Full text of Review (Federal Reserve Bank of St. Louis) : August 1972

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

F E D E R A L R ESER V E B A N K
OF ST. LOUIS
AUGUST 1972

RPDs and Other Reserve Operating
Targets.............................................
An Appropriate International Currency —
Gold, Dollars, or S D R s ? ......................
L IT T L E R O C K

Digitized for Vol. 54, No.
FRASER


8

2
8

RPDs and Other Reserve Operating Targets
by CHARLOTTE E. RUEBLING

S INCE EARLY this year the Federal Reserve has
emphasized member bank reserves available for pri­
vate nonbank deposits ( RPDs) as its short-run operat­
ing target.1 RPDs are a reserve aggregate, defined as
total member bank reserves less reserves required
against U.S. Government demand deposits and net
interbank deposits. This article examines a framework
in which a reserve aggregate, such as RPDs, functions
as an operating target to achieve policy objectives.
U.S. monetary authorities —the Federal Reserve
System and the U.S. Treasury —share in the responsi­
bility for achieving the Government’s economic stabi­
lization objectives. These objectives, such as reducing
inflation and unemployment, and encouraging a sus­
tainable growth of output and income, are sometimes
summarized in a growth rate for gross national prod­
uct (GNP). From the point of view of the monetary
authorities, influencing GNP is generally interpreted
as being achieved through the intermediate step of
influencing market interest rates or growth of the
money stock. In this article we assume that policy
objectives have been translated into a desired growth
rate for the money stock and examine how the use of
a reserve aggregate as an operating target can help
the Federal Reserve achieve the desired rate of growth
of the money stock.

Definition of the Money Stock

deposits at commercial banks other than large certifi­
cates of deposit) and M;! (M-. plus deposits at nonbank-thrift institutions) r
In this article the M, measure —demand deposits
and currency in the hands of the public —is used,
where “public” is any person or institution other than
a monetary authority or a commercial bank. This
definition of public means that currency held by com­
mercial banks, the U.S. Treasury, or Federal Reserve
Banks, and demand deposits owned by the U.S. Treas­
ury ( Government deposits) or commercial banks ( in­
terbank deposits) are excluded from the money stock.
The demand deposit component of the money stock
is sometimes called “private demand deposits.” When
this terminology is used, “private” deposits must be in­
terpreted as including those owned by state and local
governments. The term “demand deposits adjusted”
is also used to refer to the demand deposit component
of money, computed as demand deposits at all com­
mercial banks other than those due to domestic com­
mercial banks and the U.S. Government, less cash
items in process of collection and Federal Reserve
float, plus foreign demand balances at Federal Reserve
Banks.

Sources of the Money Stock
T he Role of Commercial Banks

The money stock can be measured in several differ­
ent ways. Most commonly, it is measured as Mi (de­
mand deposits plus currency in the hands of the
public). Other measures include M2 (M j plus time

The public’s most immediate sources of money are
commercial banks. The ultimate sources of money in
our economy, however, are the monetary authorities
—the U.S. Treasury and the Federal Reserve System.

1“Record of Policy Actions” of the Federal Open Market

2Time deposits due to domestic commercial banks or the U.S.
Government are not included in M2 or M3.

Committee, Federal Reserve Bulletin (May 1972), p. 459.


Page 2
http://fraser.stlouisfed.org/

Federal Reserve Bank of St. Louis

AUGUST 1972

FEDERAL RESERVE BANK OF ST. LOUIS

Commercial banks are very important in the money
supply process because the public chooses to hold
money in the form of demand deposits. When cur­
rency supplied by the monetary authorities is ex­
changed by the public for a demand deposit, the form
of money changes. In addition, banks obtain addi­
tional reserves. In a fractional reserve banking sys­
tem, which is what we have in the United States,
required reserves equal only a fraction of deposits.
The remainder of funds obtained with deposits may
be used by banks to make loans or investments. By
making loans commercial banks increase their liabili­
ties (demand deposits) and assets (loans), and in a
sense “create” money on the basis of the original
money (reserves) provided by the monetary authori­
ties. The net monetary liabilities of the monetary
authorities are held either as currency in the hands of
the public or as commercial bank reserves and can be
viewed as the “base” for the nation’s money stock.

The Role of Monetary Authorities
The capacity of the commercial banking system to
create demand deposits is constrained by the total
amount of reserves it has and by the ratio of required
reserves to deposits. The Federal Reserve System can
directly influence the money stock by both influenc­
ing the amount of reserves and changing reserve
requirements.
The Board of Governors of the Federal Reserve
System establishes the proportions of various types of
deposits that member banks must hold either as vault
cash or as deposits at Federal Reserve Banks to meet
legal reserve requirements. Presently, required reserve
ratios are between 12.5 and 17.5 percent on demand
deposits and 3 or 5 percent on time deposits.3 The
higher the required reserve ratio, the smaller the
amount of demand deposits the banking system can
create with any given amount of reserves.
The Federal Open Market Committee (FOMC),
through decisions on Federal Reserve open market
operations, has an important influence on the amount
of reserves in the banking system. This Committee,
composed of the Board of Governors of the Federal
Reserve System and Federal Reserve Bank Presidents,
3Reserve requirements of member banks are listed in the
Federal Reserve Bulletin. For example, see p. A 10 of the
July 1972 issue.
Since reserve requirements for commercial banks currently
differ with respect to their classification and amounts of de­
posits, shifts in deposits from reserve city banks to country
banks or vice versa, from member banks to nonmember banks
or vice versa, or from smaller to larger banks or vice versa,
may change the ratio of required reserves to deposits for the
banking system as a whole.



meets about once a month to consider economic con­
ditions and the ways in which its actions can best
serve economic goals. The decisions of the FOMC
are carried out at the Trading Desk of the Federal
Reserve Bank of New York. The Manager of the
Trading Desk buys and sells securities on behalf of
the twelve Federal Reserve Banks in accordance with
the instructions of the “Directive” from the FOMC.
When the Trading Desk buys a security, it pays for
it by supplying either bank reserves or currency. On
the other hand, when it sells a security, payment to
the Federal Reserve results in a reduction of bank re­
serves or currency in the hands of the public.

Reserve Aggregates as Operating Targets
We have seen that bank reserves influence the
money stock and that the total amount of reserves in
the banking system is strongly affected by Federal
Reserve open market operations. Next we want to
identify the factors that determine the relationships
between changes in various measures of bank reserves,
that is, various reserve aggregates, and changes in the
money stock. For a reserve aggregate to function as
an operating target, policymakers must be able to
predict the relationship between the reserve aggregate
and money with some degree of accuracy over the
period in which they are attempting to control the
reserve aggregate. Information about this relationship
is necessary in order to select the appropriate path for
the reserve aggregate. We must also examine the in­
formation which is necessary to achieve desired
changes in the reserve aggregate, since paths of the
reserve aggregates themselves are not under the com­
plete control of the monetary authorities.
There are many reserve aggregates which bear
definable relationships to the money stock on one side
and to open market operations on the other. We will
consider three: the monetary base, member bank re­
serves, and RPDs (member bank reserves used to
support private nonbank deposits).

The Monetary Base
One reserve aggregate concept that is useful for
monetary analysis is the monetary base. The base is
defined as the net monetary liabilities of the Govern­
ment (U.S. Treasury and the Federal Reserve Sys­
tem) held by commercial banks and the nonbank
public. These monetary liabilities are member bank
reserves and currency in the hands of the public.4
4A use of the base which is excluded from the analysis in this
paper is vault cash held by nonmember banks.

Page 3

FEDERAL. RESERVE BANK OF ST. LOUIS

The monetary base is derived from a consolidated
balance sheet of the Treasury and Federal Reserve
“monetary” accounts. The total amount of base out­
standing at any time is principally determined by the
U.S. gold stock (which has been relatively constant
in recent years) and the amount of U.S. Government
securities owned by the Federal Reserve System.
Whenever the Federal Reserve buys securities on the
open market for its own account, the monetary base
increases.5
The amount of money the base will support is
largely determined by the decisions of commercial
banks, the public, and the Treasury. Each $1 of base
the public chooses to hold as currency, for example,
supports only $1 of money; whereas $1 of base held in
the banking system as reserves may support more than
$1 of demand deposits and therefore more than $1 of
money. However, if reserves are absorbed as require­
ments against Government deposits, interbank de­
posits, or time deposits, then they are not available to
support expansion of private demand deposits. The
amount of demand deposit money that the banking
system can supply to the public with a given amount
of reserves is influenced by decisions determining the
relative amounts of different types of bank deposits.
A more detailed example can help explain how the
behavior of commercial banks, the public, and the
Treasury influences the amount of money that the
base will support (see illustration). Suppose the base
equals $100 billion, $60 billion of which is used as
currency in the hands of the public and $40 billion as
bank reserves. These reserves may be used as legal
reserve requirements against time deposits, private
demand deposits, U.S. Government deposits, inter­
bank deposits,8 or they could be held by banks as ex­
cess reserves. We can assume a certain set of reserve
requirement ratios for the different types of deposits
and a certain amount of each type of deposit, and
compute the money stock which is consistent with the
$100 billion base. Assume banks must hold reserves
equal to 5 percent of time deposits and 15 percent of
5For a more detailed discussion of the monetary base, see
Leonall C. Andersen and Jerry L. Jordan, “The Monetary
Base —Explanation and Analytical Use,” this Review ( August
1968), pp. 7-11; Jerry L. Jordan, “Elements of Money Stock
Determination,” this Review (October 1969), pp. 10-19; Jane
Anderson and Thomas M. Humphrey, “Determinants of
Change in the Money Stock: 1960-1970,” Monthly Review,
Federal Reserve Bank of Richmond (March 1972), pp. 2-8;
John D. Rea, “Sources of Money Growth in 1970 and 1971,”
Monthly Review, Federal Reserve Bank of Kansas City (July
/August 1972), pp. 3-13.
6Banks have some nondeposit sources of funds against which
they must hold reserves. We ignore these in the analysis in
this article.

Page 4


AUGUST 1972

all other deposits. If the public chooses to hold $300
billion of time deposits, this uses up or absorbs .05 x
$300 billion or $15 billion of reserves. Now $25 billion
of reserves are available to support various types of
demand deposits. If Government deposits equal $4
billion and interbank deposits equal $1 billion, these
two types of deposits use up 0.15 x $5 billion or $0.75
billion of reserves. The remaining $24.25 billion of re­
serves will support a maximum of approximately $162
billion of private demand deposits.7 If banks choose to
hold $1 billion in excess reserves, then given our other
assumptions, reserves used to support private demand
deposits will equal $23.25 billion and private demand
deposits will amount to $155 billion. In this example,
the base of $100 billion supports a money stock of
$215 billion, $60 billion of which is currency and $155
billion of which is private demand deposits.
A change in the composition of deposits could be
expected to change the amount of money resulting
from a $100 billion base. Suppose the public decided
to increase its holdings of time deposits by $50 billion.
Initially this would require a $50 billion reduction in
private demand deposits, assuming other factors do
not change. Since the ratio of required reserves to time
deposits is less than the ratio of required reserves to
demand deposits, the shift would lower the average
reserve requirement for the banking system. By mak­
ing loans with reserves released in the shift of deposits,
the banking system could again raise the level of de­
mand deposits in the system. In the example, a $50
billion contraction of demand deposits would release
$7.5 billion of reserves (0.15 x $50 billion). The in­
crease in time deposits would absorb $2.5 billion of
reserves (.05 x $50 billion). The remaining $5 billion
of reserves would support $33.33 billion of demand
.
/ $5 billion \
.
, .
. . . r .
deposits I —
— I . One might look at it in the fol­
lowing way. A $50 billion increase in time deposits
would require a $16.67 billion contraction of demand
deposits, given other things do not change. Increased
preference for time deposits, with no change in the
amount of base, results in a decline in private de­
mand deposits, and hence money stock.
An increase in Government deposits at the expense
of private demand deposits, with no change in total
bank reserves, would also result in a contraction of
the money stock. One occasion in which this occurs is
when people pay Federal income taxes. With more
reserves used to support Government deposits and
7This is computed by dividing reserves available by the re. j
• / $24.25 billion
\
quired reserve ratio I q
= ■ 161.6 / billion I.
?

AUGUST 1972

FEDERAL RESERVE BANK OF ST. LOUIS

Illustrations of U ses of R e se rv e A g g r e g a t e s *
( B illio n s of D o lla rs )

MONEY STOCK = $215

A r e s e r v e a g g re g a t e can se rv e a s an e ffe c tiv e o p e r a t in g targe t w h e n m o n e ta ry a u t h o r itie s can se le c t the a p p ro p ria te g r o w t h rate o f the r e s e r v e a g g r e g a t e and
a ch ie ve th a t g r o w t h ra te w ith an acceptable d e g re e of accuracy.

Se le ctin g a d e sire d g r o w t h p ath fo r a re s e rv e a g g re g a t e in v o lv e s p re d ic tin g its re la t io n sh ip to the

m one y stock. A c h ie v in g the selected gro w th rate re q u ire s p re d ic tin g the re la tio n sh ip b etw e en m o n e ta ry a ctio n s a n d the r e s e r v e a g g re g a te .
T he b o i e s a c ro ss the top of the illu stra tio n sh o w the " u s e s " of re se rve a g g r e g a t e s . Each of the se " u s e s " influe nces either the re la tio n sh ip b e tw e e n a re se rv e
a g g re g a t e and m o ne y, or the re la t io n sh ip b etw e en open m a rk e t o p e ra tio n s a n d the path of the re se rv e a g g r e g a t e . A s s u m in g the m o n e ta ry b ase can be c o n tro lle d by
the m o n e ta ry authorities, its use a s on o p e ra tin g t a r g e t a n d the selection of its m a gn itu d e re q u ire s p red ictin g its re latio n sh ip to m o ne y, w h ich is a ffe cte d b y l-V I.
If total m em ber b a n k r e s e r v e s o re used a s the o p e ratin g targe t, the se le c tion of th e ir p a th re q u ire s that the m o ne y r e la tio n sh ip in v o lv in g ll- V I be e stim a te d .
A c h ie v in g the ta r g e t itself im plies k n o w le d g e of the m o n e ta ry b a s e and I. The use of R P D s (r e s e r v e s a v a ila b le to su p p ort p riv a te n o n b a n k d e p o sits) a s the o p e ratin g
ta rg e t re q u ire s p red ictin g a r e la tio n sh ip in vo lvin g 11,111, a n d IV to select the g r o w t h of R P D s and in fo rm a tio n a b o u t l,V, a n d V I to achieve tha t path.
*T h e illu s tr a tio n e x c lu d e s c o n sid e ra tio n of n on d ep o sit so u rce s of fu n d s a n d v a u lt cash of n on m em b e r b a n k s . R e s e r v e re q u ire m e n ts are assu m e d to b e I S p e rcen t
on p riv a te dem and, G o v e rn m e n t dem and, and in t e rb a n k d e p o sits , a n d 5 percent on time d ep o sits.

fewer available to support private deposits, the money
stock tends to decline.
If the Federal Reserve were to use the base as its
operating target to achieve the desired rate of growth
of the money stock, it would have to predict with
some degree of accuracy the relationship between the
growth rate of the base and the growth rate of the
money stock. If all the uses of the base were known
with certainty, the relationship between growth of the



base and growth of the money stock could be known
precisely, and the FOMC could translate any desired
growth rate of the money stock into a target rate of
growth for the base. Since most of the uses cannot be
predicted with complete certainty, the relationship
between the base and money is often estimated from
historical behavior as a single number, or “multiplier.”
This multiplier nevertheless incorporates conceptually
and quantitatively all of the uses of the base discussed
in preceding paragraphs.
Page 5

FEDERAL. RESERVE BANK OF ST. LOUIS

A second criterion for using the base as an operat­
ing target is that the Trading Desk must be able to
achieve the targeted rate of growth of the base. To do
this, it must predict what will be happening to factors,
other than those under its initiative, which affect the
base. These factors include changes in Federal Re­
serve float and member bank borrowing from Federal
Reserve Banks. With information about how outside
factors would result in increases or decreases in the
base, the Desk could use open market operations to
make up the difference between the change in the
base caused by uncontrolled factors and the desired
change. For example, suppose that the base were the
operating target, that the only two sources of the base
were Federal Reserve purchases of securities and
Federal Reserve lending to member banks, and that
these two sources were independent of each other. If
the desired rate of growth of the base were translated
into a $1 billion increase in a given week, and the
Trading Desk expected a $0.3 billion increase in mem­
ber bank borrowing that week, it would purchase $0.7
billion of securities to make up the difference.

M em ber Bank Reserves
Bank reserves can be viewed as the base less the
currency component of the money stock. If reserves
were used as an operating target, then the relation­
ship to be estimated would be between reserves and
the demand deposit component of money. Information
about currency, necessary to achieve a certain growth
rate of reserves, could be used to compute the change
in private demand deposits necessary to achieve the
desired money growth.
The Trading Desk must predict the public’s de­
mand for currency in order to attain the desired
growth rate for reserves, since changes in the public’s
holdings of currency change the amount of the base
available for bank reserves. In other words, the Desk
must predict the distribution of uses of the base be­
tween currency in the hands of the public and bank
reserves, as well as uncontrolled influences on the
base. If the FOMC has very good evidence about
what will be happening to currency in the hands of
the public because, say, currency has increased at a
historically steady rate, then it may be no more diffi­
cult to achieve a given growth rate for reserves than
for the base. If at the same time the relationship be­
tween reserves and demand deposits is more pre­
dictable than that between the base and money, using
reserves instead of the base as the operating target
might result in a lower error involved in achieving the
desired rate of money growth.

Page 6


AUGUST 1972

RPDs
RPDs are a particular reserve aggregate that can
be viewed as the base minus currency in the hands
of the public and reserves used to meet legal require­
ments against U.S. Government demand deposits and
net interbank deposits. This is equivalent to total
member bank reserves minus required reserves against
Government and interbank deposits. In the illustra­
tion RPDs amount to $39.25 billion.
As is the case with any other reserve aggregate,
how changes in RPDs affect the money stock is influ­
enced by its uses. Uses of RPDs include required re­
serves against all time and savings deposits, required
reserves against private demand deposits, and excess
reserves. Since RPDs exclude required reserves
against Government and interbank deposits, informa­
tion about these two types of deposits is not needed to
estimate the relationship between RPDs and private
demand deposits. However, in order for the Trading
Desk to achieve the targeted path of RPDs, it must
estimate reserves required against Government and
interbank deposits. This is in addition to forecasting
factors needed to achieve total reserves.
Since fewer factors affect the relationship between
RPDs and demand deposits than those involving the
base or total reserves, the relationship between RPDs
and the money stock may exhibit greater stability
than either of the others. If this is so, errors involved
in selecting the correct path of RPDs might be smaller
than those involved in selecting an appropriate path
for the monetary base or total reserves. If at the same
time, currency, Government deposit, and interbank
deposit uses of the base can be estimated with con­
siderable accuracy, errors involved in achieving the
targeted growth rate for RPDs would not be signifi­
cantly larger than those involved in achieving given
growth rates for the base or total reserves. In general,
reliable information about some of the uses of a re­
serve aggregate may tend to make the aggregate
which excludes those uses a better operating target.

Interpreting Changes in Reserve Aggregates
As is implied in the foregoing discussion, changes
in the growth rate of a reserve aggregate, such as
RPDs, should not be viewed as indications of the
tightness or ease of monetary actions. For example,
any increase in a reserve aggregate which is absorbed
as required reserves against time deposits is not avail­
able to support expansion in private demand deposits.
Consequently, when the public is accelerating the
pace of acquiring time deposits, larger growth rates

FEDERAL RESERVE BANK OF ST. LOUIS

in the reserve aggregates discussed will be necessary
to achieve a given growth rate of money, compared to
when time deposit growth is constant or decelerating.
In other words, the same rate of growth of a reserve
aggregate may result in different rates of growth of
money at different times.
In terms of RPDs, one might find that a 6 percent
annual rate of growth in money could require a 6 per­
cent rate of growth of RPDs in one month and an 8
percent rate in another. This suggests that one
should employ some caution in interpreting changes
in the growth rate of RPDs over short periods. A
change in the rate of growth of RPDs does not neces­
sarily suggest a change in the influence of monetary
actions on economic activity.
It is sometimes assumed, as it is here, that the rate
of growth of the money stock over a period of signif­
icant duration indicates the impact of monetary de­
velopments on the economy. Interpreting the role of
reserve aggregates in such a framework requires in­
formation about changes in other factors that influence
money growth.

Summary
A reserve aggregate is an effective operating target
for monetary policy when the Federal Reserve can
both select the growth rate of the aggregate required
to meet policy goals and achieve that growth rate
with some degree of accuracy. If the growth of the
money stock is an important intermediate link be­
tween monetary actions and the ultimate goal of
monetary policy, then the first criterion suggests that
the Federal Reserve must be able to forecast the
average relation between the reserve aggregate and
the money stock over the period in which it attempts
to control the reserve aggregate. The second criterion




AUGUST 1972

suggests that the Trading Desk must be able to fore­
cast changes in the reserve aggregate that would re­
sult from circumstances other than its own actions.
Errors are bound to exist in both the estimation of
the relationship between the growth rates of the re­
serve aggregate and money and in the forecast of
uncontrollable factors influencing the growth of the
reserve aggregate. If the errors are small, however,
they may be of little consequence to the growth rate
of the money stock over the period in which it is
significant to GNP.
Thus the choice of a specific reserve aggregate as
an operating target depends on whether:
( 1 ) the information necessary to select the appro­
priate growth path of that reserve aggregate is
better than the information necessary to select
growth paths of other aggregates;
( 2 ) the information necessary to achieve a selected
growth path of that reserve aggregate is better
than the information necessary to achieve growth
paths of other aggregates; and
(3 ) the total information errors (in selecting and
achieving) are minimized by using that reserve
aggregate as compared to other aggregates.

The FOMC adopted RPDs as its operating target
at its February 1972 meeting. RPDs are a reserve ag­
gregate defined as total member bank reserves less
reserves required against U.S. Government demand
and net interbank deposits.
Since the experiment using RPDs as the operating
target has been in existence for only about seven
months, there is insufficient evidence to judge the
extent to which it is a better target than reserves or
the base. The experience thus far indicates, however,
that the use of RPDs has been quite successful in
producing desired money growth.

An Appropriate International Currency
Gold, Dollars, or SDRs?*
by MICHAEL W. RERAN

I n THE wake of the momentous international eco­
nomic events of 1971, there is serious talk of the need
for major reforms in the international monetary
structure. Many economists consider the breakdown
of the old system in 1971 as a confirmation of the in­
herent weakness of a fixed exchange rate system. Their
reform proposals call for a general movement to a sys­
tem of flexible exchange rates. However, policymakers
of most countries have demonstrated by word and
deed their continued opposition to abandoning the
basic structure of a fixed exchange rate system.1 Op­
erating through the International Monetary Fund
(IM F) a new “Group of 20” finance ministers and
central bank governors has recently been formed to
negotiate a new monetary system.
This article considers the situation in which fixed
exchange rates continue to be the basis of any new
system (albeit with more frequent rate changes and
temporary floats) and focuses on the issue of an ap­
propriate international currency in such a system.
Specifically, two questions are investigated. Should
the role of the dollar be reduced or eliminated? If
so, what could replace it? The discussion will deal
almost entirely with international currency in foreign
official use. The private use of an international cur­
rency is both large and important. However, the
“The initial idea of this paper evolved from the author’s
association with Messrs. Peter Clark and John Makin at the
U.S. Treasury Department during the academic year 1971-72.
This article expresses the views of the author and does not
necessarily represent the views of the Federal Reserve Sys­
tem or the Federal Reserve Rank of St. Louis.
'The fixed exchange rate system is neither completely fixed
nor unchanging. The exchange rate is allowed to vary from
day to day within a margin. That margin was “temporarily”
widened from 1 percent to 2% percent around the “par
value” as part of the Smithsonian Agreement of December
18, 1971. The exchange rate can be changed within the IMF
articles of agreement when there is a “fundamental disequilib­
rium.” This characteristic has led many commentators to call
it the adjustable par value rather than a fixed exchange rate
system. For a recent defense of the fixed exchange rate con­
cept, see Samuel I. Katz, “The Case for the Par Palue Sys­
tem, 1972,” Essays in International Finance (March 1972).

Page 8


problems considered in this article are primarily re­
lated to official, rather than private dollar use.
Central banks have a demand for international re­
serves somewhat analogous to the demand by com­
mercial banks for domestic reserves.2 When interna­
tional reserves can be held in more than one form,
central banks also have a portfolio preference as to
the composition of their reserves. There are a number
of techniques available to central banks for controlling
both the quantity and the composition of their inter­
national reserves.
A desirable characteristic of an international cur­
rency system is the provision for incentives to central
banks to achieve their desired level and composition
of reserves in a way which promotes international
monetary stability. Such stability can be achieved
by prompt changes in exchange rates between na­
tional currencies which lead to balance-of-payments
equilibrium.
The first section of this article analyzes the inter­
national monetary system which existed from the early
post-war years to August 1971. It concludes that the
multiple unit international currency, such as the dollargold system, was defective in the sense that it did not
provide sufficient incentives to central banks and gov­
ernments to achieve equilibrium exchange rates. This
led to progressively larger speculative flows of short­
term capital, which eventually forced exchange rate
changes.
The second section analyzes the single unit inter­
national currency system, based on the U.S. dollar,
which has evolved since August 1971. It is argued that
this system is superior to the old one in the sense that
2See Herbert Grubel, “Demand for International Reserves —A
Critical Review of the Literature,” in Review of Economic
Literature (December 1971). Also, see Michael Keran, “De­
mand for International Money —A Micro Approach,” mimeo­
graphed (Federal Reserve Rank of St. Louis), available on
request.

FEDERAL RESERVE BANK OF ST. LOUIS

it provides incentives to move towards equilibrium
exchange rates.
The third section considers the possibility of future
changes in the international monetary system. On the
one hand, fears of future inflation in the United States
make many foreign governments reluctant to continue
under a straight dollar system. On the other hand, a
return to a dollar-gold or other multiunit international
currency system would weaken the present incentives
for achieving equilibrium exchange rates. Any future
monetary reform must simultaneously satisfy the con­
cerns of foreigners with respect to eliminating any
undesired dollar balances and the need of the inter­
national monetary system to have built-in incentives
towards achieving equilibrium exchange rates.
The movement to a system based on a purely inter­
national currency unrelated to the national currency
of any country (such as an SDR standard) has the
potential to deal with both problems. However, there
are some important obstacles to an SDR standard
which could make it an impractical alternative.

The Dollar-Gold System
In the period from the early 1950s to August 1971,
the dollar was not only the national currency of the
United States, but also a major component of the in­
ternational currency of the world. The other major
component was monetary gold. The dollar satisfied
all the conditions of an international currency: it was
a means of payment for settlement of private inter­
national transactions (trading currency); it was used
by foreign central banks to maintain the value of their
national currency in the foreign exchange market (in­
tervention currency); it was a store of value for hold­
ing international assets (reserve currency); and it
was a unit of account for measuring the value of
international transactions. In short, it was the de facto
international numeraire.3
There is a potential flaw in a multiunit international
currency system such as that based on the dollar and
gold. It is related to the fact that in this special case a
national currency like the dollar has two interna­
tional ties. One is a standard tie with other national
currencies through the system of fixed exchange rates.
The other tie is unique to a national currency which
is also a component of international reserves. This is
the “conversion rate” between the national currency
and other elements of international reserves.
3Gold continued to be the de jure numeraire in that all official
par values established with the International Monetary Fund
were defined in terms of gold.



AUGUST 1972

A problem can arise because there are no natural
market forces which will make these two links con­
sistent with each other. The system inhibits the United
States from changing its exchange rate and does not
encourage other countries to change their exchange
rates in an equilibrium direction. This has contributed
to persistent balance-of-payments surpluses for some
countries and deficits for the United States.
In this circumstance the only market force operat­
ing to achieve equilibrium exchange rate change is
private speculation with associated massive capital
flows. This eventually will force appropriate exchange
rate changes. However, these adjustments are so long
delayed and the associated market disruption so great
as to distort trade and financial flows and the effi­
ciency of the international economic system.
The lack of incentives for foreign governments to
change exchange rates is related to the fixed conver­
sion rate between the dollar and gold. This is dis­
cussed below in the section “The Dollar as Interna­
tional Currency.” The restraints on the United States
in changing its exchange rate are discussed in the sec­
tion “The Dollar and Other National Currencies.”

The Dollar as International Currency
In general, foreign central banks have maintained
the international value of their national currencies by
standing ready to buy and sell their currency at fixed
rates for the dollar.4 Thus, a foreign government’s
balance-of-payments surplus has been registered, in
the first instance, as an increase in dollar reserves, and
a deficit as a decrease in dollar reserves. If this led to
a dollar share of reserves other than that desired by
the foreign central bank, it could adjust its portfolio by
exchanging dollars for gold with the U.S. Treasury
at a fixed and known price of $35 per ounce. Since
central banks could control the stock of dollars they
acquired through this portfolio adjustment process,
there was less incentive to control the stock of dollars
via a change in the price offered for dollars, that is,
through a change in the exchange rate of the national
currency for dollars.5 In addition, when a central bank
could achieve a desired composition in its international
4The spread between the buying and selling rates represents
the margin around the par value in which the private forces
of supply and demand are generally allowed to operate. Until
December 1971, the maximum margin was 1 percent above
and below the par value. It was “temporarily’ increased to
2V percent in conjunction with the Smithsonian exchange
4
rate agreement.
5This, of course, does not mean that there were no incentives
for other countries to change their exchange rate. It only
means that there was no incentive associated with achieving
a desired level of dollar holdings relative to alternative forms
Page 9

FEDERAL RESERVE BANK OF ST. LOUIS

reserves, it was more likely to accept a higher level of
total reserves than if it were unable to control the com­
position of reserves. Thus, a given increase in the total
level of reserves under a gold-dollar standard was less
likely to induce the central bank to change the ex­
change rate.
The portfolio adjustment process by central banks
to eliminate undesired dollar balances was not, of
course, perfect. Some central bankers were well aware
that too large a demand for gold from the U.S. Treas­
ury would lead to a breakdown in the system. How­
ever, most central banks maintained remarkably
stable ratios of gold to total reserves over the twentyyear period from 1949 to 1969, in spite of the sharp
rise in the overall level of reserves. The proportion of
gold to dollar reserves for all IMF members (exclusive
of the United States) was held in the narrow range
between 45 and 55 percent from 1949 to 1969. Only
towards the end of 1970 did the gold share fall below
this range. The proportion of dollars to total reserves
was held in a range between 25 and 30 percent in the
period 1953 to 1970. Only in 1971 did the dollar
share of reserves rise above 30 percent. Sterling, on
the other hand, declined steadily over the whole post­
war period as a share of total reserves.
Each of the major industrial countries maintained
a relatively stable share of gold to total reserves during
this period. Some countries, like Germany and Japan,
held relatively small ratios of gold to total reserves,
while other countries, like France, Belgium, and the
Netherlands, held relatively large ratios. In each coun­
try, however, the ratio was remarkably stable.6
With the United States maintaining a fixed conver­
sion rate between the dollar and gold at $35 per
ounce, foreign central banks had no incentive to con­
trol the dollar share of their reserves through changes
in the exchange rate of their national currency for
dollars. This represented a significant rigidity in the
old international monetary system which permitted
many foreign currencies to become undervalued and
the U.S. dollar overvalued.
This system was viable only as long as the U.S. in­
ternational competitive position was strong, and this
of reserve assets. Given the speed by which foreign countries
changed their exchange rates after August 1971, this lack of
incentive appears to have been substantial.
6We cannot, of course, determine from direct observation
whether the desired ratio of gold to total reserves was equal
to the actual ratio. However, given that these ratios were
relatively stable over a period of almost twenty years, it is
reasonable to assume that central banks were satisfied with
the proportion of their international reserves in gold and in
other forms, specifically the dollar.

http://fraser.stlouisfed.org/
Page 10
Federal Reserve Bank of St. Louis

AUGUST 1972

Shares of International Reserves
Percent

Percent

Note-. D o lla rs a n d " ste rlin g a re r e p o rt e d a s lia b ilit ie s o f the U.S. a n d U.K. to fo re ig n
official institutions w h ile g o l d is a re se rv e a s s e t re p o rte d b y fo re ign c oun tries.
To tal in te rn a tio n a l re se rv e s a re the sum o f re se rv e a s s e t s re p o rte d b y co u n trie s
to the IM F le s s U .S. re s e rv e a ss e t s . R e s e rv e asse ts in clud e g o ld , S D R s , the n et IM F
p o sitio n , a n d fo re ig n e xc h a n g e .

depended upon the domestic economic performance
of the United States and the exchange rate between
the dollar and other currencies. To analyze the factors
which inhibited the United States from eliminating
the overvalued dollar by changing its own exchange
rate requires a look at the dollar in relation to other
national currencies.

The Dollar and Other National Currencies
The dollar is linked to other national currencies
through the system of fixed exchange rates. Concep­
tually, an internationally overvalued dollar can be
eliminated by one of three possible actions: reduction
in U.S. prices, increase in foreign prices, or changes in
the exchange rate between the dollar and other na­
tional currencies. For purposes of the current discus­
sion, we will limit ourselves to considering only the
third solution, changing the exchange rate.7 The fixed
exchange rate system permits correction of an over
or undervalued currency by changing the exchange
rate between national currencies.
The exchange rate of the dollar, in contrast to the
exchange rates of other national currencies, was not
under the unilateral control of the United States in
the old international monetary system. The rate for
the dollar was determined by each foreign govern7Most empirical studies indicate that it takes a long time and
substantial, though temporary, loss of domestic production for
monetary and fiscal policy to affect prices. Exchange rate
changes, on the other hand, can be taken quickly.

FEDERAL. RESERVE BANK OF ST. LOUIS

ment fixing the exchange rate for its national currency
by intervening in its foreign exchange market with
dollars. When all countries maintained their exchange
rate by standing ready to buy or sell a unit of their
national currency for a fixed number of dollars, the
United States, in effect, did not determine its own
exchange rate.
This is referred to as the “Nth” currency problem.
If there are “N” countries with “N” currencies, there
can be no more than “N-l” independently determined
exchange rates or prices between these currencies.
One of these countries must be passive with respect
to the exchange rate of its currency. In the old inter­
national monetary system, the dollar, because of its
international reserve role, was the “Nth” currency.8
The only unilateral action the United States could
take was to change or suspend the conversion rate
between the dollar and gold. However, either action
would have destroyed the dollar-gold international
currency system. The United States was inhibited from
taking either action because of the disruptive effects
it would have on many countries.9 Thus an important
economic policy tool available to other governments
who desire to achieve balance-of-payments equilib­
rium was not available to the United States within the
context of the old international monetary system.

Conflict Between National and International
Currency Roles
In principle, there is no reason why there should be
a conflict between the national and international cur­
rency roles of the dollar. If each foreign country had,
by taking appropriate individual exchange rate ac­
tions, insured that its currency was not undervalued,
the U.S. dollar would have automatically avoided be­
ing overvalued. No change in the dollar-gold con­
version rate would have been necessary. However,
most governments appear to hold the neo-mercantilist
view that balance-of-payments surpluses are a sign
of national strength, rather than a sign of external
imbalance.
8If N = 2 and France decides that 4 francs is the appropriate
exchange rate for $1, the United States cannot simultaneously
decide that it will exchange 3 francs for $1. These rates are
inconsistent with one another.
If N = 3 or more, the consistency rule must, of course, also
be satisfied between the French exchange rate and any third
country, like Germany. However, because the French and
Germans use the dollar, and not each other’s currency, in
establishing and maintaining the exchange rate, the ‘ cross
rates” are automatically kept consistent by private arbitragers.
9When speculation caused the U.S. balance of payments to
experience unprecedented deficits, the United States did
take action and the dollar-gold system was destroyed.



AUGUST 1972

When a national currency becomes overvalued and
that country experiences balance-of-payments deficits
and a weakened international competitive position of
export and import-competing industries, it tends to
devalue promptly. However, an undervalued national
currency leads to balance-of-payments surpluses and
increased international competitive strength, both of
which may appear to be desirable developments. Thus
without other incentives, an undervalued currency
has generally not been eliminated quickly.
The IMF articles of agreement prohibit govern­
ments from achieving an undervalued currency by
explicit exchange rate change. However, if differen­
tial rates of inflation or other factors lead to an under­
valued currency, a government is not required to
appreciate its exchange rate to eliminate the condi­
tion. Thus when domestic inflation developed in the
United States in the second half of the 1960s, the
dollar gradually became overvalued and the interna­
tional competitive position of the U.S. export and
import-competing industries deteriorated. With the
exception of Germany, the countries whose currencies
were undervalued were not willing to appreciate. It
seemed reasonable to them that if domestic inflation
in the United States had caused the overvalued dollar
and a payments deficit, the United States alone should
solve the problem. This view overlooked the fact that
the standard policy tool for dealing with an over­
valued currency —devaluation —was not available to
the United States under the old system.
With other countries unwilling to change their ex­
change rates, and the United States unable to do so,
the only unilateral action the United States could take
in the face of an overvalued dollar and massive specu­
lation (short of major trade and capital controls) was
to suspend dollar convertibility into gold.10 This ac­
tion was taken on August 15, 1971. Although, by itself,
10The United States could have unilaterally increased the
dollar price of gold. However, this action would have elim­
inated the overvalued dollar only temporarily, if at all.
Since the Smithsonian Agreements, we know that foreign
governments would allow the United States a devaluation
of no more than 9-10 percent. If the United States had
raised the price of gold by more, other countries would have
followed the U.S. action. Since speculators seem to consider
the dollar more than 10 percent overvalued from its May
1971 rates, the demands on the U.S. gold stock would have
forced suspension of gold-dollar convertibility. Finally, rais­
ing the price of gold would not have eased the balance-ofpayments financing burden, as was achieved with suspensions.
The overvalued position of the dollar developed gradu­
ally over a number of years. The pressure of events forced
things to a head in 1971. These events included: (1 ) in­
terest rate differentials and the after effects of the partial
suspension of Federal Reserve Regulation Q caused large
interest sensitive capital flows from the United States; (2 )
Page II

FEDERAL RESERVE BANK OF ST. LOUIS

it did not eliminate the overvalued dollar, this action
did put other countries in the position of financing 100
percent of future U.S. balance-of-payments deficits by
accepting U.S. dollar denominated liabilities without
the option of later converting part of them into gold
or other reserve assets of the United States.

The Dollar Standard — Post August 15
Breaking the link between the dollar and gold at
least temporarily converted the monetary system into
a pure dollar standard. Gold continues to be held as a
reserve asset by central banks. However, without the
United States actively supporting the price of gold by
acting as a residual supplier, other central banks are
not likely to be willing to sell or able to buy additional
gold at the new official price of $38 per ounce. Thus
gold represents an untradable reserve asset. The other
element of total reserves, SDRs, is too small a share of
the total to play an important role at this time.

Methods of Controlling Dollar Balances
With the suspension of gold convertibility, foreign
central banks are no longer able to make dollar-gold
portfolio adjustments to eliminate undesired dollar
balances.11 The only methods now available to adjust
holdings of dollars are changes in the price at which
they are prepared to buy dollars (that is, through
changing the exchange rates) or government imposi­
tion of capital controls.
Capital Controls —It is unlikely that many countries
will choose to control future dollar inflows via strin­
gent new capital controls. Most countries have a much
larger stake in world trade than the United States,
and a move away from the free flow of trade and
capital would be at enormous economic cost. The
actions (in contrast to the words) of most govern­
ments indicate that they are aware of this.
dock strikes made the U.S. trade balance appear worse
than it actually was in the months prior to August 1971; (3)
the German Economics Minister implied support of a floating
^
Deutsche mark, causing speculative flows which “forced
the Germans to float the Deutsche mark in May 1971.
n This change can also be stated from the U.S. point of
view. Under the old system, the United States would finance
part of its balance-of-payments deficit with gold and other
reserve assets and part with dollar liabilities. The proportion
financed in these ways depended upon the gold-dollar pref­
erences of those countries which were experiencing balanceof-payments surpluses. Under the present international dol­
lar standard, the United States has financed virtually all of
its balance-of-payments deficits with dollar liabilities and
none with gold. The exceptions to dollar financing are re­
lated to the use of the IMF gold tranche and the re-activa­
tion, as of July 19, 1972, of the central bank swap network.

http://fraser.stlouisfed.org/
Page 12
Federal Reserve Bank of St. Louis

AUGUST 1972

There has been an increase in exchange controls
since the establishment of the dollar standard. How­
ever, up until now (August 1972) they have not been
the type which would interfere with “normal” inter­
national trade and financial transactions. (See Wall
Street Journal, July 6, 1972.) These controls are
largely ineffective because speculators can “disguise”
their actions so they appear to be in normal trade and
financial form. The real test of controls will come when
governments face the issue of whether to close these
“loopholes.” If they do, then the progressive expansion
of controls will gradually have an adverse effect on
normal trade and financial transactions.
Exchange Rate Adjustments —If capital controls are
not extensively used, then exchange rate adjustment
is left as the dominant method of controlling dollar
balances for most industrial countries. In principle,
there are four methods by which exchange rate
changes could be achieved: (1) a permanent floating
exchange rate; (2) a temporary floating exchange
rate; (3) a dual exchange rate with a fixed rate for
commercial transactions and a floating rate for all
other transactions; and (4) more frequent changes
in a basically fixed exchange rate system.
The first option, a permanently floating exchange
rate, which is widely accepted by economists, has
thus far been rejected by the policymakers of almost
all industrial countries.12 Experience with floating ex­
change rates from August to December 1971 has con­
vinced them that “other governments” will see to it
that their export industries enjoy an “unfair” competi­
tive advantage by engaging in a “dirty float.” This has
meant that central bank intervention in foreign ex­
change markets has kept the national currency under­
valued relative to its equilibrium rate. If such fears
lead to competitive devaluations via “dirty float”, it
represents an unstable condition with respect to
achieving equilibrium exchange rates.
The second option (a temporary floating exchange
rate) has been used successfully by a number of gov­
ernments in recent years (most recently the United
Kingdom) when there was substantial speculation for
or against the national currency. It relieves specula­
tive pressure by allowing a change in the international
price of the currency, rather than a change in the
country’s international reserves, thus avoiding adverse
effects on domestic liquidity. This technique can in­
12For a discussion of the advantages of flexible rates, see
Darryl Francis, “The Flexible Exchange Rate: Gain or Loss
to the United States,” this Review (November 1971), and
Harry Johnson, “The Case for Flexible Exchange Rates,”
this Review (June 1969).

FEDERAL RESERVE BANK OF ST. LOUIS

crease the flexibility of the international system if
used by both surplus and deficit countries.
The third option, a dual exchange rate with a fixed
rate for commercial transactions and a floating rate
for all other international transactions, is now used by
the Belgians and the French. To the extent that the
rates are allowed to deviate only temporarily during
periods of speculation, it is very similar to the tem­
porary float described above, with a possible advan­
tage that commercial transactions may not be dis­
turbed. However, if the rates are expected to be
permanently apart, then there are incentives to incur
the cost necessary to arbitrage between the two “mar­
kets” for the national currency. Such arbitrage can
only be stopped by strict governmental supervision of
all transactions. As such, supervision could be as in­
hibiting as traditional exchange controls and will make
many countries reluctant to use this option.
The fourth option involves more frequent changes
in a basically fixed exchange rate system. This tech­
nique, along with wider bands to discourage interest
sensitive capital flows, was used as the basis of the
Smithsonian Agreements of December 1971 and cur­
rently seems to be the favored method of adjustment.

Preference for Exchange Rate Adjustments
Since the emergence of the dollar standard in Au­
gust 1971, most central banks have revealed a strong
preference to control dollar balances through exchange
rate changes rather than controls. In addition, virtually
all of the exchange rate changes have been in the di­
rection of eliminating under or overvalued currencies.
These actions alone attest to the superiority of the
dollar standard over the old system.
Before August 15, only Germany and Canada of
the major industrial countries were willing to see
their currencies appreciate against the dollar. Only a
few weeks after August 15, every major country was
willing to appreciate their national currency against
the dollar. The exchange rate changes were negotiated
in the setting of an international conference because
most currencies were undervalued against the dollar,
but were not necessarily undervalued with respect
to each other. A multilateral agreement could take
into account the effects of all the exchange rate
changes occurring simultaneously.
The important point about the Smithsonian episode
is that the realignment of exchange rates, however
justified by underlying economic conditions, would
not have occurred when it did if the United States
had not suspended gold convertibility. This put the



AUGUST 1972

world on a dollar standard and created incentives for
exchange rate changes.

Proposals for Change in the
International Monetary System
Judging by the comments of central bankers and
others involved in international finance, the recently
evolved dollar standard apparently is not considered
a suitable, permanent arrangement. The economic
rationale against a dollar standard held by foreign
central bankers is that inflation in the United States
since 1965 has generated expectations of continued
inflation.13 They appear to be reluctant to hold a
dominant portion of their international reserves in a
form which they expect to decline in real value in
the future.14
For the United States the present dollar standard
may be superior to the old gold-dollar standard. If
other countries continue to follow neo-mercantilist
policies of maintaining an undervalued national cur­
rency to encourage exports and balance-of-payments
surpluses, then the U.S. dollar will continue to be
overvalued, and the U.S. consumer will enjoy a sub­
sidy on foreign purchased goods. The resulting U.S.
payments deficits are financed almost completely with
dollar liabilities. This method of financing deficits
does not impose the type of balance-of-payments
constraint on U.S. policy actions which existed under
the old gold-dollar standard.
There is, however, a potential for U.S. dissatisfac­
tion with the dollar standard. If the dollar continues
to be overvalued there will continue to be a decline in
the relative size of U.S. export and import-competing
industries and a displacement of labor from those
industries. Some U.S. industries, which would be in­
ternationally competitive if the dollar were at its equi­
librium exchange rate, are not as competitive with an
overvalued dollar. The resulting distortion of interna­
tional trade implies a less than optimally efficient in­
ternational division of labor. An improvement in the
international division of labor would benefit not only
the United States, but all trading nations. In addition,
an overvalued dollar has led to increased Congres­
13There is another noneconomic argument which also con­
tributes to the reluctance of foreign central banks to con­
tinue on a dollar standard. A dollar standard places the
United States in a unique category which, with the rise in
the relative economic position of Europe and Japan, is no
longer justified.
14It would seem that if the interest rate on dollar denomi­
nated reserve assets included an appropriate inflation adjust­
ment, the central bank would feel compensated. However,
the evidence suggests such is not the case. See M. Keran,
“Demand for International Money —A Micro Approach.”
Page 13

FEDERAL RESERVE BANK OF ST. LOUIS

sional pressure for protectionist legislation which could
have adverse effects on world trade. Thus the United
States and the rest of the world would seem to have
an interest in international monetary reform to the
extent that it increased the probability of an equilib­
rium exchange rate for the dollar.
These observations indicate two criteria for a gen­
erally acceptable monetary reform:
(1) the concern of foreign central bankers about
possible future decline in the value of the dol­
lar should be satisfied, and
(2) the probability of equilibrium exchange rates for
the dollar should be improved.
Various proposals for monetary reform will be
analyzed on the basis of these criteria. The procedure
will be to consider the ways in which foreign central
bankers can be protected from a possible inflation in
the United States, and then whether these procedures
contribute to the achievement of equilibrium ex­
change rates.
There are three ways in which the concerns of cen­
tral bankers regarding a future decline in the value of
the dollar could be dealt with: first, by engendering
belief that there will be no future U.S. inflation; sec­
ond, by guaranteeing the real value of dollar de­
nominated international assets; and third, by reducing
or eliminating the dollar component of international
reserve assets.
The first solution is not feasible, because the U.S.
Government cannot guarantee that inflation will never
occur in the future. The second solution is not de­
sirable on the basis of the criteria mentioned above,
because it reduces the present incentive for equili­
brium exchange rate changes by other governments
without increasing the ability of the United States to
control unilaterally its exchange rate. If foreign cen­
tral banks had a guarantee that the real value of
their dollar reserves would be maintained irrespective
of the degree of U.S. inflation, the incentive on the
part of these governments to change their exchange
rates in the face of a large dollar inflow would be
reduced.15 The third solution would seem to be the
15This assumes that the maintenance of value is based on some
measure of U.S. price inflation or U.S. initiated exchange
rate change. If, however, the maintenance of value is based
on changes in the exchange rate taken at the initiative of
the foreign government, it might actually increase incentives
for equilibrium exchange rate changes. In this circumstance,
a country experiencing a balance-of-payments surplus and a
dollar inflow would benefit from the maintenance-of-value
option only if it appreciated its exchange rate. Any time a
country appreciated, the United States would stand ready
to increase the nominal value of its dollar reserves in proortion to the appreciation. Presumably, when a country
epreciated, the United States would have to reverse the
process. However, some governments which have experi­
enced chronic inflation, and therefore periodic devaluations,

http://fraser.stlouisfed.org/
Page 14
Federal Reserve Bank of St. Louis

AUGUST 1972

only economically viable one —reduce or eliminate
the dollar component of international reserves. This
would require eliminating the undesired dollar bal­
ances which accrued to central banks during 1971 and
would insure that future accruals of dollars do not
exceed desired holdings.

Eliminate Current Excess Dollar Holdings
This issue can, in the final analysis, be solved only
in the market place. If the exchange rates established
in December 1971 represent realistic values for the
major currencies in the sense of eliminating the over­
valued dollar, private speculators will come to accept
them as reasonably permanent. They will then take
their speculative profits by selling foreign denomi­
nated assets acquired in 1971 to the respective cen­
tral banks for dollars. This natural market force should
reduce dollar holdings of most central banks to more
desired levels.
If the December 1971 exchange rates have not con­
vinced speculators that the overvalued dollar has been
eliminated, then no amount of international negotia­
tion (other than on new exchange rates) will reduce
the excess holdings of international reserves of central
banks. An international conference could conceivably
negotiate a switch of reserves from dollar denomina­
tion to some other form, such as SDR denomination.
However, the only way the stock of reserves can be
reduced is if private speculators are convinced that
there are no more profits to be gained by continuing
to hold assets denominated in foreign currency, or if
the United States rims a sufficient surplus to officially
absorb the dollars. The elimination of the overvalued
dollar is the necessary and sufficient condition for
either development.

Prevent Future Excess Dollar Holdings
Most proposals for reforming the international
monetary system include as a key element the re­
newal of dollar convertibility into gold, SDRs, or both.
This is designed to allow central banks to control their
dollar holdings by converting undesired balances into
alternative reserve assets. There are basically only
two ways these proposals could be implemented. The
first would be a system similar to that which existed
prior to August 15 —either by returning to the golddollar system or by developing a new SDR-dollar
would be reluctant to participate in this type of arrange­
ment. They should have the option of, in effect, choosing
ahead of time whether or not to participate in the maintenance-of-value agreement. However, it is unlikely that this
procedure with its appearance of a “bribe” would ever be
agreed to.

AUGUST 1972

FEDERAL RESERVE BANK OF ST. LOUIS

system. The second method would require the virtual
elimination of the dollar from its reserve asset role,
with nearly all dollars acquired by central banks in
the foreign exchange market being immediately con­
verted into a truly international currency, unrelated
to any national currency.
Achievement of the first proposal would result in
the re-establishment of a multiunit international cur­
rency with the U.S. dollar continuing to be a major
element. In a world of “N” national currencies and
“N-l” exchange rates between those currencies, the
dollar would continue to be the “Nth” currency, the
exchange rate of which would be residually deter­
mined by the collective exchange rates of the “N-l”
currencies. The only unilateral action the United States
could take would be to change or suspend convertibil­
ity of the dollar into the other forms of international
currency. This would destroy the new system just as
U.S. action in August 1971 destroyed the old system.
In the case of the gold version of this system, it
would be identical to that which existed before Au­
gust 15 except that with the U.S. gold stock standing
at $10 billion, the life expectancy of such a system
would not be very long.16 The second version with
SDRs and dollars would presumably be negotiated
within the context of a substantial increase in SDR
balances for the United States. This would permit the
United States to act as a residual supplier of SDRs to
other central banks as they operated to achieve de­
sired portfolio ratios of SDRs and dollars.
This system could work only as long as inflation in
the United States ( or some other major change in the
structure of world trade) did not lead to an overvalued
dollar. However, it is possible that the dollar could
eventually become overvalued again, because a dollar-SDR system provides no mechanism to automati­
cally keep the international value of the dollar in line
with its domestic value. Central banks of countries
with balance-of-payments surpluses could adjust their
international reserve portfolios between dollars and
SDRs by means of U.S. convertibility of dollars to
SDRs. These central banks would have no incentive
to make the adjustment by changing the exchange
16The value of the U.S. gold stock could be increased if the
dollar price of gold were increased substantially. However,
as such a price change could only be accomplished by the
United States, it is not likely to occur. Although a higher
gold price may allow a dollar-gold system to last longer, it
would be at substantial economic cost to the United States.
Equally important, it would take away the present incentive
of other governments to make equilibrium exchange rate
changes. It would have the same problems as a dollar-SDR
standard as discussed in the text, plus future speculation
about another change in the dollar-gold conversion rate.



rate at which they would purchase dollars with their
national currency.
Neither a dollar-gold or a dollar-SDR system would
provide the present incentives for foreign governments
to make equilibrium changes in their exchange rates,
nor would either provide a greater opportunity for
the United States to unilaterally change its exchange
rate. Thus, these proposals would lead to a system
which is in these respects inferior to the present dollar
standard for the United States.
The only reform of the present dollar standard
which would provide incentive for equilibrium ex­
change rate changes (exclusive of freely floating ex­
change rates) would be a system based on a truly
international currency. The natural candidate for the
role of a truly international currency would be Special
Drawing Rights (SDRs) of the IMF.17 This is an
established and functioning system with $9.5 billion
already credited to the countries which are members
of the IMF.

Establishing a Truly International
Currency
SDRs
—

In a truly international currency system, the great
bulk of international reserves would be in the form
of SDRs and gold.18 Whether any minimum working
dollar balances were held by central banks would
depend upon the extent to which the dollar continued
to be used as a market intervention currency. In any
event, modest dollar holdings by central banks would
not be inconsistent with an SDR standard. The SDRgold conversion rate would be fixed and unchange­
able. There would be nothing to cause speculation
about a change in this conversion rate, because neither
unit is a national currency. All national currencies,
including the U.S. dollar, would define their exchange
rates in terms of SDRs.
There are a number of features which would make
such a system attractive. For the United States it has
the potential of increasing unilateral control over its
exchange rate, thus reducing dependence on other
countries making equilibrium exchange rate changes.
For the rest of the world, it implies U.S. financing of
17A pure gold standard would provide the same incentives
as an SDR standard. However, a gold standard is inferior
because the monetary supply of gold is constrained by
changes in gold mining tecnniques, commodity demand, and
discovery of new mine fields. None of these problems are
faced in an SDR standard because it is a pure fiduciary
currency, the supply of which is regulated by the member
countries of the IMF.
18A proposal along this line has been made by Andre Vlerick,
the Relgian Minister of Finance. See American Banker,
July 17, 1972.
Page 15

AUGUST 1972

FEDERAL RESERVE BANK OF ST. LOUIS

its balance-of-payments deficits with reserve assets
rather than increased dollar liabilities.
Advantages to the United States —The movement
from the present international currency based on the
dollar to one based on SDRs would solve the “Nth”
currency problem. With the SDR as the international
currency we would have, in effect, “N” national cur­
rencies and an “N + 1” international currency. This
would permit “N” independent prices or exchange
rates to exist. The United States could then change
its exchange rate with respect to SDRs. With SDRs
and gold as the dominant forms of reserve assets, and
dollars held to minimum working balances, a change
in the dollar-SDR exchange rate would leave an un­
changed conversion rate between the major compon­
ents of the international currency.

U.S. E xp orts a s a Percent o f W o r ld T ra d e *
P e rce n t

4

P e rce n t

/\ A
/ \l \

20

18

.

A n n u a l D ata

/

20

\

A,

18

N.

V,

16

\

14
0

.....................
19 5 0

1955

1960

16

14

......................................
1965
1970
S o u r c e : In t e r n a t i o n a l M o n e t a r y F u n d

lo t h U .S . a n d W o r l d E x p o r t d a t a a r e re p o rt e d o n a n f o b b a s i s a n d e x c lu d e

The United States would be in the same position
as other countries in being able to change unilaterally
its exchange rate and thus move towards eliminating
an overvalued national currency. This would not only
help the United States achieve balance-of-payments
equilibrium, but by indirectly reducing the underval­
ued position of other national currencies, it would help
achieve world-wide balance-of-payments equilibrium.
The United States still would not have complete
control over its exchange rate. No country ever does.
There are two factors which have inhibited this con­
trol in the past. The first is “monetary” —the dollar’s
role as an international money —and the second is
“real” —the size of the United States in world trade.
An international monetary reform could deal with the
first issue, but not with the second. However, this
“real” factor constrains every country to some extent
and is not unique to the United States. Just as coun­
tries with large and important trade ties with the
United Kingdom generally follow sterling in any
exchange rate change, so some other countries would
follow the dollar in any exchange rate change.
These “real” constraints on U.S. control of its ex­
change rate have been steadily declining because of
a decline in the relative importance of the United
States in world trade. The U.S. share of world exports
fell from 21 percent in 1953 to 18 percent in 1960
and to 14.7 percent in 1971. Although the rate of
decline may be reduced by a correction of U.S.
inflation, its direction will undoubtedly continue to
be downward for some years to come unless present
trade talks are successful in blunting the effects of
preferential trading blocks, such as the enlarged Com­
mon Market.
Advantages to Others —A major criticism by for­
eign governments of the old monetary system was the

http://fraser.stlouisfed.org/
Page 16
Federal Reserve Bank of St. Louis

S e r v ic e E x p o rts.
.a t e s t d a t a p lo t t e d : 1971

asymmetry which allowed the United States to par­
tially finance balance-of-payments deficits by increas­
ing dollar liabilities while foreign governments had to
draw down international reserve assets. This asym­
metry which appeared to be to the advantage of the
United States was, of course, matched by a parallel
asymmetry which was to the disadvantage of the
United States —the requirement that the United
States refrain from changing its exchange rate. It is in
the nature of the system that when a domestic cur­
rency is also used as the international money, that
country will have both the inability to control its own
exchange rate and the ability to finance at least part
of its payments deficits with increased liabilities rather
than decreased assets.
If SDRs replaced the dollar as the international cur­
rency, then the United States would be on the same
footing as other countries.10 Dollars which came into
the hands of foreign commercial or central banks
would automatically be converted into SDRs at the
19The use of SDRs as an international currency would also
make the United States similar to other countries in terms of
domestic monetary policy control. With the dollar as the
international currency and the Federal Reserve free from
intervening in the foreign exchange market, domestic mone­
tary actions are not impeded by international trade and
capital flows. The fixed exchange rate for the United
States is maintained by other countries. Under an SDR
regime, the Federal Reserve would have to intervene in the
foreign exchange market to maintain the fixed exchange rate
for the dollar. The purchases and sales of SDRs in the
foreign exchange market would have the same effect on the
monetary base and bank reserves as purchases and sales of
Treasury bills from open market operations.
This could be offset by the Federal Reserve in the short
run. However, as world capital markets become more in­
tegrated, it will be progressively more difficult for central
banks in general, and the Federal Reserve in particular, to
isolate the domestic monetary policy from international capi­
tal flows under a regime of fixed exchange rates.

FEDERAL RESERVE BANK OF ST. LOUIS

Federal Reserve. Thus U.S. payments deficits would
be financed completely with reserve assets.

Problems of an SDR Standard
An SDR standard would be a new and untried
system. Governments and central banks would natu­
rally be reluctant to place something as important as
the international monetary standard in an untested
framework. One cannot anticipate all the problems
which would emerge from the system until it is actu­
ally implemented. However, some foreseeable prob­
lems can be grouped under the headings “Supply”
and “Demand” for SDR.
Supply —On the supply side, the potential prob­
lem can be stated simply. What would insure that the
issuance of SDRs would be such as not to contribute
to the present international inflation? The SDR is a
fiat currency, the supply of which is determined by
the issuing agency. In this case, the agency is the
International Monetary Fund (IMF) and the amount
of each issue is decided by an 85 percent weighted
vote of the member countries. This, in effect, gives a
veto power to the United States and also to the Com­
mon Market countries (if they vote as a group). In
deed, any group of countries with 15 percent of the
weighted votes can veto an increase in an SDR issue,
which would seem to eliminate any inflationary bias
in an SDR standard.

AUGUST 1972

At present, there are no incentives to induce for­
eign central banks to hold SDRs rather than dollars.
Dollar reserves are held in interest-earning form, such
as Treasury bills or certificates of deposit. SDRs, on
the other hand, pay only a nominal 1.5 percent rate of
interest. In addition, the dollar would be in constant
demand in the foreign exchange market as the inter­
vention currency by central banks and as the trading
currency for private international transactions. Both
factors would create incentives for central banks to
increase their share of reserves in dollar form versus
SDR form over time. Indeed, this is how the dollar
assumed its reserve asset role in the old monetary
system.
There are two obvious steps which could be taken
to make SDRs more attractive for central bankers to
hold. First, the interest rate on SDRs could be in­
creased to be more competitive with dollar assets.
Second, central bank intervention in the foreign ex­
change market could be switched from dollars to
SDRs. If the interest rate on SDRs could be increased
substantially, then perhaps only this action need be
taken. However, if for institutional reasons the interest
rate on SDRs can be raised only marginally, perhaps
SDRs would also have to replace dollars as the central
bank intervention currency in the foreign exchange
markets. This would also require commercial banks to
hold SDRs. One method of implementing these pro­
posals is described in the appendix.

Demand —On the demand side, the problems are
somewhat more complex, but they all come down to
one issue, how to induce central banks to actually
hold a significant share of reserves in the form of SDRs.
It is not sufficient to just change the IMF articles
of agreement making SDRs rather than gold the de
jure currency. As long as the dollar continues to be
the de facto international currency, the world will
continue to be on a dollar standard.

With gold or SDRs making up the bulk of inter­
national reserves and dollars reduced to minimum
working balances, a change in the U.S. exchange rate
would not affect central banks as in the past. Central
banks would not incur an accounting loss on the do­
mestic value of their international reserves or a finan­
cial loss through a decline in the international purchas­
ing power of their reserves if the dollar was devalued.

The movement to an SDR standard would represent
a substantial change in the working practices of cen­
tral banks and perhaps of commercial banks and oth­
ers in international trade and finance. The SDR is a
relatively new concept and has been in active use by
central banks for less than three years. An interna­
tional SDR standard would represent a major expan­
sion from its present use, and the confidence which
comes with years of use and experience with a facility
is not present in this case. Thus, to make SDRs actually
replace dollars in international use, institutional
changes would be required to encourage the use of
SDRs at the expense of dollars. Only when central
banks actually have been induced to hold SDRs in­
stead of dollars would the SDR standard be operative.

In this SDR world, if private speculators anticipated
a devaluation of the dollar, they would attempt to
convert their dollar balances into SDRs at the Federal
Reserve ( or if the dollar continued to be an interven­
tion currency, at other central banks). The United
States could react to this development with almost
the same array of alternative actions as any other
government. If it was considered that private specula­
tion reflected a “true” overvaluation of the dollar, the
United States could devalue with respect to the new
international numeraire —SDRs. If the dollar was not
considered to be overvalued, the United States could
arrange special credit facilities with the IMF or other
governments to satisfy the speculative demand.




Page 17

AUGUST 1972

FEDERAL. RESERVE BANK OF ST. LOUIS

One action the United States could not take which
other governments could, would be to suspend con­
vertibility of the dollar into SDRs. As long as the
dollar continued to be the intervention currency of
other central banks, dollar-SDR convertibility would
be needed to insure that foreign central bank dollar
holdings did not increase beyond minimum working
balances.

Conclusions
The intention of this article has been to analyze the
role of an international currency with respect to its
contribution to achieving international monetary sta­
bility. The article assumes that the international sys­
tem advocated by most economists —freely floating
exchange rates —is not applied, and thus looks at the
effects which alternative forms of international cur­
rency can have on promoting equilibrium exchange
rate changes within a system of basically fixed rates.
The old monetary system based on a multiunit
dollar-gold international currency was deficient in that
it did not encourage foreign governments to make
equilibrium exchange rate changes and it inhibited
the United States from taking such actions. This rigid­
ity in the face of changing economic conditions made
existing exchange rates increasingly unrealistic. As a
result, private speculation reflected in massive short­
term capital flows became increasingly frequent oc­
currences in the second half of the 1960s, culminating
in the breakdown of the old system in 1971.
The international dollar standard which emerged
in 1971 is superior to the old system in encouraging
other governments to make equilibrium exchange rate
changes, even though it still inhibits the United States

from making exchange rate adjustments. The Smith­
sonian Agreements attest to the success of the dollar
standard in that direction. The act of the United States
in changing the conversion rate between the dollar
and gold from $35 to $38 per ounce has no economic
meaning as long as the dollar remains inconvertible
into gold.
Discussions of the need for further changes in the
international monetary system are largely based on
foreign dissatisfaction with a single-unit dollar stand­
ard. This is primarily due to expectations of further in­
flation in the United States which would reduce the
real value of international reserves denominated in
dollars.
Given the reluctance of major central bankers to
accept freely floating exchange rates, the great chal­
lenge for international monetary reform is to devise a
system which would simultaneously encourage govern­
ments to make exchange rate adjustments and satisfy
the legitimate concerns of foreign central bankers
about holding excessive dollar balances. One way to
accomplish these two goals would be the establish­
ment of a truly international currency unrelated to the
national currency of any country. This would permit
the United States more control over its own exchange
rate without having to rely on the actions of other
countries to eliminate an overvalued dollar and, at
the same time, insure that dollars accruing to foreign
central banks would “automatically” be converted
into the new international currency. A modified form
of the present Special Drawing Rights (SDRs) of the
IMF is a natural candidate for this truly international
currency function if the real problems associated with
an SDR standard can be overcome.

APPENDIX
Substituting SDRs for Dollars
as the International Numeraire
To replace dollars with SDRs as the international
numeraire, it would be necessary to make it more at­
tractive to hold SDRs. Two proposals along this line are
(1) increase interest payments on SDRs and (2) make
SDRs the intervention currency in the foreign exchange
market. This appendix considers how these proposals
could be implemented.
Interest Payments on SDRs —There would be two
ways to increase SDRs and two sources of interest

Page 18


payments.1 The first source would be the conversion of
presently outstanding dollar assets in the hands of foreign
central banks into SDR assets. The dollar assets could
revert directly to the United States or to the IMF. In
either case, the interest rate which would normally have
1A number of studies have indicated the theoretical desirability

of increased interest payments on SDRs. It is postulated that
it would increase the demand for international reserves with­
out adding to the pressures for world inflation. With larger
reserves, temporary deficits in the balance of payments can
be met with smaller and more time-consuming adjustment
actions which would reduce economic instability and resource
misallocation.

FEDERAL RESERVE BANK OF ST. LOUIS

been paid on the dollar assets could be used to pay
interest on the SDRs. If the dollars reverted to the U.S.
Treasury and were replaced by a special SDR issue, the
interest on Treasury bills would be reduced and that on
special SDRs increased by an equal amount. If the dol­
lars were transferred to the IMF, the Treasury bills
would still be outstanding and the interest payments
made to the IM F, which in turn would credit it to mem­
ber countries in proportion to their SDR holdings. In
either case, the interest cost to the U.S. Treasury would
probably be no greater than it is now.2
The second source of increased SDRs would be the
annual addition to total reserve assets needed to meet
the demand for international reserves of central banks.
The amount of the interest payment must be the same
on all forms of SDRs, otherwise either a “black market”
would develop with the price on low earning SDRs
lower than on high earning SDRs, or Gresham’s law
might operate. Low earning SDRs would be used to
settle international accounts and high earning SDRs
would be held for income purposes. The method of inter­
est payments could continue as at present on net addi­
tions of SDRs, with countries paying interest on SDRs
according to their allocation and receiving interest ac­
cording to their holdings.

SDRs as an In terv en tion C u rren cy —A major in­
centive for foreign central banks to hold dollars as a
reserve asset is its role as an intervention currency in
the foreign exchange market. The standard way to control
the international value of the domestic currency is to
buy and sell dollars in the private foreign exchange
market at a fixed price with respect to the domestic
2There are auxiliary issues which must be dealt with. (1 )
Would the interest rate earned on SDRs be just equal to
that earned on Treasury bills? (2 ) Would the interest pay­
ment be denominated in dollars or SDRs?
With respect to the interest rate, a case can be made that
it should be lower on SDRs than dollars because SDRs are
a less risky asset. It is not subject to change in value by the
unilateral act of one government. This same argument would
imply that interest payments should also be in SDRs. This, in
effect, would extend the maintenance of value protection
implicit in originally exchanging SDRs for dollars, to the in­
terest income on the assets.
If the U.S. Treasury paid interest in SDRs rather than
dollars, the effect on the balance of payments would be
almost identical. In the SDR case, the Treasury would have
to acquire SDRs in the exchange markets before the interest
payments were due. In the dollar case, the Treasury would
not have to settle in SDRs until after the interest payment
was made. In either case, the United States would have to
run a proportionately larger balance-of-payments surplus to
pay the interest cost.
The only difference in the two cases is that in the former
the U.S. Treasury takes the exchange risk, and in the latter
case the foreign country takes the exchange risk. It would be
in keeping with the standard market practices for the U.S.
Treasury to pay a somewhat lower interest rate if it absorbed
the exchange risk by making interest payments in SDRs.




AUGUST 1972

currency. When almost all countries do this, they, in the
aggregate, determine the international value of the dollar.
Although it is possible for the SDR to be the major
reserve asset without also being the intervention cur­
rency, SDRs would be in greater demand if they per­
formed both functions. First, it would reduce transactions
costs if SDRs did not have to be swapped for dollars
when intervention is conducted in the foreign exchange
market. There are problems of delay, and sources of
dollar balances that could be avoided if the SDR were
the intervention currency. Second, if the dollar itself
were fluctuating within its band in the foreign exchange
market, there would be serious problems of valuating
transactions within the market and distributing the ex­
change risk of dealing with a variable dollar.
From the point of view of central banks dealing with
other central banks, the IMF, the Federal Reserve, and
the private foreign exchange markets, it is less costly and
less risky to have the reserve asset also used as the
intervention currency.
Achieving an intervention role for SDRs would not
be easy. Such a role would require commercial banks
to hold working balances of SDRs. They would be re­
luctant to take this step given the present low interest
rates. An alternative to paying a higher interest rate
would be to introduce an administrative rule that all
foreign exchange transactions between the public and the
central bank must be in SDRs or the domestic currency.
Commercial banks which want the advantages of ac­
quiring foreign exchange at the official price would have
to hold SDRs. When the United States has a balance-ofpayments deficit, the excess supply of dollars which
would accrue to foreign commercial banks would be
converted into SDRs at the Federal Reserve, which
would be the only central bank allowed to deal in dol­
lars. The SDRs, in turn, could be sold by these commer­
cial banks to their national central bank for domestic
currency. The converse would happen when the United
States had a balance-of-payments surplus.
The United States could control its exchange rate by
standing ready to buy and sell dollars for SDRs at what­
ever rate it unilaterally established. The United States
could choose to deal with a deficit either by changing
the price at which SDRs were exchanged for dollars, or
by shifting the international supply and demand sched­
ules for dollars through domestic monetary and fiscal
actions.
If the dollar continued to be the intervention currency,
foreign central banks would have to take action explicitly
to validate the new dollar-SDR exchange rate by delib­
erately changing the rate at which they exchange dollars
for domestic currency. With the SDR as the interven­
tion currency, this explicit “cooperation” by other central
banks would not be required.

T his article is av a ila ble as R eprint No. 78.

Page 19

SUBSCRIPTIONS to this Bank’s

R e v ie w

are available to the public without

charge, including bulk mailings to banks, business organizations, educational
institutions, and others. For information write: Research Department, Federal
Reserve Bank o f St. Louis, P. O. Box 442, St. Louis, Missouri 63166.