View original document

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

[SUBCOMMITTEE PRINT]

THE FEDERAL RESERVE'S ATTACHMENT
TO THE FREE RESERVE CONCEPT

A STAFF ANALYSIS

SUBCOMMITTEE ON DOMESTIC FINANCE

COMMITTEE ON BANKING AND CURRENCY
HOUSE OF REPRESENTATIVES
88th Congress, 2d Session

Printed for the use of the Committee on Banking and Currency
U.S. GOVERNMENT PRINTING OFFICE
2W7S




"WASHINGTON : 1964

COMMITTEE

ON BANKING

AND

CURRENCY

W R I G H T P A T MAN, Texas, Chairman
C L A R E N C E E. K I L B U R N , New York
A L B E R T RAINS, Alabama
WILLIAM B. WIDNALL, New Jersey
ABRAHAM J. M U L T E R , New York
E U G E N E SILER, Kentucky
WILLIAM A. B A R R E T T , Pennsylvania
PAUL A. PINO, New York
LEON OR K. SULLIVAN, Missouri
F L O R E N C E P . D W Y E R , New Jersey
H E N R Y S. REUSS, Wisconsin
SEYMOUR H A L P E R N , New York
THOMAS L. ASHLEY, Ohio
J A M E S H A R V E Y , Michigan
CHARLES A. VANIK, Ohio
OLIVER P . BOLTON, Ohio
WILLIAM S. MOORHEAD, Pennsylvania
W. E. (BILL) BROCK, Tennessee
R O B E R T G. STEPHENS, JR., Georgia
R O B E R T T A F T , JR., Ohio
F E R N AND J. ST GERMAIN, Rhode Island
H E N R Y B. GONZALEZ, Texas
J O S E P H M . McDADE, Pennsylvania
CLAUDE P E P P E R , Florida
S H E R M A N P. LLOYD, Utah
JOSEPH G. MINISH, New Jersey
B U R T L. T A L C O T T , California
CHARLES L. W E L T N E R , Georgia
D E L CLAWSON, California
R I C H A R D T . HANNA, California
B E R N A R D F. GRABOWSKI, Connecticut
CHARLES H. WILSON, California
COMPTON I. W H I T E , Ja., Idaho
JOHJT R. STARK, Clerk and Staff Director
JOHN E. BARRIERK, Professional Staff Member
Alvis L E E MORSE, Counsel
ORMAN S. FINK, urmoritt Staff Msmbtr

SUBCOMMITTEE ON DOMESTIC FINANCE
W R I G H T P A T M A N , Texas, Chairman
H E N R Y S. REUSS, Wisconsin
WILLIAM B. WIDNALL, New Jersey
CHARLES A. VANIK, Ohio
CLAUDE P E P P E R , Florida
J O S E P H G. MINISH, New Jersey
CHARLES L. W E L T N E R , Georgia
R I C H A R D T. HANNA, California
CHARLES H. WILSON, California
II




JAMES H A R V E Y , Michigan
OLIVER P . BOLTON, Ohio
W. E. (BILL) BROCK, Tennessee
R O B E R T T A F T , JR., Ohio

LETTER OF

TRANSMITTAL

MAT 2, 1964.
To the Members oj the Subcommittee on Domestic Finance:
Transmitted herewith for the use of the subcommittee is a staff
analysis of the role of the "free reserves" concept in Federal Reserve
policymaking. The Federal Reserve has indicated that "free reserves"
is accorded a central position as an objective of monetary policy.
The study raises critical and stimulating questions about the validity
of assigning so important a role to this concept.
This analysis is the second part of a three-part study of the guidelines used by the Federal Reserve in formulating its policy. The
first installment, published on February 10, analyzed some of the
concepts used by the Open Market Committee in executing monetary
policy and proved very useful to subcommittee members. Because
this installment is equally relevant to the current hearings on the
Federal Reserve System, it is being published separately at this time.
It should be pointed out that many of the Open Market Committee's
activities are conducted in secret and therefore not known to the
Congress or to this subcommittee. For this reason, it is necessary for
staff investigators and other students of monetary policy to rely on
general statements of Federal Reserve authorities, announced policy
actions, reasons given for such actions by Federal Reserve officials,
and the independent analysis of factual information.
Sincerely,




WRIGHT PATMAN.

m




LETTER

OF

TRANSMITTAL

H o n . WRIGHT PATMAN,

Chairman, House Banking and Currency Committee,
House oj Representatives, Washington, D.C.
DEAR M R , CHAIRMAN: The material transmitted herewith represents the second part of a three-part study, "An Analysis of Federal
reserve Monetary Policymaking," that is being prepared for the
committee. These chapters attempt to develop in more detail the
nature of the conception that guides Federal Reserve actions, the
of free reserves in their analysis, and the relevance of the
modified free reserve conception as an explanation of changes in
money and credit.
f ^T? s , u b c o m m i t t e e print of February 10, "Some General Features
th I ' e<*era* Reserve's Approach to Policy," we discussed some of
™ diverse and often disconnected strands that play a prominent
role m policy discussions by spokesmen and officials of the System,
we found that one recurrent theme appears to occupy a dominant
position in their notions about the monetary process: the asserted
relation between the level of free reserves and the rate of credit
expansion. While this relation has by no means been developed into
a coherent frame for analysis of the monetary process, the persistent
references suggest that it occupies a central position in the Federal
•Reserve's antdysis.
section 1, ' T h e Federal Reserve's Attachment to the Free Reserve
v>oncept: Evidence From Published Statements," develops the
meaning of the free reserve doctrine as it is used within the System.
£_oree separable roles are assigned to free reserves. First, ana most
important, free reserves are regarded as an important causal factor
anecting credit expansion ana contraction. Second, free reserves
are used as an indicator of a given monetary situation. Third, they
a*e used as a target of Federal Reserve policy. Much of the confusion
Rendered by the contemporaneous denials and explicit affirmations
oi tiie role of free reserve results from the three uses of the term,
or example, the use of free reserves as a target may be denied without
meeting the role of free reserves as the centerpiece of the causal
mechanism used by the Federal Reserve.
an u sect, *? n indicates that the free reserve conception evolved out of
* older notion developed most coherently by Riefler in the twenties,
*
of flkaracter of this evolution is of particular importance, since many
the notions that formed a part of older views continue to dominate
thinking, long after their rationale has disappeared. Moreer, the discussion of the evolving notions helps to explicate some of
actions ofe v e r e lFederal Reserve in the thirties and forties. For
the
s
y deflationary action of the mid-thirties, the doublin^ f '
.
of reserve requirements, was appraised by the Federal Reserve
foil
of the dominant Riefler notion. Had this notion been well
unaed, and applicable to the then current events, the doubling of




VI

LETTERS 01- TRANSMITTAL

reserve requirements would not have bad severely deflationary consequences. But tbo not ion was inapplicable and incorrect.. Increased
unemployment and a reduction in the pace of economic activity ensued.
.
. .
Dramatic evenls, like tbe doubliug of reserve requirements in tne
thirties, do not take place at frequent intervals. But they serve to
indicate the importance of a validated conception of the monetary
mechanism and to illustrate the costs to society resulting from the
application of invalid, untested theories.
The remainder of section I presents evidence from published Federal
Reserve statements and from the as yet unpublished responses, by the
members of the Board of Governors and the presidents of the Reserve
Banks, to a series of questions posed in connection with this study.
The recent modifications of the doctrine, associated with the recognition that the demand for free reserves play a role in the monetary
process, are discussed. We find that the recognition of demand factors
has introduced important additional elements into the Federal Reserve's analysis.
A major conclusion of the section is that many of the new notions
that have been introduced conflict with older views. These conflicts
remain unresolved within the Federal Reserve, llad the Federal
Reserve attempted to develop and test their conception, many such
conflicts would be recognized, and resolved 011 the basis of evidence;
analysis and understanding would be improved, and the foundations of
monetary policy making would be strengthened. The failure of the
Federal 'Reserve to carry out systematic appraisals of the mechansim
that has been entruste3 to tlicir control perpetuates incorrect and
poorly developed views, and renders monetary policy less useful as a
tool of economic policy.
In contrast to section 1 that looks at. the evidence from the professed
views of oflicials and spokesmen, section 2 concentrates principally
on the actions taken by the Federal Reserve." We noted in chapter II
that the Federal Reserve has an extremely short-run policy focus,
that actions are taken in response to weekly, daily, and even hourly
events on the financial markets. Section 2 builds 011 the earlier
discussion and reveals the way in which concentration on short-run
occurrences and the absence of systematic analysis leads to a substantial grant of authority to the Manager of the System Open
Market Account.
A principal piece of supporting evidence for our view of the position
of the Manager, his reliance on free reserves as an indicator, and his
use of free reserves as a target, is of particular importance. A comparison of decisions by the Federal Open Market Committee with
the recorded movement of free reserves indicates that the level of
free reserves quite often moved decisively in advance of a decision
by the Open Market Committee to "ease" or "restrain." The
observed pattern strongly supports our interpretation that the absence of a systematic framework and the concentration on extremely
short, run market events has resulted in a substantial grant of authority
to the Manager. At major turning points, in the post-Accord period,
it, has often been the Manager's action that reversed the direction
of policy. This action was tlien ratified at a meeting of the Federal
Open Market Committee. Contrary to published - statements by
oflicials of the System, the Manager appears to occupy a major
policymaking role.



LETTERS OF TRANSMITTAL

VII

Our appraisal of the Federal Reserve's record at post-Accord turning
points in economic activity suggests that, in this respect, the Federal
Reserve litis compiled a good, even excellent record. They have been
alert and sensitive to a variety of indicators, and the}^ have made
timely and appropriate judgments about the pace of economic activity. We contend that in the present state 01 knowledge, it would
be difficult to improve upon their record in this respect.
However, recognition of the turning points in the pace of economic
activity must be accompanied by appropriate action to alter the money
supply, if discretionary monetary policy is to have an appropriate
countercyclical influence. The Federal Reserve does not directly
control the stocks of money or credit. Without a valid conception of
the relation between their actions and the stock of money ana credit,
the usefullness of their judgment of the timing of turning points in
economic activity is diminished. Unless appropriate action is taken,
our economy does not benefit from their timely recognition of the
turning point.
The type of action that is taken depends on the conception of the
monetary process that is held. If that conception is seriously deficient,
it is quite likely that correct judgment of turning points will not be
accompanied by action appropriate to reduce unemployment or to
prevent inflation. Section 3, "The Relation of Free Reserves to
Changes in Money and Credit/' therefore provides some evidence on
the validity of the modified free reserves conception.
The evidence is quite clear. The modified free reserves mechanism
hears almost no relation to changes in the stock of bank credit or
money. Indeed, the relation is so poor that it raises questions about
the usefulness of Federal Reserve policy as a means of controlling
money or credit. Judged in terms of the Federal Reserve conception,
an overwhelming proportion of observed changes in money and credit
are outside the control of the Federal Reserve. If their view of the
monetary mechanism were the only admissible view, we would be
forced to concede that monetary policy is little more than a futile
Exercise.
Fortunately, alternative conceptions of the monetary mechanism
substantially greater validity can be formulated. One such
conception will be presented in a later chapter. It will suggest that
federal Reserve policy has an important influence on the stock of
money. But to obtain this influence, i.e., to carry out the congresS1
°nal mandate, the Federal Reserve must abandon the modified free
reserves conception and operate in terms of a markedly different
inception.
.. The evidence presented in Section 3 supports our contention that
the conception that dominates Federal Reserve discussion and thinkabout the monetarv mechanism is woefully inadequate and without
factual foundation, t h e failure of the Federal Reserve to develop an
^equate appraisal of the monetary mechanism seriously reduces the
Jue of their recognition of turning points, leads to inappropriate
P°ljcy actions and renders discretionary monetary policy a less useful
l
°ol for carrying out the congressional mandate.




KARL BKUNNER.
ALLAN H . MELTZER.




CONTENTS
Letter of transmittal to the members of the Subcommittee on Domestic
Finance
Letter of transmittal to Hon. Wright Patman, chairman, Committee on
Banking and Currency
—
Section 1. Evidence from published statements
Genesis and development of the free reserve conception of monetary
processes
Riefler's contribution to the analysis.
Burcess' views.
Golden weiser's views
-—
The causal position of free reserves: Free reserves as an index of a
monetary situation and free reserves as a policy target
Free reserves as a policy target..
The indicator function of free reserves
Evidence from published statements
Some clues suggesting recent developments of the free reserve
conception
Clues from published statements
Clues from answers to questionnaires
Policy in some specific contexts
Table I I I — f e d e r a l funds rates and free reserves for selected
months of 1961-62.
Summary, the modified free reserves doctrine
t
oection 2. Evidence from announced changes in policy
The FOMC and the Manager
The directive
The consensus
.
Policy objectives and free reserve levels
The FOMC and free reserves at post-Accord turning periods
1. Peak dated July 1953
2. Trough dated August 1954
3. Peak dated July 1957
4. Trough dated April 1958
5. Peak dated May 1960
6. Trough dated February 1961
----Table IV-1. A summary of policy actions and movements of free
reserves at post-Accord turning points
The FOMC and free reserves in 1963-63
Table IV-2. Scaling of the policy record for 1962
Table IV-3. Free reserve ranges during 1962 and early 1963
Appraisal of the post-Accord record of policy actions
The timing of policy changes
The autonomy of the Manager
—
.
Free reserves as an indicator of desired ease and restraint
Summary
-

2^-678—64




2

in
v
1
2
3
7
8
10
12
15
17
21
21
22
25
27
28
31
31
32
34
36
37
38
38
38
39
40
41
42
42
43
43
43
44
47
48
49

X

CONTENTS
Page

Section 3. The relation of free reserves to changes in money and credit—
The demand for free reserves by banks
_
The percentage change in deposits resulting from the interaction of
the demand for and supply of free reserves
Money, credit, and the modified free reserves doctrine
Table V - l . Measures of the relationship between monthly average free reserves and changes in money supply and member bank
credit outstanding
Table V-2. Measures of the relationship between annual moving
averages of free reserves and percentage rates of change in
money and credit
Table V-3. Measures of the relationship between monthly changes
in money and credit, and level and distribution of free reserves
and interest rates
Table V-4. Measures of the relationship between monthly
changes in money and credit, and level and distribution of free
reserves and interest rates, 3 postwar periods
Table V-5. Measures of the relationship between monthly
changes in money and credit, and the level and distribution of
free reserves and interest rates in months of expanding and
contracting economic activity
Summary and conclusion




51
52
54
56
58
59
59
60

61
63

THE FEDERAL RESERVE'S ATTACHMENT TO THE
FREE RESERVE CONCEPT
SECTION 1 — E V I D E N C E FKOM PUBLISHED STATEMENTS

One of the dominant Federal Reserve conceptions centers on the
role of free reserves in the monetary process. This idea has had an
important- influence on assessments made by the Federal Reserve
authorities and on the policies applied in concrete situations. We
contend that the Federal Reserve has viewed, and continues to view,
free reserves as an element playing a causal role of central importance
m the monetary process and simultaneously supplying a useful
summary measure of "ease and restraint." Detailed evidence is presented in this and a subsequent chapter in support of our contention.
The evidence presented in this chapter comes almost exclusively
from the statements made by members of the Board of Governors,
the FOMO, and their staffs. It would be useful, perhaps, to supplement these statements by indications of the importance attached to
free reserves, both as an index and as a causal factor, by Members
of the Congress, the banking community, the academic profession and
others. But to do so seems beyond the scope of this inquiry and adds
little direct evidence to the point. In the following section, a second
type of evidence will be presented, evidence from the record of policy
actions and the actual movements of the level of free reserves in the
postwar period.
Free reserves are defined as the difference between measured
excess reserves and member bank borrowings. They are the volume
of measured excess reserves not borrowed from the Federal Reserve.
To obtain the volume of free reserves, the amount of required reserves
and the amount of member bank borrowing axe subtracted from total
reserves.
In the context of the dominant notion to be considered, the level
of
free reserves is viewed as a causal factor affecting the rate at which
commercial banks adjust their portfolios of earning assets. An
to crease in free reserves is expected to accelerate the expansion rate
of bank portfolios; i.e., to increase the rate at which banks acquire
earning assets or decrease the rate at which they unload securities
and/or compress outstanding loans. A decline in free reserves, on
the other hand, is expected to retard the expansion rate; either the rate
°f acquisition will be lower or portfolios contract. A systematic
association thus links the level of free reserves with the rate of change
the commercial banks' portfolio of earning assets.
The association of free reserves with the banks' asset expansion,
combined with the causal role assigned to free reserves, influenced
tte choice of free reserves as an index of the monetary situation—
a
summary measure indicating relative "ease" or "restraint.
According to the Federal Reserve's notion, policy actions and other
e
vents modify the monetary process to the extent that they change-




1

FREE

RESERVE

CONCEPT {>3

the prevailing level of free reserves. Open market operations immediately change total reserves and free reserves by the same amount.
Banks respond by modifying the adjustment rate of their portfolios.
As a result, required reserves and borrowing change, and the initial
impact of open market operations on free reserves is attenuated.
However, free reserves do not return to their initial level even after
the impulse triggered by open market operations has been fully
absorbed.
Changes in reserve requirements immediately change the volume
of required reserves relative to an unchanged volume of total reserves.
There is no direct effect on borrowings from Federal Reserve Banks.
Thus there is an instantaneous change in free reserves that affects
the banks' portfolio adjustment. Subsequently, part of the initial
change in free reserves is absorbed, via the gradual shift in required
reserves associated with changes in deposits induced by the portfolio
adjustments.
The effect of other events, gold flows, currency flows, a redistribution of the public's deposits between demand and time accounts,
the division of the Treasury's balances between holdings at Federal
Reserve Banks or on tax and loan accounts at commercial banks, etc.,
can be traced in a similar manner. A necessary and sufficient condition for all these events to exert an influence on the money supply
and credit markets is the existence of an immediate impact on the
level of free reserves. This, in essence, is the foundation of the free
reserve conception of the monetary process.
GENESIS AND DEVELOPMENT OF T H E F R E E R E S E R V E
MONETARY PROCESSES

CONCEPTION OF

A short description of genesis and development of the "free reserve
doctrine" seems appropriate before we discuss the evidence supporting
our contention about the dominant role of this "doctrine" among
Federal Reserve views.1 This description focuses attention on the
^behavior patterns and operating problems that stimulated the development of the free reserve doctrine. Recent emergence of important
modifications will also be noted. However these modifications of the
free reserve doctrine have neither been systematically developed nor
absorbed into a coherent view. Conceivably, these new elements will
lead, in the future, to a reassessment of the Federal Reserve's viewpoint
about the role and significance of free reserves.
The origins of the free reserve "doctrine" may be traced to the
discovery of open market operations. Such operations emerged in
the early twenties as a result of the Reserve Banks' endeavor to
bolster their revenues with suitable earning assets. The Federal
Reserve authorities rapidly realized that open market operations
immediately affect the commercial banks' reserve position. Purchases
inject additional reserves, and sales siphon off available reserves.
Open market operations thus appeared to offer an excellent opportunity to modify the commercial banks' positions in the direction and
* The reader may usefully consult A. I . Meigs, "Free Reserves and the Money Supply" (Chicago: University of Chicago Press, 1962), ch. 2. A detailed exposition and analysis of several free reserve theories is
also presented in our forthcoming paper "Evolving Federal Reserve Conceptions About the Monetary
Process/' It is shown in this paper that a number of typical assertions made by Federal Reserve authorities
can be subsumed under these theories. Furthermore, the explicit construction of usually vague and elliptical notions permits a detailed empirical evaluation of their comparative validity. Subsequent references
will be made simply to ** Evolving * • *




{>3

FREE RESERVE CONCEPT

extent desired by the Federal Reserve authorities. But the gradually
accumulating observations concerning the commercial banks' reserve
and borrowing behavior slowly dispelled this belief. Open market
operations typically induced a response in bank borrowing from the
Federal Reserve Banks which seriously mitigated the impact of open
market operations on the total volume of reserves. Purchases
generated, a repayment of outstanding loans and sales "forced banks
into the central bank" and induced an increased volume of borrowing.
Open market operations were systematically associated with offsetting
variations in the banks' borrowing from the Federal Reserve Banks.
The banks' behavior almost annihilated any potential effect of
open market operations on the volume of bank reserves in the twenties.2
But the composition of reserves between borrowed and unborrowed
reserves changed decisively. Open market sales by the Federal
Reserve gave rise to increased borrowing by member banks that
approximately restored total reserves. Excess reserves were small
during the period and exhibited negligible variations. Hence varia*
tions in bank indebtedness were practically equivalent to variations
in free reserves. (Of course the values of free reserves and member
bank indebtedness moved in opposite directions as would be expected
from the definition of free reserves and the relatively unchanging
value of excess reserves.)
Under the circumstances of the period, the volume of bank reserves
did not appear to play any important role in the transmission of
the impulses set off by open market operations. Students of monetary
policy were seemingly forced to recognize the futility of open market
operations or to search for an alternative route transmitting the impact
generated by open market operations.
RIEFLER's CONTRIBUTION TO THE ANALYSIS

The behavior patterns summarized above appeared to hint at an
alternative route which was explored in a pathbreaking study of
our monetaiy system by W. Riefler.3 This alternative view made
variations in banks' indebtedness the focal point of the money supply
and credit market process. But Riefler fully realized that bank
indebtedness could only assume an important role in explaining the
effective transmission of Federal Reserve jpolicy if the banks had little
control over the volume of their indebtedness to the Reserve Banks.
In the early part of his book, therefore, he devoted much attention
to the rationale for member bank borrowing. Two hypotheses are
' Two regressions were computed in order to appraise the order of magnitude and significance of com{ g a t i n g variations In the Federal Reserve's "discounts and advances," denoted by A for the period of the
i«20 a. The regressions relate first differences of A between adjacent months or corresponding months of
Jdjaoent years with similar first differences of the adjusted base B*. The latter magnitude is equal to the
oase minus discounts and advances. It is thus equal to the sum constituted by the Federal Reserve s
Portfolio of securities including float, the Treasury's gold stock net of Treasury cash, Treasury currency
outstanding, the negative value of Treasury deposits and foreign deposit at Federal Reserve Bants, ana
some. minor other accounts including "other deposits" at Federal Reserve Banks. The result of the re__ — a w / u u u u
Sessions are collect cd below:
AA(t r-i-.00S6-.8742AB*t, M
(.0070) (.0962)
3697
where

AAt, r-i**At~At~i't

AAtt t^tt*'At"Afiil

059
i-i**-B*i

t~n**

j

Sample period: January 1918 to December 1929.
^ d Money Markets in the United States" (New York: Harper * Bros.




FREE RESERVE CONCEPT {>3

considered, a profit hypothesis and a "needs and reluctance" hypothesis. The alternative views may be summarized in his own words:
The most obvious theory is that member banks, on the
whole, borrow at the Reserve Banks when it is profitable to
do so and repay their indebtedness as soon as the operation
proves costly. The cost of borrowing at the Reserve Banks,
accordingly, is held to be the determining factor in the relation between Reserve Bank operations to money rates, and
the discount rate policy adopted by the Reserve Banks to be
the most important factor in making Reserve Bank policy
effective in the money markets. At the other extreme,
there is the theory that member banks borrow at Reserve
Banks only in case of necessity and endeavor to repay their
borrowing as soon as possible. According to this theory the
fact of borrowing in and of itself—the necessity imposed by
circumstances on member banks for resorting to the resources
of Reserve Banks—is a more important factor in the money
market than the discount rate * * * and open market
operations * * * contribute more directly to the effectiveness 4of Reserve Bank credit policy than changes in discount
rate.
After the sketch of the two theories, Riefler discusses the implications of the profit theory and notes that under this theory interest
rates on the open market should be 5close to the discount rate, with
only minor or transitory deviations. A confrontation of this conclusion with observable rate behavior on short-term open markets
leads Riefler to reject the profit theory.
We may easily concede the pertinent facts and admit that the open
market rates diverged markedly and persistently from the discount
rate. On the other hand, the behavior of the acceptance market was
consistent with the implication of the profit theory as formulated by
Riefler; i.e., the pertinent market rates followed closely6 and deviated
only little from the Federal Reserve's acceptance rate.
Riefler concludes his appraisal of the rival conceptions concerning
the process generating the banks7 indebtedness to Federal Reserve
Banks with a decision in favor of the "needs and reluctance" theory
of bank borrowing:
There is little question, on the whole, that the first of
the two theories outlined above, covering the relation of
Reserve Bank rates to money rates in the money markets,
applies to the rates at which acceptances sell in the open
« Ibid., pp. 19-20.
* "If member bank borrowing has been governed primarily by motives of profit during this period, money
rates would have been dominated by the discount rates ch arged by the Reserve Banks. Particular!y would
this have been true of rates in the short-term open markets where member banks can lend freely and withdraw funds entirely on their own volition without regard to the results of their actions on future lending
.operations/ As member banks had plenty of eligible paper on which to borrow at the Reserve Banks during
most of this penod, there was nothing to prevent them from 'scalping' a profit out of the open market whenever rates m those markets were above discount rates. If member bank borrowing had actually been
governed by the profit motive m this manner, ofTers of additional funds in the short-term open markets
yould_have been so plentiful whenever opportunity presented itself that rates in those markets could never
have risen far above discount rates, so long as eligible paper continued available in ample supply. Nor
•could rates in the short-term open markets have fallen much below discount rates so long as an appreciable
volume of member bank borrowing at the R e s e m Banks represented indebtedness incurred under the
profit motive, since member banks would have withdrawn funds from the short-term open markets to
repay indebtedness at the Reserve Banks whenever continued borrowing became unprofitable, and ratesin
-th0S3 markets could not have fallen much below discount rates until member banks had liquidated a considerable proportion of their indebtedness." Riefler, op. cit.. pp. 20-21
, ' " T h e relationship which this theoryJLe.y the 'profit theory') envisages between the discount rates of
the Reserve Banks, on the one hand, and the money rates in the short-term open markets, on the other, Is
•essentially that which prevailed in the acceptance market in this country." Riefler, op cit ch 2, p. 21.




FKEE RESERVE CONCEPT

f)

market. It does not, however, us might be expected, apply
with anything approaching the same precision to rates in
other short-term markets which have varied widely from
discount rates 7 * * * it is impossible to explain the movements of money rates in the open market and the levels
which they have occupied during recent yean? by the movements and levels of discount rates at the Reserve Banks
alone.8
Then his main conclusion with respect to the two alternatives:
The functioning of the Reserve Banks in the money market
must, therefore, be considered from the point of view of the
theory that changes in the volume of member bank borrowing exert a more important influence on rates than do changes
in discount rate.9
In the remainder of his book, Iliefler developed an analysis of
the monetary process connecting Federal Reserve operations with
the money supply and the behavior of the banks and the public,
jhoiigh incomplete and deficient in several important respects and
dominated by very short-run considerations, his analysis was an extremely useful and important beginning. Unfortunately tho theory
construction t hat he began has not been completed by others. Instead.
several of the notions that lie introduced were uncritically accepted
hy the Federal Reserve and have continued to appear in their discussions despite their inconsistency with other elements that have been
introduced. Of particular interest in the light of later developments,
is the emphasis nlm-eci oil the expansion or contraction of bank credit
hi response to changes in the volume of member bank borrowing at
the Keserve Banks.% While we make no attempt to present, the theory
jn detail,10 consideration of his cent nil idea is u useful introduction to
later discussions by Federal Reserve officials.
iliefler °s theory"contains four major elements. The first, and by
tor the most important, sects to explain the relation of market rates
of interest and the volume of bank indebtedness. Larger batik indebtedness and a higher discount rate are said to raise the prevailing
Market rules; smaller indebtedness and lower discount rates depress
nmrkel rates. '\Tjhe volume of member bank indebtedness at the
Keserve Bunks at anv given time is one of the most important single
"jonetury factors in the level of money rates, and * * * the prospect
°f increase or decrease in the indebtedness is one of the most important
s,
*igle factors in the rate outlook." 11 Open market operations by the
ccutral bank were at the root of the rate changes since, under the
needs and reluctance'' notion, member banks borrowed when open
pmrket operations reduced member bank reserves and repaid borrowlr
*g when the central bank increased reserves. Iliefler stated the
Point a* follows:
(Fluctuations of money rates in the short-term open
markets should be governed by corresponding fluctuations in
aggregate volume of member bank indebtedness at "the
I Ibid.. I* 23
J Idem.

I^'^IIed analysis or the Iliofl,* conception U contained in our pai>or "Kvolving • • V
"K'flcr, oi»t cit., p. 1.7. Sw p. 1 4 &. 0
J i" .




6

FREE

RESERVE

CONCEPT

Reserve Banks, increased borrowing there being reflected in a
rise of money market rates and decreased borrowing m a
decline of rates in these markets. This would be expected
because * * * member banks do not borrow in order to
increase their loans, but rather endeavor to contract their
loans in order to repay their indebtedness.12
A second element concerns the public's response to the changes in
market rates initiated by the Reserve Banks. Higher interest rates
were said to reduce the demand for bank credit either directly or
indirectly because the public sold fewer securities to the banks. Lower
rates expanded the quantity of bank credit demanded. Given the
volume of bank indebtedness to the Reserve Banks, the "quality
evaluation" of loan applications by commercial banks and the public's
behavior determined the volume of earning assets held by banks and
the total deposit liabilities.
The volume of acceptances held by the Reserve Banks was the third
main element. The Federal Reserve set a rate at which it was willing
to buy acceptances. When market rates rose relative to the acceptance rate, banks sold acceptances to the Reserve^ Banks and total
reserves increased. The Federal Reserve's portfolio of acceptances
thus was determined by the prevailing market conditions and the
acceptance rate.
The last main building block introduced by Riefler is designed to
explain the variations in member bank indebtedness. The amount
of indebtedness is shown to be equal to required reserves plus currency held by the public minus the Federal Reserve's holdings of
Government securities and float, minus the gold stock net of Treasury
cash, minus Treasury currency outstanding, plus Treasury and
foreign deposits at the Federal Reserve Banks, plus "other deposits"
and "other accounts" on the balance sheet of the Federal Keserve
Banks. For Riefler this relationship is not simply a balance sheet
identity from the consolidated Federal Reserve statement. I t reveals
a causal relation that determines the volume of member bank borrowing. Banks have no desire to borrow from the Reserve Banks;
variations in the pressure to borrow emanate from changes in the
elements described. When "favorable" circumstances permit—e.g.,
when the currency flows into the banks from the public, when there
is a gold inflow or an open market operation, etc.—the banks follow
their fundamental disposition, viz., they reduce indebtedness.
The four building blocks jointly operate to determine the response
of the monetary system to the policy actions taken by the Federal
Reserve authorities. The transmission of typical policy actions to
the credit markets and the money supply may be traced with the aid
of Reifler's framework. Open market purchases lower the banks'
indebtedness dollar for dollar; lower indebtedness induces banks to
lower the yields on money markets; the public responds with a larger
supply of earning assets to banks; the banks' asset portfolio and deposit liabilities expand. Open market purchases thus expand "bank
J f n S T t J H S ? . ™ * ! ? ® ^ augment the clues about the central relation visualized by Kiefler. For example, he writes « • • * changes in this indebtedness appear to be the initiating force in correspondinf
changes in money rates. It is this relationship apparently1which Ins given t < T ^ ^ a n k S e S n s in

tte^h^n^

A™ *hmmTbefun?

tnan 11 nas been written into Reserve banking theory * * \ Induced throw eh ouon market tmerations,
t t ^ l ^ V ^ f , ? f m e T ? ^ r , ^ Indebtedness have been ^ S c e S t t to UghtenSndtoea^
K
the money markets, independently of changes in discount rates."




FREE RESERVE CONCEPT

{>3

credit" and the money supply and lower the interest rates on the
credit markets.
Riefler's exposition of central banking theory thus made the volume
of member bank borrowing completely unresponsive to any direct
influence of interest rates. Only to the extent that these rates operated on currency flows, gold movements, or the other balance sheet
items listed above could they alter the amount of borrowing. This
view carried over, in part at least, to the initial formulation of the
free reserves doctrine.
Riefler is rather vague about the role of the discount rate in the
process. At times he seems to suggest that the discount rate has no
effect on the environment described by the four building blocks.
Other suggestions hint that the discount rate operates independently
of changes in bank indebtedness but not independent of the existence
of borrowed reserves by the banks. Also missing from the Riefler
discussion is any consideration of variations in the volume of excess
reserves held by banks. As we have noted, fluctuations in excess
reserves were relatively small during the twenties, and this may
account for the lack of attention. Finally, a reading of Riefler's book
shows that his discussion is dominated by concern with extremely
short-run money market considerations. This emphasis has an
important bearing on his acceptance of the "reluctance" theory of
bank indebtedness that occupies a vital position in his analysis.13
BURGESS' VIEWS

Shortly after Riefler's book, a revised edition of Burgess' well-known
study appeared.14 Burgess accepted most of the Riefler formulation
of the monetary process and added a slightly more explicit treatment
of the role of the rediscount rate. Like Riefler, he notes the close
correspondence between the behavior of money market rates and the
volume of indebtedness. He explains this association in terms of the
banks' reluctance to borrow or to remain in debt.
When the member banks find themselves continuously in
debt at the Reserve Banks, they take steps to pay off their
indebtedness. They tend to sell securities, call loans, and
restrict their purchases of commercial paper and other
investments. The consequence is that when a large number
of member banks are in debt, money generally becomes
firmer, commercial paper sells rapidly, and rates increase.
Conversely, when most of the member banks are out of debt
at the Reserve Banks, they are in a position to invest their
funds; and money rates, including commercial paper rates,
become easier. I'his relationship res'ts largely on the unwillingness of banks to remain in debt at the Federal Reserve
Banks.15
The central feature of Riefler's discussion—that Federal Reserve
policy operates on the monetary system by inducing variations in
member bank indebtedness—is repeated by Burgess.16 In addition,
u
ibid.,
11

p. 220.
Cf. p. 236 and p. 238forexamples.
29—678— 64——3




FREE

RESERVE

CONCEPT {>3

recognizes the operation of discount rate policy as a separate
element in the process that is reinforced by open market operations.
The effectiveness of purchases and sales of Governnient
securities as an instrument of policy lies usually in their influence on the indebtedness of member banks at the Reserve
Banks. Purchases enable member banks to pay off loans and
thus tend to make money easier; sales lead banks to borrow
more heavily and thus tend to make money firmer. Government security transactions supplement and enforce discount
policy.17
Elsewhere, after commenting on the principle of open market operations along the lines described by Riefler, Burgess notes:
It can thus be seen that buying and selling [by the Reserve
Banks] is not only an independent influence on the credit
situation, but may and often has been used as a means of
preparing for discount rate changes and making them more
effective.18
Variations in the rediscount rates were seen as an independent
influence on "bank credit." Such influences operated in conjunction
with open market policy. When open market operations reduced
reserves, banks borrowed from the Reserve Banks, as Riefler had
described. Open market operations were effective in changing interest rates and could be reinforced by fiat changes in the r e d i s c o u n t
rate that made increased bank indebtedness more or less expensive
and contributed to the variation of market rates.
The explicit recognition of the discount rate as a separate influence on market interest rates and on member bank borrowing might
have stimulated further interest in the influence of costs ana yields
on banks' reserve positions and an analysis of the demand by banks
for reserves. But the Riefler-Burgess conception was dominated
by the "reluctance theory" of bank borrowing, and this further
step was not taken. As a result, the role of "excess reserves" and
the growth of such reserves during the thirties could not be interpreted in the prevailing Federal Reserve view.

Burgess

GOLDENWEISER'S

VIEWS

The persistence of the viewpoint explored by both Riefler and
Burgess is clearly revealed by Goldenweiser's 1941 article.1® The
major change is in the direction of weakening the description of the
causal connection between open market operations, bank indebtedness, and "credit expansion" and increased emphasis on the bankers'
"frame of mind."
* * * When the System wishes to ease credit conditions
* * * it purchased Government securities in the open
market and simultaneously reduced the discount rate at
the Reserve
provided member banks with
reserve funds t o A m A ^ l ^ i indebtedness at the Reserve
ai
Banks *d also*&&d^^JSMMebtedness as remained less
11

P. 238.
M
" E.
G o l d e n w e i s e r . ' ' ' i M ^ n e t o ^ S ^ ^ ' l ^ M e i T e Policy
Board of Governors of the F e ^ ^ ^ R ^ ^ ^ ^ ^ ^ ^ i .




" in "Banking" Studies." Washington:
^ ^
'
^

9
FREE RESERVE CONCEPT

burdensome to the member banks. This policy was intended
to put member banks in a position 20 a frame of mind to
and
be more liben/.l in extending credit.
Like Burgess, Ooldenweiser is more explicit than Riefler about
the role of discount rates in the central relation. Discount rate
changes anpear to affect market rates of interest by an amount that
depends directly on the volume of bank indebtedness. This view is
particularly interesting in view of later suggestions that the primary
effect of changing the discount rate was the psychological impact
associated with the announcement of the change.
Goldenweiser also stressed an important implication of the RieflerBurgess conception which explains the Federal Reserve's policy in
1936-37. The absence of excess reserves and the existence of bank
indebtedness are presented as necessary conditions for monetary
policy to be effective. According to the Riefler-Burgess notion the
evaporation of bank indebtedness during the 1930's broke the,chain
linking 21 Federal Reserve with the credit markets and the money
the
supplyRestoration of an effectively working monetary policy
thus required that "contact be reestablished with the market/'
Reestablishment of contact meant the potential emergence of bank
indebtedness and the disappearance of large excess reserves. The
drastic increase in reserve requirements arranged in the late summer
of 1936 and the early months of 1937 seemed ideally designed, under
the ruling notion developed by Riefler-Burgess and carried on by
Goldenweiser, to render monetarv policy more potent without exerting
any deflationary damage. Goldenweiser asserted with particular
emphasis that this dramatic jump in reserve requirements "was not a
reversal of the policy of monetaiy ease pursued since the beginning
of the depression. * * * The Board's action was precautionary in
character and placed the system in a position where an injurious credit
expansion, if it should occur, coula 12 controlled by open market
be
operations and discount rate policy."
Goldenweiser thus explicitly indicated that accumulating excess
reserves and vanishing bank indebtedness broke a crucial link of the
Monetary process and rendered policy incapable of coping with potential problems.23 Under the Riefler-Burgess conception excess reserve
have no role to play in the monetary process. They are inconsistent
with this view of the monetary mechanism. It is therefore intriguing
to note that the emergence of excess reserves was immediately interpreted to mean a breakdown of policy mechanisms and not a denial
falsification of the Riefler-Burgess conception, that denies the
existence of excess reserves.
The exclusion of excess reserves from systematic consideration was
closely associated with the involuntary and imposed character attributed by the Federal Reserve to bank indebtedness. The implicit
* IbifK. p 4 0
0
A

J ! " bIter the autumn of 1933 these instruments (i.e., the discount rate and open market operation) vcre
w l ^ , e > t * ™ " * the hanks were out of debt and had a large volume of excess f ^ v e s The^banks
were therefore largely independent of the Federal Reserve System's traditional method of credit reguiaOoldenweiser, op ? c i t p ! 3 9 1 "A raUary condition for the effectiveness of such a policy <i e of
credit expansion) is that the volume ol excess reserves at the disposal of member banks be
s K '
Ibid., p. 400.
p 410
rtnlV1 I ^ i n g we note two points. One, tbe concern with the possible problem ofinflatlon appears to have
g»taatetl tbe conccrn tor t he emtio* P^biem of underutilimtion of resources. Two those who continue
"assert that monetary policy in recession is analogous to "pushing on a string" should be aware that the
of this view is the Riefler-Burgess notion that denied any influence of interest rates on member bank
"Growing or excess reserves and any relevance to the demand by banks for reserves.




10

FREE RESERVE

CONCEPT

denial of any systematic response of the volume of bank indebtedness
to market conditions was extended to cover excess reserves. It was
therefore consistent for the Federal Reserve authorities to consider
excess reserves throughout the thirties as a redundant surplus of no
use and of no function in the monetary process. But excess reserves
persisted and still exist, particularly among country banks.
The persistent occurrence of excess reserves must have slowly
eroded the old version of the Riefler-Burgess conception. We find it
difficult to obtain clear evidence of the gradual transformation of
this dominant view into a notion emphasizing the central position of
free reserves. This transformation must have occurred during the
late 1940's or early 1950's.24 In the new view excess reserves were
treated as an extension of bank indebtedness, a magnitude offsetting
the retarding influence of member bank borrowing. Free reserves
assumed the position and role which originally had been assigned to
bank indebtedness. The free reserve conception thus emanated as
a result of an adjustment in the central building block of the RieflerBurgess view of the monetary process.
Additional modifications occurred in the late 1950's. These changes
are considered in detail later in this section, along with some suggestions of the direction in which the doctrine is currently moving. I t should be noted, however, that while there are periodic
changes in the prevailing views, the basic conception is almost never
completely formulated and has never been subjected to a searching
appraisal or even to the discursive arguments that Riefler provided for
his views. No doubt, such an appraisal would show that the broad
movements of interest rates and excess reserves during the 1930's are
consistent with many of the transformations that have been made in
the old notions. But this gain in empirical relevance is bought at a
high cost; viz, it would be impossible to justify the Federal Reserve's
policy in 1936-37 with the modified free reserve conception. The
Federal Reserve authorities apparently never realized this implication
of their evolving notions that substituted free reserves in place of bank
indebtedness as a central magnitude in the causal process. Such unawareness is a typical symptom of the unsystematic and essentially
impressionistic nature of their discourse concerning these problems.
A clear and definite grasp of pertinent implications can only be obtained by tracing the conception as a whole in a coherent, systematic
manner. The interactions of the distinct blocks composing the
whole process must be carefully followed in order to fully understand
the patterns implied by a given framework. To date, the F e d e r a l
Reserve has not done this.
THE CAUSAL POSITION OP FREE RESERVES: FREE RESERVES AS AN
I N D E X OF A MONETARY SITUATION AND FREE RESERVES AS A POLICY
TARGET

To understand the role of free reserves in Federal Reserve discussions, it is important to separate three distinct meanings assigned
to the term. Each of these meanings involves the use of the term in a
.different way, with markedly different connotations. I t is symptomatic of the manner in which the Federal Reserve discusses the mone*« Irving Auerbach of the Federal Reserve Bank of New York is often given credit for the development
*of the .free reserves concept.




FREE

RESERVE

CONCEPT

{>3

tary mechanism that the three uses of the term are not distinguished.
Thus, when there is an occasional denial or affirmation of the importance of free reserves, it is not made clear which of the three
uses of free reserves is involved, and we can only judge from the context.
As we shall note, this has encouraged needless confusion about the
status of the free reserve conception and its importance in Federal
Reserve thinking.
One meaning of free reserves has to do with the causal role assigned
to the concept. Previous sections have presented the broad outlines
of the causal connection that was said to exist between free reserves
and interest rates, or changes in bank credit in the later evolution of
Riefler's notions. A second meaning gives to free reserves the function
of indicating changes in the prevailing monetary situation, particularly
modifications of the Federal Reserve's policy posture. Sustained
movements productive of large differences in the level of free reserves
usually have been interpreted to indicate more or less "ease" or
"restraint" in the monetary system. Closely associated with the
signal or indicator function, often assigned to free reserves by Federal
Reserve officials, is the third use of the term. This is the target
function or the practice of incorporating some particular range of
free reserves as a guideline for monetary policy.
The occurrence of free reserves as a policv target is heavily dependent on the assumption, implicit in the free reserves conception,
that free reserves play a central role in the causal process linking
policy actions with the behavior of the credit markets and the money
supply. But this conception of the monetary process does not imply
the use of free reserves as a policy target. Thus it is quite consistent
with the continued adherence to the free reserves conception that other
targets may replace free reserves. As we shall see, the choice of an
alternative target provides no information that permits us to conclude that the Federal Reserve has rejected the conception of monetary processes centered on the causal role of free reserves. Neither
does the use of some other target necessarily indicate that free reserves have been abandoned as a signal of changes in the monetary
situation. However, abandonment of the causal connection between
free reserves and changes in bank credit would destroy any basis for
the indicator or target functions often assigned to free reserves.
Before discussing the use of free reserves as an indicator and/or
target of policy actions, it is useful to consider the large variations in
tree reserves that have occurred since 1946. The table in the appendix* shows that the monthly average of free reserves reached a
Maximum of $1.1 billion in Januarv 1946 and dropped to a minimum
°f minus $874 million in November 1952. Since 1955 the monthly
average has moved between $500 million and minus $500 million.
For reasons that have been cited and are inherent in the definition,
the Federal Reserve cannot control the volume of free reserves on
a d a i l y or weeklv basis up to the last dollar. Banks and the public
c
an affect the value of free reserves by borrowing or repaying borrowing, converting demand deposits into time deposits, or depositing
a
*id withdrawing currency. These operations cause changes m ret i r e d or total reserves. Other factors such as float, movements of the
•Treasury balance between commercial banks and Federal Reserve
'Appendix will appear in full study.




FREE RESERVE CONCEPT {>3

Banks, etc., introduce changes in the volume of free reserves available
on any given day. Movements of deposits from Reserve city banks
to country banks also change the volume of free reserves since reserve
requirements for country banks are lower than for Reserve city banks.
As a result of movements of deposits into country banks, fewer
reserves are classed as required reserves, more reserves are measured
as excess reserves, and free reserves are larger. Movements of deposits
in the opposite direction reduce free reserves.
The Federal Reserve is principally concerned with extremely shortrun market influences, but it cannot hope to anticipate with precision
all of the movements occurring each day or week. It can and does
make projections designed to offset some or all of the anticipated
changes, as the earlier discussion of "defensive" operations pointed
out. But it must be satisfied with a level of free reserves that is
subject to daily and weekly variation.
The amount of variation that must be accepted has been reduced
in recent years. Better information, improved coordination with the
Treasury, and elimination of the higher reserve requirement for central
Reserve city banks have all contributed to the reduction in weekly
fluctuations. Most of the week-to-week changes in free reserves in
1962 were less than $50 million. Ten years earlier, week-to-week
changes of $100 or $200 million were not uncommon.
FREE RESERVES AS A POLICY TARGET

The Federal Open Market Committee at times specifies a range for
the value of free reserves. At other times, they may specify some
other criterion or a set of criteria. These alternative criteria are
often vague, e.g., concepts such as "tone" or "feel," described earlier,
may be used. The Manager of the System Open Market Account must
then translate these statements into an operative concept, i.e., into a
range of free reserves or some other magnitude that he will attempt
to maintain. Quite often it is a range of free reserves that is chosen.
The choice of free reserves as the manager's target is very likely
to emerge under the circumstances. The account manager's position
on the credit market is similar to the position of the commercial banks'
money desk men in important respects. Both appraise events in the
context of a single bank's frame, and both focus on extremely short-run
occurrences. The operational duties imposed on both the a c c o u n t
manager and the money desk men channel their views in the direction
mentioned, as discussed in "The Federal Reserves' Approach to
Policy." Free reserves provide the manager with a concept that is
analogous to the "money (or Federal funds) position" that plays a
dominant role as a target for the money desk men. Variations ii/free
reserves consistently tend to play a major role in the account manager's
considerations. And these same considerations will frequently lead
the FOMC to incorporate levels or ranges of free reserves among the
policy targets, or will influence the account manager to translate other
policy targets into a range of free reserves that he attempts to
maintain.
We do not contend that free reserves are the only or most important
policy target of the Federal Reserve authorities. As we noted, the
acceptance of the free reserve doctrine as a conception about the
structure of monetary processes, the choice of free reserves as a specific



FREE

RESERVE

CONCEPT

{>3

signal or indicator of monetary situations, and the choice of free
reserves as a policy target are distinct and only partly dependent issues.
The dependence of these issues is quite asymmetrical in the sense that
rejection of the free reserve doctrine would effectively remove free
reserves both as indicator and polic}^ target, whereas acceptance of the
doctrine does not entail the other choices, except by a purely nonlogical
connection. The use of other policy targets, accompanied by a modified or fading emphasis on free reserve targets, is perfectly compatible
with the recognition that free reserves play a causal role in the monetary mechanism and are also used as an indicator of "the degree of ease."
The modification and adjustment of the policy targets, however,
poses some problems not fully or explicitly appreciated by the Federal
Reserve authorities. A close control over some market rates would
preclude a close control over free reserves, unless the discount rate and
the reserve requirement ratios are continuously adjusted to evolving
circumstances similar to the FOMC's open market operations. But to
the extent that requirement ratios and the discount rate are maintained
at certain levels, the close control of some interest rates near a target
level implies abandonment of free reserves as a closely controlled
target. One goal must be sacrificed, in part at least, to the other.
For example, in the pre-Accord period, when interest rates were
pegged from above to prevent bonds from selling below par, some
control of free reserves was lost. Most recently, when bill yields
have been pegged from below to prevent a decline in short-term money
m
^rket rates, some of the control of free reserves has been lost again.
When the Federal Reserve withdraws reserves from the banks
through open market operations to maintain a given bill yield, total
reserves decline. Some reduction in free reserves will result either
because member banks borrow additional reserves to restore some of
the reserves removed, or because there is not an equal concomitant
deduction in required reserves through a decline in demand deposits
? r a shift of demand to time deposits. Pegging bill yields is therefore
inconsistent with tight control of the level of free reserves.
The pegging of short-term rates in recent years has changed the
role of free reserves as a policy target. But that does not mean that
they have not remained an important indicator or measure of ease
and restraint. An illustration of the effect of choosing bill yields as
a guide to policy is contained in a graph accompanying an article
py the present Xlanairer of the Open Market Account.25 The graph
js entitled "Free Reserves fluctuated from week to week while ShortTerm Rates moved narrowly."
The graph shows the effect of choosing bill yields and rejecting free
reserves as the primary target of Federal Reserve policy. We have
?een that a necessary condition for the choice of free reserves as an
indicator and target is the belief that a high level of free reserves
induces a rapid rate of credit expansion and that persistently low
levels decelerate credit expansion. By choosing the bill yield as
Policy target, the Federal Reserve surrenders some control of the level
of free reserves. The primary instrument becomes the bill yield;
the level of free reserves must be adjusted to maintain the bill yield at
0T
near some minimum level.
lg?R.

W. Stone,"Federal Reserve Open Market Operations in 1962," Federal Reserve Bulletin, April




FREE RESERVE CONCEPT {>3

The choice of Treasury bill yields as a primary target of System
policy is one indication that the goals of System policy have changed.
The goal is now related to the so-called balance-of-payments problem.
The Federal Reserve has adopted the position that the difference
between domestic and foreign (especially European and Canadian)
short-term interest rates is an important source of the outflow of gold.
To reduce the outflow of gold, domestic short-term interest rates are
pegged from below.
I t is indicative of the absence of analysis as a base for Federal
Reserve policymaking that the choice of the new instrument and the
new policy have not been supported by any detailed study that confirms or even strongly suggests that short-term capital movements are
highly sensitive to interest rate differentials. This does not mean that
the relationship does not exist. Some indirect evidence from general
economics supports the relation, but no direct evidence has been
adduced thus far that confirms a close and sensitive response of shortterm capital to differentials in short-term interest rates. Nevertheless, the new policy and the new target appear to have supplanted free
reserves and the goal of achieving ease or restraint; i.e., full employment and price stability.
The adoption of a new policy and a new target does not imply
that free reserves have been rejected as a measure or indicator of
ease and restraint. Rather it suggests that domestic expansion has
been relegated to a position of lower priority. In the current and
past euphemisms, we have accepted the "discipline of the balance of
payments" in place of the policy of "leaning against the wind."
Does the modified free reserve doctrine continue to measure ease
and restraint? We suggest that the bulk of the evidence supports
the view that it does. The choice of a new target of System policy
is important, because it indicates a change in the aims of policy. If
domestic expansion will be encouraged by the System only to the
extent that the constraints on short-term interest rates permit, then
the use of free reserves as a target has been suspended, while the
causal role of free reserves is not affected.
This is not the first time in recent years that the System has used
an instrument other than free reserves as a target. During the
period in which Chairman Martin and others continued to refer to
free reserves as an indicator of ease and restraint, other targets were
mentioned at times in the "Record of Policy Actions." For example,
at the meeting of May 27, 1958, the manager's targets were "to
maintain the current posture 26 monetary ease, without further deof
pressing Treasury bill rates."
At other times, particularly during
Treasury offerings, the manager is told to "maintain an even keel."
Numerous other targets have been used. At the meeting of
December 19, 1961, a principal target of short-run policv was the
level of "available reserves." This measures the "net change in
total reserves after allowing for reserves provided or absorbed to offset
seasonal factors 27
and changes in Treasury tax and loan balances at
member banks."
References to total reserves, unborrowed reserves,
and a variety of other instruments are found also. At other times, no
specific credit market target is mentioned.
M"Record of Folicy Actions," Annual Report of the Board of Governors for the year 195S.
Guy Noyes, "Short-Run Objectives of Monetary Policy," Review of Economics and Statistics Supplement, February 1963, p. 148. Mr. Noyes was Director of the Division of r S ^ ^ ^
Governors at the time.
'




FREE RESERVE CONCEPT

{>3

THE INDICATOR FUNCTION OF FREE RESERVES

Our previous discussion of "defensive operations" and the importance of random elements in the volume of free reserves implies that
a single policy target specified in terms of free reserves does not assure
even approximate control of free reserves in the shortest run. Moreover, banks do not respond immediately to variations in the volume
of free reserves according to the best explications of the free reserve
conception. The very short run, or instantaneous, relation between
free reserves and "credit expansion" is, therefore, admittedly loose.
A longer policy horizon lowers the relative importance of the random
components and raises the extent to which the level of free reserves
can be controlled. These differences between the Federal Reserve's
ability to control the short- and longer-run variations in free reserves
are of importance for a discussion of free reserves as an indicator or
signal of changing monetary conditions.
Day-to-day variations in free reserves are relatively large. Banks
receiving increased reserves may hold them for a day or more, or
lend them in the Federal funds market rather than purchase securities
or reduce loan rates to stimulate borrowing. That is, banks may
interpret the inflow of reserves as "transitory," the result of a
^defensive" operation when the Federal Reserve had in mind a
"sustained" change in a reserves, a "dynamic" operation or a combination of the two. Similarly, losses of reserves may be treated as
"transitory" rather than "sustained" by the banks and thus lead to
borrowing in the Federal funds market. For this reason the Federal
funds rate and the movements of Federal funds become additional
indicators of the effect of policy operations.
Experiments with the reported free reserve totals suggest that a
3-week moving average of total free reserves provides a relatively
reliable indicator of Federal Reserve policy. We will consider the
evidence in more detail in section 2. It is pointed out here to suggest
that there may be a lag of several weeks between a change in the range
of free reserves and an evaluation of the change as a sustained change
by the market.
. Several bankers have indicated to us that their staffs perform
similar smoothing operations to obtain a moving average of weekly
reported free reserves. Three weeks is often used as the period of
the moving average. It is probably not a coincidence that the 3-week
l o v i n g average permits observers to isolate the periods between
meetings of the FOMC. In any case, it suggests that bankers and
other interested observers of Federal Reserve policy may not respond
immediately to a change in the level of free reserves by increasing
or decreasing their outstanding loans and investments. Much may
depend on the size of the change in free reserves, the rate on Federal
funds, the discount rate, the rate on Treasury bills, and the direction
m which these rates move. These and other signs are carefully
watched for clues to infer the composition of the changes experienced.
Our discussion in a previous chapter indicates that the response of
banks to variations in reserve positions substantially depends on their
interpretation of these changes. Modifications of reserve positions
interpreted as only transitory will not induce the portfolio adjustments typically associated with changes in reserve positions deemed
to be persistent and systematic.
29-07$—64




4

FREE RESERVE CONCEPT {>3

From the viewpoint of the banks' money desk men or the Federal
System's account manager the central building block of the
free reserve conception—connecting free reserves to "bank credit"
expansion—seems most natural and rather obvious. Continuous
exposure to the daily variations in reserve positions and the associated
portfolio adjustments seems to support the relationship. But the
reader is once more cautioned against such impressionistic evidence
that contributes little to the discriminating evaluation of accustomed
beliefs. Less subjective procedures will be used in section 3 to examine the central core of the free reserves conception and will lead us to
reject it. Such rejection immediately invalidates both the causal
role and the signal or indicator function attributed to free reserves.
At the present we consider an essentially logical issue, viz, the appropriateness of assigning an indicator, character to free reserves in the
context of the free reserve conception. We are thus not questioning,
at the moment, the empirical relevance of this conception. We
accept it for the moment and question the signal and indicator function attributed to free reserves by the Federal Reserve authorities.
The logical issue cannot be settled by a direct critical examination of statements made by the Federal Reserve authorities.
Their pronouncements are usually too vague to permit a searching
analysis without first translating them into a more coherent ana
definite analytical context. Our appraisal of the compatibility of
the indicator character with the free reserve conception is therefore
based on a specific translation which we undertook for the examination of this question and related issues.28 Once more we emphasize
that there is no guarantee that our explanation of the Federal Reserve's notion is "correct." It is offered as a substitute for the product never supplied by the Federal Reserve authorities. Its comparative adequacy, however, can be judged by the extent to which
typical Federal Reserve statements can be successfully subsumed by
our analysis.
The explicit construction of the Federal Reserve's free reserve
notions yields a remarkable result. It turns out that free reserves
could rationally serve as a signal or indicator in the manner used by
the Federal Reserve only if we possess detailed and reliable information about crucial links in the monetary process that are presently
beyond our disposal. The required information must be sufficient
to separate the strands composing the observable behavior of free
reserves. In particular, the component attributable to policy actions
should be separated from the influences on free reserves emanating
from the economy via the public's asset supply to banks. But in our
present situation, i.e., in the absence of sufficiently detailed and reliable information concerning the structure of the monetary process,
no useful indicator function can be rationally assigned to free reserves
on the basis of the free reserve conception. Indeed, it can be shown
that even large variations in free reserves cannot be safelv interpreted
as modifications of relative "ease" or "tightness" with the usual
connotations of accelerated or decelerated rates of "credit expansion."
Moreover, even if all the required information were available, other
elements (to be explained in section 4) supply a simpler and more
useful indicator of the prevailing monetary situation and the Federal
Reserve's posture.

Reserve

» Cf. our paper "Evolving * * *




op. cit.

FREE RESERVE CONCEPT

{>3

Detailed analysis thus exhibits the inadequacy of assigning free
reserves an indicator function, even if die central! causal role of free
reserves is acknowledged. Older these circumstances, how can we
explain the attention focused on free reserves by the Federal Reserve
authorities? The answer must be found in "the piecemeal nature
and uncoordinated character of the Federal Reserve's conception.
i«e Federal Reserve apparontlv ne\er viewed the simultaneous
operation and interaction of all the building blocks. Emphasis was
placed on the central block, relating market rates, or more recently,
the banks9 port folio adjustment with free reserves and the discount
rate. I he other building blocks were vaguely disregarded.
Attention limited to the single relation emphasizing the causal role
of free reserves generated a view of free reserves as a summary measure
oi the monetary situation. The relevant feedbacks and interactions
generated by the other relations constituting the monetary process
jvere neglected, although many of them were a part of tlie Riefler
tradition. The Federal Reserve's procedure may be likened to an
explanation of price and output in terms of supply only, disregarding
tlmt pnee and output emerge from the joint interact ion of demand
ami supply forces.
EVIDENCE FROM PUNLISHKl) STATEMENTS

Our discussion has recognized three separable aspects in the Federal
Koserve's view of free reserves, viz: (1) the recognition of their
central position in the transmission of monetary impulses, (2) their
weeptanee us a summary measure of a monetary situation, the signal
indicator function, and (3) their use at times"as a target for monepolicy. Two statements made in 1958 provide a relatively
wear exposition of the free reserve doctrine as seen bj- the Federal
Kwerve» The article published by the Federal Reserve Bank of
:^Cvv ^ °rk appears, on superficial reading, to be a criticism of "the
iree reserves doctrine." But careful reading yields a different result.
A
causal role of central importance is explicitly recognized, and a
qualified indicator function is acknowledged.
'he qualifications that are introduced take the form of explicit
recognition that other elements shape the banks' portfolio adjustments
Jointly with the volume of free reserves. The joint operation of free
reserves and other elements at times modifies the interpretation that
be placed on the effect of a particular level of free reserves on the
Jinks' behavior. For example "while excess reserves have been
,ai
nv stable, free reserves have moved over a wide range, marking
? u t the major swings between monetary ease and restraint." An
important qualification is introduced to explain why a given level of
We reserves does not always mean the same thing. Country banks
"J* willing i 0 hold larger excess reserves than money market banks.
i
«e accrual of free reserves at money market banks is expected, in
general. to have a different effect on portfolio adjustments than a corresponding accrual at country banks.
Other qualifying factors must be considered. "At times when
gjjgk* have, for example, higher ratios of loans to deposits, or of long-




Ig

FREE RESERVE CONCEPT

term to short-term loans (measures of liquidity), they may be less
responsive to higher levels of free reserves." However the report
notes that the qualifications are not denials of the role of free reserves;
they simply indicate that the process of influencing "credit" is not
instantaneous. For example, note the following:
.
"When free reserves are held for some time at a relatively high level,
member banks will not only continue to make loans available to their
customers—and probably more readily available than at lower levels
of free reserves—but they are also likely to seek out new investment
opportunities aggressively * * *." "It is in this way, through the
pressure of an enlarged supply of bank funds seeking investment
against a reduced demand for bank credit, that there is a tendency
for a high level of free reserves to be associated with falling interest
rates, increased liquidity in the banking system, expansion of credit,
and growth in the money supply."
After a discussion of the effects of a reduction in free reserves the
New York Bank adds the following indication that it is the level and
not the rate of change of free reserves that influences credit expansion.
"(E)asing or restraining effects are related primarily to the level offree
reserves that is being maintained and are not dependent on continuing
further changes in that level. To be sure, as noted above, a given
magnitude of free reserves may induce different degrees of ease at
different times, depending on a variety of influences. But in maintaining whatever degree of ease or restraint has been achieved under the
conditions prevailing at any particular time, it is not necessanj for free
reserves or net borrowed reserves to rise continuously to higher and
higher levels, as has sometimes been supposed."
On occasion one may encounter passages that seem to deprecate
the causal role or the indicator function of free reserves. The New
York Bank article refers to the "loose fit of any specific level of free
reserves to any degree of credit ease or restraint * * *." We have
considered earlier the way in which random variations impose a "loose
fit" in the very short run. But the formulation also refers to the
modification of the free reserves doctrine mentioned towrard the end
of the last section. Other elements supplement free reserves as causal
factors operating to shape the banks' rate of portfolio adjustment.
The report suggests that "consideration must be given to the distribution of bank assets among loans and investments of varying
degrees of liquidity, the size and composition of bank liabilities, ana
the level and structure of interest rates." Other factors may be mentioned on other occasions. A perusal of Federal Reserve pronouncements supplies ample evidence that the range of arguments in the
central relation was extended beyond free reserves and discount rate,
but the nature of the extension remains ambiguous.
Nevertheless, the conclusion is quite clear. "For all its limitations,
the free reserves concept remains a useful guide to the interpretation
of credit policy." A very similar conclusion with many fewer qualifications is reached in the Riefler article:
"Federal Reserve operations also affect the prices and yields of
Government securities becauset hey change the volume of free reserve$
available to member banks." Indeed, Riefler's discussion of Federal
Reserve operations provides an indication of their beliefs about the
quantitative impact of any change in reserves. "For example, if, &
general analysis suggests, something like seven-eighths of the effect




FREE RESERVE CONCEPT

{>3

of an open market operation on the availabilit}' of funds in the market
represents the effect of that operation on bank reserve positions * * *
while only one-eighth reflects the fact that bills were simultaneously
put into or withdrawn from the market, it follows that a comparable
change in the level of net free reserves from whatever came ultimately
should affect the general credit situation and interest rates to roughly
the same extent as the open market operation or within seven-eightlis
of the same extent."
The above remarks pertain to the recognition of free reserves both
as an indicator of current policy ("marking out the maior swings")
and as a target of policy ("held for some time," "the level of free
reserves that is being maintained"). Riefler also introduces some
qualifications about the interpretations of free reserves as a measure
of "ease and restraint." To remove doubt that statements published
m 195S are applicable to the interpretation of current policy operations, more recent quotations are provided. These indicate that the
free reserve concept has played a major role during the past 5 years.
One very clear set of statements appears as a part of the answers
that the Federal Reserve gave to questions asked by the Commission
on Money and Credit. Under the heading "Operating Guides and
Procedures," we are told:
"The figure of 'free reserves' or its negative counterpart 'net
borrowed reserves' provides a convenient and significant working
measure of the posture of policy at the time. * * * It is also a device
that is better adapted than its components taken separately for
estimating and projecting the net impact of regular variations in
factors affecting reserves.
. "The general level of free reserves prevailing over a long period of
time may be viewed as an indicator of the degree of restraint or ease
that exists in the money market." 30 The writer elaborates on some
qualifications that must be made. "The particular level of free
reserves that may be needed to achieve the objectives of policy may
vary from time to time depending on changing economic conditions."
Recognition is given to the possible changes in required and borrowed
reserves.31
Further recognition of the causal position of free reserves is given
(P» 8) in the statement "The Federal Reserve restrains (or encourages)
bank credit expansion by reducing (or increasing) the banks' primary
liquidity." The latter term is defined (p. 6) as follows: "Primary
bank liquidity relates to the net reserve position of commercial banks."
These remarks clearly indicate that the policy of the Federal Reserve
to operate on the level of free reserves in order to effect an increase
or decrease in the rate of credit expansion. However, qualifications or
Modifications are shortly introduced (p. 9).
"The significance at any given time of net borrowed reserves (or
free reserves) as a factor tending to restrain (or encourage) bank credit
expansion depends on at least five things: (1) the magnitude of free
reserves (or net borrowed reserves); (2) the level of short-term money
r
ates relative to Federal Reserve discount rates; (3) the vigor of
actual current demands for bank credit; (4) the existing level of total
,0

Commtssion on Money and Credit, "The Federal Reserve and the Treasury. * *
Op. cit., pp.
[italics h&re been added here and in the following quotations.]
tfc S p - cit-» P- 20- The pafre numbers in the text that follows refer to the Federal Reserve's answer to
De
Commission on Money and Credit.




20

FREE RESERVE COXCF.PT

bunk liquidity; and (5) the variations among different classes and
groups of banks with respect to the conditions just named/'
The last quotation, reportedly written by Woodlief Ihomas of
the staff of the Board of Governors," makes quite explicit the role of
free reserves as a causal entity of central importance and as a measure
or indicator of policy. It should be noted thai few of tlie.se statements
suggest that only free reserves influence the rate of credit expansion.
But. all of the statements assign an important role to free reserves as a
measure of credit policy and as a magnitude to be modified by policy
action in order to achieve desired changes in credit markets or the
money supply.
- . 1 1 Additional testimony to this effect appeared in an article by 1 outig
and Yager published in I960. Their statement, assigns a primary role
to free reserves and omits many of the qualifications.
" * * * ITjhese mechanical aspects of monetary regulations find
their summary in the movement of net reserve positions of the banks.
This quantity, in effect, may be thought of as tlie rudder by means of
which the monetary ship is made to" 'lean against the winds.' * * *
Since operations to increase or decrease the System's portfolio are
undertaken to change the direction of tbe rudder- - that is to influence
the net reserve position of the banking system on the basis of either
short- or longer-term corisiderations or both * * *." *
The evidence that free reserves have been used as both a measure
and, in their causal role, as an instrument of policy is relatively clear.
However, one final quotation is introduced to establish (1) that the
interpretation of free reserves as an indicator of policy has been sugested within the last several months and (2) that our interpretation of
ederal Reserve views is confirmed bv statements of their staff.34 In
particular they state that increased borrowing by member banks has
a contractive effect on the stock of credit or its rate of adjustment.
"One of the most sensitive measures of the day-to-day interaction
of mouetary policy and market forces is tho so-called net reserve
osition of banks. This measure is computed by subtracting member
ank borrowing at the Federal Reserve from excess reserves." A "persistent change in net reserve positions over a period of several weeks
often indicates a basic shift in the credit climate."
"In fact such market conditions are likely to stimulate growth
in total bank reserves by increasing the willingness of member banks
to borrow from the Federal Keserve. However, if a rise in total
reserves is composed largely of borrowed reserves, it is less likely to
be sustainable than if it is composed mainly of nonborrowed reserves.
Member bank borrowing at Federal Reserve Banks is generally
regarded as a temporary' source of reserves both by the borrowing
bank and by^ the Federal Keserve oflicials who administer discount
operations. This transitory or emergency nature of borrowed
reserves, * * * tends to limit the volume' of credit that can be
supported by such reserves."
These statements, as a group, are sufficiently clear that no summary
is required. Our interpretations and some criticisms have already

f

SumjKi"nl F o & y ^ S ^ i ? 1 * * * " Moni!larsr Po,lcy'"
Kconornic, and statistics'
A YaSW l,ThC Kc0n mIcs of mi
E^um^
'
'
°
> »'refrn>!»ly," Quarurly Journal*
onppMuTuS s^n^tivd^ Bf5Sm'C8'"
July IW». The quotations appear




FREE RESERVE CONCEPT

{>3

been made in an curlier section. Additional indications of the attachment of the Board of Governors, the Federal Open Market Committee,
and their staffs to the free reserves concept are readily available,
however.35
Some clues suggesting went developments oj the Jree reserve conception
Recently, a series of statements with somewhat different import
have been made. These statements are somewhat dilficult to interpret; they may indicate a change of views in the System, or they may
reflect existing dissents or discussions. Some people within the System apparently reject the level of free reserves as an indicator and
question the causal role customaiily assigned to free reserves. Despite the very recent statement of the Manager of the System Open
Market Account describing policy operations in 1002 in terms of the
modified free reserves doctrine, there is some evidence that the usefulness of the free reserve concept is not accepted throughout the
System.
Clues from published statements
A recent publication of the Federal Reserve Bank of Atlanta asks
rhetorically what the Federal Reserve System controls. The answer
given is clearly not free reserves but total reserves. It is total reserves that influence the expansion and contraction of "bank credit." 36
1-est some doubt remain that the free reserves conception is being
reconsidered by some officials, the article discusses excess reserves and
borrowings as'indicators of policy. It rejects these measures largely
for the usual reason that the distribution of excess reserves and borrowing is important. Free reserves are then discussed. The report

also faulty as a measure of the intensity of credit demand. Moreover,
they are not usually very indicative of actual bank credit trends." 37
On other recent occasions, Federal Reserve spokesmen have assigned
°nly the shortest run significance to levels of free reserves, and longer
run significance to total reserves. Such statements suggest a modification, but not a rejection, of the free reserves conception. They can
he reasonably interpreted in the context of a coherently formulated
free reserve conception that explicitly traces the interaction of the
relations composing the monetary process. The affirmation of the
longer run significance of total reserves docs not necessarily reveal a
r" 1 1 ),^. McC. Martin, "Statements to Coiuiresw: Monetary Policy and the KcpTiomy." rj-primed in
I M t i l l r t ! : . . Febru.iry
See e v i l l y p. 124. _ V2) "ftmawfetf J**fciw."Monthly
tilii?£ofl,M! Fwl«,r!j
H n of Framisoo. i». 113. '^meiii-hai
ak
hive ^ ^
iV^Ji
ft ? ^
w s from Si » uil!)i«n io Si n million. i3) Ilohsl O. KjiU-c. in " lleview of the Annu.il Ke/wrt
Jf W ?- Joint Kiwi'iinic Corn::iitlee. Wns'iliirton. 1IAI. p. 31. (4) * m MC Martin ibid., p. O . pp.
w,
c
S
W. tl-w. " Federal Ke-ervi- Open Muikoi Operations In II^." Foderul Ie*rve 1 Julian.
April
«ill(i,(-ltlv<s o( uM. mildriiss of the shift In einph:isi- tuwhrd U* e.i-r in June, weekly
i*
p
g»«w free Wrves xmTi>lirn ulovl\i in a kmr of abMit * 0 million to S N milhoii from mid-June to midW
KI
la?iCbPrContinuing attention was p.iln to free ream's but not to the extent of jMirsiinv-p^iiciiwftverc^en-o l#>vel* at t!ie eiiinx? of wide swings in the enteral tone of theinor.try nurk« t Other f.ietors
^"tinned. Tliy ale™ the luruil<>n of rv^rvi". (6) il.e awl!.ililllty of Federal funds, jc) clever
InSf?*
ireiMl'in .«hi»rt-terui r.ite.«. if) ll» pattern of wplt.il iiiarfcw <lrveljj«^or.ts</) credit
J5Mn>lon. and < ) m-uth in the money supply.
a
Various, meeting of the l-edera Oivr Mykci C < m&JI2 «:(>>' M SInslem Hcconl of Policy Action.*" in Annual Bejwrt of the 1 o.;r< '
"Tise
a
m-T^ >' - **to*"tuples the rejKirt of the meeting on July II. I9»l, Mar. 25,19.*, and
1

l&VVi
* Ibid

A greater \
other " Harry Brandt. ••Oonirtllin*
lta
* SH>tcmlwr IMS, p. 1.




Ig

FREE

RESERVE

CONCEPT

disposition to reject the free reserves conception. More likely it
suggests that some of the building blocks described earlier that had
been discarded, are now explicitly considered and acknowledged as
relevant.
.
A more fundamental attack on the free reserves doctrine was mentioned earlier. The statement "use of the term excess reserves to
indicate a supply of readily available funds or unused lending power
is probably misleading" attacks the root of the free reserves concept
as an indicator of potential expansion. It substitutes the view that
"a relatively high average level of excess reserves that persists for
several months does not necessarily indicate that there is an expansive
force on bank credit and money; instead it may reflect a weak credit
demand, low interest rates, or an increased desire for liquidity by
bankers." 33 The banks' demand for cash assets or reserves is thus
introduced as an influence on the rate of expansion of money and
credit and interpreted as a response to the prevailing level of interest
rates. As suggested above, this position is inconsistent with the
analysis based on free reserves and the use of free reserves as an
indicator of "ease and restraint."
The admission of a systematic response in the banks' cash asset
position to prevailing interest rates is not a minor adjustment of
prevailing views. It is a major break with the Riefier-Burgess tradition that assigned a nonvolitional character to banks' indebtedness
and later to free reserves. The abandonment of this position is a
rejection of the causal role and the indicator function assigned to free
reserves by the Federal Reserve authorities. The discovery of bank's
demand for free reserves by the Federal Reserve authorities must lead
eventually to a fundamental readjustment of the Federal Reserve's
conception of the monetary process. Such a readjustment will involve a radical break with an accumulated heritage of views and
pronouncements assessing monetary situations and guiding monetary
policy.
Clues from answers to Questionnaires
These signs of dissension or disagreement within the Federal Reserve System suggested that there may be serious questioning of the
causal role and indicator function assigned to free reserves. To obtain
more information, a questionnaire was sent to each president of the 12
Reserve Banks and to each member of the Board of Governors. The
presidents and the Board members both answered the questionnaire as
a group and indicated substantial agreement within each group.
Both replies are reproduced in the appendix to this study.
What do their answers indicate? Questions II, V, and VII in
effect asked the FOMC members to explain the substantive c o n t e n t
of "ease" and "restraint," to describe the monetary mechanism, to
specify the role of free reserves in the monetary process and the
meaning that they assign to changes in the level of free reserves.
Question III asked that the analysis be applied to a particular c o n t e x t ,
the year 1962. The responses to these questions are extremely helpful
m clarifying the state of the Federal Reserve's thinking. M o r e o v e r ,
the answers provide evidence of disagreement between the two groups,
since important differences of emphasis appear. These differences
suggest the emergence of a "conceptual interregnum." The free
reserve doctrine is no longer accepted by all as the primary building
w "Excess Reserves/' Review, Federal Reserve Bark of St. Louis, April 1963, p. 15.




FREE RESERVE CONCEPT

{>3

block in the analysis. The answers provided multiply the signs that
the inherited doctrine is being reconsidered. But the answers also
reveal quite clearly that the reconsideration has not proceeded far
enough to provide a firmer foundation for monetary policy or a more
appropriate analysis of the monetary situation. Residues of the free
reserve doctrine, emanating from the Riefier-Burgess tradition, continue to hold a prominent place in the Federal Reserve conception.
The 12 presidents indicate that free reserves are one of the indicators
but usually not the most important indicator of "credit conditions"
even in the very short run. Short-term market interest rates are
important in relation to the discount rate (II).3a All of these factors
are summarized in the statement that appears to define changes in
the degree of ease as "an availability of reserves relative to the
economy's demand for credit" (V).
Levels of free reserves above S500 million have occurred in periods
of ease since the "Accord." But the degree of ease is not the same
each time the level of free reserves gets above $500 million because
there is no unique association between free reserves of $500 million
or more and a particular level of borrowing, short-term interest rates,
or credit expansion (II.4). The demand for reserves must also be
considered (II.2 and V.2). In fact it is total reserves, not free reserves,
that is 4 'relevant from a longer term point of view" (II.2).
The Board of Governors notes that "ease" and "restraint" are
relative terms. T They must be interpreted in the light of the demand
for reserves b} banks. "Interaction between the supply of and
demand for free reserves gets reflected in the rate of expansion in
total required reserves." The distribution of free reserves is at
times an important indicator of short-run behavior of the monetary
system (II.1). "The level of free reserves is * * * an indicator
of the degree of ease or restraint if interpreted in the light of prevailing
demand conditions" (II.2; italic in the original). But the most
important fact is the demand and supply for loanable funds.
The principal supply factor subject to Federal Reserve policy
is said to be the supply of total reserves. This is reflected m the
level of free reserves, on the supply side, but must be judged relative
to the demand for free reserves. The latter shows primarily changes
in the desired borrowing of member banks since the desired excess
reserve position of member banks "changes only infrequently" (II.2
and II.1).
Measurement of the demand for free reserves cannot be precise.
But the factors influencing the demand are provided "by changes in
bank loans, especially business loans, and oy the level of interest
rates on short-term securities and the Federal funds rate and their
relationship to the discount rate." The context suggests that other,
unnamed factors might be important also (VII).
The Board appears to place substantially greater influence on the
level of free reserves than do the Reserve Bank presidents. Both
emphasize a number of other factors, that modify the "free reserves
doctrine," but the presidents seem to suggest that free reserves are a
relatively poor indicator of short-term policy and that total reserves
are better both as a target and as an indicator of policy. Both groups
" Roman numerals appearing in the text will refer to the numbers of the questions in the appendix, when
the answers have been paraphrased or quoted. If a particular subsection is paraphrased or quoted, the
reference will be given as IL3. The appendix will appear in the full study to be published later. •

2M7— 4
1 J8 &



6

Ig

FREE RESERVE CONCEPT

seem to agree that if free reserves are used as an indicator of policy, the
following must be used to interpret the meaning of the level of free
rGssrvcs i
(1) The rate of expansion of bank loans and investments, the
rate of expansion of bank loans, or both;
(2) The Treasury bill rate relative to the rediscount rate;
(3) The Federal funds rate relative to the rediscount rate;
(4) The distribution of free reserves, an important factor in the
very short run only. To these the presidents would add "tone"
or "feel."
There is substantial agreement in the two statements that free
reserves are generally not the target of monetary policy. At times
free reserves may be used for this purpose. But a variety of other
measures and concepts are also used from time to time. This topic
will be considered in a later section when we discuss the procedures
at the FOMC meetings and the information that is given to the
Manager.
Both groups recognize that "ease" and "restraint" are relative
matters. Both define these concepts in terms of demand and supply.
For the presidents, it is the demand for credit relative to the supply of
reserves; for the members of the Board, it is the demand for and supply
of loanable funds that determines the prevailing degree of ease and
restraint. But this is probably more a difference in wording than in
content. The supply of loanable funds is influenced by the supply of
reserves. I t is in this way that monetary policy is said to operate.
A part of the mechanism underlying and responding to monetary
operations is described by both groups. The 12 presidents separate
the effects of changes in the supply of reserves from those associated
with the demand for reserves by banks, although they suggest that
the mechanism is similar in both cases. Central to the discussion of
ease is the implicit assumption of a small short-run response by
business borrowers to a reduction in interest rates occurring as a
result of increases in reserves (V.l). No evidence has been provided
to support this contention, but it is assumed to be a basic feature of
the process. As a result banks restore and increase their earning
assets by buying securities. This reduces interest rates and adds to
the stock of money. The 12 presidents seem to recognize that the
rate of monetary expansion or contraction may differ from cycle to
cycle even if the rate of credit expansion or contraction is the same
(V).
The major point of interest that seems to emerge from the discussion is that the stock of credit or its rate of change is the focus of
policy. Money is said to respond to an unspecified and complex set
of other factors (V. 1). A part of any increase in reserves stemming
from open market purchases will be used to advance credit (VII).
Indeed this is true of any increase in nonborrowed reserves. Borrowed
reserves apparently are not used to support or increase credit according
to the 12 presidents.
The Board's reply explicitly lists the ways in which banks will use
reserves in periods of ease following periods of restraint. The details
differ slightly, and the discussion is less informative, but the conclusions are approximately the same as the presidents' (V). The
effect of time deposit rates and rates of interest paid by savings and
loan associations are explicitly recognized as important factors that



FREE RESERVE CONCEPT

{>3

influence the rule of monetary expansion and the. distribution of
deposit balances between demand and time deposits.
These statements, when road in detail, seem to suggest that some
earlier criticisms of Federal Reserve pronouncements inade in this
report arc not applicable to the present FOMC. Both the presidents
and the Board of Governors appear to be aware of the differences in
the rates of credit and monetary expansion and even appear to explain
the differences in these rates in a manner somewhat similar to the
analysis presented in a previous section. Weaknesses in the free
reserve concept are more or less implicitly acknowledged and the
demand for reserves by banks is introduced "as an important influence
in the process, difficult to measure, but nonetheless capable of being
approximated by reference to observable market entities. Among
the important influences affecting the demand for reserves, shortterm interest rates relative to the discount rate appear to be prominent. Thus despite the many earlier indications to the contrary,
the replies suggest, that the Federal Reserve has abandoned, perhaps
recently, much of the previous analysis of the monetary process that
has been criticized here.
But some disturbing elements remain. Recall that a basic feature
of the free reserves concept is the view that borrowing by member
banks exerts a restrictive effect on the expansion process and that
repayments of borrowing have an expansive effect. This interpretation of borrowing remains as an anachronism. Moreover, the
emphasis is still on the rate of credit expansion, not on the rate of
monetary expansion. "Monetary policy is concerned with the overall
availability of credit." Although it "is recognized thai credit expansion does not mean the same thing wherf the rate of monetary
expansion is slow as when it is fast, "credit"- not money—is regarded
fts the factor transmitting System policy to the economy. This
means that if credit is expanding at a rapid rate, case is occurring,
even if the money supply us raluccd. The answers to some specific
questions make this clear.
Policy in some specific contexts
The questionnaire asked alxmt some specific policy situations,
1049 and 1001. The answers provide strong evidence that the
'credit" view remains dominant. The reformulation of "ease" and
'restraint" in terms of the interaction of the demand for and supply
of reserves or free reserves has not been systematically absorbed.
The analysis that plavs a prominent role in response to more general
questions is nowhere iu evidence when specific questions were
answered.
During 19-19, a year of recession, the money supply declined,
" o m the end of June 1948 to the end of June 1949, the stock of
money decreased bv more than SI billion. For the calendar year
JM9f the decline was smaller, less than one-half billion dollars.
During the earlv months of the recession in 1957, the stock of money
Ml at the rate of 2.7 percent per annum. In 19G1, a year of recovery,
foe monev supplv grew at about 2.6 percent per annum. The
members of the FOMC were asked to explain these differences in rates
of change in question 1.6.
,
,
. „ .
„
The Board of Governors replied that they were following an
active countercyclical monetary policy in both years, i.e., 1949



20

FREE RESERVE CONCEPT

and 1001. The fuel thai the stock or money did not increase during
the vcar of recession but did increase in the year of recovery was
due apparently to other forces. The 12 presidents stilted that if
the Federal Reserve were to attempt to force an increase m the monev
supply at a faster rate than the public was willing to add to its cash
balances at prevailing price levels, the result would be rising prices
and aggravation of the bulance-of-payments situation rather than
promotion of sustainable economic growth."
.
.
The Board characterized policy in 1910 and 1901 as "stimulative.
This assessment follows if one accepts the inherited notions composing
the "free reserve doctrine," described in previous sections. Toward
the middle of 1919 free reserves moved from a level of approximately
SG00 million to a level of approximately SS00 million. Again in
1901 free reserves would be interpreted to reflect a "stimulative
policy." The radical difference in the behavior of the money supply
observed during the recession of 19-19 and the upswing of 1901 is
therefore attributed to the operation of "other factors." "Among
the most important of these fact ore are the economy's demands for
bank credit, public, preferences for holding liquid assets in particular
forms, and the incentives for banks to make loans and purchase investments." But the Board's answer supplies no clues or reference?
to explanations of how these other factors operate to affect the money
supply. Xo information is given about the relevance of these "other
factors" in the money supply process.
Our own analysis\>f the two periods, based on a framework summarized in section 4, yields a radically different result. The
difference in tho behavior of the money stock in the two periods is
dominated by the difference in the Federal Reserve's policy behavior.
The reader is invited to consult the chart on the growth rate of the
extended base in the appendix for a summary measure of policy action.
An inspection of the chart indicates that'the extended biuse had a
negative growth rate of about — S50Q million during most of 1949; in
1901 the growth rate was positive and rapidly accelerating.
Monetary policy was thus strongly stimulative in 1901 and strongly
deflationary in 1949. The differential behavior of the money supply
thus reflects the difference in the policy pursued by the Fecferal Reserve authorities. "Other factors" do not explain tho deflationary
policy pursued in the 1949 recession. Indeed, some relevant "other
factors" helped to offset the deflationary consequences of the. Federal
Reserve's behavior in that year. Foremost among these factors is
the public's reallocation of its "payment monev" between currency
and checking deposits.
The decline in the money supply in 1949 was not the result of
"other factors" compensating a "properly "stimulative policy.'*
Neither did it Teflect, the working of a substantial lair between policy
actions and the responses of the monetary svstem. The decline in the
money supply reflected the policy pursueel bv the Federal Reserve
authorities. There simply was no "stimulative policy" during the
recession of 1949.
The presidents' answer indicates that an increase in the money
supply beyond the volume desired bv the public would onlv raise
prices and aggravate the balaneo-of-pavments deficit. The last point
of course has no bearing for 1949. Ilut the central portion of the
presidents' answer appears to deny anv effect of increases in the
money supply on real output. The effect would be completely ex


FREE RESERVE CONCEPT

{>3

hausted by rising prices. The answer provides, of course, no analysis
or evidence supporting the contention that an increase in the money
supply during a recession only raises prices without raising real income.
To our knowledge no analysis has been performed by the Federal
Reserve in order to present a reasonable case for this contention. A
mass of contrary evidence suggests the opposite conclusion.
Question III asked the FOMC to interpret some published statements by the present Manager of the System Open Market Account.
It was asserted that policy has shifted toward "slightly less ease"
in 1962. Our observations indicate that free reserves declined with
little noticeable effect on the stock of money and credit or on interest
rates. The replies of the Board of Governors and the 12 presidents
were similar except for one point. Both agreed that policy contributed to expansion in 1962, that the policy of "slightly less ease" was
reflected in money market rates, particularly on Treasury bills and
Federal funds, and that it was not clear that "credit" was restricted.
The presidents add that they did not intend to restrict credit but
only to increase interest rates. This appears to deny any effect of
increased interest rates on the demand for loans, a position that is
inconsistent with their explanation of the operation of monetary policy by inducing changes in interest rates.
AVhat are the facts about interest rates and free reserves in 1962?
The monthly averages of daily figures indicate the following for
particular months:
TABLE I I I - L . — F e d e r a l funds

rates and free reserves for selected months of
Federal funds
rate in
percent

Month
December 1961
January 19G2
March 1962
June 1962
July 1962
August 1962
November 1962
December 1962

,

2.35483
2.15322
2.84677
2.68333
2.70967
2.92741
2.94583
2.92741

1961-62
Free reserves in
millions
$419
546
379
391
440
439
473
268

Federal funds rates increased strongly from January to March.
Thereafter, they fluctuated in a rather narrow range. When the
System allegedly shifted toward "slightly less ease" in mid-June, there
is almost no sign of an increase in the Federal funds rate, and there
is a rise in free reserves on a month-to-month basis. It was not until
August that the Federal funds rate rose above the rate prevailing in
March. Thereafter it fell slightly, rose in November, and was the
same in December as in August to five decimal places. In late
December, when the System again shifted to "slightly less ease," free
reserves fell noticeably. But that is the only indication of slightly less
ease in the monthly figures.
Treasury bill yields were higher in December than in several
previous months. But the peak for the year occurred in July, and
yields were lower at the end of the year than in the middle. The same
is true in general for 6-month bills, 1-year bills, longer term Government bonds, and municipal bonds. Moreover, the annual rate of
change of the money supply was one of the largest for any 6-month
period since late 1951, 6 percent in the latter part of 1962 against



Ig

FREE

RESERVE

CONCEPT

minus 0.9 percent in the early part of 1962. 40 ^Aside from the change in
free reserves in December, it is difficult to find any indication of
"slightly less ease" in the information which the Federal Reserve
replies referred.
One last answer should be mentioned. Question IV asked both
groups to explain what was meant by an "even keel." References
to the "even keel" policy are not uncommon in System statements,
but we had not been able to find any explanation of what was supposed
to remain "even." The presidents replied that maintaining an "even
keel" meant that no action would be taken that would alter conditions
in the financial markets before, during and shortly after Treasury
financing operations. A minor policy shift might be undertaken.
The Board of Governors' reply was slightly more explicit. There
are no changes in rediscount rates, reserve requirements or reserves,
and money market conditions large enough to cause a change in
expectations. We interpret this to mean that there are no "dynamic"
operations.
The presidents add that free reserves wrould not be kept constant.
Some effort would be made to keep them in a range. But, it is added,
free reserves are a highly imperfect indicator of market atmosphere
(IV.3). Emphasis is on "reserve availability/' in a context that
seems to suggest that total reserves, less the reserves required to support Treasury deposits, are taken as the reserve guideline. However,
"every market situation is unique" and no general conclusions can
be drawn (IV.1).
The Board of Governors do not so clearly deny the relevance of
free reserves as a market indicator. Instead, they repeat that constant free reserves are not inconsistent with an unchanged monetary
position. This suggests that some operations on free reserves may
be attempted.
At first glance it is difficult to reconcile these differing replies to
specific questions. How can a single committee have differing interpretations of the meaning of policy? How can a particular measure
mean slightly different things to different groups serving on the same
committee and making policy decisions? The answer must be that
policy directives are made in relatively broad form and that specific
meanings have not been assigned or agreed upon. Consideration of
the available information on the procedures of the FOMC suggest
that this is the case. After a summary of the evidence in this section,
we will return to that discussion.
Summary: The modified free reserves doctrine
The evidence from the statements submitted by the members of the
FOMAC is subject to a number of interpretations, particularly when
read m the light of earlier statements and of applications of their conceptions to particular events. It seems clear that no attempt has
been made to write down an explicit statement of the mechanism that
relates the actions of the Board of Governors and the FOMC to the
stock of money or credit, and to investigate how much can be explained by the particular mechanism. Even if one firmly believes that
judgment" is the most important element in decisionmaking, i*
does not follow that facts and evidence are useless. Even " j u d g m e n t "
must be shown to be relevant.
Au^^^l^^an^O^obcr m ™ ™ *




S U p p , y 1116

^ ^

i

n

the

Fcderal

Reserve Bank of St. Louis,

FREE RESERVE CONCEPT

{>3

Yet there is little or no evidence provided that shows or denies that
a particular measure of reserves, perhaps augmented by some measures
of interest rates, has a clear and definite relationship to the stock of
credit or money or to some other magnitude. Evidence of this kind
would seem to be a prerequisite for policy and for the belief that a
certain conception is valid as well as the belief that a particular
measure of reserves is more useful than certain other measures.
Without analysis and detailed evaluation of the evidence, it is extreme^ difficult to improve understanding of the mechanism connecting monetary policy with the economy.
For these reasons, we attempt to set down in explicit form the
mechanism that seems to emerge from the statements quoted. Enough
has been said to indicate that there are areas of disagreement within
the FOMC, that at times some factors are considered to be more
important than others. But unless detailed evidence is used to
support these assertions, there is little chance that monetary policy
operations can be improved, that the "degree of control" can be
increased. Moreover, the statements quoted above generally do not
deny that free reserves are an indicator and, at times, a target of
policy. Generally, they suggest that free reserves are only one of the
elements that must be used. It is for that reason that the conception
is referred to here as "the modified free reserves doctrine."
We believe that the free reserves concept is regarded as one of the
crucial links in the so-called credit mechanism or the money supply
process. To represent the view that a particular level of free reserves
does not always bear the same relation to the rate of change of credit,
three interest rates are explicitly mentioned as modifying factors.
These are the rate on Federal funds, the rate on Treasury bills, and the
discount rate. But the references usually suggest that two of these
rates, the Treasury bill rate and the Federal funds rate, must be measured relative to the third interest rate, the discount rate. Thus, it is
not the absolute yield on Treasury bills or Federal funds that matters;
these interest rates must be judged relative to the prevailing discount
rate. In the very short run, the distribution of free reserves is said to
be of importance also. In addition, some ambiguous fringe elements
remain, their position unresolved in Federal Reserve thinking.
The core of the many and often divergent statements of the Federal
Reserve may be summarized as follows:
(1) Given the prevailing economic conditions, the demand for
and supply of reserves can be expressed in terms of the level and
distribution of free reserves, and the two measures of relative
yields. These factors are influenced by monetary policy and in
fact, dependent, at least in part, directly on policy operations.
(2) The rate of expansion of bank credit is dependent, at least
in part, on the way in which monetary policy operations influence the factors affecting the demand for and supply of reserves. The rate of credit expansion is dependent on monetary
policy also, but indirectly rather than directly. The intervening
relations are the measures used to summarize the demand for
and supply of reserves.
(3) Recent pronouncements of the Federal Reserve authorities
do not deny or reject the inherited free reserves conception. Instead, they compound the confusion by introducing new notions
that contradict the old in important respects. But the older
ingredients remain, and the inconsistencies are not resolved.



30

FREE RESERVE CONCEPT

Anions the most fruitful of the "new ideas" is the recognition that
bank indebtedness and the prevailing level of excess reserves
result from systematic choices made by bankers 1 1 response to pre1
vailing market conditions. But the older conception, emphasizing
an alternative view of these entities as elements imposed on the
banking system, remains. There is no evidence tlmt this conflict
has been recognized. If the Federal Keserve attempted to
develop a coherent and useful conception, the conflict would
become apparent. More important, it is doubtful that the new
conception has reached the policymaking bodies.
The dependence of the rate of expansion of bank credit in the short
run on the level and distribution of free reserves and relative interest
rates does not imply an absence of other influences. There may be
changes that arise by chance, e.g., errors in the interpretation of
Federal Reserve policy, predominantly local influences that do not
balance out nationally, and other random events. Moreover, there
may be lags in the effect of some of these factors 011 the rate of expansion of bank credit or other influences that have not been mentioned in their statements. No amount of private speculation and
guessing can substitute for a critical examination of the extent to
which the rate of expansion of bank credit or of money has been
influenced by the factors summarized above. Only after the attempt
has been made to examine the evidence in relation to the presumed
mechanism can we intelligently accept, reject, or modify the conception.
The heart of our intended criticism of the Federal Reserve System
is that we have found no evidence which suggests that any detailed
tests or systematic evaluations have been performed. This has had
three important results:
(1) It has prevented a clear formulation of the relation between
monetary policy and the rate of credit expansion that can be
used as a guide to future policy operations;
(2) It has prevented any thorough internal evaluation of the
successes and failures of past policy action as a guide to improved
understanding of the process and the avoidance of future errors;
(3) It has given Congress and the public very little real understanding of the power or lack of power of monetary policy as a
means of promoting employment and price stability.
Does the modified free reserves doctrine permit the Federal Reserve
to predict the rate of expansion of bank credit? Has it worked well
in the past? Can it be relied on in the future? How much room must
be left for judgment or "feel" in the short run or in the long run?
Does it work best in expansion or in contraction? Does monetary
policy merely "push on strings" as has been asserted? Is one measure
of reserves a better indicator of the impact of policy than another?
useful answers to these and other questions are not obtained by
intuition or by unsupported judgment. Analysis and evidence are
required, but before such evidence can be usefully appraised, the
conception must be specified clearly. If our interpretation of the
* ederal Reserve s view of the monetary process is incorrect or deficient,
our test of the presumed conception will be inapplicable or irrelevant.
If this is the case, we ask only: What is the Federal Reserve's view
ot the monetary mechanism and where is the analysis and the evidence
J
to support it?



SECTION

2—EVIDENCE

FROM

ANNOUNCED

CHANGES

IN

POLICY

Another source of information bearing on the use of the doctrine
that is centered on free reserves comes from the "Record of Policy
Actions" published in the Annual Reports of the Board of Governors.
These records contain indications of the policy that was agreed upon
at the meeting of the FOMC. If the indicated changes in policy are
quickly reflected in the prevailing level of free reserves, this would
suggest that free reserves are used as either a target or indicator of
policy. If there is no clear relation between changes in the level of
free reserves and changes in announced policy, the evidence that free
reserves are a target or indicator of policy is weaker.
It might appear that this second source of evidence is redundant.
Didn't the previous section conclude that the "modified free reserves
doctrine" is a formulation of the conception that is used by the
Federal Reserve? While the question is answered in the affirmative,
it does not resolve the issue. The statements that we have quoted
come mainly from the remarks of the staff and the members of the
Board of Governors and the FOMC. It is the Manager of the S3^stem
Open Market Account who carries out the policy. The evidence in
this section is presented to indicate the way in which he interprets the
policy decision made at the FOMC meeting. An indication of his
interpretations of the policy decision can be observed by comparing
movements of free reserves with changes in FOMC policy. A systematic association of policy changes and changes in free reserves would
support our contention about the role of free reserves in the Federal
Reserve's policy conception.
To evaluate the eviaence in this section, it is helpful to understand
the relationship between the Manager and the FOMC. Unfortunately, the detailed records of the FOMC meetings are not released,
so we must rely on the condensed information that is made available
and occasional comments about procedure made in System publications and at congressional hearings. From these we can attempt to
assess how mucli discretion is left to the Manager in the choice of
targets and indicators or in the magnitude of open market operations.
Before presenting the evidence relating free reserves to the Record
of Policy Actions, we discuss the role of the Manager and the control
over his operations exercised by the FOMC.
THE FOMC AND THE MANAGER

The Manager is not a member of the FOMC. He has no vote in
the policymaking process, but he has an important voice in the deliberations. He briefs the members and other participants about
the details of System operations and the factors affecting reserves in
advance of each meeting. He provides a detailed weekly summary
of operations, and a less detailed daily summary, to each of the presidents and each member of the Board of Governors. He brings the
^formation up to date at each meeting of the FOMC in a written




31

Ig

FREE RESERVE CONCEPT

and verbal statement. He provides and discusses the estimates of
the short-run movements of float, Treasury balances, currency etc.,
that are deemed to be important between meetings of the *OMU
With the members of the staff of the Board of Governors, who describe
the prevailing domestic and international economic climate, he provides the essential background information that is relevant to the
decision taken.
Each member of the FOMC and each president, whether currently
a member of the Committee or not, has the opportunity to express his
views about past policy and desirable future policy. The Manager
makes notes for his future guidance. His notes, a statement of consensus, a rather vague directive, and an unofficial set of notes taken
by the staff of the Board of Governors are the written guidelines
available to the Manager between meetings. By the time of the next
meeting, currently at the end of 3 weeks, the notes taken by the staff
of the Board of Governors are summarized in a report that is published
as a part of the annual Record of Policy Actions. Such records do
the Manager little good. He must rely on the more informal
documents.
The participants in the discussion offer a rich variety of suggestions
and criteria. There is no official format for statements. Even if
there is agreement about the direction of policy, there may be disagreement about the method of bringing about greater "ease or restraint"
or the size of the desired change. One participant may conclude that
greater restraint should occur and may mention an amount or range
of amounts by which total reserves should be reduced. Another may
clarify his statement by suggesting a range of free reserves, lower than
the range prevailing, as an indication of the increased degree of tightness. Still a third may indicate an increase in Treasury bill rates or
a level of such rates that would in his judgment represent the desired
increase in "restraint." A fourth participant may summarize his
position in terms of "tone" or "feel" of the money market.
There are 19 participants at the meeting. There is little or no
apparent attempt^ at most of the meetings, to summarize the statements of the participants in terms of an objective. (Occasionally,
as indicated in the preceding section, a specific target is named, for
example, free reserves, bill yields, or total reserves.) Instead, a
"directive" is issued or reissued to the Federal Reserve Bank of New
York, as the agent for the System, and a statement of consensus is
made to summarize the discussion. We consider each of these in turn.
The directive
This formal document has been a curious admixture of detailed
restrictions and broad policy goals. In the immediate postwar period
and until mid-June 1955, the directive was addressed to an executive
committee of the FOMC. The most common form contained an
instruction to "provide for the credit needs of commerce and business"
or to relate the supply of funds in the market to the needs of commerce
and business." This was the "a" part of paragraph 1. It referred to
the seasonal aspects of the problem. The "b" part of the paragraph
contained a broad indication of economic or monetary policy. For
example,^the Executive Committee and later the Federal Reserve
Bank of New York would be directed to "maintain orderly markets,"
prevent disorderly markets," "avoid deflationary tendencies/'




FREE RESERVE CONCEPT

{>3

"avoid deflationary tendencies without encouraging a renewal of
inflationary developments." or some similar generality.
In contrast, the remainder of the directive contained very explicit
statements. The Executive Committee or the Federal Reserve Bank
was told that they could buy or sell in amounts that would change
the holdings in the System account by at most a specified number of
dollars, that they could only hold a specified dollar amount of Treasury
certificates of indebtedness issued directly to the Reserve Banks, that
they could exchange directly with the Treasury only a specific amount
of securities for gold certificates. During much of the period, agreement was reached that only short-term securities with specific maximum maturity would be used in regular operations, the so-called
"bills onlv" policy.
The "li" clause of paragraph 1 was sufficiently broad that in many
years it was possible to cliange the policy without changing the directive. The converse was also true. At times, the directive would be
changed, but it would be noted in the Record of Policy Actions that
there was 110 change in policy. One example of the former type
occurred in 1950-52.' A new directive was adopted at the meeting of
August 18, 1050. Thereafter, in the words of the Committee, the
directive issued was "in the same form" as the directive issued at the
previous meeting. This phrase reappears from meeting to meeting
until late in 1952. Even the famed "Accord" did not require a
change in the form of the directive. At times during the period, the
Record refers to the need for greater "restraint"; at other meetings,
the Committee expresses satisfaction with the degree of "restraint"
and states that more restriction is unnecessary; at the meeting of
August IS, 1950, strong language was used in the statement accompanying the directive to indicate that the FOMC and the Board
"are prepared to use all the means at their command to restrain
further expansion of bank credit." At times the policy is described
as one of "neutrality." Conversely, the report in 1957 notes that
''four chamrcs in the wording of the directive of the Open Market
Committee were made during 1957. * * * The January 8 and March 5
changes continued policies * * * in effect * * *."
The most recently published directives, for the year 1962, are
somewhat less vague. They provide a more detailed statement of
the framework in which the Manager must operate. For example,
was not uncommon in 1962 for the FOMC to add a paragraph
jpdicating how the policy should be implemented. Phrases such as
. Provide moderate reserve expansion," "avoid downward pressure on
interest rates," "foster a moderately firm tone in the money market,
or "offset the seasonal casing of Treasury bill rates" have been added
to the directive.
.. .
One wonders why it is desirable to make very explicit statements
about purchases of gold certificates from the Treasury or the use of
Wis onlv and very vasyue statements about the obiectjves of monetary policy. One wonders, also, why it is possible to indicate a very
cI
*ar direction at times and no clear direction at other times. I<or
^ample, the directive of the meeting of December 19, 1961, contained
the comparatively explicit statement that the Manager should proa -somewhat slower rate of increase in total reserves than
d
*nng the recent months. Operations shaU place emphasis on
continuance of the 3-month Treasury bill rate at close to the top of the



Ig

FREE RESERVE CONCEPT

range recently prevailing. No overt action shall be taken to reduce
unduly the supply of reserves or to bring about a rise in interest rates."
If the Committee at times can issue a clear statement of the specific
intent of policy, why must it be vague at other times?
The consensus
Part of the answer to the question lies in the role of the consensus.
At almost every meeting of the FOMC, a statement of consensus,
that is not part of the directive, is attached to the Record. This statement summarizes the views of the Committee. After each participant
expresses his views, the Chairman indicates the direction of policymore or less "ease or restraint" or no change in direction. The statement of consensus and accompanying remarks provide a direction for
policy until the next meeting. Only on rare occasions does the
"consensus" take note of a specific target like reserves or bill yields.
More often it takes refuge in vague phrases to indicate the direction
of change in policy, if any.
The most obvious reason for the vague directive and the rather
broadly stated consensus has already been noted. There is incomplete agreement about the immediate target of monetary policy.
This decision is left to the Manager because agreement about a desire
for "greater ease or restraint" does not mean that there has been
agreement about the exact meaning of the policy. If, as we are told
by the 12 presidents (II.1, last paragraph) a variety of indicators are
used, it is not unlikely that the Manager is left relatively free to
choose the immediate target or measure that seems correct to hini.
To return to an earlier example, when the various members or participants use total reserves, market interest rates, free reserves, and perhaps "tone" as their indicators of desired policy, it may be impossible
for the Manager to satisfy all of them. Meeting the interest rate
goal suggested by one participant may increase total or free reserves
by more or less than some other member thinks desirable. Under
such circumstances, it is difficult for the presidents and Governors to
reprimand or criticize the actions taken. 1
Furthermore, the Committee is concerned with day-to-day operations. Errors in float projections, shifts in the Treasury "balance,
and numerous other short-term changes can and apparently are used to
explain deviations from the policy desired by a particular member or
members. (For a recent example, see the report of the meeting of
January 1962.)
If the Committee cannot or does not agree upon a specific increase
or decrease in reserves or some other target, and if it does not agree
upon any specific target, it is left for the Manager to decide whether
the actions that he takes are appropriate in the light of the general
policy statement or consensus. This provides a large measure of
autonomy for the Manager that is further encouraged by the reference to numerous and possibly divergent indicators with which
participants amplify their statements. It is not uncommon for total
reserves and bill yields to move in the same direction for several days.
Which criteria does the Manager follow if both have been used by the
Committee members as indicators of desired tightness? The choice
he was never
***** to explain his action in tenon of tbe




FREE RESERVE CONCEPT

{>3

must be made by the Manager. If five, or six different criteria are
used, the discretion left to the Manager is enlarged.
In practice, some devices have been developed to exercise a measure
of control over the autonomy of the Manager in carrying out the
consensus. Each morning there is a telephone conversation between
the trading desk at the New York Bank, the Board of Governors staff
in Washington, and one Reserve bank president. The Manager, or
one of his principal assistants, outlines the plans for the day. The
members or staff of the Board of Governors or the president of some
Reserve Bank can question the decision of the Manager and the extent
to which it fits within the framework of the directive and the consensus.
But the Manager can always point to a large number of market occurrences that indicate to him that the decision is an appropriate
interpretation of the sense of the last meeting. If there is no clear,
tested conception of the process by which open market operations,
interest rates, and other observable market phenomena affect t i e
desired portfolios of banks and the public and no clear statement of
FOMC objectives, only unanalyzed judgment can be used to interpret the events that the market is recording and their relation to the
FOMC's consensus. Responsible men often differ in their personal
judgments. It would be surprising if the views of the Manager, who
has the responsibility for the final decision, did not generally prevail.
Other devices used to inform the committee members; e.g., a telegram outlining the telephone discussion for those who did not participate, suffer the same weaknesses. Ultimately, the Chairman or a
member of the FOMC must either accept the judgment of the Manager, attempt to modify it, or substitute his judgment for the opinion
of the Manager. There is no record of the number of times that such
differences have occurred or have been appealed to the Chairman of
the Board of Governors, but it would be surprising if the Chairman
interfered frequently to reverse or alter the daily decisions of the
Manager.
A clear goal for monetary policy has recently been formulated, the
maintenance of a particular short-term interest rate. The members
of the Committee are consequently in a better position to judge the
actions of the Manager. In essence, the Committee has taken a
longer run view of policy operations. On the basis of a belief—largely
unsupported by detailed examination of direct evidence—that higher
short-term interest rates in the United States wrill reduce the gold
outflow, they attempt to maintain a particular interest rate. The
administrative effect of this policy is to instruct the Manager that
whatever "defensive" operations are taken daily must be tailored to
the longer run policy of maintaining the Treasury bill rate. Any
member who wishes to judge the actions of the Manager need only
look at the bill rate prevailing in the market to see whether or not the
primary goal is being achieved.
Very similar procedures could be followed in principle if a particular
rate of monetary or credit expansion is taken as the desired goal of
monetary policy. But to do so requires an understanding of the
precise effect of changes in reserves on the rate of monetary or credit
expansion. This requires an explanation of the behavior of the stock
of money or credit that usefully predicts future movements with
reasonable accuracy. Until there is internal agreement within the
FOMC on the relevance and utility of a particular conception of



Ig

FREE

RESERVE

CONCEPT

monetary mechanisms, there can be no agreement about the change
in some measure of reserves required to achieve a particular longrange goal. We believe that it is primarily because understanding of
the monetary process has not been developed that the FOMC has
been concerned with the day-to-day operations in the past and the
Manager has developed a large measure of autonomy.
Our understanding that the Manager has exercised substantial
autonomy in the past is supported by evidence. The same evidence
also suggests that during much of the postwar period, the Manager
interpreted the directive and the consensus in terms of a level or
range of free reserves. That is, he operated in the market to achieve
short- and long-range objectives by altering the level of free reserves.
We turn to consider the evidence.
Policy objectives andjree reserve levels
Each author of this report independently read relevant portions of the
Record of Policy Actions for each meeting of the Federal Open
Market Committee from 1946 through 1962. Since the language is
often vague, a question of judgment is involved at times as to whether
or not there is a minor adjustment of policy. At o t h e r times, the
signal is quite clear. Using a scale ranging from + 1 (decisive easing)
to —1 (decisive tightening), we recorded our independent judgments of
tiie meaning of the directive, the consensus, and the accompanying
remarks. We then compared these judgments and arrived at our own
consensus about the interpretation of the "Record of Policy Actions"
for each meeting.2 At times, the Record indicated no change in open
market policy but referred instead to actions taken by the Board
of Governors. These actions—changes in reserve requirements, in
discount rates, in preferential buying rates for particular securities—
were of importance particularly in the pre-Accord period. Our
records place these changes at the time of the open market meeting.
A 3-week moving average of the level of total free reserves was
computed. This eliminated some of the extremely short-run fluctuations and permitted a clearer indication of the timing of changes in
level. Without reference to our previous dating of changes in the
announced policy of the FOMC, we dated the changes in the level (or
direction of change) of the 3-week moving average of total free
reserves. The two series were then compared to indicate the relation
of announced policy changes to changes in free reserves. For the
period 1946-51, open market operations were used primarily to
control bond prices. The analysis of policy operations for these years
will be considered when we discuss the rationale and the results of
changes in reserve requirements in a later section.
During the years 1951-62, we recorded 163 meetings of the FOMC
including special meetings called for a variety of purposes. At 76
meetings, we recorded a change in policy. All changes are not of the
same importance. For example, some of the policy changes are
given a scale value of 1/8. This value is used to note a minor modification that is indicated by a statement such as "doubts should be
resolved on the side of ease during a period of Treasury financing" or
any deviation should be on the side of less restraint-" But at times,
££ese
differed at times, but they never ewent in opposite directions At most one of as felt
th wo
^ f r ^ ^
rdm gto mean a^l igh t change^icase o r Sstraiflt
By re-reading the report for the relevant meeting, a consensus was reached.




FREE RESERVE CONCEPT

{>3

such minor changes are of importance since they may be an initial
indication of a reversal of policy.
In keeping with the extremely short-run orientation of the FOMC,
most policy decisions are made for the period between meetings.
Thus our scaling of the magnitude of policy changes refers only to tne
change from the previous meeting and cannot be interpreted as an
indicator of the absolute level of free reserves. The desired policy
holds for a period of 3 weeks under present arrangements. If unexpected events occur, a special meeting may be held, often by telephone, and members of the Committee may decide on a new policy
or a new directive. Like the regular meetings, the special meetings,
generally make decisions that are relative to the prevailing policy
and are not absolute. Even when a specific criterion is mentioned,
e.g., a particular range of free reserves, the decision must be understood as an agreement that holds only until renewed or modified at
the next meeting of the FOMC.
Several questions can be considered using the moving average of
free reserves and our scaling or index of policy actions. First, how
promptly did the FOMC recognize changes in economic events and
respond to them according to the scale that we have developed?
Second, how promptly were these actions reflected in the moving
average of free reserves? We have already noted that there are more
than 70 indicated changes in the direction or magnitude of policy
operations. There are a similar number of identified changes in the
moving average of free reserves. Many of these are, as noted, minor
movements. We will consider, first, the more important movements
revealed by the record of Federal Reserve action at or near turning
points of economy activity.
The FOMC and free reserves at post-Accord turning points
The National Bureau of Economic Research provides a record of
cyclical turning points that shows the month in which the economy
is judged to have moved from expansion into contraction, or viceversa.3 In the post-Accord period, six turning points have been noted.
A brief history of the record of Federal Reserve policy decisions and
the movement of free reserves is provided for each turning point. To
indicate the decision taken, we will present a quotation or summary
statement taken from the published Record of Policy Actions for the
meeting before, during and/or after the turning point. If the direction of policy did not change at any of these meetings, the date and
uotation are given for the first meeting or meetings at which a
esired reversal of policy was indicated.
Information is also given for the movement of free reserves during
the period. It should be noted that our 3-week moving average is
not centered in the middle week. The average for the 3-week period
is given the date on which the period ended. This was done to assure
that the changes in the level of free reserves would not be reflected in
the moving average earlier than they occurred. They may have
occurred earlier, and in cases where the exact dating is of some consequence for later discussion, information is also given for the
unadjusted weekly level of free reserves.
3
This record, like our index of policy action, is based on judgement. As an alternative, the monthly
index of industrial production could be used for this purpose. At times the latter would produce slightly
different findings about the speed of response of the FOMC.




Ig

FREE

RESERVE CONCEPT

U Peak dated July 1053
Last mooting of the FOMO before the peak, June 11:
Decision: "It was the view of the Committee * * * that
policy should be one of aggressively supplying reserves to the
market." (Scale of the decision: + 1)
First meeting of the FOMC after the peak, September 24.
Decision: "Further easing would be needed to assure ready
availability of credit." (Scale: + 'Q
Movement, of free reserves during the period: The moving average
had remained between — 8500 and —$700 million since February. In
the week ending May 27, the moving average rose S200 million. Free
reserves continued to rise for several weeks and became positive in
the week ending June 10, i.e., btjore the change in policy had been
decided upon at the FOMC meeting. From late June until late July,
the moving average remained in a narrow range around i- 8500 million.
During August there was a sharp decline, but in the week ending
September 23, the moving average of free reserves increased by S200
million to -j-S267 million. Again the change occurred before the
FOMC meeting.
2. Trough dated August 1054
Last meeting of the FOMO before the trough, June 23:
Decision: A reduction in reserve requirements was to be
partially compensated by open market operations. The net
effect was further ease. " (Scale: + ) i )
First meeting of the FOMO after the trough, September 22:
Decision: "* * * Resolve doubt on the side of ease * * V
(Scale: +JJ)
First meeting announcing a clear change in policy. December 7:
Decision: "A reexamination of the policy of 'active ease'
* * * led the Committee to the conclusion that the developing economic situation did not warrant continuing as active a
program of supplying reserves as had been followed during
the preceding year * * *." (Scale: —
Movement of free reserves during the period: The moving average
remained between +8500 and +8700 million from March 24 through
June 10. In the week ending June 23, the moving average increased
to approximately S740 million and remained between 8050 and 8800
million until the week ending November 24. In the week ending
December 1, the moving average declined bv 8250 million to S541
million. Both of the changes noted in this period occurred before the
meeting of the FOMC.
8. Peak dated July 1057
Last meeting before the peak, June 18:
Decision: " * * * a firm policy of restraint should be continued for the present * * *." (Scale of the decision: 0)
Two meetings in the peak month, July 9 and 30:
Decision on July 9: 11 * * * to maintain but not to increase
the existing degree of pressure * *



FREE RESERVE CONCEPT

{>3

Decision, July 30: " * * * To keep the banking system under
substantial pressure * * V
(Scale of these decisions: 0)
First meeting after the peak, August 20:
Decision: " * * * the System account would have flexibility
in providing reserves * * V
(Scale: % (also % for meeting
September 10)) " * * * doubts would be resolved on the
side of less rather than greater restraint."
First meeting announcing a clear change in policy, October 22:
Decision: " * * * although general policy was not to be
changed appreciably, it would tend on the easier side from
where it had been in recent weeks." (Scale: +%)
Movement of free reserves during the period: For several months
prior to July 24, the moving average of free reserves remained between
-$400 and —$600 million. In the last week of July and the first
week of August, the moving average increased sharply but then
returned to the range —$400 to — S525 million until the week ending
October 16. During the week ending October 23, the moving average
of free reserves rose SI SO million to the level —$321 million. It
remained in the range of —$200 to —$350 million until December 11;
Either the level of free reserves moved in advance of the meeting of
October 22, or a sufficient volume of free reserves was supplied on
the day following the meeting of the FOMC to raise the 3-week
moving average.
No clear indication of the August 20 and September 10 decisions
are observable. Two additional indications of an easier policy were
made at the meetings of November 12 and December 3. The moving
average of free reserves fails to record any significant response to these
changes. However, there is a response to the modification of the
directive at the meeting of December 17. Free reserves respond in
the week ending December IS, gradually moving toward a positive
level and ultimately to the range +$450 to +$550 million in midMarch 1958.
Trough dated April 1958
Last meeting before the trough, March 25:
Decision:
* * operations in the System account should
be directed toward maintaining a slightly larger volume of
free reserves and money market conditions slightly
easier * * V (Scale: +JQ
Meeting during the month of the trough, April 15:
Decision: "Easing" was "contemplated" in the form of
lower discount rates and reserve requirements. (Scale:

+ }0

First meeting after the trough, May 6:
Decision:
* * the prevailing policy of ease should be
continued * * *." (Scale: 0)
First meeting indicating a slight change of policy, May 27:
Decision:
* * maintain the current posture of monetary
policy without further depressing Treasury bill rates * *
(Scale: —



40

FllKE RESERVE CONCEPT

First recognition of a major change in policy direction, August 19
(on July 29 a smaller policy change was indicated also):
Decision, July 29: "Absorb redundant reserves generated
bv emergency purchases of securities." (Scale: —
"Decision, August. 19: "that the rale of expansion in the
money supply * * * should be tempered and that operations
for the System Open Market Account should move in the
direction of lower free reserves * * *." (Scale: — fj)
Movement of free reserves during the period: The level $4.50 to
8550 million that had been reached in mid-March was retained
throughout the spring. Neither the moving average nor the unadjusted weekly data show any significant effect" of the decisions taken at
tlio meetings on March 25 and April 15. The meeting of July 8
indicated no change in policy, but free reserves moved up slighilv in
the week ending July V), perluips for seasonal or holiday reasons. 'Hie
range of 85*10 to 8600 million was maintained until the week ending
August 0. The decision taken at a special meeting on Julv IS, to
ease the money market in response to the "disorderly conditions,"
has very little effect on the moving average.
The first sign of change toward a policy of increased restraint
appcai-s in the moving average of free reserves in the week ending
August Hi. Once again this change prcctdcs the decision to restrict
the rate of growth of the money supply that was made at the meeting
of August 19.
Between the week ending August 13 and the week ending September
17, the moving average declined steadily to a range between 850 and
S125 million. No further indications of policy change or increased
restraint are noted at the FOMC meetings until December 2 and
December 10 when a desire for further tightening is recorded. As
if in anticipation of these decisions, the level of free reserves began to
fall in the week ending November 2(>, became negative in the week
ending December 3, and remained in the range 0 to - 8 1 0 0 million
until mid-March 1959.
5. Peak dated May 1960
Last meeting before peak, April 12:
Decision:
* * the consensus favored easing furthei
« i e reserve position of member banks * * V
(Scale:
+J0
Meetings during the month of the peak. May and 21:
Decisions taken: (May 3) "* * * moving moderately in
the direction of increasing the supply of reserves available to
the banking system" (scale: K). (Mav 24) "The consensus
a furt,ler
»"l>pl.v of reserves * *
(scale:
-rJa).
First meeting after the peak, June 14:
Decision:
• • any deviation should be on the side of
ease * * V
(Scale: +!i)




FREE RESERVE CONCEPT

{>3

First reoognition of a major cha.ige in poliev direction. March 1
(taken in advance of the peak):
Decision: "The Committee concluded that it would he
appropriate to supply reserves to the banking system somewhat more readily.1' This was characterized as a poliev of
"moderately less restraint/' (Scale: r}\)
Movement of free reserves during the period: In early June 1959,
the moving average of free reserves"fell below —8370 million and remained below that level, with the exception of 2 weeks in late January,
until the week ending March 2, 1900. During most of this period,
average free reserves were below —$450 million. Concurrent with
or in advance of, the meeting of the FOMC1 on March 1, the moving
average rose until it reached the range — S100 to - $225 million where it
remained from March 23 to May 17. The decision to ease further
taken at the meeting of April 12*has no perceptible influence on the
moving average. During the weeks ending April 0, 13, and 20, free
reserves are near the lop of the range indicated. There is some slight
increase in the following 2 weeks and a further increase following
the meeting on May 3. In advance of the decision on May 24, free
reserves rose during the week ending on that date and remained
between 0 and +8200 million during most of the summer.
ff. Trough data! February 1961
Last meeting before the trough, January 24:
Decision: 14* * * there should be no change in the existing degree of monetary ease * * V
(.lose attention
to the bill rate was urged"for balance-of-payments reasons.
(Scale: -}' s )
Meeting during month of the trough, February 7:
Decision: "The consensus of the Committee favored no
change in open market policy * * V
(Scale: 0)
First meeting after the trough, March 7:
Decision: "The consensus of the Committee was that the
existing monetary policy of ease should be followed * * *."
(Scale: 0)
First post-trough decision for a minor policy change, August 22:
Decision: The consensus favored continuing the policy of
early August "when a confluence of market factors contrived
to produce more firmness than had otherwise been the
case." (Scale: —!*)
First post-trough decision for a significant change in policy,
•December 19:
Decision: " • * * no substantial change from recent policies
was called for. "* * * a somewhat slower rate of increase
in total reserves than during recent months * * V
(Scale: -}.'<)
Movement of free reserves during the period: In December I960
and January 1961, the moving average remained in the range +S650
+S750 million. During the first week of February, the average



Ig
F R E E RESERVE

CONCEPT

fell $125 million to $609 million, most likely reversing a seasonal
increment in reserves during the late fall. Thereafter, the average
remained between $450 and $600 million throughout the year 1961
with very minor exceptions. There is no indication of a move toward
lower free reserves following the meeting of August 22. Indeed
free reserves rose in the week ending August 23, by $63 million from
the lower level that had prevailed in advance of the meeting. The
lower level of early August was not regained until October.
The desired increase in tightness indicated by the decision made
in late December is reflected in the average free reserves for the
week ending January 3, 1962. Thereafter free reserves returned to
the approximate range in which they had been, $450 to $600 million.
However, free reserves are generally high or rising in January as
currency flows back to the banks. The failure of the average to
rise may be an indication of the move to a tighter policy.
The record at turning points in economic activity is* summarized
in table IV-1.
TABLE I V - 1 . — A summary

of policy
post-Accord

actions and, ?novements
turning
points

of free

reserves d

1st indication of—
Date of turning point
(NBER) (month)

July 1953
August 1954._
July 1957
April 1058
May 1960
February 1961

Any change, in the
direction of policy
(day)
June 11
Dec. 7
Aug. 20
May 27
Feb. 9
Jan. 24,1961, or
Aug. 22.

Major change in
policy (day)
Juno 11
Dcc 7
Oct 22
Aug. 19
Mar 1
Dec 19

Change in the moving
average of free reserves (week ending)

Miy 27
Dcc. 1.
Oct. 23.
Aug. 13.
Mar. 2.
Jan. 3, 1962.

Before commenting upon some important issues and questions that
arise from this discussion, we will present some additional evidence of
the relation of free reserves to the desired policy changes published
in the Record of Policy Actions.
The FOMC and free reserves in 1962-63
Our earlier discussion of the targets and indicators of Federal
Reserve policy suggested that policy actions in 1962 were r e f l e c t e d
in the movement of free reserves despite the attention paid to the
level of Treasury bill yields as a target of policy action. A c a r e f u l
Teading of the Record of Policy Actions and the changes in the moving
average of free reserves for the year largely confirms our earlier
statement. Moreover, it provides additional indication of the importance of the level of free reserves in the management of the System
Open Market Account.
Our index of policy indicates one minor and three more important
changes in desired policy recorded in the reports of the meetings of
the FOMC. These changes are shown in table IV-2. All other
meetings4 of the FOMC produced no change in desired policy and are
scaled 0.
1
< The meeting of by 23 indicates
reserve requirementsOct.the Board. seasonal easing to be accomplished by a reduction In time dep^




F R E E KESEKVE

TABLE I V - 2 . — S c a l i n g of the Policy

Date of meeting

Magnitude
of desired
change

4a

CONCEPT
Record for

1962

Quotation or paraphrase of the Record

+H The majority favored no change, but "promote further expansion of

Mar. 6
Mar. 27.
June 19...
Dec. 18

bank credit."
,
Slightly more expansion in reserve availability than had developed.
"Avoid redundant reserves." "Slightly less easy policy indicated."
A somewhat less easy policy was favored by tbo majority to firm Treasury bill rates.

IS
-H

The moving average of free reserves remained in a very narrow
range during most of the year. Certain changes in level are discernible however and are recorded in table IV~3.
TABLE I V - 3 . — F r e e reserve ranges during
Date(s)

Week end in e—
Feb. 7 . „ .

Feb. i4 to Mar/iv.:::::":"

war. 21 to May
..
May 23 to Juno 1 3 . . / _
June 20 to Jul v 4
I
/
July 11 to D e c . 5

Dec. 1 to Jan. 30, iiift."}/"'
2

Feb. 6, i m to March 1963.

1962 and early

1963

Range of the level of free
reserves (millions)
$504
$425 to $450
$375 to $425
$440 to $490
$360 to $390
$100 to $450
$300 to $375
$2S0 to $315

wnT ^ a t a i l n t , l e t w o tables suggest that the level of free reserve
as not reduced in response to either indication by the Committee
a slight desire to ease. Instead, a movement toward a lower free
serve level occurred between the two decisions to ease. However,
tI?A l
changes toward tighter money markets are reflected in
e
at | - j o f f r e e c ireservess in advance of the meeting of the FOMC
Darf ii
^ s l o n w a i n ade. The change in June was reversed
P tially m July, and the moving average of free reserves remained in
n
arrow bounds during the next 5 months,
th 1 R e c o r d o f Policy Actions for 1963 is not yet available. Neverfirst months, the moving average of free reserves
ess,
the m o v e m e I 1
a
ea^ e S f?
t toward increased restraint occurred in
addV ^ r u a r y . Perusal of the unadjusted data suggests that an
aaitional desired policy change toward increased restraint was made
a
*>out the middle of May.
APPRAISAL OF THE POST-ACCORD RECORD OF POLICY ACTIONS

Three principal conclusions about the record of FOMC actions
uring the post-Accord period emerge from the data. A discussion
each of these permits additional appraisal of the policymaking
of°th r e s o f t l l e F O M C a n d t h e Federal Reserve's understanding
tjie monetary mechanism. We will consider in turn (1) the timing
J , decisions at turning points or speed of response of the FOMC r
f'
autonomy of the Manager, and (3) the meaning of the findings
or the role of free reserves in the Federal Reserve's view of the
Monetary mechanism.




Ig

FREE RESERVE CONCEPT

The timing of policy changes
The FOMC's record at post-Accord turning points, summarized
in table IV-1, is most impressive. Much of the academic criticism
of the Federal Unserve has suggested that the FOMC or the Board
is slow (o respond to changes in economic indicators. Our appraisal
of the .evidence su<r<rests tlie opposite. In particular, when economic
activity has readier! a peak and discretionary policy should move
toward "ease." the FOMC has been quick to recognize the need for
a change in policv. Indeed, our index or scaling of the Record of
Policv Actions suggests that the Federal Ueserve indicated a desire to
reverse the direction of policy in advance of the peak recorded by the
National Bureau at two of the three postwar peaks.
\Ve submit that this record is remarkably good. It should be
recalled that the turning points recorded by the National Bureau are
chosen with hindsight. But the desired direction of monetary policv
must be made by considering the detail of present and past events and
by attempting io assess the near-term future. The record at peaks
suggests an extremely competent assessment of economic data by the
stall and the use of excellent judgment by the Committee.
Some writers have presented ail alternative interpretation of a part
of this record, It has been suggested that the reversal of policy in
1951* was accidental, a response to the "disorderly conditions" that
had developed in the bond market. We do not believe that detailed
examination of the record supports this conclusion. First, the level of
free reserves began to increase before the development of disorderly
markets in June. The unadjusted weekly data for free reserves record
an increase in the level of free reserves of more than $400 million in
the week ending May 20. The moving average of free reserves places
the change in policy a week later. In any case, the change was
initiated before there was any indication of cliHiculty with the newly
issued S^-percent bonds. Second, the increase in the moving averaire
of free reserves during the "disorderly period" was reversed, while
the earlier change was not. In early August, the moving average of
free reserves returned to the range in which it had been ill early June
and remained between —8100 and -rSlOO million in every week from
August 5 to September 16. Thereafter, free reserves increased perhaps in anticipation of the decision bv the FOMC at the meeting on
September 24. Third, the response to the disorderly markets in July
1958 proceeded in a rather similar way. In 1953, the unadjusted data
show that almost $750 million of additional free reserves was supplied
during the 2-week period June 10 to 24. The newlv created reserves
were withdrawn by early August as noted above, "in Julv 195S. the
System was maintaining a policy described as "ease" and frnd not yet
indicated a desire for any significant increase in "restraint/1 Nevertheless, the unadjusted weekly data show that onlv 8250 million was
supplied temporarily and withdrawn within 2 weeks. The Svstem
acted with greater restraint., and reversed more quickly in 195$. but,
in both 'disorderly markets/' the previously existing level of free
reserves.was restored. Later additional policy actions were taken.
It is difficult to understand why a temporary increase in reserves
should ease the banking system in one case but "not in the other. But
that conclusion seems implicit in the argument of those who regard
the prompt response by the FOMC to the developing i ©cession in




FREE

RESERVE

CONCEPT

{>3

1953 as a fortunate accident. Finally, the FOMC showed again in
I960 that it was capable of recognizing a deceleration in the pace of
economic activity before the cyclical peak was reached.
Table IV-1 also suggests that the lag between the trough and the
indicated desire to tighten is longer than the lag between the peak
and ihe indicated desire to ease. More importantly, the table suggests
that these lags are comparatively short. The latter conclusion is
directly opposed to the finding of 13rown, Solow, Ando, and Lvareken,
in a study prepared for the Commission on Money and Credit.* The
difference between the findings here and findings of Brown et. al.
arises from a difference in measurement procedure associated with
different conceptions about the monetary process. Our procedure is
related to tlie dominant notion guiding "the Federal Reserve's evaluations and policy actions. Brown et al. select some maximally achievable stock of bank credit as an index of modifications in monetary
policy. The lag in the appropriate motion of this index behind cyclical
turning points is then interpreted to measure the Federal Reserve's
"recognition lag," its habitual lag in recognizing changing economic
circumstances. But the la<r measured by Brown et al. permits an alternative interpretation which denies the significant occurrence of a recognition lag. at least for the peak. We have already submitted
evidence indicating a recognition lag not longer than the period required for purely 'technological" reasons to collect and prepare the
necessary information. This period is substantially shorter than the
la«r estimated by Brown et al. We do not Question the existence or the
relevance of the lag obtained by Brown et al., but we do contend that it
is not. attributable to a lag in recognition. The lag observed by Brown
et al. is t he natural outcome of policy actions based on a misconception
about the structure of the monetary process. In case of an onset ting
recession the Federal Reserve observes the rapid upsurge of free
reserves and feels that it has pursued a "stimulative policy." It has
rapidly become aware of the change iu circumstances and adjusted
the prevailing policy nocture according to its own conceptions. But
guidance based on* the free reserve doctrine frequently leads the
Federal Reserve authorities into a position where they believe that a
countercyclical monetary policy is underway, while for many months
almost no relevant actum is' taken. Consider, for example, open
market purchases intended by the Federal Reserve to "ease reserve
Positions" and thus exert an expansionary effect. Suppose that the
injected reserve funds are used to repay borrowings. Total reserves
Jire therefore unchanged, and Brown et al. would indicate no change
in Federal Reserve policy, although the FOMC, judging policy actions
{n terms of the modified free reserve doctrine, would believe that it
}|ttd moved toward "ease." While we have indicated that the modified free reserve doctrine is a defective tool, we have suggested quite
strongly that it is the doctrine that the FOMC uses. However corr
W or incorrect the views propounded by Brown et al., one cannot
measure the lair in recognition according to a theory which does
n
<>t guide actual policymaking. For this reason, the measurement
the lag between turnimr points and Federal Reserve recognition
action should not be measured by the rate of change of maximum
J

K

: <•«'>• Mrown. R. M. Solow. A. Ando. and J. Kiuvkcn. "La* in
for Hie Cmmislssloii on Money an<l Crwlil.




M m k W rolicy." pns-

Ig

FREE RESERVE CONCEPT

bank credit. The findings of Brown et al. simply indicate again one of
the problems with the free reserve (or modified free reserve) doctrine."
The fact that the lag in changing policy at troughs is longer than
the lag at peaks does not necessarily indicate a slower recognition of
recoveries. I t is doubtful that more rapid movement t o w a r d ^
strain t" would be desirable from the viewpoint of either the FOMC
or the economy. There is always the danger that a more rapid
reversal of policy at the trough—'"tightening" faster—would smother
the incipient recovery. Since the FOMC and the staffs of the Federal
Reserve have not attempted to appraise carefully the relation of
monetary policy to the stock of money and the pace of economic
activity, they have no information about the length of the lag between
their decisions and their effects on money and national income. Still,
it would appear to the outside observer that the Federal Reserve
authorities assume the lag to be very short. This assumption is at
least consistent with the quick reversals in policy direction which
may be observed on occasion. The Record of Policy Action for the
period 1955-57 is most instructive in this respect.
Several reversals of policy were made during that period since the
judgment of the FOMC indicated that the economy may have reversed direction. The year 1956 is particularly interesting in this
regard. There were 10 major changes or minor adjustments in
policy at the IS meetings held. Only a few of these changes are
considered here to indicate the frequency of some major policy
revisions that the FOMC desired to institute.
January 24: "a shift in emphasis seemed desirable"
"some relaxation of restraint appropriate in the near
future" (Scale: +%)
March 27: "The supplementary clause which was introduced in January 24, was eliminated" "instructions to
take into account deflationary tendencies * * * was not
consistent with the existing situation." (Scale:
May 23: "the Committee agreed that during the immediate future additional reserves should be supplied to take
care * * * of growth needs." The qualifying phrase deleted from the instructions in March 27 was reintroduced.
(Scale: +%)
August 7: The qualifying phrase reintroduced on May 23
was deleted. Instructions required that attention be directed
toward inflationary developments. (Scale: —
In the space of 8 months the desired direction of policy changed
at least four times. For the many reversals of policy in 19*56 to have
an important bearing on the pace of activity three conditions must be
satisfied: (1) the Federal Reserve's control mechanism must operate
with very short lags; (2) the money supply must respond rapidly to
changes m the level of free reserves and (3) the pace of economic
activity must respond very quickly to changes in the stock of money.
« Brown et al. recognize in their appendix the problem raised here They dismiss free n e r v e s for a reason
analogous to the one mentioned in thet h e t h c oThis7 seems to miss the point Th? l ^ b c t ^ n recognition
test, r y 0
if
TcT^ln^ZUiZ'used by the F O MC has m t T e ^ S S ^
S J S i S ^ d S S S r S t t t h e 0maximum tstock of bank credit, criticism should J*
f
tha
S t a S
^ e FOMC takes to r e s i z e changes
For further evidence of the problem raised by inappropriate judgment of its actions by the FOMC, see
our discussion in "The Federal Reserve's Approach to Policy," of
^
changes in money and "credit" during economic expansion and wntmrtion




FREE RESERVE CONCEPT

{>3

But if the control mechanism operates only slowly on the level of
economic activity, these frequent reversals in the direction of policy
nave no justification. The inclination to reverse policy direction
easily and rapidly is likely to generate uncertainty and raise interest
rates comparatively.
^ Wkep the turning point did occur in the summer of 1957, the
federal Reserve moved more slowly than it had at the start of the
earlier or later postwar recession. There was recognition that the
economy had been "moving sidewise" for several months, but more
tnan 3 months passed before there was a major change in desired
policy. It is not unlikely that the experience of 1956, and the judgment by members of the Committee that their response to the events
of 1956 had contributed0 1 "inflationary pressures" in 1957, delayed
to
*n o a r t u r ™ n £ P " *- The record indicates that one member
ttnw-i ® 4
Governors opposed all movements toward ease at
rOAlC meetings until December and opposed a reduction in the discount rate in November because of his fear of "inflation."
1 he lesson from this experience seems to be that the FOMC has had
a good—and perhaps excellent—record in judging the timing of postAccord turning points. Whether or not its judgments between turning points have been appropriate depends on the relevance of the
Modified free reserves doctrine. This is not solely a judgmental
flatter; it is one that requires detailed appraisal of evidence. Some
information on that question will be presented shortly. We will
reopen the question at that point.
The autonomy oj the Manager
The tuning of the responses of free reserves shown in table IV-1,
and the discussion in the text about the movement of free reserves at
or near cyclical turning points, suggested quite strongly that the free
reserve levels often change in advance of meetings of the FOMC.
Additional evidence often pointing in the same direction is provided
at tunes of other policy changes or modifications. Some of the
policy actions in 1956 have been indicated in the preceding section.
Ranges in free reserve levels around these dates are considered
here, using the 3-week moving average to date changes in level:
,
The decision of January 24: The moving average remained
m the range — S350 to — $500 million in the fall of 1955. It
rose to —$192 million in the week ending December 28 and to
—$50 million in the first 2 weeks of January. Following the
FOMC meeting indicating a desire for further ease, the free
reserve level began to decline. If there was any movement
toward ease, it appears in advance of the meeting, not after, if
judged by the moving average of free reserves.
2. The decision of March 27: A decision was made to tighten.
The moving average indicates a reduction of S140 million from
the level prevailing in the previous 2 weeks. This reduction
came in the week ending March 28 and returned free reserves to
the level existing in mid-December. Again the Manager appears
to have moved in advance of the FOMC decision.
3. The decision of May,23: A movement toward greater ease
was indicated at the meeting. The moving average had remained between —$450 and —$600 million from late March
until the time of the meeting. Free reserves increased slightly



Ig

FREE RESERVE CONCEPT

in the week of the meeting. but did not leave the prevailing
range until the week ending June 0, when they rose to
million. Thereafter tliev continued to rise for several week
until tlicv attained the range —SI00 to — $200 million where they
remained until early August. This policy change is not reflected
in the moving average until after the meeting of the FOMC.
4. The decision of August 7: The FOMC instructed the
Manager to increase restraint. The moving average responded
rather promptly, but again the response came after the meeting.
Tn the week ending August 8. there is a slight fall in the level, hut
free reserves remain in the previous range. In the following
week, ending August lo. the moving average fell S125 million.
We interpret this response ns one that occurred after the meeting,
although it could be an advance indication of the decision taken
at the meeting on August 21 that called for additional restraint.
At two of the four meetings in 11)50, at five of the six turning point?
shown in table IV -1. as well as at other times, the moving average
of free reserves appears to have changed direction in advance of the
decision by the FOMC. Moreover, the close correspondence between changes in the moving average of free reserves and the decisions
of the FOMC is unlikely to reflect the operation of chance factors
or solely the behavior of the banks and the public. Instead the
relation between decisions of the FOMC and the changes in the moving
average suggest that free reserves arc an important part of the control
mechanism used by the Federal Keserve.
Most important* the evidence suggests that the Manager has much
wider latitude for policy operations than has generally been conceded.
We have noted that he is largely responsible for the day-to-day
operations that arc an important part of open market operations.
And we have seen that an important analyst of System operations.
11. V. Itoosa, has concluded that the "dynamic" or*policy operations
"emerge from the day's confusion as a dominating forcc.7 " But only
by examining the evidence of the relation of FOMC decisions to the
movement of free reserves lias it been possible to observe that the
FOMC often ratifies a decision that has already been made rather
than directing policy operations.
If there is no clear guide to policy operations and no clear understanding of the relation of policy operations to the rale of change of
the stock of money and credit, it becomes extremely difficult for
members of the Committee to make independent judgments about
the state of the market or to interpret the prevailing policv. While
our evidence does not indicate that the Manager" is making the
policy decisions, it does suggest that frequentlv someone or some
group other than the full Committee is making policv decisions that
Congress has entrusted to the Federal Open Market Committee.
Free reserves as an indicator oj desired fcwe and restraint
The evidence on the relation of policy decisions to changes in the
moving average of free reserves bears out the conclusions of the
lengthy discussion on the Federal Reserve view of the monetary
mechanism. We contended that there is no single, unified, consistent
view that can be characterized as the Federal Reserve view. But
we noted also that many of the statements made are consistent with
7

Koosa, op. cit., p. 105.




FREE RESERVE CONCEPT

{>3

our interpretation that free reserves are regarded by the Federal
Reserve as a major element in the monetary mechanism.
Consideration of the details of policy operations in this section
provide strong confirmation of the importance of free reserves in the
Federal Reserve's policy operations. Although there may be many
different and changing interpretations of the usefulness of particular
indicators within the FOMC, there is a single Manager. We have
now found that there is a strong indication that his actions are more
than a reflection of the policy views of the Committee. Often the
reverse is true; the Manager permits or encourages changes in the
level of free reserves, and the Committee often ratifies his prior
decision.
However useful it may be for the Committee to change the guides
to desired policy or to refer to a variety of targets and indicators in
their discussion, it is extremely difficult for the Manager to continually readjust his operations to a new target or indicator every few
weeks. Market events do not have the same interpretation in terms
of all of the criteria that are proposed. In self-defense the Manager
must choose and retain a particular criterion or set of criteria by which
he can judge the effect of his operations. He then translates the
vague and often changing suggestions of the Committee into the
framework that is useful to him m his operations. Moreover, it is the
Manager who furnishes the principal information on the "tone" of
the money market to the Committee. It is not difficult to understand, therefore, why the Committee is often in the position in which
n can do little more than ratify his prior judgments. Since the
FOMC as a group does not have any explicit criteria or analytic
frame for independent judgment, it is not clear that they are aware
of the autonomy exercised oy the Manager.8
Thus it occurs that the Manager is relatively free to make adjustments in policy in advance of the FOMC meetings or to avoid adjustments [judged to be desirable by the Committee. Since both Managers
nave testified about the importance of free reserves in their view of the
mechanism, little doubt remains about the importance of free reserves in the actual operations of the System.®
Summary

We have found that the moving average of free reserves is often
an adequate guide to System policy and that movements in the level
of free reserves are often made in anticipation of decisions of the
FOMC.10 But at other times, we have found that changes in desired
policy are not noticeably reflected in the level of free reserves. Small
changes in emphasis; e.g., modifications characterized by the statement "resolve doubts on the side of ease," are rarely observable in
the moving average. We must conclude that either (1) these small
J ?he procedures described do help to explain why the Committee does not M b t o e the Manager for
Mistakes o ri nat l t e r h i g j u d | ; i r i e n t about the state of the market." See the testimony of the former Manager
i «his R ?
"Review of the Annual Report * * V* op. cit., pp. 31-32.
'See the statement by Robert Rouse, ibid., p. 34, discussing the relation of free reserves to the rate of
JJange o f bank credit and summing the importance of free reserves in his understanding of the monetary
^chanism. See also the statemeS of Robert Stone, -Federal Reserve Open Market Operations.to119€2£
?P. eit. As we have Indicated above, Stone's discussion of policy changes is largely in terms of thejfree reserve concept
10
a
We have found only one writer who has noted that the level of
rXmb?i
£>bcy decisions. Cf. Daniel S. Ahearn. "Federal Reserve Policy Reappraised"^(^ew Yorfe. Columbia
University Press 1963) D. 218-N.6. It is not clear from Aheam's discussion whether he is referring to disa r m Policy only o ^ o ^ n m^ket ^ l i c y as well. In any case, he offers very little evidence m support of




50

FREE RESERVE

CONCEPT

changes are eliminated along with many other random variations
when the moving average is constructed, (2) that free reserves are not
used to effect these small changes; e.g., only the distribution of reserves is affected, or (3) that the Manager ignores some of the instructions given by the Committee. In view of large week-to-week
-changes in unadjusted free reserves and the relatively wide range o
f
values for the moving average that characterizes a particular policy,
it is likely that the first or second interpretation is correct.
More troublesome for our interpretation is the absence of changes
in the level of free reserves when somewhat larger changes in policy
are directed. For example, we noted that the level of free reserves
shows little change toward ease either before or after the meetings o
f
March 25 and April 15, 1958, or March 27, 1962. Our index gives
each of these statements a value of
On the basis of our analysis,
we cannot reach a firm conclusion about these counter-examples.
Nevertheless, the detailed examination of System policy seems to
indicate that the timing of many of the major changes in monetary
policy actions in the post-Accord period can be observed using the
moving average of free reserves. Moreover, the evidence suggests
that the FOMC moves rather quickly at times of change in the direction of economic activity. In current academic parlance, the "inside
lag" in monetary policy appears to be extremely short. On two o
f
the three occasions when the economy turned toward recession, the
T
"recognition lag" was negative; w hen the economy turned toward
recovery, the "recognition lag" was longer, averaging 3 to 4 months.
But this longer lag is most likely a reflection of the desire on the part
of the FOMC to avoid stifling an incipient recovery. The "action
lag"—the length of time that it takes for decisions to be carried outis at most zero and often negative, if we choose the moving a v e r a g e o
f
free reserves as the measure of System policy. We conclude, therefore, that the System's post-Accord record of recognizing and acting
at turning points can only be regarded as splendid.
The size of the response by the System and the speed with which
the change in free reserves affects the rate of change in money and
credit have not yet been considered. Recognition and action at
turning points are undoubtedly important. But it is also important
to take action in terms of the best available instruments that an
understanding of the monetary mechanism can provide. We turn,
therefore, to consider the relation of free reserves and the " m o d i f i e d
free reserves mechanism" to the rate of change of money and credit.




SECTION

3—THK

R E L A T I O N OF F R E E R E S E R V E S
M O N E Y AND C R E D I T

TO

CHANGES

IN

Analysis of the effectiveness of monetary policy can be divided into
three subtopics: (1) The timing of the recognition of the need for policy
changes and the decision to act; (2) the choice of appropriate action to
influence promptly the stock of money and credit; and (3) the effect
of changes in money and credit ou the "pace of economic activity. We
have seen that the post-Accord decisions of the FOMC have been
timed coinmcudably and that action has often been taken by tho
manager in advance of the Committee's decision to act. But effective
monetary policy depends also on the extent to which the action taken
by the Federal Reserve is capable of altering the stock of money in the
appropriate direction. This iu turn depends on an understanding of
the mechanism relating policy actions to the stock of money; i.e., on
the use of an appropriate concept as a measure and indicator of monetary policy.
Much of the discussion in the earlier sections attempted to describe
the prevailing Federal Reserve view of the monetary mechanism.
Evidence was presented to support the contention that the dominant
notiou guiding the Federal Reserve in the post-Accord period has been
centered on the role of free reserves as a measure of the impact of
policy. Examination of the details of policy operations, supported
by statements of the managers and other oflicials, indicates that free
reserves are used as a target and signal of policy as well. We have
called this view of the monetary mechanism the "modified free
reserves doctrine," since at times there is clear recognition in official
statements that the effect of a particular level or range of free reserves
is modified by prevailing interest rates and by the distribution of free
reserves among classes of banks. In this chapter evidence bearing on
the relation of free reserves and the modified free reserves mechanism
to the stocks of money and credit is presented.
Two principal sources of evidence will be considered. The first,
based on the findings of Meigs* study,1 considers the effect of interest
rates and open market operations on the demand for free reserves and
the rate of change of deposits. These findings are concerned with
the evidence for a position that some Federal Reserve spokesmen seem
to have accepted—that the supply or level of free reserves must be
considered in relation to the bands' desired holdings. A second set
of findings was developed as a part of the present study. These
seek to isolate the effects of interest rates and the level and distribution of free reserves on three measures of money and credit—
demand deposits plus currency held by the public, total deposits plus
currency held by the public, and total loans and investments of
member banks.
The evidence presented confirms in large part some assertions made
earlier in this study that the Federal Reserve has failed to develop a
1

A. J. Meigs, "Free Reserves and the Money Supply" (Chicago: University of Chicufto Press, IMS).




51

52

l-RKK RESERVE CONCEPT

useful working knowledge of llic monetary mechanism. After 50
years, the System's degree of control over changes in the stock of
money or credit, judged in terms of the modified free reserves doctrine,
is so pitifullv small that retention of free reserves as an important
measure or indicator of policy appears to be completely unwarranted.
T H E DEMAND FOR F R E E R E S E R V E S IVY RANKS

Metes' book began as a study of the factors determining the money
fiupphv* But he did not develop an explanation of the money supply.
Tho closest lie came was to consider some possible determinants of
the rate of change of demand deposits. llis preliminary results led
liiin to investigate the relation between the ratio of free reserves to
deposits, interest rates, and the rate of change of unborrowed reserves.3
In the process, he broke important new ground in our understanding
of the monetary process by developing and testing a theory of the
demand for free reserves by banks.
Meigs' results 011 this topic? can be summarized succinctly. Three
of his findings arc of particular interest. First, he found that open
market operations had only a small positive direct effect on the demand for free reserves. 'Phis effect is observable within the month
in which the operation is conducted. Kxperiments with lags suggested that tho direct- response of desired free reserves to open market
operations was substantially stronger in the month following the operation, but- the effect- remained relatively small withal. Second, the
yield on Treasury bills appeared to have a much more important iniluence on the desired level of free reserves than the direct effect of
open market- operations. An increase in Treasury bill yields was
accompanied by a reduction in the desired level of free reserves. The
effect, of lagged Treasury bill yields on the demand for free reserves
was in the same direction as current yields, but in general the lagged
relation was no stronger. Third. Meigs was able to explain by far
the larger part of the month-to-month change in the ratio of free
reserves to deposits by the proximate determinants of the demand
for and supply of free reserves.
Wc have seen that the Rieller-Burgess view of policy operations,
from which the simple free reserves notion appears to have emanated,
considered the direct effect of open market operations on reserves
to be largely offset by changes in member bank borrowings. Tt wa?
through such changes in the proportion or volume of borrowed reserves that monetary policy was said to be made effective.
It was previously indicated that during the twenties variations in
member bank borrowing and changes in the adjusted base (dominated
by the gold stock, Treasury currency and the Federal Reserve banks'
portfolio of Government securities) were closely correlated. A
dollar change in the adjusted base was associated on the average
with a dollar change in the opposite direction of member bank borrowing. This pattern vanished in the thirties and has not reappeared.
Even the fifties, though exhibiting substantial variations in the
volume of member bank borrowing, show no significant correlation
between these variations and changes in tho adjusted base. Thus,
open market operations immediately modified the supply of free




FREE RESERVE CONCEPT

{>3

reserves in the postwar period by changing the banks' volume of
reserves and excess reserves.
Meigs1 investigations reveal, on the other hand, a comparatively
sinall direct effect of open-market operations ou the bank's desireil
level of free reserves. This magnitude appears to respond most
decisively to changes in prevailing interest rates. Variations in
credit murket conditions, expressed by a spectrum of interest rates,
induce banks to adjust their reserve position. Open market operations (or other events affecting the magnitude of the adjusted base)
thus immediately create a divcrgcnce between the banks' desired and
actual free reserve position. This divergence triggers a process
involving readjustments in the banks' balance sheets and generates
modifications in both the moncv stock and interest rates. The
variations in interest rates form" an essential part of this process,
induced by the banks' endeavor to adjust their actual reserve position
to their desired position. This endeavor generates losses iu free
reserves, via changes in required reserves and currency flows. These
losses occur in response to deposit liabilities created of destroyed during the process. Desired free reserve positions respond to the changes
in interest rates that accompany the process and also contribute to
the elimination of the difference between actual and desired free
reserves.
Meigs' investigations suggest a view of the monetary process
radically different from the Riefier-Burgess heritage. Our previous
discussion emphasized the peculiar character attributed to member
bunk borrowing and free reserves under the Kiefler-Burgess notions
and the subsequent evolution of the free reserve conception. Free
reserves were typically visualized as a magnitude emerging from a
process imposed on banks und independent, of any choice behavior on
the part of banks. The analysis developed by Meigs, supported by
his statistical results, strongly emphasizes the neglected volitional
aspects of free reserves. Banks are shown to hold free reserves in
response to market conditions.
It follows from Meigs' analvsis and results that banks adjust their
asset portfolios in response to prevailing levels of free reserves relative
to their desired volume office reserves. Some of the answers provided
m the context of the questionnaires published in the appendix reveal
tt
partial acknowledgment of Meigs' and similar results. Such
acknowledgment is a decisive break with past Federal .Reserve conceptions, in particular with the Riefier-Burgess heritage. Numerous
Policy statements and evaluations made in the past, and repeated in
Jhe questionnaire, are inconsistent, with the acknowledgment that
banks modify their desired free reserve position in response to market
conditions.
Meigs' analysis and results also bear significantly on the Federal
Reserve's use of various "liquidity measures." Such measures have
no meaning bv themselves but must be interpreted within an appropriate conception. The Federal Reserve's conceptions bearing on
liquidity" and "liquidity measures," discussed in chapter 11, are inconsistent with the best' validated portions of economic theory and
seriously challenged bv Meigs' results. While there is no need to
T
*peat the previous discussion of the use made by the Federal Reserve
the concept of liquidity, it is worth noting (I) that the evidence
Meigs' study supports the argument made earlier and (2) that the



54

FREE KESERVE

CONCEPT

Federal Reserve's use of "liquidity measures" is an indication of their
failure to fullv understand the meaning of a demand by banks for free
reserves. Despite the important role that has been assigned to the
demand for free reserves in the responses of the presidents and the
governors (see appendix), the evidence suggests that they have not
drawn the logical conclusions with respect to "liquidity" and the
monetary mechanism.
T i l t PERCENTAGE CHANGE IN DEPOSITS RESULTING FROM THE INTERACTION OF THE DEMAND FOR AND SUPPLY OF FREE RESERVES

•\Vo have noted above that Meigs did not fully develop and test a
theory of the relation of free reserves to the money supply. Nevertheless, his findings, about the cfTcct of interest rates, tree reserves, and
open market operations on the monthly change in deposits, provide
important information about the inadequacy of the Federal Reserve's
conception. Some of Meigs' findings arc summarized here:
1. Meigs constructed a comparatively simple explanation of the
monthly percentage change in bank deposits in terms of the interaction of the supply of anddemand for free reserves. He was able to
account for two-thirds of tho variability of the percentage change of
bank deposits for the period 1947-58 in terms of nis framework.4 His
results suggest that one-third of the variations in the rate of change of
demand deposits is outside the control of the Federal Reserve, oince
the currency component of the money supply is excluded from consideration in Meigs' study, wc can reach no" firm conclusions as yet
about the meaning of those findings for the rate of change of the money
supply.
2. Free reserves have a negative effect on tho monthly percent change
in demand deposits in Meigs' formulation. The higher the ratio of
free reserves to demand deposits, the lower the percent change in
deposits, holding interest rates and open market operations unchanged.
This implication was strongly supported by the evidence, a finding that
flatly contradicts the Federal Reserve's interpretation of free reserves
as an indicator of "ease" and "restraint." A rise in free reserves
means an increase in measured excess reserves of the banking svstem or
a fall iu member bank borrowing. Thus a decline in borrowing i*
associated with a reduction in the percentage change in demand
deposits; an increase in member bank borrowing raises the percental
change in demand deposits. Increased member bank borrowing add*
to the total reserves of the banking system and contributes to the
growth of demand deposits. Moreover, a rise in excess reserves
relative to deposits appears to reduce tho percentage iucrcase in
demand deposits. The Federal Reserve's interpretation of free reserves is directly contrary to these findings. The evidence suggests
that their conception is incorrect.
3. The effect of an increase in the percentage of unborrowed
reserves (open market operations,) like the increase in borrowed reserves, has an expansive effect on the rate at which demand deposit*
increase. But the direct effect of open market operations is relatively
small in each of Meigs' equations. We have noted above that spokesmen for the federal Reserve often refer to a six or seven dollar expan4

Ibid., p.

eqimtiou T 7. To obtain this explanatory power. Moles allowed for * «*ii*mtc effect ol &




FREE

RESERVE

CONCEPT

{>3

sion of Hie amount of deposits per dollar of increased reserves.5
Mcigs| results suggest that the appropriate value of the reserve multiplier is smaller; a 1-percent change in unborrowed reserves is
associated with at most a ^-percent change in the rate of change of
demand deposits for given interest rates. "Under the conditions prevailing in recent years, Meigs' results suggest that an open market
operation had a multiple effect 011 the clmnge in demand deposits,
but the multiplier is between 2.5 and 3. not G or 7.
These examples of the evidence available from Meigs' study raise
broader questions about the Federal Reserve's conception of the
monetary process. Meigs' finds that the direct effect of open market
operations on changes in the ratio of free reserves to deposits is much
smaller than the indirect effect through interest- rates. This raises
doubts about the Federal Reserve's rationale for attaching significance
to random and often self-reversing changes in float, Treasurv balances,
etc.
Clearly, more needs to be known about the effect 011 interest rates
of changes in the supply of reserves and the timing and magnitude of
the effect of changes in Treasury bill rates on other interest rates.
Then a firmer conclusion can be drawn as to whether many of tho
''defensive" operations are stabilizing rather than destabilizing. Preliminary evidence presented earlier on the variability of monthly
changes in the supply of money and the stock of bank credit- suggest
that these "defensive" operations may be the source of increased
instability in the stock of money. Additional evidence on this subject
comes from the computation of the monthly changes in the money
supply, currency plus demand deposits. The simple correlation of
the change in one month with the change in the following or preceding
month is negative (—0.20). This suggests the interpretation that an
increase in the money supply this month will more likely than not be
followed by a decrease in tlie next month. A similar result is found
'or monthly changes in bank credit. It is difficult to find any rationale
for such variability. Bv smoothing the extreme variability in bank
reserve positions, "the Federal Reserve seems to contribute to the
variability of money supply and bank credit changes. It is difficult
to judge,"iii the present state of knowledge, the effect of such shortrun variabilitv in the financial variables on the pace of economic
activity. But there has been 110 analysis or supporting evidence
adduced to suggest that the variability contributes to economic
stability.
If variations in float or Treasury balances affect individual banks
adversely, there is no reason why"open market operations must be
used to offset- these temporary disturbances. Banks that arc under
temporary pressure can, if tliev desire, pay a price to obtain reserves
temporarily in the Federal funds market, *If 110 sales of Federal funds
are offered at a price less than or equal to the discount rate, banks
c
an borrow from the Federal Reserve banks, i.e., use the discount or
collateral loan facilities of the Reserve banks. This may require a
change in the Federal Reserves long-cherished notions about- borrowalthough it would seem to be in keeping with the spirit of the
original Federal Reserve Act to permit- banks to borrow for these
fhort-tcrm purposes. In any case, it is difficult to understand why
NoTrRi°flcr'

,<0pen Markct




Operations in Long-Term Securities," Federal Reserve Bulletin, vol. 41,

Ig

FREE RESERVE CONCEPT

the Federal Reserve is willing to supply reserves through open market
operations that it might he unwilling to supply through the discount
window, since borrowed reserves and unborrowed affect the money
supplv ill a similar wav. We will return to a discussion of alternative
means of eliminating" tho undesirable effects of transitory reserve
changes in the concluding chapter.
MOSEY, CREDIT, AND THE MODIFIED FREE

RESERVES

DOCTRINE

By asking and answering a number of questions and suggesting some
interpretations of the evidence, we can assess the usefulness of free
reserves as an indicator of a monetary position and the relevance of
the modified free reserves doctrine. The questions that will be asked
concentrate primarily on the relation between interest rates, free
resrves, the modified "free reserves mechanism, money, and bank credit.
The answers to the questions are given in terms of computed coefficients of determination. This computation permits us to measure
relations between magnitudes and to express the percentage or fraction
of yearly or monthly variation in one magnitude that accompanies,
i.e., occurs jointly with, another magnitude.
it should be noted that the coefficient of determination g i v e s no
indication of cause or effect. If we find that the coefficient of determination between interest, rates and free reserves is relatively large,
for example, we cannot judge from this observation alone whether
changes in free reserves caused changes in interest rates or whether
interest rates caused changes in free reserves. All that we can infer
from the given observation is the extent to which the two moved together or in opposite directions, perhaps under the influence of &
common causal factor. But the observation of such correlations yields
support for conceptions asserting a systematic (or causal) association
between the magnitudes under consideration. Further observations
may be gathered in order to discriminate more sharply between different- conceptions compatible with the given gross correlation.
The coefficient of determination between monthly free reserves and
the yield o n Treasury bills is 0.42 for the 170 months from N o v e m b e r
1948 to December 10(32. Furthermore, the correlation between the
two is negative. These observations support- the contention that desired free reserves rapidly adjust to the prevailing volume of free reserves and depend on market, conditions. The observed correlation
indicates that, under this conception, 42 percent of the variation observed in free reserves during the period are explainable bv the concomitant variation in interest rates. Since the correlation is nesrative,
we infer, in addition, that large positive levels of free r e s e r v e s are
associated with low yields on Treasury bills.
These data suggest that there is* a closer relation between the
level of free reserves and the Treasury bill yield than between the
level of free reserves and changes in the stoclfof monev or credit that
arc considered below. This is similar to the finding of Mei^s' studv.
But, wo should note, in pacing, that the addition oftho vears J959-©
to the data that Meigs used has increased the correlation between free
reserves and Treasury bill yields; A possible explanation for the




FREE RESERVE CONCEPT

57

increased coefficient of determination is that the relation may be
stronger in times of high interest rates than in periods of low interest
rates.
To £o beyond the simple facts provided by the coefficients of determination requires detailed tests of alternative theories about the
monetary mechanism. This is not our concern in the present chapter.
Here we are interested only in the presentation and interpretation of
some elementary facts bearing on the adequacy of prevalent Federal
Reserve notions (theories) about the factors influencing the supply
of money and bank credit.
The measure that we have chosen, the coefficient of determination,
must be between zero and one. The closer that the computed value
of this measure comes to one, the closer the correspondence in the
movements of the magnitudes under consideration. Conversely,
when the computed coefficient, of determination is near zero, there "is
no evidence of any systematic relation between the magnitudes, and
there is no indication" of any influence running from one to the other.
Thus, if we find that the coefficient of variation between free reserves
and total reserves is very close to 1, it might make very little difference which of these measures was used in the explanation of the
monetary mechanism or interpretation of the events in the money
market. Of course, the two might be closely related while neither
has a close relation to changes in money and credit.
In addition to the questions and answers about the relation of one
monetary factor to another, we will use the coefficient of determination to evaluate the combined effects of a series of separate factors
operating jointly on a particular measure of monetary change. For
example, we can consider the combined effects on the change in bank
credit of (1) the level of free reserves, (2) the distribution of free reserves among classes of banks, and (15) interest rates. The extent to
which the positions or levels of these three factors contribute jointly
to an explanation of the change in bank credit will be measured by
the computed coefficient of determination. The following questions
and answers present the evidence on the Federal Reserve's conception
of the monetary mechanism in terms of these computed coefficients.
Quest iou 1. Is the monthly change in the supply of money or
in member bank loans and"investments closely related to the
monthly average level of free reserves?
Answer. No. There is almost no evidence of a relation between the level of free reserves and changes in money and credit.
The coefficients of determination in table V-l suggest that the
level of free reserves has almost no influence on the change: in
money supplv or member bank credit and vice versa. There is a
slight" indication that the relation is stronger during periods of
expansion than during periods of contraction in the economy.
This is particularly true for the monthly change in demand deposits plus currency.




58

FREE

RESERVE

TABLE . V-L.—TMeasures of the relationship
and monthly changes in money supply
November 1948 to December 1962 1

CONCEPT
between monthly average
and member bank credit

Item

Monthly change in currency plus demand deposits—
Monthly change in currency plus total deposits
Monthly change in total loans and investments in member
banks.
.
.
-

Months of
contraction
in the
economy *

free reserve*
outsiandity,

Months of
expansion
in the
economy *

All months

0.02
0

0.14
.06

0.04
.04

.01

.02

.03

i Figures in the table are coefficients of determination.
i 46 months from National Bureau peaks to troughs during the period November 1948 to February 1»1.
* 108 months from National Bureau troughs to peaks during the period October 1949 through May I960,

Question 2. Is the relatively poor explanation evidenced by
the values shown in table V - l largely a reflection of extremely
short-run money market variations? Wouldn't an average of
free reserves taken over a period longer than a month show a
substantially stronger relation?
Answer. No. We recall that the 3-week moving average of
free reserves was a useful indicator of changes in the direction
of Federal Open Market Committee policy. It was sufficiently
smooth for major swings in desired policy to be revealed, as
indicated in the previous chapter. The monthly averages of
free reserves are less erratic than the weekly moving averages.
They also mark out clearly the changes in desired Federal Reserve
policy. The more appropriate interpretation seems to be that
free reserve levels tell almost nothing about the changes in
money and bank credit.
Annual moving averages of free reserves were used to provide more
evidence on this point. Much of the short-run variation that remained
in the monthly data was eliminated by the use of annual data.
Seasonal variations in money and credit were eliminated by comparing free reserves to annual percentage rates of change from one
month to the corresponding month in the following year. The
explanatory power of the relation between free reserves and changes
in money and credit improved very little when annual data were
used.
Allowing for the distribution of free reserves among classes of
banks and allowing for the role of Treasury bill yields as additional
explanatory factors did improve the annual results quite a bit. But
the best explanation was for the annual percentage rate of change of
money, demand deposits plus currency held by the public, and not
for total bank credit that is so much emphasized in Federal Reserve
discussions of the monetary mechanism. These results are shown in
table V-2.




F R E E RESERVE CONCEPT

{>3

TABLE V - 2 . — M e a s u r e s of the relationship
between annual moving averages of free
reserves and annual percentage rales of change in money and credit, November 1948
to December 1962
Coefficient of determination
Item

Using (1) and
Using annual distribution Using (2) and
average of
Treasury
of free refree reserves serves by
bill yields
bank class

Annual percent rate of change in demand deposits plus currency
Annual percent rate of chanpo of total deposits plus currency..
Annual percent rate of change of total loans and investments of
member banks .

(3)

(2)

a>
0.08
.03

0.33
.24

0.45
.42

.08

.29

.34

Question 3. Do measures of the distribution of free reserves
and interest rates have a similar effect in improving the explanatory power for monthly changes in money and bank credit?
Answer. No. For the 170 monthly changes in bank credit and
money between November 1948 and December 1962, there is
almost no improvement when we allow for the distribution of
free reserves between classes of banks, the yield on Treasury bills,
and/or the ratio of Treasure bill yields to the prevailing rediscount
rate. This is shown in table V-3.
TABLE V-3.—Measures of the relationship
between monthly changes in money and
credit and the level and distribution
of free reserves and interest rates,
November
1948 to December
1962
Coefficients of determination

Monthly change in currency plus demand deposits-.*
{jonth y change in currency plus total deposits
•Monthly change in total loans and investments of member
banks
_

Using (1)
and
Treasury
bill rates

Using (1)
and the ratio
of Treasury
bill rates to
rediscount
rates

(1)

Item

Using free
reserves and
their distribution

(2)

(3)

0.04
.07

0.05
.10

0.05
.07

.05

.06

.06

-

the data the last column of
provide
Suestion 4. Dochanges ininmoney and credit? table V-3explanaence about the modified free reserves doctrine as an
tion of monthly
Answer. Yes. The evidence in column (3) suggests that the
modified free reserves doctrine provides almost no explanation of
monthly changes in money or credit.
Question 5. Does the relatively weak relation for both monthly
changes in money and monthly changes in bank credit suggest
that while neither is closely related to free reserves, changes in
money and credit are closely related to each other?
Answer. No. Changes in money and credit have a monthly
coefficient of determination of 0.01 when money is measured as de


Ig

FREE

RESERVE CONCEPT

mand deposits plus currency. These data indicate that there is
no support whatsoever for ihe position that monetary expansion
and credit expansion are one and the same, as the Federal Reserve
spokesmen' have maintained.
Question 0. Is the relatively weak support for the modified
free reserves doctrine and the relation of free reserves to changes
in money and credit partly explained by the relatively low
iuterest rates of the pre-Accord period and the early post-Accord
period?
Answer. Yes, there is some support for this interpretation
from the evidence on changes in the stock of money. For changes
in the stock of credit, the evidence suggests the opposite conclusion as shown in table V-4. But the explanatory power remains
small. Moreover, there is very little evidence of anv steady
direction of change in either the stock of money or credit. The
correlation between the successive changes in the money supply
in adjacent months is small and negative. The same is true for
monthly changes in bank credit as noted earlier. Despite the
fact that both of these changes have been positive on the average
over the 14-year period, an increase in money or. credit in one
month is not a reliable indication that there will be an increase
in the following.month.
TaiUiK V- -4.— -Measures of the relationship
between monthly changes 1 i n money and
credit and the level and distribution
of free reserves and' interest rates, S postwar
periods
Coefficients of determination
TVIIIK free
reserves and
their distribution

utfng (1>
a:.d
Treasury
bill ratas

U«:nF (I).,
and the rati J
of Treaty
bill rates to
rediscount
rate

(1)

Item

(2)

(3)

November IMS to July 1953
Monthly changes In currency plus demand deposits
Monthly chanpe In currency plus Total deposits
Monthly change in total loans and Investments of member
banks

0.05
.10
.24

"

0.0*
.10

0.0*
".TS

.21

.35

Jul* 1053 to July 1837
Monthly changes in currency plus demand deposits
Monthly cluiii;;e in currency plu* total deposits
Monthlytiltangttill total loans and iiivfe&nonts of member"
baufes

0.14
.09

0.20
.09

0.1J
.05

.05

.00

.25

July 1957 to May 1900
Monthly chaises in currency phis demand deposits
Monthly chan-c; in currency plu« total deposiis
Monthly change in total loans and investments of inemlier
banks
.
i Tsinp monthly percentage rates of cliauee i:ivos very similar results.




0.32
.3d

0.32
.37

0.41
.37

.10

.11

.10

FREE

RESEllVE

61

CONCEPT

It should be noted also that the monthly variation in free reserves
lias been smaller iu the past few years than in the years of the early,
fifties when swings in free reserves of more than SoOO million were
more frequent-. Accompanying the greater stability in the level of
free reserves, there has been a growth in the number of banks participating actively in the Federal funds market. These changes in
market- conditions, or arrangements, probably help to account for
the steady improvement in the relation between free reserves and
changes in money supply and in the influence of the modified free reserves mechanism on changes in the money stock. But, of course
this explanation cannot account for the very poor relation between
free reserves and bank credit that we observe from July 1957 to
May 1960.
Question 7. Is there any additional evidence that the facLors
included as part of the "modified free reserves doctrine have,
different effects in periods of relatively low interest rates and
economic activity and periods of relatively high interest rates
and economic activity?
Answer. Yes; the evidence in table V-5 suggests this conclur
sion for the changes in currency plus demand deposits. Inters
est rates were higher on the average in 10S months from trough
to peak than in the 40 months from peak to trough. The modified free reserve doctrine is a better explanation of changes in the
stock of money in months from trough to peak.. We have
noted earlier that- the level of free reserves was more closely
related to interest rates in recent months when interest rates
have been higher on the avorage.
IABI.E V - 5 . — M e a s u r e s of the relationship
between monthly changes 1 in
and credit and the Icirl and distribution of free reserves and interest rates in
of expanding and contracting economic
artivily

money
months

Cwlllcicnts of determination
Ufihk free
reserves und
thi-ir distribution

loin* (1)
imd
Troiu-ury
hill rales

Using O)
and the rr»tio
cf Treasury
bill raies to
rediscount
rate

(I)

I torn

(2)

(3)

40 months from pc-Jik to trough
0.00
.12

0.00
.17

0.10
.IS

.02

\t? if y c t a W In currency pliu demand de;>OAits
l i S . 1y
in currency plus total ticpojii?
monthly chance In total loans and imeatmisita of member
"i»iiks

.02

.OS

JOS months from trough to peak
Month v L'liaiiQt |„ currency p!us demand deposits.
vJSH! y clwiine in ctirrciiry plus total dupable-—.
..
b t i chaiiRc In total loans and investments of member
1

VsiiiR monthly iwrccntoRe rates of change gives very similar results.




0.14 I
.07|

0.LM
.10

0.15
.07

.04 |

.05

.04

Ig

FREE RESERVE CONCEPT

This circumstance might explain the relative improvement in the
relation between the level and distribution of free reserves and changes
in money supply during the 108 months from peak to trough (col. 1).
Data for the three periods shown in table V-4 suggest that the
association between changes in the money stock and free reserves is
stronger in high interest rate periods than in periods of comparatively
low interest rates. The average level of interest rates rises from
period to period as we move down the table and the explanatory
power measured in columns 1, 2, and 3 rises also. But the improvement in explanatory power holds only for changes in the stock of
money; it does not appear when we consider changes in member
bank credit.
Table V-5 again indicates that the improvement in the explanatory
power of the modified free reserves doctrine in periods of comparatively high interest rates applies to changes in the stock of money
but not to changes in the stock of credit. Reasons for the Federal
Reserves attachment to, and emphasis on, credit cannot be obtained
from the evidence that we have examined. None of the evidence
gives any indication that there is a reliable association between free
reserves, or the modified free reserves mechanism, and the change in
bank credit.
Question 8. Do the findings suggest that further modifications
and tests of the Federal Reserve explanation of the monetary
mechanism, centered on the free reserves doctrine, would be
useful?
Answer. We can never be certain that additional modifications
would not improve the explanatory power. One can only try.
But the accumulated evidence falsifies so many of the features
of the Federal Reserve conception that time can more usefully
be spent developing an alternative explanation that avoids these
errors.
Question 9. Do the poor results obtained necessarily indicate
the irrelevance of a conception inherited from the RieflerBurgess tradition and centered on the free reserve mechanism?
Answer. No. Such results might be obtained under a coherently formulated free reserve conception if the adjustment
of the banks' portfolio of earning assets is very rapid. In
particular, little association between the variations in money
stock or bank portfolios and free reserves would remain in monthly
data, if the adjustment process is concentrated within a month.
Under such circumstances the central relation of the Federal
Reserve's conception, associating "credit-expansion" and free
reserves, would operate significantly only in the shortest run
and could not be detected in the monthly data used here. But
m this case some other problems arise.
If this formulation of the free reserve conception is advanced,
free reserves must be abandoned as an indicator of a monetary
situation and as a target of monetary policy. The interaction
of the relation centered on free reserves with other pertinent
relations, constituting the structure of a rapidly adjusting




F R E E RESERVE CONCEPT

{>3

process, implies that free reserves cannot be interpreted as
an indicator or used as a target by the Federal Reserve authorities. Furthermore, other implications of the reformulated
free reserve conception can be shown to be seriously incompatible
with validated portions of economic analysis. Thus either the
central relation of the Federal Reserve conception is seriously
invalidated by the observations presented, or a reformulated
conception, compatible with the data in tables V-l to V-5, is
incompatible with the burden placed on free reserves as an
indicator and target by the Federal Reserve authorities.®
SUMMARY AND CONCLUSION

The four chapters discussing and analyzing the Federal Reserve
conception of the monetary mechanism are now complete. In our
Sgeneral overall view and summary in chapter II, we noted some of the
ogical and factual errors that mar the Federal Reserve's understanding of the mechanism. And we looked at some of the reasons for these
errors—particularly their extremely short-run orientation, their
concern with daily or weekly defensive operations, their tendency to
view the banking system as analogous to a single bank. We noted
repeatedly that if the Federal Reserve had looked seriously at the
evidence, they would not persist in repeating these errors.
Sections 1 and 2 attempted to demonstrate that particular factors,
summarized in the modified free reserve doctrine, have dominated
the prevailing Federal Reserve view. Inconsistencies, qualifications,
and modifications that appear from time to time made this task
laborious and difficult. But the evidence from the "Record of
Policy Actions" and the repeated references to the same factors
suggested quite strongly that the modified free reserve doctrine comes
reasonably close to a statement of some of their dominant views.
In arriving at this judgment, we noted that the absence of a clear,
generally accepted statement of the mechanism, the concern with
hourly, daily, and weekly money market changes, and the emphasis
on "defensive" operations contributed to a substantial grant of
authority to the Manager of the System Open Market Account.
In this section, some of the evidence on the Federal Reserve's
conception of the mechanism controlling the stock of money has been
assessed. We found that the relation between Federal Reserve policy
and the change in money and credit is quite poor judging from the
monthly and annual data used in our tests.
A new question arises, therefore: Does it really matter very much
What the Federal Reserve does? By far the larger part of the monthly
and annual changes in money ana credit seem to be outside their
control, judged by their conception of the mechanism.
The proviso of the last sentence is the crux of the matter. If the
Federal Reserve's conception of the monetary process, centered on
the position of free reserves, were the only admissible view, we would
nave to concede that monetary policy is little more than a futile exerstatements In the above answer are based on an underlying analysis to be published in another
context; "Evolving Federal Reserve Conceptions Concerning the Money Supply Process.




6 4

FREE

RESERVE

CONCEPT

cise. But alternative conceptions of the monetary mechanism have
been formulated, and it is essential to consider them before accepting
such a negative conclusion. I t should be noted, therefore, that our
analysis does not suggest the futility of monetary policy but only
supports our contention about the failure of the Federal Reserve to
develop a coherent, validated conception.
The failure of the Federal Reserve to develop a useful conception
of the monetary mechanism does not mean that one cannot be developed. The following chapter will present an alternative view of the
mechanism that places emphasis on substantially different factors
and suggests a much more reliable association between the money
supply and policy actions.




O


Federal Reserve Bank of St. Louis, One Federal Reserve Bank Plaza, St. Louis, MO 63102