuss each of these briefly, with emphasis on
investment at the industry level.
INVESTMENT BY THE TOTAL ECONOMY

It is not easy to appraise the performance of our economy from
the standpoint of the total level of investment. The only criteria
that we could use are that investment should be high enough to maintain full employment and to produce a satisfactory rate of growth
for the economy as a whole. We have, to be sure, ample evidence
that investment has been high enough to maintain full employment—
or to put this another way, has been high enough to employ the savings from a full employment level of income—since the end of World
War II. The question of a satisfactory rate of growth is complicated
by the fact that we have no good standard of what is satisfactory or
unsatisfactory in this field. Many individuals argue that our postwar rate of growth is not satisfactory because it is substantially lower
than the Eussian rate of growth. If, as has been suggested, Russian
output were to catch up with ours in the next 15 to 20 years, this
might well be a more serious blow in our contest with communism
than was the recent Russian success in conquering space.
It seems clear that our economy will not grow as fast as the Russian
economy has been growing unless we take special measures to stimulate investment. However, it is not my intention to advocate that
such measures be taken, particularly since I believe this decision to be
a political one rather than an economic one. As an economist, I see
no cause for alarm at the rate of growth which our economy is now
achieving. Furthermore, I should like to issue a warning to those
who think that a race with the Russians in this area is necessary. It
will not be easy to devise methods of stimulating investment which
will not destroy the freedom which is essential to a capitalist economy,
and which will not create a much more serious problem of inflation
than we have been facing for the last 10 years.
INVESTMENT BY THE INDIVIDUAL FIRM

Recent studies 1 of investment decisions of the firm stress the importance of profits as a requirement for investment. These studies do
not, to be sure, indicate that the firms with the highest profits are
always the ones which invest the most, and we should not expect to
find that our general economic principles are followed to the letter in
each and every instance. The price system can be said to work satisfactorily if by and large there is a tendency for investment to flow into
those firms and industries which are most profitable; and the evidence
supports the conclusion that this is the case.
Perhaps the most important result of the recent studies has been the
shift in emphasis from considerations of the demand for investment
funds to considerations of the supply of such funds. Earlier treatments of the investment problem stressed the importance of profits in
creating a demand or desire to invest and increase production. The
1
See for example, Meyer and Knh, The Investment Decision, Harvard University Press,
1957. and the National Bureau of Economic Research Conference on Research in Income
and Wealth, vol. 19, Problems of Capital Formation, Princeton University Press, 1957.

ECONOMIC STABILITY AND GROWTH

337

idea was that entrepreneurs wish to make profits, and therefore seek
to invest money in profitable enterprises. The assumption seemed to
be that the funds would be available for the entrepreneurs to borrow,
and that they would naturally channel the funds into the most profitable places.
More recent studies of investment decisions of the firm stress the
importance of profit in providing a source of investment funds, rather
than as signaling a profitable place to use funds. That is, a firm which
wishes to invest must either Dorrow the money, or get more money
from stockholders, or use money which it has saved itself. The two
sources of internally saved money are depreciation and profits which
have not been distributed to stockholders as dividends. Firms which
have large profits are in a better position to retain some of these profits
and to use them for investment purposes.
This emphasis on the supply of investment funds stems in part from
the way in which investment decisions are usually made by the firm.
There are usually individuals in the lower levels of management in
any firm who think of many ways in which to invest the firm's funds.
Therefore, the problem of top management usually seems to be one
of reducing the requests for funds on the part of lower management
to the level of funds which are available to the firm. That is, the
firm usually looks as though it wants to invest more money than it
actually has available. Therefore, the availability of the funds seems
to be of crucial importance.
It is clear, of course, that for the economy as a whole it is not
always the availability of funds which limits investment. In times
of recession, such as these, the funds are available but there is apparently not enough demand. This is simply another illustration of
the familiar theme that things do not always look the same from the
point of view of the individual firm as they do from the point of
view of the economy. Profits are clearly important on the demand side as well. This is because a firm which has been making
profits usually feels more optimistic about the future and wants to
invest more than a firm which has been having difficulty making
profits. This association on the demand side would, of course, increase the apparent correlation between investment and profits as an
internal source of funds. In a situation of this kind, it is frequently
difficult, if not impossible, to determine how much of the association
between profits and investment is due to the increased availability
of funds which is associated with the increased profit, and how much
is due to the increased demand for funds which is also associated with
the increased profit.
From the point of view of this analysis, however, it makes no difference why profitable firms invest. The price system is functioning
properly if resources are flowing into the profitable firms, since these
are the firms which know how to use the resources to best satisfy
consumer demands. The evidence of recent studies shows that profitable firms do indeed attract investment and expand.
INVESTMENT BY INDUSTRIES

Most of the interest in investment in recent years has been either
in the investment of the firm or the investment of the economy as a
whole, and little has been said about investment at the industry level.

338

ECONOMIC STABILITY AND GROWTH

One reason for this is that investment decisions are made by firms,
not industries. Another reason is that better data are available for
firms than for industries.
I have therefore analyzed two sets of data with a view to seeing
how the profit system and the price mechanism are apparently working at the industry level. All of my data are confined to manufacturing, and the number of industries into which the manufacturing
sector of the economy is divided is quite small. In no case do I consider more than 21 manufacturing industries, and in much of the
data only 11 industries are available. The industries are therefore
very broadly defined, which has certain advantages as well as disadvantages. The main advantage is that the broader the industry,
the less likely are firms to engage in activities which cut across industry lines. Therefore, the data are somewhat more reliable for
broad industries than they would be for more refined industries since
the data came from firms and each firm is placed in only one industry.
On the other hand, broadly defined industries can cover up many
interesting things which are happening within the industry. It may
be that some segment of the industries are exceedingly profitable
and are expanding rapidly, while other segments of the industry are
stationary or even declining.
The two sets of data which I have analyzed differ in the length of
the period covered, and in the definition of investment, as well as in
industrial breakdown. The first set of data, from the National Industrial Conference Board, covers the entire period from 1925 to
1954 for 11 industries, and has a finer industrial classification since
1939. The Census Bureau data covers only selected years from 1947
to 1954.
The conference board data uses a very broad definition of investment, including as investment all the assets of a firm except holdings
of securities and securities of other firms. The Census Bureau data
is concerned with expenditures which are more commonly thought of
as investment.
Finally, the conference board has information on the total investment tied up in an industry at any time, as well as information on
the amount of investment during a particular year. This permits
measurement of the rate of return on investment, and the rate of
increase of investment. The Census Bureau data give only plant
and equipment expenditures during a particular year, and other
Department of Commerce sources give profits (not of income tax)
for the same years. The Census Bureau data permits only measurement of the relative change in investment from 1 year to another
as compared to the relative change in profits between the same years.
In spite of these important differences in definitions and in treatment of the data, both sources indicate that, by and large, investment
is being channeled into the more profitable industries. Again, as in
the case of the investment of the firm, we find that there are individual
exceptions from time to time, but still in general the industries which
realize the highest profits, or which show the greatest improvement
in profit position, are the ones which expand investment the fastest.
Thus both sets of data apparently present the results of an economy
which is functioning the way a free enterprise society is supposed to
operate.

ECONOMIC STABILITY AND GROWTH

339

THE CONFERENCE BOARD DATA

The first set of data which I have analyzed is put out by the National Industrial Conference Board, and gives investment and rate
of return on investment for 11 manufacturing industries from 1925
through 1954, and for about 20 manufacturing industries since 1939.
The 11 industry data has the advantage of giving a long historical
sweep. The finer industrial classification permits a somewhat more
detailed analysis of postwar developments.
The definition of investment which the conference board uses is
the broadest possible one. They define investment as the total assets
of a firm, less only holdings of the securities of other firms and of the
Government. This means that the investment figure includes not only
plant, equipment, and inventories, which are usually thought of as
investment, and are included as investment in our national income
accounts, but also such items as cash and accounts receivable, which
are not included as investment in the national income accounts. The
conference board includes all assets because the firm needs to raise
money to obtain these other assets in much the same way that it needs
to obtain money to finance plant, equipment, and inventories.
Profits for the industries are given after income taxes have been
paid, and also after deducting any interest or dividends which the
firm collected from its security holdings. The rate of return on
investment in any year is then the profit earned during the year
divided by the total amount of money which is invested in the firm
as of the end of the year. Thus the profit is earned during the year,
but the investment may have been made at any time prior to the end
of the year.
The question which I seek to answer is whether those industries
with the highest rate of return on investment are also the industries
which increase their investment the most. In order to do this, it is
necessary to measure investment during any given year relative to
the size of the industry; that is, relative to the amount of investment
in the industry at the start of the year. This means really that we
are measuring the rate of increase of investment, rather than the
absolute amount of investment. I also sought to reduce the influence
of irregular factors by averaging both the rate of increase of investment and the rate of return on investment over 2-year periods. This
means that if automobiles, for example, had a particularly good year
followed by a poor one, these 2 years would be averaged together.
The idea in back of this procedure is that investment is usually made
for the long term, and on the assumption that some years will be
good and others poor. A 2-year average gives a somewhat better
indication of the normal rate of return, and also of the normal rate
of increase of investment.
I also experimented with somewhat longer periods, such as 4 and
8 years, and found that the relationship between profits and investment improved for these longer periods.
Table I shows the correlations between rates of increase of investment and rates of return on investment, by industry, for various time

23734—58

340

ECONOMIC STABILITY AND GROWTH

periods.2 It should be noted that the correlations are quite good
in the postwar years, and in the early thirties, and are somewhat lower
during the war and recovery from the depression of the early thirties.
The 4- and 8-year correlations are all high, except for those that
include the mobilization period of World War II.
The data not only confirm the idea that industries which make
high profits in any particular period are those which invest during
this period, but also confirm the idea that a particular industry invests more when its profits are high than when its profits are low.
For our broader industrial classification of 11 manufacturing industries, we have data for investment and profits covering 14 2-year
periods. We find that the correlation between the rate of investment
and the rate of return through time is uniformly high for all industries. This is a reflection primarily, of course, of the business cycle.
During periods of prosperity both profits and investments are high,
while those are low during periods of recession and depression. In
some industries the war years caused a certain amount of discrepancy
in the correlation between rates of return and rates of investment.
However, even in these industries, the disturbance was not enough
to reduce the correlation significantly.
It is also worth noting that the profit position of an industry
changes relatively slowly through time. That is, the industry with the
highest rate of return on investment in any 2-year period is likely
also to have a comparatively high rate of return on investment during the following 2-year period, and probably had a comparatively
high rate of return on investment during the preceding 2 years. In
spite of the relatively slow change during 2-year periods, very significant changes are possible over longer periods. In fact, industries
which showed high rates of return during the war period are unlikely
to be showing high rates of return today and probably did not show
high rates of return during the late 1920's. Nevertheless, there is
apparently a pattern of rates of return which appears during normal
peacetime prosperous periods. During the period under study this
pattern was broken first by the great depression, and second, by the
World War. Nevertheless, this normal pattern has reemerged to a
noticeable extent in the 1950's.
2
These correlations are of ranks, rather than of actual rates of return and investment.
My hypothesis is that the most profitable industries should invest the most, and I can test
this by correlating the ranks of the industries according to profitability and increase in
investment.

341

ECONOMIC STABILITY AND GROWTH
TABLE

I.—Correlation of rate of return on investment with rate of increase of
investment for manufacturing industries
[Selected time periods, 1927-54]
Correlation
Time period

2-year periods:
1927 and 1928
1929 and 1930
1931 and 1932
1933 and 1934
1935 and 1936
1937 and 1938
1939 and 1940
1941 and 1942
1943 and 1944
1945 and 1946 1947 and 1948
1949 and 1950
1951 and 1952
1953 and 1954
4-year periods:
1927 through 1930
1931 through 1934
1935 through 1938
1939 through 1942
1943 through 1946
1947 through 1950
1951 through 1954
8-year periods:
1931 through 1938
1939 through 1946
1947 through 1954

11-industry
classification

0.15
.69
.75
.75
.27
.30
.33
.35
.37
.75
.37
.54
.51
.58

Finer industry classification

i 6.257
1.065
2.426
3.574
3.455
3.187

.61
.86
.57
.28
.56
.77
.54
.83
.27
.62

* There are 19 industries in these years.
2 There are 18 industries in these years.
3 There are 21 industries in these years.
Source: National Industrial Conference Board.

It should also be noted that the price system has apparently performed its function quite well during the post war period. At the end
of the war there was a substantial shortage of productive capacity
and likewise, a shortage of many consumer goods. At this time the
rate of return on investment was exceedingly high, and so was the
rate of increase of investment. In fact, the rate of return was so
high that it induced so much investment that the rate of return
finally fell back to a more normal level during the early 1950's. Thus,
it appears that the high rates of return during the reconversion period
served a useful function.3
3
This particular conclusion is confirmed by the data from the First National City Bank
of New York, which shows a very high rate of return on investment in the reconversion
period which gradually becomes smaller and reaches what could be considered a rather
normal level in the 1950's. This is all the more interesting because the National City Bank
does not measure the rate of return in the same way t h a t the conference board measured
rate of return. Instead of taking the total asset of the firm as a measure of investment, the
National City Bank uses the net worth. 'Therefore, the rate of return is profit as a percent
of net worth.

342

ECONOMIC STABILITY AND GROWTH
THE CENSUS BUREAU DATA

The Census Bureau has published data on plant and equipment
expenditures by industry for several years since 1947. The 1947 data
came from the census of manufacturers, while data for 1949, 1951,
1953, and 1954 were obtained from the annual surveys of manufacturers. This data gives plant and equipment expenditures by industry during the particular year, and gives us no information as to the
total level of investment of the firm either during the year or at the
beginning of the year. Accordingly, I have not sought to associate
investment in plant and equipment in any particular year with the
profits in that year, but rather to associate changes in plant and
equipment expenditures with changes in the level of profits in the
various industries. Data on profits for the same industry classification that are used by the Census Bureau in these reports are available
from the Department of Commerce in their national income statistics.
The hypothesis which I wish to test is whether the industries which
have shown the greatest growth in profits are also the industries
which have shown the greatest growth in plant and equipment expenditures.
In order to reduce the effect of fluctuations in profit from year to
year, profits were not calculated for a single year but rather for
a 2-year period, in much the same way that the conference board data
were handled. Thus, for example, in seeking to explain the change
in the rate of plant and equipment expenditures between 1947 and
1954, I did not calculate the increase in profit between 1947 and 1954.
I chose rather to take the average profit in the 2 years, 1946 and 1947,
and compare these with the average profits for the 2 years, 1953 and
1954. The purpose of averaging profits over a 2-year period is, as
before, to make irregular or unexpected changes in profit less important. Unfortunately, a similar averaging of plant and equipment
expenditures was not possible because data was generally available
only for alternate years.
Two difficulties which arise in making this comparison should be
noted. First, it seems clear that the industries which are large and
have large profits will probably also be the ones which have large
increases in profit and large increases in plant and equipment expenditures. In order to avoid this difficulty, therefore, I considered the
relative change in profit and in plant and equipment expenditures.
That is, I found out, for example, what the percentage increase in
profits was between 1947 and 1954, and I also found the percentage
increase in plant and equipment expenditures between 1947 and 1954.
I then sought to determine if these industries had the highest percentage increase in investment. The second difficulty arises because I
am taking a ratio of investment (and profits) in 1 year to the investment (and profit) in another year. If the investment in either one of
these years did not bear its normal relationship to profit in that particular year, then the ratios which I get will be misleading. Let us
suppose that we have an industry in which investment is usually
closely tied to profit. Even so, there will be some years in which the
investment is higher than what one would expect on the basis of
profits and some years in which the investment is lower than what
one would expect on the basis of profits. If we now calculate the
rates of growth of investment and profit between any 2 years, we will

ECONOMIC STABILITY AND GROWTH

343

find that the ratios are misleading if either 1 of the 2 years did not
bear its normal relationship of profit to investment.
TABLE II.—Correlation of percentage changes of plant and equipment expenditures

and profits, for manufacturing industries—Selected time periods, 1947-54
Percentage changes between 1947 and*—
1949
1951
1953
1954
Percentage changes between 1949 and 1 —
1951
1953
1954

Percentage changes between 1951 and *—
1953
1954
Percentage changes between 1953 and 1954

Correlation
-0.04
.59
.86
.69
.59
.55
.45

.25
.36
. 48

1
Percentage changes in profits are calculated using a 2-year base period, including the
year preceding the base year used for plant and equipment expenditures, and a 2-year final
period, again including the year preceding the final year for plant and equipment expenditures.
Source : Census Bureau and Department of Commerce.

Nevertheless, the relationship between changes in profits and changes
in investment is quite good. Table I I shows the correlation coefficients
of the rank of the percent increase in investment and the rank of the
percent increase in profit. These coefficients are quite high except for
the changes which took place between 1947 and 1949; 1947 was a
period of rapidly rising prices and an extreme shortage of capital
goods. There were many firms which would have liked to invest a
good deal more than they were able to invest, because they wished to
be able to increase their output. We have reason to suspect, therefore,
that 1947 was not a year of normal relationship between profit and
investment. Therefore, it should not surprise us too much to find
that the changes in investment between 1947 and 1949 were not correlated with the change in profit in these years. In fact, I find it
somewhat surprising that the correlations between 1947 and some
of the later years are as high as they are. I would rather have expected
to find that any changes which use 1947 as a base were rather abnormal
and misleading. This is not the case, however, and the correlations
between 1947 and 1953 and 1954 are among the highest. This may
arise because one would rather expect these correlations to improve
with age. This is because changes from year to year tend to be rather
smaller than changes between 5-year periods. Thus, any observation
which tends to be higher or lower than "normal" will cause a larger
change in year to year rates of increase than it will cause in rates of
increase for years which are 5 or 10 years apart. Indeed, this
improvement of correlation as the time period grows longer is apparent
in the data.
CONCLUSIONS

Investment by private enterprise in our economy has been large
enough in the postwar period to keep the working force fully employed, and to achieve a satisfactory rate of growth for the economy.
It is true that we have not had continuously full employment, and that
we fall short of this goal at the present moment, but these lapses from

344

ECONOMIC STABILITY AND GROWTH

full employment have been neither serious nor prolonged. It is not
reasonable to expect a free economy to be devoid of fluctuations in
output and employment.
A more legitimate concern is that the Russian economy has apparently been growing faster than ours, and will eventually surpass us if
this continues. It would indeed be unfortunate if we were no longer
able to combat communism with the argument that we have both more
freedom and a more abundant economy. However, it is certainly not
worth giving up our freedom to seek economic abundance, even in the
unlikely event that we could achieve faster growth by doing So. Our
fundamental argument must be for the freedom and dignity of the
individual, and our cause will surely be lost if we abandon these.
Eecent analyses of investment by firms and industries indicate that
our economy has been performing satisfactorily in these areas as well.
Investment and resources are generally flowing into the more profitable uses, even though there are individual exceptions to this general
tendency.

RESOURCE ALLOCATION, ECONOMIC STABILIZATION,
AND PUBLIC POLICY TOWARD PRICES
Carl F. Christ, The University of Chicago
I. INTRODUCTION AND SUMMARY

In this paper I shall argue that public policy in the United States
should try to maintain a stable price level and a high stable level of
employment, and that employment is the more important of the two,
but that apart from certain important qualifications public policy
should pay no attention to the prices of particular goods and services,
or to tne quantities consumed by particular individuals. Thus the
monetary and fiscal stabilization policies that I shall endorse are explicitly designed to minimize Government action directed at particular goods and services.
Other contributors to this compendium will surely pay competent
attention to many technical aspects of prices. I have chosen to deal
with the relationships between public policy toward prices and some
relevant economic principles, in the belief that this will be helpful in
arriving at useful public policies.
My argument has six main points, some of which are controversial.
They may be summarized at the outset, in a somewhat oversimplified
fashion, as follows:
(1) With certain important qualifications that I shall discuss, in a
private-property free-contract economy public policy is not properly
concerned with the relative prices at which particular goods and services are exchanged in the market, such as 1,000 haircuts for one stripped
1958 Plymouth; nor, therefore, is public policy properly concerned
with the absolute prices of particular goods and services, such as $1.75
for a haircut, or $1,750 for a stripped 1958 Plymouth. This is because
relative prices have a job to do m allocating resources, and, with the
qualifications alluded to, this job is done in what economists call an
optimal way if the relative prices are determined by free contract.
(2) Again with certain important qualifications that I shall discuss,
in a private-property free-contract economy public policy is not
properly concerned with the quantities of particular goods and services produced by the economy as a whole, or consumed by particular
individuals. This is because, with the qualifications alluded to, freedom of contract leads to what economists call an optimal allocation of
resources.
(3) There is an important area within which public policy is
properly concerned with particular relative or absolute prices, and/or
with the quantities of particular goods and services consumed by particular individuals. This area provides the qualifications referred to
under points (1) and (2). These qualifications arise in part from the
question of wealth distribution, a question not answered by the
economists' definition of "optimal allocation of resources." In addi-

345

346

ECONOMIC STABILITY AND GROWTH

tion, most of these qualifications involve one or more of the following
situations: An individual is not well informed about the alternatives
facing him, or about the consequences of his actions; an individual is
not the best judge of his own interests; an individual cannot sell or
mortgage himself as he can other wealth; an individual's deed brings
about gains or losses to himself that are greater or less than the gains
or losses to society, so that an individual who pursues his own private
gain is not advancing the interests of society as much as he might.
Familiar examples of the last are theft, fraud, extortion, tax evasion,
monopolization, etc.; others will be discussed below.
(4) Public policy in a private-property, free-contract economy is
properly concerned with stabilizing the absolute price level, or simply
the price level for short, that is, with stabilizing an appropriate
average of the absolute prices of all goods and services. (We now have
three price indexes for measuring the price level: the gross national
product deflator published annually—not quarterly, alas—by the Commerce Department, and the Wholesale and Consumer Price Indexes
published monthly by the Bureau of Labor Statistics.) The reason
we ought to strive for a stable price level is not that a rising price
level or a falling level is bad in itself—indeed, I shall argue that
almost any kind of changes in the price level would be fine as long as
everybody foresaw them accurately in advance; the reason is rather
that the price level has a job to do that it can only do well if its future
movements are accurately foreseen by everyone, and that stabilizing
the absolute price level is the best way to permit everyone to foresee
its future movements accurately.
(5) Public policy in a private-property free-contract economy is
properly concerned with maintaining a high stable level of employment of resources, especially human resources. The reason for this
is that full employment is better than unemployment, and that in the
absence of public policy to the contrary we have experienced occasional
periods of substantial unemployment.
(6) Although there appears to be some conflict between the objectives of a stable price level and a high stable level of employment,
particularly in recent years, appropriate monetary and fiscal policy
can come tolerably close to achieving both of these objectives at the
same time. If it is necessary to choose, I would be inclined to accept
some price level instability rather than very much unemployment,
because while both bring about a redistribution of income and wealth
among people, unemployment brings about a loss of income and wealth
to the economy as a whole, in addition to a redistribution.
Sections I I to VI below discuss some economic principles fundamental to these 6 points, and the remaining sections (VII to X I I ) discuss the 6 points in more detail.
II. THE PROBLEM OF THE DISTRIBUTION OF WEALTH

Any economic society must decide, explicitly or implicitly (the
latter is the more typical), what kind of distribution of income and
wealth among its people there shall be. This is a highly controversial
issue. Most people prefer receiving a larger to a smaller share of the
total social income and wealth. In addition, many if not most people
have ideas about what constitute fair principles for distributing income and wealth, and these ideas are not all consistent, to say the least.

ECONOMIC STABILITY AND GROWTH

347

Fundamental conflicts of interest arise here. Social and economic
institutions and laws are involved in the resolution of this issue, not
once for all, but again and again, year after year.
I I I . PRIVATE PROPERTY AND FREE CONTRACT

A society based on private property and freedom of contract settles
this issue by taking the initial distribution of property as given, and
letting the subsequent distribution arise from free contracting, in a
manner familiar to all. There are rules for deciding who owns each
piece of property. Typically, each person initially owns himself (and
within limits his immature children), and whatever property he has
been given by his parents or others, by inheritance or otherwise. Each
person has control over the use of his own property, except that we do
not permit a person to sell or mortgage himself or his children. Property can change hands only by free contract agreed to by all parties to
the contract (this usually involves exchange), or by free gift. A person's wealth increases rapidly if he receives large gifts, if he produces much, if others become willing to offer high prices for what he
owns, and if he saves much. Because of freedom of contract, no person
is forced to accept a material position that he thinks is inferior to the
one that his original property will provide for him. If anyone can
produce property that he regards as more valuable than the resources
he consumes in the process, then he can gain from production. And if
he can find other people who will give him in trade things that he
regards as more valuable than the things he gives up in return, then he
can gain from exchange.
Under certain conditions, this kind of an arrangement would lead
to what economists call an optimal allocation of resources. The conditions, of course, are the source of the qualifications mentioned in
connection with points (1) to (3) at the beginning of the paper. Let
me now turn to the economists' definition of "optimal allocation of resources," then state the conditions, and then argue that if they are met,
private property and freedom of contract do lead to the economists'
optimal allocation of resources. This will occupy sections IV-VI:
IV. DEFINITION OF AN OPTIMAL ALLOCATION OF RESOURCES

An optimal allocation of resources, in the technical language of the
economist, means a situation in which no individual can be made better
off without making some other or others worse off. It may help to
give examples of situations that are not optimal allocations. For
example, if I own a house and I want to sell it provided I can get at
least $25,000 for it, and if you want to buy it provided that you can
get it for $30,000 or less, then (if we are both well informed about the
matter, and if the interests of third parties are not affected) the existing allocation of resources is not optimal, and it can be improved by a
contract for the sale of the house from me to you at some price in the
range of $25,000 to $30,000. Our respective gains from the improvement in allocation will depend on the price we agree upon: the higher
the price, the larger will be my gain and the smaller will ]be yours.
Take another example that is more typical in that it involves production as well as exchange: Suppose that you are willing to buy a
widget for $120 or less, and that I can cause a widget to be produced

348

ECONOMIC STABILITY AND GROWTH

with resources that I can buy for $90 or more from their owners (including myself). Then, again if all parties are well informed and
the interests of persons not party to the action are not affected, the
existing allocation of resources is not optimal; it can be improved by
a contract for the sale of a widget from me to you for a price somewhere in the range from $90 to $120, and another contract for the
sale of the required resources from their owners to me at a combined
price somewhere in the same range, but at or below the price of the
widget. Gains from the improvement in allocation will accrue to you,
me, and the owners of the resources that are used to produce the
widget, in a manner depending upon the prices agreed on for the
widget and the resources. An optimal allocation is a situation in
which no more contracts agreeable to all parties remain to be made.
Notice that I said "an optimal allocation," not "the optimal allocation." This is not an accident; it is important. There is an infinite
number of optimal allocations, according to this definition. For example, suppose that the existing allocation is optimal; i. e., it is impossible to make anyone better off without making someone worse off.
Then suppose that you give a thousand dollars to me. The new allocation is still optimal provided that my wants lead my to buy with
my new wealth the same things that you decide to forego because of
being poorer: the people whose wares you no longer buy now sell them
to me instead, so that no one is affected but you and me, and it is again
the case in the new situation as in the old that no one can be made
better off without making someone else worse off.
If wealth is transferred from you to me, it is more typical if I do
not choose to buy with my new wealth just the same things that you
decide to give up, but I choose to buy at least some other things instead. This means that if the resource allocation existing just before
the transfer was optimal, then the one existing just afterward is not,
because resource owners and I can now make mutually agreeable contracts providing for the production of more of the things that I want
and can now pay for, and less of the things that you want and can no
longer pay for. But, if free contract is permitted, a new allocation
of recourses will arise that is optimal in the same sense as was the old:
nobody can be made better off without making someone worse off.
The foregoing should make it clear that when an economist says
that an allocation of resources is an optimal one, he is taking the currently existing distribution of wealth as given, and is saying that the
economy is satisfying efficiently the wants that are regarded as most
important by those people who happen to have the wealth. No judgment is made at all on the question of whether the currently existing
distribution of wealth is too unequal, too equal, fair, proper, justified,
or anything of the sort. Thus in saying that a particular allocation
of resources is an optimal one, the economist completely sidesteps an
important public policy question, namely, what should the current
distribution of wealth be ?
It is important to notice that if the current distribution of wealth
is accepted, and if individuals are allowed to contract freely thereafter, then the distribution of wealth may and almost certainly will
change, in favor of those who receive gifts, who produce much, whose
property becomes highly valued by others, and who save much.
It may seem strange to use the term "optimal" in a way that does
not raise the important policy question of wealth distribution. But

ECONOMIC STABILITY AND GROWTH

349

on reflection, I think it is not so strange after all. If economics is to
be a science, then it must contain propositions that will be recognized
as valid by scientifically trained people, regardless of the value judgments that they hold or the vested interests that they have. The
statement that it is better for people to get more rather than less satisfaction from their wealth must be acceptable to nearly everybody,
Socialist or private enterpriser, rich or poor, and hence it is useful to
have a term that describes a situation where the best has been achieved
in this respect. Optimal allocation of resources is that term, and its
meaning is therefore perfectly clear.
The policy question of what kind of wealth distribution is proper
is one concerning which economics as a science has no well-established
answer. Indeed, in most respects it is not a scientific question at all,
but a question of value judgments.
V. CONDITIONS FOR AN OPTIMAL ALLOCATION OF RESOURCES UNDER PRIVATE
PROPERTY AND FREE CONTRACT

Now it is time to state the conditions under which private property
and free contract will lead to an optimal allocation of resources, thus
defined. The following discussion draws heavily on Prof. Frank H.
Knight's account of the assumptions necessary of the existence of
perfect competition, set forth on pages 76-80 of his book, Risk, Uncertainty, and Profit, first published in 1921 as one of the Hart,
Schaffner & Marx prize essays in economics. For completeness, I
shall first repeat that an optimal allocation of resources is a situation
in which nobody can be made better off without making someone else
worse off.
The required conditions are as follows (all are clearly unrealistic
in some way) :
(a) Motivation.—Each individual seeks intelligently to satisfy his
wants as completely as he can with the means he has. In particular,
he will always buy from the lowest priced seller he knows of, and
will always sell to the highest priced buyer he knows of.
(b) Information.—Each individual knows what his present and
future wants are and how important they are to him; he also knows
what present and future actions are open to him, and what the consequences of his actions will be. In particular, he knows everyone
else's present and future bid and asked prices for all wealth, he
knows what and how much he can produce now and in the future
with any given combination of resources, and he knows to what extent
his present and future wants will be satisfied by each kind of wealth.
(c) Mobility.—Resources are perfectly mobile, that is, it takes no
time and. costs nothing to convert from one kind of productive activity to another, or to exchange goods in the market.
(d) Competition.—Every individual is such a small buyer and
seller of the things he deals in, relative to the total transactions of
the market, that his decisions to bid or ask or buy or sell have no
perceptible effect on the prices bid or asked by others.
(e) Third parties.—Every individual deed affects only the doer,
and every contract affects only the parties to the contract. In particular, this excludes contracts by which third parties are injured,
and also contracts by which third parties are benefited.

350

ECONOMIC STABILITY AND GROWTH

(/) The institution of private property and free contract as we
know it is modified to permit individuals to sell or mortgage their
persons in return for present and/or future benefits.
These conditions have been deliberately designed so that if they
were fulfilled, then each individual would seek and find others with
whom he can make mutually advantageous contracts, and such contracts, and such contracts would be made, until no more remain to be
made. When that state of affairs occurs, an optimal allocation of
resources has been achieved.
VI. THE ROLE OF RELATIVE PRICES I N RESOURCE ALLOCATION

It is now easy to explain the role of relative prices in the process
of resource allocation. First, note that if everyone seeks the best price
he can get, and no one can control prices, and everyone knows what
prices are being quoted, then it will turn out that for each good or
service there is a single market price: No one will pay more and no
one will accept less. Market prices may change in response to changes
in resource availability, technology, tastes, or wealth distribution,
but they are not perceptibly influenced by any individual's action.
Market prices, and in particular relative prices, have several jobs
to do in allocating resources; these jobs can conveniently be grouped
into three related but more or less distinct main groups. The first is
to decide how much of each product is to be made, and for whom.
The second is to decide what combination of resources is to be used in
making each product. And the third is to decide how much of current
output shall be consumed and how much shall be saved and invested
for increasing future output.
Kelative prices perform these jobs in several ways. They carry
information throughout the economy concerning products that are being made in larger or smaller than optimal amounts, and concerning products that are not being produced with the optimal combination
of resources. They provide the incentive for people to shift toward
an optimum. And they distribute the product among resource owners,
since each resource owner is paid the market price for his resources,
thus receiving an income that he can use in the way most satisfying
to him.
The price of a product, relative to the combined price of the resources used to produce it, tells whether the right amount of that product is being produced or not. If the relative price of a product is high,
so that the rate of return to capital in that industry is higher than
elsewhere in the economy, then too little of this product is being produced, because buyers can induce resource owners to shift resources
from elsewhere to make more of it, at prices that will satisfy both the
buyers of the extra product and the owners of the transferred resources. (Recall the example of the widgets in sec. IV.) And the
fact that the rate of return in that industry is higher than elsewhere
in the economy provides the incentive for more resources to come into
it. Similarly, if the price of a product is low relative to its cost, that
is a sign that too much of it is bein^ produced, and provides the incentive for resources to leave that industry.
If the price of a resource is low, relative to the value of the output
it can produce in a certain firm, that is a sign that too little of the
resource is being used in that firm, and the firm will induce owners of

ECONOMIC STABILITY AND GROWTH

351

the resource to shift some of it from elsewhere in the economy to that
firm, at a price that will satisfy both the firm and the owners of the
resource. Similarly, if the price of a resource is high relative to the
value of the output it can produce in a firm, then the firm is using too
much of the resource, and the firm is led to cut down on the use of it.
Thus, in the end, producers are induced to make for each buyer the
things he values most highly. Resource owners are induced to apply
their resources where they can produce things that are most highly
valued by buyers. And each producer is induced to make his product
with the combination of resources that costs least; i. e., that is the least
useful to other producers.
Interest rates form the relative prices of future versus present
goods and services (if the absolute price level is stable—of that more
later). Through interest rates, consumers decide how much to consume and how much to save, and entrepreneurs are induced to invest in
activities whose products will be most highly valued relative to their
cost; i. e., relative to the things that must be foregone to obtain them.
It is important to remember that relative prices established under
free contract will not lead to these optimal results unless resource owners are permitted to receive the market price for each unit of resources
they own, for otherwise the incentive to allocate optimally is affected.
This means that large incomes will be received by those who own large
amounts of valuable resources. Policy directed toward the distribution of the ownership of resources will be considered in section VIII.
One of the most important advantages of free contract, barely mentioned so far in this paper, is that it provides automatic adjustments
to changes in tastes and technology. Resources shift out of industries
whose products have become less desirable or more expensive to produce, and into industries whose products have become more desirable
or cheaper to produce. Equally important, producers who feel that
they have ideas for useful new products or better techniques are free
to try them out, and if buyers find they are worthwhile, they will be
adopted.
VII. PREVIEW OF THE REST OF THE PAPER

If the foregoing two sections were a perfectly accurate picture of
the real world, there would be no need for public policy at all, except
to maintain the institution of private property and free contract, and
possibly to modify the initial distribution of wealth (of which more in
the next section). Though the real world is not just like the picture
drawn, there is a significant resemblance.
In the remainder of this paper, I shall argue first that there are
important areas within which givernment is justified in modifying the
results that would follow from free contract alone (this is the third
of the six points that I stated at the beginning). Then I shall argue
that after that is taken account of, there remains a significant field
within which public policy should not be concerned with prices of
particular goods and services, or the quantities consumed by particular individuals (these are the first 2 of the 6 points). Then I shall
argue that public policy should be concerned with stabilizing the price
level and with maintaining a high stable level of employment, that
both objectives can be tolerably well attained, and that if it is necessary to yield a little on one of them, it is better to yield on the stable
price level (these are the last 3 of the 6 points).

352

ECONOMIC STABILITY AND GROWTH
VIII. PUBLIC POLICY TOWARD THE DISTRIBUTION OF WEALTH

Let me distinguish two kinds of justifications for public policy to
affect the prices of particular goods and services, and/or the quantities consumed by particular individuals. One kind deals with policy
to affect the distribution of wealth, which, as pointed out earlier, is
a subject not dealt with in the analysis of an optimal allocation of
resources. The second kind deals with policies designed to correct
misallocations of resources that arise, under free contract, from the
fact that not all of the conditions (a\ to (/) in section V above are
satisfied in the real world. This section is devoted to the first kind,
regarding the distribution of wealth, on the assumption that conditions (a) to (/) are met. The next section is devoted to the second
kind.
The definition of an optimal allocation of resources that I adopt
provides no ground at all for deciding that one initial distribution of
wealth is better than another. If a choice among wealth distributions
is to be made, it must be made in some other way. Most people, including me, feel that the wealth distribution is a matter of public
concern, at least within limits, and that it is proper for government
to move toward a greater degree of equality than would result from
the institutions of private property, free contract, and unrestricted
inheritance (though there is wide disagreement about how far to
move). This feeling is based partly on a belief in equality of opportunity (this applies particularly to restrictions on inheritance), and
partly on the belief that it is fair and right to compel the fortunate
to help the unfortunate.
There are many different policies that can redistribute wealth. I t
is clear that the best ones are those that do not misallocate resources
at the same time. After all, the object of wealth redistribution is to
make someone poor better off than before at the expense of someone
rich; not to make everyone worse off. This is one of the most important things economists have to say to policymakers.
Consider taxes on income, wealth, gifts including inheritance, and
consumption—either total consumption, or the consumption of specific
luxury commodities. It is clear that a tax on income can be avoided
to some extent by not earning as much income as without the tax;
taxes on wealth and inheritance can be avoided in part by accumulating less wealth; and taxes on consumption can be avoided to some
extent by saving or by shifting to the consumption of untaxed commodities. Any of these taxes has some misallocative effects, as economic analysis can show; the only tax that does not is one that cannot
be avoided in any way. The only one I can think of is a sudden expropriation and redistribution of wealth that people were completely
unprepared for, and that they firmly believe will never happen again.
Of course, once a government levies any tax, it creates expectations
that it may do so again, so even this will not work perfectly except
the first time. The taxes that have the least misallocative effects are
those that have the least effect on relative prices, namely taxes on
inheritance, total consumption, or income. (The bad effects of income
taxes can be reduced by allowing taxpayers to average their incomes
for tax purposes over their lifetimes, as suggested by Prof. William
Vickrey in his book, Agenda for Progressive Taxation, published in

ECONOMIC STABILITY AND GROWTH

353

1947.) Worst are taxes on the consumption of particular goods and
services only.
Subsidies are just like taxes, except that they encourage rather than
discourage the activity singled out. Direct cash payments to people
with low incomes or wealth are accordingly better than subsidies for
the consumption of particular goods and services. (I am still assuming that people are well informed about how to serve their own
interests.)
A particularly good way to transfer wealth to a person may be
to subsidize his education, on the grounds that wealth redistributed
in this way is likely to stay redistributed. (Here I am assuming
that, at least before he was educated, he was poorly informed about
how to serve his own true interests—but this anticipates the next
section.)
IX. PUBLIC POLICY TO CORRECT RESOURCE MISALLOCATTON UNDER FREE
CONTRACT

I now turn to policies designed to correct resource misallocations
that arise, under free contract, from the fact that in the real world the
conditions under which free contract leads to optimum resource allocation do not always hold.
A familiar case is that of monopoly, which is distinguished by the
seller's ability to affect the price of his product significantly by selling
more or less of it. A monopolist, if he exploits his position, turns out
less product than would a competitive industry in his place, and
charges a higher price, in order to make more profit than would a
competitive industry. In effect, he prevents some resources from being
used to make his product, even though there are would-be buyers who
are willing to pay more for the product than the resource owners could
get for the necessary resources if they sold them elsewhere. Economic
analysis can demonstrate that the gains that the resource owners and
would-be buyers would get, if they were allowed to exchange, are great
enough to pay the monopolist's profit and still leave something over.
Hence the monopolist creates a misallocation of resources, because he
creates a situation in which someone could be made better off without
making anyone worse off. Public policy to prevent monopoly is accordingly proper.
A second example is that of protection against common disasters,
e. g., activities such as national defense, fire and police protection, and
sanitation. Here the distinguishing characteristic is that if one individual contracts for protection for himself, he thereby protects his
neighbors at least to some extent, whether they agree to help pay for
the protection or not: He cannot have the fire in his own house put out
without reducing the risk that his neighbor's will catch fire, he cannot
protect himself from attack without reducing the likelihood that his
neighbor will be attacked, and he cannot practice sanitation without
reducing the chance that his neighbors will become infected.
If everyone else is going to pay for protection, it may be in the
private interest of a single individual not to do so, because he gets
substantial protection anyway from the fact that others are providing
it for themselves. This is not fair. More important, since many
people pursue their interests in this way, the cost of protection for the
remaining ones may be so great that they do not think it is worth

354

ECONOMIC STABILITY AND GROWTH

while. For this reason, some private multiparty contracts that would
benefit all parties never get made. Therefore, provided it can be
shown that a community enterprise financed by appropriately apportioned taxes or assessments could convey to each person benefits that
he values more highly than he values his share of the cost, public policy
for this purpose improves the allocation of resources and is proper.
Indeed, in such a case, the public policy is merely an effective substitute for the private contract that does not get made. However, the
proviso is crucial: If it is not possible to apportion the costs of the
enterprise in such a way that everyone will value his benefits more
highly than his share of the costs, then the enterprise represents a
misallocation of resources, and public policy to create it must be justified, if at all, on some other ground.
Suppose it has been decided that the costs of a proposed public
enterprise can be apportioned in such a way that every individual
will experience a net gain. This of course may include situations
where some people lose by the enterprise itself, but can be compensated by the gainers so that everyone can still be better off on balance.
Then the question arises: Should the costs be so apportioned? Or
should they instead be apportioned equally or in some other simple
way, with a resulting transfer of wealth to the large gainers from
the small gainers and/or losers ? I would argue that if a large extra
cost were required in order to make each individual's net gain positive, and if the net losses of the losers under a simpler cost allocation
scheme were small, then the simpler allocation method ought to be
used, and the small wealth transfer ought to be tolerated. On the
other hand, if the wealth transfers would look large to the losers 5 1
would argue that compensation ought to be paid. Any substantial
wealth transfers that are made as a result of public policy ought to be
a part of the explicit wealth-redistribution policy, and be debated
and judged as such.
When a government decides to collect taxes or assessments for some
activity, the question still remains as to whether the activity should
be carried out by the government directly, or whether it should be
performed by some private agency or agencies under contract with
the government. This is not as important a question of principle as
the question of whether to plan and finance the activity publicly; it
should be decided essentially on grounds of efficiency.
A third case is that of resources whose competitive exploitation
would result in a waste of the resources. For example, when many
people own land lying above an oil deposit, it is in the private interest
of each person to drill and pump oil as fast as possible, before his
neighbors have time to get a very big share. But when everyone
does this, total costs will be greater and less of the oil in the deposit
will be recovered than if just a few wells were drilled in the right
places. Everyone could be better off under a unified drilling and
pumping operation, and there is a clear case for community action
requiring this. Fishing is another industry where this problem can
arise.
There are many public policies whose justification rests on the
premise that people are not well informed, either about their own true
interests, or about the alternatives open to them, or about the consequences of their actions, or about all of these. Eules requiring the
disclosure of information about the contents of food and drugs, and

ECONOMIC STABILITY AND GROWTH

355

about the nature of corporations whose stock is offered for sale, are
of this kind. So also are certain health, safety, and building regulations.
^ A more basic policy that is justified this way in part is the prohibition against a person's selling or mortgaging himself: Freedom is a
paramount value, and whenever a person feels that he wants to sell
himself for something else offered in return, he should be protected
against his own poor judgment. I feel that there is another quite
different justification for this policy, namely that human beings are
simply not like other property, that they have a dignity that demands
they not be treated so, even though it is true that human beings are
an important form of wealth, constituting a large share of total
wealth.
The policy of compulsory education and tax-supported schools rests
on several of the justifications suggested above. The compulsory
aspect is based in part on the presumption that parents who would not
provide for the education of their children are not sufficiently aware
that education is a powerful force for the welfare of those who have it,
both increasing their knowledge of what opportunities exist, and widening the range of opportunities open to them in particular. Both
compulsory education and tax-supported schools are based in part on
the disparity between private and total net benefits. If I provide for
the education of my children, you are made better off than if I do
not, whether you have children or not. The benefit to my family
alone, from the education of my children, may not be great enough
in my estimation to cover the cost, but the combined benefit to my
family and others is great enough so that if a suitable apportionment
of cost is made among us all, we can all get a net gain from the education of my children. And similarly with other children.
The tax support of schools makes sense for another reason. It is
clear that investment in education has a high payoff from a social
point of view, especially if particularly promising students are permitted to continue beyond the minimum prescribed level of education.
Some of these promising students come from families that do not have
the resources to send them to school. The capital market is not organized in such a way that loans for the purpose of education are easy to
get, partly because people are not permitted to mortgage themselves:
a promising but impecunious student has nothing besides himself
that he could mortgage. Hence tax support for schools is a way of
making investments in education that, productive as they are, might
not be made otherwise.
X. THE PROPER PROVINCE OF FREE CONTRACT

After taking notice, in the preceding section, of the area within
which government can properly modify the results that would come
from free contract alone, I now turn to argue that there remains a
significant field within which the allocation of resources under free
contract is better than if public policy concerns itself with the prices
of particular commodities or the quantities consumed by particular
persons (these are the first 2 of the 6 points stated at the beginning).
Clearly there is a significant field where individuals do know what
they want better than the government does, and seek it fairly intelligently; where they are reasonably well informed about their oppor23734—58

356

ECONOMIC STABILITY AND GROWTH

tunities; where resources are reasonably mobile between different uses;
where substantial competition prevails; and where effects of private
contracts on third parties are unimportant by comparison with what it
would cost to administer compensating payments. This field includes
a large part of the consumer goods area of the economy, and of the area
where techniques and capital goods and other resources are marshaled
for the production of desired goods and services. The mechanism by
which relative prices established under free contract allocate resources
is described briefly in section VI above; that section is an important
part of the argument here.
The economic history of the last hundred or 200 years supports
this argument in two ways. For one thing, it is possible to point to
many public policies that have sought, with public-spirited motives
but with unfortunate results, to do a better job of allocating resources
than free contract would do. When the FCC allocates valuable property in the form of television channels to private citizens, it would be
better to auction them off, and thus allocate them to people who have
the highest opinion of their value, and at the same time to collect
some revenue so that taxes could be lower, rather than giving them
away free on the say-so of officials—a procedure that provides strong
temptation to dishonesty. If regulation is needed to prevent railroads from exercising monopoly power, it should be administered in
such a way that a railroad can abandon a service that uses resources
more valuable than the service itself. Since the community wants to
assist low-income families who are in agriculture, it should do so in a
way that does not promote the production of additional farm products at costs greater than consumers are willing to pay. There are
many other examples.
The second way in which economic history supports free contract
in a wide field is in terms of the results of relying on it. I know of
no social and economic arrangement in the history of the world that
has succeeded in increasing the standard of living of the average man
as rapidly as has the system of private property and free contract that
has prevailed in the Western World. Other systems have demonstrated the capacity to challenge it in the rate of growth of total output, but not in the rate of growth of the output and distribution of
the things that consumers want.
XI. GOALS OF STABILIZATION POLICY

I now turn to my last three points, concerning stabilization policy.
I take it for granted that the maintenance of a high stable level of
employment is highly desirable, and that if we can attain it without
giving up other valuable things, we certainly should do so. I shall
argue that a stable price level is also desirable, not for itself, but because it is desirable that changes in the price level be accurately foreseen.
Consider first the relationship between flexible prices and unemployment. Suppose that the price level were perfectly flexible both upward and downward, so that whenever more goods and services were
demanded with money than supplied in the economy as a whole, the
price level would rise immediately to remove the discrepancy, and
so that whenever less goods and services were demanded with money
than supplied in the economy as a whole, the price level would fall

ECONOMIC STABILITY AND GROWTH

357

immediately to remove the discrepancy; then there would be no unemployment problem. (There would of course be some normal minimum amount of unemployment, perhaps 3 or 4 percent of the labor
force, associated with the transfer of workers from one job to another,
but this is no cause for concern.) Therefore from the point of view
of minimizing unemployment alone, perfect downward flexibility of
the absolute price level appears desirable. We do not have it, we
never have had it, and as far as I can see we never will have it, although we seem to have a fair degree of upward price flexibility, as
evidenced by the fact that general shortages of goods and services are
rare and brief except when they are created by the imposition of
price ceilings or quotas.
If the price level is to be quite flexible upward but not very flexible
downward, as appears to be the case, then, in the absence of public
policy to the contrary, it is likely that in periods of strong demand the
price level will rise, and in periods of weak demand there will be
more than normal unemployment while the price level falls only
slowly, as has occurred more or less regularly in the past.
Even if we could have a perfectly flexible price level, I believe we
would not want to rely on it, because in practice rapid changes in the
price level are bad. Why are they bad ? Only because they cannot
be accurately foreseen by everyone. When people make contracts
whose performance will stretch out over time, such as loans, rentals,
insurance, annuities, pensions, etc., or even when they simply hold
cash, they cannot do so wisely unless they know what the price level
will be at relevant points of time in the future as compared with its
present level. For example, suppose I lend you $100 today, to be
repaid after a year, and suppose we want to agree on terms that will
make it possible to buy 1.05 times as much a year from now, with the
principal and interest that you are going to pay me then, as can be
bought now with the $100 I am about to lend you. If we know that
the price level will not change during the year, we will agree that
you are to repay me $105, which is $100 with 5 percent interest. If
we know that the price level is going to double during the year, we
will agree that you are to repay me $210, which is $100 with 110 percent
interest. If we know that the price level will fall to half its current
value during the year, we will agree that you are to repay me $52.50,
which is $100 with minus 4 7 ^ percent interest. As long as the future
behavior of the price level is known, it does not matter much what
it is going to be; we can adjust our loan contract to allow for it. But
if we don't know what will happen to the price level, then a contract
for the repayment of $100 at an interest rate of, say, 5 percent is of
unknown worth in terms of real purchasing power * * * it will be
worth more if the price level falls than if it rises. An escalator
clause could be written into the contract, as is done with some wage
contracts, but this device is cumbersome, and it could not very well
be used to protect the holder of cash from fluctuations in its real
value.
The same principle applies to any other contract stretching over
time: if the future course of the price level is known, then the contract can be written to take account of these changes, and thus to provide for the payment of the amount of real purchasing power that
the parties to the contract want to agree upon.

358

ECONOMIC STABILITY AND GROWTH

If the real value of future payments to be made under such contracts is uncertain, because future price levels are uncertain, then the
use of such contracts will be discouraged, with a consequent misallocation of resources. Also, if the price level rises unexpectedly, there
will be an unexpected transfer of wealth from lenders to borrowers,,
which is unfair, and vice versa if the price level falls unexpectedlyy
which is also unfair.
The argument for a stable price level, then, is essentially an argument against imperfectly foreseen changes in the price level. It
is much easier to persuade people that the price level will be stable in
the future than to persuade them that it will rise or fall at some definite
rate, or that it will follow some prescribed path. An additional advantage of a stable price level over a varying one, assuming that foresight is equally good in both cases, is that with a stable level it is not
necessary to do quite so much computing in order to write out a contract that does what one wants in real purchasing power terms. Further, there is a large volume of outstanding contracts for insurance,
loans, etc., that were presumably made on the expectation that the
price level would stay approximately constant, and there is no equitable reason to transfer wealth from the lenders to the borrowers
under these contracts, or vice versa, by changing the price level now.
In view of the foregoing argument, I believe it is preferable to
maintain an approximately stable price level rather than having a
flexible one, and to rely on monetary and fiscal policy to prevent rises
in the price level during periods when demand is strong and to prevent unemployment and declines in the price level in periods when
demand is weak.
XII. STABILIZATION POLICY I N PRACTICE

A rising price level together with full employment is due to the
fact that more goods and services are demanded with money than
are supplied by resource owners at the existing price level. Unemployment is due to the fact that fewer goods and services are demanded
for money than are supplied by resource owners at the existing price
level. The cure for the former is to diminish the demand for goods
and services, and the cure for the latter is to increase it. The Federal
Government has several tools in its policy arsenal that can do this
without price controls or other serious interferences with resource
allocation as established by free contract. These include the following (I describe them in terms of what one wants to do in a case of
unemployment; in a case of inflation the reverse of each would apply) :
(i) Increase the quantity of money (deposits and currency) in the
hands of the public without doing anything else—this is perhaps
done most simply by a Government deficit produced by either a cut
in taxes or an increase in transfer payments to the public (particularly the unemployed).
(11) Reduce interest rates and increase the quantity of money in
the hands of the public through open market purchases of outstanding Government securities by the central bank. (A reduction in
reserve requirements for commercial banks makes possible an essentially similar operation in which banks buy the public's promissory
notes.)

ECONOMIC STABILITY AND GROWTH

359

(iii) Increase Government purchases of goods and services by
means of deficit financing, thus creating a direct demand for production and increasing the quantity of money in the hands of the public.
Of these three, the first already occurs to a considerable extent
automatically, due to the operation of various built-in stabilizers
such as progressive income taxes and unemployment benefits. All
three can be applied in a deliberate discretionary fashion, however.
I believe that if they are pursued vigorously, as soon as substantial
unemployment becomes evident, they can cure any depression
promptly; and of course if they are pursued too vigorously and too
long, they can turn any depression into an inflation. Similarly, I
believe that, if they are pursued vigorously in reverse, they can stop inflation promptly; and if overdone can turn any inflation into a depression.
I do not believe that the great depression of the 1930's affords a fair
test of the effectiveness of countercyclical monetary and fiscal policy,
because either wrong or too weak policies were followed in the first
4 years of the decline. For one thing, the quantity of money in the
hands of the public was permitted to undergo an almost steady decline
of about 25 percent from the fall of 1929 to the middle of 1933; it
should have been kept fairly stable instead. Also, shortly after industrial production had begun to rise in the early spring of 1931, and
then had turned down again the summer, the Federal Reserve raised
its discount rate from iy2 to Sy2 percent, and then maintained it at
or above 2% percent for the next 2 years, when it should have been
kept low. Thus it was only after the depression had become very
deep, and private spending had fallen drastically, that public policy
began to come strongly to the rescue. This may explain why it was
not very effective when it did come.
The declines in 1948-49 and 1953-54 were much milder, especially
the second one. I believe that this was partly because in both of them
public countercyclical policy was more appropriate. In the 1948-49
decline, taxes were cut in the latter half of 1948, and the Federal
Reserve at least did not raise the discount rate after the decline
began (though it did not lower it). In the 1953-54 decline, taxes
were cut early in 1954, and the Federal Reserve lowered the discount
rate quite promptly and kept it low until the spring of 1955.
If I may be permitted to express the opinion, it seems to me that
early in February of 1958, when the January unemployment figure
of 4.5 million came out and showed an increase of 1.1 million from
December, about twice as large as the expected seasonal increase at
that time of year, it was then time for a tax cut.
There is one cloud on the horizon of the picture I have painted of the
effectiveness of stabilization policy, and that is this question: If unemployment is to be kept fairly stable at a level as low as 4 percent of
the labor force (which is approximately the postwar average), can the
price level be kept from rising? Or put another way, if the price
level is to be kept stable, can unemployment be kept as low as about
4 percent of the labor force ?
If one compares the histories of recent depressions, one finds it hard
to give a clear "ves" answer to this question. In depressions between
the two World Wars, prices apparently fell more readily than since the
second war. And if one compares the three postwar recessions, which
(at this writing) are about equally severe, the current one being a

360

ECONOMIC STABILITY AND GROWTH

bit deeper than the others, one finds the following declines of the
Wholesale Price Index, the Consumer Price Index, and average
hourly earnings in manufacturing, each measured from its own peak
month to its own trough month:

Wholesale Price Index
Consumer Price Index
Average hourly earnings in manufacturing

1948-49

1953-54

1957 (latest)

Percent
-11
-4
-1

Percent
-1
-1
-1

Still rising (February).
Do.
—ft percent (February).

This suggests that prices are becoming less and less flexible downward. If so, and if prices rise at all in the prosperous periods that
intervene between recessions, then we are in for a gradually rising
price level on a sort of ratchet principle: little or no change during
recessions? and increases in between. If this is really the case, I
believe it is tolerable if the price level does not rise more than, say, 20
percent in a generation. I would prefer to accept an average price
increase of 1 percent a year, rather than having to accept an average
level of unemployment of 5 percent instead of 4 percent, if the choice
came down to that, because the price rise is largely redistributive,
whereas the extra unemployment represents foregone production of
goods and services.

INFLATION, THE WAGE-PRICE SPIRAL AND ECONOMIC
GROWTH
Otto Eckstein, Harvard University
The relation between economic growth and inflation is not a simple
one. Investment adds to effective demand, helping to drive up prices;
it also adds to the capacity of the economy, making possible an increased supply of goods in the future. Inflation increases the opportunities for profit, but may discourage saving and distort the directions of investment. This paper examines this relationship, primarily emphasizing postwar experience in the United States. Most
of our attention will be devoted to the effect of investments and growth
on inflation,
a relationship which has been discussed less fully than the
reverse.1 Much of our analysis will run in terms of the movements
of the wage-price spiral.
I. LONG-RUN TRENDS OF OUTPUT AND PRICES I N ADVANCED ECONOMIES

The ambiguity of the relationship is brought out most clearly by the
long-run experience of many countries. Table I, which draws on the
long-run growth data assembled by S. Kuznets, gives the rates of
growth of output per decade for 8 countries, and the rates of price
change that accompanied the growth. It can be seen that periods
of rapid growth occurred with and without inflation, and that periods
of stagnation also saw a very wide range of price changes. Thus, as
a long-run phenomenon,
there is no historical association between
growth and inflation.2
!But see A. Smithies, The Control of Inflation, Review of Economics and Statistics,
August 1957, pp. 272-281, and S. Slichter, Thinking Ahead: On the Side of Inflation,
Harvard
Business Review, September-October 1957, p. 15.
2
For the United States, the correlation coefficients for the two series of decade-by-decade
changes of prices and output are 0.08 and —0.05. For most of the other countries, the correlations are dominated by extreme values of inflation associated with little growth.

361

362

ECONOMIC STABILITY AND GROWTH

TABLE 1.—Rates of growth of output per decade and rates of change of prices
per decade in some advanced countries
From preceding decade
Decades

Output
growth

Percent price
change

UNITED STATES

Output
growth

Percent price
change

UNITED STATES

1869-78 _ _ _ _ _ _
1879-88
1889-98
1899-1908
1909-18
1919-28 .__
1929-38
1939-48
1950-541

88.0
38.2
56.4
35.5
39.2
6.2
71.7
29.0

-19.5
-12.9

+9.3
+34.6
+46.3
-18.0
+34.2
+34.5

UNITED KINGDOM

1874-83 _ _ .
1884-93
1894-1903
1904-13
1914-23
1924-33
1934-43
1944-53

54.8
45.2
50.6
29.0
29.1
26.0
52.0

-16.0
-7.1
+20.7
+64.0
0
-7.1
+84.8

4.2
19.4
11.8
30.7
9.3
40.6
8.8
23.9

+13.1
-8.7
-2.6
+11.4
+260.0
+35.0
+18.5
+4442.4

ITALY

1865-74
1875-84
1885-94
1895-1904
1905-14
1915-24
1925-34
1935-44
1949-53

21.1
37.6
29.2
16.5
-0.8
21.1
30.2
22.7

-15.0
-0.9
+9.6
+106.0
-16.6
+17.1
+51.3

1864-73
1874-83
1884-93
1894-1903
1904-13
1914-23
1924-33
1934-43
1950-54
SWEDEN

JAPAN

1883-92
1892-1902
1903-12
1913-22
1923-32
1933-42
1950-541

64.8
34.7
45.3
67.4
52.1
10.6

+35.4
+38.2
+86. 5
-7.1
+28.6
+18856.3

NORWAY

1864-73
1874-83
1884-93
1894-1903
1904-13
1914-23
1924-33
1934-43
1950-541

+7.5

30.6
21.7
40.5
40.3
25.1
23.2
36.3
60.1

+2.7
+13.9
+99.2
-11.3
+13.0
+62.4

24.3
38.4
41.9
35.8
36.7
16.1
28.5

-6.2
0
+3.5
+126. 5
-26.7
+25.6
+119.3

-12.0

DENMARK

1900-1908 . _
1909-18
1919-28
1929-39
1950-54

33.0
29.1
38.1
33.5

+71.9
+60.8
-37.5
+174.6

25.1
40.1
.2
33.7

+58.3
-12.7
-7.1
+145.4

NETHERLANDS

1904-13
1914-23
1924-33
1934-43
1950-541

From preceding decade
Decades

1874-83
1884-93
1894-1903
1904-13
1914-23
1924-33
1934-43
1950-54

__ _

i Last figure for all countries is not a decade-by-decade comparison.
Source: S. Kuznets, Quantitative Aspects of the Economic Growth of Nations, Economic Development and Cultural Change, vol. V, No. 1.

The explanation of these figures is quite simple. First in the late
d

Full text of The Relationship of Prices to Economic Stability and Growth : Compendium of Papers Submitted by Panelists Appearing Before the Joint Economic Committee

FRASER