Full text of New England Economic Review : November/December 1969

View original document

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

November/December

FEDERAL RESERVE BANK OF BOSTON

NEW
ENGLAND
ECONOMIC
REVIEW
Free Reserves in Monetary Policy Formulation
This article presents the results of a statistical analysis of the impact of free
reserves on the growth of bank credit. The study concludes that when free
reserves are placed in a proper framework that includes the strength of credit
demands, they are a significant factor in explaining changes in member bank
loans and investments.
SUPPLEMENT:

The U. S. Balance of Payments Deficit and the
State of International Reserves
Remarks by Norman S. Fieleke, Asst. Vice President and Economist, at Area
Bank Conferences conducted by the Federal Reserve Bank of Boston in September 1969.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

969

Free Reserves in
Monetary Policy Formulation
by Robert E. Knight and Paul S. Anderson

POLICY has traditionally been
criticized for its alleged unfortunate effects
during certain periods; to cite examples, it has
been blamed for bringing on recessions at some
times and for fostering inflation at other times.
In recent years, however, new types of criticism
have appeared which are narrower in scope and
rather technical. These involve monetary magnitudes, or indicators which critics assert the
Federal Reserve either should, or should not, use
in policy formulation. Net free reserves, which
equal excess reserves of member banks less their
discounting from the Federal Reserve, are one
indicator whose use has been severely criticized
and are the subject of this article.

ONETARY

Another indicator which has become an even
more discussed issue in recent years is the
growth rate of the money stock. The free reserves and money supply growth issues are related since one important question is: What, if
any, impact does the level of free reserves have on
the growth of monetary aggregates?
For many years free reserves have been used
This article is based on a Ph.D. thesis, Federal
Reserve System Policies and Their Effects on the Banking System, by Robert E. Knight at Harvard University, 1968. The thesis was written with financial aid
from this Bank and will soon be available on request
to the Research Department of the Bank, 30 Pearl
Street, Boston, Massachusetts, 02106.
Dr. Knight is now a Financial Economist at the
Federal Reserve Bank of Kansas City.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

as one of the key indicators of the degree of
monetary tightness or ease experienced by the
banking system. Negative free, or net borrowed,
reserves were believed to be contractionary since
banks were expected to repay their borrowings
from the Federal Reserve promptly. Conversely,
high levels of free reserves have been considered
a stimulus to bank credit growth because banks
are not under pressure to repay borrowing to the
Federal Reserve and can use unneeded excess
reserves to expand loans and investments. A
typical view about the role of free reserves in the
monetary process is illustrated in the following
quotation:
In general, the net reserve position of banks is
an important gauge of the pressures on bank reserves. When net free reserves rise, the result is an
increased marginal availability of reserves, which
the banking system can readily use to expand
credit. But when member bank borrowings grow
relative to excess reserves, credit expansion comes
under restraint. In this process individual banks
find extra reserves more difficult and expensive to
obtain, and they come under increasing pressure
to repay advances from the Federal Reserve. 1

Even if the use of free reserves as a gauge of reserve availability were considered acceptable, they
could still be used improperly if the demand for ·
loanable funds were not fully considered in policy
making. Thus, a certain level, say $500 million,
1 Board of Governors, Purposes and Functions of the
Federal Reserve System (5th ed; Washington: Federal
Reserve System, 1963) p. 224.

November/ December 1969

of free reserves may have at times in Federal
Reserve history been viewed as a sufficient contribution of policy toward expansion during a
recession period, when a much higher level of
free reserves might have been required to foster
desired growth in monetary aggregates (such as
bank credit).
An important current issue, however, is
whether free reserves have a value even as a
gauge of reserve-supply conditions. This has
been questioned both inside and outside the
Federal Reserve System. To cite a recent article
in the Federal Reserve Bank of St. Louis Review,
. . . We [do not] consider free reserves to have any
causal impact on bank behavior. The evidence
marshalled against free reserves as an important
causal link in the monetary process is impressive. 2
If such views as this are correct, then free reserves are not a good measure of supply conditions in the money market and their usefulness
in monetary policy formulation may be seriously
questioned.
This article presents the results of a statistical
analysis of the impact of free reserves on the
growth of bank credit. The conclusions are that
free reserves are a significant factor in explaining
changes in member bank loans and investments
when they are placed into a proper framework
which includes the strength of credit demands.
Critics who have obtained contrary statistical
results have not explicitly recognized the demand
side of the monetary process in their analyses.

Outline of Monetary Policy
Operations
To place the free reserves controversy into
context, we must begin with an overall view of
monetary policy operations. The Federal Reserve has three traditional tools of monetary
2 Michael

W. Keran and Christopher T. Babb, "An Explanation of Federal Reserve Actions (1933-1968)," Federal
Reserve Bank of St. Louis Review, July 1969, p. 9.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

policy, namely, open market operations, the
discount rate, and reserve requirements. In
addition, it has in recent years used its authority
to set maximum rates on savings and other time
deposits as a tool to affect the ability of banks to
secure funds. The chief tool and only one used
daily is open market operations. In these operations, the Federal Reserve buys Government
securities when it wants to expand the supply
and availability of money and credit and sells
when it wants to diminish such supply and
availability. These sales and purchases are made
to influence the monetary environment of the
economy in such a way as to achieve certain desirable results with respect to prices, employment
and growth.
In actual day-to-day operations, however, it
is obviously impossible to determine or judge
the correct magnitude of sales or purchases of
Government securities by observing the impact
on these ultimate goals of prices, employment
and growth. This impact is so long delayed estimates of relevant lags generally range from
6 months to several years - that it is simply of
no help whatsoever as to what sale/purchase
action should be undertaken on any given day.
In this situation, as a substitute for its longterm goals, the Federal Reserve has chosen intermediate goals which it wants to achieve. These
intermediate goals are usually growth rates for
monetary magnitudes like bank credit or the
money stock. To serve their function well, these
intermediate goals should meet two requirements: 1) they should have a predictable effect
on the economy with respect to the ultimate
goals of prices, employment and growth and
2) they should respond within a reasonable time
to Federal Reserve purchases and sales of
Government securities. Both these requirements
have proven difficult to meet satisfactorily.
The money supply has traditionally been the
intermediate goal of monetary policy. However,
the experience of the Great Depression after the
3

New England Economic Review

Banking Holiday of 1933 when the money supply
rose fairly rapidly but economic activity expanded very slowly cast much doubt on the
ability of money to influence prices, growth, and
employment. Until well after World War II
money remained out of favor as a goal and monetary policy in general tended to be ignored as an
influence on the economy. Fiscal policy was
believed to be much more potent. Nevertheless,
when inflation became a problem during the
1950's the role assigned monetary policy again
gained attention and importance. But how to
make it most effective has continued to be
disputed.
Intermediate goals of policy which have been
suggested in the more recent postwar period include interest rates and bank credit as well as the
money stock. All of these indicators have the
major shortcoming that none has a proven relationship to movements or trends in total
economic activity. Interest rates have fared
particularly poorly in empirical analyses; for
example, surveys of changes in business investment in 1966 as a result of the (then) very high
interest rates showed that the impact was quite
small. 3
Bank credit and the money stock have had a
somewhat better relationship with economic
activity. Research carried out at the Federal
Reserve Bank of St. Louis4 showed the closest
association between these variables and GNP.
Analysts generally have not fully accepted
these St. Louis results, however. They have
3Crockett, Jean, Friend, Irwin, and Shavell, Henry, "The
Impact of Monetary Stringency on Business Investment,"
Survey of Current Business, August 1967, p. 10 ff.
4 Andersen, L. C., and Jordan, J. L., "Monetary and Fiscal
Actions: A Test of the Relative Importance in Economic
Stabilization," Review, Federal Reserve Bank of St. Louis,
November 1968, pp. 11-21. Only the regressions between the
money stock and GNP are shown but unpublished results
using bank credit had an even higher R2 than the .60 or so
obtained for using the money stock as the main explanatory
variable.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

questions regarding the chain of causation from
money to GNP, the lack of proportionality between the growth rates of these magnitudes and
GNP- the money stock since World War II
has risen much more slowly than has GNP, a
sharp turnabout from the relation of the previous
70 years - and whether these magnitudes would
have the same relation to GNP if they were controlled within narrow limits rather than allowed
to vary rather freely as they have been in the past.
Nevertheless, while the relationship between
intermediate monetary goals and ultimate economic goals has not been perfect, yet it has been
good enough to warrant their continued use. In
any case, there probably is no alternative but to
observe relevant intermediate variables and to
rely on subjective judgment at those times when
these various measures do not agree reasonably
well in their movements.

Short-run Associations Between
Policy and Indicators
The second requirement for a well-functioning
intermediate indicator is to respond predictably
to Federal Reserve policy actions, particularly
to sales and purchases of Government securities.
Put the other way, the day-to-day operating aim
of the Open Market Desk which does the selling
and buying of Government securities is to conduct affairs in such a way that the chosen indicators of policy behave as desired. Since numerous factors in addition to open market operations interact to determine the magnitude and
use of bank reserves, open market operations
cannot directly determine movements in the level
of intermediate variables such as bank credit and
money supply. Although the Federal Reserve
generally attempts to offset reserve changes
caused by nonmanaged factors, the willingness
of banks to expand credit and of bank customers
to borrow also influences movements in the intermediate targets. Open market operations are
intended to induce banks and the public to be-

No11ember/ December 1969

have in certain ways, but they cannot force any
specific behavior, at least in the short run.
Many slippages can occur between open
market sales and purchases and the behavior of
banks and the public because the chain of causation is so long as the following listing shows:
1. Open market sales and purchases by the
Desk determine the security holdings in the
Federal Reserve portfolio.
2. These securities, plus member bank borrowings, float, gold stock, Treasury currency,
and some other minor items govern the
amount of total reserves supplied to commercial banks. Only about two-thirds of
total reserves supplied are provided by open
market operations.
3. Of total reserves supplied, more than half is
absorbed by currency in circulation, and a
small additional amount is absorbed by
minor factors leaving only about a third
available for member bank reserves.
4. Member bank reserves
three parts according to
demand deposits, 2) to
posits, and 3) unused or

can be split into
use: 1) to support
support time deexcess.

This listing shows that the connection between
open market operations and a monetary indicator such as bank credit is very loose. It is not
surprising that there is almost no correlation
between day-to-day or week-to-week changes in
open market operations and bank credit. 5
While open market operations exert little precise control on total bank credit in the short run,
these operations can be used to control total
member bank reserves within fairly narrow
limits. This control is accomplished by offsetting
variations in "external" factors like currency in
5See, for example, Sherman J. Maisel, "Controlling Monetary Aggregates," in Controlling Monetary Aggregates, proceedings ofa Monetary Conference held in June 1969 (Boston:
Federal Reserve Bank of Boston, October 1969).

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

circulation which affect reserves. If, for example,
some external factor increases reserves above the
current goal level, the Desk can immediately sell
securities and reduce reserves back to the goal
level.
Although total reserves are controllable within
fairly narrow limits even on a daily basis, there is
still substantial slippage between them and commercial bank credit or the money supply. This
is shown in Chart 1 where these series are
plotted on a weekly basis for 1969, together with
the bank credit proxy which equals total member
bank deposits less interbank deposits and cash
items in process of collection. It is quite apparent
that movements in the money stock and bank
credit do not follow movements in reserves at all
closely on a weekly or even monthly basis. In
almost every case, the largest movements in total
reserves are not paralleled by similar movements
in the other series.

Free Reserves Target
In view of the, at best, loose connection in the
short run between open market transactions and
indicators of policy, the Federal Reserve System
has considerable latitude for the day-to-day conduct of these transactions. Under these circumstances, it has developed what has been termed a
"money market strategy" in which it instructs
the manager of the Open Market Account to
maintain a certain degree of ease or firmness in
the money market. The level of free reserves is
an important measure of this degree of ease or
firmness but in recent years it has no longer been
considered the sole measure as before. Shortterm interest rates, particularly those on Federal
Funds and Treasury bills, have been increasingly
used together with the level of free reserves to
measure money market conditions.
Even though free reserves do not occupy their
former unique position as a measure, they nevertheless still reflect quite accurately the desired
5

New England Economic Review

money market stance of the Open Market Committee.6 Therefore, in this article, the level of
free reserves can be viewed as a representation
of this money market strategy. Beyond their
value as a measure of policy intent, the level of
free ( or net borrowed) reserves is an indicator of
conditions in the credit markets, particularly the
ability and willingness of banks to extend loans.
While free reserves are equal to excess reserves
less member bank discounting from the Federal
Reserve Banks, the fluctuation in their level is
accounted for mainly by discounting since excess reserves have tended to remain rather stable
cyclically though having a declining trend. Thus
cyclical changes in the level of free reserves essentially reflect changes in the level of discounting.
When discounting is at a low level, banks are
being supplied their required reserves by the open
market operations of the Desk. In addition, they
usually are able to acquire Treasury Bills and
other securities which can be liquidated later if
necessary to meet loan demand. But when banks
are discounting heavily, their reserve needs are
not being met by open market operations and
they may not have salable assets which can be
liquidated in order to obtain lendable funds.
Banks are expected to use other methods of
meeting customers' credit needs and their own
reserve needs and to turn to discounting only as
the last resort.
Discounting places banks under constraint.
First, they are expected to repay their borrowing
within a short period so they must manage their
operations in order to obtain alternative funds
or else reduce their requirements by reducing
deposits. Second, discounting brings on the
"surveillance" of the Federal Reserve Bank
which involves a review of the bank's reserve
6See Keran and Babb, op. cit., pp. 8-9. The authors present
an index which measures the policy stance called for in the
directives of the Federal Open Market Committee. This
index correlates highly with movements in the actual level of
free reserves.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

management. Banks find this "looking over the
shoulder" uncomfortable and embarrassing to
some extent so they strive to limit their discounting as much as possible. Thus the level of discounting, and of free reserves, is a direct indication of the reserve situation and lending stance of
member banks.
The monetary inducement or pressure strategy
entailed in a free reserves target implies that reserves should act as a short-term shock-absorber
for temporary fluctuations in demands for reserves. Thus as demands for bank credit fluctuate around a growth trend; operations with
a free reserves target accommodate such fluctuations since reserves supplied will follow variations in demands for reserves. By contrast,
operations with some fixed quantity target like
total reserves would not allow such variations
around the allowed growth trend, consequently
the reserve supply would be deficient one period
and "excessive" the next. Not only would such
a rigid reserve supply course cause wide fluctuations in short-term money market interest rates,
but it would also frustrate some borrowers who
were trying to secure financing when the quota
was being absorbed by others. Even though
these disappointed borrowers would be accommodated in the following period, they would
have been penalized for no good purpose.
Assuming the same total volume of financing
under alternative targets, the use of the free reserves target would seem to improve efficiency
by causing no "unnecessary" disruption.

Criticisms of the Free Reserves Target
The money market strategy in general as well
as the free reserves target has been criticized on
two main grounds. The first is that concentration on smoothing hourly and daily fluctuations
in the money market is believed to divert attention from the longer-run goal of providing a
favorable monetary environment for the economy•

November/December 1969

While it is conceivable that anxiety over these
hourly or daily fluctuations could overshadow
concern with broader problems, it is unlikely in
actual practice. The main reason is that the
policy makers in the Federal Open Market Committee pay little attention to the smoothing operations of the Open Market Desk. A rather
clearcut division of function has developed between the Committee and the Open Market
Desk. Typically, the Committee votes on a directive specifying the degree of inducement or
pressure which should be applied,7 and then the
Desk conducts its smoothing, or defensive,
operations around, or centering on, the target
levels of free reserves and other measures implied
in this directive. It is immaterial for the smoothing operations whether the free reserves target is,
say, zero or minus $200 million of free reserves.
In fact, since defensive activities typically require more sale or purchase activity than does
implementation of policy such as reducing free
reserves from zero to minus $200 million, policy
implementation is carried out easily and conveniently within the general context of smoothing.
Another criticism of the free reserves target is
its ineffectiveness in changing the economic
climate. That is, variations in free reserves do
not cause corresponding changes in such policy
indicators as bank credit or the money supply.
The explanation for this presumed ineffectiveness
stems from the idea that banks wish to hold a
certain level of free reserves and that this desired
level of free reserves moves in the same direction
as the actual level of free reserves, possibly nullifying the expansionary or constrictive impact of
changes in the actual level of free reserves.
The reasoning behind the perverse impact of
changes in desired free reserves hinges on possible reactions of banks to variations in interest
71n recent years a "proviso'.' clause has usually been added
which instructs the desk to vary the degree of restraint if results, generally as measured by the bank credit proxy, appear
to be deviating significantly from projections.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

and discount rates. When money market interest rates rise relative to the discount rate,
as usually occurs during business expansions,
banks become less reluctant to borrow. Consequently, borrowings will rise, reducing the
level of free reserves unless offsetting action is
taken by the Federal Reserve. Since banks have
chosen to go deeper into debt, we can say that
the desired level of free reserves has declined. If
the Federal Reserve is aiming at a lower level of
free reserves at the same time, the actual level
will decline right along with the banks' desired
level and the Federal Reserve could mistakenly
interpret its action as being contractionary.
According to this line of reasoning, the crucial
determinant of monetary and bank credit expansion is the differential between actual and
desired free reserves. If actual free reserves exceed the desired, bank efforts to reduce free reserves will result in credit and deposit growth.
However, if actual free reserves are less than desired reserves, banks will restrain credit, as they
attempt to equate the two. Under any circumstances, according to critics of the free reserve
concept, monetary policies must make allowance
for shifts in desired free reserves if effective control over the growth of bank credit and the
money supply is to be maintained.

Testing the Free Reserves Target
Whether free reserves are an effective short-run
target can be tested statistically. One possibility
is to correlate the level of free reserves with
changes in bank loans and investments. (Bank
credit, rather than the money supply, has been
selected as the dependent variable because the
minutes of the Federal Open Market Committee
imply that the Committee, at least over most of
the period since the 1950's, has shown greater
concern over movements in bank credit than in
the money supply. Changes in the money supply
are always reported and considered at these
7

New England Economic Review

meetings, but they do not appear to have often
formed the basis for policy prescriptions.) However, in this test, the free reserves target fares
rather poorly; it explains or determines only
about a tenth of the changes in bank loans and
investments. 8 This result can hardly be considered satisfactory and appears to support the
criticisms of the free reserves target.
Another test correlating free reserves with the
Treasury bill rate indicates that the two variables
are quite closely related. This suggests that the
level of free reserves banks are willing to hold
may fluctuate with rates on earning assets and
that the levels of free reserves observed may
largely be determined by adjustments of banks
to bring the free reserve portion of their portfolio to the desired level. 9
These tests, however, ignore the strength of
demand for bank credit. Since free reserves are
used as a measure of inducement or pressure,
they must be seen as working with, or against, a
certain strength of demand for credit. When
demands are weak, a given level of free reserves
will naturally induce less credit growth than
when demands are strong. If this were not the
case, the Federal Reserve could fix free reserves
at the level allowing the desired rate of growth
in bank credit and then leave them unchanged
indefinitely, assuming it wanted constant growth
in credit.
Therefore, if statistical analysis is used to
evaluate the impact of free reserves on changes
in some intermediate indicator, a measure of the
strength of demand must be included in the
analysis. In this study the variable which has
been chosen to represent the strength of demands
8 See, for example, A. James Meigs, Free Reserves and the
Money Supply (Chicago, 1962), and Karl Brunner and Allan
H. Meltzer, An Alternative Approach to the Monetary Mechanism, Subcommittee on Domestic Finance, Committee on
Banking and Currency, House of Representatives (Washington: U . S. Government Printing Office, 1964).

9 /bid.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

for bank credit is the volume of external funds
obtained by all nonfinancial corporations.lo

Results of the Tests
The test results are described in detail in the
Technical Note on page 15 and can briefly be
summarized here. First, when the credit demands variable is included in a regression equation with free reserves, both are highly significant
as explanatory variables in determining changes
in member bank loans and investments (statistical results are shown in Table 1 on page 15). Thus
there can be little question that the level of free
reserves does have a strong influence on growth
in bank credit. The test results also show that
this influence cannot be accurately measured
alone, but must be analyzed in the proper context; in particular, the strength of credit demands
must be taken into account.
The statistical results indicate that a certain
rise in the average quarterly level of free reserves
is associated with a quarterly rise of almost four
times as much in member bank credit, assuming
the strength of credit demands remains unchanged. In like manner, a quarterly rise of a
given volume in total external credit obtained
by nonfinancial corporations is associated with
a bank credit rise of one-fourth as much, assuming the level of free reserves remains unchanged.
In a real-world situation, both these factors
would be operating, of course, and the results
would reflect the net impact of their joint influence. For example, credit demands may be
tending to raise member bank credit by, say, $4
billion per quarter at the same time that the
Federal Reserve System is attempting to slow
down the growth of bank credit by reducing free
reserves to a minus $300 million. The average
net result expected from this combination would
be growth of not quite $3 billion in member bank
10For a detailed explanation of this choice as a proxy to
represent the strength of demand see page 15.

November/December 1969
Chart 1

WEEKLY MOVEMENTS OF TOT Al RESERVES & VARIOUS MONET ARY INDICATORS
Index,

Jan.1969=100

103
102
101
100
99
98
97

TOTAL RESERVES

96
103

102
101

LOANS & INVESTMENTS AT
LARGE COMMERCIAL BANKS

..,.,"~
\_"\

100

.,,-'"

\/ '-,

'--'

,,-'

'v"v"'~

98
97
96

TOTAL

103
102
101
BANK CREDIT PROXY

100

\ ,..,

,,
98

....,,,

97
96
Jan .

Feb.

Mar .

Apr.

May

Jun .
1969

Jul.

Aug .

Sept.

Oct.

Nov .

Dec.

Note - All series shown a re based on d ata used prior to revisions caused by redefinition of Euro-dollar checks as
deposits subject to reserves.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

New England Economic Review

credit. A casual observer might wonder why
bank credit grew at all with such a tighter policy,
not realizing that without the tightening, bank
credit would have grown $1 billion more per
quarter.
The statistical results are illustrated in Charts
2 and 3. Shown in Chart 2 are the actual
quarterly changes in member bank loans and
investments since 1954 and the changes which
would be predicted from the joint impact of the
level of free reserves and the strength of credit
demands. Actual growth in bank credit fluctuates tremendously from quarter to quarter, yet
the predicted changes correspond surprisingly
closely. The period 1954-1966 was used as the
base for deriving the predicting equation but it
has continued to perform about as well in the
period since, thus supporting the reliability of
the equation. The predictive accuracy since 1966
is especially notable in view of the extreme
changes that have taken place in policy and
banking operations.
Shown in Chart 3 are the estimated contributions of free reserves and credit demands to bank
credit growth. These estimations are based on
regression equation (1) in the Technical Note
which is,
Quarterly Growth in Member Bank Loans and
Investments = 3.68 Level of Free Reserves
+ 0.23 Quarterly Credit Demands + Seasonal
Factors.
According to Chart 3, credit demands were
quite stable from 1954 to 1964, and their impact
accounted for member bank credit growth of
$2-$3 billion per quarter, or an annual rate of
around 5 percent. Since 1964, however, credit
demands have skyrocketed so that they have
contributed $5-$7 billion a quarter to bank credit
growth, or an annual rate approaching 10 percent.
The contribution of free reserves to growth in
member bank loans and investments averaged

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

out to about zero from 1954 to 1960, being positive in the recession years of 1954 and 1958 and
negative during the other years of the period.
Thus, over the 1954-1960 period as a whole,
monetary policy as measured by free reserves
was a stabilizing influence on credit growth although criticism has been directed at certain
episodes, particularly the excessive ease in 1954,
the late move toward ease in 1957, and the excessive tightness in 1959. The net result in terms
of credit growth over this period probably can
be considered satisfactory, however, since growth
averaged close to the noninflationary rate of 4-5
percent annually.
From 1960 to 1964, free reserves were adding
substantially to credit growth at a time when
credit demand contribution alone was achieving
a 4-5 percent growth rate. This added growth,
however, seemed necessary to stimulate an
underemployed economy. After 1964 free reserves did turn negative but the restraint was
wholly inadequate to keep credit growth within
noninflationary bounds. The positive contribution of free reserves to credit growth in 1967
seems especially inappropriate in light of the
tremendous strength of credit demands that
year.
Since the statistical analysis supports the
effectiveness of free reserve levels in influencing
extensions of bank credit, it likewise supports the
reluctance theory of discounting. The coefficient
for free reserves is positive and highly significant
in all regressions, indicating that reserve additions obtained through Federal Reserve Open
Market operations or reductions in reserve requirements, other things equal, lead to expansion
of bank loans and investments while increased
discounting results in credit restraint. Banks
evidently are reluctant to borrow from the
Federal Reserve and experience pressure to repay
quickly. For rising levels of discounting to
facilitate credit expansion, the coefficient of free
reserves should be negative.

November/ Decemher 1969
Chart 2
ACTUAL AND PREDICTED CHANGES IN MEMBER BANK LOANS AND INVESTMENTS
1954 • 1969

BIiiion• of Dollar•

12
11

10
9

ACTUAL CHANGES

e
IS
4

PREDICTED CHANGES

-1

19154

191515

19!57

19158

19!59

19150

19151

Test Results for Other Variables
When interest rate variables, like changes in
bill rates or in the difference between the bill rate
and the discount rate, are added to regression
equations containing free reserves and credit
demands, they add almost nothing to the explanatory power of the equations. Furthermore
their coefficients are negative, and significantly
so in the case of the bill rate. Admittedly, the
difference between the bill rate and the discount
rate was generally not large during the period
analyzed and it is possible that bigger differences
would have yielded different results.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

19152

19153

19154

191515

19157

1988

19159

Nevertheless, the results suggest that the significance of short-term interest and discount
rates in affecting bank behavior has been greatly
overemphasized and that as long as the Fed
keeps the differential between the two rates relatively small, it need not worry about changes in
the desired level of free reserves due to this
factor. The demand for bank loans clearly seems
to be a more important determinant of growth in
bank credit than are interest rate differentials.
The final group of statistical tests involved a
comparison of the impact of alternative policy
targets, namely, nonborrowed reserves, total reserves, and the monetary base, which is total re-

New England Economic Review
Chart 3
ESTIMATED CONTRIBUTIONS OF CREDIT DEMANDS AND
FREE RESERVES TO BANK CREDIT GROWTH 1954 . 1969

BIiiion• of Dollar•
10

CREDIT DEMANDS .

(Alwaya Poaltlva)

FREE RESERVES ~

(May Be Negative)

NET OF BOTH

1964

1966

195e

1967

1968

1969

19eo

19e1

19e2

1953

19e4

19e5

19ee

19e7

19ee

1959

Explanatory Note -The chart is based on Equation (1) in the Technical Note. The free reserves effect equals 3.68
times their average quarterly level, while the credit demands effect equals 0.23 times total net external funds obtained
by non-financial corporations. When free reserves are positive, their effect is added to the credit demands effect, but
when they are negative (net borrowed), their effect is subtracted.

serves plus currency outside banks. These alternative targets do not generally test much better
than do free reserves despite the fact that they
have a direct relation to member bank credit.
This direct relation arises from the fact that required reserves are determined by member bank
deposits and the latter are directly related
through the balance sheet to member bank loans
and investments.

Some Implications for Policy

growth in bank credit. But even if this were
granted, it must be recognized that major
problems remain in policy formulation. Free
reserves operate only on the supply side of the
supply-demand interaction that determines the
level of monetary aggregates like bank credit and
the money supply. The level of free reserves can
be fixed within a rather narrow range, but demands for credit vary widely and subsequently
wide fluctuations occur in the growth of bank
credit even with a given level of free reserves.

The main conclusion of this article is that free
reserve levels do have a significant impact on

Assuming that an intermediate objective like
growth in bank credit or money were adopted

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

November/ December 1969

by the Open Market Committee, its primary
problem would be the varying and often unpredictable strength of demands for bank credit
and money. For example, suppose at a certain
point in time, the desired annual growth rate of
bank credit, say 4 percent, had been satisfactorily achieved for several months with a
target level of min us $100 million of free reserves.
Then, if in a subsequent month bank credit begins rising at an 8 percent annual rate, a decision
has to be made whether to continue the previous
target level of free reserves. The difficulty is that
the 8 percent growth may be a random fluctuation which would be offset by zero growth in the
following month or it may reflect a stronger demand for bank credit than had been anticipated
which would result in another month of 8 percent
growth if no changes in policy were made.
The basic problem here then is judging the
true underlying strength of demand for bank
credit from monthly observations of actual rates
of credit growth. One approach would be to
treat the situation as a quality control problem
in production where no action is taken until
there is a certain "run" of observations outside
specified control limits. Thus the Open Market
Committee might take corrective action when
actual monetary growth deviates from the desired trend by a certain amount. Such an approach could evolve into a relatively mechanical
and routine operating procedure once some
growth rate for a monetary magnitude were
adopted.
The choice of definite limits for growth of a
intermediate goal like bank credit or the money
supply would probably be a much more important decision than the mode of day-to-day
operations. If such an intermediate goal were
followed with determination, wider month-tomonth swings in money market conditions would
occur than under the money market strategy
which is followed with a relatively fixed free reserve target. For example, to correct for higher-

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

than-anticipated demands for bank credit which
resulted in a growth rate of, say, 8, rather than a
desired 4, percent would require a decrease in the
level of free reserves of around $800 million.
(According to Equation (1) in the Technical
Note each change in the level of free reserves of
$100 million implies a corresponding quarterly
change in bank credit of $368 million which
represents an annual rate of change of about
one-half of one percent.) Thus much more
vigorous monetary policy changes would be
necessary under a definite growth rate goal than
have occurred in the past.
Financial observors have often set up the free
reserves target and intermediate goals such as a
certain growth rate of monetary magnitudes as
opposing modes of policy formulation. But as
we have seen, there need be no conflict since the
free reserves target can be a useful tool in reaching a growth rate objective. Nevertheless, free
reserves are not an absolute necessity in the context of Open Market operations. A variety of
short-term target variables are feasible.
While the free reserves target is useful in implementing a policy focusing on growth rate
objectives, it is readily adaptable to other intermediate goals as well. Not all policy formulators
and advisors accept monetary and credit growth
rates as sole, or even primary, intermediate objectives; it still remains to be proven that monetary policy acts only through specific monetary
quantities or aggregates. It is possible, for example, that general conditions in the money and
credit markets affect anticipations and business
planning quite apart from developments within
monetary aggregates. The free reserves target is
well suited to accommodate such policy assumptions since it both measures sensitively credit
availability and quite accurately reflects movements in the cost of credit.
A fact which tends to be overlooked in disputes over the free reserves target is that the
13

New England Economic Review

magnitude called free reserves, whether used as a
target or entirely ignored, occupies a central .
place in the chain of operations involved in
monetary policy. Insofar as monetary policy
entails action to influence the size of financial
quantities such as member bank reserves, bank
loans and investments, and the money supply
or total credit flows, it must first affect free
reserves.
As the System increases (or decreases) total
Federal Reserve credit, which is essentially the
only quantity action it can take, this immediately
and necessarily tends to move free reserves in the
same direction. Similarly, if reserve requirements
are changed, free reserves are concurrently
changed. If the Federal Reserve were to use interest rates as its goal, its operations to influence
rates (other than the largely symbolic effect of
the discount rate) would again directly affect
free reserves. Thus even though free reserves
could be totally ignored in carrying out policy,

fluctuations in their level would continue as in
the past as banks adjusted to Federal Reserve
policies and to changes in the demand for credit.

Summary
The use of the level of free reserves as a shortterm target for monetary policy operations has
been criticized in recent years. The two chief
bases of criticism are that the use of such a shortrun target diverts attention from the longer-run
influence of policy and it is not functional, meaning it does not affect more basic monetary variables in desired ways.
When the free reserves target is considered in
its proper context, however, it appears to be a
reasonable target for guiding daily and weekly
operations. In statistical tests, if the strength of
demand for bank credit is included in regression
equations, the level of free reserves is a significant
variable explaining changes in member bank
loans and investments.

TECHNICAL NOTE:

Formal Analytical Model
The model underlying the empirical analysis in this article
is an excess demand equation for free reserves. The growth in
bank loans and investments is assumed to be functionally
related to the difference between the actual level of free reserves and the level desired by the banking system. Consequently:
~L + ~I = S + X 1 (FR - FR*)
where:
~L

~I = Quarterly Change in Member
Bank Loans and Investments
S = Seasonal Variables
FR = Actual Level of Free
Reserves

FR•

Desired Level of Free
Reserves

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

If the actual level of free reserves exceeds the desired, banks
will grant loans and purchase securities in an effort to reduce
free reserves. Conversely, if the actual level of free reserves is
less than the desired, efforts by banks to increase the actual
level would result in contraction in loans and investments.
Some observers have concluded that the relatively high
correlation between the Treasury bill rate and the level of free
reserves indicates that the level of free reserves desired by
banks is highly sensitive to interest rates and that the observed
level is likely to approximate the desired level. But if monetary
policy in conjunction with the demand for credit determines
both free reserves and interest rates, a high correlation between free reserves and interest rates could merely mean that
policies influence more than one variable, but would imply
nothing about the relationship between actual and desired
free reserves. In the model presented here, the actual level of
free reserves is taken as a marginal measure of Federal Reserve
policies.

November/ December 1969

The level of free (or net borrowed) reserves desired by the
banking system is a function of numerous variables, but the
most important would appear to be the demand for loans and
interest and discount rates. As the demand for loans increases,
banks are increasingly willing to risk the possibility of being
forced to discount and the desired level of free reserves should
decline, producing an increase in bank credit. On the other
hand, the higher the level of money market interest rates relative to the discount rate, the greater could be the incentive of
banks to secure additional loanable funds by borrowing from
the Federal Reserve. A relative rise in the level of money
market rates, therefore, may be associated with a decline in
the desired level of free reserves and lead to an expansion of
bank loans and investments.
A measure of the strength of demand for loans is a crucial
variable in this model. Unfortunately, it is difficult to measure
demand for bank credit or for any other good because demand is not observed by itself. What can be measured is the
amount "sold" and this is the amount demanded at a certain
"price." If the price were lower, more would be demanded
and vice versa.
The best that can be done in measuring demand for bank
credit is to obtain some indicator which tends to vary with
the strength of the demand. One possibility is total credit obtained in the Nation from all sources by all borrowers Federal, State and local governments, business, and consumers, the latter including both consumer credit and mortgage credit. But the difficulty with this measure is that the
credit obtained by some of these groups is not closely associated with the demand for bank credit which fluctuates
quite closely with the cycle. Credit obtained by the Federal
government and mortgage borrowers, for example, tends to
be contracyclical, generally reaching highs in the latter half
of recessions.

The most pronounced cyclical pattern among these groups
is shown by businesses so their total borrowing should serve
well as an indicator of variations in bank credit demand over
the cycle. Also the demands of large businesses for external
funds are quite inelastic so that the total business funds obtained are not usually influenced greatly by variations in
interest cost. Furthermore, although business firms typically
obtain less than a quarter of their external funds from banks,
they own the largest deposits and are the banks' prime customers. Thus banks would be quite anxious to satisfy their
demands for credit if at all possible. Despite these favorable
factors, the total of external funds obtained by businesses is
not, of course, a foolproof measure of the strength of demands
for bank credit. It is used here simply as a reasonable indicator
and with further study a better measure conceivably could be
derived.

Results of the Tests
When total credit obtained by corporate non-financial
businesses is inserted into the regression equation with the
level of free reserves, results were obtained of which equation
(1) in Table 1 is typical. The coefficient of free reserves is
highly significant and explained variance, R 2, is up around
two-thirds. Thus, when the strength of credit demands as
measured by total external funds obtained by businesses is
taken into consideration, the tests show that free reserves
have a potent impact on changes in bank credit. That is, with
a given intensity of demand for bank credit, the actual amount
of credit extended by banks is significantly affected by the
level of free reserves. These results strongly support the functional usefulness of a free reserves target.
The coefficient of 3.68 for free reserves signifies that member bank loans and investments will rise $300-$400 million

Table 1

REGRESSION EQUATION ESTIMATING QUARTERLY CHANGES
IN MEMBER BANK LOANS AND INVESTMENTS
Coefficients
Free
Reserves

Credit
Demands

(1)

3.68
(5.9)

0.23
(6.0)

(2)

3.27
(5.1)

0.25
(6.5)

(3)

3.15
(4.2)

0.25
(6.2)

Equation

NOTE:

Change in
Treasury
Bill Rate

Treasury Bill
Rate Min us
Discount Rate

- 0.81
(2.0)
- 0.96
(1.3)

DurbinWatson
Statistic

.67

1.91

.70

1.92

.69

1.93

Quarterly seasonal variables are not shown ab'ove; t-values are in parentheses.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

New England Economic Review
per quarter for each additional $100 million in the level of
free reserves, assuming no change in the demand for credit.
Similarly, the coefficient of 0.23 for credit demands means
that, assuming no change in free reserves, on the average,
bank loans and investments will rise about one-fourth as
much as total credit obtained by businesses. 1
'All regression results reported in this article are based on quarterly
data from 1954 through 1966. The coefficients obtained for regressions
run for somewhat shorter time periods are about the same, suggesting
that the estimated coefficients are relatively stable.
One way of validating a model is to test its performance in predicting
future values of the dependent variable. This has been done for 11
quarters following the fourth quarter of 1966. The predictions for the
first six quarters are very accurate and the cumulative error in the
prediction is less than $140 million. During the third and fourth
quarters of 1968, however, the model predicts considerably less growth
in bank credit than actually occurred. During this period money market interest rates declined relative to the rates banks were paying on
certificates of deposit and the volume of CD's grew very rapidly. In the
three quarters of 1969 the accuracy of the predictions again improves,
but on balance the model predicts more growth in bank credit during
these quarters than actually occurred. This divergence is partially
attributable to the large runoff of CD's which banks have experienced.
Despite the errors in some quarters the model performs very well.
For the period since 1966 as a whole, the predicted growth in loans and
investments differs from the actual growth by less than $60 millions.
This accuracy is remarkable in view of the sharp dislocations in credit
flows which occurred during these 11 quarters.

Influence of Interest Rates
Other equations in the table show the impact of the Treasury bill rate and the difference between the bill rate and the
discount rate. As contrasted with results obtained by critics
of the free reserves target, here both have coefficients which
are either statistically insignificant or have the wrong sign.
As stated earlier, critics said that changes in the bill rate could
totally offset the influence of free reserves, so that, for example, rises in the bill rate would lead to increases in bank
credit even if free reserves declined to negative levels.2
Since the addition of the credit demand measure causes the
interest rate variables to become insignificant or negative as
contrasted with the critics' regression results, the interest
rate in the critics' equations evidently is a proxy for the
strength of credit demands. This suggests that the critics'
equations are misspecified.

Alternative Policy Targets
Critics of the free reserves target have suggested alternate
policy targets which they claim are more effective for monetary stabilizfltion purposes. The most frequently recommended alternatives are non borrowed reserves, total reserves,
ZSee for example, A. James Meigs, op. cit. and Karl Brunner and
Allan H. Meltzer, op. cit.

Table 2

REGRESSION EQUATIONS FOR ALTERNATIVE POLICY TARGETS TO INFLUENCE
QUARTERLY CHANGES IN MEMBER BANK LOANS AND INVESTMENTS
1954-1966
Equation Number

Explanatory Variables

Coefficients & t-values

(1)

Free Reserves
Credit Demands

3.68 (5.9)
0.23 (6.0)

.67

(4)

Nonborrowed Reserves

4.43

(4.2)

.51

(5)

Nonborrowed Reserves
Credit Demands

4.75 (5.2)
0.15 (4.0)

.64

(6)

Nonborrowed Reserves
Credit Demands
Free Reserves

2.82 (3.0)
0.21 (3.5)
2.59 (3.8)

.73

(7)

Total Reserves

8.73

(7.5)

.69

(8)

Total Reserves
Free Reserves
Credit Demands

6.33 (5.1)
1.70 (2.7)
0.16 (4.7)

.79

(9)

Monetary Base

5.23

(9.7)

.78

(10)

Monetary Base
Free Reserves
Credit Demands

4.32 (6.5)
1.95 (3.7)
0.04 (0.9)

.83

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

November/December 1969
and the monetary base. No definitive answer can be given as
to the desirability or effectiveness of these alternatives as
compared to a free reserves target but statistically and theoretically, free reserves seem to do at least as well as their
corn peti tors.
Shown in Table 2 are the comparative statistical results of
using free reserves and the other suggested ta rgets. Analysis
of these results entails comparison of the resulting R 2 , or
percentage of variation in loans and investments "explained"
by the several equations, and diagnosis of the factors causing
a higher R 2 in equations (6) through (10) than in the free
reserves equation ( 1).
The free reserves credit demand equation ( 1) yields an R 2 of
.67, which is higher than that of equation (4) with non borrowed reserves as the explanatory variable or that of equation
(5) where credit demands are joined with nonborrowed reserves. The higher R 2 for equation (1) suggests that free reserves are rather effective in inducing or restraining credit
growth as the case may be. Nonborrowed reserves comprise
about 97-98 percent of total reserves on average (discounts
account for the remaining 2-3 percent). Since total reserves
must grow with member bank loans and investments (when
adjustrnen ts are made for decreases in reserve requirernen ts),
there is a built-in, mechanical connection between nonborrowed (as well as total) reserves and bank credit. (In theoretical terms, required reserves make up all but a small fraction
of total reserves and are, therefore, very highly correlated
with nonborrowed and total reserves. To the extent that the
Federal Reserve has followed a money market, or free reserve, strategy, it has supplied reserves as banks expand loans
and investments. Thus, the relationship between changes in
bank credit and an aggregate reserve is very close.) There is
no similar necessary connection between free reserves and
bank credit so the R 2 of equation (1) is impressive.
The same observation about a built-in connection can be
made concerning equation (7) where total reserves are the explanatory variable. While its R 2 at .69 is a little higher than
that of equation (1), the edge is certainly minor and probably
is largely due to the built-in factor. As equation (8) shows, the
addition of free reserves and credit demands to total reserves
improves the R 2 substantially. These equations suggest that
if the Federal Reserve were to seek to regulate movements in
total reserves, its control over the growth of member bank

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

credit would be about as close as with a free reserve target.
However, the extent to which this would result in wider and
perhaps potentially disrupting fluctuations in interest rates
and other money market conditions cannot be known with
certainty.
The remaining equations in Table 2, (9) and (10), have
higher R 2 than equation (1). The monetary base of these
equations equals total reserves plus currency in the hands of
the public. Thus the higher R 2 of equation (9) than of (7) is
due to the inclusion of currency. This raises an analytical
problem: how does currency In the hands of the public affect
member bank loans and investments?
The Federal Reserve accommodates the public's demand
for currency and the net impact of an increased demand is a
rise in its holdings of Government securities. Usually the increased currency is obtained initially from commercial banks
by a liquidation of a demand deposit. The banks in turn replenish their vault cash by selling Government securities to
the Federal Reserve. The Federal Reserve may then provide
enough additional reserves to enable commercial banks to
restore the decline in the level of bank credit caused by the
sale of securities to replenish vault cash.
It is difficult to understand why such a rise in currency holdings should lead to an increase in bank credit, but that is
what the higher R2 for equations (9) and (10) than for equation (1) seem to indicate. The most likely explanation is that
changes in currency holdings for some reason reflect changes
in general demands for credit. This is supported by a comparison of equations (8) and (10). The monetary base in (10)
differs from total reserves in (8) essentially by the amount
of currency holdings of the public but when it is substituted
for total reserves in the equation, the chief result is that the
coefficient of credit demands is cut to one-fourth its previous
magnitude and becomes statistically insignificant.
Whether currency in the monetary base is meaningful or
not seems, however, to be irrelevant so far as its serving as a
monetary policy target. A variable used to carry out policy
should be directly influenced by open market sales and purchases. Currency in the hands of the public is not such a
variable. The public controls the amount of currency it wants
to hold and open market operations have no discernible impact on this decision. To this extent the monetary base seems
not to be a good operational policy target.

New England Economic Review

Remarks by NORMAN S. FIELEKE, Asst. Vice President
and Economist of the Federal Reserve Bank
of Boston, at Area Bank Conferences
in September 1969.*

The U.S. Balance-of-Payments Deficit
and the
State of International Reserves
a number of years the world has had to
wrestle with two closely related problems in
the field of international finance, namely, the
U.S. balance-of-payments deficit and the state of
international reserves. These remarks briefly discuss a few salient aspects of these two problems
and suggest the nature of the relationship between the problems.

The U. S. Balance-of-Payments
Deficit
To start with the U. S. balance of payments, it
continues to be a matter for concern, although
some of the data are rather confusing. Indeed,
as shown by Chart 1, one measure of the balance
indicates that the United States enjoyed a substantial surplus during the first three quarters of
1969, while another measure indicates an awesome deficit. Specifically, the official reserve
*Slightly revised for presentation in this format.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

transactions measure recorded a surplus of $1.5
billion, while the liquidity measure recorded a
deficit of $7 .9 billion.
In order to make a judgment about the
balance-of-payments situation, one must know
what accounts for the difference between these
two measures. The difference is accounted for
largely by Euro-dollar borrowings, or, more
generally, by flows of private short-term foreign
capital into the United States. Short-term private loans by foreigners to U.S. residents are not
counted as balance-of-payments receipts in computing the liquidity balance, but these loans are
counted as receipts in computing the official
reserve transactions balance. Since there has
been a large volume of such loans in the recent
past, the two measures of the balance of payments have diverged widely.
In our opinion, the official reserve transactions
measure is generally the better indicator of the
U. S. balance-of-payments position at any par-

November/ December J9f,9
Chart 1

U.S. BALANCE OF PAYMENTS
Bllllons of Dollars

HS
OFFICIAL RESERVE
TRANSACTIONS BALANCE

~,___
~

\
\

-e

LIQUIDITY BALANCE \

\
-10 ....__ _..__ _"""'-_ _ _ _ _ _ _ _ ____._ _ _....__ _ _ _ _ _ _ _ _ _ _ _ _---'!
1963
1966
1960
1969

*First thre• quarters, preliminary, seasonally adjusted

ticular point in time. After all, to take the recent
experience, if the U. S. position during the first
three quarters of 1969 had been as bad as the
liquidity measure suggests, there should have
been much concern over the strength of the dollar
internationally, and a flight from the dollar might
have developed. Instead, there have not even
been rumors of such a flight, and the dollar
has been a strong currency. On the other hand,
in respect to the future, lending by foreigners to
U. S. residents appears to be diminishing, and
past foreign loans may be repaid rather than
renewed. Therefore, the official reserve transactions measure will probably move into deficit
unless some random occurrence, such as further

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

Soviet aggression in Europe, drives foreign capital into this country again.
In any event, the U. S. balance of payments
has been a matter of concern for some time. One
aspect of the problem has been the deterioration
in the U. S. balance of trade, a matter which is
considered briefly in the following section.

The Disappearance of the
U. S. Trade Surplus
Since 1964 the U. S. trade surplus has dwindled
away. Exports and imports are now running
almost neck and neck, and if one were to deduct
from the export total those exports financed by
19

New England Economic Review

CONSUMER AND WHOLESAL~ PRICE INDEXES FOR SELECTED COUNTRIES
1963 -

June, 1969
Wholesale Price Index

Consumer Price Index
Year

1963

United
States

Canada

United
Kingdom

France

West
Germany

Japan

United
States

100
101

100
103
108
113

100
103
106
109

100
102
106

100
104

100
100
102

1964
1965
1966

103
106

100
102
104
108

1967
1968

109
114

112
117

115
121

112
117

111
113

120

122

128

124

116

June,

1969

1Index

110

111
116
121

106
106

United
Canada Kingdom
1

100
100
102
106

128

109

108
110

133

113

116

France

West
Germany

Japan

100

100
101

100

104
105

101
103

100
104
105

102
103

108
108
117

105
105
106

104
99

105
106

121

113

100

108

for basic materials.

foreign aid, he would discover a commercial
trade deficit. Such a reduction in a country's
trade surplus (or an increase in its trade deficit)
suggests that the country's competitive position
in world markets has weakened.
Further evidence that the U. S. competitive
position in world markets has weakened is offered
by the behavior of the U. S. share of total merchandise exports of non-Communist countries.
Between 1964 and 1968 this U. S. share declined
slightly, from about 17.5 percent to about 16.3
percent. If U. S. merchandise exports are compared with those of just the industrial nonCommunist countries, there appears to be a
somewhat sharper deterioration, as the U. S.
share of these exports has fallen from 24.7 percent in 1964 to 22.2 percent in 1968. Data for
the first half of 1969 do not alter the general
picture.
20

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

Why has the United States experienced this
deterioration in its competitive position in world
markets? Qne possible answer is that the rate of
inflation has been higher here than in other
countries. However, this answer is not altogether correct, as the accompanying table reveals.
These data on price trends suggest that since
1963 the United States has not experienced more
inflation than the typical major advanced
country.1 The U. S. comparative performance
1lt is difficult to select the proper measures of price change
to use in comparing rates of inflation between countries and
in estimating the extent to which exchange rates between
national currencies depart from their equilibrium levels.
If a single measure is to be used, an index which gives appreciable weight to non-traded items is in our opinion generally superior to an index which includes virtually nothing
but internationally traded items. For an introduction to this
subject, see Gottfried Haberler, A Survey of International
Trade Theory (Princeton, N. J.: Princeton University, 1961),
pp. 48-50.

100

NoFemher/ December 1969

over this period would of course have been even
better had not the rate of inflation in this country
accelerated after 1967 ; in the fiscal year ending
June 30, 1968, increased Federal defense spending and delay in raising tax rates contributed to
a huge Federal budget deficit, with seriously inflationary effects. The lowest rate of inflation
among the countries in the table has occurred in
West Germany, a fact which helps to explain
why that country has run a balance-of-payments
surplus in recent years.2
It is clear that price indexes do not tell the whole
story. For example, Canada and the United
Kingdom appear to have experienced more inflation than the United States has since 1963, a
development which tended to improve the U. S.
balance of trade. But Canada and the United
Kingdom, and France as well, have devalued
their currencies at one time or another during
the past 8 years, and when a country devalues
the effect is to make its goods more competitive
in world markets.
To be more specific, there were devaluations
of 14 percent by the United Kingdom in 1967,
11 percent by France this year, and about 9 percent by Canada when it abandoned its floating
rate in 1962. Each of these devaluations, taken
by itself, acted to reduce the U. S. trade surplus
(although the full effects of the French devaluation are still to be experienced). To be sure,
working in the opposite direction were upward
currency revaluations of 5 percent by both West
Germany and the Netherlands in 1961 and of 9
2 Inflation in Japan has been relatively great in terms of the
consumer price index but relatively small in terms of the
wholesale price index. Because Japan's trade balance has
been in substantial surplus recently, the wholesale price
index, which gives more weight to productivity gains in the
production of traded goods, may be the better indicator in
Japan's case. [Cf. Bela Balassa, "The Purchasing-Power
Parity Doctrine: A Reappraisal," The Journal of Political
Economy, LXXII (December, 1964), 593-95.] On the other
hand, the adverse impact on Japan's trade surplus of the
sort of inflation suggested by Japan's consumer price index
could have been at least partly offset by Japan's intensive
use of trade controls.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

percent by West Germany again in October of
this year. However, not only are these revaluations generally smaller than the devaluations;
they also affect a much smaller volume of U. S.
trade. The upward valuations directly affect only
about 10 percent of U. S. trade, but the devaluations, which made U. S. goods less competitive,
directly affect more than a third of U.S. trade. 3
The implication is not that France, Canada,
and the United Kingdom devalued in order to
impair the trading position of the United States.
• On the contrary, countries which devalue usually
are driven to do so by depletion of the reserves
with which they support the exchange value of
their currencies. The point is that their devaluations may help to explain why the U. S. trade
surplus has contracted even though the U. S.
record on inflation seems better than theirs. In
this connection, it should be noted that the option to devalue is not nearly so readily available
to the United States, because many other countries would be strongly affected; quite a few
countries hold a large portion of their international reserves in the form of dollars, a matter
which is elaborated below, and carry on a significant share of their trade with the United States.
Another factor which may help to explain the
deterioration in the U. S. trade surplus is the
automotive agreement which this country concluded with Canada in 1965. This agreement
substantially reduced the barriers to trade between the two countries in automotive products.
The announced goal was to integrate the automotive industries in the two countries, with each
country specializing in the production of those
items it could make the more efficiently. While
such a goal may seem praiseworthy, the impact
on the U. S. surplus in automotive trade with
Canada appears to have been adverse, as that
3"U. S. trade" is here defined as U. S. general imports in
1968. A precise analysis of the effects of exchange rate changes
would also have to consider price elasticities of supply and
demand, among other things.

New England Economic Review
Chart 2

U.S. BALANCES ON GOODS AND SERVICES
AND ON CAPITAL TRANSACTIONS 1
Billions of Dollars
12

NET AUTONOMOUS CAPITAL OUTFLOWS

4
NET EXPORTS OF GOODS & SERVICES
2

01--------.J------- --...&..------------ -.......- - - - - - - - - " -- - - t
1969
1966
1963
1960
ba lances a r e as defined by Fritz Machi u p in his The Transfer Gap of the United States (Pri nceto n, N. J.: Pri nceto n
University, 1968).

1 These

*First half preliminary , seasonally adjusted

surplus has declined from about $583 million in
1964 to about $164 million in 1968.4
One reason for this result is that part of the
agreement serves to restrict the U. S. components which are incorporated into vehicles made
4T hese data exclude tires and tubes. Over the same period,
a U. S. surplus of $776 million on all merchandise trade
(excluding military) with Canada was converted to a deficit
of $453 million, indicating that factors other than the automotive agreement have also contributed to the deterioration
in the U. S. trade balance with Canada. These other factors
may also have affected trade in automotive products.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

in Canada, so that U . S. exports of automotive
products to Canada have been restrained. This
restriction of U. S. exports under the agreement
was intended to be merely a temporary measure
to help the Canadian automotive industry adjust
to the agreement; and the President has suggested that the time has come to remove the restriction, considering how rapidly the Canadian
industry has grown. Of course, even if the restriction on U. S. exports were removed, the
agreement would still work to reduce the U. S.

November/ December 1969

trade surplus if Canada turned out to be more
efficient than the United States in the production
of enough automotive items.
In our view, still another factor which has reduced U. S. net exports is the Federal restraint
on U. S. lending and investing abroad. As
Chart 2 shows, there is a fairly close correlation
between U.S. net exports and the net capital outflows from this country for lending and investing
abroad. When U.S. capital outflows decline, net
exports usually also decline; and in recent years
the sharp reduction in capital outflows caused
largely by the Government control programs has
been paralleled by a reduction in net exports.
One reason for this relationship between
capital outflows and net exports is that foreigners
who receive dollars usually spend some of the
dollars directly on U. S. goods and services.
When their dollar receipts are diminished, their
expenditures on U. S. goods tend to decline as
well.
In addition, when firms (including banks) are
prevented from transferring funds abroad, they
tend to spend or lend the money in this country.
They are not likely to hold the money in idle
balances. The result is a tendency to raise spending and prices in this country, and also to reduce
spending and prices abroad below what they
would have been if the funds had been transferred. If spending and prices rise more rapidly
here and less rapidly abroad, the U. S. trade
surplus suffers.
Still another reason for the relationship is that
when the transfer of funds abroad is reduced, the
balances of payments of foreign countries are
weakened, so that the governments of these
countries tend to pursue more deflationary
policies than they otherwise would follow in an
attempt to reverse part of the change in their
balances of payments. The result, again, is to
diminish U. S. net exports. 5

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

There is little debate among economists over
these general tendencies, but much debate over
their strength and over the importance of other
factors. For example, in explaining changes in
net exports one finds it difficult to isolate the
influence of the capital outflow controls from the
influence of other factors tending to increase the
rate of inflation in this country relative to the
rate of inflation abroad. To take another example, some economists have maintained that
U. S. investments abroad go largely into plants
whose output competes with U. S. exports. To
give some idea of the range of opinion on these
matters, Professor Machlup of Princeton has
argued that U. S. net capital outflows are usually
matched almost dollar for dollar in the same year
by net exports of goods and services, while G. C.
Hufbauer and F. M. Adler have maintained that
U. S. direct investment overseas may act to augment the U. S. balance-of-payments deficit for
many years.6 No attempt is made to settle this
controversy in the present brief remarks; here
we simply note that our own research, together
with that of Rolf Piekarz and Lois Stekler, 7 lead
us to a view closer to that of Machlup than to
that of Hufbauer and Adler.
Of course, the fact that the U. S. trade surplus
has dwindled away does not mean that the surplus cannot be enlarged again. In particular, if

5Other factors, especially differences in capacity utilization
through time between this country and the rest of the world,
also contribute to the correlation between capital outflows
and net exports. In these remarks we mention only a few
factors which act to make net exports partly dependent on
capital outflows.
6See Fritz Machlup, The Transfer Gap of the United States,
Reprints in International Finance, No. 11 (Princeton, N. J.:
Princeton University, 1968), and G. C. Hufbauer and F . M .
Adler Overseas Manufacturing Investment and the Balance
of Pa;ments, U.S. Treasury Department Tax Policy Research
Study Number One (Washington, D. C.: U . S. Government
Printing Office, 1968).
?Rolf Piekarz and Lois Ernstoff Stekler, "Induced Changes
in Trade and Payments," The Review of Economics and
Statistics, XLIX (November, 1967), 517-26.

New England Economic Review
Chart 3

LIABILITIES OF U.S. AND
INTERNATIONAL RESERVES OF U.S.
In Billions of U.S. Dollars
35

25
RESERVES

10
LIABILITIES TO FOREIGN OFFICIAL HOLDERS

O.__....__ _..___

1958

__.._ ___,__ _...___ _. . _ _ _ - - J ._ _....__ _..___

1960

1962

1964

1966

____.....__--J._ _ __ _

June

1969
End of Period

the current rate of inflation in this country were
moderated relative to the rate abroad, the trade
surplus would surely expand, other things remaining the same. A relaxation of the restraints
over capital outflows would probably also enhance the trade surplus; at least in the short run,
however, the increase in capital outflows would
probably exceed the resulting increase in the
trade surplus, so that the immediate impact on
the U. S. balance of payments would be to augment the deficit.
The foregoing discussion does not consider all
of the influences which may be responsible for
24

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

the shrinkage of the U.S. trade surplus. For example, the formation of the European Common
Market and the European Free Trade Association, as well as changes in tastes, in technology,
and in resource availabilities, would also have to
be examined in the course of a thorough analysis.
The purpose in this article is merely to set forth
some of the more important or commonly
mentioned influences. In the same spirit, the
following section briefly considers the state of
international reserves and the relationship between these reserves and the U. S. balance-ofpayments deficit.

NOl'ember/ Decemher 1969

International Reserves and the U. S.
Balance-of-Payments Deficit

cept them as legal tender in settiing international
accounts.

As a result of the continuing deficits in the
U. S. balance of payments, the country's international reserves have declined and its liabilities to
foreign officials have risen, although in recent
years these trends have been interrupted, if not
reversed. (See Chart 3.) The Nation's international reserves, of course, are held by the Federal
Government and consist of gold, convertible
foreign currencies, and automatic borrowing
rights at the International Monetary Fund.
Liabilities to foreign officials take the form of
U. S. Government securities and other U. S.
dollar assets held by these officials.

Not only have international reserves been expanding very slowly, but Chart 4 also shows
that official holdings of gold have declined since
1965. As a result, what growth there has been
in international reserves in recent years has
been in holdings of foreign exchange. Now the
foreign exchange which is held by governments
as international reserves consists almost entirely
of claims on the United States and on the United
Kingdom, and these claims, of course, are the
result of U. S. and U. K. deficits.

Chart 3 clearly reveals why the U.S. Government has been so concerned with the balance of
payments. In theory, foreign officials could present their claims for payment at any time, and
U. S. reserves would be barely adequate to meet
their demands. In practice, of course, it is most
unlikely that all would demand payment
simultaneously.
Chart 4 shows not merely the international
reserves of the United States but those of the
entire non-Communist world; international reserves, of course, are reserves which are used to
settle accounts between countries. It is clear that
these reserves have grown very slowly in recent
years. However, more than 70 nations have recently agreed to create a new form of reserves,
"special drawing rights," in the amount of $3 ½
billion in 1970, $3 billion in 1971, and $3 billion
again in 1972. These special drawing rights, or
SDR's, will probably constitute the major part
of the additions to international reserves in these
years. They will carry a gold value guarantee,
will earn interest, and will be allocated to participating countries in proportion to their quotas
in the International Monetary Fund. Subject to
certain limits, participating countries must ac-

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

In other words, virtually the only sources of
new reserves for the non-Communist world have
been the deficits of the United States and the
United Kingdom. But these two countries have
been under pressure to reduce or eliminate their
deficits. If they succeed, as the United Kingdom
has recently done, they will no longer supply a
large volume of reserves to the rest of the world.
On the other hand, if they fail to curtail their
deficits, other countries may lose confidence in
the strength of the dollar and the pound and try
to obtain gold in exchange for them, in which
case international reserves might decline sharply
as holdings of dollars and pounds were reduced.
Thus the problem of the U. S. balance-ofpayments deficit is closely related to the problem
of providing adequate international reserves.
It is to meet this dilemma that special drawing
rights were designed. Unlike the dollar and the
pound sterling, they will not be the obligation of
any one country, but will have the backing of
the entire non-Communist world, so that they
should not be subject to the loss of confidence
that might afflict the dollar or the pound.

The Appropriate Volume of Reserves
and Their Distribution
Exactly what volume of international reserves
is needed is a difficult question to answer. It is
25

New England Economic Review

Chart 4
fNTERNATIONAL RESERVES
In Billions of U .S. Dollars

( Excluding Communist Countries)

100

1955

1960

1965
End of Period

sometimes argued that these reserves should expand as rapidly as world trade, in order to
finance the growing volume of trade. If this
argument is correct, international reserves have
been growing too slowly, for reserves taken as a
percent of world imports have declined from 47
percent in 1955 to 35 percent in 1968 (not
counting reserves or imports of Communist
countries).
Actually, reserves probably do not need to
increase as rapidly as trade, because international payments are netted, or cleared, so that
it is only temporary balance-of-payments deficits,
not all trade, which have to be financed by the
use of reserves. Moreover, there has been too
much inflation in the world in recent years to
26

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

June

1969

believe that reserves have been seriously inadequate; a serious reserve shortage would
probably have induced more governments to follow deflationary policies in an attempt to realize
balance-of-payments surpluses and thereby to
accumulate reserves. Still, as trade grows, the
magnitude of the balance-of-payments deficits in
the world probably also tends to grow, and since
these deficits must be temporarily financed, international reserves should probably expand a little
more rapidly in the future. This, of course, is
another argument for SDR's.
How will the new SDR's be distributed? Before answering this question, we might note that
the distribution of international reserves has
changed quite radically since 1953. As shown in

November/ Decemher 1969

Chart 5

DISTRIBUTION OF INTERNATIONAL RESERVES
( Excluding Communist Countries)
Percent of Total

______ ..... _

INDUSTRIAL EUROPE
EX CL U DING U . K .

~
_,. _,,,,.

--·--

-------·-·

_..-•--•

LESS DEVELOPED

••

.. .. .

---·-·-•---·-·-·-·-·-.--·-·---·-·-·-·-·-·-·L•
/

19531958

1960

1962

1964

1966

June

1969

End of Per i od

Chart 5, the U. S. share of the total has fallen
from about 44 percent to about 21 percent, while
over the same period the share of industrial
Europe, not counting the United Kingdom, has
risen from 15 percent to 39 percent. Recently,
however, industrial Europe has been losing reserves to other areas.
SDR's will be created and distributed through
the International Monetary Fund, or IMF.
Because the IMF will administer these new reserve assets, it is interesting to see how the voting
power within the organization is distributed
among the member countries. (See Chart 6.)
While the United States has a much larger share
of the votes than any other country, it is not in a

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

position to dictate to the other members. Even
joining forces with other large countries would
not permit the United States to have its way
easily, since key decisions require a very large
majority. For example, 85 percent of the total
votes is required to authorize the creation of
special drawing rights. On the other hand, since
such large majorities are required to make major
changes, the United States is in a position to veto
a fundamental action which it regards as mistaken.
The votes within the IMF are allocated approximately in proportion to the financial contributions, or quotas, which members have paid.
Because SDR's will also be allocated in propor27

New England Economic Review

Chart 6

Percent

PERCENT OF VOTES IN IMF
AND PERCENT OF WORLD IMPORTS, 1968

40
-

PERCENT OF VOTES IN IMF, 1968

~ PERCENT OF WORLD IMPORTS, 1968
30

0
U.S.

U.K.

WEST

FRANCE

GERMANY

tion to these quotas, the United States will receive a little less than 25 percent of all SDR's
created. Thus when $3 ½ billion in SD R's is
created in 1970, the U. S. share will be about
$850 million.
Some countries have objected that they are
now underrepresented in the IMF and that they
will not receive a fair share of the new SDR's.
They argue that they should have more votes and
larger quotas in the IMF to reflect their importance in the world economy. What standards
to use in measuring a country's economic importance is a debatable matter, but it is clear
from Chart 6 that the voting power of several
countries is much lower than their share of world
trade. In this category are West Germany,
France, Canada, and Japan, each of which has
grown rapidly since the IMF was established.
28

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

INDIA

CANADA

JAPAN

OTHER

COUNTRIES

Consequently, it appears that such countries will
soon experience relatively larger increases in
their IMF quotas than other countries.
In conclusion, to return to an earlier theme
of this article, the creation of SDR's may help
the United States to reduce its balance-ofpayments deficit. The reason is that SDR's will
add to the international reserves of all countries,
and after receiving these reserves other countries
may decide that they can afford to run deficits
or smaller surpluses in their balances of payments; and if with the cooperation of the
United States they do run deficits or smaller surpluses, the U.S. balance of payments with them
will of course improve. Once again, the relationship between the U. S. balance-of-payments
deficit and the state of international reserves becomes apparent.

New England Economic Review
INDEX-1969

Banking

Bank Holding Companies and Public Policy Jan./Feb., p. 1.
Commercial Bank Price Competition: The Case of "Free" Checking
Accounts. Sept./Oct., p. 1.
International Economics

The Buy-American Policy of the United States Government: Its Balanceof-Payments and Welfare Effects. July/Aug., p. 1.
The Euro-dollar Market: Its Nature and Impact. May/June, p. 1.
Pax Americana and the U. S. Balance of Payments. Jan./Feb., p. 41.
The U. S. Balance-of-Payments Deficit and the State of International
Reserves. Nov./Dec., p. 18.
Monetary Policy

Defining the Money Supply: The Case of Government Deposits.
March/April, p. 21.
Free Reserves in Monetary Policy Formulation. Nov./Dec., p. 1.
Transportation

Toward a More Efficient Railroad System. March/April, p. 1.
Urban Economics

Increasing Job Opportunities in Boston's Urban Core. Jan./Feb., p. 30.

https://fraser.stlouisfed.org
Federal Reserve Bank of St. Louis

Full text of New England Economic Review : November/December 1969

FRASER