Full text of Economic Policy Review (Federal Reserve Bank of New York) : Winter 1982, Vol. 7, No. 4

View original document

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

Federal
Reserve Bankof
NewYbrk
Quarterly Review

W inter 1982-83

Volum e 7 No. 4

2
10
16

M oney and C redit: E xploring A lternatives
I C redit A ggregates as P olicy Targets
II Control of a C redit A ggregate
III Reserves against Debits

Im pact of IRAs on Saving

Econom ic Effects of Enforcing
Due-on-Sale Clauses

New Y ork C ity’s Property Tax

C hina’s Rapid Trade G rowth and Im pact
on the W orld Econom y

Treasury and Federal Reserve Foreign
Exchange O perations

The Quarterly Review is published by
the Research and Statistics Function
of the Federal Reserve Bank of New
York. Among the members of the staff
who contributed to this issue are
JAMES FACKLER and ANDREW
SILVER (on credit aggregates as
policy targets, page 2); MARCELLE
ARAK (on control of a credit aggre
gate, page 10); JOHN WENNINGER
(on reserves against debits, page 16);
ROBIN C. DeMAGISTRIS and CARL J.
PALASH (on the impact of IRAs on
saving, page 24); HOWARD ESAKI (on
the economic effects of enforcing
due-on-sale clauses, page 33); MARK
A. WILLIS and DANIEL CHALL (on
New York City’s property tax prob
lems, page 37); SUSAN A. HICKOK
and ROSANNA ARGUELLES (on
China’s rapid trade growth and impact
on the world economy, page 41).
An interim report of Treasury and
Federal Reserve foreign exchange
operations for the period of August
through October 1982 begins on
page 52.

Money and Credit:
Exploring Alternatives

Rapid and sweeping changes during the past few years
in financial instruments available to the public have
raised questions about basing monetary policy on a
narrow monetary aggregate (M-1). Already, past rela
tionships between narrow money and the economy at
large have started breaking down. In the future, the
relationships are likely to change even more. Because
of recent regulatory changes, an expanding portion of
narrow money pays interest, and over time a large por
tion may pay interest virtually at money market rates.
Thus, narrow money has increasingly become a hybrid,
including some balances that are used primarily to
make transactions and other balances used primarily
as savings instruments. Of course, other hybrid invest
ment instruments are also available (notably money
market funds and the money market deposit account).
They are instruments with some transactions features,
though they are outside the definition of narrow money.
If narrow money as presently defined and measured
threatens to become an increasingly unhelpful aggre
gate for policy purposes, the question naturally arises

whether monetary policy should focus on something
else over the longer term. This issue of the Quarterly
Review begins with three articles that seek to analyze
some aspects of that question. James Fackler and
Andrew Silver have taken a new look at the evidence
on the links between measures of credit and the econ
omy. Marcelle Arak has asked what means could be
applied to attain targets for credit aggregates. And
John Wenninger has explored the case for putting re
serve requirements on debits to transactions accounts,
as well as on the accounts themselves— an idea which
first surfaced over a half a century ago, but which may
have relevance for future monetary policy.
To a considerable extent, these articles step back
from the discussion of current monetary policy. Rather,
they are an attempt to inform policy debates that may
emerge somewhere down the road. Obviously, they do
not begin to exhaust the possibilities for further work
on these topics. But they do raise some new themes
and bring some fresh insights into subjects that will
require increasing attention.

FRBNY Quarterly Review/Winter 1982-83

Money and Credit: Exploring Alternatives

Credit Aggregates as
Policy Targets

In recent years, financial innovations have raised ques
tions about the usefulness of narrowly defined money
as a policy target. The establishment of nationwide
NOW (negotiable order of withdrawal) accounts, the
rapid growth of money market mutual funds, and,
more recently, the creation of Federally insured money
market deposit and super-NOW accounts have repeat
edly altered the form in which the public holds trans
actions and savings balances. Because of the chang
ing composition of the monetary aggregates, the rela
tionship between GNP and those aggregates— espe
cially narrowly defined money— have become less
certain. As a result, the usefulness of monetary mea
sures in general, and M-1 in particular, has declined
in recent years. Furthermore, the difficulty with money
likely will continue during the transition period of ad
justment to the new accounts.
In such an environment, it may be worthwhile for
policymakers to look at other financial measures in
addition to the monetary aggregates. Credit aggregates
are one kind of potential alternative. This article sum
marizes existing evidence on the relationship between
GNP and credit. In addition, it investigates how the
reaction of policymakers might have been different in
the past if the Federal Reserve had targeted credit
along with money. Such an analysis can provide some
clues to the effectiveness of monetary policy in the
future if a credit aggregate is added as a target.
Our main finding is that there is no evidence from
the past suggesting that weighting credit more heavily
in the policymaking process would have resulted, on
balance, in a path of GNP more desirable than the one
actually experienced. However, this result should be

FRBNY Quarterly Review/Winter 1982-83

viewed cautiously with regard to its implications for
the future. Because the economy can change as a re
sult of financial innovations, past relationships may not
be reliable indicators of current or future behavior.
Moreover, correlations also can change just because
the Federal Reserve places more emphasis on a vari
able than it did in the past. Furthermore, conceptual
considerations suggest that, in times of substantial
shocks to the components of broad monetary aggre
gates, targeting credit or broad money instead of nar
row money may result in smaller fluctuations in GNP.

Conceptual issues
Until recently, conceptual models investigating what
variables monetary policy should target consid
ered only the choice between targeting interest rates
and the money stock. More recently, a number of
analysts have pointed out that one should also con
sider aggregates from the liability side (i.e., credit
aggregates) of the balance sheet of the nonbank pub
lic. However, the few models developed in the past
several years that do incorporate credit unfortunately
have not been able to provide any general conclusions
on whether money, credit, or the interest rate is best
suited for policy targets.1 The general problem is that
in some situations economic theory suggests that
’ See, for example, Franco Modigliani and Lucas Papademos, “ The
Structure of Financial Markets and the Monetary Mechanism”
(Columbia University Discussion Paper No. 90, February 1981), and
Andrew Silver, “ Choosing Among Narrow and Broad Monetary and
Credit Aggregates: An Evaluation of ‘The Structure of Financial
Markets and the Monetary Mechanism’, by Modigliani and
Papademos” (Federal Reserve Bank of New York Research Paper
No. 8110, June 1981).

money would be superior to credit as a policy target;
in other situations, however, credit would be preferred.
Thus, to choose optimal policies, it is necessary to
understand which of the various theoretical situations
best approximates the economy and then to tailor policy
accordingly. For example, assume that the only un
certainties about the economy were the random shifts
that could occur between two components of a broad
monetary aggregate (say, between M-1 and money
market mutual funds). Then, controlling one of those
components (e.g., M-1) would be inferior to controlling
a credit aggregate or a broad money aggregate, both
of which would be stable under this assumption. On
the other hand, suppose the demand for money were
stable and interest inelastic but the level of credit
demand was uncertain (due, e.g., to shifts in invest
ment spending between firms that typically depend ex
tensively on credit and those that do not). In this case,
controlling M-1 would be better than controlling credit.2
The obvious problem for the policymaker in inter
preting these results is that the actual structure of
the economy cannot be categorized easily as belong
ing to either of these special cases or, for that matter,
to any special case. In a world in which random
shocks occur to the components of money and credit
demand, as well as to other sectors, the policymaker
needs to know the expected magnitudes of all those
random occurrences. In addition, it is necessary to
know how all economic agents respond to changes in
the economy to understand how the shocks are trans
mitted from market to market. Unfortunately, existing
empirical models are not complete enough to provide
the necessary estimates.
Despite lack of such precise knowledge, if it is true
that shocks to the monetary measures resulting from
financial innovations are increasing relative to those
affecting the credit aggregates, then this relative in
crease in uncertainty about the monetary aggregates
probably moves us closer to a situation in which a
credit target may be superior to a money target. In
such a situation, attention to a credit aggregate in
addition to the usual concern with the money aggre
gates may guide us to policies that are better (or at
least no worse) than policies that focus exclusively
on money.

Empirical evidence
It is apparent from the previous discussion that the
case for a credit target, either instead of or in addi
tion to a money target, is not strong enough to be
made exclusively on conceptual grounds. Even if we
3 Silver, “ Choosing Among Narrow and Broad Monetary and
Credit Aggregates".

fully understood qualitatively how disturbances in any
particular market were transmitted to other markets in
the economy, it would still be necessary to estimate
empirically the magnitudes of various transmission
mechanisms. In view of the complexities inherent in
the economy, such a detailed estimation of the struc
ture of the economy is extremely difficult.
An alternative to trying to identify and to estimate
the relevant links among the various markets for
money, credit, and output is to focus an empirical
analysis on a very small subset of economic variables.
This approach has been used recently in numerous
studies in an attempt to sort out the impacts of money
and credit on output.
The results from these “ reduced-form” studies are
very sensitive to the time period under consideration,
the particular sets of variables in the analysis, and the
form in which the data are analyzed.3 Consequently,
these studies often show conflicting results. In some,
credit measures do better in explaining movements
in nominal GNP, or prices, or real output than do
monetary measures, while in others, money does better.
Furthermore, interest rates were shown to affect the
relationships among the variables in ways that raise
the question of whether either money or credit is linked
directly to the ultimate goals of monetary policy— price
stability and output growth. The basic conclusion from
these empirical findings, therefore, is a rather weak
one: because of the lack of a consistent and durable
set of empirical results, the relative usefulness of
money and credit as policy targets cannot be deter
mined on the basis of such evidence. We will now re
view some of these results, and the underlying meth
odologies, that have led us to this conclusion.
Much of the discussion on the use of credit aggre
gates has concentrated upon three different financial
measures: bank credit, formerly an associated pol
icy target; the debt proxy, an aggregate recommended
by Henry Kaufman;4 and nonfinancial domestic credit,
a variable proposed by Benjamin Friedman.5 Bank
credit is the narrowest measure of the three and con
sists of loans and investments of commercial banks;
the debt proxy represents financial claims held by the

3 For example, different results are obtained if one uses end-of-quarter
data or averages of adjacent end-of-quarter data. See E.K. Offenbacher,
R.D. Porter, and E.F. McKelvey, "Empirical Comparisons of Credit and
Monetary Aggregates Using Vector Autoregression Methods” , mimeo
graphed (Board of Governors of the Federal Reserve System,
July 1982).
4 Henry Kaufman, Testimony before the House of Representatives
Committee on the Budget, February 6,1978.
5 Benjamin Friedman, "Time to Reexamine the Monetary Targets
Framework” , New England Economic Review (Federal Reserve Bank
of Boston, March/April, 1982).

FRBNY Quarterly Review/Winter 1982-83

nonfinancial domestic sectors; and nonfinancial do
mestic credit represents credit market funds raised by
all domestic nonfinancial sectors, including local, state,
and Federal government units .6 Detailed definitions of
these aggregates are provided in the box on page 6.
The point of departure for a variety of studies in
vestigating the credit-GNP relationship 7 is the striking
constancy of the ratio of income to each of several
broad credit aggregates, i.e., constant income veloci
ties (chart). Indeed, at face value, this constancy sug
gests that GNP can be accurately controlled if
sufficiently close control over credit is maintained.
Nonetheless, a velocity which is trending but predict
able would be just as helpful to the policymaker as
a velocity which is constant. What is important from
a policy viewpoint are the sizes and predictability
of the fluctuations around the trend of velocity. How
ever, when one adjusts the velocities for their trends
and respective means, there is not much difference
in the variability of the financial aggregates under
discussion (Table 1).8
Since the income velocity of a financial aggregate
is the ratio of nominal GNP in a particular period
to the value of the financial aggregate in that same
period, a further shortcoming of an analysis of veloci
ties is that time lags are ignored. To allow for time
lags in a simple way, researchers have run equations
in which GNP is regressed on both current and
lagged values of the financial variables of interest.
These types of relationships between GNP and M-1,
as well as those between GNP and the credit aggre
gates similar to those described above, have been
investigated previously by Richard G. Davis.9 Each pair
of variables was analyzed by regressing GNP growth
on a weighted average of current and four lagged
growth rates of the respective financial aggregates
and by regressing GNP growth on a weighted average
of only the lagged growth rates of the various financial
variables. The latter type of regression, which excludes
values of the financial aggregates contemporaneous to
the dependent variable, is used to reduce the ambigu6 Nonfinancial domestic credit measures the liabilities of particular
sectors, while the debt proxy and bank credit measure the assets of
various agents. Nonetheless, for the purposes of this article, we will
refer to all three of the aggregates as "credit" aggregates.
7

Specifically, Richard G. Davis, "Broad Credit Measures as Targets
for Monetary Policy", this Quarterly Review (Summer 1979); Frank E.
Morris, “ Do the Monetary Aggregates Have a Future as Targets of
Federal Reserve Policy?" New England Economic Review (Federal
Reserve Bank of Boston, March/April 1982); and Benjamin Friedman,
‘‘Time to Reexamine the Monetary Targets Framework” .

A useful graphical comparison of variability in the M-1 and debt proxy
velocities, for example, is in Davis, "Broad Credit Measures as Targets
for Monetary Policy” , page 17.

Davis, "Broad Credit Measures as Targets for Monetary Policy".

FRBNY Quarterly Review/Winter 1982-83

Table 1

Coefficients of Variation of Velocities:
Money and Credit Aggregates*
1960-1 through 1982-1

Velocities
M-1 ......................................
Bank credit ...........................
Debt proxy ...........................
Nonfinancial c r e d it .............

Raw
data

Detrended
data

0.205
0.058
0.015
0.013

1 .2 0

1.32
1.17
1 .2 2

* The coefficient of variation is defined as the standard deviation
divided by the mean.

ities with regard to the “ causality” in the relationships
between the variables; thus, attention is focused upon
whether current GNP movements are determined strictly
by prior movements in the explanatory variable.
In his analysis, Davis studied the 1961-1 through
1977-IV period as well as the 1961-1 through 1969-11 and
1969-111 through 1977-IV subperiods. He found that, when
contemporaneous values of the financial aggregates
were included, the debt proxy provided the most ex
planatory power for GNP (a multiple correlation co
efficient of 0.35) over the full sample period. M-1 pro
vided slightly less explanatory power, followed by
total credit 10 and bank credit (which explained only 4
percent of the GNP variation). In each of the two sub
periods, the debt proxy also explained a substantial
portion of GNP movements; however, total credit pro
vided the most explanatory power in the 1970s (39
percent).
Benjamin Friedman reports similar results from in
vestigating the relationships between GNP and various
credit and monetary aggregates in similar types of equa
tions; broad credit aggregates do about as well as M-1
in explaining GNP.11 Friedman found, however, that the
explanatory power of all the aggregates, including the
broad credit aggregates, declined sharply in the 1970s.12

Davis's measure of total credit is similar to what we have called
nonfinancial domestic credit, but his measure excludes Federal
Government borrowing.

Benjamin Friedman, "The Roles of Money and Credit in Macroeconomic Analysis” , mimeographed, September 1981.

In contrast, Davis found that the explanatory power of total credit rose
in the 1970s as compared with the 1960s and that the explanatory
power of the debt proxy was about the same in the two periods.
Possible explanations for this discrepancy include the form of the lag
distribution, presence of other explanatory variables, and different
estimation periods and definitions of variables.

Further, it has been found that excluding the con
temporaneous values of credit aggregates from the
regressions results in a sharp drop in the ability of
these measures to explain income .13 Analogous results
were found by Davis for total credit as well as for M-1
and the debt proxy. When only lagged values of these
financial aggregates were used, the aggregates’ ex
planatory power for GNP declined dramatically in the
1970s.
This type of equation has been criticized by econo
metricians. In particular, in these equations the tem
poral relations between GNP and the financial aggre
gates are constrained to run from money or credit to
GNP. Of course, such a constraint may not be valid; it
may be that GNP changes, in fact, precede credit
changes or that the two are jointly determined. If the
data are inappropriately constrained, then unreliable
estimates of the financial aggregate-GNP relationship
may emerge.14
In part due to the potential problems associated with
inappropriately constraining the temporal relations
among variables, a new statistical technique— vector
13

James Fackler and Ken Guentner, “ Money, Credit and Income” ,
mimeographed (Federal Reserve Bank of New York, January 20, 1982).

For an exposition of the biases inherent in ignoring feedback from
the economy to policy variables in the context of a structural model,
see S.M. Goldfeld and A.S. Blinder, “ Some Implications of Endog
enous Stabilization Policy” , Brookings Papers on Economic Activity
(3, 1973).

autoregression— recently was developed to help avoid
imposing such incorrect constraints. This technique is
in many ways a natural extension of the simpler method
ology referred to above. The technique allows each
variable to depend potentially upon prior values of all
the variables under analysis .15 For example, GNP would
be regressed on prior values of itself as well as prior
values of the financial variable(s) of interest. Davis also
ran equations of this type .16 He found that, when lagged
values of either M-1 or the debt proxy were added to
lagged values of GNP, the explanatory power for cur
rent GNP improved significantly. Neither bank credit nor
total credit, however, provided similar contributions in
explanatory power.
Other analyses using the general vector autoregres
sion technique often distinguish explicitly between the
real and price components of nominal GNP. This dis
tinction between the real and price components allows
for analysis of issues such as whether financial aggre
gates directly influence real output or prices, or both.
For example, if it is found that some financial aggregate
is linked directly to prices, but not to output, then that
aggregate might be a candidate as an intermediate
15

Vector autoregression is so named because it investigates the relation
ship between a set of current variables (i.e., a vector) and prior values
of that set of variables (i.e., an autoregression).

16-However,

Davis did not provide a full vector autoregression analysis
since analogous equations explaining the financial aggregates were
not simultaneously estimated.

Income V e lo c itie s of M-1 and Various C re d it A ggregates
Velocity
7
6
5
4
3
GNP/Bank credi t

2
GNP/Debt proxy
1

I I i l l II I II l I I
1960
1962

-LL

1964

.
.
,
| GNP/Nonfi nanci al domest i c cr edi t .
i
i
i
i
i
i
I
i i i 111 l l .i I I l I I l l l I I I I l I I i I i l i I I I i i I L i i I I l l i I I I i l l l l l l l l I l l l l l I l l l l l I l l I

1966

1968

1970

1972

1974

1976

1978

1980

1982

Sources: Board of Governors of the Federal Reserve System; United States Department of Commerce,
Bureau of Economic Analysis.

FRBNY Quarterly Review/Winter 1982-83

prices, however, neither set is individually significant.
This suggests that M-1 and credit contain similar in
formation for future prices, and therefore both are not
necessary to explain prices. On the other hand, when
M-1 and credit are included in an equation explaining
output, both financial measures are significant.19 There
fore, information on money and credit possibly can
be combined usefully in determining future output.
However, when the rate of interest is substituted for
credit in a four-variable analysis (i.e., output, prices,
money, interest rate), money is no longer directly im
portant for explaining output.20 Money, however, does
retain importance for explaining prices. Furthermore,
in a five-variable analysis {i.e., money, credit, the rate
of interest, output, and prices), the interest rate helps
explain movements in output and prices. But, while
money and credit explain some of the variations in
the interest rate, neither financial aggregate directly
influences prices, a result similar to the four-variable
case referred to above. In addition, neither money nor
credit appears to contribute directly to the determina
tion of output; this is a markedly different result from
the one obtained when the interest rate is excluded .21
The interest rate, therefore, appears to be an impor
tant but heretofore relatively ignored variable in studies
of the relationships among money, credit, prices, and
output. The complete set of relationships needs to be
explored and explained further before the target ques
tion can be resolved.

Definitions of Selected Credit Aggregates
Bank credit

.
Federal and state and local gov
ernment obligations and total
bank loans. These include m ort
gages, consumer credit, ag ricul
tural loans, open market paper,
com m ercial loans, loans to other
financial institutions, and loans to
foreign banks.

Debt proxy . .
nonfinancial investors of cur
rency, checkable deposits, large
and small time deposits, money
fund shares, repurchase agree
ments, Federal Government se
curities, state and local govern
ment obligations, open market
paper, corporate and foreign
bonds, and other loans.
Nonfinancial
dom estic
cred it .............
nonfinancial sectors,
including
funds raised by Federal, state,
and local governments, corpo
rate bonds, mortgages, consumer
credit, and open market paper.

Would credit aggregates have helped policy
in the past?
Although the Federal Reserve monitored a variety of
economic and financial indicators in deciding upon
monetary policy in the past, M-1 was usually the pri
mary aggregate used in the period since 1970. In this
section, we look at whether giving more weight to
credit would have resulted in a “ better” monetary
policy. That is, would more attention to credit have
resulted in reactions by policymakers that would have

target in a policy aimed at controlling inflation.
An analysis of the temporal relations among real
GNP, prices, and money has shown that lagged values
of money (M-1) help explain prices, but not output.17
When a broad credit aggregate (nonfinancial domestic
credit) is substituted for M-1, so that the set of
variables under consideration includes output, prices,
and credit, then credit, like money, explains prices
but not output.18 When sets of lagged values of both
M-1 and credit are included in an equation explaining

In addition, both lagged output and lagged prices influence the current
stock of money, so that significant feedback from the economy to
money exists. One implication of this result is that single equation
models may inappropriately constrain the income-money relationship
to one in which changes in money temporally precede changes
in income.
Whether money or credit is used as the explanatory variable in the
output and price equations makes very little difference in the explana
tory power, as measured by the coefficient of multiple determination
(R2) and the standard error. (See Benjamin Friedman, “ The Roles of
Money and Credit” .) In the systems of three-variable equations referred
to above, neither money nor credit is significant at the 1 0 percent level
in explaining output. In the price equations, money is significant at the
1 percent level, but credit is significant only at the 5 percent level.

6 FRBNY Quarterly Review/Winter 1982-83

These results, along with those in the preceding paragraph, are
presented in Friedman, “ The Roles of Money and Credit". The R2s
and the standard errors of the equation in the system including both
M-1 and credit are essentially the same as those in the systems that
include only one of the financial variables. The exception is that the
standard error of the output equation is slightly lower in the system
that includes both financial variables.

For instance, Friedman, “ The Roles of Money and Credit” , and
Christopher Sims, “ Comparison of Interwar and Postwar Business
Cycles: Monetarism Reconsidered", American Economic Review
(May 1980).

See James Fackler, “ An Empirical Model of the Markets for Money,
Credit, and Output", mimeographed (Federal Reserve Bank of New
York, revised December 1982).

led, on balance, to a better performance, in terms of
output and inflation, since 1970?” The answer, we find,
is that it probably would have not.
To see if a joint M-1 and credit target would have
resulted in a better policy than one based on money
alone, it is first necessary to determine when those
two policies would have been different. That determi
nation depends on how policymakers were to incor
porate credit into the decision-making process. One
possible way would be to make decisions based on
M-1 alone as long as the signals emanating from M-1
and credit were not very different. That is, one could
assume that there is some “ normal” or “ average” re
lationship between M-1 growth and a particular credit
aggregate growth and that some variation around this
relationship is to be expected and not to be viewed
as “ unusual” in any policy-oriented sense. When the
signals were very different, though, policymakers
would weight those conflicting signals and adjust pol
icy accordingly.
For example, if credit growth greatly exceeded money
growth, the credit aggregate would indicate that a
“ looser” policy was actually being employed than if
policymakers looked just at money growth. Hence, to
achieve a given desired goal, policy would be tight
ened, either by raising interest rates or by reducing
nonborrowed reserves, relative to the situation in which
money was used as the sole indicator. This relative
tightening would occur whatever the policymakers’ re
action would have been to money growth alone. If money
growth was deemed unsatisfactorily low, policymakers,
looking only at M-1, might react by loosening policy;
looking at credit (growing rapidly) in addition to money
(growing slowly), they might still loosen but loosen
less. On the other hand, if money was growing faster
than desired or anticipated, policymakers might tighten
policy. Adding credit as a target in this situation would
lead policymakers to tighten even more. Analogously,
when credit growth was “ abnormally” low relative to
money growth, the indication would be that policy was
tighter in fact than that signaled by money alone;
hence, taking credit growth under consideration would
lead to a relatively looser policy.
Table 2 shows the periods in which such a joint
policy of M-1 and each of the credit aggregates, re
spectively, would have been different from one based
solely on M-1. The specific criteria used to judge
when the relationships between the growth rates were
“ abnormal” were differences between the four-quarter
growth rates of M-1 and the credit aggregates of more

22 For expositions! simplicity, we refer to actual policy in the post-1969
era as an M-1 based policy and to a hypothetical policy, with more
weight on credit, as a joint M-1 and credit policy.

than two standard deviations.23 For those periods in
which “ abnormal” differences in growth rates existed,
and hence a joint target-based policy would have been
different from an M-1 based policy, we have indicated
the direction of the difference in resulting policies. That
is, if policy would have been more expansionary using
a joint target, the term “ looser” has been entered in
the appropriate column; if policy would have been
more contractionary, the term “ tighter” has been used.
We have not indicated the exact magnitudes of the
resulting policy shifts because to do so would require
precise knowledge of, or assumptions regarding, how
policymakers would have weighted conflicting signals
given by M-1 and credit. However, as long as credit is
given some weight, we can determine the direction of
policy change. Furthermore, we have looked only at
short-run changes in policy. Any hypothetical policy
that differs from the historical record would have af
fected the entire subsequent evolution of the economy,
at least partly due to the reaction of subsequent
policy. However, the study of such potential long-run
responses is beyond the scope of this article.
While it is relatively easy to determine when policy
would have been different, as well as the direction of
those differences, it is much harder to conclude
whether the alternative policies would have been an
improvement over actual policy. Both economists and
policymakers have widely divergent opinions on the
efficacy of monetary policy on inflation and real eco
nomic activity, on the timing of any influence that does
exist, and on the relative costs of inflation and sacri
fices in real output. Therefore, in evaluating the joint
target policies, our conclusions are limited to those
cases in which we view the evidence as overwhelm
ing. Even so, we recognize that those conclusions are
by no means the only ones that could be reached.
To help readers follow our analysis of the alternative
economic policies and to aid in the formation of in
dependent judgments, we have listed the quarterly
growth rates (at annual rates) of real GNP and con
sumer prices in the last two columns, respectively, of
Table 2.
Nonfinancial domestic credit
A joint target of nonfinancial domestic credit and M-1
would have created a policy different from a strict M-1
target in three periods. In the first period, 1973-111
through 1973-IV, nonfinancial domestic credit was

23 The mean and standard deviation for a given quarter were calculated
from the data for the five years prior to that quarter. Thus it is assumed
that, given the changing relationships in an evolving economy, policy
makers would have looked only at the five most recent years of
behavior in evaluating signals in a given quarter.

FRBNY Quarterly Review/Winter 1982-83

Table 2

Differences in Policy Signals between M-1 and Credit
Policy change indicated by:
Nonfinan
cial domesDebt
Bank
tic credit
proxy
credit

Yearquarter

Real
Consumer
GNP price index
Quarterly growth, at
compound annual rate

1970-1 . .

Looser

— 1.54

6.70

1970-11 .

Looser

0.59

5.73

1970-111 .

3.86

4.45

1970-IV .

-3 .1 1

5.94

1971-1 . .

10.26

3.30

1.96

3.84
3.92

1971-11

1971-111 .

3.19

1971 -IV .

3.49

2.87

1972-1 . .

7.90

3.63

7.57

2.50

5.05

3.36

1972-IV .

7.48

4.20

1973-1 . .

10.96

6.34

Tighter

0.48

8.42

Tighter

1972-11 .

Tighter

1972-111 .

1973-11 .

Yearquarter
1976-11
1976-111
1976-IV
1977-1 .
1977-11
1977-111
1977-IV
1978-1 .
1978-11
1978-111
1978-1V
1979-1 .
1979-11
1979-111
1979-1V
1980-1 .

1973-111 .

Tighter

2.43

8.14

1973-1V .

Tighter Tighter

3.32

10.53

1980-11

1974-1 . .

Tighter

- 3 .9 8

12.57

1980-ill

0.45

1 0 .8 6

1980-IV

Tighter

1974-11 .

Policy change indicated by:
Nonfinan
cial domesDebt
Bank
tic credit
proxy
credit

Looser
Looser
Looser

Tighter

Looser

Real
Consumer
GNP
price index
Quarterly growth, at
compound annual rate
2.74
2.31
3.74
8.87
6.72
6.77
0.75
3.38

3.22
6.55
6.28
7.39
7.26
5.58
5.96
7.38
9.42
9.28

* 1 1 .0 0
3.33
5.52
1.15
-0 .9 2
4.82
0.73
1.51

10.61
12.71
13.69
14.10
16.48

- 9 .5 6

13.52

1 0 .0 1

1 .6 6

7.62

4.33

12.83
10.96

1974-111 .

- 2 .4 9

11.87

1981-1 .

Looser

7.90

1974-IV .

-5 .1 9

13.37

1981-11

Looser

-1 .4 8

7.81

1975-1 . .

- 8 .1 9

8.56

1981-111

2 .2 0

11.81

4.94

4.48

1981 -IV

8.55

1982-1 .

1975-11 .

Looser

1975-111 .

Looser

9.23

1975-IV .

Looser

3.64

8 .1 1

9.11

4.33

1976-1 ..

growing much faster than M-1, so that, had credit been
considered along with M-1, policy would have been
tighter than it actually was. The arguments for the
appropriateness of a tighter policy center around the
rapidly accelerating inflation rate of the period and
the high level of capacity utilization in various indus
tries. On the other hand, some of the inflation was
caused by rising food prices due to bad weather in
the 1972-73 winter and, in the fourth quarter of 1973,
by rising fuel prices due to the OPEC oil embargo. In
addition, the lower rate of economic growth experi
enced in the final three quarters of 1973 (compared
with the 1972-1 through 1973-1 period), if allowed to
persist, may have eventually led to less inflationary
pressures. Therefore, it is not clear what kind of policy
change would have been desirable in the second half
of 1973.

8 FRBNY Quarterly Review/Winter 1982-83

-5 .2 7

7.80

-5 .1 1

3.16

1982-11

2 .1 1

4.59

1982-111

0 .0 0

7.60

Tighter

In the second quarter of 1980, nonfinancial domes
tic credit again would have led policymakers to tighten
policy. Such an action, coming on top of the credit
control program adopted in March 1980, may have
intensified and extended the sharp drop in real GNP
that occurred in the second quarter (9.6 percent at
an annual rate). The benefit, of course, may have been
an even sharper drop in the inflation rate than actually
occurred from the first to third quarters of 1980.
The final instance in which nonfinancial domestic
credit would have affected policy was in the 1980-IV
to 1981-11 period. Had the credit measure been fol
lowed during that period, policy would have been looser
than the one actually employed. Although this may
have moderated the subsequent fall in economic ac
tivity, a looser policy may also have resulted in less of
an improvement in the inflation picture than was experi

enced. Thus, it is not clear that a looser policy would
have been desired.
Debt proxy
The joint target system based on the debt proxy mea
sure of credit would have called for actions different
from those based on M-1 alone in five different peri
ods in the post-1969 era. In 1972-11 the debt proxy
measure signaled that policy was looser than indicated
by money growth; hence, had policy been based at
least partially on the debt proxy, policy would have
been tighter. This, in fact, may have been a more
appropriate policy to pursue, given the fact that the
economy was in the midst of an extremely rapid
expansion and at the beginning of a period of accel
erating inflation.
A policy based on the debt proxy in the 1973-111
through 1974-1 period would have been similar to,
and had the same problems as, one based on nonfinancial domestic credit. Furthermore, since the pe
riod of tighter policy called for by the debt proxy
extended into the beginning of 1974, following that
signal would have led to a tighter policy at the start
of what was already developing into a very severe
recession.
From 1977-11 through 1977-1V a debt proxy-based
policy would have been more expansionary than an
M-1 based policy. At the time, the economy was ex
panding at a rapid rate, and a period of sharply accel
erating inflation was about to begin. A looser policy,
therefore, was probably not desirable.
In the second quarter of 1980 the debt proxy again
would have led to the same change in policy as indi
cated by nonfinancial domestic credit, and thus would
have had benefits and pitfalls similar to those de
scribed above. In addition, the debt proxy would have
led policymakers to tighten policy in the fourth quarter
of 1981, another period of very weak economic activ
ity and slackening inflation.
Bank credit
A joint bank credit and M-1 target would have led to a
looser policy than an M-1 target in the first two quar
ters of 1970. Although the economy was In a minor
recession at that time, the relatively high rate of
inflation may have made a more expansionary policy
undesirable.
The next two periods in which bank credit and M-1
would have given divergent signals were in 1972-11
and 1973-11 through 1973-111. In both periods, bank
credit would have led to a tighter policy. As in the case

of the debt proxy, a tighter policy probably would have
been appropriate in 1972-11. However, the same may
not be true for the latter period, as noted above for
nonfinancial domestic credit.
The final period in which bank credit and M-1
would have given different signals was from 1975-11
to 1975-1V, when bank credit would have led to a
looser policy. Given that this period marked the start
of a vigorous economic recovery, coincident with a
falling rate of inflation, it is not clear that any alterna
tive policy would have been superior to the one ac
tually pursued.
Summary
One possible argument on the part of proponents of
credit aggregates is that, even if the credit aggregates
on average do not “ outperform” the monetary aggre
gates, they may in fact give important signals at criti
cal points in the business cycle. However, the evidence
described in this section does not support this argu
ment. None of the credit aggregates described above
would have consistently improved an M-1 based policy
in the pbst-1969 period. Even in two of the three
periods in which more than one credit measure would
have led to the same modification in policy, those
modifications may have led policy in the wrong direc
tion. Furthermore, none of the aggregates indicated
that a tighter policy should have been pursued during
the most recent period of accelerating inflation, 1977
through the beginning of 1981. Finally, at points dur
ing the most recent contractionary period, nonfinan
cial domestic credit would have led to a relatively
more expansionary policy, while the debt proxy would
have called for a tighter policy; which policy modifi
cation would have been more appropriate is still an
open question.

Conclusion
In our view, there is no strong empirical evidence from
recent history supporting use of a credit aggregate.
However, such a view is an insufficient reason for dis
carding proposals to use credit aggregates as policy
targets. In particular, even though credit would not
have unambiguously aided the policymaker at critical
points of the business cycle, limited existing theoreti
cal work suggests that, in periods of substantial shocks
to money, credit targets are more likely to be useful
policy targets. If the recent financial innovations are
viewed primarily as distorting the monetary aggregates,
then credit aggregates are likely to be relatively use
ful, at least until the shocks to money subside.

James Fackler and Andrew Silver

FRBNY Quarterly Review/Winter 1982-83

Money and Credit: Exploring Alternatives

Control of a Credit Aggregate

In the debate about which financial aggregates the
Federal Reserve should target, a key question is how
the Federal Reserve would go about controlling a
credit aggregate. While the Federal Reserve can limit
the supply of reserves, it is difficult to see a close
connection between reserves and a broad financial ag
gregate, much (or all) of which is not reservable. Ad
mittedly, some nations’ central banks directly restrict
the quantity of credit that their banking system may
lend. And, in the United States under the special credit
restraint program of 1980, guidelines were set for
permissable expansion in loans and credit. Neverthe
less, in this country, direct restrictions on the quantity
of credit, particularly for extended periods of time,
have not generally been regarded as either a desirable
or a feasible way to operate.1
How, then, could a credit aggregate target be
achieved in the United States? One approach, advo
cated by some Wall Street economists, is to impose
a high capital requirement on banks. (This capital re
quirement would be set above the level demanded for
prudential purposes by the bank supervisor.) The ad
ditional need for capital to support bank credit expan
sion could, according to the proponents, act as a

I owe thanks to Henry Kaufman whose ideas provided much of the
impetus for this paper and to Don Kohn of the Board of Governors of
the Federal Reserve System as well as to many of my colleagues at the
Federal Reserve Bank of New York who provided useful comments.
1 Irving Auerbach has suggested that such limits on credit expansion
be adopted to replace the present system whereby reserves must be
maintained against liabilities. According to his proposal, banks would
be able to buy or sell their allocations.

FRBNY Quarterly Review/Winter 1982-83

substantial brake on both bank lending and total credit.
A second approach is to impose reserve requirements
on credit expansion, making it more costly for firms
and households to borrow. A third possibility is use
of a “ shadow” reserve requirement with some broad
measure of credit; the implicit “ reserves” calculated
in this fashion could be used to guide the actual
amount of reserves provided by the Federal Reserve.
In this article, I examine the effectiveness of capital
ratios as well as these alternatives as control devices.
The analysis indicates that
• Capital ratios on bank assets (above the pru
dential level that banks would maintain to
satisfy the supervisors or their equity holders)
would curb domestic bank lending. But much
of any increase in credit demand would be
accommodated in other markets.
• To the extent that credit needs were easily met
elsewhere, capital ratios would put little pres
sure on the general level of interest rates and
would have little effect on spending and on a
broad credit measure.
• Reserve requirements on bank credit suffer
from some of the same problems as capital
requirements.
• Shadow reserve requirements on a broad
credit aggregate, however, might be an effec
tive mechanism to influence both spending and
credit in the economy.

specified. For simplicity, assume that capital ratios
would be superimposed on a system in which there is
still a monetary target, implemented through reserve
provisions.
When the demand for loans increases, banks typi
cally raise funds through repurchase agreements on
securities in their portfolios or through the issue of
more certificates of deposit (CDs); they may also sell
some securities or buy fewer than they had planned.
The additional need for funds, if it is widespread, will
cause the CD rate to rise since investors need to be
induced to hold a larger volume of CDs. Also, if re
duced bank holdings of securities push up the rates
on these securities, CD investors will also typically
require a higher return. The higher cost of funds will, in
turn, induce banks to raise the rates charged on loans.
How would capital requirements affect this process?
Capital requirements, if they exceed the capital ratios
that banks would otherwise maintain, represent an ad
ditional cost attached to expanding a bank’s asset
portfolio. The markup on the cost of funds will there
fore probably be greater than in the absence of capital
ratios. As a consequence, some borrowers who have
direct access to funds in the commercial paper market
or bond market will elect to raise funds that way in
stead of through the intermediation of banks. In addi
tion, they may seek to borrow from foreign banks or
other institutions not subject to the capital require
ment.
The results of an expansion of loan demand under
this regime are:

Diagram

M a rk e t for Bank Loans and
C e rtific a te s of D eposit

CD0

Assume banks have core deposits on which they pay an
average rate of interest q. Banks raise additional funds through
certificates of deposit (CDs). In turn, banks make loans, basically
charging a rate of interest that is a constant markup over the CD rate.
Thus, L|, the supply function for bank loans before capital re
quirements, is represented in the diagram as parallel to the demand

• The loan rate rises by more than in the uncon
strained case;

for CDs (CDd). Ld is loan demand. In equilibrium, loan volume will be
L°, and the interest rate charged on loans will be r£. Part of L° will be
funded with deposits (M) and the remainder with CDs, in quantity CD0.

• Loans may expand but not by as much as in
the unconstrained case;

With the imposition of capital requirements, the banks' required
return on loans jumps, shown by the Lg curve. L* is the level of loans
above which marginal capital requirements take effect. In the new
equilibrium, the interest rate charged on loans increases to rj., and
loan volume falls to L1. A smaller volume of CDs is thus necessary
(CD1), causing the CD rate to fall to r^D from r ° Q.

Loan volume and capital requirements
Suppose there is an increase in the demand for loans
stemming from a planned increase in spending. How
would the results differ if expansions of bank credit
beyond some point were subject to high marginal
capital ratios?
To examine the effect of capital ratios, some back
ground assumptions about monetary policy must be

• Domestic bank profits will not rise as much;
• The capital requirement does not act as a com
plete bar against loan expansion except in the
extreme case where banks cannot raise capital
at all (even through retained earnings) and they
have no securities in their portfolio that they
can sell.2

Quantitative impact of capital requirements
The impact of capital requirements on the volume of
loans and the rates charged depend upon a number
2

A bank may wish not to sell securities whose market value is below
par value because, by selling them, they would be forced to show a
loss on those securities.

FRBNY Quarterly Review/Winter 1982-83

of factors including (a) the ease of raising capital
(either through equity or long-term subordinated debt)
and (b) the level of the capital requirement relative to
the capital ratio that would otherwise be maintained.
Suppose, for example, that on bank credit expansion
which exceeded the target a bank had to maintain
a marginal capital requirement of 2 percent above
the capital ratio that banks needed for prudential rea
sons. Further, suppose capital would be raised in the
form of equity rather than subordinated long-term
debt and that expected annual earnings of 20 cents
per dollar invested would need to be offered to new
equity buyers. (In August 1982, the thirty-five large
banks included in the Salomon Brothers Inc average
had an earnings-price ratio of 19.2 percent.) Since the
new equity represents a source of funds which could
substitute for liabilities, such as deposits or other
forms of borrowing, the net extra cost of the capital
acquired would be 20 percent minus the aftertax inter
est cost on these liabilities. Suppose that the interest
rate on these liabilities is 15 percent per year and that,
after Federal income taxes, this costs the bank roughly
8 percent (state and local taxes are ignored for this
calculation). Then the net cost of financing with equity
rather than with debt would be 12 percent. The capital
requirement would therefore add 24 basis points (the
0.02 undesired capital requirement multiplied by the
12 percent marginal cost of equity finance) per dollar
to the cost of expanding loan volume.
Some of the 24 basis points would be reflected in
a higher loan rate, some in a lower CD rate, and some
in a reduction of bank profits (diagram). The fewer
the alternatives to bank loans, the more willing poten
tial borrowers would be to pay a higher rate and still
obtain credit from banks. In this case, where the de
mand for loans is “ inelastic” , borrowers will end up
paying almost 24 basis points more. In contrast, in the
case where borrowers have good alternatives to bank
loans, they will pay an increase which is much less
than 24 basis points.
Naturally, if the original assumptions were not a
good description of the real world, the loan rate could
increase by more (or less) than 24 basis points. For
example, if the supply curve of equity capital to the
bank were upward sloping, so that the bank had to
offer successively higher expected returns to share
holders to raise more capital, the effect on the loan
rate could be greater. Further, there may be sub
jective costs involved in raising new capital. For ex
ample, a bank may be reluctant to issue new stock if
the market price of its stock is below book value. In
this case, current earnings would be a key factor in
capital expansion: banks with high earnings could
retain earnings to finance expansion, while those with

FRBNY Quarterly Review/Winter 1982-83

lower earnings would not expand their portfolios.
Moreover, if capital ratio requirements reduce earn
ings, banks would have to reduce dividends in order
to retain earnings. Aversion to reducing dividends may
create a lower desired capital ratio than would be true
in the unconstrained case. Thus, the capital require
ment could exceed the desired ratio by more than an
initial comparison would suggest.
If current earnings were a key factor in capital
growth, then the bank’s earnings per dollar of equity
would provide an upper limit on asset growth. For ex
ample, earnings on equity of, say, 20 percent would
permit an expansion of capital and assets of something
less than 20 percent. Capital growth above this level
would then be very “ expensive” in terms of stock
holders’ preferences. At the extreme, the bank would
raise the loan rate enough so that loan demand was
at or below the point where the marginal capital ratio
applied.
In any event, the loan rate would tend to rise and
some customers of banks would then try alternative,
cheaper ways of raising funds. Those that could issue
commercial paper or bonds might do so. Others would
seek to borrow from foreign banks located abroad.
Still others would make financing arrangements with
either suppliers or customers who had access to the
commercial paper market. In addition, if bank guaran
tees were not subject to reserve requirements, banks
could insure credit extended by other parties to busi
nesses with which the banks are familiar. Letters of
credit are one way the banks could provide such
guarantees. In this fashion, banks could, in principle,
continue to perform the role of rating customers not
known to the general public and putting themselves
as guarantors between the public and those customers.
In this way, they would continue to facilitate the ex
pansion of credit.
The higher the capital ratio imposed, the more
would loan demanders seek these alternative routes.
Thus, the reduction of domestic bank loans would be
offset, at least in part, by expansion in other sources
of credit. Twenty years ago when many of these al
ternative markets were either undeveloped or com
pletely nonexistent, the demand for bank loans was
less elastic. Then the imposition of capital ratios would
have raised loan rates more substantially and the re
duction of loans would have been offset only to a
small extent by other credit sources. Today, however,
because of the availability of substitutes, the rise in
the loan rate is likely to be small and the offset to
bank credit provided by the alternatives is likely to
be sizable.

Effects of capital ratios on credit and spending growth
Marginal capital ratios would presumably be applied
to banks when credit expansion was above a specified
target range. (These special marginal capital ratios
would have to be set above the level that would be
maintained otherwise to have an effect on interest
rates.) Bank asset growth which is subject to the
marginal capital ratio would be accommodated by
banks only at a higher interest rate. And higher inter
est rates— if they occur— are likely to influence the
spending decisions of households and firms. Then, as
spending responds and credit needs change, total
credit as well as bank credit would be reduced.
The train of events, however, might not follow this
pattern. First, if a sufficient number of borrowers had
low-cost alternative sources of credit, these borrowers
would not be willing to pay higher rates to banks.
Instead, they would take their funding needs else
where, leaving banks with no expansion in assets that
was subject to the capital ratio. Interest rates would
differ little from what they would be without capital
ratios, and spending decisions that depend upon in
terest rates would also be little affected.
Another problem with the simple system of marginal
capital ratios is that it works only when credit growth
is rapid. When credit growth is low or negative, such
as during a recession, the capital ratio would not
work to lower interest rates more than they would fall
naturally: a marginal capital ratio below the prudential
level will not change the capital ratio that banks main
tain and therefore will not change their costs.

Other noteworthy effects of capital ratios
Capital ratios could have relatively little effect on the
general level of interest rates and on total credit. But
they could nevertheless have a large impact on the
banking industry, and related industries, as well as on
the financial markets. The banking system’s profits, in
aggregate, would probably be reduced somewhat and
the banking system would have a smaller relative asset
volume as potential borrowers shift to alternative
sources of funds.
Another effect of a high marginal capital requirement
is to reduce a bank’s leverage. Lower leverage means
a lower average return on equity and a greater degree
of safety. Both factors will tend to induce banks to
increase risk and raise return by altering their port
folios away from investments and toward loans and by
making loans to riskier borrowers. Indeed, Koehn and
Santomero3 have argued that constraints such as cap
ital ratios can actually increase the probability of bank
3 See M. Koehn and A. Santomero, "Regulation of Bank Capital and
Bank Portfolio Risk” , Journal of Finance (December 1980).

failures. On balance, though, capital ratio requirements
are likely to make banks safer institutions.
Capital ratios will affect different banks to different
extents. One factor is the relationship between the
height of the marginal capital ratio and the capital
ratio a bank would otherwise seek to maintain: a bank
with a relatively low initial capital ratio would tend to
be affected more than a bank with a higher capital
ratio. Another problem with marginal capital ratios on
rapid expansion in bank credit is that they penalize
banks in regions where there is rapid economic
growth; at the same time, they have no effect on
banks in areas which are growing slowly or contract
ing. It is possible, however, that banks in growing
areas would induce others to do loan participations
or that banks in growing areas would concentrate on
loans, reducing their securities holdings.
Another effect of capital ratios might be to change
the relationship between various interest rates. Com
pared with the situation that could prevail without such
capital ratios, commercial paper rates would probably
be higher as more firms seek nonbank financing. CD
rates would be lower because there would be less
need to issue CDs.

Reserve requirements
In principle, reserve requirements could be imposed
on bank assets or even on other types of domestic
credit such as finance company credit or bonds issued
in the United States. Marginal reserve requirements
were, in fact, imposed upon certain types of consumer
loans during 1980.
Reserve requirements on bank assets, or marginal
reserve requirements on expansions in bank assets,
would have effects very similar to those of capital
ratios. A reserve requirement on the increases in asset
volume would make it more expensive to expand
loans, just as did a marginal capital ratio. For example,
a marginal reserve requirement of 2 percent would
mean that the bank would have to raise $1.02 for each
$1 it lent out. Thus, its borrowing cost would effective
ly be raised 2 percent. At an interest rate of 15 per
cent, say, this cost is 30 basis points. As a conse
quence of the greater effective cost, loan rates would
tend to be higher and the loan volume smaller than in
the absence of the reserve requirement.
In several other respects, too, reserve requirements
on bank credit are similar to capital ratios. For ex
ample, a marginal reserve requirement on bank credit
expansions would encourage borrowers to circumvent
the domestic banking system. It would also tend to
reduce bank profits. In contrast to capital ratios, how
ever, reserve requirements are unlikely to improve
bank safety. In fact, in the effort to improve their aver

FRBNY Quarterly Review/Winter 1982-83

age returns, banks would probably choose riskier loans
and investments. Reduced safety is thus the likely
result.
If reserve requirements were also imposed upon
nonbank credit, borrowers would not gain by shifting
to nonbank sources of funds. For example, assuming
that commercial paper issues were also reservable,
borrowers would not shift from bank loans to com
mercial paper. Nevertheless, such reserve require
ments could be circumvented by borrowing abroad
or, if only public issues were covered, by arranging
private deals.

Shadow reserve requirements
Another way of implementing a credit target would be
to utilize a system of shadow reserves. (Indeed, a
system of shadow reserves can be used with any
aggregate containing nonreservable components.) As
sume that reserve requirements remain on deposit
liabilities; the tightness in the reserves market, how
ever, depends not on money growth but rather on the
growth of a selected credit aggregate.4 One way of
connecting reserves availability to credit growth is to:
• Set a target for the credit aggregate. This
credit aggregate could include credit raised
from nonbank sources— it would not matter
whether the institutions (or the market) were
under the purview of the Federal Reserve;
• Calculate the deviation of credit from its target
level;
• Apply a shadow reserve requirement to that
deviation to obtain the adjustment to the Fed
eral Reserve’s objective for nonborrowed re
serves;
• Reduce the path for nonborrowed reserves by
this amount. (If the adjustment were negative,
nonborrowed reserves would be increased.)
For example, if credit moved $10 billion above its
target range and the shadow reserve requirement was
5 percent, the nonborrowed reserves path would be
lowered by $0.5 billion (equal to 0.05 x $10 billion).
Thus, rapid growth of credit would be translated auto
matically into reserves shortages which would put up
ward pressure on interest rates.
Because of the generalized effects on interest rates,
* Tightness in the reserves market could be made to depend upon a
combination of money and credit growth. For simplicity of exposition,

it is here assumed that only credit affects tightness in the reserves
market (i.e., deposit growth is accommodated).

FRBNY Quarterly Review/Winter 1982-83

this shadow reserves mechanism is more likely to affect
total credit usage than a system which imposes capital
ratios or reserve requirements only on banks— since
a mechanism focused on domestic bank lending may
have little effect if alternatives to bank loans are readily
available.
Of the three mechanisms considered— capital ratios
on bank assets, reserve requirements on bank assets,
and shadow reserves on total credit— the shadow re
serves mechanism has some clear advantages. First,
since it is a variant of the reserves targeting mechan
ism currently in use, a rough estimate of its impact on
interest rates could be based upon the experience of
the last few years: the typical spread between the
Federal funds rate and the discount rate that results
from that reserves shortage. In contrast, the impact of
capital ratios or reserve requirements on interest rates
depends upon the elasticities of loan demand and
CD demand whose magnitudes are now well-known.
Second, it is a system which discriminates less be
tween credit expansion by banks versus nonbanks.5
Fundamentally, though, control of a broad financial
aggregate, whether through capital ratios or through
shadow reserve requirements, would be quite indirect.
The mechanism by which shadow reserve requirements
(or for that matter capital ratios) influence the volume
of credit is:
• Interest rates are altered by overly rapid (or
overly slow) growth of the financial aggregate;
• These interest rate changes affect spending
decisions of households and firms and the
credit demands that go along with those spend
ing decisions.
Also, the level of shadow reserve requirements is
not much easier to set than an actual reserve require
ment or a capital ratio. In a general sense, the higher
the shadow reserve requirement, the more the Federal
funds rate will change when the financial aggregate
deviates from the target range. And the shadow re
quirement would have to be set high enough to ensure
that, when the rate of inflation accelerates, interest
rates rise by more than the rate of inflation. That is,
rapidly expanding credit must produce an increase in
real rates of interest or else GNP and credit demands
will tend not to recede. Economic models at the present
time, however, do not yield a unique answer on exact
ly how much of an interest rate change is needed to
5 But any rise in short-term interest rates increases the cost of
reserves, which yield no interest, and thus affects the institutions
that must hold reserves.

produce the credit reduction that is sought. The re
sponsiveness of spending decisions and credit usage to
interest rates needs to be studied further to design ap
propriate implementation procedures for credit targets.

Concluding remarks
In this article, I have examined various mechanisms
for using a credit measure in monetary policy. Other
key issues, such as the potential problems involved
in focusing on credit, have not been addressed here.
For one thing, pressures for special treatment of one
category of credit or another are bound to arise. In

addition, changes in the distribution of income or
spending or in the tax laws may produce an increased
demand for borrowing through the credit markets
while total spending is unchanged. (For example, firms
doing the bulk of the investment in one year may be
those who have poor earnings and as a consequence
need to borrow a lot, whereas in other years it could
be firms with large profits who are doing most of the
investment with retained earnings.) How the monetary
authorities would deal with these shifts in demand
could develop into an important issue if credit aggre
gates become the primary focus of monetary policy.

Marcelle Arak

FRBNY Quarterly Review/Winter 1982-83

Money and Credit: Exploring Alternatives

III

Reserves against Debits

Financial developments during the last few years have
provided some valuable insights into the problems the
Federal Reserve is likely to encounter in targeting a
narrow monetary aggregate over the next few years.
The rapid growth of NOW accounts since 1981 illus
trates how difficult it is to interpret M-1 growth when
a single account contains both savings and transac
tions balances. Working in the opposite direction, the
increased emphasis on cash management by the cor
porate sector has spurred the development of financial
innovations, resulting at times in weak growth of M-1
relative to income and interest rates since the mid1970s. Analysts are increasingly questioning the future
usefulness of M-1 as a guide to policy.1
As the financial system continues to evolve, that is,
as interest rate ceilings are phased out and consumers
and corporations continue to find new ways to hold
transactions and investment balances, it will become
increasingly difficult to measure a transactions con
cept of money. This raises the question whether mone
tary policy should be formulated in terms of a reserves
path linked to balances in certain types of accounts
or whether the reserves path should be tied to deposits
in some way other than daily average balances.
This issue has been raised before, and it seems

1 For example, Frank E. Morris, “ Do the Monetary Aggregates Have a
Future as Targets of Federal Reserve Policy?", New England Economic
Review (March/April 1982); Anthony M. Solomon, "Financial
Innovation and Monetary Policy", Annual Report-1981 (Federal
Reserve Bank of New York).

FRBNY Quarterly Review/Winter 1982-83

worthwhile to reexamine the solutions that were pro
posed as a first step toward finding answers for current
policy problems. In 1932, the Committee on Bank Re
serves of the Federal Reserve System proposed that
reserve requirements be placed, not only on the level
of deposits, but also on the volume of debits to those
deposits:2
These withdrawals, which are shown by debit
entries on the books of member banks, are the
only real test of the activity of a deposit account
and furnish the only basis by which that activity
can be equitably and effectively reflected in re
quirements for reserves. Under this proposal,
therefore, each deposit will carry a total reserve
based on its activity as well as on its amount.
Could a reserves path approach to policy be improved
if reserve requirements were also placed on the daily
average outflows or debits from certain accounts
rather than only on balances in accounts used for
transactions purposes? Debits, it can be argued, give
a good indication of the volume of transactions a given
account is used for whether or not it contains
some savings balances as well as transactions bal
ances. Hence, with reserve requirements on debits in
addition to balances, an account of a given size used
extensively for transactions would have a higher re2Annual Report (Board of Governors of the Federal Reserve System,
1932), page 262.

serve requirement than one of the same size not
used for transactions as much because it also contains
savings balances. Because the volume of incomerelated transactions or debits should reflect the state
of the economy, reserve requirements on debits would
automatically give a signal about undesired strength
or weakness in the economy, as the demand for re
serves deviates from path. Moreover, the structure of
reserve requirements on debits and balances could be
so designed that financial innovations would not nec
essarily result in an easing in policy which could
occur when reserve requirements are placed on bal
ances only. The additional debits necessary to manage
balances at lower levels would still generate reserves
pressures as would the income-related debits. In some
sense, it could even be said that such an approach
would bridge part of the gap between those arguing
that the Federal Reserve should focus on money and
those arguing that the Federal Reserve should target
nominal GNP directly.
Debits have been growing much more quickly than
GNP or M-1 since about the mid-1960s (Chart 1).3 The
more rapid growth of debits, however, should not be
surprising during a period of increasing inflationary
pressures. As higher inflation rates were reflected in
higher nominal rates of interest, the increased em
phasis on cash management not only reduced de
sired cash balances but also increased the volume
of debits required to keep money balances at the
lower desired levels. This has resulted in much more
rapid growth of velocity, measured as the ratio of
debits to money, than in the more conventional mea
sure of velocity calculated as the ratio of GNP to
money balances.
In the next section of this article, the potential
use of debits in the policy process is explored in
detail. Some econometric results are presented in
Appendix 1, using the debit statistics to show the
effects of financial innovation on the demand for
money since the mid-1970s. The results suggest that:

Chart 1

W ith th e volum e of debits to tra n s a c tio n s
accounts growing much fa s te r than M-1 or
n o m in a l GNP . . .
70
1946=1.0
60

40
Debits to transactions /
accounts
/
30

20
/
S

Nominal
GNP
M' 1

0l- i'i'i
1947

i i i i i i 1-1 1 i I.X )

i i i i
65

i i i l I l 1I I 11 111 1
70

80 82

. . . the v elo city of M-1 m easu red in term s
of d e b its has in creased much m ore rap id ly
than velo city m easu red in term s of GNP.
16
1946=1.0
14

• Prior to 1973 debits explained the transactions
demand for money about as well as income.
• Since 1973, however, debits appear to be a
better proxy for financial transactions under
taken to reduce money balances while still
allowing the same volume of income-related
transactions.
3

The debit statistics are published in the Federal Reserve Bulletin.
The series is based on a limited sample which has been changed
over the years. On a month-to-month and quarterly basis, the debits
statistics are very volatile. For a detailed discussion of this series,
see George Garvy, “ Debits and Clearing Statistics and Their Use”
(Board of Governors of the Federal Reserve System, 1958).

GNP/M-1

^^ ^

___

o l I I Ll l I Li. I i i 1.1 l l I I I I I I I I I l I I I I l I l I I I I

1947

Source:
System.

80 82

Board of Governors of the Federal Reserve

FRBNY Quarterly Review/Winter 1982-83

Table 1

Effects on Required Reserves of Alternative Reserve Requirements on Debits4
In dollars; figures in parentheses are percentage deviations from initial conditions.

Example

Daily
average
balances

Daily average debits
Income
related
Financial

Initial con
ditions . . .

200

400

( 1 ) .........

130

400

( 2 ) ........

130

260

(3)

........

130

120

(4)

.........

130

400

(5)

........

200

r = 0.15
y= 0

Levels of required reserves!
r = 0.05
r = 0.02
y = 0.05
y = 0.065

30.00

19.50
(-3 5 )
19.50
(-3 5 )
19.50
(-3 5 )
19.50
(-3 5 )
30.00
( 0 .0 )

27.00
( - 10)
2 0 .0 0

(-3 3 .3 3 )
13.00
( -5 6 .6 7 )
30.00
( 0 .0 0 )
2 0 .0 0

(-3 3 .3 3 )

30.00
29.25
(-2 .5 )
20.15
(-3 2 .8 3 )
11.05
( -6 3 .1 7 )
33.15
(10.50)
17.00
(-4 3 .3 3 )

Velocity =
income debits
total balances
2 .0

3.1
2 .0

0.9
3.1
1.0

* The table was constructed assuming that the $200 in NOW account balances in the initial conditions case is equally divided
into savings and transactions balances ($100 each). In all the examples, the assumption is made that transactions balances
are held in a constant proportion to income debits (4 = $400/$100). Savings balances are assumed not to be related to the
level of income debits. The figures in parentheses represent percentage deviations in required reserves for examples (1) through (5)
from the levels in the initial conditions case.
t Assuming reserve ratios (r on balances, y on debits), first, for a 15 percent reserve requirement on daily average balances only, second,
fo ra 5 percent reserve requirement on daily average balances and debits, and, third, for a 2 percent reserve requirement on daily
average balances and 6.5 percent on daily average debits.

*
This part of the table is intended for readers who
desire a more general understanding of the example
above. It is not necessary for following the main points
of the article. If money (M) has both a savings (S) and
a transactions (T) component, then total income-related
debits (D) are equal to debits to the savings com po
nent (DS) plus debits to the transactions component
(DT).
(1)
(2)

M = S + T
D = DS + DT

The income turnover or velocity of transactions bal
ances (VT) is equal to the volume of income-related
debits to the transactions component (DT) divided by
the level of transactions balances (T). Sim ilarly, the
income turnover of savings (VS) is calculated as
income-related debits to the savings com ponent (DS)
divided by savings balances.
(3)
(4)

VT = D T /T
VS = D S/S

Required reserves (R) are equal to the reserve ratio
on deposits (r) m ultiplied by the daily average level
of deposits plus the reserve ratio on debits (y) m ulti
plied by the daily average volume of debits.
(5)

R = rM + yD

FRBNY Quarterly Review/Winter 1982-83

*
Equation (5) can also be w ritten as:
(6)

R = rS + rT - f yDS + yDT

Solving equations (3) and (4) for DT and DS and then
substituting into equation (6) yields:
(7)

R = (r + yVT)T + (r + yVS)S

If VS is equal to VT, then a change of a given size in
either savings or transactions balances w ill result in the
same short-run movement in the demand for reserves.
(This is also the result when there are reserve require
ments on balances only.) But, since VT is considerably
greater than VS, reserve requirements on debits give
greater weight, in effect, to movements in the trans
actions com ponent.* In the numerical example above,
VT (income de bits/tra nsa ctions balances) was set equal
to 4, and VS equal to zero by assuming there were no
income debits to savings balances. VT and VS were
treated as constants. But they cou!d vary in practice.
If so, the impact on the demand for reserves would
depend upon the relative sizes of S and T, and whether
the change in VS or VT was caused by a movement in
income debits or in balances.
' Here, the focus has been on income-related debits. If
the analysis is extended to include the financial debits as
well, total velocity of each component (income debits
plus financial debits/balances) would determine the relative
weights given to the transactions and savings components.

Appendix 2 contains an analysis of placing reserve
requirements on debits in the context of a simple
IS-LM model. The main conclusions are:
• Reserve requirements on debits could help
stabilize income from shocks originating in
both the monetary and real sectors.
• At what level to set reserve requirements on
debits relative to requirements on balances is
an open question. It depends upon the type of
shocks the economy is likely to encounter.

Potential use of debits in monetary policy
Would monetary policy be more responsive to changes
in economic activity if reserve requirements were
placed against debits as well as balances? In this sec
tion, this question is explored for both the corporate
and consumer sectors. Suppose that an increase in
the demand for transactions balances occurs because
income is expanding more rapidly than expected.
Businesses are assumed to manage demand deposits
at minimal levels on a continuous basis. Thus, they
would increase the volume of debits to their demand
accounts, as sales improved and more funds than
usual accumulated in their accounts and needed to be
invested daily in liquid (overnight) instruments. Like
wise, debits would increase as corporate payrolls and
other variable business costs rose along with the more
rapid growth of aggregate demand and production. Un
der such circumstances, the increased demand for re
serves relative to the target path caused by the greater
volume of debits would push interest rates higher.
Conversely, as the economy slowed in response to
higher rates, a smaller volume of funds would ac
cumulate each day in corporate accounts for over
night investment, and firms would cut back on their
spending for labor and other variable factors of
production. This would cause the volume of debits
to fall, thereby easing reserve pressures. Under
the assumption used here— that firms in the future
will be very effective in managing cash balances—
reserve requirements against debits would automati
cally apply pressure to correct deviations in the de
mand for “ money” and hence reserves from target.
In contrast, under the same assumption that firms
manage balances at constant minimal levels, reserves
on balances only would have little automatic effect on
controlling “ money” held by the corporate sector.
For consumers, who would be holding both sav
ings and transactions balances in their NOW ac
counts, the consequences of having reserve require
ments on both debits to and balances in NOW ac
counts are more difficult to analyze. To illustrate how
it might work, a simple example was constructed

(Table 1). The numbers were chosen for the sake of
ease of illustration. In the first row, the example shows
three different ways a given level of required reserves
($30) could be generated. The first column shows
daily average balances, the second column daily aver
age income-related debits, and the third column finan
cial transactions. The numbers in the middle group
of figures of the table are required reserves, first under
the assumption of a 15 percent reserve requirement
on daily average balances only, then under the as
sumption of a 5 percent reserve requirement on both
daily average debits and daily average balances, and
finally required reserves when there are reserve require
ments of 2 percent on daily average balances and 6.5
percent on daily average debits. In each of the next
three examples, it is assumed that balances in NOW
accounts are reduced by $70 from the levels in the
“ initial conditions” case as a result of a single sub
stitution out of NOW accounts into some other financial
instrument at the beginning of the week. Therefore,
daily average financial debits for the week increase by
$10 ($70 + 7). Depending on the source of that $70
reduction of NOW account balances (savings or trans
actions balances), the percentage drop in the demand
for required reserves will vary in the short run accord
ing to how reserve requirements are set.
The first example is constructed to show what would
happen to the demand for reserves in the short run
when NOW account balances decline from $200 to
$130 because of a $70 reduction of the savings com
ponent of NOWs. Since the reduction stems from a
decline in savings balances, income debits remain at
the same level and velocity measured in terms of total
balances rises from 2 to 3.1 (final column of table). If
reserve requirements are on balances only, there is a
35 percent reduction of the demand for reserves, com
pared with the initial conditions. But, if there are re
serve requirements on debits also, a much smaller
percentage decline in the demand for required re
serves would occur. The latter seems to be a more
correct result when the savings or investment compo
nent, not the transactions component of NOWs, is
being reduced. In other words, if the reserves levels in
the initial conditions are target levels, then the interest
rate reduction with reserve requirements on debits and
balances would be smaller than with reserve require
ments on balances only. And, since economic activity
has not declined even though total NOW account bal
ances have, the smaller reduction of rates seems more
appropriate.
In the second example, savings and transactions
balances are reduced in proportional amounts. As a
result, the decline in total NOW account balances
is proportional to the fall in income debits, and velocity

FRBNY Quarterly Review/Winter 1982-83

Chart 2

Peak Three-Month Treasury Bill Rate and
V e lo c ity of M-1 in Term s o f D ebits
30

1947=1.0

15
Peak 3-month
p+v
Treasury bill r a t e - '

/
/

*L )e b its /M -1

.L l I I I l I I -L l I I I I I I I I I I I 1.-Li I i I I I i.i. 1 1
1947 50
55
60
65
70
75
80 82
Source: Board of Governors of the Federal Reserve System.

is stable. Thus, in this very special case, the decline in
total NOW account balances is an accurate indicator
of what is happening to economic activity, even though
NOW accounts contain both savings and transactions
balances. Under these circumstances, the decline in
the demand for reserves is much the same, regardless
of how reserve requirements are set.
Turning next to the third example, it is assumed that
the $70 reduction of NOW account balances occurs be
cause of a reduction of the transactions component of
NOW accounts. Income velocity— measured in terms of
total balances— declines to about one as a result. In
other words, the economy declined more sharply than
total NOW account balances. As in the previous two
examples, if reserve requirements are on balances only,
the demand for reserves in the short run declines by
35 percent. But, when there are reserves on debits
also, the percentage decline in reserves demand is
considerably larger because of the greater percentage
decline in income debits than in total balances. In this
case, if the reserves levels in the initial conditions case
are the target levels for reserves, interest rates would
decline considerably more, when there are also re

FRBNY Quarterly Review/Winter 1982-83

serve requirements on debits, than when reserves are
on balances only. Again, this seems to be a more
correct policy response.
Example four is designed to show what happens if
a financial innovation is developed that enables the
consumer to undertake the same volume of incomerelated debits with lower average transactions bal
ances. In this example, balances are reduced 35
percent by a $70 overnight investment at the end of
each business day. It is assumed that the next day
the funds are brought back into the account and can
be used again for transactions purposes. Thus, rather
than having a daily average volume of $10 in financial
debits as was the case in examples (1) through (3),
financial debits average $70 because withdrawals are
done daily, not just once, to attain the lower average
balance. In this case, reserves only on balances give
completely the wrong signal, that is, the demand for
reserves falls 35 percent below the assumed reserves
path even though the economy has not weakened.
With equal reserve requirements on debits and bal
ances, the reserves freed by lower average balances
are absorbed by the higher volume of financial debits,
and there is no deviation from path as a result. When
there is a higher ratio on debits than on balances, re
quired reserves would increase. The latter would pro
duce an unnecessary tightening in the reserves market.
In the fifth example, the economy is assumed to be
weakening. Transactions balances decline as a result,
but total balances remain unchanged because it is
assumed that consumers increase precautionary sav
ings balances. Income velocity falls because, while
balances remain unchanged, income debits decline
along with the drop in transactions balances. With
reserves on balances only, the demand for reserves
remains unchanged. If there are reserves on debits
also, the demand for reserves falls, leading to an
easing in the reserves market, an appropriate policy
response since the economy has weakened. In the
bottom panel of Table 1 is the basic model underlying
the numerical examples just discussed. It is not nec
essary to read that part of the table to understand the
rest of the article.
When the proposal was made back in 1932 to put
reserves against debits as well as balances, stable
growth of money did not seem to be an issue. Ba
sically, the concern was that, even if the Federal
Reserve controlled the reserves base and therefore
in some sense money, income could still expand more
rapidly than desired because of an increase in velocity
that might not be recognized immediately. When re
serve requirements are against debits as well as bal
ances, policy would automatically tighten when ve
locity increases and ease when it falls. Moreover, the

higher the reserve ratio on debits is set relative to the
reserve requirements on balances, the greater would
be the policy response to changes in velocity. The
1932 Federal Reserve committee, in fact, proposed set
ting reserve requirements on balances at 5 percent and
on debits at 50 percent. However, at that time the
distinction between income-related debits and cashmanagement debits had not been made, and the above
examples suggest that under certain circumstances it
might be better to set the same reserve requirement
ratio on both debits and balances (Appendix 2).

Is such an approach practical?
It appears at least in theory that monetary policy might
be more responsive in the future to changes in eco
nomic conditions if required reserves were linked to
debits against transactions accounts as well as to bal
ances in those accounts. But the practical implemen
tation of such an approach would still require far more
research.4 If an increase in financial debits not related
to cash management occurred, both the balances in
transactions accounts (as is the case currently) and
debits could increase, causing the demand for reserves
to be above path. To the extent that this represented
“ speculation” in the stock, bond, money, or commodi
ties markets, some tightening of monetary policy might
be desirable. But to the extent the increase in financial
debits represented a financial panic or crisis, during
* For additional discussion of difficulties with such an approach:

W.L. Smith, “ Reserve Requirements in the American Monetary System” ,
Monetary Management (Commission on Money and Credit), 1963.

which the public shifted from one type of financial
asset to another, such a tightening in policy would not
appear to be appropriate.
It is also possible that the volume of transactions
debits associated with a given level of GNP could
change over time, causing monetary policy to ease
or tighten for reasons not related to the general level
of economic activity. For example, the degree of inte
gration of different business activities within one com
pany influences the number of transactions routed
through banks. A shift toward smaller firms would
cause an increase in the volume of debits associated
with a given level of GNP. Moreover, if banks explicitly
charged customers according to the volume of debits
against their accounts, there would be an incentive
to use currency more for transactions and perhaps
find other ways to avoid debiting accounts when
making transactions.
In addition, the debits statistics, as currently com
piled, are very volatile on a month-to-month and even
on a quarterly basis. This would probably add to inter
est rate variability. However, it is difficult to know how
much of a problem this would be in practice because
the statistics are now based on a limited sample.
Moreover, much of the volatility of the debits statistics
might be eliminated if reserve requirements were not
placed on the accounts of certain dealers and brokers
of financial instruments. All in all, it appears that
much more work is needed on this topic before any
firm conclusions could be reached, but it does appear
to be a proposal that is worthwhile reconsidering
some fifty years later.

John Wenninger

FRBNY Quarterly Review/Winter 1982-83

Appendix 1: Debits and the Demand for Money
From the point of view of money demand, it is not
clear how debits fit in. Some analysts have argued
that debits are a better measure of the total trans
actions demand for money than GNP because debits
would also allow fo r demand for money fo r non-GNP
transactions and financial transactions.* However,
those financial transactions (debits) undertaken to
manage money balances more efficiently, i.e., over
night repurchase agreements (RPs) or sweeps into
money funds, would be actions that reduce money
balances, as measured at the close of business, not
increase the demand for money. To the extent that
the growth of debits since the mid-1960s and the even
more rapid growth of debits since the mid-1970s
(Chart 1) represent increased emphasis on cash man
agement, the growth of debits might serve as a
reasonable proxy fo r the effects of financial inno
vation on the demand for money. Some analysts have
proxied for the effects of financial innovation by using
an interest rate ratchet variable, i.e., a variable that
rises to new peaks but never d e clin es.t Chart 2 shows
the close tim ing in the movements of this variable as

* For example, Charles Lieberman, “ The Transactions Demand
for Money and Technological Change” , Review of Economics
and Statistics (August 1977).
f Richard D. Porter, Thomas D. Simpson, and Eileen Mauskopf,
“ Financial Innovation and the Monetary Aggregates", Brookings
Papers on Economic Activity, 1979.

compared with the debits/M -1 ratio, suggesting that
perhaps the volume of debits might also be a relatively
good proxy for the effects of financial innovation on
money demand once the levels of income and interest
rates are allowed for. Reported in Table 2 are some
results using a standard money demand equation.
Equations (1) and (2) are money demand equations
estimated through 1982-111 and 1973-IV, respectively.
As has been well documented in other studies, the
equation deteriorates when the sample period is ex
tended, with the coefficient on the lagged dependent
variable rising to almost one and the income elasticity
declining sharply. Equations (3) and (4) show the re
sults of using debits rather than income as the trans
actions variable. For the shorter sample period, debits
work about as well as income— compare equations (2)
and (4)— but for the longer sample period the esti
mated coefficient on debits becomes insignificant. In
equations (5) and (6) both debits and income are
used. Debits are insignificant in the shorter period
— equation (6)— suggesting debits and income are both
measuring transactions and com peting for explanatory
power prior to 1974. In the longer sample period,
however, both debits and income are significant,
but debits have a negative coefficient. This could
be because debits are increasing relative to income
as a result of financial transactions geared toward
managing money balances more efficiently, and hence
more aggressive cash management has changed the
relationship between debits and money demand.

Table 2
R e g re ssio n R esults
(1)

M = - 0 .1 2 + 0.02Y - 0.02R + 0 .9 7 M (-1 )
(2.3)
(2.6)
(4.1)
(36.6)

R2 = 0.96

(2)

M = - 0 .5 5 + 0.12Y (3.6) (3.4)

p2 = o 97

(3 )

M = -0 .0 0 2 + 0.005D - 0.01 R + 0 .9 9 M (-1 )
(0.1)
(1.7)
(3.3)
(41.0)

R2 = 0.96

M = 0.04 + 0.05D - 0.01 R + 0 .7 0 M (-1 )
(1.0)
(3.2)
(2.0)
(6.2)

R2 = 0.92

M = - 0 .6 3 + 0.13Y (3.7)
(3.7)

0.04D - 0.01R + 0 .8 8 M (-1 )
(3.1)
(3 4)
(22.1)

R2 = 0.97

M = - 0 .6 3 + 0.13Y (2.8)
(2.8)

0.01D - 0.01R + 0 .7 5 M (-1 )
(0.6)
(2.9)
( 6 .8 )

R2 = 0.98

(4)

(5 )

(6 )

0.01R + 0.71 M ( — 1)
(2.7)
(6.4)

M=
Y=
R=
D=

In
In
In
In

(M-1 /GNP deflator)
(nominal GNP/GNP deflator)
(three-month Treasury bill rate)
(debits/GNP deflator)

Regressions were run with quarterly data, adjusted for first-order autocorrelation. Each equation was run for two time periods
1959-11 through 1982-111 and 1959-11 through 1973-IV. The shorter sample period equations are reported in the second position
in each group. Statistics in parentheses beneath coefficients are t-values.

FRBNY Quarterly Review/Winter 1982-83

Appendix 2: Economic Consequences of Reserve Requirements on Debits
A simple IS-LM model can be used to illustrate some
of the econom ic consequences of putting reserve re
quirements on debits. Table 3 shows the basic model
— equations (1), (2), and (3)— as well as the resulting
reduced-form equations fo r income with and w ithout
reserve requirements on debits, equations (4) and (5). In
this sim ple model, it is assumed that debits can be
divided into three groups: income-related debits, cashmanagement-related debits that result in a shift in the
money demand equation, and all other debits such as
purely financial or non-GNP transactions.*
Equation (4) is the reduced form fo r income with
reserve requirements on debits and balances, and
equation (5) is the reduced-form equation when reserve
requirements are placed on balances only. Comparing
the coefficients on autonomous expenditures (X) in
equations (4) and (5), it can be seen that the numera
to r is the same in both cases but the denom inator is
larger in equation (4). Thus, when there are reserve
requirements on debits, a given increase in auton
omous expenditures w ill have a sm aller impact on
income than when reserve requirements are on bal
ances only. The same holds true fo r shifts in the
demand fo r money (Z). This is because, when income
begins to increase in response to an exogenous shock
(an increase in autonomous expenditures or a reduc
tion of money demand), it begins to absorb reserves
and, given the fixed supply of reserves, the money
stock is reduced.
An increase in cash management debits (CD) at
first glance would seem to reduce income because it
would absorb reserves. But, since it results in a re
duction of the demand for money of the same size,

* For the sake of ease of presentation, it is assumed that the
volume of income-related debits is equal to the level of income.
In practice, however, there would be a proportionality factor
involved that would not alter the results as long as it remained
constant. The nonincome-related debits are assumed to be
exogenous

there would be no effect on income as long as the
reserve ratio on balances (m) is equal to the reserve
ratio on debits (y). Hence, if more aggressive man
agement of cash balances is expected to reduce the
demand fo r money further in future years, it m ight be
worthw hile to consider setting the same reserve re
quirem ent on balances and debits. For this particular
problem, the level of the reserve requirement ratios
does not matter, only whether or not they are e q u a l.t
However, for changes in autonomous expenditures
or shifts in money demand not related to a greater
volume of cash management debits, a greater reserve
ratio fo r debits than for balances would help stabilize
income. For example, looking at the coefficient on X,
the greater the reserve requirement on debits (y), the
sm aller the im pact a change in autonomous expendi
tures w ill have on income. If y is set equal to m, how
ever, the reserve ratios cancel out of both the numer
ator and the denom inator and it does not matter what
level is set.
Thus, how to set the reserve ratios remains an open
question. If it is felt that the m ajor problem will be a
greater volume of cash management debits being used
to lower money demand, then it would pay to set the
reserve ratios at the same level. On the other hand, if
the m ajor problems are likely to be other types of
shifts in the demand for money or changes in autono
mous expenditures, then the reserve ratio on debits
should be set higher than the one on balances. This,
of course, was the position taken by the 1932 Commit
tee on Bank Reserves. Finally, if debits other than
those related to income and cash management (OD)
are very large and volatile, it might be better not to
place reserve requirements on debits.
t It is im plicitly assumed that the cash management debits are
overnight investments only. Thus, the reduction of balances
on a daily average basis equals the increase in cash manage
ment debits. To the extent that the investments are
longer than overnight, the reduction of balances would be
greater than the increase in daily average debits.

Table 3

Comparison of Reduced-Form Equations for Income with and without Reserve Requirements on Debits
(1 )

M = — ar + b Y + Z

M = narrow money stock
r = the interest rate
Y = income
Z = money demand shift
x _ autonomous expenditures

(2)

Y = — cr + X

(3)

i
R = mM + yY + yCD + yOD

(4)

c
ma
mc
X — ------ r---------— - —
Z
Y = ------ :---------j— -------- RH--------- :---------r~r—
yc + ma + bmc
yc + ma + bmc
yc + ma + bmc
________ y£______ C D — _______— ______ OD
yc + ma + bmc
yc + ma + bmc

/c \

y —
c
r -}ma
x —
mc
~ ma + bmc
ma + bmc
ma + bmc

-7

R = total reserves, determined exogenously
CD = cash management debits
QD = debjts for Qther {han income
or cash management
m = reserve ratio on deposits
y = reserve ratio on debits
a, b, c = structural parameters

FRBNY Quarterly Review/Winter 1982-83

Impact of IRAs on Saving

The Congress passed legislation in 1981 broadening
the eligibility criteria governing individual retirement ar
rangements. IRA contributions for 1982— the first year
that the new rules applied— may have reached $24 bil
lion, substantially more than the estimated $4 billion
placed in IRAs in 1981 under the old legislation.
Such a popular program (Appendix 1) could have vast
implications for financial markets, in both the short
and long term, and for the total volume of saving in
the economy.
Indeed, one of the aims of the new law was to in
crease the amount of saving in the economy. This
article considers to what extent the 1981 IRA legisla
tion is likely to satisfy this objective. While personal
saving has been increasing since the change in the
law, it is not possible to say whether the two are really
connected. The increased availability of IRAs in 1982
may have contributed to the expansion of saving, but
other factors, such as a desire to hold greater pre
cautionary balances during the recession, may have
played a role as well.
In fact, our analysis shows that the individuals who
were affected most by the new legislation are those
who on average have substantial accumulated wealth
and who save considerable sums each year. Their con
tributions do not have to and may not consist of ad
ditional saving. Instead, these contributions may reflect
shifts of funds from other assets that these house
holds already hold or simply the placement of saving
that would have occurred anyway.
Thus, the amount of new saving induced by the ex
tended IRA program may be significantly smaller than
the level of IRA contributions might suggest.

FRBNY Quarterly Review/Winter 1982-83

Impact on saving: IRA incentives
The incentives to save embodied in the IRA program
consist of the tax deferral on the contribution itself
and on the earnings from it. There are two advan
tages for delaying the taxation of income until retire
ment. Deferred taxes on both annual contributions
and their subsequent earnings accrue interest, part of
which the individual may keep. The value of these de
ferred taxes can be substantial. Consider a 35-year-old
individual in the 50 percent tax bracket. The present
value of the tax saved on maximum annual contribu
tions over thirty years (figured at a 15 percent interest
rate) is about $12,000. The present value of the tax
saving on the interest earnings of this investor is more
than $88,000 (Appendix 2). in addition, since retire
ment income is likely to be lower than that earned
during an individual’s working years, the applicable
marginal tax rate in retirement might be lower. Thus,
the expected rate of return from an IRA contribution
should be higher than the yield from the same contri
bution to an identical but nontax-deferred instrument.
Households may take advantage of this higher re
turn, however, without increasing saving. For exam
ple, individuals with sufficient non-IRA assets might
shift these assets into IRAs rather than increase
saving. Again, consider a 35-year-old individual in the
50 percent tax bracket with at least $2,000 in non-IRA
assets. These assets, invested at 15 percent, earn
7.5 percent per year after taxes. An investment of
these funds in an IRA, however, would earn an after
tax yield of at least 14 percent (Appendix 2). Abstract
ing from the illiquidity of an IRA, it would be advanta
geous for this individual to shift $2,000 of existing

Table 1

Eligible Individuals Contributing to Individual Retirement Arrangements (IRAs)*
In percent
Annual income
(dollars)

1975

1976

1977

1978

1979

1980|

Less than 5,000 ................................
5.000-9,999
10.000-14,999
15.000-19,999
20.000-49,999
50,000 and a b o ve .............................

0.11
0.94
2.40
4.58
19.28
33.92

0.20
1.26
3.05
5.24
21.04
41.94

0.20
1.33
3.38
5.97
23.65
45.87

0.20
1.50
3.65
6.19
24.44
49.82

0.18
1.15
3.07
5.89
23.01
51.01

0.11
0.82
2.55
5.45
22.57
48.78

$ All income g ro u p s ...........................

2.7

3.7

4.6

5.2

5.3

5.6

* Assumes the same proportion eligible in each year where eligibility is defined as not being covered by another pension plan,
t Preliminary data.
t Weighted average.
Sources: Estimated by authors using data reported in Internal Revenue Service, Statistics ot Income: Individual Income Tax Returns;
Bureau of the Census, Perspective on American Husbands and Wives (Special Studies Series No. 77); Report of the President's Commission
on Pension Policy.

Tabie 2

IRA Contributions in Perspective

Year

Eligible number
of individuals
(millions)

1975
1976
1977
1978
1979
1980
1981

52.7
54.2
56.4
58.1
59.9
61.1
62.3

Arrangements
established
(millions)
1.3
1.9
2.5
2.7
t
*
t

Potential level of
IRA contributions*
(billions of dollars)
54.6
56.1
57.8
62.6
64.9
66.2
67.5

Total
contributions
(billions of dollars)

Tax revenue
lossf
(billions of dollars)

1.4
2.0
2.6
3.0
3.2
3.4
3.8§

t
0.6
0.7
0.9
1.0
1.2
1.3§

* The potential amount of IRA contributions in a given year was estimated by multiplying the number of eligible workers in each income class
by their respective maximum permitted annual IRA contribution,
t Tax revenue loss reflects only deductions for contributions not deferred tax on interest earnings,
t Not available.
§ Estimated from preliminary data.
Sources: Estimated by authors using data reported in Internal Revenue Service, Statistics of Income: Individual Income Tax Returns;
Bureau of the Census, Perspective on American Husbands and Wives (Special Studies Series No. 77); Report of the President's Commission
on Pension Policy, and Joint Committee on Taxation.

assets into an IRA. After this is done, he or she would
still earn only 7.5 percent after tax on any additional
saving. Consequently, for this individual there is no
incentive to increase saving .1
1

The possibility that households shift assets in response to tax-incentive
savings programs may explain to a large extent the Canadian
experience with these types of programs; see Gregory V. Jump, "Tax
Incentives to Promote Personal Saving: Recent Canadian Experience",
Saving and Government Policy (Federal Reserve Bank of Boston,
Conference Series No. 25, October 1982), pages 46-64.

Alternatively, since both the current and expected
future incomes earned on their existing assets, now
including IRAs, are higher, these investors might ac
tually increase consumption. In other words, since their
stock of wealth is accumulating faster because of the
earnings from the deferred taxes, IRA investors might
decrease their rate of saving.
Although such households may not increase saving
to fund an IRA, reallocation of their wealth into IRAs

FRBNY Quarterly Review/Winter 1982-83

Table 3

Distribution of IRA Contributions among Income Groups
In millions of dollars; numbers in parentheses represent percentage shares of total for the year.
Annual income
(dollars)
Less than 5,000 .........................

(50.0)
152.2
( 1 0 .6 )

(3.8)
282.1
( 8 .8 )
370.2
(11.5)
1,872.1
(58.1)
557.0
(17.3)

14.5
(0.4)
57.5
(1.7)
215.1
(6.3)
324.3
(9.6)
2,045.7
(60.5)
726.4
(21.5)

1,436.4
( 1 0 0 .0 )

1,968.5
( 1 0 0 .0 )

2,447.8
( 1 0 0 .0 )

2,969.3
( 1 0 0 .0 )

3,223.5
( 1 0 0 .0 )

3,383.5
( 1 0 0 .0 )

(18.8)
182.1
(12.7)

20,000-49,999 .............................

19.5
( 0 .6 )

1980

40.6
(1.4)
152.1
(5.1)
370.2
(12.5)
311.3
(10.5)
1 ,6 6 8 . 0
(56.2)
427.1
(14.4)

10,000-14,999 .............................

All income g ro u p s .....................

1979

43.7
( 1 .8 )
123.5
(5.1)
364.0
(14.9)
273.1
( 1 1 .1 )
1,343.0
(54.9)
300.5
( 1 2 .2 )

( 6 .6 )

50,000 and a b o v e .......................

1978

39.0
( 2 .0 )
119.3
( 6 .1 )
316.3
(16.1)
263.5
(13.4)
1,003.4
(51.0)
227.0
(11.5)

17.5
( 1 .2 )

5,000-9,999 .................................

15,000-19,999 .............................

1977

1976

1975

1 2 2 .6

Source: Internal Revenue Service, Statistics of Income: Individual Income Tax Returns.

might raise the share of assets being held for retire
ment. The higher relative return on an IRA may per
suade individuals to hold fewer assets for the near or
medium term, e.g., a car or house, and more assets for
retirement.
IRA contributions can come from sources other than
existing assets and still not constitute increased sav
ing. For instance, an IRA contribution can be financed
by borrowing through a personal loan or against some
other asset, although the IRA itself may not be used
as collateral.2 Even payroll deduction contributions
do not necessarily represent new saving. Although
the deposit added to an IRA is drawn from current
income, participants may correspondingly reduce other
saving from income or liquidate assets to finance
consumption.
Another source of funds is the tax saving associat
ed with an IRA contribution. An individual who con
tributes to an IRA pays less in taxes for that year. This
tax saving may be used, in part, to finance an IRA.
These lower taxes represent a transfer from the Federal
Government to households. While the Government’s
saving declines [i.e., the Government’s deficit rises), by
the associated IRA tax loss, household aftertax income
rises by the same amount. With individuals saving ex
actly their additional aftertax income, net saving for the

Because of the tax deductibility of interest payments, individuals in
high enough tax brackets may be able to make a profit by borrowing
funds to place in tax-deferred IRAs.

FRBNY Quarterly Review/Winter 1982-83

economy as a whole— the sum of private and Govern
ment saving— is unchanged.
People who do not have enough assets or cannot
borrow to fund IRAs would obtain the higher yield of
an IRA only by increasing saving. Whether they do
so depends on the responsiveness of their saving to
the expected rate of return and the illiquidity of IRAs.
There have been many attempts to estimate the rela
tionship between interest rates and personal saving.
While some researchers have found that people tend
to save more when the rate of return is higher, others
have found just the opposite— that people tend to save
less when the rate of return is higher. There has been
no definite indication that any additional saving is
generated by an increase in the rate of return .3
Moreover, even if individuals would tend to save
more in response to a higher rate of return, they may
not necessarily save more because of the availability
of IRAs which are less liquid than other types of as
sets. An IRA could have drawbacks if it had to be
cashed before retirement. Except in special cases, e.g.
disability or death of the investor, the drawdown of an
IRA before age 591/2 is subject to a 10 percent penalty
as well as to the payment of ordinary income tax on
these withdrawals. Individuals might be better off in-

William Jackson, "Saving and Rate of Return Incentives: Estimates
of the Interest Elasticity of Personal Savings” (Congressional Research
Services, Report No. 81 -198E, 1981).

vesting in a taxable asset that carries no penalty for
withdrawal rather than an IRA if they expect to need
these funds before the earnings on the deferred tax
exceed the IRA penalty (Appendix 2). Thus, individuals
who want to retain access to their assets in the near
future may decide against contributing to an IRA be
cause of its illiquidity. For instance, people may save
as a precaution against unexpected declines in income
or to accumulate the wherewithal to purchase highpriced items or services. Both of these motives may
require saving to be held as assets which can be con
verted to cash more readily than IRAs.
The combination of the illiquidity of IRAs and the
possibility of funding an IRA contribution by shifting
assets suggests that the liberalization of the IRA law
may have only a limited impact on saving. Those in
dividuals with sufficient assets to shift into IRAs may
do so without expanding saving. People without
enough assets may decide that the attractive return
on an IRA is not sufficient compensation for its illi
quidity. In addition, even households who participate
in the IRA program eventually may choose to discon
tinue making contributions when the liquid share of
their assets reaches a minimum level.
Interestingly, another relatively new saving-incentive
program, known as the 401 (k) deferred compensa
tion plan, may encourage more new saving than the
IRA program. Participants in 401 (k) plans may be per
mitted to borrow against their funds. In addition, par
ticipants may be able to withdraw funds for several
purposes before their retirement. Thus, 401 (k) assets
are not so illiquid as IRAs. Moreover, an individual may
elect to have his or her employer defer as much as
25 percent of his or her income, up to a maximum of
$30,000, to a 401 (k). For example, a person earning
$50,000 could contribute $12,500 to a 401 (k), 6.25
times the maximum IRA contribution. In general, per
sons with an annual income above $8,000 can con
tribute more to a 401 (k) than to an IRA. These people
may find that they -do not exhaust the attractive re
turn of a 401 (k) by shifting assets. Thus, they may
decide to increase their saving as well as to reallo
cate assets. Because of these advantages, 401 (k)s
might be expected to grow strongly, perhaps surpas
sing IRAs, as more firms offer them to their employees.
Historical evidence
While last year’s tax legislation changed retirement
arrangement rules in several respects, the past record
of IRAs provides some insights into their likely growth
and possible savings impact.
Participation rates have been very low. In no year
did participation exceed 6 percent of eligible indi
viduals (Table 1). Part of the explanation may be that

most eligible persons earned a low income and, con
sequently, were in a sufficiently low tax bracket to
make the IRA tax incentive relatively small. Also, the
illiquidity of IRAs may have dissuaded some individu
als from participating. This last reason might explain
why only about half of eligible individuals with in
comes greater than $50,000 a year contributed to IRAs.

Table 4

Average Holdings of Liquid Assets by
Income Group, 1977
In dollars

Income group

Maximum IRA
contribution
minus associated
tax saving

Less than 3,000 ..
3,000-4,999 ..........
5,000-7,499 ..........
7,500-9,999 ...........
10,000-14,999
15,000-19,999
20,000-24,999 , . ,
25,000 and above .

1,500
1,290
1,260
1,245
1,215
1,170
1,125
915

Average
holding of
liquid assets
2,650
2 ,1 0 0

3,300
3,700
5,100
5,500
6,700
12,700

Sources: Estimated by the authors using data reported in
Board of Governors of the Federal Reserve System, 1977
Consumer Credit Survey; Internal Revenue Service, Statistics of
Income: Individual Income Tax Returns.

Table 5

Estimated IRA Contributions in 1982
January-October; by institution

Financial institution

Billions
of dollars

Commercial banks ..................................................
Mutual savings b a n k s ..............................................
Savings and loan associations...............................
Credit u n io n s .............................................................
Mutual funds ............................................................
Life insurance com panies........................................

7.2
1.1
6.5
0.7
2.6
1.5

Total ...........................................................................

19-6

Sources: Estimated by the authors using data reported by
Board of Governors of the Federal Reserve System, Statistical
Release H.6 ; Federal Home Loan Bank Board; Investment
Company Institute, American Life Insurance Council; Credit
Union National Association.

FRBNY Quarterly Review/Winter 1982-83

Table 6

Impact of 1981 Legislation on Potential IRA Contributions
Before and after the enactment of the Economic Recovery Tax Act of 1981 (ERTA)

Annual income
(dollars)
Less than 7,500 .........................
7,500-14,999 ................................
15.000-22,999
23.000-29,999
30.000-74,999
75,000 and a b o v e .......................
All income groups ...................

Number of eligible individuals
m illions
Before ERTA
After ERTA
24.3
12.3
13.8
6.0

Potential level of IRA contributions*
(billions of dollars)
Before ERTA
After ERTA

28.6
17.8
23.0
13.5
26.6
1.9

6 .6

0.5
63.5

111.4

10.7
14.9
21.8
9.8
10.7
0.8
68.7

12.6
21.8
48.0
28.5
56.6
4.0
171.5

* The potential amount of IRA contributions in a given year was estimated by multiplying the number of eligible workers in each income class
by their respective maximum permitted annual IRA contribution.
Sources: Estimated by authors using data reported in Internal Revenue Service, Statistics of Income: Individual Income Tax Returns-,
Bureau of the Census, Perspective on American Husbands and Wives (Special Studies Series No. 77); Report of the President's Commission
on Pension Policy.

would have amounted to less than 11 percent of their
average liquid assets— the assets generally easiest to
shift (Table 4).6 In contrast, for individuals earning less
than $20,000 a year, a maximum IRA contribution net
of the associated tax reduction would have been closer
to 39 percent of their liquid assets .7
Despite the large amount of assets individuals held,
they still may have increased their saving to fund IRAs.
They may have wanted to retain liquidity of their ex
isting assets for near-term purposes. Nonetheless, it
can be concluded that in the past most IRA contrib
utors were part of the income group that typically had
enough assets to fund IRAs without saving more. To
the extent that they held these assets solely for retire
ment, they most likely would have shifted them into
IRAs instead of increasing saving. Analysis of recent
data suggest that these conclusions may apply to IRA
contributions made in 1982 as well.

Reflecting the relatively low participation in this
program, the total IRA contributions in any given year
were small compared with their potential level.4 In
1975, for instance, $1.4 billion out of a possible $54.6
billion was placed in 1.3 million IRAs (Table 2). In
1981, IRA contributions totaled $3.8 billion compared
with the maximum permitted of $67.5 billion. Between
1975 and 1981, annual IRA contributions averaged
4.4 percent of their potential level.5
There is no hard evidence, however, on the amount
of new saving that was stimulated by IRAs. Investors
may simply have shifted assets. Indirect evidence sug
gests that this was at least a possibility for most IRA
contributors. More than two thirds of all IRA contri
butions between 1975 and 1980 were made by indi
viduals with over $20,000 in annual income (Table 3).
For these individuals, a maximum annual IRA contri
bution typically would not have represented a large
share of their wealth. For instance, a $1,500 annual
IRA contribution minus the associated tax reduction
4

The potential amount of IRA contributions in a given year was estimated
by multiplying the number of eligible workers in each income class by
their respective maximum permitted annual IRA contribution.

The Government’s tax loss associated with the annual deduction of
IRA contributions was correspondingly small, never exceeding
$1.3 billion a year. The tax reduction typically represented about a
third of annual IRA contributions. Besides the tax losses produced by
annual IRA contributions, the Government also lost revenue because the
interest earned on outstanding IRA funds was tax deferred. By 1981,
annual interest on IRA accounts amounted to about $5 billion. Applying
a tax rate of one third produces an estimated revenue loss of about
$1.7 billion in that year.

FRBNY Quarterly Review/Winter 1982-83

Recent expansion of IRAs

Subsequent to the enactment of the Economic Recov
ery Tax Act of 1981 (ERTA), IRA contributions in 1982
have exceeded their level in any prior year. Based on
6

These shifts could continue for many years. Saving that would occur
in any case could replenish these assets as well as fund IRAs.

Besides liquid assets, many individuals have other forms of wealth.
The distribution of ownership of these other assets also tends to be
tilted toward upper income groups. Thus, for individuals with at least
$20,000 in annual income, a maximum IRA contribution may have
required a much smaller share of assets to be shifted than our calcu
lations with liquid assets suggest.

several surveys, we estimate that from January through
October, IRA contributions at commercial banks, mu
tual savings banks, savings and loan associations,
credit unions, mutual funds, and life insurance com
panies amounted to $19.6 billion (Table 5).8 Continued
growth at this rate over the balance of the year would
result in new IRA contributions in 1982 of $23.5 bil
lion .9 Several factors, however, may cause the amount
of IRA contributions for the year to be above or below
this figure. On the one hand, the pace of monthly IRA
contributions at commercial banks and mutual savings
banks has slowed since April. Continuation of such a
slowdown among all financial institutions would lead
to a lower level of IRA contributions for the year. On
the other hand, many firms are beginning to offer IRAs
through voluntary payroll deduction plans. These may
encourage people to participate. In addition, as the
law permits an IRA contribution for a given tax year to
be made up to April 15 of the following year, some
individuals may be postponing participation to retain
the liquidity of their saving until the last moment.
The rise in IRA contributions followed an increase
in potential IRA contributions. Since participants of
an employer-provided pension are now permitted to
contribute to an IRA, the number of individuals eligi
ble for an IRA expanded by about 75 percent, from
63.5 million people to 111.4 million people (Table 6).
In addition, the increase in the maximum annual IRA
contribution per worker— by about a third on average—
enlarged the potential level of IRA contributions. Taking
account of the greater eligibility and higher maximum
annual contribution, the aggregate pool of new funds
that can be placed in IRAs more than doubled, from
$68.7 billion to $171.5 billion.
Relative to their potential level, the annualized
amount of IRA contributions was larger in 1982 than
it was in earlier years— about 14 percent of potential in
1982, compared with under 6 percent between 1975
and 1981. Does this mean that the IRA program this
year attracted more individuals from lower middleincome groups who might actually need to save to
set up an IRA?

Besides placing their IRA contributions in the special IRA accounts
established by financial institutions, individuals may place their con
tributions in any other type of qualified investment. Data on these IRA
contributions may not be included in the available surveys. However,
the level of these contributions is not considered to be significant in
the aggregate.
The Treasury’s projection of the associated tax loss of $2.5 billion
for 1982 seems to reflect an underestimate of the growth of IRAs. Using
the relationship between tax loss and IRA contributions during the
1970s— adjusted for the 1982 10 percent individual tax cut— the tax loss
resulting from IRA contributions made in 1982 most likely will be
between $7 billion and $ 8 billion.

The expanded eligibility affected individuals earn
ing $30,000 or more annually to a greater extent than
others. (Most people in this group are covered by
pension plans and were not eligible under prior legis
lation.) In the past, this income group had the highest
rate of participation in the IRA program. Now the
number eligible in this group is four times larger than
before the new law. These individuals also own a large
share of assets with which they might fund IRAs
(Table 7). In contrast, among individuals who earn
less than $30,000 a year eligibility increased by only
47 percent.
While information on IRA contributions in 1982 by
income group is not yet available, calculations can
be made to ascertain whether increased participation
rates are needed to explain the rise in IRA contribu
tions. Alternatively, the expanded eligibility, holding
participation rates constant, may provide the answer.
When the 1980 participation rates for different in
come groups are applied to our estimate of eligible
individuals in 1982, the amount of IRA contributions that
results is about $18 billion, just below the annualized
1982 level. Thus, the primary reason for the program’s
apparently greater appeal seems to be the fact that
ERTA expanded IRA eligibility the most for the in
come group that in the past had the highest rate of
participation. It is not known whether these individuals
increased saving to fund IRAs. However, since this
group contains those individuals who already own

Table 7

Average Holdings of Liquid Assets by
Income Group, 1982
In dollars

Income group
Less than 4,500 ----4,500-7,499 ...............
7,500-11,249 .............
11,250-14,999 ...........
15,000-22,499 ...........
22,500-29,999 ...........
30,000-37,499 ...........
37,500 and above . . .

Maximum
IRA contribution
minus associated
tax saving
2 ,0 0 0

1,760
1,730
1,710
1,680
1,630
1,580
1,340

Average holding
of liquid assets
3,900
3,100
4,800
5,400
7,400
8 ,0 0 0

9,800
18,540

FRBNY Quarterly Review/Winter 1982-83

many assets, the chances that they increased saving
in response to the availability of IRAs are small.
Concluding remarks
Although only a year has passed since the legislated
expansion of the IRA program, some observations can
be made concerning its impact on household saving.
IRA contributions in 1982 may total about $24 billion.
As a percentage of potential level, these IRA contribu
tions are about twice as large as the contributions of
earlier years. Much of the improved popularity prob
ably reflects the fact that the liberalization of eligibility

requirements affected mostly income groups with the
highest participation rates in the past. Individuals in
these groups on average already have accumulated
assets that may be used to fund IRAs. IRA contribu
tions that reflect only shifts of assets do not constitute
increased saving. Thus, the gain in new saving may be
well below the level of IRA contributions. However,
shifts of assets to fund IRAs decrease tax revenue.
For example, the tax loss resulting from IRA contribu
tions in 1982 may fall between $7 billion and $8 billion.
Consequently, the IRA program may not be the most
effective policy approach to stimulate saving.

Robin C. DeMagistris and Carl J. Palash

Appendix 1: Savings Incentive Plans
The Economic Recovery Tax Act of 1981 (ERTA) signi
ficantly increased the availability of IRAs. As of January
1982, any employed person under 701/2 years of age
is eligible to open a tax-deferred individual retirement
arrangement (IRA). The original legislation creating
IRAs— the Employee Retirement Income Security Act
of 1974 (ERISA)— had limited the availability of IRAs
to persons not covered by any other retirem ent plan.
In addition, the new law raised the maximum annual
deductible IRA contribution to $2,000 or 100 percent
of earned income, whichever is less. ERISA had set the
ceiling at $1,500 or 15 percent of earned income. An
amendment effective in 1977 raised the maximum de
duction fo r an eligible individual with a nonworking
spouse to $1,750. ERTA increased this latter ceiling
to $2,250.
Few restrictions have been placed on the type of in
vestments held as IRAs. Three types of IRAs were
established: accounts at financial institutions, annuities
offered by insurance companies, and retirement bonds
issued by the Treasury.* The accounts must be ad
ministered as trusts by a financial institution or other
organization approved by the Treasury. Allocation of
funds among asssts within these trusts can be arranged
by individual investors. However, IRA monies cannot
be used to purchase life insurance or collectibles.
* Retirement bonds were discontinued in April 1982.

FRBNY Quarterly Review/Winter 1982-83

Sim ilar to an IRA, a Keogh or H.R.10 plan, estab
lished by law in 1962, allows a self-employed individual
to deduct annually a certain amount of earned income
fo r investment and defer the tax on it as well as on
interest earnings until retirement. ERTA increased the
ceiling on deductions from $7,500 to the lesser of
$15,000 or 15 percent of yearly income through 1983.
The ceiling is scheduled to be even higher thereafter.
The 401 (k) or deferred compensation plan is an
arrangement which is part of a firm ’s profit-sharing
or stock bonus plan. The 401 (k) was created by a
change in the tax law in 1978, but only recently has
the Internal Revenue Service issued guidelines govern
ing these plans. An individual may choose to have his
or her em ployer make payments as contributions to a
trust on his or her behalf. These payments may repre
sent up to 25 percent or $30,000 of the pa rticipa nt’s
annual income. For many individuals this may be a
larger proportion than the $2,000 lim it set fo r IRA con
tributions. Participants may borrow against 401 (k) funds.
They may also w ithdraw th e ir funds w ithout penalty
before attaining 591/2 years by meeting a need or
“ hardship” requirement. Further, distributions from
401 (k) plans may qualify fo r the favorable tax treatment
of ten-year averaging not afforded to IRA distributions.
Thus, 401 (k)s are less illiquid and, because of ten-year
averaging, can provide a higher return to investors than
IRAs.

Appendix 2: Improved Rate of Return from an IRA
To measure the improvement in rate of return from
an IRA, two hypothetical investment choices may be
compared. The first investment choice is an annual
$2,000 deposit in a fully taxable instrument earning
15 percent per year. The alternative investment is an
annual deposit of $2,000 minus the part financed by the
tax deduction in a tax-deferred account also earning
15 percent annually. The m aturity of both investments
was set at the retirem ent age of the investor which was
assumed to be 65 years. It was also assumed that no
funds would be withdraw n prior to retirem ent and that
the income tax faced by the investor during the w ork
ing years was constant. The future value of each of
these contributions was obtained.
Then, assuming a fifteen-year period of retirement,
an annual income stream from an annuity based on
the accum ulated funds was derived. Since the IRAtype investment is taxable upon withdrawal, appropri
ate tax rates were applied to the retirement stream
generated by IRA funds to obtain an aftertax income
stream. Both a marginal tax rate based solely on in
come derived from the IRA and a tax rate based upon
the rate paid during the working years were used.*
The income stream resulting from the hypothetical
annuity based on the non-IRA investment was not
taxed. In fact, the interest earned on an actual annuity
would be taxed, lowering the stream of retirem ent
income. However, this bias serves only to understate
the spread between the returns on the two invest
ment options. Each retirement annuity was assumed
to earn the same rate as the original investments,
15 percent per year. All compounding was annual.
The rate of return on each investment was obtained
using the stream of annual outlays and the retirement
income stream.
The gains in return from annual IRA contributions
can be substantial. For investors currently in the
50 percent tax bracket, the rate of return can be double
or triple that available on a taxable investment (de
pending upon the holding period of the IRA) when
retirement income tax is based solely on IRA funds
(Table A). Interestingly, the oldest investors in this
income group gain the most from IRA contributions
because the retirem ent income stream generated by
IRA funds is sufficiently small as to be tax free.
If investors faced the 50 percent tax rate both
before and after retirement, the return is tw o-thirds
higher to about double that otherwise available, again
* Assuming an inflation rate ot 10 percent per year of investment
and perfect indexing of current tax rates, the tax rate based only
on IRA income was calculated. In most cases, the retirement
tax rate was significantly lower than that faced while working.

Table A

Expected Improvement in Aftertax Rates of Return
from IRA Investment*
In percent

IRA investment
(years to retirement)

Retirement tax
Retirement tax
rate based
rate equals
only on IRA
working
income
tax rate
Marginal tax rate during working
years (percent)
20
35
50
20
35
50

40 ......................................... 21
30 ......................................... 21
20 ......................................... 22
....... 33
10

54
56
59
81

116
124
140
201

21
20
18
13

48
45
40
29

95
92
85
65

* Improvement expressed as a percentage increase over
taxable yield.
Table B

Value of Each Type of Tax Saving when IRA is
Held to Retirement*
In dollars
IRA invest
ment (years
to retirement)

Contribution
Interest
Marginal tax rate during working years (percent)
20
35
50
20
35
50

40 .......... ....3,300
30 .......... ....3,220
20
....2,990
10
....2,260

7,010
6,740
6,060
4,350

12,59025,510 77,060
11,81015,330 40,220
10,2007,690 17,640
6,860
2,380
4,780

199,700
88,820
33,550
7,900

* Present value is calculated using the aftertax rate of return
based on 15 percent before-tax rate as the discount factor.
Assumes contributions of $2,000 each year.
Table C

Number of Years Non-IRA Investment Return Exceeds
IRA Return
In years
Before-tax rate
of return
(percent)
5 .................................
7 .................................
9 .................................
11
13
15

Marginal tax
rate during working years (percent)
25
30
35
40
45
50
22
16
13
11
10
9

2019
15
12
10
9
8

18
14
11
10
9
8

18
13
11
9

14
11
9

Source: J. Snaiier, "IRAs: A Nonretirement Investment” ,
Federal Resen/e Bank of New York memorandum dated
April 26, 1982.

FRBNY Quarterly Review/Winter 1982-83

Appendix 2: Improved Rate of Return from an IRA (continued)
depending on the length of tim e the IRA investment is
held. Lower income individuals obtain sm aller tax ad
vantages and consequently reap smaller but still signi
ficantly improved rates of return from investing in IRAs.
Their gain ranges from about 20 to 60 percent.
The difference in rate of return between the IRA and
the taxable investment is a consequence of both the
deduction of the annual contribution from taxable in
come and from the deferral of tax on the interest
earnings. Each of these tax advantages can be viewed
as a stream of future payments to the investor. The
present value of the annual deductions and that of the
deferred tax on earnings may be compared.
On average, the value of the deferred tax on the
annual contributions exceeds that of the deferred
tax on interest fo r the first nine to fourteen years
of the investment. For older individuals or those
planning only a short-term IRA investment (subject to
penalties), the present value of the deferred tax on the

FRBNY Quarterly Review/Winter 1982-83

contribution outweighs that of the deferred tax on
interest. For younger investors or investors planning
longer term investments in IRAs, the tax saving on
interest accum ulates rapidly, far surpassing that of
the contribution by the retirem ent age (Table B). In
general, the higher the return on the IRA and the higher
the working tax rate, the faster the tax saving on in
terest overtakes the tax saving on the contributions.
For some investors, however, the higher return avail
able from an IRA may not be sufficient to offset its
illiquidity. The 10 percent excise tax fo r early w ith
drawal reduces the rate of return on an IRA to below
that of a non-IRA asset if funds are withdraw n before
the compounded interest on the deferred income tax
exceeds the penalty. For a 15 percent rate of interest,
drawdown in less than seven to nine years (depend
ing on tax bracket) would make the IRA a less de
sirable investment (Table C). For lower interest rates,
this tim e period can be longer.

Economic Effects of Enforcing
Due-on-Sale Clauses

The housing industry in the United States has been in
a severe slump over the past three years. High interest
rates have sharply reduced both new housing starts
and sales of existing homes. Although the housing
industry recently has begun to show signs of recovery,
existing single-family home sales are still nearly two
million below the peak of almost four million reached
in 1978.
Potential sellers of houses have developed several
methods of “ creative financing” in an attempt to make
their homes more attractive to buyers. The most popu
lar creative financing technique has been the assump
tion of an existing mortgage by the home buyer. This
technique enables the buyer to continue to make pay
ments on the existing mortgage of the house that is
purchased. In 1981, about one million home sales—
almost half of the sales of existing homes— involved
assumptions of existing mortgages.1 Many of these
assumptions took place at interest rates substantially
below market interest rates.
In October 1982, the Congress passed the Garn-St
Germain Depository Institutions Act, which among
other things will permit enforcement of due-on-sale

The author wishes to acknowledge the assistance of David Kroop
and Daniel Rossner of the Legal Department in preparing the box.
1 National Association of Realtors, Attitudes of Real Estate Industry
(November 1981), page 9. This figure includes homes purchased
"subject to" an existing mortgage. This technique is similar to a
mortgage assumption except that liability in event of default lies with
the original owner rather than with the new home buyer.

clauses in many mortgage contracts. These clauses
allow lenders to require full payment of the remaining
mortgage debt when a home is sold (box). The Con
gressional action was a response to actions taken by
the legislatures and courts in several states which
limited due-on-sale enforcement.
This article examines the impact of the Congres
sional action and concludes that stricter enforcement
of due-on-sale clauses could lower mortgage rates
and stimulate housing activity. In addition, the earn
ings of thrift institutions that issued low-rate mortgages
may be improved substantially by the repayment of
these loans. However, there are some losers, namely,
homeowners who formerly could offer attractive financ
ing via assumptions of old mortgages issued at low
interest rates. The net result most likely will be a bene
fit to home buyers as a group, lending institutions, and
the construction industry and a loss to homeowners
who are no longer released from the due-on-sale
clauses in their mortgages.

The value of an assumable mortgage
At times of high interest rates, assuming a lowinterest mortgage is similar to obtaining a new loan at
an interest rate below the market rate. A loan carrying
an interest rate lower than the market rate on new
loans is a valuable commodity to someone who needs
a loan. That is, a home buyer would be willing to pay
a premium to obtain such a loan. In 1981, three quar
ters of the mortgages assumed had an interest rate
more than 2 percentage points below the going market
rate.

FRBNY Quarterly Review/Winter 1982-83

The value of the assumable mortgage to a home
buyer is the difference between the present discounted
cost of payment streams on the old assumable mort
gage and a new market-rate mortgage for the same
amount. The table shows some sample calculations of
the value of a below-market-rate mortgage. For exam
ple, the present discounted cost of a $ 100,000, 8 per
cent mortgage with a ten-year maturity is $78,140 if the
current interest rate is 14 percent, whereas a mortgage

of 14 percent has a present discounted cost of
$100,000. This means that a home buyer would be
willing to pay up to $21,860 to assume a $ 100,000
mortgage with an 8 percent interest rate. Ignoring
tax considerations for the moment, a buyer should
be indifferent between ( 1 ) paying $121,860 for a house
and assuming a $ 100,000 ten-year mortgage at 8 per
cent and (2 ) paying $100,000 for the same house and
obtaining a $100,000 ten-year mortgage at 14 percent.

Present Discounted Value of a Below-Market-Rate Mortgage
In dollars

Market rate
(percent)

years

8 percent original rate
15 years
2 0 years

Original Mortgage Rate and Years Remaining to Maturity
10 percent original rate
12 percent original rate
years
15 years
2 0 years
15 years
2 0 years
1 0 years

1 4 ...................

78,140

71,760

67,260

85,110

80,690

77,600

92,400

90,120

88,550

1 6 ...................

72,430

65,070

60,120

78,890

73,170

69,360

85,650

81,720

79,140

1 8 ...................

67,330

59,340

54,200

73,340

66,730

62,530

79,620

74,530

71,350

Entries show the estimated present discounted value of a payment stream for a $ 1 00,000 mortgage with an original
10, or 12 percent contract rate when the market rate is 14, 16, or 18 percent.

Source: Thorndike Encyclopedia of Banking and Financial Tables, Revised Edition (Boston: Warren, Gorham, Lamont, Inc., 1980).

Recent Developments Affecting the Due-on-Sale Clause
Many conventional mortgages have clauses requiring
imm ediate payment of the entire mortgage debt upon
sale of the home. The enforcement of these so-called
“ due-on-sale” clauses was restricted or challenged
by state law, including court rulings, in eighteen
states.* In these states, due-on-sale clauses were gen
erally unenforceable and mortgages could be “ as
sumed” from the previous owners by home buyers.
On June 28, 1982, the United States Supreme Court
ruled that certain due-on-sale clauses could be en
forced. Specifically, Federally chartered savings and
loan associations could, in accordance with a 1976
Federal Home Loan Bank Board (FHLBB) regulation,
require full payment on outstanding mortgages con
taining a due-on-sale clause if the property were trans
ferred or sold. On O ctober 1, 1982, the Congress
enacted the Garn-St Germain Depository Institutions
Act of 1982, which among other things preempts state
* The states are Arizona, Arkansas, California, Colorado, Florida,
Georgia, Illinois, Iowa, Michigan, Minnesota, Mississippi,
New Mexico, New York, North Dakota, Ohio, Pennsylvania,
Utah, and Washington. For a state-by-state summary, see
“ Due-on-Sale— The National Picture” , Mortgage Banking
(October 1981), pages 24-27.

FRBNY Quarterly Review/Winter 1982-83

laws restricting the enforcem ent of due-on-sale clauses.
The measure was signed into law by President Reagan
on October 15, 1982.
The act gives lenders the right to enforce due-on-sale
clauses contained in real property loan agreements in
most cases, notwithstanding state constitutional, statu
tory, and ju dicia l restrictions on such enforcement.
However, a state’s restrictions on enforcement of dueon-sale clauses w ill continue to apply until October
15, 1985 to loans made by lenders during a “ window
period” specified as the period prior to the a ct’s en
actment when the state’s restrictions were in effect.
State legislatures may act prior to O ctober 15, 1985
to regulate the terms of w indow-period loans made by
non-Federally chartered lenders and, in so doing,
can extend state restrictions on the enforcem ent of
the loan’s due-on-sale clauses to the period after Oc
tober 15, 1985. S im ilar authority to regulate windowperiod loans is given to the C om ptroller of the Cur
rency with respect to national bank loans and to the
National Credit Union Adm inistration Board w ith re
spect to national cred it union loans. No window period
applies to loans by Federal savings and loan associa
tions or Federal savings banks.

(This assumes the buyer is able to finance the $21,860
premium at a rate of 14 percent.) The assumable
mortgage adds $21,860 to the value of that house.
Several factors reduce the value of an assumable
mortgage below the present discounted value of the
difference in payment streams between market-rate
and low-rate assumable mortgages. First, tax consid
erations tend to reduce the value of an assumable
mortgage. An individual not deducting interest pay
ments from income is indifferent between two dollarequivalent payment streams that have differing pro
portions of principal and interest. However, a home
buyer deducting interest prefers a payment stream
with a higher proportion of interest. Assumption of
a low-rate mortgage involves trading off lower interest
payments for higher levels of payments of principal.
Thus, the value of the low-rate mortgage is less to
high tax bracket buyers (itemizing deductions) than
to lower tax bracket buyers.
Another factor to consider is the need to obtain
funds in addition to the assumable mortgage. The
remaining balance on an assumable mortgage might
be substantially less than the value of the house
because of repayments of principal and increases in
home prices. A buyer assuming a mortgage might
have to obtain a second mortgage to finance the dif
ference between the value of the house and the re
maining balance on the assumable mortgage, plus
any premium paid for the assumable mortgage. A
low-rate assumable mortgage combined with a second
mortgage may entail a higher monthly payment stream
for some period due to the shorter maturity of the
assumed mortgage. This increased monthly payment
stream may affect buyer qualification for a mortgage
or create cash-flow problems, both of which would
reduce the value of the assumed mortgage.2

Economic impact of due-on-sale enforcement
The buyers of homes that originally carried low-rate
mortgages probably will be no worse off with the
enforcement of due-on-sale clauses since the benefits
of below-market-rate mortgages are likely to be re
placed by lower housing prices.3 The losers are the
2 See “ Accelerating Inflation and Nonassumable Fixed-Rate
Mortgages: Effects on Consumer Choice and Welfare” , Patric
Hendershott and Sheng Hu, Public Finance Quarterly
(April 1982), pages 158-84.
3 If markets are efficient, the drop in the housing price should
exactly compensate the buyer for the increased present discounted
value of the payment stream from the market-rate mortgage. For
empirical evidence that the value of the below-market-rate mortgage
is capitalized into housing prices, see Kenneth T. Rosen, "Creative
Financing and House Prices: A Study of Capitalization Effects”
(University of California, Berkeley, Center for Real Estate and Urban
Economics), Working Paper 82-52, August 1982.

owners of houses in states that had prevented the en
forcement of due-on-sale clauses, who are no longer
able to capture the value of the low-interest mortgage
when their houses are sold. In short, wealth is redis
tributed by more stringent enforcement of due-on-sale
clauses. State actions restricting due-on-sale enforce
ment produced windfall gains to some home sellers at
the expense of lending institutions. The Depository In
stitutions Act prevents this wealth transfer.4
Effects on the supply and demand for mortgage finance
The reduction of the number of mortgage assumptions
is likely to have a stronger effect on the demand for
mortgage funds by home buyers than on the supply of
funds by lending institutions. The enforcement of the
due-on-sale clause means that, when the holder of the
low-rate mortgage sells the house, the existing mort
gage is repaid in full. This results in a flow of funds
to the lending institution. If the lending institution
channels all these new funds into the mortgage mar
ket, the total supply of mortgage money is unchanged
in the short run.5
Total demand for mortgage funds, however, is likely
to be affected by more widespread enforcement of the
due-on-sale clause. If a due-on-sale clause is enforced,
the price of a house with a low-rate mortgage will not
incorporate a premium attributable to the desirable fi
nancing. Thus, a smaller amount of financing would be
required by buyers of existing homes with due-on-sale
clauses. Initially, then, the total demand for mort
gage funds would fall. If the supply of funds offered
by lending institutions is unchanged, mortgage rates
probably would be lower than they would be without
the increased amount of due-on-sale enforcement.
Effects on thrift institutions
More rigorous enforcement of the due-on-sale clause
should have a favorable impact on the earnings of the
thrift industry. The approximately one million existing
home sales in 1981 that involved assumption of a mort
gage amounted to an estimated dollar volume of as-

4 Those who gain from the Congressional action may have a lower or
higher demand for housing than the wealth losers. For example, if the
sellers of houses with low-rate mortgages had planned to use all their
profits to buy more expensive houses and the gainers (e.g., savings
and loan association stockholders) invested all their gains in Treasury
bills, then housing demand would fall when due-on-sale clauses are
enforced. To the extent that beneficiaries of higher lending-industry
profits are identical to the home sellers, the net wealth effect is
diminished.
5 This is a good assumption for thrift institutions, since most new
lending by savings and loan associations and mutual savings banks
is in the form of mortgages. This assumption may hold even for
banks, if they wish to maintain a constant fraction of their asset
portfolios in mortgages.

FRBNY Quarterly Review/Winter 1982-83 35

sumable mortgages of about $20.8 billion.* Since about
60 percent of the dollar volume of all mortgages made
is held by thrift institutions, about $12.4 billion of mort
gages from thrift institutions was assumed in 1981. This
figure includes Federal Housing Administration and
Veterans Administration (FHA/VA)-insured mortgages,
which do not contain due-on-sale clauses. FHA/VA
mortgages constituted about 20 percent of home mort
gage debt in 1981, declining from about 30 percent in
1972. Allowing that 25 percent of mortgage assump
tions in 1981 involved FHA/VA mortgages gives an es
timate of about $9.3 billion in mortgages containing
due-on-sale clauses which were assumed in 1981.
Data from the National Association of Realtors sug
gest that the average mortgage assumption made in
1981 was 4.3 percentage points below the market rate.
If these assumed mortgages would have been replaced
by mortgages at the then market rate, thrift revenues
would have increased by 0.043 times $9.3 billion, or
about $400 million.7 This amounts to about 7 percent of
their losses in 1981. For 1983, the increase in thrift
revenues will depend on the difference between the
level of market mortgage rates and the rate on mort
gages which will have due-on-sale clauses enforced as
a result of the Congressional action.
Another estimate of the effect of increased due-onsale enforcement on thrift institution earnings is avail
able from a Federal Home Loan Bank Board study.
The FHLBB has estimated that the “ potential earnings
[gains] two years after a nationwide [enforcement] on
due-on-sale clauses run from $1.0 billion to $1.3 billion
for all Federal and state associations” .8 This estimate
is somewhat larger than the one given above, but it is

4 The following assumptions were made: (1) the average age of
assumed mortgages was ten years and the initial maturity was
twenty-seven years, (2) the initial sales price was $32,500, of which
78 percent was financed using a mortgage with a contract rate of
7.50 percent.
* The $400 million figure is an overestimate for at least two reasons.
First, many mortgages with due-on-sale clauses have been renegotiated
at higher rates when the underlying property was sold. The assumed
mortgage is often combined with a second mortgage at a “ blended"
rate between the contract rate on the original mortgage and the
market rate. These types of agreements may continue after the
Congressional action. Second, some mortgages which are not
FHA/VA insured do not contain due-on-sale clauses.
• “ Final Report and Technical Papers of the Task Force on
Due-On-Sale” (Federal Home Loan Bank Board, March 1982), page 2.

not far different since it is based on two years’ worth
of assumable mortgages being replaced by market-rate
mortgages. The FHLBB has projected that, without
these transfers, the number of savings and loan as
sociations encountering “ net worth deficiencies” would
be about 17 percent higher than if the due-on-sale
clause were enforced.9
Effects on the housing market
The enforcement of due-on-sale clauses could have a
favorable impact on the housing market. As noted
earlier, due-on-sale enforcement reduces the demand
for mortgage money since the value of the belowmarket-rate assumable mortgage no longer need be
financed by the buyer. This initially lowers the mort
gage rate, making it less expensive to finance a home
purchase. As a consequence, demand for housing
should increase. The greater demand for housing in
creases its price and encourages new construction.10

Summary
The Garn-St Germain Depository^ Institutions Act,
passed by the Congress in 1982, contains a provision
which will enable many lending institutions to enforce
due-on-sale clauses in mortgage contracts in states
that had previously prohibited such enforcement. The
major impact of the stricter enforcement of due-on-sale
clauses will be a redistribution of wealth from owners
of homes with mortgages (which had due-on-sale
clauses restricted by state actions) to the lending
institutions holding these mortgages. This transfer
of wealth could amount to several hundred million
dollars, depending on the amount by which market
interest rates exceed contract rates on outstanding
mortgages. In addition, the increased enforcement of
due-on-sale clauses may well lower the rate on new
mortgages, thereby increasing the demand for housing
and stimulating home building.

9 These increased profits allow the thrift industry to compete more
readily for funds and thus might increase the supply of money available
for mortgage finance. This effect would reinforce the lowering of
mortgage rates described in the previous section.
10The rise in the price of housing offsets some of the gain to new
buyers from the lowered mortgage rate. Offsetting new construction
is a possible reduction of the supply of existing homes. For any
given market price of houses, the seller who had a low-rate formerly
assumable mortgage obtains less on a home sale when a due-on-sale
clause is enforced.

Howard Esaki

FRBNY Quarterly Review/Winter 1982-83

New York City’s Property Tax
Problems in an Era of Changing
Price Trends
From 1979 to 1981, New York City property tax rev
enues barely grew though inflation and economic
growth pushed property values sharply upward. What
retarded the growth of these tax revenues was the
state-imposed ceiling on the amount of revenues the
city may raise from its property tax. Over this period
the ceiling actually dropped slightly. As a result, the
city lost out on some $1.2 billion in property tax reve
nues it would have been able to collect if the ceiling
had kept pace with property values. While some
analysts might argue that this slow response of prop
erty tax revenues to an increase in property prices was
a healthy restraint on expenditures, the revenue short
fall came at a particularly bad time for New York City.
Not only was the city trying to balance its budget, but
also it was faced with the higher costs of providing
public services as a result of inflation. Under the
present system, the city will continue to lose revenues
during periods of inflation as increases in the ceiling
lag the upturns in property prices. Thus, if respon
siveness of revenues to property price increases is
considered to be a desirable attribute for the tax sys
tem, changes are needed in the method for determin
ing the property tax ceiling.

The ceiling: five-year averaging and full values
The state constitution restricts the amount of prop
erty tax revenues that New York City may raise yearly
Mark A. Willis, on leave from the Bank, is an assistant to New York
City’s Deputy Mayor for Finance and Economic Development.

for operating purposes.1 This ceiling is set at 2Vz per
cent of the full (market) value of taxable real estate
in the city, averaged over the latest five years. With
the amount of property taxes collected by the city at
or near the maximum allowed by the ceiling, changes
in its level effectively determine the city’s ability to
raise revenues from this tax (Chart 1). If the ceiling
grows at a slower rate than property prices, the city
must either delay increasing assessed values or lower
the statutory tax rate.2 If the ceiling rises faster than
property prices, the city can capture this increase in
taxing power through either higher assessments or
tax rates.
The movement in the ceiling from one year to the
next is critically affected by two factors: the constitu
tional requirement for five-year averaging and the
method used to estimate full values. These factors
result in a ceiling which varies widely as a fraction of
the current full value and responds poorly to changes
in property price trends.

1 New York State Constitution, Article VIII, Section 10. Property tax
revenues raised for debt service— approximately two fifths of the total
levy— are not covered by this restriction. These debt-service revenues
have stayed fairly constant lately and so have not had much impact
on year-to-year changes in the total property tax levy. (It should be
noted that the state constitution does place restrictions on the total
amount of debt the city may have outstanding.)
2 Property tax liability is determined by multiplying the statutory tax
rate by the assessed value recorded on the city’s tax rolls for the
particular piece of property.

FRBNY Quarterly Review/Winter 1982-83 37

used to derive full values for the city’s taxable real
estate. To estimate full values, the State Board of
Equalization and Assessment (SBEA) uses market
surveys— appraisals of properties sampled from the
city’s tax rolls. In the past, these surveys were con
ducted relatively infrequently. For the 1979 fiscal year,
for example, the latest available survey related to July
1974. Recently, however, surveys are being conducted
annually and completed within the year. But, even now,
the last two full values used in the average postdate
the latest survey results .3
To bridge this information gap, the SBEA extrap
olates price growth based on a weighted average of
previous property price changes .4 As a result, full
value estimates for recent years are based on price
changes many years out of date. For example, the
fiscal year 1979 ceiling relied on growth rates dating
as far back as 1968.
This lack of up-to-date information on full values
and this reliance on old growth rates to project current
full values produce a ceiling which behaves poorly
during periods of wide swings in property price growth.
The last fifteen years were just such a period for
New York City. Property prices rose rapidly in both
the late sixties and th 3 late seventies but grew very
slowly in the interim period (Chart 2).

Chart 1

Property Tax Levy for O perating Purposes
as a Percentage of the Ceiling
1979-83
Percent

1979

1980

1981

1982

1983

Source: New York City Council Tax-Fixing Resolutions
for fiscal years 1979, 1980, 1981, 1982, and 1983.

Averaging would not be a problem if property prices
were basically trendless. In fact, under such circum
stances, averaging would help smooth out short-term
fluctuations. But, when prices are on an upward
(downward) trend, the averaging requirement imparts
a downward (upward) bias to the ceiling. The
faster prices are rising over the five years spanned
by the average, the lower the ceiling as a percentage
of the latest full value. For example, when property
prices grow at a steady 10 percent per year, the fiveyear average is almost one-sixth smaller than the
latest full value, thus lowering the effective level of
the ceiling from 2.5 percent to 2.1 percent.
Averaging also slows the response of the ceiling to
changes in price trends. New full values are only grad
ually incorporated into the average. As a result, the
average is slow to reflect the new growth rate. For
example, the first year of 10 percent price inflation
following a period of stable prices triggers a mere
2 percent rise in the average. Prices must grow at
the same rate for five consecutive years before the
average will move in step.
Estimating full values
Another source of divergence between property price
movements and those of the ceiling is the method

FRBNY Quarterly Review/Winter 1982-83
Digitized for38
FRASER

A look at 1979-83
Over the fiscal years 1979 to 1981 the ceiling fell
slightly while property prices rose (Chart 3).5 The
main reason for this decline in the ceiling was the
lack of up-to-date data on property prices. Only one
survey was completed during this period— in time
for computing the 1980 ceiling— and it covered an
earlier time when property price growth was slowing,
thus leading to downward revisions of the full values
previously estimated. Five-year averaging created an
additional downward bias, and the ceiling dropped
farther and farther below 2Vz percent of actual full
value. The result was a loss of taxing power of almost
$1.2 billion over these three years.
The ceiling for fiscal years 1982 and 1983 rose
rapidly as survey completions revealed for the first

The lack of information on two years’ worth of full values results
because ( 1 ) the ceiling must be computed in advance of the start of
the fiscal year (July 1) to which it applies and (2) surveys are con
ducted on July 1 while the full values used in the five-year average
are dated as of January 1, the midpoint of the fiscal year.

Property prices are regressed on a time trend to obtain the growth
factor.

Properly taxes levied for operating purposes increased over the 1979-81
fiscal years (the fiscal year for New York City ends on June 30), but
only because the city was able to tap some unused taxing power
(Chart 1).

21/2 percent of actual full values. In 1986, for example,
the average would consist of the actual values for
fiscal years 1982, 1983, and 1984 and projected full
values for the following two years. Even if property
prices were stable from now until 1986, the growth
rate used to project these latter two full values would
still be positive. The result would be “ projected”
values which exceed their actual levels.
Although the new, faster survey schedule helps re
duce the erratic movements in the ceiling, it does not
eliminate them. As long as the underlying trend of
property prices varies over time, the continued use of
the present projection system and of five-year aver
aging ensures that growth of property tax revenues
will diverge from property prices.

time the renewed growth of property prices of a few
years earlier. The five-year averaging requirement,
however, substantially dampened the rise in the ceiling.
Over the 1981-83 period, the ceiling as a percentage
of the latest full value then being projected fell from
95 percent to 84 percent.
Outlook: growth rates of property tax revenues
and property prices likely to diverge again
Property price growth in the city is apparently enter
ing a new, more moderate phase. If property prices
should stabilize, the ceiling would not stop growing
immediately but would keep on increasing for several
more years (Chart 3). As a result, the ceiling would
rise toward the 2Vi percent level and could even sur
pass that level by a small margin. But any “ excess”
taxing power would hardly begin to offset the large
losses earlier in the decade.
Continued growth of the ceiling would result from
both the five-year averaging and the projection sys
tem. Even with prices holding steady at their July
1981 survey level, the five-year average would rise
until all the earlier years with their lower full values
were eliminated. Further pushing up the average
would be full value estimates based on the previous
rapid price growth. In fact, with the present projec
tion method, property taxes may eventually exceed

Resolving the problem
The problems caused by the present projection sys
tem could be reduced greatly through two modifica
tions. First would be a further speedup in survey
processing. Then, only the value for the upcoming
fiscal year would have to be projected to set the
ceiling for that year.
Second would be an improvement in the projec
tion method. The present reliance on an average of
past price movements to project into the future makes
little sense during periods of changing price trends.

Chart 2

P roperty Price G rowth Rates as D eterm ined by the State Board o f E qualization and Assessm ent
1968-81
Percentage growth

16------------------------

1968

1969

1970

------

1971

1972

1973-74

1975

1976

1977

1978

1979

1980

1981

Growth rates for 1968 through 1972 are for calendar years, determined from surveys conducted in January of 1968, 1970, and
1973. The growth rate for 1973-74 is the annual rate of change from January 1973 through July 1974, determined from surveys
conducted in those two months. Growth rates for 1975 through 1981 are for fiscal years (July 1 to June 30), determined from
surveys conducted in July of 1974, 1976, 1978, 1980, and 1981.
Source: New York State Board of Equalization and Assessment, computer printouts of Table A1 on the computation of special
state equalization ratios, selected years.

FRBNY Quarterly Review/Winter 1982-83

A great deal of data is available on current inflation
trends and levels of economic activity. Use of this
more up-to-date information should reduce the likeli
hood of missing changes in price trends and so would
help eliminate this source of divergence between
movements in property prices and those of the ceiling.
Improved estimates of full values would still leave
the ceiling with problems caused by five-year averag
ing. Unless changes are made in this requirement, the
ceiling will continue to fall below 21/2 percent of the
current full value during periods of rising prices and
fail to respond quickly to changing price trends. Such
behavior is particularly troublesome during inflation
ary times .6
The ceiling could in fact be made more inflation
neutral. One way would be to eliminate the averaging
requirement, thus basing the ceiling only on the latest
full value. Another way to deal with the problem would
be to base the average on full values which are ad
justed for subsequent changes in the overall price
level. For example, if the general price level increased
by 10 percent from 1981 to 1982, then the 1981 full
value would be revised up by 10 percent before being
used in computing the 1982 average. With this ap
proach, therefore, year-to-year changes in the relative
price of real estate would be smoothed, but general
inflation would not affect the city’s property taxing
power. Combining either of these options with an im
proved method for estimating the full value for the
6 New York is not alone in having a property tax ceiling which can
timit the growth of revenues below that of inflation. Both California and
Massachusetts, for example, have adopted property tax restrictions
which, not only roll back tax levies, but also establish maximum growth
rates for any subsequent increases. Proposition 13 in California limits
the annual increases for an individual property to 2 percent a year,
unless the property is sold, and Proposition 2 1/2 in Massachusetts sets
the limit on growth at 2 V2 percent a year.

Chart 3

Property Tax C eiling versus 2Vi Percent
of C urrent Full Value
Assuming property prices stabilize beginning in 1982
Billions of dollars
3 .2 0 ---------------------------------------------------------------------------------3.00—

2 1/2

percent of

*C eiling values for 1979 to 1983 are actual (from New
York City Council Tax-Fixing Resolutions for fiscal years
1979 through 1983). Ceiling values for later years are
simulated using current State Board of Equalization and
Assessment procedures and assuming that property
prices are constant in 1982 and later years.

current year would produce a ceiling which holds
closely to 2 1/> percent of the current full value and
reflects more quickly the actual changes taking place
in the full value of the city’s taxable real estate.

Mark A. Willis and Daniel Chall

FRBNY Quarterly Review/Winter 1982-83

China’s Rapid Trade Growth and
Impact on the World Economy

China’s foreign trade tripled in just five years (1977-81)
and is continuing to grow rapidly. This growth rate for
the People’s Republic outstripped that for all other
countries except Mexico, a major new oil exporter.1
By 1981 China’s exports and imports totaled more
than $40 billion. Only six other nonindustrial economies
traded more. However, aside from Saudi Arabia, all
these countries had significant trade deficits, with large
import purchases boosting their trade figures. China,
in contrast, ran a large trade surplus. On exports alone,
the People’s Republic ranked fourth among developing
countries (Table 1).
China’s emergence as a major world trader is clearly
highlighted by the current trade dispute between the
People’s Republic and the United States. In January
America unilaterally set textile import quotas on China,
now its fourth largest textile supplier, when negoti
ations failed to produce a bilateral quota agreement.
China, in turn, banned all further imports of American
cotton, soybeans, and chemical fibers. Although this
ban will cause a relatively small loss to the United
States (at most $300 million), given the low level of
previously anticipated sales of these goods to the
People’s Republic, China has warned it is considering
banning the purchase of other American products.
With total American sales to China about $3 billion in
1982, exceeding Chinese sales to the United States by
over a half billion dollars, further trade bans could
1This paragraph refers only to countries who report their trade figures
to the International Monetary Fund. Chinese trade figures reported in
this article are those released by Chinese trade authorities. Trade
partner figures often differ due to accounting practices, category defi
nitions, timing, and similar discrepancy causes.

potentially have a significant impact on American ex
ports. This article will not deal with the specifics of
this bilateral dispute but rather with the broader pat
terns and implications of China’s emergence as a
major trading country.
China’s sharp trade growth follows the radically
changed economic behavior of the People’s Republic
since the end of Mao Zedong’s rule six years ago.
Since then, China has seen the succession of two
leaders— Hua Guofeng after Mao and Deng Xiaoping
following Hua— as it moved gradually to more prag
matic economic policies. It has formalized its inter
national position by entering the United Nations, the
International Monetary Fund (IMF), and the World
Bank. It normalized relations with the United States
(1979). It also has opened its borders much more
widely to trade. Although China had made move
ments in this direction earlier (President Nixon paid
his historic visit in 1972), the changes initiated in 1977
marked a distinct and dramatic shift in China’s focus
of attention from seeking domestic political “ purifica
tion” during the tenure of Mao to pursuing rapid eco
nomic growth thereafter. China is currently reorienting
its whole economy. Increased foreign trade is an im
portant part of the transformation.
The rapid emergence of China as a major trader in
international commerce raises many important issues:
• China’s trade growth has been both fast and
unexpected. Chinese exports and imports have
grown at varying rates, however, with the Peo
ple’s Republic currently running a large trade
surplus. Are the policies behind this growth

FRBNY Quarterly Review/Winter 1982-83 41

Table 1

Trade of China, Compared with Selected Developing Countries in 1981
In billions of U.S. dollars
Total
trade

Exports

Imports

Trade
balance

C h in a ..............................................................

41.4

2 1 .6

19.8

1 .8

Saudi A ra b ia ..................................................
Mexico ..........................................................
S in g a p o re ......................................................
Hong Kong ....................................................
Korea ............................................................
Brazil .............................................................
Indonesia ......................................................
V enezuela......................................................
United Arab E m irate s.................................
India ...............................................................
Malaysia ........................................................

148.6
50.3
48.6
46.6
45.3
42.7
35.6
32.4
29.6
22.9

113.3

—
-

2 2 .1

35.3
29.1
27.6
24.8
25.1
23.0
13.5

19.6

1 2 .8

2 0 .1

7.8

9.5
15.1

2 2 .8

1 1 .2

1 1 .6

Country

Number of developing countries exceeding
C h in a ...............................................................
Of which: n o n o il.......................................

2 1 .2
2 1 .0
2 1 .8
2 0 .2

19.7

78.0
7.9
6 .6

3.0
4.9
3.3
8 .6
6 .8
1 0 .6

—
—

7.3
0.4

Sources: International Monetary Fund, International Financial Statistics and Direction of Trade Statistics Yearbook 1982.

Table 2

Examples of Special Trade Arrangements
Term
Compensation tr a d e .......................

Export p ro ce ssin g ..........................

Cooperative p ro d u ctio n ...................

Joint ve n ture s...................................

Description
A foreign firm provides machinery, equipment,
technology, supervision, and at times
raw materials and is paid back with finished
products, usually at an agreed-upon price.

Mitsubishi Corp. with two other Japanese
firms agreed to be compensated with talcum
powder by two Chinese mining units.

A foreign firm pays a fee to a Chinese
enterprise to process or to assemble products
for which the former supplies raw materials,
semi processed goods, and other components.

Sanhill Co. of Hong Kong contracted for a
Chinese radio factory to assemble radio
receiver instruments.

A foreign firm provides technical assistance
to a Chinese enterprise to produce an item
under license from the foreign firm. Payment
is made in the form of the goods produced.

Aalborg Vaerft of Denmark agreed
to coproduce ship auxiliary boilers along
with a Chinese shipbuilding firm.

A foreign firm provides equity capital— usually
50 percent or more— with profits distributed
according to capital shares.

Gillette Co. of the United States agreed with a Chinese firm jointly to produce razor blades, plastic
razors, and uncoated blades for industrial use.

Source: The National Council on U.S.-China Trade, Chinese Business Review, various issues.

FRBNY Quarterly Review/Winter 1982-83

Example

and current surplus position the result of shortrun trading decisions or do they reflect more
fundamental change in the economy?

Chart 1

Trade G rowth Rates
• China’s trade growth has been relatively con
centrated among three partners: Japan, the
United States, and Hong Kong. The composition
of China’s trade differs among the three, how
ever. How have these partners fared in terms of
import and export growth? What are the
future prospects for the three as China’s trade
composition shifts over time?

Percentage
Exports

China

• The large population of the People’s Republic
makes it a significant world producer and con
sumer of certain goods. China may thus affect
international markets for some of these items.
To what extent has China’s new presence
altered conditions in individual product mar
kets and what is likely to happen in the future?
• China has borrowed relatively little in world
credit markets even though it has been able to
obtain commercial loans on very favorable
terms. Rather, the People’s Republic has relied
primarily on the concessional funds available
from official lenders. This could have signifi
cant implications for other poor-country bor
rowers as China assumes a large interna
tional presence and position at concessional
lending institutions. Where does China stand in
both world official and private credit markets?
• Where does China stand in general? How does
the experience of the People’s Republic com
pare with those of other countries who have
experienced export surges? How will China’s
trade growth affect its development prospects?
The following sections will offer some answers to these
questions.
China’s trade accomplishment
China’s exports and imports each grew in value by an
average of about 30 percent a year during 1977-81.
This was double the average annual growth of trade
for industrialized economies and 50 percent greater
than the average for nonoil-developing countries
(Chart 1). China, consequently, moved from being the
twentieth largest nonindustrial country trader in 1976 to
being the seventh largest in 1981.
China’s exports and imports did not expand uni
formly during this period, however. Exports grew rel
atively strongly throughout the five-year period. Even

*L e s s developed countries.
Source: International Monetary Fund. Direction of
Trade Statistics Yearbook 1982.

during the world recession year of 1981, they were up
20 percent notwithstanding a V/2 percent decline in
total world trade. Only in 1982 did their growth slow
significantly to about V2 percent. Imports, in contrast,
grew extremely fast only until the end of 1980. Then
in 1981 imports fell slightly below their 1980 level. In
1982 they again declined, this time by about 2 percent.
As a result, China moved from substantial trade deficit
($2.7 billion in 1980) to substantial surplus ($1.8 billion
in 1981).
The evolution of China’s trade reflects the country’s
pragmatic economic policies since 1976. One early
/ aspect of this policy approach was the recognition that
imports are essential to China’s rapid modernization
because they embody important technology. However,
in 1980 China’s leaders made a reassessment of gen
eral economic progress to date. They appear to have
decided that the large trade deficits running from 1978
through 1980 should not be continued in 1981 or 1982

FRBNY Quarterly Review/Winter 1982-83

because the benefits of the large import purchases did
not justify the finance costs involved.
China’s general pragmatic philosophy since 1976
has led to some broad changes in the economy which,
although not specifically meant for this purpose, did
result in spurring foreign trade and, later, improving
th.e trade balance. A major change was emphasizing
decentralization of economic institutions, giving prov
inces and other government agencies more autonomy.
As a consequence, these units have become more at
tuned to market prices and profits and more trade
oriented. Since 1980, increasing emphasis has also
been placed on agriculture and consumer goods pro
duction. There has been, correspondingly, a decreased
emphasis on industrial goods production and major
capital investment projects which have a higher import
content. This switch contributed to a fall in China’s
import bill.
Specific measures dealing with trade have been
taken as well. Special economic zones catering to ex
port markets have been established. Legal structures
related to trade have been put in place. These include
regulations dealing with foreign banking and finance,
joint ventures and the taxes they must pay, oil explora
tion and development, and labor issues (hiring, firing,
compensation) in the special economic zones. Material
incentives, such as tax advantages, are given to firms
manufacturing for export. China has eagerly sought in
novative trade and financing arrangements, including
compensation trade (a foreign firm provides capital and
expertise in exchange for finished products), export
processing, cooperative production, and joint ventures
(Table 2). The People’s Republic has even encouraged
the establishment of wholly owned foreign enterprises
on its soil, provided they manufacture for export. On
the import side, China has dramatically cut purchases
since 1980, using its centralized planning system.
Almost all foreign exchange must be turned over to the
Bank of China. A central government decision is made
on how to disburse this money. The central government
also sets tariffs and quantitative restrictions to manage
the demand for foreign goods. More directly, it pur
chases the large import items, such as plant and equip
ment. Starting in 1981, the government also cut import
growth by postponing or canceling large industrial pro
jects that entailed heavy foreign imports.
China’s trade partners
The most rapid expansion in China’s trade has been
with Japan, the United States, and the developing coun
tries. Most of the developing country trade, however,
is channeled through the entrepot of Hong Kong.
Japan, the United States (particularly after official
relations were normalized in 1979), and Hong Kong

FRBNY Quarterly Review/Winter 1982-83
Digitized for44
FRASER

accounted for most of China’s trade growth since
1977, and they are continuing to grow in importance.
By 1981, China sold over half its exports and purchased
over 60 percent of its imports from these three trading
partners. Export and import values and balances, how
ever, differ among the three (Charts 2 and 3).
Japan is the largest exporter to the People’s Re
public and the second largest market for China’s
goods. Imports from Japan are primarily machinery
and equipment. The People’s Republic sells most of
its exportable fuels (crude oil and coal) to Japan in
return. Sino-Japanese trade was near balance in 1981
but turned more in China’s favor in 1982 as its ex
ports continued to expand while its imports shrank.
The United States is the second largest exporter to
China and the third largest buyer of its products. SinoAmerican trade has doubled since full-fledged diplo
matic relations were established in 1979. The com-

position of Sino-American trade differs from that of
Sino-Japanese trade. The United States supplies mainly
agricultural products, rather than machinery. Wheat,
corn, and other grains made up about 40 percent of the
$4 billion worth of goods China bought from the United
States in 1981. On the other side, textiles and other light
manufactures dominate China’s exports to the United
States. (Japan and the United States import about the
same total amount of Chinese textiles, but the United
States does not make the large fuel purchases that
Japan does.) In 1981 the United States had a substan
tial trade surplus— about $2 billion— with the People’s
Republic. This surplus fell to about a half billion dollars
in 1982.
The third largest supplier of goods to China is Hong
Kong. Chinese imports from there include both prod
ucts made locally— about one quarter of the total—
and goods reexported to the People’s Republic via
Hong Kong from producers in other countries. In re
turn for these exports, Hong Kong provides the largest
market for China’s goods. But 40 percent of sales to
Hong Kong is reexported from there to other countries,
particularly to developing economies. The larger part
of this trade is in manufactured goods. In 1981, China
had a large trade surplus of over $3 billion with Hong
Kong. The surplus in 1982 will probably be even greater.
The remaining 40 percent of China’s exports are split
about evenly between other industrialized and other
developing economies. Textile sales predominate with
the first group and simple manufactured goods sales
with the second. On the import side, China buys about
one quarter of its foreign goods, primarily machinery
and fertilizer, from industrial countries other than Ja
pan and the United States. About 15 percent of Chinese
imports, mainly food and raw materials, come from de
veloping countries other than through trade with Hong
Kong. In 1981 the People’s Republic had a $11/2 billion
deficit with these other industrialized economies and
a $2 billion surplus with these other developing
economies.
China’s trade pattern— surplus with developing coun
tries, moderate deficit with industrialized countries—
leaves the People’s Republic in a strong medium-term
position. It is in surplus in the developing country
market, which has had the fastest growth rate in world
trade and will likely continue to grow fastest in the
next decade despite the prospect for some near-term
weakness. Other economies with income levels as low
as China’s, in contrast, typically sell very little to one
another. On the other side, China benefits from the
fact that it has oil and coal it can sell to industrialized
countries, enabling it to hold down its deficit with
those economies (Chart 3). More development can be
expected in these fuel exports.

The evolving structure of China’s trade
As China’s total trade expanded and the People’s
Republic consciously moved from trade deficit in 1980
to trade surplus in 1981, the structure of both its ex
ports and imports has changed (Chart 4). The most
striking development on the export side is the rise in
oil and coal sales. These energy exports now account
for a quarter of total Chinese exports, up from about
10 percent in 1977. Currently, growth of both oil and
coal production has leveled off. The volume of these
exports is not expected to change much until offshore
011 fields start producing at the end of the decade.
This should slow the rapid growth of Japanese im
ports from the People’s Republic, since Japan is the
major purchaser of these fuels.
Chinese manufacturing exports have also done well
in the past five years. They now account for over half
of total exports, up from 45 percent in 1977. China is
currently emphasizing light manufactures production,
which requires less investment funds per factory and
quickly generates foreign exchange earnings. Textiles
and other light manufactures (such as rubber shoes,
carpets, porcelain, paper) account for about two thirds
of manufactured exports. But these products are ex
pected to decline in importance, as exports of simple
machinery and electrical equipment (e.g., electrical
components, fans, sewing machines) increase in com
ing years. Sales to developing countries, the primary
buyers of these latter products, should increase ac
cordingly.
The most striking change on the import side is the
sharp drop in machinery and construction goods pur
chases since 1980. This is primarily due to the can
cellation or postponement of a number of large in
dustrial projects. Imports of capital and construction
materials fell 20 percent in 1981. In the first half of
1982, they fell even more sharply. Concomitant with
this decline was an increased share in total imports
of consumer products and inputs for agriculture and
light industry. These goods now account for 60 per
cent of total foreign purchases, up from 42 percent in
1979.
China is likely, however, to reverse this import com
position pattern somewhat in the next few years. The
People’s Republic has acquired a high level of inter
national reserves (about $10 billion) and may start
increasing imports rather than continue running large
trade surpluses. Materials for the canceled or post
poned industrial projects will probably be given first
priority when purchases pick up. Negotiations have
already resumed for some of these items.
The considerable manipulability of China’s trade
structure contrasts with many other developing coun
tries. Large natural resource endowments and com-

FRBNY Quarterly Review/Winter 1982-83

Chart 3

China’s Trade Pattern in 1981

China’s exports to the United States

China’s im ports from the United States

Machinery, equipment,
^ a n d intermediate
goods

^Fuel^x

Machinery, equipment,
\ , a n d intermediate
goods

;15.5%>
Food and: fertilizer*
5^44.3%o<

Textiles
clothing)
Other

v 22.1%

Textile
yarns
29.8%

18.5%

^ Textile yarns,
fibers, and fabrics

C hina’s exports to Japan

China’s im ports from Japan

Textiles
/ 4.5%

Machinery, equipment,
and intermediate
goods
4-9%

Other
0.5%'

Other
4.8%

,FT ' " izer

Textile yarns
Textile
yarns
^

11. 8 %

Other
19.2%

Machinery, equipment,
and intermediate
goods
v
79.3%
y

China’s exports to Hong Kong

China’s im ports from Hong Kong

Breakdown not available
Machinery, equipment,
. and intermediate
^
goods

'extiles
12 . 6 %

Sources:

Other

Textile
yarns

18.2%

21. 0 %

United States Department of Commerce; Business America; and China Trade Report.

46 FRBNY Quarterly Review/Winter 1982-83

petitive prices (discussed in the concluding section)
have helped the People’s Republic expand and diver
sify its export market. Significant domestic production
of most essential goods— food, oil, raw materials— has
kept China’s essential import needs relatively low.
There was considerable room, consequently, to com
press the total import bill in 1981, when China decided
it no longer wanted to run trade deficits.
China’s relative size and impact on major
goods markets
Increased trade growth has meant increased market
size for China in a number of areas. For some traded
goods, there is potential for China to affect world prices
or to cut significantly into other producers’ market
shares or to become a major world consumer. As
total trade continues to grow, this may become an
even more important factor in the future. Trade now
equals about 15 percent of China’s approximately
$250 billion gross domestic product (GDP). China’s ex
ports and imports are, therefore, substantial. However,
China’s GDP per capita is a low $250. As the People’s
Republic grows, exports and imports will likely grow as
well.
Textiles are the most prominent case of China’s
increased market size. The People’s Republic has al
ready faced quota restrictions on some particular tex
tile exports to both the European Community and the
United States. As noted in the introduction, China is
currently involved in a trade dispute with the United
States over the issue. In general, because the United
States and Europe follow policies of providing a cer
tain level of overall protection to their domestic textile
producers, increased Chinese textile exports have
had some crowding-out effects for other sellers.
Although China’s petroleum sales are a significant
part of its total exports, they were well below 1 per
cent of world oil trade in 1981 and are unlikely to
change much during the decade. China, consequently,
has too small sales to have major impact on the world
oil market.
But China could potentially affect world market
conditions for a number of strategic minerals. The
People’s Republic has major deposits of tungsten, tin,
antimony, zinc, vanadium, titanium, lepidolite, gold,
copper, lead, molybdenum, bauxite, mercury, nickel,
and iron. Deposits of the first seven metals are
thought to be the largest in the world. China’s iron
deposits rank third in the world, and its gold output
ranks fourth. The People’s Republic helped depress the
world price of vanadium in 1981 with its expanded
sales.
China also has had a significant impact on world ag
ricultural trade. In 1981-82 China was one of the three

Chart 4

C h in a ’s T rad e C om position C hanges
Selected commodities traded with m ajor trade p a rtn e rs*
Billions of U.S. dollars
4
Exports
Textiles
"

- ------------

I----------- —

— — "
t

L ............

' Iron, steel, and
engineering products

Fuel

--------

Food

....... I

Imports
Iron, steel, and
engineering products

Other
manufactures

1978

1979

1980

1981

* Canada, the European Community, the European Free Trade
Association members, Hong Kong, Japan, Singapore, and
the United States.
"^Includes both heavy and light industry goods.
Source: General Agreement on Tariffs and Trade, International
Trade, 1980-81 and 1981-82.

largest wheat importers in the world, accounting for
13 percent of total world wheat imports. China ac
counted for 13 percent of world raw cotton imports and
about 10 percent of world rubber imports at that time.
It has also been the largest importer of chemical fertil
izers in recent years. On the other side, Chinese rice
exports amounted to 5 percent of the world’s total rice
trade in 1981-82.
China is expanding exports rapidly in certain chem
icals, certain pharmaceuticals, and shipbuilding. Some
trade analysts in these fields argue that the Peo
ple’s Republic may have a significant effect on
world trade for these goods. For imports, China is
unlikely, however, to have significant market influence
through its purchases of goods other than agricultural
products in the near future. Its import demands are
for fairly diverse items. For none of them is China likely
to account for a significant proportion of world trade.

FRBNY Quarterly Review/Winter 1982-83

China’s international borrowing position
China has not yet been a major borrower in inter
national credit markets, although in the past three
years it has attracted credit commitments estimated
as high as $30 billion. The People’s Republic has
drawn only a small fraction of these funds. For 1981 it
reported lining up about $9 billion in official and pri
vate long-term foreign credits but drew only $2 billion
of this. Japanese sources provided approximately one
third of the 1981 capital inflow, the IMF one quarter,
Hong Kong sources about one fifth, and American
sources perhaps one tenth.
Chinese external debt outstanding is very small by
international comparison (Table 3). Total debt is esti
mated at $8 billion. Long-term debt of about $61/2 bil
lion equaled only 2 percent of China’s 1982 gross
domestic product (GDP) and 25 percent of its goods
and services exports. This contrasts sharply with most
other developing countries. For these countries as a
group, debt averaged 25 percent of GDP and over 100
percent of exports. For the lowest income countries
(comparable to China), the figures are 25 percent and
over 200 percent, respectively. In terms of net debt,
China’s holdings for foreign exchange exceeded its
foreign obligations by about $2 billion in 1982, making
the People’s Republic a net creditor to the world. In
1981, China had long-term net debt of about $ 11/2 bil
lion, equaling Vz percent of its GDP, whereas this ratio
for all developing countries, as well as for the lowest
income countries, was 20 percent. Banks in industrial
countries reported only $1.3 billion in claims on China
in mid-1982, accounting for less than 1 percent of
all the external claims on developing countries held
by these banks. China’s 1982 long-term debt service
payments amounted to about $2 billion. This equaled
only 7 percent of Chinese exports. In contrast, the
overall developing country ratio of long-term debt
service payments to export receipts was over 20 per
cent. For lowest income countries the figure was nearly
15 percent.
China’s low overall borrowing level reflects its
strongly expressed sensitivity to high interest rates as
well as to high debt burdens. China has been able to
obtain about $2 billion, around one third of its long
term borrowing, on concessional terms, through inter
national agency loans, intergovernment loans, and
suppliers credits arranged with foreign export-credit
agencies. Japan has been the primary source of such
low-cost funds.
Although a substantial amount of China’s long-term
borrowing is already on concessional terms, the Peo
ple’s Republic could borrow a great deal more from
official sources. This it may do in the next few years.
Trade partners have been willing to offer conces-

FRBNY Quarterly Review/Winter 1982-83

Table 3

International Bank Claims on China and
Selected Developing Countries*
In billions of U.S. dollars, December 1981

Country

Liabilities
to banks

Net position
with banksf

China .............................

2.3

3.3

India ................................
Indonesia .......................

1.4
7.2

1.5
0.3

M a la ysia .........................
Brazil .............................
Korea .............................
t Hong K o n g .....................
$ S in g a p o re .......................
Mexico ...........................

4.4
52.7
19.9
31.3
36.6
56.9

— 1.2
- 4 7 .5
— 16.3
— 2.7
0.1
— 44.6

* Banks in industrialized countries.
f Country assets held by the banks minus country liabilities
owed to the banks.
| Offshore banking centers.
Source: Bank for International Settlements.

sional export financing and other subsidized credits,
partly to sell their goods. China could borrow more.
From its largest bilateral aid source— Japan— the Peo
ple’s Republic has drawn less than 5 percent of the con
cessional loan money offered. The low actual drawings
in part reflect postponements and cancellations of
large plant and equipment imports from Japan. This
may change if China decides to cut its current large
trade balance surplus.
International institutions are only beginning to in
clude China in their fund disbursements following its
formal admission to these organizations. From the
World Bank’s soft loan window— the International
Development Association (IDA)— China has borrowed
$459 million to date. (This contrasts with India, the
country most comparable to China, which in recent
years has been drawing about $1.5 billion annually
from the IDA.) The People’s Republic was admitted by
the institution only in 1980, after most of the currently
available funds had been committed. China will likely
obtain increased concessional money from the World
Bank after settlement of the IDA replenishment now
being negotiated. In the future, large Chinese borrow
ing from concessional lenders could have a significant
impact on the availability and international distribution
of concessional loan funds to other borrowers.

Long-term commercial bank borrowing accounts for
only about one third of China’s total debt. To avoid
high interest charges, China has repayed in advance
large amounts of commercial bank debt, especially in
1981. The People’s Republic was able to do this be
cause of its strong current account position. In turn,
China’s low borrowing level has buttressed its current
account position as interest payment outflows have
been kept low.
China’s low level of commercial borrowing helps
make it very attractive to banks. By the end of 1981,
banks in industrial countries reported $5 billion in bind
ing commitments to China, a figure greater than for
any other developing country and more than twice the
value of current outstanding claims on the People’s
Republic. China received interest rate spreads as low
as Va to 1/2 percent over LIBOR (the London interbank
offer rate) on 1981 loans. This compares with an aver
age spread of 1 percent for developing countries as a
group that year. China continues to receive favorable
terms even as spreads for most developing countries
are widening.
China’s trade growth and its development strategy
from an international perspective
Several economies have achieved high income growth
rates from export-led growth strategies, notably econ
omies in Asia. Export-led growth may be defined
as a conscious policy effort to promote exports and
to attain rapid GDP growth. Fast GDP growth should
result from the benefits gained by trade. Increased
employment for the production of export goods is
one benefit. Another is the rearrangement of produc

tion to concentrate on goods that can be produced
efficiently— i.e., those goods a country has a compara
tive advantage in producing— while trading to obtain
other desired commodities. Trade also allows a coun
try to manufacture goods in sufficiently large volume
to benefit from the cost savings of large-scale pro
duction. Finally, an economy gains vitality from pro
duction subject to international competition.
Factors within specific economies determine the de
gree to which those economies gain from each of
these benefits. For example, larger economies may
have less to gain from the potential economies of
scale benefit. Due to different country situations, then,
the boost to GDP growth obtainable from fast export
growth varies greatly.
China may try to pursue a pattern of rapid export
growth followed by rapid GDP growth, as have other
countries. A clue as to whether it will follow this pat
tern may be found in comparing China’s recent trade
expansion with similar spurts in six other developing
economies that experienced comparable periods of
growth and in analyzing the degree to which China
will be able to benefit from this expansion (Table 4).
China’s export performance is impressive, even by
comparison with the export spurts in other countries,
particularly since it occurred during a time of slow world
trade. The international environment was much more
favorable during the trade spurts of the other econo
mies. China may or may not continue with rapid export
growth. If export growth continues, the question arises
whether China can successfully convert to sustained
rapid income growth. The question does not have a
clear-cut answer. However, certain characteristics of

Table 4

Trade and Domestic Economic Growth of China,
Compared with Other Developing Countries with Export Surges

Country

Years

Average
export volume
growth rate
(percent)

Average
export value
growth rate
(U.S. $)

Average
real GDP
growth rate
(percent)

C h in a ................................................................................. ...........1978-80

22.3

34.2

7.5

B ra z il................................................................................. .......... 1968-70
India ................................................................................. ...........1974-76
Is r a e l................................................................................. ...........1959-61
J a p a n ............................................................................................1954-56
K o re a ................................................................................. ...........1964-66
Malaysia ........................................................................... .......... 1976-78

10.6
10.5
19.4
25.2
36.1
25.1

18.4
23.4
26.9
29.8
42.2
12.7

10.3
6 .6

10.1
7.3
9.2
8.7

Source: International Monetary Fund, International Financial Statistics.

FRBNY Quarterly Review/Winter 1982-83

the Chinese economy may affect the degree to which it
will benefit from trade.
China’s size distinguishes it from other export-led
growth economies. Large size will probably dampen
the benefits China can expect to achieve from an
export-led growth strategy. Three reasons may be cited.
The first is that China may not be able to sell interna
tionally the goods that it has a comparative ad
vantage in producing without adversely affecting the
market for these goods. If China produced for export
the same amount of textile fabrics and garments, for
example, per member of its labor force as did the
other Asian economies when they successfully pursued
export-led growth, it could easily swamp world mar
kets. The resulting sharp price declines and/or in
creased trade restrictions would significantly cut into
China’s trade benefits. A second consideration is that
China has the potential to reap the benefit of large-scale
production with manufacturing for domestic use. It,
therefore, should gain less of this benefit from interna
tional trade than have smaller economies with smaller
domestic markets. A final factor is that China could po
tentially have gained significant comparative advantage
benefits through its own internal trade. This would
contrast with the situation of smaller countries before
they opened their international trade doors. China,
therefore, may have less to gain from this benefit often
provided by expanded international trade.
China’s type of economic regime also distinguishes
it from other export-led growth economies which are
typically capitalist. The distinction will be less pro
nounced the more flexible China makes its price sys
tem and the more adeptly it handles its managerial
and incentive issues. However, the more there is such
a distinction, the more likely it is that the People’s
Republic will find less benefit in the export-led growth
strategy. This may be seen from the problems experi
enced by other centrally planned economies in East
ern Europe and the Soviet Union. Four problems have
typically arisen:
• With domestic prices determined apart from
market forces, trade patterns may depart from
what comparative advantage analysis would
recommend. This would cut potential interna
tional trade gains.
• With nonmarket determined prices, investment
decisions may cause subsequent trade patterns
to be even further away from the comparative
advantage criterion, again cutting potential
trade gains.
• Problems of quality control often have been

FRBNY Quarterly Review/Winter 1982-83
Digitized for50
FRASER

encountered. These are generally attributed to
strict adherence to production plan levels
rather than product marketability. These prob
lems dampen world demand for the products
and, hence, trade growth.
• Problems of significant consumer goods ra
tioning also have occurred because prices are
relied on to a very limited extent as a means to
allocate goods. This may cut the incentive to
produce in order to buy since increased income
need not yield increased consumption. This
lack of incentive may be an obstacle to greater
trade growth.
These constraints may explain why the unweighted
average ratio of hard currency exports to GDP is
relatively low, at about 10 percent, for the Eastern
European and Soviet economies. The ratio for the
Soviet Union is about 2 percent. China’s ratio is esti
mated to be 8 percent.

China’s trade strategy
What does China’s recent trade performance suggest
about its trade plan relative to other countries? China
may be following a straightforward pattern of export
ing the goods it can produce cheaply in exchange for
the goods that would be expensive or prohibitive to
make. On the other hand, there may be reasons why
China’s trade may reflect more than this simple type
of comparative advantage analysis.
China has been selling its exports, generally laborintensive products, at very competitive prices relative
to other countries’ exports of the same goods. In
return, the People’s Republic has been purchasing
technological products (plant and equipment) more
advanced than those which it produces domestically
(as well as food staples). China’s pricing strategy is
clearly indicative of a desire to gain market share.
However, it also indicates that the People’s Republic
is willing to accept a less favorable relative price ratio
between its Chinese labor (embodied in Chinese ex
ports) and imported technology than most other de
veloping countries accept. China’s lower price ratio
may simply reflect standard analysis of China’s com
parative advantage in trade. The People’s Republic
may have a cost advantage resulting from its abundant,
fairly well-educated labor force. It may also reflect a
well-developed organization of production units and an
established goods distribution network. Perhaps due to
these factors China can produce more goods for the
same amount of effort than can other countries and,
thus, is able to charge less for each item produced.
However, the very competitive price ratio the Peo-

pie’s Republic is setting between its own labor and
imported technology may also be attributable to cer
tain aspects of technology purchases and centrally
planned economic regimes. One aspect is that bene
fits of technology generally become widely diffused
throughout an economy, rather than captured entirely by
the initial purchaser. For example, imported technology
can be copied by others once a prototype is supplied.
Licensing and other controls can reduce this problem,
not completely eliminate it. Only an authority respon
sible for the total economy will, consequently, gain all
the benefits of purchased technology. This, of course,
is exactly what a centrally planned economic regime
is. Such a regime or authority will be most likely to
pay a price that reflects the total value of the benefits
received from the imported technology. This contrasts
with firms in a private market economy who are likely
to pay only in accordance with the benefits they
privately obtain from the technology purchase. A cen
trally planned economy may be willing to supply more
labor effort than a free market per. unit of foreign
exchange earned to increase its foreign exchange
revenue and foreign technology purchases.2
There is another reason why a centrally planned
economy might act this way. Innovation may be
slower for this type of regime than for a free market
economy. This may be because individual inventors
are unable to capitalize privately on their innovations
in a noncapitalist economy. Imported technology may

*This is another way of saying a centrally planned economy may be
willing to sell its labor-intensive exports at a cheaper price than a
capitalist economy would. Implicit in this is the assumption that the
price elasticity of excess demand in the rest of the world for these
goods is greater than one. This assumption is generally valid for laborfntens/ve manufactures.

as a result be more important there than it is for a
free market economy.
To what extent each argument— standard cost advan
tages or placing a distinctly high premium on tech
nology imports— explains China’s competitive prices
and export success cannot be determined. China, of
course, is not alone among developing countries in
relying on other than free market mechanisms. How
ever, most other developing economies do not take
such direct responsibility for total investment or such
a direct role in pricing exports (through central setting
of input prices and the exchange rate). They are,
thus, less likely to place such a premium on imported
technology relative to the product of their labor.

What lies ahead?
China’s future pricing strategy as well as its future
trade performance could change significantly. The
current strategy and trade surge occurred after a
dramatic change in political outlook by the People’s
Republic. Such sudden wide changes could occur
again and alter China’s present trade orientation. Given
current policies, however, it is likely that China’s trade
will continue to grow although at a declining rate.
Exports will rise because of the recent growth of
foreign investment (which is primarily aimed at ex
ports) and the large number of compensation trade
and cooperative production projects that are yet to be
paid off. But, as the absolute size of China’s export base
increases and market size limitations become more
pressing, the growth rate should slow. Imports by the
People’s Republic will rise with exports because
China’s demand for foreign technology and basic food
stuffs is likely to continue. However, conservative policy
should limit growth of imports to China in line with that
of exports to avoid large current account deficits.

Susan A. Hickok and Rosanna Arguelles

FRBNY Quarterly Review/Winter 1982-83

August-October 1982 Interim Report
(This report was released to the Congress
and to the press on December 9, 1982.)

Treasury and Federal Reserve
Foreign Exchange Operations
By the end of the August-October period under review
the dollar had risen to record highs or to levels not
seen in many years against several major currencies,
strengthening even as U.S. interest rates dropped
sharply and as interest differentials favoring dollardenominated assets narrowed appreciably. Favorable
prospects for the U.S. economy relative to other in
dustrial countries, apprehension about the interna
tional banking system, and concern about economic
and political conditions abroad resulted in an increased
global preference for dollar-denominated assets which
pushed dollar exchange rates sharply higher.
Concern over international credit exposures and
developing financial strains in various markets around
the world were sustaining factors behind the dollar’s
rise throughout the period. During August, market at
tention focused on Germany where a large multinational
company was being forced into receivership and on
Mexico where a foreign exchange crisis was unfolding.
During September, concern over the international fi
nancial situation mounted as developments in Mexico,
particularly in light of the unexpected move to national
ize domestic banks, raised doubts in the market about
the ability and willingness of the government and other
public-sector institutions in that country to meet their
external obligations. Meanwhile, the list of countries
experiencing payments arrears expanded, and there
were well-publicized problems of various commercial

A report by Sam Y. Cross. Mr. Cross is Executive Vice President in
charge of the Foreign Group of the Federal Reserve Bank of New
York and Manager of Foreign Operations for the System Open
Market Account.

FRBNY Quarterly Review/Winter 1982-83

banks here and abroad. In this environment, traders did
worry about the relatively large exposures of U.S. banks
to Mexico and other Latin American countries, and de
veloping pressures on the U.S. banking system were
reflected, to an extent, in a widening of yield spreads,
between U.S. Government obligations and private credit
instruments. But, with so much of the total international
credit exposures made up of dollar-denominated
claims, dollar-based institutions were thought to be
in a better position than others to deal with emerging
liquidity strains. Moreover, individual institutions
sought to augment their liquidity positions, especially
in dollars, against potential funding and cash-flow
problems and in advance of important statement dates.
Meanwhile, prospects for economic recovery re
mained gloomy, and concerns intensified that many
of the industrialized countries would tend to rely more
on protectionist measures to deal with high and rising
levels of unemployment and slack business invest
ment at home and would welcome improvements in
international competitiveness in increasingly restricted
export markets. These concerns tended to coalesce in
Europe when several Scandinavian countries devalued
their currencies, at times by more than private and
official observers thought necessary to regain compet
itive equilibrium. Market speculation developed that
several European governments would seek to adjust
their currencies downward, Involving a realignment of
the European Monetary System (EMS) joint float.
Within that arrangement speculative selling pressures
— largely against the French and Belgian francs, the
Italian lira, and the Danish krone— intensified around
mid-October. But they tended to moderate late in the

Table 1

Drawings and Repayments by
Foreign Central Banks and the Bank for
International Settlements under
Regular Reciprocal Currency Arrangements
In millions of dollars; drawings ( + ) or repayments ( — )

Bank drawing on
Federal Reserve System

Out
standing
July 31,
1982
700.0

Bank of M e x ic o ........

August 1
through
October 31,
1982

Out
standing
October 31,
1982

[+ 7 0 0 .0
700.0

700.0

j—

Data are on value-date basis.

Table 2

Drawings and Repayments by the
Bank of Mexico under Special Reciprocal
Currency Arrangements
In millions of dollars; drawings ( + ) or repayments ( - )
Out
standing
July 31,
1982

August 1
through
October 31,
1982

Out
standing
October 31,
1982

United States Treasury
special temporary
facility for
$ 1 ,0 0 0 million ..........

-0 -

f +825.0
J-8 2 5 .0

-0 -

Drawings on
special combined
credit facility:
Federal Reserve
special facility for
$325 m illio n ...............

-0 -

f + 236.3
} - 43.8

192.5

United States Treasury
special facility for
$600 m illio n ...............

-0 -

f +438.8
} — 81.3

357.5

Drawings on

Data are on value-date basis.

Table 3

Drawings and Repayments by the Bank of Brazil
under Special Reciprocal Currency Arrangement
with the United States Treasury
In millions of dollars; drawings ( + ) or repayments ( — )

Drawing on

Outstanding
July 31,
1982

August 1
through
October 31,
1982

Out
standing
October 31,
1982

-0 -

+350.0

350.0

United States Treasury
special facility for
$500 m illio n ...............
Data are on value-date basis.

period after official actions were taken by several
countries to raise domestic interest rates, to adopt do
mestic austerity measures, and/or to increase inter
national borrowings. The monetary authorities of the
EMS member states intervened heavily as sellers of
dollars and, to a lesser extent, of currencies trading
at the top of the joint float arrangement. Nonetheless,
the EMS currencies as a group declined substantially
against the dollar.
In addition, there were other international develop
ments which reinforced the demand for dollars. These
included uncertainties over the future political sov
ereignty of Hong Kong, which reportedly generated
flows of capital to North America, and aggravated hos
tilities in the Middle East, which kept alive fears of a dis
ruption of the flow of internationally traded oil. Certain
currencies that had previously offered clear alternatives
to investment in dollar-denominated assets also came
under sometimes unfavorable exchange market scru
tiny, as participants focused on unresolved political
divisions over economic, social, and foreign policies
in a number of countries. In Germany, Chancellor
Schmidt’s coalition government collapsed over disputes
about economic policy. At first, the prospect of a new
government generated expectations that the policy
stalemate would be broken. But soon the market con
cluded that the new coalition government might face
serious difficulties in winning a majority at upcoming
federal elections next spring and that, in the interim,
it had less room to reorient policies than had first
been hoped. Also, in Japan, Prime Minister Suzuki
unexpectedly announced that he would not seek reelection, and uncertainty over his successor clouded
the outlook for the course of Japanese economic
policy.
To some extent, developments in the U.S. current
account also continued to support the dollar, largely
because weaker than expected economic activity
tended to limit the deterioration in U.S. trade perfor
mance associated with the eroding price competitive
ness of U.S. exports. Thus, although many forecast
ers projected a modest current account deficit in the
third quarter of 1982, few participants anticipated a
major shift from equilibrium in the U.S. current account
until the domestic economy moved decidedly out of re
cession. At the same time, Germany’s current account
had slipped from surplus to near balance, and some
analysts, perceiving structural weaknesses in the Ger
man economy, predicted only limited further improve
ment in Germany’s balance of payments in the absence
of a recovery in world demand and output. At the same
time, earlier optimistic forecasts of Japan’s current
account surplus were scaled back further.
For these various reasons, the United States was

FRBNY Quarterly Review/Winter 1982-83

viewed relatively favorably on economic and political
grounds, and market participants bid up the value of the
dollar. On occasion, however, the impact of these con
cerns on the dollar was offset, as market participants
focused on actual and expected declines in U.S. inter
est rates. In late August, for example, a shift in the out
look for U.S. interest rates occurred. Whereas at
midyear Federal Reserve authorities had indicated that
they would tolerate monetary expansion at somewhat
higher than the targeted annual rate in view of excep
tional economic uncertainty and strong liquidity de
mands, market participants were skeptical that declines
in interest rates would be sustainable so long as they
expected an early recovery in economic activity. By late
summer, however, evidence suggested a deepening of
the U.S. recession, a weakening in short-term busi
ness credit demands, and a slowing in money supply
growth that brought the narrow monetary aggregate—
M-1— within the 21/2-51/2 percent annual growth range.
By end-August, therefore, short-term U.S. market rates
had dropped some 5 percentage points from end-June
peak levels, the Federal Reserve had reduced its dis
count rate in four steps from 12 to 10 percent, and
market participants gained confidence that these de
clines would stick. In addition, with inflation abating
and with the Congress passing a tax increase, bond
yields dropped as much as 2 percentage points amid
an unusually strong debt-market rally, accompanied
by record price increases in the stock market. Abroad,
interest rates did not recede by nearly as much, al
though production and output declines continued and
unemployment advanced further with a deepening of
the recession in major foreign economies. As a result,
interest differentials favorable to the dollar narrowed
dramatically, for instance, on three-month Eurodeposits from IV 2 to 31/4 percentage points vis-a-vis
the German mark and from 91/2 to 4 percentage points
against the Japanese yen, and the dollar moved lower
in the exchange markets.
Early in October the dollar’s strengthening trend
was again temporarily interrupted. Following the Fed
eral Open Market Committee meeting early that month,
it was announced that less emphasis would be placed
in the immediate future on M-1 as an operating target
of monetary policy and that somewhat more rapid
growth of the broader aggregates would also be tol
erated in an environment of extreme economic and
financial uncertainty. As explained by Chairman
Volcker, financial innovation and institutional change
— such as the large volume of all savers certificates
about to mature and the new money market deposit
accounts to be introduced late in 1982— coupled with
the still appreciable strengthening in the desire for
liquidity served to distort M-1 as a reliable policy

FRBNY Quarterly Review/Winter 1982-83

Table 4

United States Treasury Securities
Foreign Currency Denominated
In millions of dollars equivalent;
issues ( + ) or redemptions ( — )
Amount of
commitments
July 31,
1982

August 1
through
October 31,
1982

Amount of
commit
ments
October 31,
1982

G erm any...........................
S w itzerland.......................

2,610.6
458.5

— 671.2
-0-

1,939.4
458.5

Total ..................................

3,069.1

— 671.2

2,397.9

Issues

Public series

Data are on a value-date basis.

Table 5

Net Profits ( + ) or Losses ( — ) on
United States Treasury and Federal Reserve
Current Foreign Exchange Operations
In millions of dollars

Period

Federal
Reserve

United States Treasury
Exchange
Stabilization
General
Fund account

August 1 through
October 31, 1982 ___

-0-

-0 .6

+ 30.6

Valuation profits and
losses on outstanding
assets and liabilities as
of October 31, 1982 ..

-7 7 7 .9

-1 ,4 7 2 .9

+619.3

Data are on a value-date basis.

guide. Also, the rigid pursuit of targets in view of
these developments would have had the practical
effect of a more restrictive policy than intended when
the targets were initially set out. Shortly following
these statements deemphasizing the role of M-1, the
Federal Reserve cut the discount rate another 1/2 per
centage point to 91/2 percent. In the market, these ac
tions were widely interpreted as a shift toward greater
monetary accommodation by the U.S. authorities and
generated expectations that declines in U.S. money
market and official interest rates, which had stalled
during September, would again resume. Once again
the dollar came on offer in the exchange market.
But, as in August, the dollar’s decline proved tem
porary and market psychology toward the dollar re

mained positive. Few market participants regarded the
shift in operating procedure as an abandonment of
the fight against inflation. Moreover, substantial prog
ress had already been achieved in moving toward
greater price stability in this country, with wage, salary,
and price increases slowing markedly and unit-labor
costs even more dramatically. In response, interest rates
in longer term markets dropped another 1 percentage
point in October alone. Yet, compared with other coun
tries, the decline in U.S. nominal interest rates still
lagged behind the reduction of inflationary pressures,
so that real U.S. interest rates remained high, both ab
solutely and relative to other countries. Furthermore,
foreign monetary authorities were expected to take
fuller advantage of what by this time appeared to be
sustainable declines in U.S. interest rates to ease
credit conditions in their economies. These expecta
tions were confirmed when official and market interest
rates in major European countries declined consider
ably in the last weeks of October. Under these cir
cumstances, financial markets were impressed with
anecdotal evidence suggesting that foreign investors
sought to benefit from the continuing potential for
price appreciation in U.S. domestic capital markets by
investing in longer term dollar-denominated securities.
While foreign purchases of these securities were
apparently financed largely out of existing dollardenominated assets, talk of foreign investment activity
nonetheless had a positive psychological effect on the
dollar and may have been associated with renewed
bidding for dollars in the exchange market.
By end-October the dollar reached record highs
against several of the Continental currencies, levels
not seen in nearly six years against the pound ster
ling and the Japanese yen, and a 141/2-month high
against the German mark. On balance, for the threemonth period under review the dollar rose 81/4 percent
against the Japanese yen, 6 percent against the Swiss
franc, 5 percent against the German mark, and 4 per
cent against the pound sterling. With respect to the
Canadian dollar, however, the dollar declined about
2 percent. On a trade-weighted basis the dollar rose
4% percent.
The U.S. authorities intervened on four trading days
during the period when the dollar was bid up sharply
to higher levels in unsettled markets. The Federal Re
serve and U.S. Treasury intervened early in August
and again early in October to purchase $45.0 million
equivalent of German marks and $57.0 million equiv
alent of Japanese yen. The German mark purchases
were split evenly between the Federal Reserve and
the Treasury. Of the total Japanese yen acquired,
$38.5 million equivalent was for the Federal Reserve
and $18.5 million equivalent was for the U.S. Treasury.

In the August-October period, various short-term
financing arrangements were concluded in support of
Mexico’s efforts to strengthen its economic and finan
cial position. At the beginning of the period, the Bank of
Mexico had outstanding a one-day $700 million drawing
on its swap line under the Federal Reserve’s reciprocal
currency arrangements used to finance a short-run
liquidity need which was repaid on August 1. Then,
with the Mexican authorities proceeding with the im
plementation of a previously announced stabilization
program, the Bank of Mexico again drew $700 million
under its reciprocal swap line with the Federal Re
serve on August 4, this time for a period of three
months. The Mexican authorities also arranged a
temporary new $1 billion swap facility with the U.S.
Treasury over the August 14-15 weekend, drew $825
million, and then on August 24 repaid the entire
drawing using an advance payment for oil from the
U.S. Department of Energy. Meanwhile, negotiations
among Mexico, the U.S. Treasury, Federal Reserve,
and major foreign central banks resulted in a multi
lateral package to provide bridge financing to an
International Monetary Fund (IMF) standby credit. The
credit facility totaling $1.85 billion comprised $325
million with the Federal Reserve, $600 million with the
U.S. Treasury, and $925 million with the Bank for Inter
national Settlements. During the period under review
the Bank of Mexico drew for three months $105 million
and $195 million on the Federal Reserve and U.S. Trea
sury swaps, respectively, as part of the first $600 million
it took down on the combined facility. The Mexican
authorities also made one overnight drawing of $250
million on the combined facility which was repaid. The
drawing comprised $43.8 million on the Federal Re
serve, $81.2 million on the U.S. Treasury, and $125 mil
lion on the Bank for International Settlements. Subse
quently, the Bank of Mexico also drew for three months
$87.5 million on the Federal Reserve and $162.5 mil
lion on the U.S. Treasury, leaving $1 billion still available
on the entire combined credit facility as of October 31.
In other developments the U.S. Treasury provided
$1.23 billion of short-term financing to Brazil by ar
rangements which were under discussion since Octo
ber. This additional short-term liquidity was made
available in conjunction with economic policies
adopted by Brazil at the October meeting of its Na
tional Monetary Council. The financing was provided
under three swap facilities. One drawing on the first
$500 million facility was made on October 28 for
$350 million. Other facilities made available in Novem
ber, when combined with the above-mentioned $500
million, totaled $1.23 billion and were announced by
President Reagan during his visit to Brazil in the first
week of December. The swap arrangements represent

FRBNY Quarterly Review/Winter 1982-83

bridging loans to Brazil’s drawings under the Compen
satory Financing Facility of the IMF as well as on its
reserve position with the IMF.
On September 1 the U.S. Treasury redeemed fur
ther German mark-denominated securities equivalent
to $671.2 million. After this redemption, the Treasury
had outstanding $2,397.9 million equivalent of foreign
currency notes, public series, which had been issued
in the German and Swiss markets with the cooperation
of the respective authorities in connection with the
dollar-support program of November 1978. Of the
notes outstanding as of October 31, 1982, a total of
$1,939.4 million equivalent was denominated in Ger
man marks and $458.5 million equivalent was de
nominated in Swiss francs.

FRBNY Quarterly Review/Winter 1982-83

In the three-month period from August through
October, the Federal Reserve had no profits or losses
on its foreign currency transactions. The Exchange
Stabilization Fund (ESF) lost $0.6 million in connec
tion with sales of foreign currency to the Treasury
general account which the Treasury used to finance
interest and principal payments on foreign currencydenominated securities. The Treasury general account
gained $30.6 million on the redemption of German
mark-denominated securities. As of October 31, 1982,
valuation losses on outstanding balances were $777.9
million for the Federal Reserve and $1,472.9 million
for the ESF. The Treasury general account had valua
tion gains of $619.3 million related to outstanding
issues of securities denominated in foreign currencies.

U.S. MONETARY POLICY AND FINANCIAL MARKETS
Paul Meek, Vice President and Monetary Adviser, has
written a comprehensive review of the formulation and
execution of monetary policy. In this 192-page book, he
discusses open market operations with primary empha
sis on the post-October 1979 period. The financial in
stitutions and markets within which the Federal Reserve
operates are also described.
This book is intended primarily for economists, seri
ous economic students, bankers, participants in the
financial markets, and other “ Fed-watchers” .
Single copies are available free of charge. Additional
copies are $4 each for shipments in the United States.
However, reasonable quantities are available free of
charge for classroom use. Such orders will be sent only
to U.S. college or university addresses. For additional
copies mailed to those outside the United States the
charge is $9, and foreign residents must pay in U.S.
dollars with a check or money order drawn on a U.S.
bank or its foreign branch. Write to:
Public Information Department
Federal Reserve Bank of New York
33 Liberty Street
New York, N.Y. 10045

Subscriptions to the Quarterly Review are free. Multiple copies in reasonable
quantities are available to selected organizations for educational purposes. Single
and multiple copies for United States and for other Western Hemisphere sub
scribers are sent via third- and fourth-class mail, respectively. All copies for
Eastern Hemisphere subscribers are airlifted to Amsterdam, from where they are
forwarded via surface mail. Multiple-copy subscriptions are packaged in envelopes
containing no more than ten copies each.
Quarterly Review subscribers also receive the Bank’s Annual Report.

Quarterly Review articles may be reproduced for educational or training purposes
only, providing they are reprinted in full, distributed at no profit, and include credit
to the author, the publication, and the Bank.

Library of Congress Catalog Card Number: 77-646559