Full text of Economic Policy Review (Federal Reserve Bank of New York) : Winter 1983, Vol. 8, 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 Bank of
NewYbrk

Quarterly Review

W in te r 1983-84

Volum e 8 No. 4

State and Local Governments
An Assessment of their Financial
Position and Fiscal Policies

Twelve Improvements in the
Municipal Credit System

Commercial Bank Investment in
Municipal Securities

Neighborhood Changes in New
York City during the 1970s
Are the “ G entry” Returning?

Currency Misalignments
The Case of the Dollar and the Yen

In Brief
Economic Capsules

Treasury and Federal Reserve
Foreign Exchange Operations

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 PETER D.
SKAPERDAS (on an assessment of the
financial position and fiscal policies of
state and local governments, page 1);
AARON S. GURWITZ (on twelve
improvements in the municipal credit
system, page 14); ALLEN J. PROCTOR
and KATHLEENE K. DONAHOO (on
commercial bank investment in
municipal securities, page 26); DANIEL
E. CHALL (on neighborhood changes
in New York City during the 1970s,
page 38); and SHAFIQUL ISLAM (on
currency misalignments—the dollar and
the yen, page 49).
Other staff members who contributed to
In Brief—Economic Capsules are M. A.
AKHTAR, A. STEVEN ENGLANDER,
and CORNELIS A. LOS (was the 198082 inflation slowdown predictable, page
61); M. A. AKHTAR, CORNELIS A.
LOS, and ROBERT B. STODDARD
(surveys of inflation expectations:
forward or backward looking, page 63);
CARL J. PALASH (initial claims: a
reliable indicator of unemployment,
page 66); SANDRA C. KRIEGER,
(NOW accounts and the seasonal
adjustment of M-1, page 67); and
MICHAEL D. ANDREWS (FASB 52:
corporate response and related foreign
exchange market effects, page 69).
An interim report of Treasury and
Federal Reserve foreign exchange
operations for the period August
through October 1983 starts on
page 71.

State and Local Governments
An Assessment of their Financial
Position and Fiscal Policies
State and local governments are at a crossroads in
fiscal planning. The sector as a whole finished 1983 with
a budget surplus of $15 billion, the highest ever. The
surplus could be even larger in 1984. Nevertheless,
most state and local policymakers are taking a cautious
view toward their financial outlook. Their concern has
raised two issues. The first is the degree of austerity
state and local governments should maintain as they
formulate their budgets for fiscal 1985 and lay the
groundwork for 1986. The second, at a broader level,
is the impact the mix of policies they select will have
on the economy over the course of future business
cycles.
The caution being exercised by state and local gov
ernments in ordering their priorities for reducing taxes,
increasing spending, and retaining large surpluses is
understandable. During the past three years, a record
number of them faced annual budget crises, and in 1982
the sector had a deficit of $2 billion. Over the course
of 1983, three factors accounted for the $17 billion
improvement in their financial position. One was the
strength of the recovery which began early in that year.
The other two, starting in 1981 and escalating through
1983, were a series of tax increases and a determined
effort to restrain the growth of spending. Neither of
these were easy steps to take.
The budget decisions facing state and local govern
ments are further complicated by the fact that several
The author would like to express his appreciation to
Marie A. Chandoha for her econometric assistance,

of the circumstances which set the stage for their earlier
financial troubles are prevalent once again. One is the
projected large Federal budget deficits. Another is the
combination of state and local surpluses, their relatively
high level of taxes, and their low levels of real spending.
Under similar situations from 1978 to 1980, the Fed
eral Government slowed the growth of intergovernmental
transfers, and slates and localities enacted the largest
tax cuts in their history. These two courses of action,
especially the second one, significantly contributed to
the severity of their subsequent financial problems. If
they were pursued vigorously again today, it could leave
state and local governments more financially vulnerable
than they appear to be now.
The concern over the potential macroeconomic impact
of state and local fiscal policies is also based on recent
experience. During business cycles prior to 1980, state
and local government policies had an effect similar to
that of a shock absorber, working to cushion recessions
and to dampen recoveries. In doing so, their policies
tended to support the automatic stabilizing effects of
Federal fiscal policies. This was not the case during the
two recessions between 1980 and 1982. As the analysis
in this article shows, the state and local sector had only
a very slight moderating impact on the 1980 recession
and tended to aggravate the recession from mid-1981
through the end of 1982.
The effect state and local governments have had on
past economic cycles follows in part from the manner
in which they plan their budgets. Most of them forecast
revenues over a one- or two-year horizon and then

FRBNY Quarterly Review/Winter 1983-84

specify their spending levels accordingly. But, during a
recession, tax receipts are likely to fall short of their
projected levels. If the downturn does not last too long,
though, the intended level of expenditures can still be
financed by drawing down previously accumulated bal
ances. In a recovery, the growth of tax receipts will
accelerate, spending can increase at the planned rate,
and s u rp lu s e s a re b u ilt up o n ce a g a in . In both
instances, the net effect of state and local fiscal policies
is countercyclical.
Between 1980 and 1982, however, the budgetary
process which enabled state and local governments to
moderate earlier downturns was not effective. In par
ticular, their accumulated balances prior to the 1980 and
the 1981-82 recessions were already at low levels fo l
lowing their own tax cuts and the reduction of Federal
grants. Throughout the last two recessions, then, in
contrast to previous ones, states and localities were
fo rc e d to m a in ta in or to fo rtify th e ir s u rp lu s e s by
reducing the growth of spending, raising taxes, or both.
The consequences on the overall economy of states
and localities having to rebuild their surpluses in a
rece ssio n becam e e s p e c ia lly apparent in the most
recent downturn. During the 1981-82 recession, the
Economic Recovery Tax Act of 1981 (ERTA) is esti
mated to have cut total Federal taxes by $39 billion. At

the same time, however, this article shows that ap
proxim ately 37 percent of that reduction was offset by
discretionary tax increases at the state and local level.
The state and local surplus in the current recovery is
indicative of a policy mix which is, once again, a mod
erating influence. If pressures should build to reduce
this surplus through lower taxes, increased spending, or
reduced Federal aid, then the reinforcing impact the
state and local sector had on the last recession may not
prove to be a fluke. The outcome could be critical from
the standpoint of coordinating fiscal policies across all
levels of government.
The fiscal planning situation in which state and local
governments now find them selves did not come about
overnight. It is the outcome of a series of events which
continually reshaped the economic activities of state and
local governments and their interrelationships with the
U.S. economy and the Federal Government. The pur
pose of this article is to analyze these events and to
shed light on the decisions facing state and local pol
icymakers.

Fiscal profile of state and local governments
At the outset of this analysis, it is useful to review the
concept of financial status for state and local govern
ments. Ideally, one measure of that status would be

C h a rt 1

S tate and Local Revenues and Expenditures as a Percentage of Total Government
Revenues and E xpenditures*
P e rc e n t

1950

A ll fig u re s are on a n a tio n a l in c o m e and p ro d u c t a c co u n t b a s is .
as d e fin e d by th e N a tio n a l B u re a u o f E co n o m ic R esea rch.

U.S. D e p a rtm e n t o f C om m erce, B ureau o f E c o n o m ic A n a ly s is .

FRBNY Q uarterly Review/W inter 1983-84
Digitized for 2FRASER

S h a d e d a re a s re p re s e n t p e rio d s o f re c e s s io n ,

♦ S o c ia l in s u ra n c e re v e n u e s and e x p e n d itu re s a re e x c lu d e d .
S ource:

suitable for quantifying changes in financial conditions
for individual state and local governments, for the sector
as a whole, and for the sector with respect to the rest
of the economy. Unfortunately, no such single measure
exists.
The national income and product accounts (NIPA)
provide a basis by which the state and local sector can
be analyzed as a component of the overall economy.
However, this concept of the sector’s budget surplus or
deficit can give a very different impression of financial
conditions than one obtained by looking at the budgets
of state and local governments. There are several rea
sons for this.
First, the NIPA definition of surplus and deficit
includes capital expenditures such as those for schools
or roads on the spending side of the state and local
sector’s budget. But it does not include the funds bor
rowed to finance those projects on the revenue side.
Most states and localities, though, in contrast to tire
Federal Government, have a capital budget which is
separate from their current operations budget. This
means that, if a sufficiently large amount of capital
expenditures were financed by issuing debt, the NIPA
measure could show a deficit, even though state and
local governments had surpluses in their current oper
ations budgets.
Second, the NIPA budget measure does not include
accumulated balances which state and local govern
ments can carry over from the previous fiscal year into
the current year. Yet, states and localities consider these
balances when assessing their financial outlook and
when planning their budgets. For example, a state which
had ended the previous year with a large accumulated
balance, could run a small operating deficit in the cur
rent year and, by its method of accounting, still end the
year with a surplus.
Finally, it is important to separate, as the NIPA do, the
state and local sector’s social insurance funds (primarily
pensions) from its current operations budget. The sur
plus in the social insurance fund, which amounted to
more than $36 billion in 1983, is an important compo
nent of credit markets in the United States. However,
the surplus is reserved for future pension obligations
and cannot be used to finance the current operations
of state and local governments.
When social insurance programs are excluded, states
and localities in the United States account for about half
of the economic activity for what is commonly referred
to in other countries as the general government—Fed
eral, state, and local governments combined (Chart 1).
Of course, in terms of their fiscal operations, state and
local governments differ from the Federal Government
in a number of areas.

Budget requirements
Unlike the Federal Government, all state and local
governments (except Vermont) are required by law to
enact balanced operating budgets.1
There are three implications of the balanced budget
requirement. First, a deficit is not the only budget out
come that can constitute a potential problem. Projec
tions for a small surplus or a balance at the beginning
of the fiscal year could also be cause for concern. If
unanticipated revenue shortfalls or additional spending
needs were to arise, it could necessitate further legis
lative measures during the year in order to avoid an
end-of-year deficit. For this reason, most governments
prefer to design their budgets so that their projected
balance at the end of the year equals 5 percent or more
of expected outlays.
Second, without deficit spending, state and local
governments cannot be so responsive to deteriorating
economic conditions as can the Federal Government.
Third, the level and composition of Federal grants can
have a decisive effect on state and local taxing and
spending decisions. When assessing the impact of the
state and local sector on economic cycles, then, it is
important to distinguish the effects due to its own dis
cretionary actions from those due to changes in Federal
aid.
Policy objectives
For the most part, state and local governments are
concerned with providing the desired level and distri
bution of services for their constituents without creating
an unfavorable tax climate. The desire of a state or
locality to increase its spending must be weighed
against the risk of losing business and household
income to other jurisdictions, given that it may also have
to raise taxes.
Revenues
State and local revenues can be divided into two cat
egories. The first—own-source receipts—is comprised
of taxes, charges, and fees and currently generates just
over 80 percent of the sector’s revenue. In 1983 this
amounted to nearly $360 billion, or 13.1 percent of
aggregate personal income in the United States. In
contrast, Federal taxes excluding social insurance con
tributions amounted to about 15 percent of personal
income. The remainder of the state and local sector’s
total revenue is provided by the Federal Government in
1lf a deficit should arise during the fiscal year, it can be financed by
drawing down accumulated balances, raising taxes or accelerating
their collection, and reducing or postponing expenditures. To avert
cash-flow problems during the year, states and localities can issue
short-term debt in the form of either tax or revenue anticipation
notes (TANs or RANs).

FRBNY Quarterly Review/Winter 1983-84

the form of grants. Federal aid was just over $86 billion
during 1983.
The primary contributors to state and local own-source
receipts are sales taxes and the property tax. Together
they presently account for about 56 percent of the total.
Next in order of magnitude are taxes on individual
income (17 percent) and corporate income (5 percent).2
Other components of own-source receipts include estate
and gift taxes, severance taxes, hospital and health
charges, rents and royalties, user fees, and licenses.
State and local governments administer two types of
sales taxes. One is the broad-based general sales tax
which is specified as a percentage of the price of a
product or service. The other is made up of selective
sales taxes—sometimes referred to as excise taxes—
which are levied on a unit of output. The principal
sources of revenue in the latter category are taxes on
motor fuels, tobacco products, and alcoholic beverages.
An important trait of the state and local tax system
is that it does not obtain nearly so much revenue from
income-based taxes as the Federal Government does.
Currently, taxes on individual income and corporate
income generate about 21 percent of state and local
own-source receipts but over 85 percent of Federal
revenue (excluding social insurance tax receipts). The
significance of this disparity is that income-based taxes
tend to be more elastic than other taxes. Of the major
taxes at the state and local level, only the individual
income and corporate income taxes have elasticities
with respect to inflation which are greater than one.3
Because of this structural difference in tax systems,
state and local own-source receipts are less responsive
to changing economic conditions than Federal revenues
are.
Two points are often overlooked in regard to the
second source of state and local revenue, Federal
grants. The first is that almost half goes for direct
transfer payments to individuals, such as Aid to Families
with Dependent Children (AFDC) and medicaid benefits.
Although they are measured as state and local
expenditures, the primary role of these governments is
to administer the programs. For the most part, they have
little or no discretion as to how or where the funds
should be spent. Second, while roughly 75 percent of
total Federal grants goes to state governments, much
of this is eventually passed down to local governments.

2Within the state and local sector, roughly 85 percent of the sales tax
receipts, 90 percent of the individual income tax receipts, and
virtually all of the corporate income tax receipts are collected by
state governments. Over 95 percent of total property taxes are paid
to local governments.
3This means that, if inflation increased by 1 percent, tax receipts will
grow by more than 1 percent.

4 FRBNY Quarterly Review/Winter 1983-84

Therefore, localities depend more heavily on intergov
ernmental aid than states do.
Expenditures
In 1983, state and local governments spent over $430
billion. The single largest category of expenditure is
education. After that comes income support and welfare,
health and hospitals, and transportation. Over 95 per
cent of their total spending goes to purchases of goods
and services, including wages and salaries.4 In contrast,
only about one third of all Federal expenditures go to
purchases of goods and services as over half of all
Federal outlays now go to benefit payments to individ
uals, e.g., social security and medicare, and to interest
payments on the debt.
Since virtually all state and local expenditures are for
purchases, the spending side of their budgets is not
automatically affected by cyclical changes in economic
activity to the same extent as the Federal Government.
When considering all categories of Federal expenditures,
many types of transfer payments, such as social secu
rity benefits, are likely to have automatic cost-of-living
adjustments (COLAs). The level of interest outlays is
also directly affected by economic fluctuations, because
nominal interest rates adjust to higher or lower rates of
inflation. Purchases, however, are usually not automat
ically indexed or as responsive to changes in inflation.
Trends
The fiscal profile of state and local governments has
changed considerably over the postwar period. This has
been evident in their financial position as well as in the
types of public services they provide and how they fund
them. Three general phases in the sector’s economic
activity have occurred since 1950. Each one is identified
by the rate of expansion for the sector in relation to the
Federal Government and to GNP and by the changing
role of states and localities in the U.S. federalist system
of government.
In the first phase, from 1950 through 1971, the state
and local sector grew considerably faster than either the
Federal Government (Chart 1) or GNP (Chart 2). During
this period, it also absorbed a continually larger share
of the economy’s income and real resources. For
instance, in 1950 state and local own-source receipts
were about 8 percent of aggregate personal income. By
the end of 1971, they had risen to 13.4 percent. Since
most of this went for purchasing goods and services,
the sector provided a strong stimulus for economic
growth. In real terms, its purchases increased from
about 10 percent of GNP in 1950 to 13 percent in 1971.
4As defined in the NIPA, purchases of services include employee
compensation. This convention is used throughout this article.

C h a rt 2

S t a t e and L o c a l E x p e n d it u r e s , R e ve n u e s , and S u r p lu s or D e fic it as a P e r c e n t a g e o f G N P *
P e rc e n t

A ll fig u re s are on a n a tional in com e and p ro d u c t a cco unt b a s is .
as d e fin e d by the N a tion al B ureau o f E cono m ic R esearch.

S had ed a re a s re p re s e n t p e rio d s o f re c e s s io n s ,

♦ S o c ia l in s u ra n c e re venue s and e x p e n d itu re s are e x c lu d e d .
S o u rc e :

U.S. D epartm ent o f C om m erce , B ureau o f E conom ic A n a lysis.

By contrast, over the same period, Federal tax and
nontax re ce ip ts fe ll from 19.4 percent of personal
income to 16.7 percent, and real Federal purchases
grew from 8.8 percent of real GNP to only 9.2 percent.
The rapid growth of the state and local sector was,
in part, the natural consequence of shifting economic,
dem ographic, and political conditions. Some of the
contributing factors were the effect of the baby-boom on
school enrollm ent, the construction of the interstate
highway system, the growing need for other forms of
in fra s tru c tu re due to p op u la tio n m ig ratio n, and the
increased demand for income support and public welfare
programs.
The c a p a c ity of state and local g overnm ents to
respond to each of these conditions was enhanced by
the proliferation of Federal grants. As shown in Table 1,
grants rose sharply in current and constant dollars
throughout the period. More importantly, perhaps, Federal
aid as a percentage of state and local revenue nearly
doubled, reaching over 20 percent in 1971. This meant
that, although state and local governments were providing
a broader range of goods and services, their ability to do
so was becoming increasingly dependent on Federal
budgetary policy.
Throughout the 1950s and 1960s, the state and local
sector ran a continual deficit (Chart 2). However, this

was not an indication of fiscal distress. Rather, it was due
both to the fact that states and localities were borrowing
to finance capital projects and to the method by which this
is accounted for in the NIPA.
There were several changes worth noting in the com
position of state and local expenditures and revenues
during this phase. Most of them took place toward the
end of the period and had a bearing on the direction of
state and local fiscal policies during the decade that fol
lowed.
On the spending side, the major development was the
growing emphasis on outlays for income support and
welfare between 1965 and 1971 (Table 2). Prior to that,
the share of state and local expenditures for education
had grown the fastest. Throughout the period, the portion
of outlays for transportation declined as the interstate
highway system neared completion. These developments
were augmented by similar modifications in the compo
sition of Federal grants (Table 3).
On the revenue side, starting in 1965 and continuing
through 1971, there was a concerted effort by states to
broaden their tax bases. In that time alone, seven states
adopted the individual income tax and eight added the
general sales tax. Furthermore, in 1971 especially, a
number of states that already had one or both of these
two taxes raised their rates.

FRBNY Q uarterly R eview/W inter 1983-84

The move by states toward broad-based taxes was
due in part to the fact that they are more elastic and,
particularly in the case of the individual income tax,
more progressive. Their adoption, as well as the sub
sequent tax rate increases in 1971, was also motivated
by the desire to supply local governments with additional
revenues while providing their constituents with relief
from the property tax. As the role of state and local
governments in providing goods and services expanded,
they could rely more on econom ic growth for the nec
essary additional revenues rather than on discretionary
tax increases.
The state and local tax reforms had a significant
impact on the level and composition of their own-source
re c e ip ts . The p e rio d from 1968 to 1971 saw the
sharpest sustained rise in own-source receipts in rela
tion to Federal revenues (Chart 1) and personal income
(Chart 3) in the postwar era. Almost all the increase was
accounted for by sales taxes (primarily the general sales
tax) and the individual income tax.
The second phase, starting in 1972, brought about a
slowdown in the average rate of expansion of the state
and local sector. A principal reason for the deceleration
was that the pressure from dem ographic factors that led

Table 1

Federal Grants to State and Local Governments*

Calendar
year
1950
1955
1960
1965
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
*1983

Current
dollars
(billions)

..................
..................
..................
..................

..................

2.4
3.2
6.5
11.1
24.5
29.0
37 5
40.6
43.9
54 6
61.1
67.6
77.3
80 5
88.7
87.9
83.9
86 5

Constant
dollars
(billions)

Grants as a
percentage ot
total state and
local revenuef

44
5.2
9.5
14.9
26.8
30.2
37.5
38.4
38.2
43.4
46.2
48.2
51.4
49 3
49.7
45.1
40.6
40.1

11 6
10.6
14 0
15 8
19.4
20.3
22.4
22.3
22 3
24.7
24.6
24.5
25.5
24.8
24.9
22.8
20.8
19.4

'O n a national income and product account basis.
tTotal revenue equals tax and nontax receipts plus Federal grants.
Social insurance contributions are excluded
^Federal Reserve Bank of New York estimates.
Source: U.S. Department of Commerce, Bureau of Economic
Analysis.

FRBNY Q uarterly Review/W inter 1983-84

to the rapid growth in the 1950s and 1960s, subsided
in the 1970s. For example, the enrollm ent of schoolaged children from kindergarten through high school
peaked in 1971.
By the end of 1977, state and local expenditures and
revenues were about the same as they had been six
years earlier in terms of either the general governm ent
(C h a rt 1) or GNP (C h a rt 2). The c o m p o s itio n of
expenditures also did not change by much (Table 2).
Although the level of state and local own-source receipts
remained fairly stable with respect to personal income,
the wave of tax reforms at the end of the first period
had altered the composition. The contribution of the
individual income tax to total own-source receipts dou
bled (lower half of Chart 3). At the same time, the share
accounted for by the property tax fell.
The slowdown in state and local econom ic activities
occurred despite the continued growth of Federal aid
(Table 1). By 1977, grants had reached record levels in
both nominal and real terms, and they accounted for
nearly one quarter of state and local total revenues.
There were two important changes in Federal grant
policy during this period. One was the advent of Federal
revenue sharing in late 1972. It gave state and local
governments more of a voice in deciding which social
service and welfare programs should be provided for
and the extent to which they should be funded. This is
evident in Table 3 where a sharp drop in Federal aid
for income support and w elfare in 1973 was more than
offset by the increase in grants going to the “ other”
category. The second was a 24 percent surge in Federal
grants in 1975. This time, much of it went to specific
programs in the areas of social services and medical
care.
Perhaps the most important budget developm ents at
the state and local level in this period were the flu c
tuations in the sector’s financial position (Chart 2). The
fluctuations were mainly the products of discretionary
policies at all three levels of government. However, state
and local budgets had also become more responsive to
changing economic conditions after the tax reforms of
1965 to 1971.
There were three swings in the status of state and
local budgets between 1972 and 1977. The first was a
peak in a g g re g a te s u rp lu s e s d u rin g the re c o v e ry
between the 1970 and the 1973-75 recessions. Besides
the healthier economy, the Federal revenue-sharing
funds and a number of state tax increases accounted
for the buildup. The surplus soon turned to a deficit,
however, as a result of the 1973-75 recession, a series
of state and local tax cuts, and an acceleration in their
expenditures. The third swing started after the second
quarter of 1975. Own-source receipts were bolstered by
the recovery, the second large in crea se in Federal

Table 2

State and Local Expenditures*

Calendar
year
1952
1955
1960
1965
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
§1982
§1983

Total
expenditures
(billions
of dollars)
24.9
32.2
48.7
73.5
131.2
147.9
162.1
178.3
200.9
228.5
247.3
265.6
293.0
317.9
352.8
381.1
406.0
430.5

.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................
.........................

Education

Transportation

33.3
37.0
37.8
41.0
41.5
41.0
40.9
41.2
40.7
40.7
40.6
40.5
39,9
40.4
40.2
39.9
39.9
39.8

18.5
18.9
18.3
16.5
13.2
12.4
11.7
11 2
11.6
10.2
8.9
8.4
8.7
9.0
9.2
9.0
9.1
9.2

Income
support and
w elfaret

13.7
12.1
11.5
11.0
14.6
15.5
15.9
16.0
15.1
15.5
15.9
16.0
15.5
15.2
15.5
15.7
16.0
16.2

Housing
and
Health and
community
Other
services
hospitals
(As a percentage of total expenditures)^
5.2
3.7
3.9
3.7
2.8
2.6
2.3
2.5
2.7
3,0
3.0
3.0
3.6
3.6
3.6
2.9
3.3
3.6

9.2
8.4
8.2
8.0
8.4
8.7
8.7
8.8
9.2
9.0
9.0
9.2
9.4
9.6
9.8
10.0
10.2
10.4

20.1
19.9
20.3
19.9
19.6
19.9
20.6
20.3
20.7
21.5
22.6
22.8
22.9
22.2
21.8
22.4
21.5
20.8

*On a national income and product account basis. Data for state and local expenditures by category are not available on a NIPA basis prior to 1952.
^Figures may not sum due to rounding.

flncludes medicaid.

§Federal Reserve Bank of New York estimates.

Table 3

Composition of Federal Grants to State and Local Governments4'

Calendar
year
1952
1955
1960
1965
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
111982
11983

Total
grants
(billions
of dollars)

2.7
3.2
6.5
11.1
24.5
29.0
37.5
40.6
43.9
54.6
61.1
67.6
77.3
80.5
88.7
87.9
83.9
86.5

Education

Transportation

7.5
8.0
6,3
7.5
15.0
13.5
11.8
9.7
11.3
10.2
7.4
8.1
7.9
9.0
8.9
9.0
9.4
9.1

18.5
22.9
38.7
35.1
19.4
17.9
13.8
12.2
12.9
12.8
12.1
11.2
11.1
12.7
13.6
13.3
13.0
13.5

income
support and
welfaret

57.8
55.8
40.2
37,7
42.1
44.5
44.2
35.9
35.5
37.5
39.1
38.2
36.9
38.2
40.5
45.4
47.4
49.6

Housing
and
community
Health and
Otherf
hospitals
services
(As a percentage of total Federal aid)§
0.3
1.2
2.1
4.4
7.4
7.9
6.8
6.9
8.7
8.6
9.7
9.6
8.1
9.7
10.0
9.4
10.7
11.5

4.4
3.1
4.7
6.4
5.3
4.9
4.0
5.1
5.6
4.7
4.5
4.2
3.8
3.6
3.7
3.9
4.1
3.9

11.5
9.0
8.1
8.9
10.7
11.3
19.4
30,2
25.9
26.2
27.2
28.8
32.3
26.9
233
19.0
15.4
12.4

*On a national income and product account basis. Data for Federal grants by category are not available on a NIPA basis prior to 1952.
flncludes medicaid.

^Includes revenue sharing.

§Figures may not sum due to rounding.

IIFederal Reserve Bank of New York estimates.
Sources for Tables 2 and 3: U.S. Department of Commerce, Bureau of Economic Analysis.

FRBNY Quarterly Review/W inter 1983-84

grants occurred, and the growth of expenditures slowed.
By 1977 the state and local sector had an operating
budget surplus of over $10 billion. At the time, it was
the largest one the sector had ever run.
The end of this second phase marked a critical turning
point for state and local governments. Financially, the
sector, as well as the individual governments, had never
been better off. In spite of that, a number of factors
were complicating the budget decisions facing state and
local policymakers. One was that state and local taxes
were near record levels as a share of personal income
(Chart 3). At the same time, though, state and local
governm ents were not providing increasing levels of
goods and services in real terms. Their purchases were
at a ten-year low in relation to real GNP, and in 1976

and 1977 their real per capita purchases fell for the first
time in the postwar period. Between 1950 and 1975 real
per capita purchases had risen steadily from $375 to
$840, but by 1977 they were down to $824. Finally,
states and localities continued to be highly dependent
on grants at a time when the Federal Governm ent was
running its largest deficits to date.
In many respects, state and local governments were
confronted with the same set of circum stances in 1977
as they are now. At issue was the extent to which they
should reduce taxes, increase spending, or m aintain
large balances. Their ensuing decisions brought about
the most sweeping changes in the fiscal profile of the
state and local sector in history. The outcome of those
decisions also provide the basis for the two sets of

C h a rt 3

S ta te and L o c a l O w n -S o u rc e R e c e ip ts as a P e r c e n ta g e of P e rs o n a l In c o m e
P e rc e n t

C o m p o n e n t s of O w n -S o u rc e R e c e ip ts

1950-1 to 1971-IV

C o rp o ra te
p ro fits
ta x e s
3.5%
In d ivid u a l
incom e
ta x e s

1972-1 to 1978-11

C o rp o ra te
p r o fits
ta x e s
4.5%

S o u rce :

C o rp o ra te
p ro fits
ta x e s

In d iv id u a l
incom e
taxe s

A ll fig u re s are on a n a tio n a l incom e and p ro d u c t a c c o u n t b a sis.
as d e fin e d by th e N a tio n a l B ureau of E c o n o m ic R e s e a rc h .

In d ivid u a l
in com e
ta x e s

S haded a re a s re p re s e n t p e rio d s of re c e s s io n

U.S. D e p a rtm e n t o f C o m m e rce , B ureau of E co n o m ic A n a ly s is .

8 FRBNY Quarterly Review/W inter 1983-84

1978-111 to 1982-IV

concerns regarding the policy decisions state and local
governments face today.
.In the third phase, from 1978 through 1982, three
different types of events had a substantial impact on the
state and local sector—“the tax revolt”, a sharp cutback
in Federal aid, and two recessions.
The tax revolt was more than likely due to the com
bination of circumstances surrounding the financial
position of state and local governments in 1976 and
1977 rather than to any one of them. The movement
was ushered in by Proposition 13 in California in June
1978. That legislation was specifically designed to
reduce property taxes. It received much of its impetus,
though, from the state’s $4 billion budget surplus which
enabled taxes to be cut without necessarily requiring the
provision of goods and services to be reduced as well.
Between 1978 and 1980, many other state and local
governments, also running high surpluses, cut taxes and
placed a ceiling on the rates of growth of own-source
receipts and spending by indexing them. In most cases,
the rates of growth of specific tax receipts and
expenditures were indexed to the growth of personal
income, the assessed value of property, or the growth
of population. From 1978 to 1980, thirty-two states
enacted a total of fifty-four reductions of a major tax.
Most of their efforts were aimed at the individual income
and general sales taxes. Local governments concen
trated primarily on the property tax.
The impact of the tax revolt on the scope of state and
local economic activities was immediate and substantial.
From 1978 to 1980, the tax cuts reduced state and local
own-source receipts by about $13.6 billion.5 They also
led to a deceleration in the growth of expenditures. For
the first time in the postwar period, total revenues and
expenditures at the state and local level declined on a
sustained basis relative to the general government
(Chart 1) and to GNP (Chart 2). State and local ownsource receipts fell to a ten-year low as a share of
personal income (Chart 3).
When state and local governments cut the effective
rates in the general sales and individual income taxes,
they also dampened the responsiveness of each tax to
inflation. Between 1978-111 and 1980-IY for instance, the
average rate of inflation in the United States was over
9 percent, as measured by the GNP deflator. At the
same time, the consumer price index rose by an
average of over 12 percent. Yet, for state and local
governments, individual income tax receipts remained

•Federal Reserve Bank of New York estimate. The Tax Foundation has
estimated that tax cuts enacted solely by states from 1978 to 1980 amounted
to about $4 billion. Over the same three years, figures from the Bureau of
Economic Analysis show that, for selected taxes at the state and local levels,
the cuts were worth roughly $9.5 billion.

flat and general sales tax receipts fell with respect to
personal income.
From 1979 to 1982, attempts by the Federal Govern
ment to reduce its budget deficit led to a reduction of
grants to states and localities in real terms first and then
in nominal terms as well (Table 1). By 1982, real Federal
aid was less than it had been in 1975. As a share of
state and local revenue, grants had fallen to 20.8 per
cent, the lowest amount in eleven years. The discon
tinuation of Federal revenue-sharing funds for state
governments in 1981 was the major change in the
makeup of grants. As shown in Table 3, it was offset
by an increase in the portion of aid going to income
support and welfare programs.
The decline in grants intensified the squeeze on state
and local revenues initiated by their own tax cuts.
However, state and local governments replaced only a
small percentage of lost Federal funds with their ownsource receipts. Ultimately, then, the cuts had a greater
effect on reducing the growth of both revenues and
expenditures for state and local governments than they
had on the severity of their financial problems.
The first quarter of 1980 marked the beginning of
state and local government financial troubles as the
economy fell into a recession. By the second quarter of
the year, the overall sector had registered its smallest
aggregate surplus in four years (Chart 2). In contrast
to the sector’s deficits between 1950 and 1971, the
decline in the NIPA surplus in 1980 signaled the
beginning of financial problems for individual state and
local governments as well. At the state level, for
instance, governments opened fiscal 1980 with a bal
ance of $11.2 billion left over from fiscal 1979. By the
end of fiscal 1980, the balance was $11.3 billion, indi
cating that state operating budgets ran an aggregate
surplus of only $0.1 billion.
The four quarters of recovery following the 1980
recession did little to ease the financial pressure on
state and local governments.6 In addition to the loss of
receipts from the tax cuts, indexation had reduced the
potential revenue gains from inflation. At the end of
fiscal 1981, just prior to the start of the second reces
sion, sixteen states had either a deficit or a balance
equal to less than 1 percent of outlays.7
The recession from 1981-111 through the end of 1982-1V
left the state and local sector in its worst financial
position in six years (Chart 2). Three states ended fiscal
1982 with a deficit. As a share of outlays, the balance
•The recovery ended in the second quarter of 1981, the same time
that fiscal 1981 ended for all but four states.
7AII survey data on the financial condition of state governments at the
end of a fiscal year reported in this article were obtained from the
National Conference of State Legislatures.

FRBNY Quarterly Review/Winter 1983-84 9

in fifteen other states was no greater than 1 percent.
In fact, only thirteen states did not face financial prob
lems as they ended the year with balances of 5 percent
or more. Conditions at the local level were only some
what better.
Since the recovery was widely expected to begin by
the middle of 1982, most state and local governments
formulated their fiscal 1983 budgets on that basis. As
the recession persisted throughout 1982, though, it
became apparent that many budgets would have to be
altered to avoid ending tne year with a deficit. Once
again, taxes were raised and the growth of spending
was restrained. Nevertheless, eleven states still
recorded deficits in fiscal 1983. As a percentage of
outlays, balances in fifteen states were 1 percent or less
and no more than 3 percent in fourteen others. Only
nine states had a balance equal to 3 percent or more
of outlays.
In response to their deteriorating financial position,
state and local governments adopted strict austerity
measures. From the beginning of 1981 to the end of
fiscal 1983, they raised taxes by over $18 billion.8 They
also placed further restrictions on the growth of
spending.
At first, the taxes that states raised were not the same
ones they cut between 1978 and 1980. As a rule, states
raise their general sales tax or individual income tax
only as a measure of last resort. Of the seventy-five tax
increases enacted by thirty-five states during fiscal 1981
and 1982, fifty-four were for excise taxes and half of
those were for the gasoline tax. As their financial posi
tions worsened through fiscal 1983, though, many states
were forced to turn to a broad-based tax for additional
revenue. Between 1981 and 1983, twenty-six states
raised their general sales tax, their individual income
tax, or both. The result was the most significant
increases in these taxes since 1965-71. Back then,
however, the goals were to broaden tax bases and to
reduce the burden of local property taxes.
The structure of the corporate income tax was prob
ably altered mpre than any other tax at the state level
during this period. The Accelerated Cost Recovery
System (ACRS) contained in the Economic Recovery
Tax Act of 1981 greatly liberalized depreciation allow
ances and reduced Federal corporate income tax lia
bilities.9 Since all states except California had been
following the Federal depreciation standards, they also
•Federal Reserve Bank of New York estimate. For the full three-year
period from 1981 through 1983, the Tax Foundation estimated that
tax increases at the state level were worth more than $14 billion.
•It was estimated that the Federal Government would lose 40 percent
of its corporate income tax receipts by 1986 as a result of the
provisions in ERTA.

10 FRBNY Quarterly Review/Winter 1983-84

stood to lose a large percentage of their corporate
income tax receipts. To avoid that loss, twenty-one
states either partly or fully decoupled their systems from
Federal depreciation standards while four others raised
the corporate income tax rate.
At the local level, some governments raised the
property tax following the 1980 recession and well into
the 1981-82 downturn. More often, though, localities
increased the individual income tax, sales taxes, user
fees, or a variety of other taxes and charges to raise
revenue.
On balance, then, there were two distinct sets of
policy combinations at the state and local level from
1978 through 1982. The first one, applied between 1978
and 1980, contained reductions of both taxes and the
growth of expenditures. In the second one, pursued from
1981 through 1982, there was an abrupt reversal in
aggregate tax policy, but the stance on spending
remained the same. Starting in 1978 and continuing
through 1983, both sets of fiscal policies had a signif
icant effect on the performance of the U.S. economy.
Economic impact of state and local fiscal policies
As a first step in examining the effects of state and local
fiscal policies on the 1980 and the 1981-82 recessions,
it is useful to look at how the aggregate components of
real GNP have varied during each of the eight postwar
recessions. Table 4 shows the changes in real GNP
from the preceding peak to the trough of each recession
as well as the underlying changes in real consumption,
investment, net exports, and government purchases.
In the first six recessions after World War II, state and
local government purchases stand out as the one
component of real GNP that always served to reduce
the magnitude of the downturn. The two most recent
recessions were a departure. In 1980, state and local
purchases in real terms registered their first postwar
decline during a recession. In the 1981-82 downturn
they increased by only $0.2 billion. This last recession
was also unique in comparison to the other postwar
recessions in that it contained the largest decrease in
real Federal grants and the largest increase in real
Federal purchases.
Focusing on the changes in the real purchases of
state and local governments, though, gives an incom
plete picture of the effect the sector had on past
recessions. For example, the breakdown of figures in
Table 4 makes it tempting to conclude that state and
local governments exacerbated the 1980 recession and
had virtually no effect on the most recent one. The fol
lowing analysis, however, shows that both conclusions
are incorrect. The reason is that the additional economic
effects caused by changes in state and local tax policies
and Federal grants are excluded in Table 4.

The changes in state and local governm ents’ tax
policies had an impact on both their real purchases and
on certain economic variables which, in part, determine
several other components of real GNP. For instance,

when state and local governments rebuild their balances
during a recession by raising taxes, disposable income
is reduced and, other things being equal, real con
sumption is lowered. Furthermore, not all the changes

Table 4

Changes in the Components of Real GNP in the Eight Postwar Recessions4
In billions of 1972 dollars, seasonally adjusted annual rates

Eight postwar recession periods
Real GNP and its
com ponents
Real GNP ...............................
Personal consumption .........
Nonresidential investment
Inventory investment ...........
Federal purchases ...............
State and local purchases ..
Net exports ...........................
Addendum:
Real Federal grants to state
and local governments

1981-111
to
1982-IV

1980-1
to
1980-111

1973—IV
to
1975-1

1969-1V
to
1970-1V

1960-11
to
1961-1

1957-111
to
1958-11

1953-11
to
1954-11

1948—IV
to
1949—IV

-4 5 .1
16.7
-1 6 .5
-3 8 .8
12.6
0.2
-1 6 .8

-3 2 .2
-8 .9
-8 .2
- 9 .6
03
-0 3
3.6

-6 1 .8
- 3 .4
-2 0 .0
-3 8 .0
0.6
6.1
10.9

- 0 .9
10.0
-7 .1
- 5 .6
-1 1 .6
7.5
1.8

-0 .7
-0 .1
- 2 .6
-7 .4
2.2
4.8
3.6

-1 8 .6
- 0 .5
- 8 .5
- 9 .9
2.5
4.8
- 6 .3

-2 0 .2
1.6
-1 .1
- 9 .2
-1 9 .4
4.7
2.3

-7 .1
7.1
- 8 .2
-1 3 .0
0.2
6.4
- 2 .3

- 4 .0

0.1

2.4

2.2

0.8

2.1

-1 .1

0.3

'Changes are expressed as the first difference from the peak to the trough of each recession.
Source: U.S. Department of Commerce, Bureau of Economic Analysis.

Table 5

Sources of Change in State and Local Government Revenues and Expenditures
In billions of dollars, seasonally adjusted annual rates

Sources of change
In revenues .................................................................
Due to:
Discretionary policy ..............................................
Federal grants .....................................................
‘ All other .................................................................
In expenditures ...........................................................
Due to:
Discretionary policy ..............................................
fFederal grants .....................................................
’ All other .................................................................
In real purchases (1972 dollars) ............................
Due to:
Discretionary policy ..............................................
fFederal grants .....................................................
’ All other .................................................................

Period 1

Recession

Period 2

Recession

1978-111 to
1980-1V

1980-1 to
1980-1 If

1981-1 to
1983-IV

1981 —III to
1982-IV

68.5

12.4

82.7

25.8

-1 3 .6
15.2
66.9

-3 3
3.3
12.4

21.4
- 4 .0
65.3

14.3
- 1 .4
12.9

67.2

15.3

68.1

33.1

-7 .2
22.9
51.5

-0 .6
4.3
11.6

- 2 .8
1.6
69.3

- 6 .5
1.8
37.8

2.6

- 0 .3

- 2 .0

0.2

-2 .4
3.0
2.0

-0 .8
-0 .6
1.1

- 0 .9
- 6 .8
5.7

- 3 .3
-0 .4
3.9

*These are due primarily to changes in both real and nominal economic conditions. Demographic factors also have a small impact.
-(-Includes both the actual change in grants plus the induced changes in expenditures due, for example, to matching requirements for
state and local governments.
Source: Federal Reserve Bank of New York estimates.

FRBNY Quarterly Review/W inter 1983-84

in state and local governments’ real purchases were due
to discretionary action on their part. A portion of the
changes came about because of the cutbacks in Federal
grants.
In assessing the total impact of the state and local
sector on the last two recessions, then, it is important
to isolate the effects attributable to state and local dis
cretionary tax and expenditure policies, Federal grant
policies, and economic conditions. Otherwise, mis
leading conclusions could be drawn as to whether or not
the actions of state and local governments were coun
tercyclical or if they supported Federal efforts at re
storing economic recovery. For purposes of evaluating
fiscal policies at all levels of government, this distinction
is crucial.
The changes in revenues, expenditures, and real
purchases of state and local governments due to their
own discretionary policies, the level of Federal grants,
and all other factors are presented in Table 5.10 Two time
periods are analyzed. The first, from 1978-111 to 1980IV, corresponds to the period in which state and local
taxes were being cut and the growth of spending was
restrained following the tax revolt. This period includes
the recession from 1980-1 to 1980-111. In the second
period, from 1981-1 to 1983-1V, state and local taxes
were raised and, at least through the beginning of 1983,
spending was further restrained. The six quarters from
1981-111 to 1982-IV is the recession in which this second
mix of policies was pursued.
The overall impact of state and local governments’
discretionary policies from 1978-111 through 1980-IV was
expansionary, as they reduced taxes by more than they
reduced spending. During this time, Federal grants
increased in nominal terms but fell in real terms. Even
so, the changes in state and local governments’
expenditures resulting from the change in grants were
10The figures in Table 5 were derived from a ten-equation quarterly
econometric model of the state and local sector estimated by the
author. The model is patterned after the specification of the state
and local sector in the Federal Reserve-MIT-Penn (FMP) model and
was estimated using NIPA data. The model works as follows. First,
total state and local spending is determined from equations
estimated for employee compensation, structures, other purchases,
and transfer payments. The explanatory variables for these
expenditure equations include measures for personal income,
relative prices, interest costs, population, unemployment, Federal
grants, and lagged expenditures. Next, the portion of expenditures
that would have to be financed through state and local own-source
receipts is estimated, defining the sector's net revenue requirement.
Then, each component of state and local own-source receipts is
expressed as a share of the total and estimated as a function of
personal income, household wealth, corporate profits, inflation, and
the change in the net revenue requirements. By estimating each
component as a share of the total net revenue requirement, the
sector’s budget constraint is imposed. The components for ownsource receipts include the individual income tax and other taxes,
sales taxes, indirect business taxes (includes the property tax), and
the corporate profits tax.

12 FRBNY Quarterly Review/Winter 1983-84

positive, as many of them failed to anticipate fully the
extent of the grant cutbacks when planning their
budgets.
Only a small portion of the expansionary effect of
state and local fiscal policies between 1978-111 and
1980-IV was felt during the 1980 recession. The decline
in revenues due to the tax cuts had slowed, and the
cuts just outweighed the discretionary reductions of
expenditures and real purchases. Federal grants were
not a significant factor in determining the sector’s total
revenues and expenditures.
In the second period, from 1981-1 to 1983-IV, state
and local discretionary policies were clearly contrac
tionary. The wave of tax increases generated over $21
billion in additional receipts. Moreover, the limits and
absolute cuts in spending continued to reduce expen
ditures and real purchases. The Federal Government
reinforced the impact of state and local policies
throughout the period. As grants declined in real terms
in 1982 and 1983, so did the state and local purchases
tied to those funds.
The contractionary effects of state and local govern
ment policies were especially strong during the 198182 recession. Revenues grew by over $14 billion as a
result of discretionary tax increases. At the same time,
state and local government spending policies led to a
reduction of $6.5 billion in expenditures. Over the course
of the downturn, real state and local purchases aver
aged about 11.7 percent of real GNP, the lowest since
1965.
Near-term outlook
A new phase in the fiscal profile of state and local
governments may now be under way. Their recent tax
increases provided a strong boost to revenues, and
most of them continued to hold the line on spending.
In addition, the recovery was stronger than most ana
lysts and policymakers had originally expected. As a
result, by the end of calendar year 1983, the state and
local sector as a whole registered a large operating
surplus for the third consecutive quarter (Chart 2).
Furthermore, a survey of state budget offices revealed
that only three states anticipate deficits for fiscal 1984.11
In spite of their improved financial conditions, state
and local governments still face a number of difficult
decisions in planning their budgets over the next several
years. This is because the same set of troubling cir
cumstances existing in 1976 and 1977 exist today.
• By postwar standards, the state and local sec
tor’s surplus is the largest ever. State and local
"Steven Gold and Corina Eckl, State Fiscal Conditions Entering 1984
(National Conference of State Legislatures, Denver, Colorado, 1984).

governments’ taxes are as high as they have
ever been with respect to personal income, and
their real purchases are at a 21-year low as a
share of real GNP.
• In 1983 the Federal Government ran its largest
unified budget deficit in history—$195.4 billion,
or 6.1 percent of GNP. Under current policies,
the Federal deficit is expected to average over
5 percent of GNP throughout the rest of this
decade. By 1989, this could amount to approx
imately $300 billion.
State and local policymakers are certainly aware of
these circumstances and, given their experiences since
1978, are attempting to prepare for them. In some
cases, their preparations involve rather new and inno
vative policies.12 For example, twenty states now have
“ rainy day” or "budget stabilization” funds. Many of
these were established in the last several years. States
can draw upon these funds during lean economic times
and rebuild them during prosperous times. Some states
are also considering the adoption of “trigger” taxes that
go into effect automatically if budget problems arise.
Just how well prepared state and local governments are
should be tested soon. For instance, the recent tax
increases which led to the sector’s current surplus could be
scaled back over the next few years. Although thirteen
states raised their individual income tax in 1983, the
increases were only temporary in seven of those states and
will expire in either 1984 or 1985. Five of the fourteen
general sales tax increases enacted in 1983 will expire
during 1984. Finally, if voters believe that the Federal deficit
will be reduced by higher taxes, they may call for an off
setting reduction of state and local taxes.
On the spending side of their budgets, state and local
governments may find it increasingly difficult to hold the
line on the growth of expenditures. In the immediate
term, pressure to increase expenditures will come from
the area of education and from state and local
employees who, in the past few years, have settled for
either partial wage adjustments or none at all.
Perhaps the most serious circumstance facing state
and local governments is what may be the most sig
nificant infrastructure financing needs in their history.
The bulk of the expenditures related to the capital
projects they undertake will be financed by issuing debt.
Even so, state and local current operating budgets may
have to be adjusted to cover additional expenditures

related to infrastructure projects or a portion of the
interest costs from the increased borrowing.
identifying infrastructure financing needs and pro
jecting the potential costs is difficult. Nevertheless, the
Congressional Budget Office estimates that, from 1984
to 1990, annual capital outlays by all levels of govern
ment will have to be about $28 billion for repairs,
rehabilitation, and replacement on existing infrastructure
systems.13 To meet growing demands on existing sys
tems, they estimate an additional $25 billion will be
needed each year for new construction. That totals to
nearly $375 billion over the seven-year period.
Presently, Federal, state, and local governments
spend about $36 billion a year for capital outlays. Under
current policy, the Federal Government would finance
about half of the estimated additional needs. That share
could fall below half, though, if the Federal Government
decided to reduce its deficit by limiting its involvement
in funding infrastructure projects. How state and local
officials restructure their borrowing, taxing, and
expenditure policies to finance the remaining portion
could be their most severe test of all.
Concluding remarks
Any examination of fiscal policy in the United States
excluding state and local governments is incomplete.
State and local fiscal policy actions have had a partic
ularly significant impact on the economy during the past
five or six years. In view of the budget decisions state
and local governments must make in the next few years,
their actions will in all likelihood continue to be an
important factor in economic growth.
What this suggests is that, for purposes of macroeconomic analysis and policy, we must consider the
economic activities of all levels of government: Federal,
state, and local. Too often, only Federal financial prob
lems and policy decisions are evaluated. Certainly, one
of the most critical issues over the near term will be
selecting a course for the Federal budget. However,
given the interrelationships between Federal policies, the
economy, and state and local financial conditions, the
course which is eventually chosen could have a sub
stantial bearing on the direction of state and local fiscal
policies as well. Recognizing this, and incorporating it
into the decision-making process, would be an important
first step toward coordinating fiscal policies across all
levels of government.

Peter D. Skaperdas
12For a more detailed discussion of these policies, see Steven Gold,

Preparing for the Next Recession: Rainy Day Funds and Other Tools
for States (National Conference of State Legislatures, Denver,
Colorado, 1983).

13Congressional Budget Office,

Considerations for the 1980s

Public Works Infrastructure: Policy

(Washington, D C., April 1983). All cost
estimates are in 1982 constant dollars.

FRBNY Quarterly Review/Winter 1983-84

Itoelve Improvements in the
Municipal Credit System

The United States faces the huge task of renovating its
public capital infrastructure. There are several signs of
political willingness to get on with the job, such as
Congressional passage of the five cents a gallon gas
tax and voters’ approval of the “ Rebuild New York” bond
referendum. However, over the last few years several
changes in Federal policy and state and local govern
ment practices may have raised the cost of capital to
finance infrastructure projects at just the time when it
has become apparent that more such investment is
needed.
Improving our infrastructure will be costly in any
event, but it will be more difficult than it needs to be
without some successful effort to improve the operation
of municipal credit markets. A number of changes in the
municipal credit system are occurring or are being dis
cussed. If some combination of these changes were
implemented and if they were successful, it is conceiv
able they could produce a 20 to 25 percent savings in
the cost of servicing debt for infrastructure financing.
Three items of evidence indicate that there is room
for improvement in the way municipal credit markets
work. First, yields on municipal bonds have never been
as low, relative to corporate or Treasury yields, as they
“ should” be, given the advantage of tax exemption.
Furthermore, since 1979 municipal (tax-exempt) yields
have risen markedly relative to taxable yields (Chart 1).
Although the extremely high values of this ratio in 1982
are not unprecedented and the ratio has been falling,
few observers expect it to return to the low levels of the
late 1970s. Second, as Chart 2 suggests, the share of
Digitized for 14
FRASER
FRBNY Quarterly Review/Winter 1983-84

credit market borrowing flowing to state and local gov
ernments tends to rise when interest rates are relatively
high. This may mean that the borrowing behavior of
state and local governments is less sensitive to credit
market conditions than that of some other borrowers.
Finally, through the past decade the proportion of new
tax-exempt issues for “ nontraditional” or “ private” pur
poses has been rising (Chart 3).
In part as a result of these trends in the municipal
bond market and in part because of other forces, state
and local borrowing specifically dedicated to traditional
infrastructure projects has been held to relatively low
levels through most of the last twelve years. An effort
to reduce the cost of financing public capital projects
relative to the cost of capital for other purposes might,
therefore, be a useful element of any overall strategy
for dealing with the infrastructure problem. This would
involve a series of efforts aimed at reducing the ratio of
yields on tax-exempt bonds to yields on taxable bonds.
The yield ratio between instruments of equal riskiness
“ should” be equal to (1-m), where m is the marginal
income tax rate faced by the marginal investor in taxexempt securities. Because since 1971 the marginal
corporate tax rate has been between 46 and 48 percent,
the exempt/taxable yield ratio should have been as low
as 0.52 at those times when commercial banks were the
marginal purchasers of municipal bonds. The ratio of
yields on municipal to those on corporate bonds of
equal rating has never been lower than about 0.60 after
the early 1950s and, at times, the ratio has risen above
0.80 (Chart 1).

C hart 1

C hart 2

T a x a b le Bond Y ie ld C o m p a r e d with
E x e m p t/T a x a b le Bond Yield Ratio

C o m p a ris o n of M u n ic ip a l Bond Y ield to
S ta te and Local O b l i g a t i o n s ’ S h a r e of
T otal New C re d it M a r k e t D e b t

A nnual ave rage s, 1950 th ro u g h 1983
Ratio

Y ield

A nnual a ve ra g e s, 1969 th ro u g h 1983
P erce nt
12---------

Aaa m unicipal bond y ie ld as a p ro p o rtio n of

M o ody’s Aaa
c o rp o ra te bond y ie ld S c a le -

L j j j j j J.i I l I m

l i I i i II i i i I i I i II i i I

1950

* "N ew " po stw ar peak o f M o ody's Aaa c o rp o ra te
bond rate.
S o u rce :

M o o d y 's In v e s to rs S e rv ic e , Inc.

Reductions of the cost of capital to state and local
governm ents, w ithout new direct intergovernm ental
subsidies, could be realized by working toward the fo l
lowing broad goals:
• Increasing the liquidity of municipal bonds as
investm ent vehicles.
• Decreasing the riskiness, from the inve sto r’s
point of view, of bonds issued for infrastructure
purposes.
• Increasing the demand for traditional purpose
municipal bonds relative to the demand for other
vehicles with sim ilar risk and liquidity charac
teristics.
• Improving the flow of information to potential
investors.
• Relaxing constraints on municipal financial o ffi
cers th at lim it th e ir a b ility to econom ize on
financing costs.
• Reducing the cost of underwriting and marketing
services to issuers and investors.

1969

*1983-1 to 1983-111 a ve ra g e .
S o u rce : B oard of G o ve rn o rs of the F e d e ra l R eserve
System (u n a d ju ste d flo w -o f-fu n d s data) and M o o d y’s
In v e s to rs S e rv ic e , Inc. (A aa m u n ic ip a l b o nd y ie ld s ).

Twelve changes in the municipal bond market
There are at least twelve potential improvements in the
operation of the municipal bond market or in the prac
tices of participants in that market which offer promise
of reducing the cost of capital for traditional infrastruc
ture purposes. But without extensive analytical effort it
would be impossible to know whether any one or com
bination of these changes would have a beneficial net
effect. The purpose here is to advance that effort and
to suggest how additional work might be organized.
The first four potential improvements require Federal
Government action. The next three involve private-sector
initiatives. Four more suggest state and local govern
ment action, and the final innovation involves the cre
ation of a new type of institution.
The taxable bond option (TBO)
Under the TBO, a perennial reform proposal, m unici
palities would have the option of issuing taxable debt
instruments1 but, whenever a taxable municipal bond
was issued, the Treasury would guarantee the issuer a
stream of payments equal to a prestated proportion of

’ For a full discussion of the TBO, see David C Beek, "R ethinking
Tax-Exempt Financing for State and Local G overnm ents” , this
Review (Autumn 1982).

FRBNY Quarterly Review/W inter 1983-84

C hart 3

Real L o n g -t e r m D e b t Iss u e s of S t a t e and
L o c a l G o v e rn m e n ts , by Use
19 72-83
B illio n s o f 1983 d o lla rs
100

I D e d ic a te d to tra d itio n a l u s e s ’

0|---- 1---- 1---- 1---- 1---- 1---- 1---- --------------------------1972 73

* G a s and e le c tric ; h o s p ita l; m u ltip u rp o s e bond is s u e s .
^ In d u s tria l aid; p o llu tio n c o n tro l; s ta te and m u n ic ip a l
h o u s in g fin a n c e .
^ S c h o o l; w a te r and se w e r; high w ay, brid ge, and tu n n e l.
S o u rce :

The B ond B u y e r.

the interest cost of the taxable bond. Because issuers
would opt for taxable bonds only when it paid them to,
net in te re st costs to m unicipal borrow ers w ould be
reduced. In addition, the TBO would be more efficient
than tax-exempt bonds from the Treasury’s point of view.
When exempt bonds are issued, the Treasury loses
more in tax revenues than state and local governments
receive in terms of interest cost savings. Given a TBO,
when the option is exercised, under some assumptions,
the cost to the Treasury is exactly equal to the benefit
to the issuer (box). However, it is not obvious how the
market would receive a taxable municipal bond. Some
of the support for the TBO is based on the assumption
that ta xab le issues w ould provide a way in which
m unicipalities could tap the pool of capital held by
untaxed in s titu tio n s , e sp e c ia lly the rap id ly growing
pension funds. However, given some of the other prob
lems associated with m unicipal bonds— especially the
thinness of the secondary market and the lack of widely
recognized inform ational standards in the industry— it
could be that pension fund managers would buy taxable
bonds only at a substantial premium over the yields on
“ s im ila r” corporate issues. Furtherm ore, if untaxed
inve sto rs did purchase large volum es of m unicipal
bonds, some of the expected benefits of this proposal
to the Treasury would not materialize.

16 FRBNY Quarterly Review/W inter 1983-84

Opposition to the TBO focuses on concern over pos
sible increases in Federal control over state and local
government finance. It might be possible to design TBO
legislation so that the Treasury reim bursem ent was
perfectly automatic, but many observers are skeptical
about divorcing Federal funding from Federal regulation.
Other opponents are unwilling to concede a Federal
constitutional right to tax interest payments by state and
local governments.
Lim iting “ private use” tax exemption
In the first session of the 98th Congress, action on one
pending tax bill was delayed by the controversy sur
rounding provisions affecting the use of so-called “ pri
vate purpose” tax-exem pt bonds: mortgage revenue
bonds and small issue industrial development bonds, the
two fastest growing segments of the tax-exem pt bond
market. These instruments provide a means through
which home buyers and private firms can benefit from
the Federal tax exemption of municipal bond interest
payments.
Either of these “ private purpose” uses of tax-exem pt
financing may or may not make sense as instruments
of public policy. Our concern here, however, is the effect
the expansion of these forms of financing may have on
the cost of borrowing for more traditional state and local
government activities. It is commonly believed that the
market will not absorb large volumes of new municipal
issues without large increases in the tax-exem pt yield
relative to the yield on taxable securities. Therefore,
1982 issuance of $16 billion of tax-exempt debt for state
and municipal housing finance and another roughly $3
b illio n in in d u s tria l d e v e lo p m e n t b o n d s — to g e th e r
accounting for about 20 to 25 percent of the tax-exempt
market— may have had a substantial effect on the cost
to state and local governments of borrowing for more
tra d itio n a l purposes. E stim ates of the e ffe ct of the
aggregate supply of municipal bonds on the yield of
these securities, if the taxable yield were held constant,
vary from 0.6 basis points to 7 basis points per each
additional billion dollars of municipal bonds.2 Hence a
halving of the issuance of m ortgage revenue and
industrial development bonds might reduce municipal
yields by between 6 and 67 basis points, or by up to
about 7 percent of current yields.
Commercial bank underw riting of revenue bonds
Under the Glass-Steagall Act, commercial banks are not
allowed to participate in most revenue bond under
writing. Legislative proposals that would expand the role
2Roger C. Korm endi and Thomas T Nagle, “ The Interest Rate and Tax
Revenue Effects of M ortgage Revenue B onds” , in G eorge G.
Kaufman, ed., E fficiency in the M u nicipal B ond M arket (JAI Press,
Greenwich, CT 1981), pages 117-48.

The Taxable Bond Option: Interest Cost Savings and the Efficiency of the Subsidy
Suppose the Federal Government had, over the forty quarters
through 1983-111 made a binding offer to pay state and local
governments 31 percent of their interest payments on all taxable
municipal bonds they issued. The 31 percent figure is used
because over that period municipal Aaa yields averaged 69
percent of corporate Aaa yields. Suppose further that all issuers
exercised this option whenever and only when the yield ratio
exceeded 69 percent, but that the volume of new issues and
the series of taxable and exempt interest rates was unaffected
by the availability of the taxable bond option. Assume, finally,
that coupon yields on taxable municipal bonds were identical
to corporate yields on similarly rated issues and that all bond
holders’ marginal tax rate is 0.50.
Under these rather strong assumptions, two effects of the
taxable bond option may be observed. First, the average net

interest cost of municipal borrowing would have been lower
than it actually was (Chart 4-A). Second, the efficiency of the
subsidy to state and local governments, as measured by the
dollars lost to the Federal Government divided by the dollars
of interest cost saved by tax-exempt issuers, would increase.
When a tax-exempt bond is issued, the Treasury loses all the
taxes it would have collected on a taxable bond, but the locality
benefits only by saving the difference between the tax-exempt
yield and what it would have paid on a taxable issue. If the
typical marginal tax rate on municipal bondholders were 50
percent, then the subsidy to issuers would be less than the
cost to the Treasury whenever the yield ratio was greater than
0.50. The efficiency gain associated with a taxable bond option
with a 31 percent subsidy rate (Chart 4-B) would have been
roughly 46 percent.

Chart 4-A

C h a rt 4 -B

A v e ra g e In te r e s t R ate on M u n ic ip a l Issues
W e ig h te d by A nnual N e w -ls s u e V o lu m e

E ffic ie n c y of S ub sidy:
Loss to
T r e a s u r y p e r D ollar S av in g s to Issuers

1973-IV th ro u g h 1983-111

1973-IV throug h 1983-111

P erce nt pe r y e a r
9

A ctual

F u lly e x e rc is e d TBO
w ith 31% s u b s id y ra te

A ctual

Fully e x e rc is e d TBO
w ith 31% sub sid y rate

S o u rce s: S ta ff c a lc u la tio n s based on data fro m the B o a rd of G o ve rn o rs o f the F e d e ra l R e se rve S ystem
(u n a d ju s te d flo w -o f-fu n d s d a ta ) and M o o d y 's In v e s to rs S e rvice , Inc. (A aa b o nd y ie ld s ) .

FRBNY Quarterly Review/W inter 1983-84

of commercial banks in municipal bond underwriting
have been analyzed periodically over the past fifteen
years. Proponents of commercial bank underwriting
argue that it would bring greater competition to the
municipal bond underwriting industry, reducing coupon
yields. Opponents argue that the commercial banks'
advantages as underwriters are so overwhelming they
would soon drive the investment houses from the field,
ultimately reducing competition and driving up yields.
Commercial bank underwriting could reduce interest
costs faced by revenue bond issuers, but it is difficult to
estimate how great the reduction would be. Past empirical
studies of this question suggest that yields on new rev
enue bonds could be reduced by up to 6 percent—or
roughly 50-60 basis points at current yields.3 But it is
worth bearing in mind that, however useful this change
might be for revenue bonds, the impact on the costs of
financing infrastructure would be smaller, since infra
structure projects tend to be financed through general
obligation bonds, not revenue bonds. On the other hand,
even in the unlikely event that the commercial banks
drove the investment houses from the underwriting field,
the result need not be to reduce competition, so long as
the banks compete vigorously among themselves.
New tax laws and deregulation
The Federal Government’s influence on the tax-exempt
market is not limited to policies directly concerning
municipal bonds. General tax and regulatory policies
also have a substantial effect. Any reduction of high
bracket marginal tax rates on corporations or wealthy
individuals affects the exempt/taxable yield spread.
Whenever the Congress tries to encourage any type of
investment by granting special tax treatment, there is
a chance that some taxpayer, otherwise disposed to
investing in municipal bonds, will not buy them. One
important example of this phenomenon is the effect of
the accelerated depreciation provisions of the corporate
tax law on commercial banks’ choice of tax shelters.4
Regulatory changes, such as those that have increased
interest rates on time deposits, and that have the effect
of reducing commercial bank and property and casualty
insurance company taxable profits also lower the
demand for municipal bonds.
Some combination of tax law and regulatory changes
might return commercial banks, along with property and
casualty insurance companies, to a dominant role in the
3Phillip Cagan, “ The Interest Savings to States and Municipalities
from Bank Eligibility to Underwrite All Nonindustrial Municipal
Bonds” , Governmental Finance (May 1978), pages 40-48; Michael H.
Hopewell and George G. Kaufman, "Commercial Bank Bidding on
Municipal Revenue Bonds: New Evidence” , The Journal of Finance
(December 1977), pages 1647-56.
4See Allen J. Proctor and Kathleene K. Donahoo, this
Review, pages 26-37.

Quarterly

18 FRBNY Quarterly Review/Winter 1983-84

municipal bond market. If corporations dominated the
market, then the exempt/taxable yield spread might be
much wider than it is. In fact, in the late 1970s, the last
time institutions purchased the lion’s share of new
issues, the exempt/taxable yield ratio reached a record
low. If the municipal/corporate yield ratio had been 0.61
in December 1983, as it was on average in 1979, then
the tax-exempt yield would have been reduced by about
18 percent.
More aggressive marketing
The change in the municipal bond market that is prob
ably most obvious to the general public, especially in
the New York metropolitan area, is the new aggres
siveness with which municipal bonds, municipal funds,
and municipal unit trusts are being marketed. Extensive
advertising in the print and broadcast media have
stimulated more awareness of the advantages of
municipal bond investment. Furthermore, the products
offered by mutual bond funds and municipal unit trusts
have allowed investors with smaller portfolios and less
sophistication to realize these advantages.
Expansion of the demand for municipal bonds through
aggressive marketing probably has made it easier to
finance a record volume of new municipal issues at a
time when the institutions were playing a small role.
However, creating a new market through media adver
tising is an expensive undertaking. Most likely, the costs
of advertising have been divided among the dealers, the
investors who pay the dealers’ commissions, and the
issuers.
Third-party guarantees
Third-party guarantees of interest and principal pay
ments on individual municipal bonds or on municipal
bond portfolios have become much more common over
the past four years. There are several forms of these
guarantees. State government backing, in one form or
another, of local government or public authority obli
gations has been familiar for a number of years.
The newer forms of third-party guarantees are issued
by private-sector firms: commercial banks and municipal
bond insurance companies. Commercial bank backing
usually takes the form of an irrevocable letter of credit
in an amount sufficient to meet all outstanding interest
and principal payments on the guaranteed bond. Letter
of credit backing is more typically associated with short
term securities than with the long-term issues that are
the focus of this paper, although some letters of credit
irrevocable for ten-year periods have been written. Pri
vate guarantees of long-term municipal bonds are pro
vided by one of the three municipal bond insurance
companies. The recent performance of the two oldest
of these firms—the American Municipal Bond Assurance

Corporation (AMBAC) and the Municipal Bond Insurance
Association (MBIA)—reflects the remarkable growth of
this form of third-party guarantee. AMBAC, for example,
insures new municipal issues and the portfolios of
investors. Total insurance in force grew 770 percent
from about $6 billion in 1978 to $52 billion late in 1983.
The incidence of insurance coverage has risen from not
much more than 1 percent of new issues in 1979 to
close to 15 percent in 1983.
Municipal bond insurance companies provide two
services. First, like all insurance companies, they pool
the risk associated with their covered municipal bonds.
Second, insurance companies provide a service of
special value to those municipalities that can prove to
knowledgeable analysts that their bonds are less risky
than the market perceives them to be. In fact, since
Standard and Poor’s automatically assigns a AAA rating
to bonds insured by either of the currently active
insurance companies and Moody’s shows signs of
recognizing the credit enhancement provided by insur
ance, the insurance companies may take over part of
the rating agencies’ traditional functions. Standard and
Poor’s and Moody’s would devote their efforts to ana
lyzing uninsured issues along with the financial sound
ness of the insurance companies themselves.
Finally, third-party guarantees generate the additional
benefit of increasing the liquidity of the insured bonds.
The market for obligations of small municipalities or
obscure agencies may be extremely thin and the illi
quidity premium on their obligations, therefore, very
high. However, all the bonds insured by, say, MBIA
might trade as freely as the obligations of MBIA itself.
In other words, availability of insurance backed by
widely known AAA-rated financial service corporations
may introduce some needed uniformity into a market
with about one million separate issues.
An illustrative computation suggests the magnitude of
the savings available to issuers. In 1982 the yield on
Moody’s Aaa-rated twenty-year general obligations
averaged 10.30 percent and the Baa yield 11.58 per
cent. Suppose a Baa borrower issued $1 million worth
of bonds at a yield of 11.58 percent on the entire issue.
Suppose further the issue was designed like a home
mortgage: to be retired in equal annual payments over
twenty years. The annual payments would be $130,370.
Now suppose that by purchasing insurance, with a
premium equal to 0.8 percent of all interest and principal
payments, the issuer could have offered a coupon yield
of 10.30. The annual payments, including the premium,
would then be $120,833, or a savings of about 7 per
cent.
With the advantages introduced by third-party guar
antees, it is not surprising that their use continues to
grow rapidly. We cannot, however, be certain that this

expansion has been or will be trouble free. Roughly half
of the new municipal issues of 1982—those rated A or
Baa—could have benefited from and might have been
eligible for insurance. If, eventually, even half of these
Baa- and A-rated issues obtain insurance and if the total
value of new issues reaches $100 billion per year, then
this branch of the insurance industry will be writing
policies with face values of some $25 billion dollars a
year. The criteria for soundness and prudence in the
municipal bond insurance business may be very dif
ferent from the criteria used in evaluating more tradi
tional lines of the insurance industry and, in any case,
current regulations have not yet met the test of time.
As this industry develops, insurance regulators will have
to develop and expand this new, specialized form of
expertise.
A more troublesome potential problem concerns
municipal bond insurers who are, quite prudently,
unwilling to take all risks. As the incidence of insurance
becomes more widespread, municipalities unable to
obtain coverage may come to bear an additional stigma
in the market. In other words, a Baa-rated uninsured
issue might require an even higher premium yield than
marginal investment grade issues do now. If these
stigmatized municipalities are the ones with the most
severely dilapidated infrastructure, the advent of thirdparty guarantees might make it more difficult to solve
an important part of the infrastructure problem.
Municipal bond futures trading
Municipal bonds are generally considered relatively illi
quid investments. For one thing, market turnover is
small relative to the volume of outstanding issues. For
another, the relatively wide bid-ask spreads for bonds
listed on a regular basis raises the cost of buying and
selling tax-exempt bonds. The bid-ask spread for even
such widely held securities as seasoned Municipal
Assistance Corporation (MAC) bonds is typically
between 3 and 4 percent of the asking price. This is a
narrower proportional spread than is typical of, say, the
bid-ask differences in the daily over-the-counter quo
tations for equity prices of small, new, relatively spec
ulative companies. However, the MAC spreads are much
wider than the typical spreads of less than 1 percent
on the Federal National Mortgage Association issues, for
example. And the bonds of corporations with substan
tially smaller total indebtedness than MAC trade on the
New York and American Stock Exchanges at single
publicly quoted prices with no bid-ask spread. There is,
then, a substantial relative penalty associated with
selling even the most frequently traded municipal bond.
The illiquidity of municipal issues is, not only a
problem in and of itself, but in addition the thinness of
the secondary market for many outstanding municipal

FRBNY Quarterly Review/Winter 1983-84

bonds makes it prohibitively risky to agree to a contract
to deliver one of these bonds at some time in the future.
Without futures contracts, it is difficult for holders of
large municipal bond portfolios to hedge their positions
against market risk. An investor with a large municipal
portfolio could hedge against rises in general interest
rates by taking an appropriate position in Treasury bond
futures. However, exempt/taxable yield spreads fluc
tuate. The simple correlation between changes in the
yield on twenty-year Treasury bonds and in Moody’s
index of Aaa municipal bonds is 0.70.5 By comparison,
the correlation between changes in Treasury and in Aaa
corporate bond yields is 0.91. Therefore, the risk left
uncovered by a Treasury bond hedge against a position
in municipal bonds could be substantial.
The absence of futures trading in bonds may be a
substantial impediment to expansion of the market.
Dealers unable to cover the market risk of holdings
might be unwilling to maintain substantial inventories of
municipal bonds. Without inventories of outstanding
issues, the secondary market remains thin, reinforcing
the initial problem of illiquidity.
The desirability of some sort of hedge against adverse
fluctuations in the municipal-Treasury yield spread has
led to widespread active planning to initiate trading, not
in futures contracts for specific municipal bonds, but for
contracts based on a municipal bond index. It is likely
that trading in such a contract will commence shortly.
A rough estimate of the potential benefits to borrowers
associated with this innovation can be derived if we
assume that futures trading could make municipal bonds
as liquid as corporate bonds. Suppose further that,
given equivalent liquidity, municipal and corporate bonds
would be perfect substitutes in portfolios, except for tax
exemption. In that case, if a corporation were the mar
ginal municipal bond buyer, municipal bonds would yield
0.54 times the corporate bond rate. Over the last
decade the lowest actual yield ratio between long-term
municipal and corporate bonds was about 0.60. A
reduction of the ratio to 0.54 is equivalent to a 10 per
cent decrease in the exempt yield, the taxable yield held
constant.
There is reason to be skeptical, however, about some
of the potential benefits of this financial innovation. The
“technical” problems, making it difficult to decide on the
“ right” municipal bond index, may be more than merely
technical. There are many different participants in the
municipal bond market who might make use of a hedge,
but each group of participants is exposed to different
types of risk on different types of portfolios. A single
index may not be appropriate for all portfolios.

•Monthly average yields from January 1965 through October 1983.

Digitized for 20
FRASER
FRBNY Quarterly Review/Winter 1983-84

More flexibility for municipal finance officers
Private corporations have at their disposal a wide variety
of mechanisms for financing capital expansion and
replacement. Corporations may, as municipalities usually
do, issue long-term fixed-income debt instruments.
However, corporations may also issue preferred or
common equities, borrow directly from banks at home
and abroad, tailor the maturities of their debt to market
demand, finance projects temporarily through commer
cial paper markets, “ borrow” from their employees
through profit-sharing or stock option plans, and so on.
State and local governments have had a more limited
set of financial options; they usually finance long-term
obligations only by issuing long-term bonds. Given this
relative inability to tailor financial strategy to market
conditions, it would not be surprising if municipalities
missed opportunities to economize on financing costs.
In recent years some of the more sophisticated seg
ments of the municipal bond market began to design
new types of debt instruments to meet the requirements
of the market. Among the new mechanisms are put
option bonds, which can be “ put back” to the issuer at
various times, variable interest rate bonds, municipal
warrants, and tax-exempt commercial paper. Many of
these new instruments were designed to meet the
demand of tax-exempt money market funds for munic
ipal paper with short maturities.6
The incentive to design tax-exempt securities with
shorter effective maturities is strong. The municipal yield
curve has historically been positively sloped and steeper
than the Treasury yield curve (table). Over the past two
to three years, agencies that borrowed short or at
floating rates did better than those that borrowed long
or at fixed rates. During 1982, on average, for example,
the one-year yield on tax-exempt securities was only 68
percent of the twenty-year yield. Of course, short-term
borrowing to finance long-term obligations is risky. Given
the generally rising interest rates through the 1970s and
early 1980s, on average, it would not have paid
municipalities to finance long-term obligations by rolling
over short-term debt. For example, an AA-rated bor
rower could have issued twenty-year revenue bonds at
6.42 percent in 1979 but might have been tempted by
the 15 percent savings on the coupon yield associated
with a one-year maturity at that time. By 1982 that

•Capital markets, state and local governments, and the general public
have been wary of short-term municipal financing since New York
City's fiscal crisis of 1975. Indeed, New York City did issue a huge
volume of short-term instruments in the early 1970s. The basic
problem, however, was not the term structure of the city’s debt as
the fact that New York was financing current operations by
borrowing, with little or no plan or prospect for balancing its budget.
This is quite different from the evolving practice of financing part of
a capital improvement budget through short-term money markets.

borrower would have seen the short-term rate rise to
7.60 percent.
Still, there is some reason to suspect that there is an
endemic “ shortage” of short-term municipal paper. The
re la tiv e ly s te e p and a lw a ys p o s itiv e slop e of the
municipal yield curve is usually explained, with mixed
empirical success, by the strong demand of commercial
banks fo r ta x-e xe m p t, but re la tive ly liquid, assets.
However, another contributing factor may be the insti
tutional constraints that prevent municipal issuers from
providing the mix of m aturities the market would most
like to buy.
For the most part the innovations allowing shorter
borrowing were developed and exploited by nontraditional municipal borrowers: public authorities, mortgage
revenue authorities, and private firms borrowing through
industrial revenue bonds. State and local governments
borrowing for traditional purposes have been slower to
innovate. Im po rtan t im p ed im e nts to more crea tive
m unicipal financing are state laws limiting the use of
short-term financing of capital projects and the restric
tions on interest rates public borrowers may pay that
effectively preclude variable yield issues. It is easy to
understand why these m anifestations of risk aversion
were written into many state laws. There is, after all, a
substantial risk of rapidly rising interest costs to state
and local governments whenever any of these innova
tio n s are a d o p te d . Som e b a la n cin g of risks and
expected savings is necessary, but it is unlikely that the
optimal plan would include no variable rate borrowing
and no financing of capital projects through short-term
securities.
To date, most of the creativity in municipal finance has
focused on shorter maturities and floating interest rates.
There are other dim ensions of innovation that might be

Ratios of One-Year to Twenty-Year Yields on
Aaa General Obligation Municipal and
U.S. Treasury Securities

Year
1978
1979
1980
1981
1982

Aaa
m unicipal
securities

U.S.
Treasury
securities

.76
.89
75
.71
,68

91
105
.96
96
.86

Sources: Public S ecurity A ssociation, S tatistical Yearbook of
M unicipal Finance (various issues) and Federal Reserve Bulletin
(various issues).

p ro fita b ly explored. A few m unicipal issue rs have
experimented with small issue municipal bonds sold
directly to the public. In general, the “ entry fe e ” for
municipal bond purchasers is several thousand dollars,
w hether investors buy individual bonds or invest in
mutual funds or unit trusts. This large initial investment
excludes many potential investors from this market,
namely, those with high current incomes but relatively
small liquid portfolios. If m unicipalities could raise bor
rowed funds through instruments marketed, for example,
by commercial banks as no minimum deposit tax-exempt
passbook accounts, a potentially large new market for
these securities might open. As an alternative, small
denomination tax-exempt bonds could be sold directly
by m unicipalities to local residents through utility bills
or the property tax collection mechanism.
Another departure might allow m unicipalities to issue
something more like an “ equity” rather than the trad i
tional fixed-income security. For example, purchase of
a municipal “ equity” might entitle the investor to some
fixed percentage of the aggregate value of real property
in the municipality. From the municipality’s point of view,
such instruments might be attractive because they tie
debt service to the growth of the local tax base, that
is, to the m u n ic ip a lity ’s a b ility to pay. S p ecu lative
investors whose need for tax-exempt income is likely to
increase over time might generate a reasonable level
of demand for such instruments.
More uniform accounting, registration systems,
and legal standards
If municipal finance officers are to be allowed more
flexibility in instrument design than their private-sector
co un terp arts enjoy, then m unicipal accounting and
reporting practices should adhere to standards as strict
as, if not necessarily identical to, those the Securities
and Exchange Commission requires of private-sector
issuers of debt instruments. One of the clearest benefits
to New York City of its grueling experiences of the mid1970s was the adoption by the city gove rn m e nt of
generally accepted accounting practices (GAAP). New
York, though, remains one of a small, but growing,
number of governments whose accounts are certified to
have met this standard.
In addition, more uniform and efficient mechanisms for
registering municipal securities and transferring own
ership might reduce the adm inistrative cost of issuing
and servicing municipal debt. Federal law now requires
that the ownership of all newly issued municipal bonds
be registered. Registration adds to the adm inistrative
costs of issuers, especially if secondary market activity
expands. A number of proposals for such innovations
as pure book entry of municipal bonds are being actively
considered. If implemented, such proposals could

FRBNY Quarterly Review/W inter 1983-84

reduce administrative costs, risk of loss, and by facili
tating trading enhance the liquidity of many issues.
Finally, the default of the Washington Public Power
Supply System raises questions on the legal status of
a number of projects financed by municipal bonds. Part
of the problem lies in differences in relevant laws across
states, and it is likely that investors would feel more
confident if these laws had more national uniformity.
A better mix of revenue bonds and
general obligations
In recent decades the use of revenue bonds has
increased markedly, not just for what have been called
“ private” purposes, but also for such public purposes
as road and sewage system construction and renova
tion, and construction of higher education facilities.
Public purpose revenue bond financing has several
advantages over general obligation financing. From the
economist’s point of view, because revenue bond
financing is usually associated with user fees rather than
general taxation, there is an initial presumption of
superior efficiency. From the political leader’s point of
view, revenue bonds typically have the advantage of not
requiring voter or legislative approval of specific issues.
However, revenue bonds have one distinct disadvan
tage, i.e., investors consider them riskier than general
obligation bonds. The evidence is the spread between
the yields on the two types of issues, which averaged
about 6 percent of the general obligation yield over the
past ten years. In a sense, then, the market penalizes
the financing mechanism which, in many ways, is more
efficient.
One way of combining the advantages of revenue and
general obligation bonds would be to provide some sort
of general fund backing to revenue issues. Often, rev
enue bonds of a public agency are backed by the
“ moral obligation” of a legislature to meet any revenue
shortfall. Moral obligations, however, are of dubious
legal status.
One alternative to straight-out revenue bonds or moral
obligations is the so-called “ double barrel” security,
pledging the general obligation of the state government
to meet any revenue shortfall. Most states make such
a commitment very difficult. The purpose of restrictive
legislation is to prevent the state from becoming too
deeply indebted. However, one state with very strict
limitations on general obligation borrowing—New York,
which requires a voter referendum for each general
obligation bonding authorization—also has a very high
state and local debt per capita.7 The main effect of New
York’s strict general obligation limitation may have been
7See Appendix for a discussion of the recent history of bond
referenda in New York State.

Digitized for 22
FRASER
FRBNY Quarterly Review/Winter 1983-84

to increase the share of state debt in the form of rel
atively expensive revenue bond obligations.
Some consideration might be given, therefore, to a
relaxation of restrictions on general obligation borrowing.
One way to relax restrictions, without making general
obligation pledging too easy, might be to make it easier
for states to issue bonds with double barrel security.
Thus, for example, if a general obligation bond required
a referendum, then contingent general obligation
backing of a revenue bond might require only a vote of
the legislature.
State bond banks
Several states—Vermont, Maine, Alaska, and Puerto
Rico, among others—have established bond banks.
These financial intermediaries issue their own bonds
and distribute the proceeds to local governments for
capital projects. The banks’ bonds are backed by their
state government’s credit, usually either as a moral or
a general obligation.
Attaching the state’s name to a locality’s bond issue
allows small local governments to borrow at rates based
either on pooled risk or, if the bonds are in some sense
state obligations, at a yield appropriate to the state’s
credit rating. In addition, the state bank’s bonds are
likely to be more homogeneous and, therefore, probably
more liquid than a local government’s issues. A rough
indicator of the potential for savings associated with
substituting state for local credit is the difference
between the average net interest cost of new state
borrowing, which was 10.16 percent in 1982, and the
average net interest cost to all other borrowers of 11.09
percent in the same year: about a 9 percent difference.
Some Congressmen and Senators are attracted to the
state bond bank idea, as well. Several bills have been
introduced in the Congress—for example, the “ Public
Investment Incentive Act of 1983” (S.532) by Senators
Domenici, Bradley, Andrews, Gorton, and Randolph. The
bills authorize Federal appropriations to capitalize
infrastructure banks in the states. Initial Federal appro
priation, perhaps with required matching funds from the
states, would be allocated to infrastructure projects by
state authorities. Local “debt service” to the bank, which
might issue its own bonds to supplement its initial cap
italization, would replenish the initial Federal appropri
ation on a revolving basis.
The bond bank idea is not universally popular. Some
local leaders dislike the idea for the same reason state
leaders like it: it would reassign some of the power to
set infrastructure policy to the state from the local level.
A Federal secondary market maker
Another type of bank-like agency that might enhance the
marketability of municipal bonds would be a secondary

market maker in the municipal bond field. This would
work in a similar way to the Federal agencies that, in
effect, make secondary markets for home mortgages
(Fannie Mae, Ginnie Mae, Freddie Mac) or student loans
(Sallie Mae). A “ Muni Mae” for example, like Fannie
Mae, might issue its own securities and use the pro
ceeds to purchase certain types of municipal bonds, say,
bonds funding certain approved infrastructure purposes.
If this Muni Mae’s interest payments were taxable,
some annual appropriation would be necessary to make
up the difference between taxable and exempt yields.
Under these circumstances, the intervention of Muni
Mae would have some of the same effects as the TBO.
As with the TBO, the effect of Muni Mae would be to
remove some tax-exempt securities from the market,
replace them with taxable securities, and have the
Treasury pay a direct (or passed-through) subsidy to
qualified issuers. The difference would be that, under
the TBO, the Federal Government would play a passive
role in the secondary market. Whenever the exempt/
taxable yield spread was narrow, taxable municipal
bonds would be issued and the Treasury would begin
paying out the requisite subsidy. With a Muni Mae the
Federal Government could play an active role in influ
encing the exempt/taxable spread—and, therefore, the
relative cost of capital to municipal borrowers—by bid
ding a proportion of available municipal bonds away
from marginal purchasers. In addition, Muni Mae might
finesse some of the opposition to the TBO that exists
among municipal finance officers unwilling to concede
a Federal constitutional right to tax municipal interest
payments.
If the interest on Muni Mae were tax exempt, then
Muni Mae might run a surplus, given the higher risk
premium on municipal than on Federal Government
securities.
A Federal secondary market maker has at least one
important advantage over the state bond bank idea.
Local government authorities value their financial inde
pendence highly. Reliance on a state bond bank for
direct financing limits that independence of action. A
secondary market making agency would accomplish
many of the same objectives as the bond bank without
significantly changing the current balance of power
between state and local governments. One possible
disadvantage of this type of Federal intervention, how
ever, is the potential politicization of Muni Mae’s decision
on whether or not to purchase a specific municipality’s
debt instruments.
Some interactions among these changes
The potential effectiveness of each of these changes in
reducing the cost of capital for infrastructure purposes
depends on which combination of them are implemented

and their success. To illustrate these interactions, con
sider how nine of the other eleven8 changes would
affect the operation of a TBO.
Certainly, it is difficult to imagine the Treasury
Department supporting the passage of a TBO unless
some strict limit were placed on the issuance of private
purpose tax-exempt bonds. The TBO would increase the
benefits of tax exemption by insuring that the exempt/
taxable spread never narrowed to less than some pro
portional amount. Without some limitation this increased
subsidy would attract even more sophisticated private
purpose borrowers to the exempt market. That this
increasing, and even more direct, subsidy would be
more efficient than traditional tax exemption would be
small consolation to the Treasury.
One key design feature of the TBO is the subsidy
rate, i.e., the proportion of a municipality’s taxable
interest reimbursed by the Treasury. The “ right” subsidy
rate depends, in part, on what the yield spread would
be. But the yield spread, in turn, depends mostly on tax
law and regulatory policy. Thus, the design of a TBO
must be mindful of the likely evolution of tax and reg
ulatory policy.
Most analyses of the TBO are based on the
assumption that taxable municipal bonds would trade at
the same prices as corporate bonds of similar credit
rating. There are good reasons to suspect, however, that
AAA taxable municipal bonds would not be treated by
portfolio managers as a perfect substitute for the senior
obligations of AAA-rated corporations. Given the thin
ness of the secondary market and the fact that munic
ipal bonds are not backed by attachable collateral,
investors might demand a premium on taxable municipal
yields. Furthermore, portfolio managers, who are
accustomed to the relative uniformity and transparency
of corporate financial statements, might initially be put
off by the work it takes to understand the finances of
the typical municipality.
How well a taxable municipal bond does on the
market might depend on the outcome of the changes
discussed previously that could lead to greater uni
formity and greater liquidity. A portfolio manager might
be more receptive to bonds guaranteed by a well-known
corporate third party and to the bonds of municipalities
that issue debt frequently and are operating under
GAAP. Similarly, the taxable obligations of a wellcapitalized state bond bank might get a better reception
from institutional investors who are new to the municipal
market.
The ability to hedge a position in municipal bonds
through futures trading might also be a prerequisite
•Commercial bank underwriting and more aggressive marketing are
only distantly related to the TBO.

FRBNY Quarterly Review/Winter 1983-84 23

demanded by the managers of large pension funds. And
the flexibility of a municipal finance officer to design
obligations to meet investors’ specific requirements
might be even more important when dealing with insti
tutions whose primary interest is not tax avoidance.
Given the number and diversity of municipal issuers,
it might be, however, that all these changes would not
be enough. For example, there could be substantial
demand for the taxable bonds of larger issuers and little
or none for those of smaller issuers. Smaller munici
palities, or those with peculiar credit problems, there
fore, would be unable to realize the benefits of the TBO.
It might be that the only way for such municipalities to
issue taxable debt would be through the intermediation
of a state bond bank or a Federal secondary market
maker.
Conclusion
The improvement of municipal credit markets is a policy
making problem of considerable complexity. There are
at least a dozen different courses to follow which
interact in potentially important ways. Some of these
ongoing or potential changes fall under the purview of
the Federal Government. Others require state action,
and still others are or should be private-sector initiatives.
Regulatory agencies, trade organizations, rating agen
cies, and leagues of state and local governments all
have roles to play and axes to grind. Most of the
changes discussed here appear to be good ideas on
theoretical or rough empirical grounds. However, more
extensive policy analysis may indicate that some of
these proposals are neither cost beneficial nor practical.
Putting together a set of simultaneous initiatives with
closely related content at several levels of government
and in the private sector in a politically charged policy
arena would be a very complex and delicate under
taking. However, a more effective municipal capital
market might go a long way to help solve what many
agree to be a national problem approaching crisis pro
portions.
Very rough estimates of the most that could be saved
given universal implementation of some of these dozen
changes are possible.
• Tax and regulatory changes inducing the return
of corporate investors to a dominant role in the
market could reduce exempt yields by 18 per
cent, taxable yields held constant.

FRBNY Quarterly Review/Winter 1983-84
Digitized for24
FRASER

• If municipal bonds became as liquid as corpo
rate bonds, exempt yields might fall by 10 per
cent, taxable yields held constant.
• Eliminating half of all "private use” revenue
bonds might reduce exempt yields by 7 percent,
taxable yields held constant.
• A fully exercised TBO with a 31 percent subsidy
rate might reduce municipalities’ net interest
costs by 5 percent on average.
• Use of “ double barrel’’ security might save
revenue bond issuers about 6 percent of net
interest cost.
• State bond banks might save localities 9 percent
of net interest cost.
• Commercial bank underwriting might reduce
revenue bond yields by 6 percent, other yields
held constant.
• Finally, third-party guarantees could reduce debt
service expenditures by about 7 percent for
Baa-rated borrowers.
This array of maximum potentials suggests that a 20
to 25 percent savings of net interest cost is well within
the range of possibility. As the alternatives to municipal
credit reform—large increases in current taxation, an
even greater Federal deficit, or continued infrastructure
deterioration—are all unattractive, an attempt to design
and implement an integrated set of changes in the
municipal credit system is probably worthwhile.
One way of beginning this task would be to establish
a national commission including representatives of all
levels of government and all participants in the munic
ipal bond industry. The commission would have an
independent staff of sufficient qualifications and size to
analyze the relevant issues in depth. The task of the
commission would be to design a set of proposals
including actions to be taken by the Federal and state
governments, the private sector, and the relevant reg
ulatory agencies. Once a sound, well-balanced, and
practical set of proposals has been developed, the
commission’s job would shift to the more delicate task
of implementation.

Aaron S. Gurwitz

Appendix: Voter Approval of General Obligation Debt in New York State
New York State voters have traditionally approved few general
obligation bond referenda, and as a result New York leaders
have been reluctant to seek their approval. Since 1970, New
York authorities have asked approval for only eight bond pro
posals, yet voters rejected all but three, as shown in the
accompanying table. When state leaders have sought general
obligation financing, it has customarily been for projects that
were so large that the usually lower cost and larger size of
general obligation issues were essential. New York authorities
have requested authorization of such issues, ranging from $250
million to $3.5 billion. By comparison, the average tax-exempt
bond issue in the United States was $7 million in the 1970s.
Even with this large size, so few issues have been approved
that the general obligation debt of the State of New York
amounted to less than one fifth of New York State’s total out
standing long-term debt in 1983. Reliance on revenue bond
financing has been expensive. The average net interest
cost of New York State general obligation bonds sold in 1982
was about 10 percent. The average net interest cost for New
York statutory authority (revenue) bonds sold in the same year
was over 12 percent. Recently, however, voters appear to be
more willing to approve issues. Of the $2 billion which voters
have authorized in the past fourteen years, $1.8 billion was
approved in the last five years, and most of that in the past
four months.
Even when bonds have been authorized, the electoral support
has been generally limited (table). Out of sixty-two counties,
only two—the Bronx and New York (Manhattan)—have voted
in favor of all eight bond referenda. The bond issues approved
in the past five years won approval in no more than twentyone out of sixty-two counties and had statewide approval rates
of no more than 55 percent. In addition, in eighteen counties
the proposals have been defeated by an increasing number

of votes since 1979.*
The three general obligation bond proposals that voters have
agreed to finance have been very special, nonroutine capital
projects. The proposals in 1974 and 1979 were designed to
respond to the enormous rise in oil prices by increasing energy
efficiency through maintenance and improvement of trans
portation facilities. The 1983 proposal was designed to respond
to the severe deterioration of the state’s roads, bridges, and
tunnels. All three proposals were carefully designed to provide
benefits upstate as well as downstate in order to achieve state
wide political concensus. Even then, traditional upstate mistrust
was difficult to overcome as shown in the table by the small
number of upstate counties that approved the 1979 and 1983
proposals.
In sum, few bond proposals have provided the immediacy
and breadth of benefits that New York voters seem to require
for approval of a general obligation bond. Even then, the margin
of support was narrow and approval could not have been taken
for granted. As a result, general obligation financing has been
limited to projects with two characteristics. First, the proposed
projects are so extensive and expensive that the usually lower
cost and larger denominations of general obligation bonds have
been necessary for fiscal viability. Second, the need addressed
has been so important and immediate that a sufficient coalition
of interests could be assembled for voter approval.
'Counties in which the margin of defeat has expanded in the last five
years are Chautauqua, Columbia, Cortland, Delaware, Fulton,
Hamilton, Madison, Montgomery, Onondaga, Oswego, Otsego,
Rensselaer. Saratoga. Seneca. Tompkins, Warren, Washington, and
Yates. Counties supporting all three referenda are the Bronx,
Broome, Clinton, Kings, Nassau, New York. Queens, Richmond, and
Westchester. Counties whose support has recently reached a
majority are Monroe, Niagara, Putnam, Rockland, St. Lawrence,
Suffolk, and Ulster.

New York State Bond Referenda since 1971

Year
1971
1973
1974
1975
1977
1979
1981
1983

Project

Amount
(millions
of dollars)

Statewide
counties
approval
(percent)

Approving
counties
upstate
(out of 53)

Approving
counties
downstate
(out of 9)

Transportation
Transportation
Transportation
Housing
Economic development
Transportation
Prisons
Transportation

2,500
3.500
250
250
750
500
500
1,250

39
42
65
36
38
55
49
53

4
1
52
0
0
13
3
7

2
4
8
2
4
8
8
9

Allen J. Proctor

FRBNY Quarterly Review/W inter 1983-84

Commercial Bank Investment
in Municipal Securities

Historically commercial banks, together with casualty
insurance companies and individual households, have
been the major group of investors in tax-exempt
municipal bonds. Banks, however, are now playing a
much diminished role in the tax-exempt market. This
article examines the reasons for the change in bank
behavior.
The declining involvement of banks has taken place
at an inopportune time for state and local governments.
In 1982 and 1983, these governments issued debt at a
net rate of about $50 billion per year, more than twice
the average rate of the previous decade. Over the same
two years, banks invested at a net rate of less than $1
billion per year, about one tenth the rate of the previous
decade.
Although banks continue to participate in the munic
ipal market, their own holdings since 1971 have not
grown at the same pace as the municipal securities
market (Chart 1). Today banks hold one third of all
outstanding municipals, compared with over one half in
1971. Nor has their investment in municipals kept pace
with the growth of the rest of their investment and loan
portfolio (Chart 2).
Since 1981, banks have sharply reduced their
municipals purchases. Their net purchases dropped by
half in 1981 and remained low in 1982. They actually
sold more municipal securities than they bought in the
first three quarters of 1983. By early 1983, individual
households exceeded commercial banks as the largest
holders of municipals for the first time since 1964. And
the share of banks’ assets held in municipals also fell

FRBNY Quarterly Review/Winter 1983-84
Digitized for26
FRASER

to levels not seen since the early 1960s. If commercial
banks had instead maintained this share at 1971 levels,
they would have held $90 billion in additional municipal
bonds in 1983, over 150 percent of their actual holdings
at that time.
The decline in bank investment in municipal securities
has been broadly based. Even though small banks
generally hold proportionally more municipals than large
banks, both groups have reduced the share of domestic
assets held as municipals (Chart 3).
No single explanation accounts for banks’ diminished
role in the municipals market. Since 1979, and espe
cially since 1981, virtually every factor influencing bank
tax-exempt holdings has worked toward a decline in
bank investment in municipal bonds. Changes in tax
laws in 1981 and 1982 probably have had the largest
effects. But bank profitability, the level and volatility of
interest rates, and credit risk have also been important.
Such a highly adverse coincidence of effects is
unlikely to be repeated. And the precipitous slide in
bank demand for municipals probably will not continue.
But, if these effects are to be reversed and if banks are
to return to at least their pre-1981 role as investors in
municipals, some major changes in the financial envi
ronment or in Federal tax laws are needed. Short of
this, state and local governments can take some steps
to encourage bank investment. Most importantly, these
governments must understand the investment needs of
banks and become both more creative in designing and
more aggressive in marketing their securities specifically
to meet those needs.

The analytical framework
There are two basic determ inants of a bank’s decision
to hold municipals: the net aftertax yield it can earn from
a municipal and its desire for m unicipals at that yield
relative to other investments or loans. A simple diagram
will help organize the analysis around factors affecting
each determinant (Chart 4). By referring to the diagram
one can gain a clearer understanding of why these
factors have influenced bank investment behavior and
how they may have reinforced or offset each other in
recent years.
Of course, the supply and demand for municipals
ultim ately dictate their nominal yield. But interest rate
determ ination is not the primary concern of this article.
Accordingly, the view taken here is that of an individual
bank which observes the nominal yield available to it
and, on the basis of several other factors, decides what
its municipal holdings should be.

The effect of these factors is illustrated in the diagram
by two lines (Chart 4). Line Y represents the net aftertax
yield on municipals to a particular bank. Line D repre
sents the bank’s demand for tax-exem pt securities at a
given yield. A sim ilar diagram could be drawn to rep
resent the bank’s decisions with respect to any category
of loans or investment. But the decision to buy municipal
bonds takes on some special characteristics because
the net aftertax yield schedule that each individual bank
faces varies with the share of tax-exem pt bonds in its
total assets.
Net aftertax yield
The yield realized by a particular bank on municipal
bonds is influenced by, but is not identical to, the nom
inal coupon yield of the security. The reason is that a
municipal security is valuable largely because of its tax
implications. As a consequence, many factors other than

C hart 1

Share of State and Local Obligations Held by Various Groups of Investors
P erce ntag e sh a re o f m u n icip a ls outstanding at end o f q u a rte r
P e rc e n t

6 0 -------------------------------------------------------------------------------------------------------------------------- --------------------------------------- —

C o m m e rc ia l b a nks includ e U .S .-c h a rte re d banks, d o m e s tic a ffilia te s , Edge A c t and A g re e m e n t C o rp o ra tio n s, U.S. a g e n cie s and
b ra n ch e s of foreign banks, and b a n k s in U.S. possessions.
♦ in clu d e s open-end m utual fun ds. F ig ures fo r c lo s e d -e n d m utual fu n d s, in clu d in g unit tru s ts , sh o u ld a ls o be in this cate gory
but c a n n o t be s e p a ra te d from o th e r h o ld e rs .
S o u rc e :

B oard of G o v e rn o rs o f th e F e d e ra l R eserve S ystem , Flow o f F u n d s

FRBNY Quarterly Review/W inter 1983-84

C h a rt 2

C h a rt 3

S h a r e of N ew ly A c q u ir e d Bank Assets
H eld in th e form of M u n ic ipal B o n d s *

C o m m e rc ia l B ank H oldin gs of M u n ic ip a ls
as a S h a r e of T o ta l D o m e s tic Assets
By a s s e t size

P e rc e n t

P ercent

19521960

19611970

19711980

19811983-111

* A nnual a v e ra g e o f the ra tio of p u rc h a s e s , le s s s a le s
and re d e m p tio n s o f m a tu rin g ho ld in g s, to the ne t
in c re a s e in to ta l fin a n c ia l a s s e ts . F in a n c ia l a s s e ts
in c lu d e all ba nk a s s e ts e x c e p t c u rre n t s u rp lu s , p la n t
and e q u ip m e n t, and in te rb a n k p o s itio n s .
S o u rc e : B oard o f G o v e rn o rs o f the F e d e ra l R e se rve
S ystem , Flow o f F u n d s .

coupon yield come into play in determining the value of
a municipal as a tax shelter for an individual bank. Not
only is the income on the municipal security exempt
from Federal taxation (as it is for all investors), but also
appropriate use of municipal investments can shelter
from taxation bank profits on other operations. This tax
savings is an im portant component of the net aftertax
yield of a municipal security. The size of the tax savings
is influenced by three main factors: (1) the marginal
corporate income tax rate, (2) the bank’s interest car
rying costs, and (3) the degree to which carrying costs
are deductible from taxable profits.
The determinants of the net aftertax yield can best be
illustrated by a simplified example. Consider a bank with
$100 million of investm ents and loans which earn an
average taxable yield of 10.5 percent and are financed
by liabilities with an average cost of 9.5 percent. By the
year-end the bank will earn taxable profits of $1 million.
W ithout some “ shelter” the bank would have a tax lia
bility of approxim ately $460,000 based on the marginal
corporate tax rate of 46 percent (t). The bank could
eliminate this liability entirely if, at the beginning of the
year, it borrowed $10.5 m illion (M) at, say, a six-month
c e rtific a te of d ep osit (CD) rate of 9.5 percent and
invested the borrowed funds in municipal bonds paying

28 FRBNY Quarterly Review/W inter 1983-84

Source: B oard o f G o ve rn o rs of the F e d e ra l R e se rve
S ystem , R e p o rts of C ond ition and Incom e.

a tax-exempt yield of 9 percent (rex). The $1 million
carrying cost for these municipals (cM) is deductible
from taxable profits, reducing them to zero. Therefore
taxes too are reduced to zero.1 The total net aftertax
income from these municipal securities is the tax-exempt
earnings of about $950,000 less the carrying costs of
almost $1 million plus the tax savings of $460,000 for
a net yield of 3.9 percent (rex- c - i- tc ') .
In this example, $10.5 million is the most the bank
would invest in municipals. If the bank borrowed another
$1 m illion to buy m un icip als, the incom e w ould be
$90,000 in tax-exempt earnings less $95,000 in carrying
costs. The bank no longer has any income tax obliga
tions so that there is no tax savings from this additional
municipal investment. Thus, the net aftertax yield for
these additional m unicipals is negative. This maximum
level of bank municipal investm ent is denoted by the
drop-off, or “ kink”, in the net aftertax yield schedule
(Chart 4).
The point where this kink occurs can be expressed
in terms of the municipal-to-asset ratio at which taxable
profits are reduced to zero. In this case, the bank adds
1The Tax Equity and Fiscal R esponsibility A ct of 1982 (TEFRA) limited
this d e d u ctib ility to 85 percent of carrying costs.

municipals to its initial taxable investments of $100
million until its total assets reach $110.5 million for a
m unicipal-to-asset ratio of 9.5 percent. (A com plete
derivation of this relationship is illustrated in Chart 4.)
Although the net aftertax yield on municipals in the
absence of any tax savings is not necessarily negative,
the ratio at the kink is usually the maximum ratio a bank
is willing to maintain. The lower net aftertax yield on a
m unicipal security is almost always inferior to the cor
responding yield on taxables.
Any change in taxable profits, carrying costs, tax
savings, or nominal yields will alter the shape or position
of the net aftertax yield schedule. The direction of these
e ffe c ts can be d e m o n s tra te d by using the same
exam ple. The fundam ental factor is taxable profits.
These fall when income on taxable investments or loans
declines or when deductible expenses increase, such as
business operating expenses, the cost of borrowed
funds, loan loss provisions, and depreciation of physical
capital. When taxable profits decline, the yield schedule
shifts to the left so that the benefits of tax reduction
disappear at a lower m unicipal-to-asset ratio. In the
example, a decline in the level of taxable profits to
$500,000 would move the kink from a ratio of 9.5 per
cent to a ratio of 5 percent.
A decrease in the marginal corporate tax rate has the
effect of shifting downward the portion of the yield curve
to the left of the kink. An increase in the cost of bor
rowed funds has two effects. The entire yield curve
shifts downward because the net yield is lower. And the
kink shifts to the left because, with higher carrying
costs, a sm aller volume of municipals shelters all tax
able profits. In the example, costs of 10 percent instead
of 9.5 percent will shift the left portion of the yield line
to 3.6 percent from 3.9 percent, the right portion to a
negative 1 percent from negative 1A> percent, and the
location of the kink to 4.8 percent of assets from 9.5
percent.
Finally, when the nominal yield on municipals declines,
the net aftertax yield falls at all points and the point of
fu lly sh e lte re d p ro fits rem ains at the same ratio of
municipals to assets. In the illustration, a decline in the
coupon from 9 percent to 8.5 percent lowers the upper
and lower sections of the yield schedule by 1A> per
centage point.
Demand for m unicipals at a given yield
On average, a bank demands less than the volume of
m unicipals denoted by the kink point in its version of
Chart 4. Some banks may, in fact, choose to stay close
to the kink point, but this choice depends on where its
demand schedule (D) lies on Chart 4 relative to the net
a fte rta x yield line (Y). Its dem and schedule would
intersect the kink point if the bank aimed to pay no

C h a rt 4

N e t A f t e r t a x Y ie ld of a B a n k ’s
M u n ic ip a l H o ld in g s
P e rce n t

Y ie ld w hen
ta x a b le p ro fits
are p o s itiv e

B ank dem and fo r m u n ic ip a ls
"B ank m u n ic ip a l h o ld in g s
b e fo re 1981

Y ie ld in a b se n ce of
ta x a b le p r o fits

B ank m u n icip a l
h o ld in g s 1983

l__________

------------------Y-

M u n ic ip a l-to -a s s e t ra tio

The net aftertax yield on a tax-exempt security consists of
the nominal yield of the security (rex ) less
the interest carrying cost of financing the security (c) plus
the tax savings, i.e., the allowable deductible interest carrying cost (c')
multiplied by the marginal corporate income tax rate (t): r ex - c + tc'.
When the bank has no tax obligations, tc' = 0.
Taxable profits consist of
the yield (r) derived from the income on all taxable assets and operations
(A - M) less
the total interest carrying costs (c) of the bank's financial liabilities (L)
allocated to taxable assets (A - M) as a share of total assets (A) less
all other allowable expenses (s) of taxable assets (A - M) less
the total interest carrying costs of the bank s financial liabilities (cL) that
are permitted to be allocated (c'/c) to the share of a bank's assets (A)
held as tax-exempt securities (M):
r(A - M) - cL[(A - M)/A] - s(A - M) - cL(c'/c) (M/A).
In other words, since A = L, taxable profits are the net interest margin
adjusted for expenses (r - c - s) earned on taxable
assets (A - M) less
the deductible carrying cost (c') of tax-exempt securities (M):
(r - c - s) (A - M) - c'M.
If we set taxable profits equal to zero, we can solve for the municipalto-asset ratio at which the net aftertax yield drops to r ex - c:
M/A = (r - c - s)/c' + [1 + (r - c - s)/c'].

FRBNY Quarterly Review/W inter 1983-84

income taxes, if tax-exempt bonds were the only shelter,
if the bank had no foreign tax credits or loan loss pro
visions, and if except for their tax status municipal
bonds were perfect substitutes for other securities.
However, all these conditions are rarely met. Four fac
tors help determ ine the location of the demand
schedule:
• The availability and yield of alternative invest
ments, particularly tax-shelter investments;
• The bank’s liquidity requirements and prefer
ences and the liquidity of other assets relative
to municipal bonds;
• The risk of default, the risk of a downgraded
credit rating, and the bank’s attitude toward
these risks; and
• The size of the bank in terms of the investment
resources available.
The most important alternative to municipal bonds as
a tax shelter for bank profits is leasing. This entails the
purchase of a piece of equipment, building, or other
depreciable asset for lease to a third party. The bank
earns taxable income from the lease, but the purchase
of the physical asset entitles the bank to substantial
credits and deductions which reduce tax liabilities on
other operations. Leasing by banks was first permitted
in 1963 but did not become widespread until after 1970
when amendments to the Bank Holding Company Act
made large-scale leasing easier. Moreover, leasing is
often a highly leveraged investment by which the bank
can receive substantial tax benefits while committing
relatively few “ equity” funds (Appendix 1). As a result,
leasing can offer aftertax returns well above those on
municipals.
The bank’s need for liquidity is a second factor that
affects demand for municipal securities. Municipal bonds
are more liquid than some other assets, for example,
equipment for leasing. However, most tax exempts are
long term and the secondary market for municipals is
not nearly so well-developed as the market for some
other securities. Anything that increases a bank’s desire
for liquidity may decrease its demand for long-term
forms of tax shelter. Changes that make municipals
more liquid increase demand for tax exempts.
Third, holding municipal securities exposes the bank
to credit risk. There is some chance that a municipal
security could fall into default. There is an even greater
probability that downgrading by a credit agency will
reduce the market value of a bank’s holdings. Any per
ceptions of increased riskiness tend to reduce demand
at given net aftertax yields.
30FRASER
FRBNY Quarterly Review/Winter 1983-84
Digitized for

The size of a bank is also a factor in the level of bank
demand for municipal securities. Large banks have
access to more alternative investments and tax shelters,
such as large-scale leasing. Hence, a large bank’s need
for municipals as a source of income and tax shelter is
relatively less than that of a smaller bank. In fact, large
banks invest more in leasing and less in municipals than
small banks. By 1982, the largest 100 banks had
accumulated about 6 percent of their assets in municipal
securities whereas the small banks had accumulated
more than 10 percent in municipals (Chart 3).2
The geographic location of a bank is an additional
factor often cited as a source of varying demand.
Because of the large number and relatively small size
of most municipal bond issues, compared with corporate
or Treasury bond issues, there are fewer potential
investors for a typical municipal bond issue.3 In partic
ular, those investors are likely to be located in the same
state or locality. For that reason, many analysts char
acterize the municipal market as geographically seg
mented so that different demand curves exist for each
state. Preliminary investigation suggests that state-bystate differences in bank demand are not systematic
(Appendix 2). In other words, it appears that nationwide
factors common to all banks are the most important
influences on aggregate bank municipal holdings.
Trends over the past thirty years
Bank investment in municipal securities over the past
thirty years can be generally explained in terms of
changes in these factors. During the 1950s, bank
municipal holdings were constrained mostly by banks’
reliance on their securities portfolios for most of their
liquidity. This practice was incompatible with large

*This difference between large and small banks in the mix of tax
shelters is ironic because it does not coincide with the difference
between the proportional taxes they pay. Since leasing provides a
superior shelter from taxes, compared with municipal securities, and
since large banks do more leasing and less investment in
municipals, compared with small banks, one would expect large
banks to shelter proportionately more of their income from taxes.
This does not seem to be the case. Take the ratio of aftertax to
before-tax income as one indicator of proportional taxation. The
natural expectation is that large banks shelter more income from tax
and have a higher ratio. In fact, in 1982 the ratio was 75 percent for
the top 100 banks and 85 percent for the smaller banks. A lower
average U.S. tax rate for the smallest banks and a higher average
rate for multinational banks with operations in high-tax countries and
states may account for some of the difference. However, the rest
remains a puzzle.
3For example, during the 1970-78 period, there was an average of
7,845 new issues of municipal securities with an average value of
$7 million per issue. By contrast, corporate bond issues over the
same period averaged 493 new issues per year with an average
value per issue of over $50 million. (Robert Lamb and Stephen
Rappaport, Municipal Bonds: The Comprehensive Review of TaxExempt Securities and Public Finance (New York: McGraw-Hill Book
Company, 1980), page 8.

loldings of long-term municipal securities. As a con
sequence, bank investment in municipals was modest
relative to the municipals market and to the banks’ own
portfolios (Charts 1 and 2). By the end of 1960, banks
held 25 percent of all municipals, but municipals rep
resented only 8 percent of their financial assets.4
The growth of markets for Federal funds and large
CDs during the 1960s freed the banks from exclusive
reliance on their securities portfolios as a source of
liquidity.5 With the liquidity constraint relaxed, other
factors, such as relative yields and tax strategies, in
creased in importance as determinants of the municipalto-asset ratio. Bank municipal holdings surged over this
period to a peak of 15 percent of bank assets and 51
percent of all outstanding municipals by 1971, nearly
double the levels of a decade earlier.
During the 1960s, municipal bonds were essentially
the sole vehicle by which banks could shelter their
profits on domestic operations. The tax benefits of
investment in physical assets were available in principle,
but the use of leasing as a tax shelter did not begin in
earnest until 1971. Through the early 1970s, banks’
involvement in leasing increased rapidly and the
municipal-to-asset ratio declined.
This trend continued through the mid-1970s. By the
late 1970s, however, the growth of leasing activity by
commercial banks stopped, possibly owing to changes
in tax laws. The decline in bank participation in the
municipal market also slowed through the late 1970s
with a municipal-to-asset ratio of 11 percent and with
banks holding 43 percent of outstanding issues in 1978.
Most recent trends
Over the past four years, and particularly since 1981,
almost all factors affecting the appeal of municipals
4The discussion of the principal factors affecting bank municipal
holdings through the mid-1970s is based largely on Herman Kroos
and Martin Blyn, A History of Financial Intermediaries (New York:
Random House, 1971); Marcia Stigum, The Money Market, rev. ed.
(Homewood, IL: Dow-Jones-lrwin), 1983; John Petersen, "Changing
Conditions in the Market for State and Local Government Debt"
(U.S. Congress, Joint Economic Committee), April 16, 1976; and
Ralph Kimball, "Commercial Banks, Tax Avoidance, and the Market
for State and Local Debt Since 1970”, New England Quarterly
Review (Federal Reserve Bank of Boston, January/February 1977).
•Perhaps because the transition was so gradual from bank municipal
demand primarily determined by liquidity restrictions to demand
determined by a more diverse set of factors, empirical studies
based on the 1950s and 1960s often concluded that bank demand
for municipal securities was a residual. In other words, banks first
satisfied their demand for loans, Treasury securities, and other
investments. Whatever funds were left over were used to buy
municipals. An example of this residual approach to municipals is
Donald Hester and James Pierce, Bank Management and Portfolio
Behavior (New Haven: Yale University Press, 1975). Patric
Hendershott and Timothy Koch, "The Demand for Tax-Exempt
Securities by Financial Institutions’’, The Journal of Finance (June
1980) provide an example of a later rejection of this approach once
data for the 1970s became available.

discouraged bank investment.6 The decline in the
municipal-to-asset ratio continued in 1979, paused in
1980, and then accelerated. Banks, increasing the ratio
by 1 percent in 1980, reduced it by almost 8 percent
over 1981 and 1982. In the first three quarters of 1983,
the share of municipals in bank financial assets fell
another 7 percent.
Declining net aftertax yield
Even though the average nominal yield on municipals
increased substantially after 1979, increases in bank
costs and less favorable tax treatment of net income
from municipals made the net aftertax yield on munic
ipals much less attractive to banks.7
Banks’ need to shield profits from taxes has declined
in the 1980s.8 in 1981, bank pretax profits were flat after
a decade of virtually continual growth. In the following
year, profits declined by about $1 billion. Even though
strictly comparable figures are not yet available, bank
profits in general do not seem to have increased very
much in 1983. In particular, it appears that gains some
banks achieved in 1983 in the net yield of their loans
and investments were at times offset by increases in
loan loss provisions.
Changes in Federal tax laws in the past four years
have also had profound negative effects on the net
aftertax yield of bank-held municipal securities.9 The
changes began in 1979 when the maximum corporate
income tax rate was reduced from 48 percent to 46
percent. This change lowered the tax shelter value of
a municipal bond.
The greatest change was due to the 1982 tax act
(TEFRA). That legislation disallowed part of the interest
deduction for municipal carrying costs. Disallowance
reduces the value of a municipal as a shelter against
•In terms of Chart 4, the net aftertax yield line has moved down and
to the left (line Y,) and banks reduced their holdings from point 1 to
point 2. In addition, bank holdings were further reduced to point 3
by a decline in bank demand for municipals at that yield (line D,).
^ h e factors that Marcelle Arak and Kenneth Guentner, “ The Market
for Tax-Exempt Issues: Why Are the Yields So High?”, National Tax
Journal (June 1983) argue had contributed to this increase in
nominal yields are the same factors that made municipal securities
less attractive to banks. This may account, in part, for the sub
stantial increase in municipal investment by individuals, whose net
aftertax yield from municipals may have increased over this period.
•Because banks must make their tax-shelter investment decisions well
before actual taxable profits are known, they base their decisions on
anticipated taxable profits. Unfortunately for the analyst, anticipated
taxable profits, the most relevant variable, is not directly observable.
But some inferences can be made by looking at banks' total actual
profits before they paid taxes, as reported in the Board of
Governors of the Federal Reserve System publication, Reports of

Condition and Income.

•While state taxes may be important to the return that a bank
receives on a municipal security, preliminary investigation suggests
that state taxes do not play an important role on average in
explaining the share of assets which a bank holds as municipals
(Appendix 2).

FRBNY Quarterly Review/Winter 1983-84 31

taxes on other bank operations. Specifically, as of Jan
uary 1, 1983 banks have been able to deduct from
taxable profits only 85 percent of the interest costs
incurred to purchase municipals. Municipals held as of
the end of 1982, however, still benefit from the preTEFRA full deductibility. Thus, while the net aftertax
yield of bank holdings of municipals as of 1982 was not
affected, the yield of purchases in 1983 was reduced,
contributing to the absence of net bank purchases over
the first three quarters of 1983.10 For example, because
of the disallowance, the yield in 1983 for purchases of
municipals paying the recent market return and being
financed by six-month CDs was almost 20 percent lower
than it would have been with full deductibility of carrying
costs.
The net yield a bank can earn by investing in munic
ipals has also been eroded in the last four years by a
substantial increase in banks’ interest cost of funds. An
annual survey by the Federal Reserve shows that, in
general, banks’ average cost of money rose from
interest payments of less than 5 percent in 1979 to
almost 8 percent in 1981 and 1982. Part of the reason
for the increased costs was the general rise in interest
rates after 1979. An additional factor in banks’ cost of
funds in recent years, and for years to come, has been
the deregulation of the interest rates banks pay on
deposits. In 1981, interest-bearing transaction deposits
became widely available for the first time at commercial
banks. In 1983, deposits paying market rates of interest
were permitted in the form of transaction deposits
(Super NOW accounts) and time and savings deposits
(money market deposit accounts and CDs).
The increasing interest costs reduced the net aftertax
yield a bank can earn from investing in municipal
securities in two ways. First, higher costs contributed
to the reduction of taxable profits discussed earlier.
Second, they reduced the spread a bank could earn by
borrowing money to invest in tax-exempt securities.
Even though the bank receives back some of the higher
costs in the form of lower taxes through interest
deductibility, it must still absorb more than half of the
increase (1 -0.46, the marginal tax rate).
For example, small banks, who acquire funds primarily
through time deposits, saw their net aftertax yield on
municipals decrease over 2 percentage points from

10This lowers the left portion of the yield curve in Chart 4. In addition,
if part of its interest deduction is disallowed, the bank would have to
increase its holdings of municipals in order to create enough
deductions to shelter all its taxable profits, and the yield curve in
Chart 4 will shift to the right. If most banks were at the kink point in
Chart 4, then the effect of TEFRA could have been to increase
municipal holdings. Because most banks do not hold enough
municipals to cover fully their taxable profits, the yield effect of
TEFRA probably dominates.

FRBNY Quarterly Review/Winter 1983-84
Digitized for32
FRASER

1979 to 1982. Large banks, who finance many of their
assets through purchasing funds in the money market,
saw their net aftertax yield on municipals decline by 1
to 2 percentage points.11 Because funds available to
banks will increasingly require market rates of interest
in future years, the erosion of the spread a bank can
earn by investing in municipal bonds may continue.
Reduced bank demand at a given yield
Large and small banks have been attracted away from
municipals by several factors aside from the decline in
yield, among them the availability of tax-sheltered
leasing. Tax law allows a high degree of leveraging of
investment in physical assets so that banks receive the
tax benefits associated with a $5 investment with only
a $1 “ equity” stake in the capital asset (Appendix 1).
This magnifies the effect of accelerated depreciation and
investment tax credits on the net aftertax yield from
purchasing an asset to lease.
Tax legislation in 1981 (ERTA) liberalized the lever
aging requirements and increased the rate of cost
recovery through depreciation. Some of the changes are
subtle, and it is difficult to calculate their effect on
average returns. But it is likely that the provisions in
ERTA made leasing more attractive to banks and con
tributed to the drop in bank demand for municipals.
In fact, there is evidence that primarily large banks
responded to these changes quickly by increasing their
leasing activity substantially. In 1982, the share of total
operating income provided by leasing operations at the
largest banks increased by over 10 percent (Chart 5).
What makes the increase so impressive is that this
measure probably understates the increase in bank
leasing activity. Taxable lease income is a small part of
the net aftertax yield from leasing, and it usually does
not become sizable until at least a year after the lease
arrangement begins. The largest and most immediate
benefits from leasing are the tax credits and deprecia
tion deductions which are not reflected in this
measure.12
On the other hand, the decline in small banks’
demand for municipals was probably not influenced by
this change in ERTA. Small banks generally do not have
the resources necessary to overcome some of the dis
advantages of leasing. First, small banks often may not

"Small banks absorbed 54 percent of the increase in time deposit
costs from 6.4 percent in 1979 to 10.2 percent in 1982 as reported
in the Federal Reserve Board of Governors publication Functional
Cost Analysis. The cost of nondeposit funds rose from 6.3 percent
to 9.2 percent for medium banks and from 8.2 percent to 10.2
percent for large banks between 1979 and 1982.
12Other attempts to measure the extent of bank leasing activity can be
found in Ralph Kimball, op. cit.

C hart 5

In c o m e from L ea s e F in a n c in g as a
S h a r e of T o ta l O p e r a t in g In c o m e
By a s s e t s iz e
P e rce n t
1.4------------------------------------------------------------------------------

2------------------------------------------------------------

0I____ I_____ l_____I____ I____ l_____l____
1976

1977

1978

1979

1980

1981

1982

S ource: B oard of G overnors o f the F ederal R eserve
S yste m , R e p o rts o f C o n d itio n and In c o m e .

have large and diverse enough portfolios to absorb the
greater risk and lower liquidity of leasing over highgrade m unicipal bonds. Second, leasing requires a
specialized staff, which small banks generally cannot
afford, and it is most efficiently done in volume. As a
result, small banks have engaged in relatively little
leasing, and income from leasing has remained stable
at about 0.3 percent of their total operating income.
A second factor that reduced both large and small
banks’ demand for m unicipals at a given yield was the
rapid and unpredictable change in interest rates from
1979 to 1982. Holders of substantial volumes of long
term fixed-interest bonds saw the value of their port
folios fluctuate substantially. In this environment, banks
became wary of investing in long-term fixed-interest
securities, which constitute the majority of municipal
bonds, and sought to reduce their interest risk exposure.
One way to do this was to shorten the average maturity
of their municipal portfolios and thereby reduce their
demand for a large proportion of municipals.
The adverse impact of the increased appeal of leasing
and shorter term securities may have been limited to
some extent by the increasing availability of a special
form of tax-exem pt security known as an industrial
d e v e lo p m e n t bond (ID B ). Is s u e rs of ID B s have

pioneered the introduction of floating interest rates and
m edium -term m atu ritie s to enhance th e ir appeal to
spread- and liquidity-conscious investors.
Moreover, one type of the IDB— the small-issue IDB—
often carries a special attractiveness to banks, not
usually associated with trad ition al m unicipal bonds.
Small-issue IDBs have this appeal to small banks in
particular because, in many ways, a sm all-issue IDB is
merely a local business loan that is structured as a
bond to achieve tax exemption for the interest earned.
Small banks are used to making commercial loans, so
that they may feel particularly comfortable with smallissue IDBs. As with loans, the terms are negotiated, the
bond is held to maturity, and the bank often receives
compensatory balances from the borrower.
The desirability of these characteristics to all inves
tors, and presumably to banks as well, is suggested by
the increasing popularity of all IDBs. Although they
presently account for only 18 percent of outstanding taxexempt bonds, IDBs have grown from less than 1 per
cent of the net increase in long-term municipal bonds
in 1971 to more than one third since 1979, according
to the Federal Reserve Board’s Flow of Funds.
A final factor that has contributed to declining bank
demand for municipals is the increase in credit risk
associated with holding municipal securities in the last
few years. Unlike Federal Government securities, there
can be in te rru p tio n s in the paym ent of p rin cip a l or
interest on municipal securities. Although the probability
of default may be small, bank perceptions of the riski
ness of municipals may have increased. As a response,
banks may have kept unchanged or even lowered their
portfolio exposure limits for municipals.
Recently, many investors, including banks, have seen
the value of some of their m unicipals decline because
of default or a downgrading of their credit rating. The
most famous default recently, and the largest tax-exempt
default ever, has been in the municipal bonds supporting
WPPSS Projects 4 and 5. Unfortunately, there are few
data available to support or to refute the widespread
notion that credit risk is greater now than five years
earlier. Two indicators of increased risk are that Moody’s
Investors Service has reduced the number of investment
grade-rated bonds and that more issues have had their
ratings reduced than increased by M oody’s for the last
five years.13
A bank may try to minimize losses due to higher credit
risk by shifting its portfolio to higher quality municipals,
or it may choose to reduce municipal exposure and,
consequently, to reduce its demand for tax exempts. The
first option has become increasingly difficult because of
13Robert Lamb and Stephen Rappaport, op. cit., pages 70-71, and
M oody’s Investors Service.

FRBNY Quarterly Review/W inter 1983-84

the greater scarcity and expense of the top-grade
municipal bonds. In 1978, eighteen states issued
securities rated Aaa. By 1983, only twelve states did.
Moreover, top-rated municipal securities have become
increasingly expensive relative to riskier, minimum
investment grade municipals. A bank had to forego a
yield 14 percent higher if it wanted Aaa-rated municipals
rather than Baa-rated municipals at the end of the
1970s. By 1982, this loss in yield increased to 15 per
cent of the Aaa rate and reached 17 percent in 1983.
Faced with these difficulties in upgrading their portfolios,
many banks may have opted in part to reduce their risk
exposure and, consequently, their demand for municipal
bonds.
in addition, demand for municipals may also have
declined because some banks are choosing to take a
portion of their municipal exposure in the form of letters
of credit rather than ownership of municipal securities.
As part of an effort to improve the creditworthiness of
their debt, some issuers are asking for and receiving
irrevocable bank letters of credit as a form of insurance
that interest or principal will be paid to investors. These
letters of credit earn banks a fee, but they do not tie
up funds as investment in a municipal security would.
However, banks which consider these letters of credit
to be municipal exposure may reduce the amount of
municipals they are willing to own. There are some
estimates that, as of mid-1983, as much as $40 billion
in letters of credit had been written by the largest banks
as backing for municipal bonds. A portion of this may
have displaced bank demand for municipals.
In summary, since 1981 virtually all factors have
worked against bank demand for municipal bonds. The
net aftertax yield of a bank-held bond has been reduced
by a decline in bank profits needing shelter, a decline
in the tax rate and the deductibility of municipal carrying
costs, and a rise in the cost of financing investments
in municipals. This was offset, but not entirely, by an
increase in nominal yields. In addition, bank demand for
municipals at that yield was reduced by an increase in
the attractiveness of tax-sheltered leasing as well as
increases in the interest risk and credit risk of holding
traditional long-term fixed-interest municipals. However,
demand may not have declined for some industrial
development bonds whose similarity to loans, shorter
maturity, or interest flexibility could limit some of these
effects.

so many separate factors combining adversely again is
not high. And some factors may actually have begun to
move in more favorable directions.
• First, among the limitations to bank profit mar
gins were increasing costs of funds fostered by
deregulation of depository interest rates and
loan loss provisions occasioned by changes in
world economic conditions. Both of these should
be one-time adjustment costs that probably will
not recur for some time.
• Further decreases in income tax rates or
enhancements in the appeal of tax-sheltered
leasing are improbable. If either are changed,
it is more likely to be in a direction that will
raise tax liabilities and enhance the appeal of
municipal securities. In contrast, the erosion of
the tax-shelter benefits of municipal bonds could
continue as part of Federal efforts to increase
tax revenues. Proposals to restrict certain types
of revenue bonds have recently been intro
duced, or possibly the 85 percent limit on car
rying cost deductibility could be further lowered.
• Although one cannot predict changes in the
credit risk of municipal issuers, there has been
a noticeable expansion of techniques to reduce
this risk for investors, primarily through insur
ance and other innovative forms of payment
guarantees.
Despite the low probability of a continued decline in
commercial bank participation in the municipal market,
a reversal of the effects of the post-1981 period would
require substantial changes in the financial, regulatory,
and Federal tax environment. The specific adjustments
have yet to be discovered, yet they will have to
encompass some combination of the following condi
tions:
• A financial environment in which long-term fixedincome securities become much more attractive
than they are now;
• A wider spread between municipal yields and
banks’ cost of funds;
• A higher level of bank profitability;

Next several years
A continued sharp decline in commercial bank invest
ment in municipal securities is not likely. Virtually all the
factors affecting bank demand contributed to the post1981 drop in bank participation, but the probability of

FRBNY Quarterly Review/Winter 1983-84
Digitized for34
FRASER

• An increase in marginal corporate tax rates;
• Tax law changes that reduce the attractiveness
of leasing while increasing the attractiveness of
municipal bonds.

It is possible, but unlikely, that a sufficient combination
of these changes will occur in the foreseeable future.14
14A recent IRS opinion a c q uiescing to an earlier Tax Court decision
granted favorable tax treatm ent for repurchase agreem ents (RPs)
backed by m unicipal bonds. It remains to be seen, however, whether
this opinion will create a new source of dem and for tax-exem pt
securities.

Nonetheless, state and local governments have many
options available to enhance the m arketability of their
bonds to banks. Taking a cue from the appeal of
industrial development bonds, they might reduce the
maturities and increase the flexibility of interest rates of
more traditional municipal bonds to enhance their appeal
to banks.

Allen J. Proctor and Kathleene K. Donahoo

Appendix 1: The Returns on Leasing and Municipal Bonds Compared
Under current law, equipment leasing dominates munic
ipal bonds, in terms of rate of return, as an investment
for a profitable bank. To understand this, consider a bank
with taxable income in need of sheltering. The bank has
borrowed $1 and is deciding whether to buy a municipal
bond with the dollar or to invest in a dollar’s worth of
equipment to lease. The bank will choose the investment
with the greatest net present value (NPV) of aftertax
returns.
In both cases the interest cost of the borrowed dollar
will be written off for tax purposes against current oper
ating income. For the bond, however, only a proportion
— (1 -a), 0
a
1 — can be deducted. Under current
law a = 0.15.
For a municipal bond with the face value of $1 pur
chased at par, the NPV is:
Equation 1

NPVp

(1 + rc)'
(1 + rc)T
t= 1
Coupon yield on the exempt bond,
Cost of funds to the bank,
r
ix = Marginal corporate tax rate, and
T - Term of maturity of the bond in years.

where: r,

For a lease of a durable good with a useful life of T
years and no scrap value at the end of that period, the
NPV is roughly:
Equation 2
T
NPVL

2
t= 1

rtx(1 ~ ^) + M-<~c
(1 + rc)*

P
+ 5(x

5(t)

+ 5k
1
t = i (l+ o *

where: rtx = Rental income,
8(t) = Proportion of the value of the asset
allowed as a depreciation deduction t
years after the investment,
k
= Proportion of the value of the asset
allowed as an investment tax credit, and
P
= Period over which depreciation may be
taken.
The benefits of accelerated depreciation and the
investment tax credit are multiplied by five because
current minimum “ at risk” provisions require an invest
ment of only 20 percent of the cost of the asset. The
remainder can be borrowed. Current “ at risk” require
ments, then, mean that a bank can issue a CD for $1,
borrow another $4 from an institution with little or no tax
liability (e.g., a life insurance company or a local gov
ernment agency), purchase a $5 investment, and claim
the full tax benefits associated with that investment.
Given these considerations, under what circumstances
would a bank choose to purchase a municipal bond
rather than enter an equipment leasing arrangement?
The bank would be indifferent if equation 3 is:
NPVB = NPVl
Data for computations of rex are based on equations (1),
(2), and (3) and on the 1982 average values of the fol
lowing variables.
rc = Six-month rate on large CDs in the sec
ondary market (12.57 percent),
rtx = Yield on Baa corporate bonds (16.11 per
cent),
T = Ten years.
P = Five years under the accelerated cost
recovery (ACR) provisions of ERTA.

FRBNY Quarterly Review/W inter 1983-84

Appendix 1: The Returns on Leasing and Municipal Bonds Compared (continued)
Under these assumptions and with these variable values, the
yield on municipal bonds would have to equal 42.25 percent
to equal the rate of return on leasing. This is a little less than
four times the average annual yield of 12.48 percent on Aaarated tax-exempt bonds in 1982. Of course, comparison of
rates of return does not reveal the whole story. Leasing
arrangements are probably riskier and less liquid investments
than municipal bonds. However, the computation indicates that
there are attractive alternatives to tax-exempt securities for
banks wishing to shelter operating profits.

A series of experimental calculations with alternative variables
and parameters indicated that the key provisions of current
tax law affecting the attractiveness of leasing are the minimum
at-risk investment requirements. If the tax benefits of a leasing
arrangement were limited to the bank’s direct "equity” investment
in the equipment, instead of up to five times that investment,
then the rex required to match the benefits of leasing would be
only 11.70. Our calculations suggest, therefore, that modifi
cations of the at-risk requirements for investment tax benefits
can have an important effect on banks’ demand for municipal
bonds.
Aaron S. Gurwitz

Appendix 2: Some Characteristics of Municipal Bond Market Segmentation
The municipal bond market has been characterized as
geographically segmented. There are two reasons to
expect that banks’ municipal-to-asset ratio will differ
systematically across states.* First, the municipal bond
market consists of a large number of relatively small
issues, many of which are sold on a negotiated rather
than on a competitive basis. A close relationship fre
quently develops between a local government and the
banks or other institutions buying its debt. To the extent
that accommodating state and local government bor
rowing requirements is a major determinant of bank
municipal holdings, one would expect the municipal-toasset ratio to be higher in the states that are the largest
borrowers.
Secondly, bank income from in-state municipals is
exempt from taxation in some states but is subject to
taxation in other states. Banks in states which exempt
bank income from in-state municipals would likely have
higher proportional municipal holdings.
Analysis of sixteen states categorized by total state
and local debt and tax treatment of banks’ municipal
income does not indicate a clear relationship between
'The m andatory p le d ging of m unicipal securities as collateral
for state and local go vernm ent deposits in banks was
form erly an im portant reason for state-by-state differences in
bank dem and for m unicipals but not after 1978. For a
discussion of other factors that may cause dem and for
m unicipals to vary by state, see Robert Lamb and Stephen
R appaport. M u nicipal Bonds: The Com prehensive Review of
Tax-Exempt Securites and Public Finance.

36FRASER
FRBNY Q uarterly Review/W inter 1983-84
Digitized for

Table 1

1982 Municipal Holdings of U.S. Commercial Banks
by State
M unicipal ho ld ings as a pe rcentage of total dom estic assets
High d e b t states*

Percent

Low debt states*

Percent

Taxable m unicipal in te re s tf
New York ...........
4.9
C alifornia .........
3.2
9.9
Pennsylvania
Florida ............... ........ 10.3
Average

7.1

M ontana ....................... 11.0
New M exico ............... 10.3
South Dakota .............
6.6
North Dakota ............. 10.7
Average

...............

9.6

Tax-exempt m unicipal in te re s tf
New Jersey
M ichigan .........
Oregon .............
Virginia .............
Average

9.4
8.9
9.6
........ 10 8
9.7

Vermont
................... 9.3
Maine ........................... 11.7
New Ham pshire
8.1
Utah ............................... 7.4
Average

...............

9.1

'H igh debt states had ou tstand ing total de bt in 1980 greater
than $5.5 billion. Low debt states had outstanding total debt
in 1980 less than $2.0 billion.
fTaxable and tax-exem pt refer to sta te s’ treatm ent of bank
m unicipal incom e from in-state m unicipals as of 1983.

Appendix 2: Some Characteristics of Municipal Bond Market Segmentation (continued)

Table 2

1982 Municipal Holdings of U.S. Commercial Banks
by State and Asset Size
M unicipal holdings as a percentage of total dom estic assets
C ategory

Percent

All banks ................................................................................ ... 8.3
‘ Top 100 .................................................................................... . . . 5.6
10.4
Other ........................................................................................
All banks (e xclu ding New York and C alifornia) ........... . . . 9.9
•Top 100 .................................................................................... . . . 8.0
10.7
Other ........................................................................................
All New York banks ............................................................. . . .
'Top 100 .................................................................................... . . .
Other ........................................................................................

4.9
4.4
8.5

All C alifornia banks ............................................................. . . .
*Top 100 .................................................................................... . . .
Other ........................................................................................

3.2
2.6
5.9

'B anks among the nation's 100 larg est in asset size as of 1982.
Sources for Tables 1 and 2: Board of Governors of the Federal
Reserve System, Reports of C ondition and Income.

these two factors and banks’ municipal-to-asset ratios.
In general, the differences between municipal ratios
within categories are at least as great as those across
categories.
The average municipal-to-asset ratios for three of the
four categories are extremely close: 9.7 percent for tax-

exempt and high debt states, 9.6 percent for nonexempt
and low debt states, and 9.1 percent for tax-exempt and
low debt states (Table 1). The nonexempt and high debt
states’ average ratio of 7.1 percent is somewhat smaller,
but much of this difference" is due to the ratios in New
York and California.
The New York and California ratios are much lower
than those of any other nonexempt state studied here.
It therefore seems likely that factors other than state tax
treatment of banks’ municipal income must account for
the relatively low municipal ratios of banks in those two
states.
One explanation of the low municipal-to-asset ratios in
New York and California is that these states contain a
disproportionate number of very large banks, and that
large banks tend to hold a smaller proportion of their
assets in the form of municipals (Table 2). Comparison
of municipal holdings of large and small California and
New York banks with similar banks in the other 48 states
reveals that disaggregation by size lessens the difference
between the New York and California ratios and those
of the rest of the country. Thus, some of the difference
between the ratios of New York and California banks and
those of all other U.S. banks is due to a size effect—
California and New York have a high concentration of
very large banks. However, there remains some residual
state effect.
In sum, regardless of the greater availability of local
municipal issues and greater tax incentives to hold
municipal bonds, banks in most of the states studied
hold remarkably similar proportions of their assets in
municipals. Geographic segmentation of the municipal
market may exist in some form, but it does not seem to
affect the proportion of their assets which banks hold as
municipals.
Allen J. Proctor and Kathleene K. Donahoo

FRBNY Quarterly Review/W inter 1983-84

Neighborhood Changes in New
York City during the 1970s
Are the “Gentry” Returning?
Since the late 1970s, a number of journalists and
scholars have been calling attention to an emerging
“ back-to-the-city” movement of high-income households.
Observers note that, after two decades of suburbani
zation, high-income households have begun to redis
cover the central city and have been buying up property
in or near low-income neighborhoods for renovation and
owner occupancy. This process is sometimes called the
“ regentrification” of the inner city, with the modern
“ gentry” characterized as young, upwardly mobile
executives.
Until now, the evidence of these changes had been
largely anecdotal. For instance, a 1979 New York Times
Magazine article heralding the “ urban renaissance” and
the “ new elite” who had been rediscovering the city
reported on the perceptions and observations of the
Times author and others.1 Even academic journals are
often short on numbers. One recent paper described
gentrification in Washington, D.C., only qualitatively, with
assertions such as “ downtown residential areas [are]
increasingly populated by rich professionals who walk
or bicycle to work”.2
Regentrification is a controversial subject. Community
groups have strongly opposed the process, citing
neighborhood disruption and displacement of the poor
’ Blake Fleetwood, "The New Elite and an Urban Renaissance” ,
New York Times Magazine (January 14, 1979).
2Stephen F LeRoy and Jon Sonstelie, "Paradise Lost and Regained:
Transportation Innovation, Income, and Residential Location” ,
Journal of Urban Economics, 13 (1983).

38FRASER
FRBNY Quarterly Review/Winter 1983-84
Digitized for

and of nonwhite households. In contrast, regentrification
has its adherents, notably government officials unwilling
to discourage economic investment and property owners
hoping for capital gains.
The analysis in this article does not—and is not
intended to—support either side of the dispute. The
social costs and benefits of regentrification in New York
City have been well discussed in many forums. To add
some quantitative evidence to the public discussion, this
article presents a detailed analysis of the extent of
regentrification over the decade of the 1970s.
Last year, 1980 census tract data for New York City
were released, making this article possible. They permit
an analysis of demographic shifts since the 1970
Census both citywide and at a geographically fine level
of detail. In general, this analysis shows some specific
instances of regentrification, but not of sufficient mag
nitude to offset the aggregate trend of continuing out
migration. Specifically,
• Citywide, New York City has not increased its
population share of the metropolitan area’s highincome households, college graduates, or any
other high-status group.
• Census tract data for several individual neigh
borhoods which have been the focus of public
discussions, however, provide some quantitative
support for the existence of gentrification. Even
so, changes in neighborhood income distribu
tions were modest.

C h a rt 1

C h a rt 2

P o p u la tio n of N ew Y ork C ity

H ig h -in c o m e F a m ilie s

1 9 6 0 ,1 9 7 0 -8 1
T h o u sa n d s o f p e rs o n s

P e rc e n ta g e cha nge
P e rc e n t

8,000

7 ,5 0 0

7 ,0 0 0

6 ,5 0 0
New Y ork
C ity

6,000
S o u rc e : U.S. D e p a rtm e n t o f C om m erce , B ure au of
E cono m ic A n a ly s is , R e g io n a l E c o n o m ic
In fo rm a tio n S y s te m s .

• Citywide, there were many areas with especially
large increases in the share of the adult pop
ula tion w ith at least fo u r years of college.
N e ig h b o rh o o d s m ost co m m on ly lab ele d as
gentrifying had some, but by no means all, of
th e la rg e s t g a in s . T h e s e in c re a s e s w ere
accom panied by corresponding occupational
shifts.
• Changes in neighborhood income distributions
were generally much smaller than the increases
in educational attainment. Some of the largest
incom e sh ifts, m oreover, occurred in areas
where regentrification has not been widely dis
cussed.
• Structural and ownership characteristics of the
housing sto ck did not change dram atically.
Rents did increase substantially faster than the
inflation rate in many areas, though. Neighbor
hoods commonly regarded as gentrifying had
some, but not all, of the largest increases.

Citywide statistics: continued decline
Aggregate data provide scant evidence that a back-tothe-city movement has begun. In fact, there is a good

New Y ork
C ity s u b u rb s

New Y ork
SMSA*

U n ite d S ta te s

* S ta n d a rd M e tro p o lita n S ta tis tic a l Area, d e fin e d as
New Y o rk C ity plus Putnam, R o ckla n d , and W e s tc h e s te r
C o u n tie s in New Y o rk and B erg en C ounty in N ew J e rs e y .
S ource: U.S. D e p a rtm e n t o f C om m erce, B ureau
of the C ensu s.

deal of evidence to the contrary. The city’s total popu
lation, which rose during the 1960s, fell continuously
from 1971 to 1980 by a total of 800,000 people (Chart 1),
or about 10 percent. Even this number understates the
magnitude of the demographic change; the decrease in
white persons, for example, was over 1.8 million, nearly
30 percent of the 1970 level.
The census data also show that overall the city did
not become more attractive to high-income families.
During the 1960s, the number of fam ilies in New York
City with incomes greater than $50,000 (in inflationadjusted 1980 d ollars) rose about 30,000, but the
growth rate was slower than in the suburbs (Chart 2).3
Accordingly, the city’s share of all high-income fam ilies
in the Standard Metropolitan Statistical Area (SMSA) fell
from 66 percent in 1960 to 62 percent in 1970.4 During
the 1970s, however, national economic growth was
3ln current dollars, the definitions for high incom e were $50,000 or
higher in 1980, $25,000 or higher in 1970, and over $18,500 in
1960. In 1980 the high-incom e group num bered 5 percent of the
c ity ’s fam ilies.
4Unless otherw ise noted, all references to the New York m etropolitan
area refer to the 1980 com position of the New York-Northeastern
New Jersey SMSA. New York’s suburbs include Putnam, Rockland,
and W estchester Counties, New York, and Bergen County, New
Jersey. (Nassau and Suffolk com pose a separate m etropolitan area.)

FRBNY Quarterly Review/W inter 1983-84

slower, and the number of high-income families in the
New York SMSA even fell. The city’s share of this
smaller total dropped again, to 54 percent in 1980. In
the aggregate, then, the city’s relative attractiveness to
high-income families with respect to the suburbs fell
during both decades.
Two large demographic groups have grown over the
decade, despite the overall population decline: there
was an increase in the city’s population between ages
twenty-five and thirty-four and in the number of its oneperson households. Both these increases reflected
demographic trends occurring nationwide rather than
increased attractiveness of the city. In fact, New York
City failed to keep up with the SMSA or the U.S. growth
rates for either group. As the baby-boom generation
ages and if the rate of household formation slows, even
these segments of the population may decline by the
1990 Census.
The city posted significant gains in educational
attainment from 1970 to 1980. Over the decade, the
number of its college graduates grew by roughly 50
percent, from about 500,000 to over 750,000. This fol
lows a national trend of soaring numbers of college
graduates, but again New York City lagged the national
pace: the number of college graduates in the United
States nearly doubled between 1970 and 1980. The
number of graduates also grew more slowly in New York
City than in the entire metropolitan area (which recorded
a 64 percent increase), so that the city’s share of the
area’s graduates declined.
By all these measures, then, New York City’s share
of the metropolitan area’s gentry decreased over the
decade. While there were increases in the number of
one-person households, young people, and college
graduates, the city did not share fully in the nationwide
or metropolitanwide growth of these groups since the
proportion of each group choosing the suburbs over the
city increased. The census data do not indicate an
aggregate back-to-the-city movement for any of these
population groups.
Census tract level analysis
Although the aggregate data indicate a continued out
flow and further declines in the relative attractiveness
of New York City to high-income families overall, it is
still possible that particular parts of the city have
become more attractive to certain identifiable segments
of the population.5 To assess the extent of regentrifi
cation, it is necessary to examine demographic changes
sSeveral city neighborhoods are widely believed to have undergone
regentrification. Among the areas most widely discussed are
Brooklyn Heights and nearby neighborhoods in Brooklyn and
Greenwich Village in Manhattan. (For a close-up of five such
neighborhoods, see box on pages 44 and 45.)

Digitized for40
FRASER
FRBNY Quarterly Review/Winter 1983-84

at a geographically disaggregated level. Such an anal
ysis is possible with the use of census tract level data.
Census tracts in New York City have an area of eight
or more square blocks, and contain an average of 3,200
people or 1,300 households. There is a good deal of
variation, however; 10 percent of the city’s tracts have
fewer than 900 residents, and 10 percent have more
than 6,200. New York City has over 2,000 tracts. This
level of analysis is sufficiently fine to pick up relatively
subtle neighborhood shifts and is useful in measuring
the demographic and housing stock shifts that have
received public attention.
The tabulations provide strong—though only indirect—
evidence on the extent of gentrification and displace
ment. The ideal study would follow successive occu
pants of individual units and would determine the extent
to which low-income families move because of eviction
or large rent increases. Private and government surveys
often take this approach, but such studies generally
provide limited geographic information. The best infor
mation available on the extent of neighborhood shifts
between 1970 and 1980, then, is obtained by comparing
tract-level tabulations for the two census years.6
Some words of warning are appropriate, however.
While these census numbers are the most comprehen
sive and detailed source available, they are not without
limitations. Observing only census years may ignore
important increases and decreases occurring in be
tween, and after 1980. Moreover, the data also must be
interpreted with caution:
• Most of the census tabulations are based on a
sampling of the population, rather than on a 100
percent enumeration. The sampling procedure
may cause percentages and totals to differ from
the actual values.
• In many cases, census tract tabulations are not
reported to preserve confidentiality. As a result,
some tracts (generally those with a very small
population) had to be eliminated from some of
the analyses.7

•Most of the data used in this article came from computer tapes
created by the Census Bureau. For 1980, the files were Summary
Tape Files 1 and 3 for New York State. For 1970, they were the
Second Count tape and the Fourth Count tapes for Population and
Housing for New York State.
7There were other technical problems which caused census tracts to
be dropped. The 1970 census files had missing records, and there
were a few partitions and consolidations of census tracts over the
decade. For some tabulations of 1970-80 comparisons, missing
observations amounted to about 1 percent of the city's 1980
population. In many cases, five tracts accounted for most of the
total.

• Many of the statistics reported here are based
on preliminary data, subject to revision.
• The socioeconomic and housing data are based
on the responses of households. There may be
systematic errors in rent and income tabulations.
For example, the Census Bureau found that
respondents tended to overstate the utility cost
component of gross rent. They also tended to
understate sources of income components which
are “ minor or irregular”. (Comparisons of 1970
and 1980 data would not be adversely affected,
however, if the extent of overreporting and
underreporting were about the same for the two
censuses.)
• Changes in census procedures may affect some
1970-80 comparisons. For example, the 1980
census form had more racial categories than did
the 1970 form, and the procedure for handling
certain write-in responses was changed.
• The New York City government has charged that
the Census Bureau disproportionally under
counted blacks, Hispanics, aliens, and poor
people in the city. This charge is currently under
adjudication in Federal District Court in Man
hattan.
To identify tracts with significant changes, the extent
of the socioeconomic and housing-stock shifts in specific
neighborhoods are measured by the changes in pro
portions of a demographic group or housing type. Arbi
trary cutoffs were selected, in most cases at a 10 per
centage point increase (or decrease). Thus, for example,
an increase in the proportion of high-income households
from 1 percent to 2 percent would not be counted as a
significant change with respect to the tract population,
even though the share doubled.
Socioeconomic shifts
Increases in high-income households
Between 1970 and 1980, the number of high-income
households (those with incomes of $50,000 or more, in
1980 dollars) in New York City fell about 16,000. Despite
the citywide decrease, however, the number of highincome households increased in almost one third of the
city’s census tracts. These tracts were distributed all
across the city, and together their increase totaled over
18,000 high-income households (Map 1).
The largest increases per square mile took place
between 70th Street and 90th Street across Manhattan.
But these inflows did not drastically change neighbor

hood income distributions. In fact, the proportion of highincome households rose by 10 percentage points in only
six census tracts—one in Manhattan and none in
Brooklyn. Only one of these six tracts had more than
a few low-income families in 1970.8
When the definition of high income is broadened to
include those over $30,000 in inflation-adjusted 1980
dollars, there still were few dramatic changes in neigh
borhood income distributions.9 The percentage of
households with high incomes so defined rose by 10
points over the decade in thirty-two tracts (Map 2).10 The
increase exceeded 20 percentage points in only four
tracts. And, despite the focus of attention on Manhattan
and Brooklyn, most of the thirty-two tracts were outside
these boroughs.
If regentrification implies income distribution shifts
specifically in low-income areas, such changes were
rare. The proportion of low-income families exceeded
the 1970 citywide average of 16 percent in only nine
tracts with significant increases in the high-income
($30,000) population share. Of these nine tracts, the
largest increase in the high-income share took place in
a low-population tract near Canal Street and Broadway
in lower Manhattan.
Decreases in low-income families
Another subject of public discussion is the displacement
of low-income households by renovation, increased rent,
or eviction for owner occupancy. These changes may
take place without significant shifts in the upper end of
the income distribution, if high-income households are
not the group moving in or if there is a delay between
the exit of one group and the entry of another—such
as for major structural renovations. Examination of
changes in neighborhood proportions of low-income
families may identify significant income distribution shifts
which do not involve high-income households.
Citywide, the number of families with incomes below
125 percent of the poverty level increased more than
50,000. However, in over a third of the city’s census
tracts, the number of such families decreased. The total
decline for these tracts was over 60,000 families. About
one third of the tracts losing low-income families gained
•Low income is defined as less than 125 percent of the Census
Bureau’s poverty level (which varies by family size and other
factors). A low-income tract is one with a concentration of lowincome families greater than the citywide average of 16 percent.
About 30 percent of the city's tracts met this criterion in 1970.
Other definitions of poverty—families below the poverty level and
households with incomes below $5,000 in 1969—gave qualitatively
similar results in these and other tabulations.
•The 1970 income cutoff was $15,000 in current (1970) dollars.
10ln two of these tracts, moreover, the number of high-income
households fell, even though the population share rose. The other
income groups left these tracts even faster.

FRBNY Quarterly Review/Winter 1983-84 41

Map 1

Map 2

H ig h -in c o m e h o u s e h o ld s in New Y o rk C ity

In com e d is tr ib u tio n s h ifts in New Y o rk C ity

10 p e rc e n ta g e p o in t in c re a s e in h ig h -in c o m e
p o p u la tio n s h a re .
■ ■
0
■ H

10 p e rc e n ta g e p o in t d e c re a s e in lo w -in c o m e
p o p u la tio n sh a re .
B oth o f the above ch a n g e s.
M a jo r p a rk s , c e m e te rie s , a irp o rts , etc.

Map 3

M ap 4

C o lle g e g ra d u a te s in New Y o rk C ity

In -m ig ra tio n to New Y o rk C ity

20 p e rc e n ta g e p o in t in c re a s e in the p ro p o rtio n
o f c o lle g e gradua tes.

25 p e rc e n t o f p o p u la tio n liv e d o u ts id e
New Y o rk C ity in 1975.

M a jo r p a rks, c e m e te rie s , a irp o rts , etc.

M a jo r p a rk s , c e m e te rie s , a irp o rts , etc.

Digitized for
42FRASER
FRBNY Quarterly Review/W inter 1983-84

high-income households. These 270 tracts gained 7,500
high-income households and lost 15,000 low-income
families during the 1970s.11
However, as with the high-income population, few
census tracts had large declines in the low-income
share of the population. Across the city, the neighbor
hood percentage of low-income families fell by 10 points
or more in thirty-four tracts. These tracts mostly had very
small populations and generally were not the same as
the tracts with significant high-income increases (Map 2).
In fact, in four of the twenty-one Brooklyn and Man
hattan tracts with large declines of low-income families,
the number of high-income households also dropped to
zero.
In summary, some parts of the city have been moving
against the aggregate trend of out-migrating high-income
people and in-migrating low-income people. Such areas
are found all over the city, not just in Manhattan and
nearby Brooklyn. Changes in neighborhood income
distributions were proportionately large in only a few
areas.
Changes in racial composition
A definition of “ gentry” based exclusively on income
may be inappropriate. Significant neighborhood shifts
might have occurred along other dimensions. Wide
spread shifts in neighborhood demographics and
extensive displacement could be taking place without
notable shifts in the income distribution.
One measure of demographic change that is highly
visible is racial composition. If predominantly nonwhite
areas were entered by whites in large numbers, per
ceptions of regentrification could arise, even if the
incoming people did not have higher incomes than
current residents.
The census data show almost no sign of increasing
concentration of white persons. About 90 percent of the
city’s census tracts recorded a net decrease in the
proportion of white persons over the decade; the white
population share rose 10 percentage points or more in
only eight census tracts, of which three were in Brooklyn
or Manhattan. Moreover, three of these eight tracts had
low-income concentrations below the citywide average.
In 1980, these eight tracts collectively encompassed
only 900 households. An influx of high-income white
persons did not materially affect racial balance in poor
nonwhite neighborhoods.

250,000 college graduates, despite the fact that its
population fell by nearly a million. The increase in
graduates was widespread within the city: the proportion
of the population with four or more years of college
increased in 85 percent of the city’s census tracts. In
fact, the population proportion of graduates rose 10
percentage points or more in over 300 census tracts
and 20 percentage points in over sixty tracts (Map 3).
Despite the size of these increases, however, the city’s
share of the SMSA’s college graduates still declined
over the decade. The city’s numbers probably indicate
a broadly based increase in the extent of educational
attainment of New Yorkers, then, rather than indicating
a sudden influx of graduates to the city.
There appears to have been a good deal of move
ment between neighborhoods, though. The sharpest
educational gains were in Manhattan and nearby
Brooklyn. About one third of the tracts with significant
gains were low-income tracts in 1970; several of these
were in neighborhoods widely cited as gentrifying. In
fact, many of these neighborhoods had concentrations
of college graduates two or three times the citywide
average (box).
These educational gains were paralleled by changes
in the kinds of jobs held by neighborhood residents.
Over the decade the number of New Yorkers with
managerial, professional, or technical occupations grew
by 74,000, almost 10 percent of the 1970 level. The
neighborhoods with the sharpest educational gains also
had significant increases in the proportion of people with
such jobs.
Census tabulations, then, lend some quantitative
support to the direct observations of gentrification. It is
unlikely that such dramatic gains found in some of these
neighborhoods could have occurred without substantial
in-migration. But the failure of the income numbers to
match these changes is noteworthy. It suggests that the
new residents may be young professionals, early in their
careers, with prospects for substantially higher incomes.
In coming years, many may choose to live elsewhere.
In-migration to New York City
Despite New York City’s population loss between 1970
and 1980, many people went against the flow and
moved into the city. In fact, 10 percent of New Yorkers
in 1980, over 600,000 people, reported addresses out
side the city for 1975. This level of in-migration was
about equal to that for the 1965-70 period.12 In the

Increases in neighborhood educational attainment
As mentioned earlier, the city experienced a gain of
"T h is comparison refers to the over-$50,000 definition of high income,
for which increases are displayed in Map 1. Almost 40 percent of
these tracts lost low-income families.

12This comparison ignores 400,000 people in 1970 who moved since
1965 but indicated no prior address. Thus, the actual extent of
in-migration might have been significantly higher during the
1965-70 period than during 1975-80.

FRBNY Quarterly Review/Winter 1983-84 43

Close-up of Five New York City Neighborhoods
Population and housing statistics, 1980

Tract
number

Popu
lation

Ages
25-34*
(%)

Inmigrantsf
(%)

Col
lege*
(%)

High
income§
(%)

Low
income||
(%)

Rental
unitsll
(%)

Gross
re n t"
($)

Brooklyn Heights
1 .....................
3.01 ...............
5 .....................
7 .....................

4,902
5,353
6,173
3,272

24
33
31
35

24
24
24
27

48
53
52
55

27
36
24
31

2
2
5
6

70
85
91
83

290
325
295
293

Brooklyn

2 ,2 3 1 ft

234

5.614
6,918
3,344
6,216

33
33
42
33

24
25
37
21

47
50
62
56

13
17
28
25

14
8
4
6

95
95
92
94

261
289
358
312

14,702
12,169
7,843
6,035

31
33
22
15

27
29
22
20

57
52
66
59

36
28
52
62

4
4
6
2

84
94
66
59

377
384
448
500 +

9,219
9,343
8,665
10,357

33
30
30
25

26
21
18
14

50
52
54
50

21
26
26
31

19
9
14
7

96
94
90
89

337
338
320
341

264

West Village
67
73
75
77

E 8 3rd S treet
CL

CM
Tf

138

C
O

136

East Side
1 ?

r
<0

E 76th S tre e t

( 175

.................
.................
.................
.................

136
138
140
142

...............
...............
...............
...............

West Side
173
169

169 ...............
171 ...............
173 ...............
175 ...............

Manhattan

1 ,4 2 8 tt

'Percentage of the total population.
fPersons outside New York City in 1975 as a percentage of the
population over the age of five.
^Persons with four years of college or more as a percentage of the
population over the age of twenty-five.
§Households with incomes over $30,000 in 1979 as a percentage of all
households.

44FRASER
FRBNY Quarterly Review/W inter 1983-84
Digitized for

|| Families with incomes below 125 percent of the poverty level as a
percentage of all families.
UPercentage of all occupied units.
“ Median, including contract rent plus fuels and utilities.
t+Borough population in thousands.

Close-up of Five New York City Neighborhoods (continued)
Neighborhood change, 1970 to 1980

Tract
number

Colle g e #

High
Low
Rental
in c o m e # in c o m e #
u n its #
(In percentage points)

Gross
rent§§
(%)

Park Slope
153
155
157
159
165
167

.........................29
.........................27
.........................28
.........................33
......................... 26
.........................25

2
0
5
6
1
4

0
1
-9
-1 0
-8
-1

-1
0
-5
-6
-4
0

149
144
149
136
134
168

-2
-4
0
-5

5
-7
-5
-3

145
110
125
143

125

Brooklyn Heights
1
.........15
3 01 ............... .........13
5 ..............................12
7 ..............................17

Brooklyn

0
.8
4
8

-3

West Village
67
73
75
77

.................
.................
.................
.................

19
14
26
19

-1
0
14
6

1
0
-2
-5

-1
-2
-5
-2

1
5
5
-1

1
-4
-2
0

-9
-1
-3
-5

-2
2
2
1

6
-2
3
0

-3
-1
-3
-3

176
160
212
141

East Side
136
138
140
142

............... ........ 12
............... ........ 20
............... ........ 22
............... ........ 12

77
124
107
801|H

West Side
169
171
173
175

............... ........21
............... ........22
............... ........20
............... ........15

Manhattan

690
122
126
108

128

#C ha nge , in percentage points, of the corresponding columns on
the facing page, from their 1970 values.

Much of the discussion about gentrification suggests that the phe
nomenon has been occurring extensively in a few specific neigh
borhoods, including Greenwich Village, Chelsea, the Upper West
Side, and the East Side in Manhattan, as well as Park Slope,
Brooklyn Heights, Boerum Hill, and other Brooklyn neighborhoods
with easy access to Wall Street.
Detailed 1980 Census data for five such neighborhoods draw
a picture of a young, highly educated, high-income population
(table). The most striking indicator is the educational attainment
in these areas: in almost every census tract, at least half of the
residents over the age of twenty-five had four years of college or
more, compared with a citywide proportion of 17 percent. The
proportions of households with incomes greater than $30,000 also
tend to exceed the citywide concentration of 16 percent, and the
proportions of families with incomes below 125 percent of the
proverty level are similarly below the citywide average of 22 percent.
And, in most of these neighborhoods, median gross rents were
significantly above the citywide median of $248 in 1980.
Comparison to 1970 data provide some evidence that gentrifi
cation has taken place. With the exception of Park Slope, the
proportion of in-migrants generally ranged from one fifth to over
one third. Moreover, these in-migrants seem to be very highly
educated. The increase in the proportion of college graduates
ranged from 12 percentage points to 33 percentage points in these
tracts during the 1970s.
In contrast, the income distribution showed much less change.
The high-income proportion of households rose 10 percentage
points or more in only one tract. Similarly, the low-income share
of the family income distribution fell in most of these tracts, but
these declines were rather modest.
Brownstone purchases and “co-op” conversions did not have
a drastic effect on the composition of the housing stock in these
tracts; the rental share of the occupied housing stock fell during
the decade but not by dramatic amounts. In the Brooklyn neigh
borhoods, for example, the rental share generally remained above
the citywide and boroughwide average of 76 percent. In Manhattan,
about a third of the tracts were below the boroughwide average
of 93 percent by roughly the amount of the ten-year decline in the
share. But in no case did the rental proportion fall by as much as
10 percentage points.
Finally, most of these tracts had larger than average rent
increases. Citywide, the median gross rents rose by about 120
percent, a figure surpassed by all but five of these census tracts.
A West Side tract had an especially large increase, as a number
of single-room occupancy (SRO) buildings, containing several
thousand units with rents below $30 per month, were eliminated.

§§Percentage change of medians between 1970 and 1980.
|||| 1980 median is reported only as over $500; actual growth rate
may be larger than the number shown.

FRBNY Quarterly Review/W inter 1983-84

aggregate, then, New York City has not been attracting
more people to relocate here.
Many neighborhoods, however, had significantly
greater inflows from outside New York City. There were
over 200 tracts, in which 20 percent of the 1980 pop
ulation (over the age of five) were recent in-migrants,
and about 130 tracts where the proportion exceeded 25
percent (Map 4). The greatest concentrations were in
Manhattan, where the boroughwide average was about
18 percent. Brooklyn neighborhoods near Manhattan
also had large concentrations of recent in-migrants.
However, significant amounts of in-migration occurred
in pockets of other areas of the city, areas out of the
focus of most popular discussions of gentrification—the
Bronx, Flatbush, Flatlands, Greenpoint, and Williams
burg. Much of midtown Manhattan also had significant
in-movements in the latter half of the 1970s.
Despite all this in-migration, the data show very little
movement to the city from its own suburbs. Ten percent
of neighborhood residents were suburbanites in 1975 in
only five census tracts—one in the Bronx near West
chester, one in Brooklyn on Jamaica Bay, one on the
waterfront near Williamsburg, and two in lower Man
hattan. The citywide total of 30,000 suburb-to-city
movers was half that of the 1965-70 period.13
It is also noteworthy that much of the migration to
New York City came from outside the United States. As
of 1980, about 5 percent of the city’s residents, or half
of the in-migrants, lived abroad in 1975. In Brooklyn, 68
percent of the people living outside the city in 1975 also
lived outside the country. In Manhattan, about a third of
the new residents were immigrants.
Housing stock changes
Significant changes in the socioeconomic characteristics
of a neighborhood also may have an impact on struc
tural and occupancy characteristics of the housing stock.
Sometimes the housing stock changes may be more
conspicuous than the socioeconomic changes, as when
several buildings on a block are being renovated at the
same time. If the income distribution is an inappropriate
measure of one group displacing another, housing stock
statistics might indicate significant shifts not reflected in
other tabulations.
Units per structure
One variable useful for investigating the extent of a
particular kind of gentrification is the number of units
per structure. If brownstones with several rental units
are bought by high-income households and remodeled
13The tabulations for 1970 referred to the then-prevailing definition of
the SMSA. In 1970, Nassau and Suffolk were included in the New
York SMSA and Putnam and Bergen were not; these changes may
explain part of the statistical decline in suburb-to-city migration.

FRBNY Quarterly Review/Winter 1983-84

for owner occupancy, there would be more one- and
two-unit structures and fewer structures with three, four,
five (or more) units.14
In the aggregate, the number of one- and two-unit
structures stayed about constant. But, in one third of the
city’s census tracts, the number of such structures rose,
by an additional 50,000 units in one- and two-unit
structures.
The proportion of units in this type of structure
increased by 10 percentage points or more in about 50
census tracts. Only one such tract was in Manhattan,
however, and about three in the areas of Brooklyn
generally characterized as gentrifying. The rest of the
tracts were throughout Brooklyn, the Bronx, Staten
Island, and Queens.
The conclusion is that remodeling of structures with
several units to structures with one or two units had a
significant effect on the composition of the housing stock
in only a few isolated pockets of the gentrified areas of
Manhattan and nearby Brooklyn. If gentrification has
brought dramatic changes in the housing stock in some
of these areas, it is necessary to look to other housing
variables for statistical confirmation.
Owner occupancy
Another commonly discussed neighborhood change
involves widespread purchase of brownstones and
similar rental properties for owner occupancy (at least
in part of the purchased structure). If high-income
households have been buying up rooming houses and
rental structures in large numbers, and these trends are
concentrated in specific neighborhoods, then these
changes should show up in the census data as signif
icant shifts in the proportions of owner occupants—
especially considering that one owner can replace three
or more renters.15
Only two census tracts in the regentrified part of
Brooklyn—the area broadly encompassing Park Slope
to Brooklyn Heights—experienced a sig n ifican t
replacement of owner-occupied units for rental units,
however.16 In fact, the rental proportion did not decline
14Renovation and subdivision of one-unit structures into multi-unit
rental properties is also consistent with regentrification. The analysis
presented here looks only for a specific renovation activity which
may reduce the rental stock and displace the most low-income
families per high-income household. If both subdivision and
consolidation are going on, regentrification may not affect the
renter-owner composition of the housing stock.
15Of course, three high-income renters can also replace one lowincome owner of an unimproved brownstone. But, if the rental
statistics are to show any regentrification, it would have to be as an
increase in "co-op" or building ownership.
1,A significant shift is defined as a 10 percentage point increase in
the rental proportion of the occupied housing stock, provided that
the number of rental units falls and the number of owner-occupied
units rises.

by 10 percentage points in any Park Slope or Brooklyn
Heights tract.
In Manhattan, one Greenwich Village tract, three in
the East 60s, and three other tracts on or near the East
Side experienced such changes. Many of these loca
tions suggest “ co-op” or condominium conversion of
housing that already was serving high-income residents.
Four Queens tracts also showed significant increases in
owner occupancy.
Rent increases
For the city as a whole, rent increases seem to have
been fairly moderate for the 1970s. Citywide, the
median gross rent (which includes fuel and utilities)
grew from $112 to $248, an increase moderately greater
than that of the general price level (which about dou
bled). And high-rent units on average had similar rent
increases. Ten percent of gross rents were at or above
$200 in 1970; the comparable figure in 1980 was about
$430.
But in many neighborhoods typical rent increases
were much larger: the growth of median gross rent
exceeded 150 percent in nearly 250 census tracts. Sixty
of the tracts were in Manhattan, generally south of 100th
Street. There were ninety-six in Brooklyn, a third of
which were in neighborhoods near lower Manhattan.
About forty-five of the tracts were in Queens, thirty-five
in the Bronx, and thirteen in Staten Island.
The change in median gross rent is an ambiguous
indicator of housing market conditions. Increases in the
median might indicate rent increases only for belowmedian units, the construction of high-rent units, or the
demolition of low-rent units. Conversion of rent-stabilized
buildings to “ co-ops” or condominiums can also
increase a tract’s median rent by removing low-rent units
from the rental stock. Moreover, differences in rent
increases may be largely a reflection of differential
coverage of rent control and stabilization, as well as the
varying impacts on units subject to these controls (due,
for example, to areas with high or low turnover).17
Nevertheless, rents increasing faster than the citywide
average may to some extent be an indication of gen
trification. To the extent that the new gentry push up
rents in some neighborhoods by outbidding the former
tenants for some units, large rent increases may be
indicative of conditions which force some of the original
occupants to leave. In Brooklyn, however, most areas
with rapid rent growth were not in prime gentrification
territory. In Manhattan, areas below 14th Street had

17There is evidence that median rent increases do not reflect uniform
rent increases. The 10th and 90th percentiles of tract rent
distributions often grew significantly faster or slower than the tract
median rents.

many tracts with large rent increases, but so did the
less-discussed Midtown South area.
Summary and conclusions
Tract-level tabulations from the 1980 Census provide the
first opportunity to investigate empirically the extent and
nature of changes of New York City’s neighborhoods.
Heretofore, much of the discussion about the shifts
during the 1970s was based on limited observations and
anecdotal evidence. An analysis of socioeconomic and
housing data has led to the following conclusions:
• The overall attractiveness of New York City to
the “gentry", by various definitions, did not grow
between 1970 and 1980. The city’s share of the
metropolitan area’s high-income households,
college graduates, and managers/professionals/
technical workers all fell over the decade.
• Neighborhood takeovers by high-incom e
households did not occur. Even moderate
changes in income distribution were fairly rare,
and most of the larger shifts generally occurred
outside Manhattan and Brooklyn.
• The population shares of low-income families
did not decline significantly except in small
pockets of some neighborhoods. Many of these
changes, moreover, took place outside the prime
gentrification areas of Manhattan and nearby
Brooklyn.
• The numerically smallest changes were of racial
composition. Only 10 percent of the city’s
census tracts posted any net increase in whites;
fewer than 1 percent experienced substantial
increases in population share.
• The strongest evidence of gentrification comes
from increased educational attainment. Park
Slope and other well-known gentrified areas
posted some of the largest gains. Corre
sponding shifts in the occupational mix were
also observed. But many of the young profes
sionals, with incomes similar to those of the
1970 occupants, may have prospects for higher
incomes in coming years. Whether they will stay
in New York City is an open question.
• There was little evidence of widespread
remodeling of multi-unit structures to create
one- and two-unit structures in Manhattan and
nearby Brooklyn. And increases in the latter
structure type took place in several areas not
associated with gentrification.

FRBNY Quarterly Review/Winter 1983-84 47

• Few instances were observed of especially large
owner-occupancy shifts. Substantial decreases
in renters were not observed in Brooklyn
neighborhoods near downtown Manhattan, and
many of the other shifts could be attributed to
“ co-op” or condominium conversion of highincome rental buildings.
• The most dramatic housing stock changes were
rent increases. Sections of the neighborhoods
most commonly cited as gentrified had some of
the largest increases, but median rent increases
in many other areas were equally large.
In summary, the census data indicate that some areas
of the city have indeed moved counter to the citywide
trends of declining high-income households and
increasing low-income families. The most discussed
neighborhoods in Manhattan and Brooklyn were among
those areas, but some of the largest income distribution

Digitized for 48
FRASER
FRBNY Quarterly Review/Winter 1983-84

shifts were found in parts of the Bronx, central Brooklyn,
and other areas not noted for much gentrification. Other
kinds of socioeconomic and housing-stock changes
generally associated with the phenomenon were
observed only in moderate degrees, were not wide
spread, and were not limited to neighborhoods widely
recognized as gentrifying.
It should be realized that regentrification is not a
simple phenomenon whose essence can be captured by
any dozen census variables. The tabulations presented
in this article can only reflect the order of magnitude of
neighborhood changes. These census variables cannot
describe the changing appearance of neighborhoods—
the departed people, the new faces, the closed ethnic
restaurants, and the new boutiques. And the statistical
significance of gentrification may not reflect the impor
tance to neighborhood residents or other interested
parties. Nevertheless, the numbers are still an important
part of the picture and the individual stories must be
balanced with an overall perspective.

Daniel E. Chall

Currency Misalignments
The Case of the Dollar and the Yen

The dollar’s dramatic rise in the last three years has
initiated an international debate involving sharply con
flicting views. The strong dollar has been largely behind
the substantial loss of U.S. competitiveness in world
markets, which has importantly contributed to the large
and still growing U.S. trade and current account deficits.
As a result, many analysts assert that the dollar is
“ overvalued”. By contrast, some analysts focus their
attention on the growing U.S. trade deficit with Japan
and on the large and rising Japanese current account
surplus. They conclude that the yen is “ undervalued”.
Still others argue that terms like overvaluation and
undervaluation are meaningless in a floating rate system
because exchange rates are basically determined by
market forces.
The purpose of this paper is to shed some light on
this debate. Theory and empirical evidence both suggest
two major conclusions. First, it does make sense to talk
about the possible overvaluation or undervaluation of a
currency even in a freely floating system. However,
these concepts mean different things to different people.
Therefore, unqualified use of these terms can lead to
confusion and unnecessary argument. Second, empirical
evidence on changes in international competitiveness as
well as on the behavior of trade and current accounts
suggests that the dollar is unusually strong but the yen
is not especially weak.

The views expressed in this article are those of the author and do not
necessarily reflect those of the Federal Reserve Bank of New York.

Some conceptual issues
When is a currency in disequilibrium? To arrive at a
sensible and operationally useful answer, it is first nec
essary to make a distinction between the concepts of
short-run or temporary equilibrium and long-run or what
may be called fundamental equilibrium.
A sequence of short-run equilibrium exchange rates
is determined by continuously shifting forces of supply
and demand in the foreign exchange market. Although
trade in goods and services contributes to changes in
the balance of supply and demand of foreign exchange,
trade in financial assets, often speculative in nature,
dominates the short-run dynamics of this market. And
market rates change constantly as market participants
assess and reassess all relevant new information. Since
transaction costs are small and buying and selling go
on at all times of the day, the foreign exchange market
is in the process of clearing virtually continuously. Thus,
for all practical purposes, it is a close approximation to
treat market exchange rates as short-run equilibrium
exchange rates, although we know that strictly speaking
not every rate quoted during a trading day represents
an equilibrium.
A long-run equilibrium exchange rate, by contrast, is
determined only when the world economy as a whole
is in equilibrium and all other economic variables are
also in equilbrium. In this kind of general equilibrium,
markets for assets, goods, and labor all clear in the
sense that supply equals demand. In addition, all
expectations are realized and all relative prices remain
constant.

FRBNY Quarterly Review/Winter 1983-84 49

This notion of equilibrium, although useful in theory,
does not have a counterpart in the real world. The world
economy is continuously adjusting, but often slowly, to
new shocks, and therefore it is questionable whether it
could ever be in an overall long-run equilibrium. For
example, slow and lagged adjustments in the goods
markets and in the current account often lead to swings
in market exchange rates that may persist over ex
tended periods at times. In addition, the long-run equi
librium values of all economic variables, including
exchange rates, are also changing over time, sometimes
very sharply. Market exchange rates are, therefore,
almost always overshooting or undershooting long-run
equilibrium exchange rates which are changing as well.
Thus, deviations from long-run equilibrium represent
a normal state of affairs, and they could be persistently
large at times. These deviations may simply reflect slow
adjustments in other markets rather than any malfunc
tioning of the exchange market per se.
But there are times when some may consider the
short-run volatility or even the medium-term swings in
market exchange rates as excessive and not consistent
with changes in underlying economic and financial
conditions (fundamentals). Exchange rate swings may
also reflect a relative absence of stabilizing speculation
or even the presence of destabilizing speculation. In
extreme cases, market exchange rates may behave like
speculative “ bubbles”.1 Admittedly, it is often difficult to
determine whether the market is being driven by de
stabilizing speculation or is just responding, with some
uncertainty, to so-called economic fundamentals. This is
so because the foreign exchange market is essentially
a speculative market, and there is no consensus, even
among economists, which variables represent funda

1Market observers and policy practitioners often voice their concern
with bandwagon and other destabilizing speculative behavior in the
exchange market. Academic economists have also recognized this
possibility long before the advent of the current float. Nurkse, for
example, wrote in 1944 "...anticipatory purchases of foreign
exchange tend to produce or at any rate to hasten the anticipated
fall in the exchange value of the national currency, and the actual
fall may set up or strengthen expectations of a further
fall...Exchange rates in such circumstances are bound to become
highly unstable, and the influence of psychological factors may at
times be overwhelming.” R. Nurkse, International Currency
Experience: Lessons of the Interwar Period (League of Nations,
Princeton, New Jersey, 1944).
Recently, Dornbusch expressed a similar view, "The idea of a
bubble is worth recognizing because it emphasizes that there is no
tendency for efficient capital markets to force a rate toward its
fundamental value...There is no reason to assume that the present
value of the dollar does not represent such a speculative trap"
(page 7). R. Dornbusch, "U.S. International Monetary Policies", a
paper presented to the Board of Governors of the Federal Reserve
System, September 30, 1982.

Digitized for50
FRASER
FRBNY Quarterly Review/Winter 1983-84

mentals, let alone how to assess their linkages with
exchange rates.2
The distinction between the two concepts of equilib
rium helps shed light on a number of aspects of the
current currency debate. For example, it appears that
those who believe that the dollar cannot be in disequi
librium in the current floating rate system are essentially
arguing that the dollar is almost always in short-run
equilibrium. They are also expressing the view that the
market rate reflects all available information and the
market assessment of economic fundamentals is more
accurate than that of anyone else.3
Those who believe that the dollar is overvalued do not
necessarily dispute the view that the dollar is almost
always in short-run equilibrium. But they stress the point
that the dollar may be severely out of line with the likely
range of long-run equilibrium values. Although this
concern sometimes focuses on the persistence of
deviations from long-run equilibrium per se, and some
may have misgivings about the market assessment of
fundamentals, the real issue is often the macroeconomic
and distributional costs that these deviations induce.
The focus of the concern and the assessment of the
costs involved, however, vary a great deal, and so do
the qualitative judgment and quantitative assessment of
currency misalignments.
Thus, terms like “ overvalued” and “ undervalued” in
and of themselves are nebulous, but they do reflect
someone’s judgment that the going rate is undesirable
for one reason or another. To some, the dollar is over
valued because it has led to a substantial loss of
international competitiveness of U.S. manufacturing.
Certainly, the industries that have been adversely
affected consider the dollar overvalued. The large and
deteriorating trade deficit and the loss of jobs resulting
from this erosion of U.S. competitiveness are seen as

2James Tobin has aptly summarized the problem, "...n o one has any
good basis for estimating the equilibrium dollar-mark parity for 1980
or 1985, to which current rates might be related. The parity depends
on a host of incalculables—not just the future paths of the two
economies and the rest of the world but the future portfolio
preferences of the world’s wealth owners....In the absence of any
consensus on fundamentals, the markets are dominated—like those
for gold, rare paintings, and—yes, often equities— by traders in the
game of guessing what other traders are going to think,” See J.
Tobin, “A Proposal for International Monetary Reform” (Cowles
Foundation Paper No. 95, Yale University, New Haven, 1980).
3A recent expression of this view comes from Treasury Secretary
Donald Regan: "In a floating exchange rate system, there can be no
correct value to any currency other than the value given to a
currency through market transactions” (quoted in Washington Post,
February 23, 1984. Secretary Regan, however, also believes that
"...it is confused thinking to describe the dollar as overvalued".

confirmation of the view that the dollar is overvalued
even from a national standpoint.
Many European analysts and officials consider the
dollar overvalued because of the increased import costs
and the inflationary impact of the associated deprecia
tion of their currencies. The argument that the dollar
deepened the European recession by inducing the
authorities to raise interest rates is sometimes advanced
as further evidence indicating the overvaluation of the
dollar. The heavily indebted less developed countries
(LDCs) view the dollar as overvalued because it has
increased the burden of debt servicing substantially
since most of their debt is denominated in dollars.
Thus, there are many groups at home and abroad that
have been adversely affected by the current strength of
the dollar, and hence they consider the dollar over
valued. However, significant exchange rate changes—
equilibrating and disequilibrating ones—are always going
to induce costs for some groups and lead to the com
plaint that one currency or another is misaligned. In the
recent period, the concept of an overvalued dollar, for
some, simply expresses the fact that a strong dollar is
costly to them in one way or another.
The strong dollar, however, generates not only costs
but benefits as well. The main beneficiaries of the
strong dollar are consumers in the United States and
foreign competitors of U.S. producers. The strong dollar
has also significantly contributed to the achievement of
one of the major policy objectives in the United States—
the reduction of inflation. These examples again illus
trate the distributional aspect of changes in exchange
rates.
There is, however, a second and quite different way
of looking at the concept of currency misalignment.
Some analysts may consider the dollar overvalued from
the standpoint of resource allocation costs rather than
a distributional concern.
In their view, the current strength of the dollar is
unsustainable, and hence resource allocations and
global adjustments resulting from the strong dollar are
temporary and likely to be reversed when the dollar
declines. They believe that the strong dollar has been
causing hardship in otherwise profitable industries in the
United States and perhaps providing incentives for
inefficient industries to spring up abroad. These
resource allocations and reallocations are, therefore,
unnecessarily costly and should be prevented.
An associated theme is the following. If the current
levels of the dollar are ultimately unsustainable
(because the large and growing U.S. current account
deficits cannot be financed indefinitely) but nevertheless
persist over the medium term, there may be a protec
tionist fallout in the United States. A variety of protec
tionist measures may be sought that would undermine

the progress made over the past three decades toward
a liberal global trading environment. Thus, persistent
currency misalignments can impose real costs by
reducing the economic efficiency that stems from a
global expansion of free trade.
If unsustainable exchange rate movements are costly
from the standpoint of efficient allocation of international
resources, then it is important that these movements be
detected and their magnitude be estimated. This is, in
fact, the way most economists view the issue of cur
rency misalignments. Although they usually define
misalignments in exchange rates as deviations from
long-run equilibrium values, many conceptual and
practical problems have discouraged the use of generalequilibrium models of the world economy for calculating
long-run equilibrium exchange rates. Instead, simple
rules of thumb involving the concept of purchasing
power parity (PPP) and considerations of current
account balance are widely used to detect currency
misalignments and sometimes to offer quantitative
estimates. These approaches are, however, beset with
many conceptual and practical problems as well, espe
cially if they are used mechanically.
Assessment of currency misalignments
Attempts to detect disequilibrium in current exchange
rates through PPP calculations must assume that, while
the economy is suspected to be in long-run disequili
brium today, it was in long-run equilibrium or at least
much closer to it in the chosen base period. In other
words, PPP methodology essentially ignores two crucial
insights that emerge from theory, namely, (1) observed
rates are almost never in long-run equilibrium and (2)
long-run equilibrium real exchange rates do change over
time. This methodology also requires the use of a price
or a cost index, and estimates of misalignments are
often very sensitive to this choice.
A second approach is to define an approximately
balanced current account, sometimes over a business
cycle, as a long-run equilibrium condition. It has some
practical appeal because it implies an absence of any
net inflow or outflow of savings, and hence an absence
of a redistribution of wealth between the country and the
rest of the world. However, this notion of long-run
equilibrium neither holds up to historical scrutiny, nor is
based on first principles of economics. The United
Kingdom, for example, ran surpluses in its current
account continuously between 1870 and 1914. Between
1946 and 1970, the U.S. current account was in surplus
in all but three years. These persistent “ imbalances” did
not necessarily point to any serious macroeconomic
disequilibrium.
In principle, any current account imbalance is sus
tainable and optimal if it reflects the saving-investment

FRBNY Quarterly Review/Winter 1983-84 51

decisions of rational individuals and profit-maximizing
firms. In other words, if current account deficits and
surpluses result from or lead to matching and voluntary
private trade in assets, then those external “ imbalances”
are both sustainable and optimal. This implies that there
need be no macroeconomic problem if a country with
a higher propensity to save or with a lower rate of return
on investment at home runs current account surpluses
for long periods of time.
Although mechanical applications of PPP and current
account considerations can lead to misleading conclu
sions, judicious use of information on changes in inter
national competitiveness, as well as careful analysis of
current account behavior, can prove very useful in
assessing currency misalignments. From a practical
point of view, it is more tractable to view currency mis
alignments as deviations from currency values that are
sustainable over the medium term rather than as
deviations from the elusive long-run or fundamental
equilibrium exchange rates.
A sustainable exchange rate can be thought of in a
very broad manner: it is a rate that can be sustained
over the medium term by policies that are appropriate
from the point of view of efficient allocation of inter
national resources. It is important to consider the
appropriateness of policies. Even if a currency value can
be sustained by inappropriate policies, it could be con
sidered unsustainable because inappropriate policies
themselves should be viewed as unsustainable.
A sustainable currency value is, therefore, one that
can be maintained by government policies that are
appropriate and sustainable in the sense of being con
sistent with such common national goals as stable
economic growth, low inflation, and low unemployment.
Assessing whether an exchange rate is sustainable from
this point of view is not an easy matter, especially if
distributional considerations—both intranational and
international—are brought to bear on this judgment. For
example, policy goals or policy mixes of a large country,
even if deemed appropriate by its residents, may well
be considered undesirable by its trading partners.
A careful analysis of changes in international com
petitiveness and the behavior of relative current
accounts, however, can help detect extreme deviations
from sustainable exchange rates. In less obvious situ
ations, these types of information can contribute
importantly to public discussion of, and private negoti
ations on, currency misalignments even if noneconomic
considerations ultimately determine the nature of gov
ernment policies. For example, if currency appreciation
leads to substantial losses in a country’s international
competitiveness, it may suggest that the current value
of the currency is unsustainable.
A narrow focus on conventional measures of inter

FRBNY Quarterly Review/Winter 1983-84

national competitiveness, however, could be misleading.
Changes in competitiveness may reflect structural shifts
in the economy and hence may not indicate that the
going exchange rate is unsustainable. Thus, it is nec
essary to supplement this type of information with a
more comprehensive analysis of the country’s “ under
lying” external payments position.
This requires a judgment on how large a current
account surplus or deficit can be considered sustain
able, given the medium-term saving/investment behavior
of the country in relation to its trading partners. It is also
necessary to take into account temporary and cyclical
factors that may be influencing the present and pro
spective behavior of the current account and the capital
account of the country’s balance of payments. If the
present and future current account balance of the
country appears unsustainable over the medium term,
even after accounting for factors other than the
exchange rate, then the going exchange rate can be
considered unsustainable.
Again, a thorough analysis of this type, especially if
it is to be consistent on a multilateral basis, requires
modeling the linkages among major economic variables
as well as a great deal of judgment. However, less
formal analysis can prove useful in extreme cases. For
example, a rapidly deteriorating current account deficit
is unsustainable if it points to a rising foreign debt/GNP
ratio that the country will be unable or unwilling to
maintain after a certain point. Or, if a large and dete
riorating current account deficit reflects policies that are
inappropriate and unsustainable from the standpoint of
the medium-term objectives of the country and/or the
international community, then the external imbalances
can be viewed as unsustainable.
The rest of the paper is devoted to an assessment
of the current exchange rates of the dollar and the yen
along these lines. Particular attention is paid to the view
that the yen is undervalued. Our assessment relies on
an examination of various available measures of inter
national price and cost competitiveness supplemented
by an informal, albeit careful, analysis of the behavior
of the current accounts of Japan and the United States.
The yen problem
There are two versions of the argument that there is a
yen problem. One version simply claims that the yen is
undervalued. The other version is more specific; it holds
that the dollar is more overvalued against the yen than
any other major currency. That is, there is a special yendollar imbalance that cannot be explained solely by the
dollar’s overall strength.
Japan’s large and rising trade and current account
surpluses are seen as prima facie evidence for an
undervalued yen. Japan’s gains in international com

petitiveness against the United States as well as the
large and growing U.S. trade deficit with Japan are
interpreted as evidence suggesting a special yen-dollar
imbalance.
For example, in April 1983, Fred Bergsten of the
Institute of International Economics (HE) stated,
Quantitatively the dollar-yen misalignment is more
severe than the misalignment between the dollar
and any other major currency....U.S. international
price competitiveness deteriorated against Japan
by over 70 percent in four years. Is it any wonder
that the U.S.-Japan trade imbalance has soared
to record levels and that a major crisis exists in
economic relations between the two coun
tries?... the dollar is overvalued against a number
of important currencies, but it is more overvalued
against the yen than against the others.4
More recently, The Economist (December 3, 1983,
page 15) expressed its view in this way.
Overprotected farmers and an undervalued yen
both anger Japan’s trading partners, especially the
Americans__ This cheapness of the Japanese
currency and dearness of the American one can
go a long way towards explaining why Japan is
headed for a current-account surplus of $25 billion
this year and America a deficit of $40 billion.
In what follows, these views on the dollar and the yen
exchange rates are assessed on the basis of the data
on changes in competitiveness and on movements in
the current account. The major conclusion is that the
dollar appears to be unusually strong, but the yen is not
particularly weak. Japan’s recent gains in competitive
ness against the United States have resulted from an
overall strength of the dollar, and not from any overall
weakness of the yen.
Changes in competitiveness
Changes in international competitiveness are examined
in a number of different ways. These changes are esti
mated for both the economy as a whole and for the
manufacturing sector. Data on both price competitive

4Testimony before the Senate Foreign Relations Committee, pages 67. However, John Williamson, also of the HE, did not identify the yen
as particularly weak in the subsequent monograph, “ The Exchange
Rate System” (October 1983).
The Organization for Economic Cooperation and Development
(OECD), in its December 1983 Economic Outlook (page 8), comes
close to identifying a separate yen problem: "Exchange rates,
assessed on the basis of current account prospects, have seemed
out of line, with the dollar high and the yen, in particular, low”.

ness and cost competitiveness are used. The calcula
tions are performed for the United States, Japan, and
to facilitate comparison Germany and France. Changes
in competitiveness are computed on a trade-weighted
(effective) basis for all four countries as well as vis-£tvis the United States for the other three countries.
The nominal trade-weighted dollar started its current
upward swing in late 1980. But the average tradeweighted value of the dollar did not rise in 1980 from
its level in 1979. Rather than choosing 1980 as a ref
erence period, we take a long-run view and compare
recent levels of exchange rates and competitiveness
with their corresponding averages for the entire 197480 period. This averaging minimizes the effects of
peculiarities of particular years on the broad conclusions
of this analysis.
The year 1983 is chosen as the terminal period for
the aggregate economy. For manufacturing, data avail
ability dictates that we use 1983-11 as the terminal
period. Later we argue that our principal qualitative
conclusions are essentially invariant with respect to any
reasonable choice of these base and terminal periods.
Chart 1 presents some preliminary evidence in sup
port of the view that the yen is not weak but that the
dollar is strong. The appreciation of the inflationadjusted trade-weighted dollar that began in late 1980
continued through 1983 and the value of the dollar in
1983 was much higher than its 1974-80 average level
(top panel).5 By contrast, the inflation-adjusted tradeweighted yen appreciated sharply in 1980 and then
more than offset this appreciation by depreciating until
late 1982. The substantial depreciation of 1981-82 has
been partly responsible for the international concern on
the weakness of the yen. The yen, however, appreciated
sharply in late 1982 and, in contrast to the dollar,
remained in 1983 near its average during 1974-80.
The bottom panel of the chart points to the overall
strength of the dollar rather than to any overall weak
ness of the yen as a source of Japan’s recent gains in
competitiveness against the United States. The chart
shows that, when compared with 1974-80 averages, the
yen is significantly weak in inflation-adjusted terms vis&-vis the dollar (because of the overall dollar appreci
ation), but it is very strong against the German mark
and the French franc. These observations are confirmed
and further elaborated by additional evidence summa
rized in the table.
Bilateral competitiveness. The major observations on
bilateral exchange rates and bilateral competitiveness
5This index uses weights from the IMF’s multilateral exchange rate
model (MERM) for 11 major countries. Wholesale price indexes
(WPI) are used in measuring inflation.

FRBNY Quarterly Review/Winter 1983-84 53

C hart 1

R eal E x c h a n g e R ates
Index 1974-8 0= 10 0

19 74-1980

1980

1981

1982

1983

W h o le s a le p ric e s ha ve been used to a d ju s t fo r
in fla tio n d iffe re n tia ls . W e ig h ts fo r e ffe c tiv e ra te s
are d e riv e d fo r e le v e n m a jo r c o u n trie s fro m the
IM F’ s m u ltila te ra l e x c h a n g e ra te m o d e l (M ER M ).

of Japan vis-a-vis the United States can be summarized
as follows:
In 1983, the yen was 4 percent higher against the
dollar than in 1974-80 on average. The German mark,
however, was 15 percent lower and the French franc
was 40 percent lower.
When adjusted for wholesale price inflation differen
tials, the yen does show a depreciation of 13 percent
against the dollar. The mark, however, depreciated over
25 percent and the franc over 30 percent. Thus, Japan’s
gain in overall price competitiveness against the United
States has been substantially less than that of Germany
and France.
In m anufacturing, Japan and Germany made sub
stantial, but similar, gains against the United States both
in terms of export price competitiveness (30 percent)
and labor cost com petitiveness (20 percent). France’s
gains have been even greater. But note that, while
France’s and to a lesser extent Germ any’s competitive
advantage against the United States can be attributed
to the depreciation of their currencies, Japan’s advan
tage is more than accounted for by the relatively slow
growth of its prices and costs. For example, if currency
values remained at their 1974-80 averages, Japan’s
gains in export price com petitiveness would have been
even greater than 30 percent, whereas Germ any’s and
France’s gains would have been a great deal less.

54 FRBNY Quarterly Review/W inter 1983-84

Effective competitiveness. The evidence on changes in
exchange rate s and c o m p e titiv e n e s s on a tra d e weighted basis (table) can be summarized as follows:
Relative to 1974-80 averages, the yen was, in fact,
as strong as the dollar in 1983. Both currencies had
appreciated around 23 percent. By contrast, the mark
was over 10 percent stronger and the franc was around
27 percent weaker. As far as nominal rates are con
cerned, it is the strength of the yen rather than its
widely perceived weakness that clearly emerges from
the data.
When the relatively low inflation in Japan is taken into
consideration, Japan’s gains in aggregate price com
petitiveness in 1983 turn out to be negligible (1 percent).
During the same period, the United States incurred a
loss of 23 percent in overall price competitiveness.
Germany and France, on the other hand, show sub
stantial gains in competitiveness, about 8 percent and
20 percent, respectively.
The comparisons above refer to price competitiveness
for the aggregate economies. But what about manu
facturing competitiveness on a trade-weighted basis?
According to the IMF index on m anufacturing price
competitiveness, Japan in 1983-11 was where it was on
average during 1974-80, while Germany had gained 5
percent and France 9 percent. According to the Morgan
Guaranty index,6 however, the manufacturing sectors of
all three countries appear to have made sim ilar gains
(between 6 to 8 percent) in price com petitiveness. By
contrast, manufacturing price com petitiveness in the
United States declined by over 20 percent.
Finally, as regards changes in labor cost com peti
tiveness in m anufacturing, Germ any and especially
France again come way ahead of Japan, whose gains
in competitiveness turn out to be minor. Again, the
United States shows a dramatic loss, over 25 percent.
Sensitivity of the results to the choice of the base and
terminal periods. It could be argued that the base period
1974-80 goes too far back in the past, that it biases the
conclusions by including information of little relevance
today. Instead, it would be more instructive to see how
competitiveness has changed since the dollar began its
rise in 1980. Further investigation, however, suggests
that the principal qualitative conclusions of the above
analysis do not depend on the choice of 1974-80 as the
base period although the quantitative estim ates are
sensitive to such a choice.

•Of the two Morgan Guaranty indexes, this one takes into account
competition in third markets (World Financial Markets, August 1983).
Morgan Guaranty regularly publishes another index which shows
Japan to be slightly more competitive than Germany. This latter
index does not take into account competition in third markets.

Changes in Exchange Rates and Competitiveness: Aggregate Economy (1983) and Manufacturing (1983-11)
In percent; ( + ) indicates a loss of competitiveness

Changes from 1974-80 period
Measure of competitiveness

United
States

Changes from 1979-80 period

Japan

Germany

France

United
States

6.3
-1 3 .4

-1 4 .7
-2 5 .7

-4 0 .5
-3 3 .6

—

-2 9 .6
-2 0 .6

-3 1 .9
-2 1 .7

-3 5 .3
-28.1

Japan

Germany

France

- 6 .5
-1 6 .7

-2 8 .5
-3 0 .2

-4 4 .2
-3 5 .3

-2 7 .8
-1 6 .2

-3 5 .4
-2 9.1

-3 7 .5
-34.1

Bilateral competitiveness
Aggregate economy:
Nominal exchange rate (dollar price of the currency)
'Real exchange rate ........................................................

—

Manufacturing sector:
fExport price competitiveness ......................................
fLabor cost competitiveness ........................................

—

Effective competitiveness
Aggregate economy:
§Nominal exchange rate ................................................
Real exchange rate ......................................................

22 7
23.0

23.5
-1 .2

11.0
-7 .8

-2 7 .2
-1 9 .5

30.9
27.3

12.8
-2 .7

- 2 .0
- 9 .9

-2 6 .9
-18.1

Manufacturing sector:
Price competitiveness:
||IMF index .................................................................
IMorgan Guaranty index ........................................
"L a b o r cost com petitiveness..........................................

21.2
23.8
27.5

- 0 .6
- 5 .8
- 2 .5

- 5 .2
-7 .2
-9 .7

- 8 .5
- 7 .5
-1 3 .2

26.9
27.5
35.7

2.4
-4 .1
3.3

- 5 .2
- 6 .9
-1 2 .3

-1 1 .7
- 7 .9
-1 4 .8

Bilateral competitiveness refers to competitiveness of the country in question vis-i-vis the United States, while effective competitiveness
measures the competitiveness of a country against its major trading partners.
’ Wholesale price indexes (WPI) are used as deflators.
tOrganization for Economic Cooperation and Development (OECD) data on export unit values are used.
fOECD data on unit labor costs are used.
§Weights are derived from the International Monetary Fund (IMF) multilateral exchange rate model (MERM) for eleven major countries.
UChanges in trade-weighted wholesale dollar prices for manufactures.
fThis index is weighted by 1980 bilateral manufacturing trade weights, adjusted for supplier competition in third markets. Wholesale prices of
nonfood manufactures are used as deflators.
**This index of relative normalized unit labor costs for manufacturing represents the ratio of the indicator for the country to a weighted geometric
average of corresponding indicators for thirteen other industrial countries, all expressed in a common currency (dollar). The country indicator
is calculated by dividing an index of actual hourly compensation per worker by an index of output per man-hour adjusted so as to eliminate
estimated cyclical swings.

The table summarizes the data taking 1979-80 (the
period immediately preceding the rise of the dollar) as
the base period. This change in the base period does
not affect any of the above qualitative conclusions. The
loss of U.S. co m p e titive n e ss now looks even more
dramatic, and Japan’s gains in competitiveness appear
even less significant. Germany and France continue to
show substantial gains in competitiveness.
To provide a sense of the robustness of these con
clusions, two other sets of computations were carried
out by using period intervals unfavorable to our principal
conclusions. First, an attempt was made to see how
weak the yen was vis-a-vis the dollar in 1982-111, com
pared with the 1974-80 average. The 1982-111 quarter
was characterized by the weakest yen against the dollar

since 1980. We find that the real exchange rate of the
yen was 18 percent lower in 1982-111 than the 1974-80
average, whereas the mark and the franc were around
28 percent lower.
On a trade-weighted basis, the pattern is the same—
the yen shows a depreciation of 8 percent in real terms
against a depreciation of 12 to 14 percent for the other
two currencies. Thus, even when the yen hit its low of
the recent period, Japan’s gain in overall price com
petitiveness against the United States and on a tradeweighted basis was smaller than that of Germany and
France.
The other computation involves measuring the extent
of depreciation of the yen since 1978-111, the quarter
characterized by the strongest real effective exchange

FRBNY Quarterly Review/W inter 1983-84

rate of the yen during the last decade. Between 1978III and 1983, the yen depreciated around 28 percent
against the dollar in real terms. The corresponding
figure is about the same for the mark and is around 33
percent for the franc.
In real effective terms, however, the yen depreciated
around 13 percent—slightly less than the franc (15
percent) but slightly more than the mark (10 percent).
So, measuring from the yen’s peak in the floating period,
the yen is somewhat weaker in real terms than the mark
on a trade-weighted basis but not vis-it-vis the dollar.
However, it would be misleading to use this evidence
from a period as brief as one quarter to argue that the
yen has become a relatively weak currency. It is espe
cially unwarranted in this case because the yen was
unusually strong in the third quarter of 1978, a period
of great speculative instability in the exchange markets.
Evidence on competitiveness, therefore, does not
support the view that there is any overall weakness of
the yen or that the yen is undervalued. It does, however,
suggest the dollar is unusually strong. This implies that
the yen’s weakness against the dollar (or, more pre
cisely, Japan’s gains in competitiveness against the
United States) reflects the overall strength of the dollar
rather than any overall weakness of the yen.
This assessment implicitly assumes that the yen was
not seriously undervalued on average during 1974-80.
It could be argued, however, that the yen was already
undervalued during the broad sweep of 1974-80.
When the base period is shifted to the sixties, evi
dence on competitiveness does not point to any
undervaluation of the yen during 1974-80. In fact, con
ventional measures point to an overall loss of compet
itiveness for Japan during the seventies. It is possible
that the available measures of competitiveness are
faulty and do not capture the “ true” changes in com
petitiveness. Besides, as argued earlier, changes in
competitiveness—manufacturing or economywide—do
not tell the whole story about currency misalignments.
A broader macroeconomic perspective on the yen can
be obtained from considering the behavior of the Jap
anese current account.
Current account considerations
As explained earlier, current account imbalances by
themselves are, at best, an imperfect guide to detecting
the presence of currency misalignments. One needs a
judgment as to how large an external surplus or deficit
can be considered normal or sustainable, given the
pattern of the country’s saving and investment as well
as policy objectives over the medium term.
Thus, a persistent surplus in the Japanese current
account per se does not suggest any undervaluation of
the yen. In fact, many argue that by the early seventies

56 FRBNY Quarterly Review/Winter 1983-84

Japan became a natural capital-exporting country, and
a persistent underlying, if not actual, current account
surplus is normal for such a country.7
Although a current account surplus in and of itself
may not point to an undervalued currency, a rapidly
growing and/or persistently large surplus may. This was,
however, not the case in Japan during the seventies.
Japan’s current account was neither persistently in
surplus nor growing every year but went through wide
swings between surpluses and deficits. During 1974-80,
Japan had surpluses in three years and deficits in foul*
years.
Moreover, the average size of the Japanese surplus
was not very large. Japan had an average surplus in
its current account of only 0.13 percent of GNP during
1974-80. This was moderately higher than the U.S.
figure (0.03 percent) and substantially less than that of
Germany (0.57 percent) and that of the United States
during the sixties (0.73 percent).
Thus, the behavior of Japan’s current account during
1974-80 does not suggest that the yen was undervalued
during that period.
Does the yen appear undervalued now if viewed in
light of the current and future path of the Japanese
current account? Since the Japanese current account
surplus is expected to be much larger than that of any
other industrial country in 1983-84, it could be argued
that the yen is now more undervalued relative to other
major nondollar currencies in the sense that its current
account surplus is large, it is growing, and it is unsus
tainable.
A number of facts may help put this view in proper
perspective. First, the Japanese current account surplus
was around 0.5 percent of its gross domestic product
(GDP) in 1981-82, and it rose to around 2 percent of
its GDP in 1983. In 1984, the surplus is likely to rise
somewhat in dollar terms, but not significantly as a
percentage of GDP. The perception that the recent
increase in the Japanese surplus results from a boom
in Japan’s exports helped by an overall weakness of the
yen is, however, incorrect.
The dollar value of Japan’s exports increased 18
percent in 1981 (without a matching increase in
7After the midsixties, Japan turned from being a capital importer to
being a capital exporter in its long-term capital account. In addition,
since the first oil shock, the private investment rate as well as
overall economic growth has declined significantly in Japan. The
decline in the private saving rate, however, has been much smaller.
For fuller discussions, see R. I. McKinnon, “ Exchange Rate
Instability, Trade Imbalances, and Monetary Policies in Japan and the
United States”, in R Oppenheimer (ed.), Issues in International
Economics (Stocksfield, England: Orill Press Ltd., 1980), and M.
Yoshitomi, “An Analysis of Current Account Surpluses in the
Japanese Economy", in E. R. Fried, P. H. Trezise, and S. Yoshida
(eds.), The Future Course of U.S.-Japan Economic Relations (The
Brookings Institution, Washington, D.C., 1983).

imports), turning the 1980 current account deficit of over
$10 billion into a surplus of $5 billion. But in 1982, a
year in which the global criticism of Japanese trade
practices as well as its financial and exchange rate
policies became intense, Japan’s exports declined 8
percent in dollar terms, leading to a slight shrinkage of
its trade surplus. Both global recession and growing
protectionism appear to have contributed to this drop in
Japanese exports.
In 1983, exports recovered sharply, in both volume
and d o lla r term s. But the d o lla r value of Japanese
exports was still lower in 1983 than it was in 1981.
Viewed in light of Japanese trade performance in 1981
and 1982, what appears remarkable about the 1983
bulge in Ja pa ne se trade su rp lu s is not a surge in
exports but an unusually low level of dollar imports. This
decline in import value appears to have resulted from
cyclical weakness of the Japanese economy, lower oil
prices, and pure valuation effects of exchange rate
changes. Volume of imports was slightly higher in 1983
than in 1981, but the dollar value of imports was, in fact,
$17 billion lower in 1983 than in 1981.
Second, Japan’s 1983-84 surplus can also be viewed
as partly resulting from substantial liberalization of
international capital flows in Japan at the end of 1980
and hence may reflect a one-shot but slow portfolio
adjustment to an increase in capital mobility at a time
when U.S. yields have been very high. The sharp rise
in capital outflows resulting from this portfolio adjust
ment was partly responsible for the 1981-82 deprecia
tion of the yen and appears to be counteracting upward
pressure on the yen that may have come from large
current account surpluses of Japan in 1983-84.
Some have suggested th at Japan s till m aintains
capital controls that depress the value of the yen by
discouraging capital inflows. It is true that capital inflows
into Japan are still not completely free. For example,
foreign ownership of Japanese companies is still con
strained by many regulations. But there are regulations
that deter capital outflows as well. On balance, it is
difficult to establish that the remaining capital controls
significantly bias capital flows in the outward direction.8
Third, a shift has been taking place in the fiscal
position of Japan relative to the United States during
1982-84 (Chart 2). A measure of this relative fiscal shift
can be obtained from estimates of discretionary changes
in general government budget balances. According to
the December 1983 Econom ic Outlook of the OECD, the

8Many argue that the elim ination of all rem aining capital controls may,
in fact, encourage further net cap ital outflow s and weaken the yen in
the short run. See, for exam ple, W. A. Niskanen, "Issues and
N onissues’’, in E. R. Fried, P. H. Trezise, and S. Yoshida (eds ),
The Future Course o f U.S.-Japan E conom ic Relations.

Chart 2

B u d g e t D e fic it an d th e C u rre n t A c c o u n t:

1970-84
In p e rc e n ta g e of g ro s s do m e stic p ro d u c t

G e n e ra l g o v e rn m e n t fin a n c ia l b a la n c e
Percent
United States

Q " ^

1970 71 72 73 74 75 76 77 78 79 80 81 82 83 84
Est.
Percent
4
Japan

□ n

I_ _ I— 1 1 . 1

I_ _ J— L

1970 71 72 73 74 75 76 77 78 79 80 81 82 83 84
Est.

C u rre n t a c c o u n t b a la n c e
P e rc e n t
2

United States

0
-1

•2
—
-3a I— I— I— L

J__ I__ I__ L

J— I

1970 71 72 73 74 75 76 77 78 79 80 81 82 83 84
Est.
Percent
4
Japan

®IH
- 2l— L

J__ I__ I__ I__ I__ I__ L

1970 71 72 73 74 75 76 77 78 79 80 81 82 83 84
Est.

FRBNY Quarterly Review/W inter 1983-84

United States is expected to provide a fiscal stimulus
of 2.5 percent of GDP over 1982-84 while Japan is likely
to contract fiscally by a similar magnitude.
In light of this information, it does not seem unusual
that over the same period the current account would
deteriorate by 2 percentage points of GDP in the United
States while improving by 1.5 percent of GDP in Japan.
Fourth, the relative cyclical position of Japan and the
United States may also have temporarily aggravated
their current account imbalances. While the U.S.
economy is experiencing a rapid homegrown recovery,
Japanese growth has been modest and appears to have
been led by the pickup in export demand.
Finally, if the overall strength of the dollar is partly
responsible for the deterioration of the U.S. current
account deficit, then the improvement in Japan’s
external surplus can be traced partly to the same factor
as well. Japan is the second largest trading partner of
the United States which itself happens to be the largest
trading partner of Japan.
This close trade relationship dictates that some of the
U.S. deficit will show up as the Japanese surplus.
Between 1982 and 1983, the U.S. trade balance dete
riorated by $25 billion. Japan picked up less than $3
billion of this directly in the form of an increase in its
bilateral trade surplus with the United States. However,
the strength of the dollar may also have enabled Japan
to outcompete the United States in some third markets,
such as Western Europe and East Asia.
When these various special and cyclical factors are
considered together in assessing the behavior of the
Japanese current account, Japan’s large and growing
current account surplus does not point to any overall
weakness of the yen. A significant part of the surplus
appears to be resulting from temporary factors and from
the overall strength of the dollar.
Indeed, once these factors are taken into account,
Japan’s underlying current account surplus does not
appear to be very large or to be growing rapidly. A
persistent surplus in the Japanese current account may
or may not lead to an overall appreciation of the real
exchange rate of the yen. That will depend on the joint
future interaction between the private portfolio prefer
ences of international investors and the public policy
choices of national governments. But the recent
behavior of the Japanese current account does not
suggest that the yen is particuiary weak.
The dollar appears too strong, however, if the be
havior of the U.S. current account is analyzed. Whereas
the Japanese current account surplus as a share of its
GDP is not likely to rise appreciably in 1984, the U.S.
current account deficit is expected to rise from over 1
percent of its GDP in 1983 to over 2 percent in 1984.
Although a significant part of the deterioration of the

58 FRBNY Quarterly Review/Winter 1983-84

U.S. deficit can be accounted for by the relative cyclical
position of the United States and the decline in demand
for U.S. goods from the heavily indebted countries,
especially those in Latin America, the strength of the
dollar is still the single most important factor.0 If the
dollar remains at the current level, the U.S. current
account is expected to continue to deteriorate. By end1985, according to recorded statistics, the United States
is likely to turn from a net creditor country to a net
debtor country.
One implication of this shift in U.S. wealth will be a
gradual change in the composition of the U.S. current
account. Net investment income, which peaked at $33
billion in 1981, has already started declining and will
continue to do so in the foreseeable future. Thus, the
large service account surplus of the seventies will con
tinue to shrink in the eighties. To achieve a balanced
current account, the U.S. merchandise trade deficit must
be significantly smaller during the eighties than it was
in the seventies. Because of this dynamic effect, the
longer the current account deficit persists, the larger is
the depreciation of the real exchange rate of the dollar
required to eliminate the deficit.
But what is of greater concern about the present sit
uation is the fact that such large current account deficits
are unprecedented in recent U.S. history, and there is
a great deal of uncertainty as to how the dollar, and
more generally the world financial markets, will react as
the United States continues to demand a greater pro
portion of world savings.
Because of the current high return and the relatively
low political and economic risk that characterize U.S.
assets, international investors have so far financed the
growing U.S. current account deficit. However, as the
U.S. current account deficit grows bigger, the perceived
exchange rate risk of holding financial dollar assets may
begin to dominate the attraction of high U.S. yields, and
market assessment may increasingly turn against the
dollar. The weakening of the dollar since mid-January
may be reflecting such a change, but how far the dollar
will fall and how fast depends on how international
investors and speculators will assess and reassess their
expectations of the future course of the dollar in light
of new events and new information.
To sum up, an assessment of the behavior of the U.S.
current account supports the view that the current
strength of the dollar may not be sustainable indefinitely
but the precise dynamics of the dollar decline is still
impossible to predict.

•According to staff estimates, if the real effective exchange of the
dollar were held constant from 1980 to 1983 at its average 1973-80
level, the U.S. merchandise trade deficit would have been over $30
billion lower in 1983, all other things remaining the same.

C hart 3

R eal S h o rt-te rm In te re s t R a te s *
Jan uary 1980—N o v e m b e r 1983

R eal in te re s t ra te s
P e rc e n t p e r annum

P e rc e n t p e r annum
Japa n

1980

I I I 1 I I 11I 1M I I I I I I 1I
1981

R eal d iffe re n tia ls :

1982

P e rce n t p e r annum
10 ------------------------ -----------------Germany

France

1 9 80

1982

lo c a l m inus U.S. re a l in te r e s t ra te

P e rc e n t p e r annum
10 ------------------------ ------------------------------------------------ -----------------------f \
Ja pan

I ll I I

1983

1980

19 83

U nit ed Kingdom

19 82

1983

19 80

1982

1 9 83

♦ The ra tes sho w n are m o n th ly a ve ra g e s of da ily ra te s on m o ney m a rke t in stru m e n ts o f a b o u t n ine ty d a y s ’ m a tu rity a d ju s te d by
an e s tim a te of e x p e c te d in fla tio n . The rate fo r Japan is the d is c o u n t ra te on tw o -m o n th (p riv a te ) bills.

FRBNY Quarterly Review/W inter 1983-84

Interest rates
Another factor often cited in support of the view that the
yen is undervalued is the low level of Japanese interest
rates.10
A closer look at the facts, however, reveals that Jap
anese interest rates were not particularly low in 1983.
For example, throughout 1983, the three-month Euro
yen deposit rate (as reported in Morgan Guaranty’s
World Financial Markets) remained somewhat higher
than the comparable Euromark rate. Chart 3 provides
additional evidence by taking into account differential
inflation rates. Throughout 1983 the United States
clearly emerges as the country with the highest level of
short-term real interest rates, but Japan’s short-term real
interest rates were higher than those of most major
industrial countries.11 An examination of long-term real
rates suggest the same conclusion.
There could be disagreements on the details of the
measurement of these real interest rates, but the central
conclusion is clear: the United States has very high
interest rates but Japan does not have very low interest
rates. Therefore, the weakness of the yen vis-d-vis the
dollar cannot be attributed to low Japanese interest
rates but appears to be partly a result of high U.S.
interest rates.
Concluding remarks
The central conclusion of this article is that, from a
macroeconomic and trade point of view, the dollar is too
strong but the yen is not particularly weak. Japan’s
recent gains in competitiveness against the United
States have resulted from an overall strength of the

10A recent expression of this view can be found in the October 19,
1983 issue of The Economist (page 77). An article entitled “ How
Japan Cheapens the Yen" maintains that Japan “ contributes to the
yen’s weakness by still rigging interest rates".
"T h e rates used are monthly averages of daily rates on money market
instruments with maturity of about three months. Expected inflation
in month t is proxied by the twelve-month rate of CPI inflation in
month t + 6.

60 FRBNY Quarterly Review/Winter 1983-84

dollar and not from any overall weakness of the yen.
In other words, there is no special yen-dollar imbalance.
Evidence on changes in international competitiveness
as well as an assessment of the present and prospec
tive current account movements both point to such a
conclusion.
The conclusion that Japan’s present and prospective
current account surpluses are not excessively large
once temporary and cyclical factors are taken into
account can, however, be criticized from an international
point of view. Since Japan is the world’s second largest
economy, a current account surplus that may not be
large from its national point of view may be considered
to be unduly large by the rest of the world. Japan’s
trading partners may not wish to incur matching current
account deficits for economic or political reasons.
If this is the issue, it needs to be clearly spelled out.
This will open an international debate on how large a
Japanese current account surplus is considered unde
sirable by her trading partners and why. Is a Japanese
current account surplus on the order of 1.0 percent of
its GDP on average too large from the point of the view
of Japan’s trading partners? Should Japan run a bal
anced current account on average? Why is a small
Japanese surplus desirable? Is it to keep the forces of
protectionism in the United States and in Europe on a
leash?
Similar questions can be raised about the size of the
U.S. current account deficit. Since the dollar is the major
international currency and the United States is the
world’s largest and richest economy, a large and
growing U.S. demand on world savings may create
unique adjustment problems for the international finan
cial system.
These are important questions. But they do not focus
narrowly on exchange rates. Rather they direct public
attention to the broader issue of the international
implications of different mixes of monetary, fiscal,
financial, and trade policies. Ultimately, the question of
what constitutes correct values of exchange rates can
be understood only in that broader context.

Shafiqul Islam

In Brief
Economic Capsules
Was the 1980-82 Inflation
Slowdown Predictable?
The steep deceleration of inflation over 1980-82, even
after taking account of the depth of the recession, is
widely believed to have been unpredictable on the basis
of standard models of inflation. Some analysts argue
that the two successive recessions in 1980 and 1981 82 altered (or were altering) the response of inflation
to demand influences, possibly by lowering wage-price
expectations faster than had been indicated by statistical
models. As a result, the historical relationships in labor
and product markets would then have become less
useful in predicting inflation rates.
Looking to the future, with the progress toward elim
inating inflation as yet incomplete, many analysts fear
that inflation may flare up again as economic expansion
continues unless the changes in wage-price behavior
turn out to be durable. This has led some to argue that
substantial further reduction of inflation and long-run
price stability would be likely only if another recession
occurred in the next year or two.
Our research suggests that the steep deceleration in
inflation during 1980-82 was very much in line with the
historical relationship between inflation and its critical
determinants—wage-price expectations, aggregate
demand pressures, and productivity. That relationship,
as embodied in a conventional two equation wage-price
model, appears to have been quite reliable since around
I960.1
1The model used is a simpler version of that presented in A. Steven
Englander and Cornelis A. Los, "Recovery without Accelerating
Inflation?”, this Quarterly Review (Summer 1983). Wages are
determined essentially by price expectations and the unemployment
rate, while prices depend on labor compensation, productivity, and
cyclical factors. Food and energy prices are exogenous.

As for the medium-term outlook, no one can rule out
the possibility that the relationship between inflation and
its major determinants may be changing at present. But
so far there is no compelling evidence to suggest that
it is. Given that past relationships hold in the future, our
research suggests that a short third recession would not
likely drive inflation out of the system. Instead, in the
absence of a sharp break with past relationships, it
would take a long time to eliminate the last traces of
inflation. By contrast, a vigorous expansion of aggregate
demand that some forecasters are predicting could pull
wages and prices toward an accelerating course.
Chart 1 provides a historical perspective on the fore
casting performance of our conventional two equation
wage-price model beyond the estimation period. The
structural relations of the model are based on pre-1977
information. Use of post-1976 information is limited to
energy and food price shocks and to developments in
demand. On the whole, this conventional model is able
to pick up both the trends and turning points in inflation.
Over the forecast period the model underpredicts infla
tion very slightly (0.04 percentage point per quarter),
while the average absolute discrepancy between actual
and forecasted inflation is 0.67 percentage point.
The ability of the conventional model to track inflation
suggests that the relationships were stable in the late
1970s and early 1980s. Given information on future
demand pressures and special factors, a forecaster in
1976 or 1977 would have been able to predict both the
upturn and the downturn in inflation quite accurately.
And, indeed, even in the absence of any information on
energy and food price shocks, a forecaster could have
predicted an acceleration of inflation from about 5.5
percent in late 1977 to about 8 percent at the end of
1979 (Chart 1). Thus, the assertion that the acceleration
of inflation over 1978-79 and the subsequent deceler
ation were unpredictable is not true.
Turning to the future, we simulate the paths of

FRBNY Quarterly Review/Winter 1983-84 61

C h a rt 1

B u ild u p and R e d u c tio n of In fla tio n a ry
P re s s u re s *
P e rce n t

C hart 2

P rice In fla tio n u n d e r A lte rn a tiv e
U n e m p lo y m e n t R a te P aths
P e rce n t
^

I n f la tio n ra te s

____ ^

/ \ ^

---B aseline

I
I i i i I i i i I i i i I i i i I i i i I i I i 1i I i 1 M i 1 l 11 111 1 1
1983 1984 1985 1986 1 9 8 7 ,1 9 8 8 1989 1990 1991 1992

* G r o w th of p e rs o n a l c o n s u m p tio n e x p e n d itu re s d e fla to r.
^ A s s u m in g h is to ric a l v a lu e s o f e n e rg y and foo d p ric e s .
^ A s s u m in g c o n s ta n t re al value s of en ergy and fo o d p ric e s .

essentially the same model over the 1984-92 period
under three alternative unemployment rate paths (Chart
2).2 The baseline brings the unemployment rate grad
ually down to 6.5 percent (roughly the rate at which
inflation is stable, commonly called the natural rate) by
the third quarter of 1986 and keeps it there subse
quently. The expansion path shows the effect of an
unemployment rate path assumed to be 2 percentage
points below the baseline throughout the period. The
recession path raises the unemployment rate 2 per
centage points above the baseline in the first half of
1985 and gradually returns it to the baseline. After the
th ird q u a rte r of 1986 the b a se lin e and re ce ssio n
unemploym ent rates are the same.
The baseline infla tion path gradually stabilizes at
about 4-5 percent. In contrast, under the expansion
path, inflation would climb close to 10 percent over time.
Under the recession scenario, a sharp downturn would
produce an imm ediate lowering of the inflation rate to
2ln the estim ation and 1977-83 sim ulations, we did not constrain the
effects of, for exam ple, a 10 pe rcent increase in unit labor costs,
food prices, and energy prices to produce a 10 percent increase in
prices. We did im pose this restriction over the 1984-92 sim uiations
because there are strong theoretical reasons to expect it to hold
over the long run.

62 FRBNY Quarterly Review/W inter 1983-84

* A ssum e s the un em plo ym en t rate re m a in s 2 p e rc e n ta g e
p o in ts b e lo w the b a s e lin e .
"^A ssum es a sh a rp re c e s s io n o c c u rs in the firs t h a lf
o f 1985, w ith th e un em plo ym en t ra te re tu rn in g g ra d u a lly
to the b a s e lin e by 1986-111.

below 2 percent, but that would be reversed quickly with
any recovery. The net gain three or four years after the
recession, therefore, is fairly small.
The three inflation paths do not represent forecasts
of the actual economy, looking a few years down the
road. They are co nstructed to illu s tra te the co n s e
quences of different unemployment paths under certain
assum ptions, such as the absence of supply price
shocks and a constant natural rate of unemployment.
Any change in these assum ptio ns w ould a lter the
baseline as well as the other two inflation scenarios. For
example, another energy price shock would lead to a
rise in the baseline inflation, while continuing energy
conservation efforts could result in a lower inflation rate
over time. Similarly, the baseline inflation path might
trend downward if there is a decline in the natural rate
of unemployment due to changes in the composition of
the labor force and/or a rise in the trend growth rate in
labor productivity.
In any case, changes in unemployment rates produce
substantial short-term movements in inflation, but unless
they are sustained the long-term effect is limited. Thus,
for a short recession to lower inflation to zero and to
maintain price stability over an extended period, the
historical relationships would have to change. While

IN BRIEF— ECONOMIC CAPSULES

recent moderation in wage settlements and strike
activity hint at that possibility, it is too early to assess
their significance.
M. A. Akhtar, A. Steven Englander,
and Cornells A. Los

Surveys of Inflation
Expectations: Forward or
Backward Looking?
Many economists believe that surveys of price forecasts
(expectations) are “ forward looking” in that they make
use of information about current and future economic
policies and about developments of other variables. This
forward-looking feature presumably distinguishes these
surveys from the purely “ backward-looking” measures
of inflation expectations based on past experience. If
households and businesses hold forward-looking infla
tion expectations, they would pay less attention, in
making their decisions, to past movements of inflation
and more to current developments and changes in
economic policies. In this sense, the nature of expec
tations behavior is an important factor in determining the
outcome of economic policies.
The evidence seems to indicate, however, that survey
price forecasts are not forward looking or “ rational” in
the sense of incorporating information about current and
future periods. In fact, they appear to be lagging indi
cators of actual inflation rates; expectations contained
in them are essentially adaptive in character. The
inflation forecasting performance of these surveys is
roughly similar to forecasts based on recent past
experience with inflation.
Two well-known surveys of price expectations are
those conducted by Joseph Livingston of the Philadel
phia Inquirer and by the Michigan University’s Institute
for Social Research. We examined both these surveys
to determine (1) whether they are forward looking or
backward looking and (2) whether their ability to track
inflation is better or worse than purely backward-looking
inflation expectations based on past experience.
Charts 1 and 2 plot the Livingston and Michigan
survey inflation forecasts and actual consumer price
inflation for the corresponding periods. In both cases the
survey expectations of inflation rates lag actual inflation
rates, particularly in upturn phases of inflation. Two other

features of the charts are also inconsistent with the
forward-looking behavior. First, the survey forecasts
appear, on average, to underpredict inflation system
atically, although in the case of the Michigan survey this
underprediction is rather small. Second, they tend to
smooth the peaks and valleys of actual inflation rates,
which is reflected in the large discrepancy (i.e., average
absolute error) between the forecasted and the actual
inflation rates. As shown in the left-hand upper corner
of the charts, the downward bias of inflation forecasts
and the average absolute error are particularly signifi
cant in the case of the Livingston survey.
More rigorous analysis of data underlying the charts
confirms the impression that the Livingston and Mich
igan survey forecasts are not forward looking. If survey
expectations were forward looking or rational, they
would tend to be free of any serious systematic
underprediction or overprediction, and any errors
between actual and predicted inflation rates would be
completely random. Both surveys failed to meet these
conditions in our formal tests. The survey forecasts
systematically deviate from actual inflation rates and do
not incorporate all available information on past inflation
rates; that is, forecasts could have been improved by
making better use of past inflation experience.
In technical terms, we tested the forward-looking (or
rationality) hypothesis by estimating the following
equation:
p = a0 + a, pe + u
where p is the actual rate of inflation, pe is the survey
expectation of inflation, and u is an error term. If the
estimated value of a0 and a, are equal to 0 and 1,
respectively, this equation implies that the survey fore
casts would be unbiased predictions of future inflation.
In addition, such forecasts would incorporate all avail
able information from the past if the prediction errors
(u’s) are random, i.e., there is no serial correlation of
residuals. In this case, one would learn nothing from
past prediction errors in forecasting future inflation.
Estimates of the above equation for the Livingston
and Michigan inflation forecast data are reported in
Table 1. They indicate that survey expectations are not
forward looking.1 In particular, an F statistic test for the
'For the conversion of the price-level forecasts of the Livingston
survey into expected inflation rates, we followed the formal
procedures of Stephen Figlewski and Paul Wachtel in their article,
"The Formation of Inflationary Expectations", The Review of
Economics and Statistics (1981), pages 1-10. Because of the timing
of the availability of data to the economists in this survey, the
predicted rates of change are actually eight-month rates of change.
Compare with Edward M. Gramlich, "Models of Inflation Expectations
Formation”, Journal of Money, Credit and Banking (1983),
pages 155-73.

FRBNY Quarterly Review/Winter 1983-84 63

C h a rt 1

In fla tio n T ra c k in g P e rfo rm a n c e o f L iv in g s to n S u rvey D a ta
S e a s o n a lly a d ju s te d a n n u a l ra te s
P e rc e n t
16
-1 .2 0
A v e ra g e e r r o r (b ia s ):
A v e ra g e a b s o lu te e r ro r:
1.66

yi
12

/
/

10-

/ V / \

-2

l l i
1953

^ 1 1x
55

1 1 1 1 1 1 1 1 I ■ I I I . I
61
57
59
63

■1 ' i
65

64 FRBNY Quarterly Review/W inter 1983-84

/A
/

X
\

/
\

/
/

✓

M
\
K

V' U i
‘' V /
V

/
-'V

_ /A V
\

1
I

/
/

A c tu a l c o n s u m e r
p r ic e in fla tio n

| I

\ /

t x p e c ie u c o n s u m e r
p r ic e in fla tio n

I I I 1 l l
67
69

| | | | | 1 1 1 1 | | 1 I 111
71
73
75
77

J
79

. i <i > i
81
83

IN BRIEF— ECONOMIC CAPSULES

jo in t hypothesis that a0 = 0 and a, = 1 is rejected by
the data, and the residuals from the regressions appear
to contain a high level of serial correlation. Thus, both
survey forecasts of inflation are biased and do not make
efficient use of information from preceding periods in
forecasting future inflation.
Even though the two survey inflation forecasts are not
forward looking in a strong statistical sense, it could be
that they contain more information than purely backwardlooking inflation expectations and provide better fore
casts of actual inflation than the latter. An examination
of the data suggests, however, that this is not the case.
The survey in fla tio n fo re ca sts fo r any given period
appear to follow closely the actual inflation performance
over the preceding periods. It is as if the survey fore
casters are projecting the recent past experience into
the future; for example, shifting the position of the actual
inflation line forward by one period in Charts 1 and 2
tends to match up actual inflation rates more closely
with the survey inflation forecasts.
To pursue this analysis further, Table 2 compares the
L iv in g s to n and M ich ig a n fo re c a s ts w ith a sim ple
extrapolation and with a “ standard backward-looking”
fo re ca st. The e xtra p o la tio n fo re cast assum es that
inflation in the next half year remains unchanged from

the last half year, while the standard backward-looking
forecast is based on a distributed lag over the last two
periods.
The Michigan survey forecasts and the two backwardlooking forecasts deviate, on average, by slightly more
than 1 percentage point from actual inflation rates. The
three forecasts also exhibit a virtually identical small
downward bias or underprediction. By contrast, the
Livingston survey fo re ca sts are c o n sid e ra b ly less
accurate and more downward biased.
The predictive power— which measures, on a scale of
0 to 1, the ability of inflation forecasts to track actual
inflation— of all four forecasts in Table 2 is virtually the
same. When the predictive power is close to 0, there
is little evidence of fo re ca stin g ability, even if the
average error is quite small. The predictive ability of all
four forecasts is significantly less for the period after
1970 than for the longer period. All four forecasts also
become considerably less accurate in the 1970s. How
ever, there is no s ig n ific a n t change in the bias of
backward-looking forecasts, whereas underpredictions
from the two survey forecasts are more pronounced

Table 2

Bias, Accuracy, and Predictive Power of Survey and
Backward-Looking Forecasts

Table 1

Tests of the Forward-Looking or Rational
Expectations Forecasting Hypothesis

Statistics

Livingston
(econom ists)
1953-83

Backward-looking
forecasts:
M ichigan
(households)
1949-83

*No correction for serial
correlation:

.............................

R2 ............................................
SEE ........................................
D.W. ........................................
§F-statistic ...............................
IlC hi-squared statistic ..........

0.81
(2 .0 3 )f
1.11
(1.26)
0.72
2.04
1.41
22.34$
31 57 }

0.77
(3.11)*
0.86
(2 .6 6 )t
0.66
1.91
0.44
9.86*
175 36*

0.30
(2 .3 9 )t

0.96
(41.46)*

Correction for serial
correlation:
Rho

........................................

'A b so lu te t-ratios for a0 and Rho around 0 and for a, around 1
are reported in parentheses beneath the coefficients.
tS ig n ific a n t at 5 percent level.
^S ignificant at 1 percent level.
§For the joint hypothesis that ao = 0 and a, = 1.
HFor tests if the resulting residual series are white noise
(not serially correlated).

Periods of
observation

Survey forecasts:

Simple
extrapolation

Standard
backward

Livingston

Michigan

Average error (bias)
. -1 .2 0
Average absolute error
1.66
fPredictive power ........
0.73

-0 .0 4
1.13
0.80

-0 .0 3
1.18
0.78

-0 ,0 4
1.17
0.78

-0 .1 5
1.44
0.51

0.04
1.62
0.52

0.04
1.62
0.51

*1953-83:

*1970-83:
Average error (bias)
. -1 .4 5
Average absolute error
2.29
tPredictive power ........
0.45

Rates of inflation are annualized six-month rates of change. The
"simple extrapolation" assumes that inflation in the next half year
remains unchanged from the last half year. The “ standard
backward” looking forecast is a two-period distributed lag
(coefficients add up to one) on the rates of inflation.
’ The forecast errors are defined as e, = pe, - pt, where pt is the
actual rate of inflation and pe( is the corresponding expected rate
of inflation; the average (mean) error, i.e., bias, and the average
(mean) absolute error are expressed in percentage points.
fThe predictive power (coefficient of determination) indicates on a
scale from 0 to 1 how closely related the forecasted and
predicted rates of inflation were. It gives the percentage of
variation in the actual rate of inflation predicted by the forecast.

FRBNY Quarterly Review/W inter 1983-84

over the 1970-83 period than over the whole period.
In sum, our analysis suggests that survey inflation
expectations are not forward looking. On the contrary,
they follow actual inflation with a lag. The average
inflation forecasts provided by the Livingston and
Michigan surveys are not very accurate and frequently
tend to be downward biased. On the whole, their per
formance appears to be essentially similar to (or per
haps slightly weaker than) that of inflation expectations
exclusively based on immediate past data for inflation.
M. A. Akhtar, Cornelis A. Los,
and Robert B. Stoddard

Initial Claims: a Reliable
Indicator of Unemployment?
Every Thursday the Bureau of Labor Statistics (BLS)
reports the number of people filing for the first time for
state unemployment insurance benefits. These data on
initial claims are examined closely by many observers
of the economy for clues on the near-term change in
the unemployment rate. Notwithstanding the close
attention they receive, this analysis shows that initial
claims do not provide much information about the
direction of unemployment. However, they are a good
indicator of employment.
First, consider the simple relationship between the
monthly changes in initial claims and the unemployment
rate. Both often change in the opposite direction. Since
1948, initial claims and the unemployment rate moved
in the same direction less than 50 percent of the time.
Their inconsistent behavior has been more pronounced
during economic expansions. But, when the economy
was in a downslide, the relationship was a little better—
nearly two thirds of their monthly movements were in
the same direction.
Looking at the level of claims rather than their change
may be a more appropriate way to predict the direction
of unemployment. This is because, given the labor
force, unemployment tends to rise or fall when the
number of people who have just lost their job—mea
sured by initial claims—is relatively high or low. To
determine whether the level of initial claims is high or
low, the cyclical component of claims has to be sepa
rated from the trend component. This trend is related
to the long-run growth of the economy and the labor

66 FRBNY Quarterly Review/Winter 1983-84

force and can be estimated by regression analysis.1 In
expansions, claims should be expected to be below
trend, as fewer people become unemployed, while in
recessions claims should exceed trend.
Even relative to trend, initial claims do not predict the
unemployment rate satisfactorily. Only half of the time
did the unemployment rate rise when initial claims
exceeded their trend or fall when claims stood below it.
The direction of unemployment was predicted with a
little more success, at close to 60 percent, during
recessions. Thus, comparing claims with trend—while
for the most part better than observing their monthly
change—still is not particularly helpful.
In addition to claims, many observers monitor the
insured unemployment rate to predict overall unem
ployment. The BLS constructs the insured unemploy
ment rate partly from initial claims data. The insured
unemployment rate is not a dependable indicator of
overall unemployment either. Its success rate in pre
dicting the direction of unemployment has been only
slightly above 50 percent, although somewhat better
during recessions.
While initial claims do not consistently predict the
direction of unemployment, they may provide a clue
around cyclical troughs as to when the unemployment
rate will fall. During most of the postwar recessions,
claims have peaked two or three months ahead of the
unemployment rate. Thus, they may serve as a leading
indicator of the decline in unemployment. Around cycli
cal peaks, in contrast, claims frequently have begun to
climb simultaneously with unemployment and by only a
small, normally inconsequential amount. As a result,
they may not be a useful indicator as an expansion
nears its end.
To be sure, there have been occasions during reces
sions when initial claims turned down but then resumed
their climb before the economy began to recover. In the
last recession, claims declined fairly sharply but tem
porarily in June and July 1982, well ahead of the peak
in unemployment. Thus, even during recessions, a drop

1The trend in claims was calculated in two stages. First, the following
estimated equation yielded a trend level of about 0.4 percent of the
labor force. Second, the predictive performances of other, parallel
trends were examined. The most successful predictions were
obtained when initial claims were compared with a level 15 percent
above the estimated trend at any point in time. This adjusted trend
is used in this analysis. It stood at 480,000 in 1983.
IC = 0.003762 + 0.0005327(U-UN) + 0.841 error_,
LF
(21.7)
(9.2)
(17.9)

R2 = 37.8

where IC = initial claims, LF = labor force, U = unemployment
rate, UN = natural rate of unemployment (numbers in parentheses
are t-statistics). For a discussion of the natural rate of
unemployment, see A. Steven Englander and Cornelis A. Los,
"Recovery without Accelerating Inflation?”, this Quarterly Review
(Summer 1983).

IN BRIEF— ECONOMIC CAPSULES

in initial claims over two or three months may incorrectly
signal a topping in the unemployment rate.
Although initial claims do not tell us much about the
unemployment rate, they are useful in predicting the
direction of payroll employment. Admittedly, the simple
relationship between the monthly changes in claims and
employment is about as weak as that between claims
and unemployment; in only a little more than half the
time did they move in the same direction. Nevertheless,
when claims are compared with trend, they serve as a
good guide to the direction of employment. In more than
80 percent of the time, employment has fallen when
initial claims have been greater than trend or has risen
when claims have been less than trend. This rate of
success has been evident in both expansions and
recessions.
Claims predict employment best when the economy
is not near a cyclical turning point. About half of the
months in which claims relative to trend flashed incor
rect signals occurred within five months of cyclical peaks
or troughs. However, around cyclical troughs, claims
have tended to turn around ahead of employment,
although the lag has been more variable than for
unemployment. Thus, claims may serve as a leading
indicator at those times. Around many past cyclical
peaks, though, claims and employment have often
reversed direction simultaneously—with the change in
claims being quite small—making claims less insightful.
In conclusion, should anyone pay attention to initial
claims? The bottom line of this analysis is that for the
most part the answer is no when the purpose is to
predict the direction of unemployment, but yes when the
purpose is to predict the direction of payroll employ
ment. In one important case, however, it may pay to
watch the pattern of initial claims to predict the direction
of unemployment. This is at times when a recession
seems to be ending and questions arise as to when the
unemployment rate will start to fall. Although the true
peak in initial claims may be difficult to determine,
claims usually have peaked about two months ahead of
the unemployment rate. Thus, at those times when
many observers strive to pinpoint an upturn in the
economy, initial claims may be of some help.

Carl J. Palash

NOW Accounts and the
Seasonal Adjustment of M-1
In general, adjusting economic statistics to remove
purely seasonal influences is an imperfect exercise at
best. It is even more difficult a task when a long-time
series upon which to base estimates of changing sea
sonal patterns is not available, or when a given time
series contains components with different underlying
seasonal patterns, but the relative size of the compo
nents cannot be determined.
In recent years, these types of seasonal adjustment
problems have been quite serious for M-1. This aggre
gate includes a large and growing component, NOW
account deposits, which were not available on a
nationwide basis before 1981. By statistical standards,
that is much too short a time period to estimate a reli
able seasonal pattern for NOW accounts. Usually, five
years or more of data are required to estimate seasonal
adjustment factors.
Moreover, NOW accounts are not like the other
deposit components of M-1, because they can fulfill two
distinct functions. They can be used for transactions
purposes as well as a savings vehicle. Indeed, in 1981
when nationwide NOWs were introduced, it was esti
mated that about 25 percent of the initial flow into
NOWs came from sources outside M-1, primarily from
passbook savings accounts. Therefore, it would appear
incorrect to adjust NOW accounts using the same sea
sonal factors that are appropriate for demand deposits.
This, of course, raises the question of whether it
would be better to adjust NOW accounts by using a
weighted average of seasonal factors for demand
deposits and for savings balances (which are part of
M-2), where the weights would be in proportion to the
degree to which consumers use NOW accounts for
savings purposes. While that might sound good in
theory, in practice it is impossible to know to what extent
NOWs are used as a transaction vehicle and to what
extent as savings accounts. Nonetheless, some calcu
lations can be made to illustrate how serious a problem
NOW account deposits could pose in the seasonal
adjustment of M-1.
The following equation uses weighted average sea
sonal factors to adjust NOWs.
OCD.SA = -------------OCD^------------- +
xSAV.SFCB+ (1 -x)DDA.SF
_________ OCDT_______
xSAV.SFT+ (l-x)DDA.SF

FRBNY Quarterly Review/Winter 1983-84 67

M-1 under Alternative Seasonal Adjustment Procedure
A nnualized one-m onth rates of grow th, in percent

1983
January
February
M arch . .
A pril . . .
May
June ___
July
A ugust .
September
October
November
December

M-1
prior to
revisions
9.8
22.4
15.9
-2 .7
26.3
10.2
8.9
2.8
0.9
1.9
0.9
6.5

t{Standard
deviation)
9.0
tCorrelation with
revised M-1 .............

Savings balance fraction
in alternate procedure
25%
50%
75%

Revised
M-1

11.8
18.6
16.9
5.8
14.2
10.7
7.5
3.0
2.6
3.5
3.5
5.8

10.8
15.4
16.9
7.7
12.0
11.4
7.5
2.8
3.7
4.0
4.6
7.4

10.2
12.1
16.7
9.7
9.8
12.1
7.5
2.8
4.7
4.4
5.6
9.2

13.9
14.6
13.0
1.9
20.5
8.8
8.5
5.8
2.9
4.6
3.2
6.3

4.6

4.0

5.7

0.77

0.80

0.81

‘ Adjusted for February 1984 revisions to seasonal factors and does
not incorporate benchmark revisions
tin percentage points.
^Correlation between the changes to M-1 suggested by the
alternative adjustment procedure and the published changes to
M-1 derived from the revised 1983 seasonal factors.

where:
OCD.SA = A lte rn a tiv e se a so n a lly a djusted interestbearing checkable deposit component of
M-1.
OCDCB= Interest-bearing checkable deposits at com
mercial banks.
OCDT = Interest-bearing checkable deposits at thrift
institutions.
SAVSFCB =S easonal factor for savings deposits at all
commercial banks.
SAVSFT = Seasonal factor for savings deposits at thrift
institutions.
DDA.SF = Seasonal factor for demand deposits.
x = P o rtio n of in te re st-b e a rin g M-1 deposits
assumed to reflect savings balances.

growth than originally reported in 1983. For example, the
standard deviation of the monthly M-1 growth rates for
1983 was 9 percentage points as first reported and 5.7
p ercentage points a fte r the annual seasonal fa c to r
revision. The assumption that something in the range
of 25 to 75 percent of NOWs are savings resulted in
standard deviations of 5.6 percentage points to 4.0
percentage points.1
Furthermore, the impacts of the seasonal factor revi
sions made recently for 1983 and of the alternative
seasonal factors calculated here are highly correlated,
suggesting that the standard seasonal adjustm e nt
process is beginning to pick up some of the changing
character of M-1, as a greater percentage of it is com
posed of interest-bearing accounts that can also be
used for savings purposes. Picking up some of the new
seasonal characteristics of M-1 is only one step in
understanding the changing nature of M-1 now that it
contains a savings component. We still do not have
enough experience to understand its cyclical behavior.
All in all, the lesson from this exercise seems clear.
M on eta ry d ata in g e n e ra l m ust be a s se sse d w ith
extreme care, particularly over intervals shorter than one
year. But even greater caution should be exercised
when looking at seasonally adjusted M-1 because it is
no longer made up exclusively of transactions deposits.
NOW accounts— since they pay explicit interest— are
likely to be used by consumers for savings purposes as
well but to an unknown degree, not only in a seasonal
sense, but over the business cycle as well.
1A sm aller standard deviation is not necessarily an indica tor of better
seasonal adjustm ent. The orig inally reported M-1 grow th rates for
1983, however, were so volatile (a range of - 2 . 7 to 26.3 percent)
that it seemed quite natural to investigate w hether alternative
seasonal adjustm ent procedures would reduce the volatility in 1983.

Sandra C. Krieger

The resultant impacts on the M-1 growth rates for
1983 are p resented in the table. No m atter which
weights are chosen in the 25 percent to 75 percent
range (the assumed share of savings in NOWs), this
procedure yields a sm oother pattern for monthly M-1

68 FRBNY Quarterly Review/W inter 1983-84

IN BRIEF— ECONOMIC CAPSULES

FASB 52: Corporate
Response and Related
Foreign Exchange Market
Effects
U.S. multinational corporations are in the midst of
responding to the second major change in foreign
exchange accounting rules in the last nine years. The
first change occurred in October 1975 when the Finan
cial Accounting Standards Board (FASB)1 issued
“ Statement of Financial Accounting Standards Number
8” (FASB 8). The FASB had felt compelled to develop
a standard set of rules to replace the diverse accounting
procedures being used by U.S. corporations following
the move to generalized floating exchange rates in
1973.
But FASB 8 almost immediately generated controversy.
In particular, it was criticized for producing a distorted
picture of a multinational company’s underlying eco
nomic situation. Thus, after much debate and a thorough
review of various alternatives, the FASB adopted in
December 1981 a vastly revised set of accounting rules
embodied in FASB 52. Corporations were required to
implement the new statement for fiscal years beginning
on or after December 15, 1982.
To determine how U.S. corporations are responding
to the significant changes of FASB 52, corporate
treasury personnel at sixteen of the largest U.S. indus
trial companies and at one of the top ten U.S. diversified
service firms were contacted and questioned about
corporate foreign exchange hedging practices, borrowing
strategies, and other matters related to managing for
eign exchange risk. In most cases, corporate respon
dents did not confine their remarks to activities of their
own firms. Instead, based on experience and conver
sations with their counterparts at other companies, they
spoke more generally about their views regarding the
reaction of U.S. corporations to FASB 52.
In addition to the corporations, corporate advisory
personnel at seven of the top twenty U.S. commercial
banks and at one foreign bank operating in New York
were contacted. Altogether, a total of twenty-seven
people were contacted.
Based on these conversations, several findings

’ The FASB is a private rule-making body in the United States with the
responsibility of setting forth generally accepted accounting
principles.

emerge about the response of U.S. corporations to the
adoption of FASB 52.
Most of those asked said that many corporations
which had hedged or offset balance sheet exposure
under FASB 8 had scaled back, or ended altogether, this
practice following the adoption of FASB 52. Balance
sheet exposure results from a mismatch between those
foreign-currency-denominated assets and liabilities
which must be translated into U.S. dollars at exchange
rates prevailing on the date of the balance sheet. The
majority said that overall corporate activity in the
exchanges had declined, although not everyone attrib
uted this to the new accounting rules. But the bulk of
respondents thought that the volume of foreign
exchange business done by corporations in the forward
market had dropped under FASB 52. By contrast, well
over half believed corporations had become more active
in the foreign exchanges during the time of FASB 8.
Virtually all of those questioned said that some com
panies, including many deemphasizing or ending the
practice of hedging balance sheet exposure, are now
focusing more on transaction and/or economic exposure.
Transaction exposure results from the possibility that
exchange rates might change between the time a
transaction is agreed to (e.g., when a sales contract is
signed) and the time when it is actually settled (e.g.,
after the goods are delivered). Economic exposure, a
broader and more forward-looking concept, stems from
the possibility that the firm’s future cash flow will be
affected by exchange rate changes.
The change in hedging strategy by many U.S. com
panies seems to have been accompanied by a shift in
corporate borrowing patterns. A m ajority of the
respondents thought that under FASB 52 some U.S.
firms are more willing, or had moved, to arrange more
foreign currency loans than before. About half felt
that some corporations had relied more on dollardenominated and less on foreign-currency-denominated
borrowings under FASB 8.
A majority of those asked believed that many U.S.
companies had already centralized, or were moving
toward centralizing, the management of foreign
exchange exposure. They felt that many corporations
use, or are looking at the possibility of using, a system
of netting exposures. Netting involves collecting at a
central location information about payments and receipts
between the different entities within a corporation. The
central office calculates a net receipt or payment figure
for each entity and issues specific payment instructions,
which result in funds flowing from entities with net
payments to those with net receipts. Netting lowers
transaction costs by reducing the number of payments
between entities within the corporation.
Most of the contacts reported that corporations also

FRBNY Quarterly Review/Winter 1983-84 69

have used, or were considering, foreign exchange
options contracts as part of their overall strategy to
manage exchange rate exposure better. However, actual
corporate use of foreign exchange options apparently
has not become very widespread as yet. And few
respondents felt that corporations were using foreign
exchange futures contracts as a tool for managing for
eign exchange exposure.
Statistical evidence to support most of these findings
is unfortunately sparse. However, the respondents’ belief
that corporate activity in the forward foreign exchange
market has dropped following the adoption of FASB 52
receives support from the latest foreign exchange turn
over survey conducted by the Federal Reserve Bank of
New York. The survey shows that outright forward
transactions reported by U.S. banking institutions with
nonfinancial customers declined 16 percent to $8.8 bil
lion in April 1983 from $10.5 billion in March 1980 even
as total foreign exchange turnover reported by the
banks rose about 44 percent. While FASB 52 may not
be the only reason for this decline, it seems to have
played an important role.
Michael D. Andrews

70 FRBNY Quarterly Review/Winter 1983-84

IN BRIEF— ECONOMIC CAPSULES

August-October 1983 Interim Report
(This report was released to the Congress
and to the press on December 7, 1983.)

Treasury and Federal Reserve
Foreign Exchange Operations
Early in August the dollar moved up sharply, reaching
a 91/2-year high against the German mark and a record
high on a trade-weighted basis. For much of the balance
of the period, market participants expected the dollar to
retreat substantially from those levels, and the dollar did
depreciate gradually through early October. But, buoyed
by the effects of greater than expected strength in the
domestic economy and political turbulence internation
ally, the dollar strengthened again during the remainder
of October to close the period little changed from its
end-July levels against most major foreign currencies.
The decline in the dollar through early October was
influenced by widespread predictions of a slowing of the
recovery and an easing of money market conditions in
the United. States. Many forecasters doubted that the
domestic economy, which had advanced at a strong 9.7
percent rate in the second quarter largely on the basis
of a rebound in consumer expenditures and residential
construction, could show sustained growth in the face
of the strong dollar and high real interest rates. More
over, growth of the narrowly defined monetary aggre
gate, M-1, had decelerated sufficiently to move within
its monitoring range for the first time this year, and price
data indicated that inflation remained relatively mod
erate. Consequently, many market participants came to

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 for Foreign Operations of the System Open
Market Account.

the view that the Federal Reserve would take this
opportunity to exert less pressure on bank reserves, and
U.S. financial markets developed a considerable sense
of optimism from late August through early October.
Short-term interest rates declined by some 3U per
centage point. Yields on longer dated securities also fell,
but by smaller margins. Some market participants were
concerned that, if interest rates should continue to ease,
financing the widening U.S. current account deficits
could become more difficult.
However, the U.S. economy continued to grow faster
than many observers had anticipated. To be sure,
housing starts and retail sales temporarily weakened
during the summer, and the release of these statistics
kept alive expectations of a significant slowing later in
the year. But demand in other sectors, especially busi
ness fixed investment and inventories, was strong
enough to support major gains in industrial production
and employment. During the third quarter, GNP regis
tered a growth rate of some 7.7 percent in real terms,
and by October it was clear that the economy retained
considerable momentum as it proceeded into the fourth
quarter.
As the economy remained buoyant, the scope for
further declines in interest rates gradually came to be
seen in the market as limited. After mid-October most
U.S. interest rates edged higher, reinforced somewhat
by uncertainties over the credit market implications of
the lack of Congressional action to raise the government
debt ceiling. In addition, the rapid reemployment of idle

FRBNY Quarterly Review/Winter 1983-84 71

Table 1

Federal Reserve Reciprocal
Currency Arrangements

C h a rt 1

T h e D o lla r a g a in s t S e le c te d
F o r e ig n C u rre n c ie s

In millions of dollars

Institution

Amount of
facility
October 31,
1982

Amount of
facility
October 31,
1983

250
1,000
2,000
250
3,000
2,000
6,000
3,000
5,000

250
1,000
2,000
250
3.000
2,000
6,000
3,000
5,000

700
325
500
250
300
4,000

700
*
500
250
300
4,000

600

1,250

30,425

30,100

Austrian National Bank .....................
National Bank of B e lg iu m .................
Bank of Canada .................................
National Bank of Denmark ...............
Bank of England ...............................
Bank of F ra n c e ...................................
German Federal Bank .....................
Bank of Italy .......................................
Bank of Japan ...................................
Bank of Mexico:
Regular facility ...............................
Special facility ...............................
Netherlands Bank .............................
Bank of Norway .................................
Bank of S w e d e n .................................
Swiss National Bank .........................
Bank for International Settlements:
Swiss francs-dollars .....................
Other authorized European
currency-dollars .............................
Total

....................................................

P e rc e n t

P e rc e n ta g e c h a n g e o f w e e k ly a v e ra g e bid ra te s fo r
d o lla rs fro m the a v e ra g e ra te fo r th e w e e k of
S e p te m b e r 2 7 -O c to b e r 1, 1982. F ig u re s c a lc u la te d
fro m New Y o rk noon q u o ta tio n s .

'Facility, which became effective August 30, 1982, expired on
August 23, 1983.

Table 2

Drawings and Repayments by the Bank of Mexico under Special Combined Credit Facility
In millions of dollars; drawings ( + ) or repayments ( - )

1983
III

Outstanding
October 31,
1983

56 0

-2 6 9 .0

+ 122.3

-1 0 4 .0

-4 9 6 .0

+ 190.0

-1 6 0 .0

-7 6 5 .0

1982
IV

1983
I

Federal Reserve special facility
for $325 million ..........................................

46.0

+ 211.2

+ 67.8

United States Treasury special
facility for $600 million ...............................

85.5

+ 392.2

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

131.5

+ 603.5

Drawings on

Data are on a value-date basis. Because of rounding, figures may not add to totals.
'Facility expired and outstanding drawings were repaid on August 23, 1983.

1983
II

Outstanding
October 1,
1982

FRBNY Q uarterly Review/W inter 1983-84

capacity began to raise some questions among market
participants over the medium-term outlook for monetary
policy, particularly in view of the continuing fiscal stim
ulus provided by a large government deficit. As the
outlook for U.S. interest rates and the economy shifted
during October, market professionals moved to cover
large short-dollar positions that they had built up earlier.
With the outlook for the U.S. economy remaining
stronger than for those abroad, capital continued to flow
into U.S. stock and bond markets. Also adding support
fo r the d o lla r w ere “ s a fe -h a v e n ” c o n s id e ra tio n s
p ro m p tin g c a p ita l flo w s in to the U nited S ta te s in
response to events that heightened international ten
sions during the period. Market participants were mindful
that such episodes had generated significant capital
inflows at tim es during the past year, and talk of safehaven influences resurfaced on September 1 following
report that the Soviet Union had downed a Korean air
liner. But that particular incident did not elicit a strong
exchange rate reaction. Later in the period, however,
intensified fighting in Lebanon, escalation of threats in
the Iran-lraq war, and a U.S. landing in Grenada were
among the events that did have a more noticeable
impact on the dollar and thereby enhanced the per
ceived risk of positioning against the U.S. currency.
Exchange market reaction to announcement of record
U.S. trade and current account deficits was subdued,
as the deficits were being easily offset by the continuing
capital inflows. Although the statistics confirmed the
existence of deficits of unprecedented size— with one
monthly trade deficit over $7 billion— the current account
issue faded into the background as a market factor,
especially when the Septem ber trade deficit showed a
reassuring narrowing.
The only currency to advance significantly against the
dollar over the three-m onth period as a whole was the
Japanese yen, buoyed by Japan’s outstanding trade and
price perform ance. The yen also benefited from the
m arket’s perception that the Japanese authorities were
committed to supporting the yen. Bank of Japan Gov
ernor Mayekawa made clear that the exchange rate was
an im portant consideration in the tim ing of the V 2 per
centage point discount rate cut which finally took place
on October 21 in conjunction with announcement of a
six-point econom ic stim ulus package. The Japanese
authorities stated that they remained ready to intervene
in the exchanges when necessary to defend the yen,
and in fact they did sell dollars in the market on several
occasions during the period. Following close consultation
with the Bank of Japan as the yen weakened late in the
period, the U.S. authorities also purchased a modest
amount of yen in a joint operation with the Japanese
central bank. These operations began on October 31
and continued the next day. In total, the U.S. authorities

purchased $29.6 m illion equivalent of yen, an amount
that was split evenly between the Treasury and the
Federal Reserve.
As detailed in the previous report covering the period
through end-July, the U.S. authorities also intervened in
the exchanges on four occasions during six business
days, buying both Japanese yen and German marks, in
coordinated operations that began on July 29 and lasted
through August 5. These operations together brought
the total of U.S. authorities’ intervention in the exchange
market from July 29 through November 1 to $283.6
million equivalent, split equally between the Treasury
and the Federal Reserve. Of this amount, $101.0 million
equivalent "was in Japanese yen and $182.6 million
equivalent was in German marks.
In other operations during the three-m onth period,
Mexico fully repaid the remaining portion of its special
combined credit facility. On August 15, Mexico prepaid
outstanding swaps of $100.8 million to the Treasury and
$54.3 m illion to the Federal R eserve. D raw ings of
$395.3 million and $214.8 m illion were repaid to the
Treasury and the Federal Reserve, respectively, upon
maturity on August 23, and the facility then expired. This

FRBNY Quarterly Review/W inter 1983-84

Table 3

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

Exchange
Stabilization
Fund

General
account

-0-

Valuation profits and
losses on outstanding
assets and liabilities
as of October 31, 1983 . . -771.9

-786.2

-0-

•
Period
August 1 through
October 31, 1983 ........

Data are on a value-date basis.

facility had originally consisted of $600 million from the
Treasury and $325 m illion from the Federal Reserve. It
was provided in cooperation with other central banks,
which together extended credit to the Bank of Mexico
totaling $1.85 billion.
During the past year, the Treasury had participated,
along with other nations, in providing liquidity support
to the Bank fo r In te rn a tio n a l S e ttle m e nts fo r c re d it
facilities that the BIS provided to the Central Bank of
Brazil and to the National Bank of Yugoslavia. This
su p p o rt to o k the form of the Treasury, through the

74 FRBNY Q uarterly Review/W inter 1983-84

Exchange S tabilization Fund (ESF), agreeing to be
substituted for the BIS in the event of delayed repay
ments. By the end of the period, contingent com m it
ments on behalf of Brazil remained at $500 m illion and
on behalf of Yugoslavia were reduced to $16 million.
Both commitments expired as the credits were repaid
after the close of the reporting period.
In the period from August through October, the Fed
eral Reserve, the ESF, and the Treasury general account
realized no profits or losses from exchange transactions.
As of October 31, cumulative bookkeeping, or valuation,
losses on outstanding foreign currency balances were
$771.9 million for the Federal Reserve and $786.2 m il
lion for the ESF. (Valuation gains and losses represent
the increase or decrease in the dollar value of out
standing currency assets and liabilities, using end-ofperiod exchange rates as com pared w ith rates of
acquisition.) These losses reflect the fact that the dollar
strengthened since the foreign currencies were pur
chased.
The Federal Reserve and the Treasury invest foreign
currency balances acquired in the market as a result of
th e ir fo re ig n e xcha ng e o p e ra tio n s in a v a rie ty of
instruments that yield market-related rates of return and
that have a high degree of quality and liquidity. Under
the authority provided by the M onetary Control Act of
1980, the Federal Reserve invested some of its foreign
currency resources in securities issued by foreign gov
ernments. As of October 31, the Federal R eserve’s
holdings of these securities were equivalent to $1,618.6
million. In addition, the Treasury held the equivalent of
$2,318.8 million in these securities as of end-October.

NEW PUBLICATION
A table— Depository Institutions and Their Regulators— is
now available from the Federal Reserve Bank of New York.
The grid-like form shows the responsibilities that national
and state regulators have in ten areas—from branching to
consumer protection—for a variety of depository institutions.
The table contains footnotes summarizing laws and rulings
affecting the activities of regulators and depository institu
tions. It is intended to provide easy reference for bankers
and advanced students of banking.
Single copies of the 11V2" x 22 V2" foldout table can be
obtained free. Quantities for classroom use are also avail
able free when ordered from a university address. Write to
the Public Information Department, 33 Liberty Street, New
York, N.Y 10045.

FRBNY Quarterly Review/Winter 1983-84

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 subscribers
are sent via third- and fourth-class mail, respectively. All copies for Eastern Hemi
sphere 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

FRBNY Quarterly Review/Winter 1983-84