The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
Federal
Reserve Bank of
NewYork
Quarterly Revie
Spring 1990
Volume 15 No. 1
1
Reforming the U.S. Financial
System: An International Perspective
15
Event Risk Premia and Bond Market
Incentives for Corporate Leverage
31
Understanding International
Differences in Lpverage Trends
43
M onetary Policy and Open
Market Operations during 1989
66
Treasury and Federal Reserve
Foreign Exchange Operations
This Quarterly Review is published
by the Research and Statistics Group
of the Federal Reserve Bank of New
York. Statement of E. GERALD
CORRIGAN, President of the Bank,
on reform ing the U.S. financial
system begins on page 1. Among the
sta ff members who contributed to
articles in this issue are STEVEN A.
ZIMMER (on event risk prem ia and
bond m arket incentives for corporate
leverage, page 15); and ELI M.
REMOLONA (on understanding
international differences in leverage
trends, page 31).
A rep ort on monetary p o licy and
open m arket operations in 1989
begins on page 43.
A quarterly report on Treasury and
Federal Reserve foreign exchange
operations for the period February
through A p ril 1990 starts on page 66.
Reforming the U.S. Financial
System: An International
Perspective
Good morning, Mr. Chairman. It is a pleasure to appear
again before this Committee to discuss the pressing
issues facing the U.S. banking and financial system
and to stress the need to promptly enact broad-based
legislation that would reform and modernize the struc
ture of the U.S. financial system and many features of
the supervisory arrangem ents associated with the
operation of that system. As has been the case in the
past, I want to state at the outset that the views I will
express today are my own and, as such, they should
not be construed to reflect the views of the Federal
Reserve Board or the Federal Reserve System as a
whole.
While I will, over the course of my remarks, address
all of the issues raised in your letter of invitation, I
intend, for the purpose of orderly presentation, to con
sider these issues in a different sequence than outlined
in your letter, starting with the international side of the
equation. Since the scope of the material to be cov
ered this morning is very broad, I have attached to my
statement a number of appendixes which I hope will be
of value to the Committee and its staff in the effort to
gain a broad perspective on the complex set of issues
bearing on how we can best adapt the structure of the
U.S. financial system over time.
Banking and financial structure abroad
The legal and institutional framework within which
Statement by E. Gerald Corrigan, President of the Federal Reserve Bank
of New York, before the United States Senate Committee on Banking,
Housing, and Urban Affairs, on Thursday, May 3, 1990. The appendixes
referred to in this statement are available upon request from the Public
Information Department of the Federal Reserve Bank of New York.
banking and financial systems operate in the major for
eign industrial countries is of importance to the United
States for a variety of reasons. Two are particularly rel
evant in the immediate context of this hearing: first,
international differences in banking structure can have
important implications for the competitiveness of U.S.
institutions both here in the United States and around
the world; and second, international differences in
financial structure can introduce complex and poten
tially dangerous elements of tension into cross-border
relationships as they pertain to the rights and privi
leges of banks and other financial firms to operate
across national boundaries.
International differences in financial structure can
introduce complex and potentially dangerous
elements of tension into cross-border relation
ships as they pertain to the rights and privileges
of banks and other financial firms to operate
across national boundaries.
In order to provide the Committee with an overview
of these structural arrangements abroad, the first
appendix to this statem ent provides a broad —and
admittedly oversimplified— summary of banking struc
tures in the Federal Republic of Germany, Japan, and
the United Kingdom as well as a brief description of
the main thrust of the Second Banking Directive that
will govern banking activities in the European Commu
nity. While I will not repeat the thrust of that appendix, I
would stress the following major points:
FRBNY Quarterly Review/Spring 1990
1
• First, as a rough approximation, financial structure
in the United Kingdom is quite similar to the finan
cial structure in other countries with close historic
ties to the United Kingdom such as Canada and
Australia.
• Second, the prevailing structure in West Germany
is very sim ilar to that in the Netherlands and
Switzerland and, to a somewhat lesser extent,
other industrial countries in continental Europe
such as France and Italy.
Looked at in that light, there are, as a rough approx
imation, three operational models of banking structure
in the industrial world today, as follows:
• First, the West Germ an-style universal bank in
which the full range of banking and financial ser
vices is provided within a single legal entity. There
is no holding company, and separate subsidiaries
are used only at the convenience of the bank or
when required by foreign regulatory authorities for
particular activities conducted outside of Germany.
In most universal banking countries, banks may,
and often do, own sizable equity stakes in com
mercial concerns, but the opposite is generally not
the case. In other words, manufacturing and other
nonfinancial firms do not typically own and control
banks. In this regard, it should also be said that in
Germany the practice of banks owning large equity
stakes in commercial firms has been the subject of
lively political debate from time to time and the
subject of renewed debate in the recent past.
• Second, the British-style universal bank, which dif
fers from the German model in that (1) separate
legal subsidiaries are more common; (2) bank
holding of shares of commercial firms is far less
common, and (3) com binations of banking and
insurance firms are far less frequent —at least to
date. But the operational character of the United
Kingdom-style universal bank has much more in
common with the German-style model than it does
with arrangements here in the United States —a
pattern which will be magnified when the EC bank
ing directive becomes operational.
• Third, the fragmented systems such as currently
prevail in the United States and Japan. In these
models there are, of course, rigid legal and opera
tional distinctions between classes of financial
institutions, including but not limited to the separa
tion between commercial and investment banking.
However, even between the U.S. and Japanese
systems, important differences exist —for example,
the fact that holding companies do not exist in
Japan and, in fact, are strictly forbidden by law. It
2
FRBNY Quarterly Review/Spring 1990
should also be noted that some would suggest that
the Japanese system shows some signs of moving
toward a British-style universal bank, at least for
wholesale banking, securities, and other financial
services.
Even between the U.S. and Japanese systems,
important differences exist —for example, the fact
that holding companies do not exist in Japan and,
in fact, are strictly forbidden by law.
With that general description of the three prevailing
models in mind, let me now turn to the major differ
ences between the U.S. system and systems in the
major industrial countries abroad. Placing aside super
visory issues, which I will come to shortly, the major
structural differences are as follows:
• First, the bank or financial services holding com
pany is unique to the United States. This is not an
incidental difference when it is remembered that in
this country such holding companies are almost
always the financial and managerial nerve center
of the entity as a whole. Abroad, those crucial
functions — including the all-im portant point of
access to capital markets —are alm ost always
housed directly in the lead bank or lead financial
institution itself. In this setting, even sophisticated
foreign market participants and officials often have
a great deal of difficulty understanding structural
arrangements here in the United States —a situa
tion that was am plified by the recent Drexel
episode.
The United States is the only major country that
does not have a true national banking system.
While state initiatives are materially reducing the
barriers to national banking in the United States,
even after state initiatives have run their course,
we will still be left with a crazy-quilt pattern of
state and federal laws and regulations governing
various aspects of interstate banking.
• Second, the United States is the only major coun
try that does not have a true national banking sys
tem. While state initiatives are materially reducing
the barriers to national banking in the United
States, even after state initiatives have run their
course, we will still be left with a crazy-quilt pattern
of state and federal laws and regulations govern
ing various aspects of interstate banking. Such
arrangements in this country will stand in increas
ingly sharp contrast to the situation in the rest of
the world, and especially relative to Europe, once
the new banking directive takes hold and all duly
licensed banking entities — including subsidiaries
of U.S. banking and securities firms —will be freely
able to provide a full range of banking and finan
cial services across the national boundaries of the
twelve countries making up the European Eco
nomic Community. No small wonder, therefore, that
prominent officials in Europe have some difficulty
understanding the restrictions placed on the scope
of geographic and product opportunities available
to European institutions operating in the United States.
• Third, with the sole exception of Japan —and that
will almost surely change in time whether or not we
change —the United States stands out as the only
country with a fragmented banking system that
severely limits or restricts the type of financial
products and services that can be offered by par
ticular classes of institutions. Once again, and
With the sole exception of Japan —and that will
almost surely change in time whether or not we
change —the United States stands out as the only
country with a fragmented banking system that
severly limits or restricts the type of financial
products and services that can be offered by
particular classes of institutions.
leaving aside what changes may occur in Japan
over time, the comparative situation in the United
States will only get worse when the banking direc
tive becomes operational in Europe.
While these differences in structure are important,
they become all the more important in the face of
related differences in supervisory or so-called safety
net arrangements in the other major industrial coun
tries. In contrasting such arrangements in the United
States with those in other countries, there are some
striking differences. The most important of those differ
ences are as follows:
• First, in every other country studied, consolidated
supervision of mainstream banking and financial
companies is the rule —without any exception that
I am aware of. It is obviously the case with the
German-style universal bank; it is unambiguously
the case in practice in the United Kingdom; and it
is the case in Japan, even in the face of Article 65.
Moreover, in the few cases where commercial
companies own banks —such as in France or Italy
— the supervisory process pierces the “ corporate
veil” between the bank and the commercial com
pany owning and controlling the bank. Maybe my
friends and associates abroad tell me what they
think I want to hear, but what they often say is that
they are bewildered by those supervisory arrange
ments in the United States that do not rely on the
principle of consolidated supervision.
In every other country studied, consolidated
supervision of mainstream banking and financial
companies is the rule —without any exception that
I am aware of.
• Second, while twenty years in this business has
taught me that there are no absolutes, the fact of
the matter is that in no case that I have been able
to discover are there firewalls in mainstream for
eign banking firms that would routinely —but espe
cially as a matter of law —restrict flows of funds
and capital among affiliated entities within the
same financial group except in extraordinary cir
cumstances. My tendency to reject absolutes tells
me there must be exceptions to this. But if there
are, they are not prominent. For example, the
absence of such firewalls is obviously the case for
universal banks, but it is also the case within finan
cial groups or firms in the United Kingdom, Japan,
and elsewhere. That is not to say that strict regula
tions governing certain intracompany activities
aimed at customer protection, competitive equality,
and the facilitation of what we would call functional
regulation do not exist, for surely they do exist.
Rather, it is to say that in all major foreign coun
tries studied, issues of liquidity and solvency are
viewed at the level of the firm as a whole by both
regulators and market participants. Thus, firewalls
that wholly preclude or limit the flow of funds or
capital within the firm are viewed as either unnec
essary or counterproductive, except in extraordin
a ry c a se s w hen th e y are im p o se d by the
a u th o ritie s in an e ffo rt to exercise “ dam age
control.”
• Third, in all of the countries studied, mainstream
banking organizations as a whole —including their
securities affiliates — have direct or indirect access
to the payment, account, and liquidity facilities of
their respective central banks. Indeed, even in
Japan with its Article 65 separation of commercial
and investment banks, the major securities firms
FRBNY Quarterly Review/Spring 1990
3
have accounts at the Bank of Japan as well as
access to its discount window in exceptional cir
cumstances. These global arrangements, among
other things, reflect the implicit or explicit recogni
tion that central banks have a unique role and
responsibility to help safeguard the effective func
tioning of financial markets and institutions in order
to help guard against disruptions that might have
systemic implications.
In all of the countries studied, mainstream
banking organizations as a whole —including their
securities affiliates — have direct or indirect
access to the payment, account, and liquidity
facilities of their respective central banks.
• Fourth, in the area of deposit insurance, all of the
countries studied have deposit insurance systems
that have some similarities —but some important
differences — compared with the system of deposit
insurance in the United States. The most important
similarity is that all such systems studied have for
mal or legal limits stipulating that only deposits up
to a certain size are insured. The most important
differen ce is that, as a rough approxim ation,
appearances would suggest that the private sector
has a larger role in the operation and administra
tion of deposit insurance schemes in foreign coun
tries than in the United States, and as a result, the
direct linkage to the full faith and credit of the gov
ernment may be less explicit than is the case in
the United States.
financial institutions in foreign countries that have
occurred over the past twenty-five years. The results of
their work are summarized in Appendix II. Undertaking
this effort was not easy, because in all such instances
my associates relied essentially on publicly available
information and data. While this approach has its lim
itations, it is also true that this information is that which
market participants must use in assessing how the
a u th o ritie s w ill behave. On the oth e r hand, this
approach of course implies that factual details, but
especially judgments about motivation, may not always
be entirely clear. Yet for the purposes at hand, even the
broad sweep of events surrounding these cases pro
vides a useful insight into the de facto working of the
safety net in other countries. Moreover, beyond these
case studies, I have from time to time informally dis
cussed this general subject with many of my c o l
leagues in the major foreign central banks. On the
basis of both the more formal research and impres
sions gained in discussions with officials abroad, I
would draw the following conclusions concerning the
de facto operation of the safety net in major foreign
countries:
• First, with the sole exception of the Herstatt failure
in 1974, I am unable to find any case in which the
authorities have been willing to permit the sudden
and disorderly failure of an important banking or
nonbanking financial institution. (As I will indicate
below, the meaning of the word “ important” in this
context clearly is not limited to size.) Indeed, the
experience with Herstatt and its long and painful
aftermath seem to have provided the authorities in
all countries with a lasting impression of the grave
dangers associated with the sudden and uncon
trolled collapse of an important financial institution,
especially one with significant —although again not
la rg e by to d a y ’s s ta n d a rd s — in te rn a tio n a l
operations.
From the above, one might be tempted to conclude
that in foreign countries depositors are, in de facto
terms, less protected than they are in the United
States, or more broadly, that foreign banking and/or
financial firm s are somehow less “ protected” and
therefore subject to a greater degree of “ market disci
pline” than is the case here in the United States. In
order to gain some insights into these and related
questions, the next section of this statement looks at
some concrete examples of the workings of the socalled safety net in other countries.
Notwithstanding the presence of legal or formal
limits on the extent of deposit insurance coverage,
authorities in all foreign countries have, as a
matter of general practice, gone to considerable
lengths to protect depositors from loss.
The workings of the safety net abroad: some case
studies
In order to gain some useful insights into the de facto
workings of the safety net in other countries, I asked
several of my colleagues at the Federal Reserve Bank
of New York to research experience in that regard with
respect to a number of highly visible cases of troubled
• Second, notwithstanding the presence of legal or
formal limits on the extent of deposit insurance
coverage, authorities in all foreign countries have,
as a matter of general practice, gone to consider
able lengths to protect depositors from loss.
Indeed, my overall impression is that the number
FRBNY Quarterly Review/Spring 1990
Digitized 4for FRASER
of instances in which depositors — large or small,
insured or not — have incurred actual losses are
few and fa r betw een. And w here they have
occurred, they have been limited to isolated cases
or to cases in which the depositors had other rela
tionships—such as being shareholders —with the
failed institutions. Having said that, I should quickly
add that we have never seen anything abroad even
remotely approaching the scope of the thrift indus
try problem, although the U.K. secondary banking
crisis had many structural characteristics in com
mon with the thrift problem in the United States.
• Third, in a number of prominent cases, some cov
ered in the appendix and some not, the authorities,
including central banks, have moved swiftly and
decisively to intervene in the cases of troubled
nonbank financial institutions — here too, some
conspicuously small in size. Such interventions
have involved the use of central bank credit facili
ties, public monies, and at times a degree of moral
suasion, if not arm twisting, that, from my experi
ence, simply would not work in the United States.
In all such cases, public intervention was appar
ently motivated by concerns about systemic risk,
but as suggested above, in several instances the
troubled institution was not particularly large and
in other cases was distinctly “nonbank” in character.
• Fourth, in a number of prominent cases, again
some covered in the appendix and others not, the
“ rescue” efforts undertaken by the authorities
entailed a joint effort with public and private enti
ties. Indeed, at the risk of overgeneralization, it
seems fair to say that the foreign official institu
tions seem better able to call upon, if not insist
upon, the participation of private entities in such
It seems fair to say that the foreign official
institutions seem better able to call upon, if not
insist upon, the participation of private entities in
such rescue operations than is the case in the
United States.
rescue operations than is the case in the United
States. In a number of cases, though, including the
silver market crisis in 1980, the LDC debt crisis,
the Continental Illinois Bank problem, and the 1987
stock market crash, joint public and private actions
were very much in evidence in the United States.
Nevertheless, my impression —unscientific as it is
— remains that private institutions either are more
willing or feel more compelled to participate in sta
bilization or rescue efforts in foreign countries
than they do in the United States.
If that impression is correct, it raises the obvious
question why private foreign financial institutions
play a larger role in such rescue operations than is
the case in the United States. Here one can only
speculate that part of the answer may simply lie in
history: that is, it has been done that way for many,
many decades. It may also be true that where a
handful of banks dominate national banking sys
tems, that handful of banks feel more directly
threatened by potential dangers of a systemic
nature than do institutions here in the United
States. Finally, when the headquarters of that
same handful of banks are typically located only
minutes away from the central bank, and when so
few must agree on a particular course of action, it is
obviously much easier to bring things together than
it is here in the United States. For example, there is
some substance to the suggestion that one of the
most potent tools available to the Bank of England in
time of stress is the Governor’s eyebrow!
In summary, experience abroad suggests that author
ities in foreign countries behave in a manner very simi
lar to the authorities in this country when faced with
problems in financial institutions. If anything, the pro
tections provided by authorities abroad seem, on bal
ance, to go further than might reasonably be expected
in this country. This, of course, is another way of say
ing that institutions in this country are subject to at
least the same —if not a greater —degree of market
discipline than is the case abroad.
Yet it seems clear to me that over the past fifteen
years, the United States has had more than its share of
banking and financial market disruptions. In those cir
cumstances, a question naturally arises why we have
had, in relative terms, so many such disruptions in
recent years.
Given the experience in other countries, it is very
hard to make the case that the de facto operation
of the safety net in this country in and of itself is
accountable for the problems and strains in our
financial system.
For many, the answer to that question comes easily:
too big to fail, de facto full insurance of deposits, over
extension of the safety net, not enough market disci
pline. W hile there is some tru th in all of these
generalizations, the answer clearly is not all that sim
ple. One thing, however, strikes me as quite clear:
namely, given the experience in other countries, it is
FRBNY Quarterly Review/Spring 1990
5
very hard to make the case that the de facto operation
of the safety net in this country in and of itself is
accountable for the problems and strains in our finan
cial system. Even greater market discipline may be
needed, but the evidence as a whole suggests that
there is something much more fundamental at work.
To complete the analysis of the international side of
the equation, one more piece of the puzzle must be put
in place. That piece, of course, relates to the interna
tional competitiveness of U.S. banking and financial
institutions.
The international com petitiveness of major U.S.
financial institutions
In recent times, the subject of the international compe
titiveness of U.S. financial firms has received increas
ing attention. In order to shed some further light on this
subject, my associates at the Federal Reserve Bank of
New York have been engaged in an effort to analyze
more systematically both the myths and the realities of
this situation. As a part of that effort, I have attached
to this statement as Appendix III a paper summarizing
the results of some of this work. That paper provides a
summary of selected performance traits of a cross sec
tion of fifty-one major and internationally active banking
and securities companies from seven major industrial
countries. Before I attempt to summarize the results of
this analysis, allow me to emphasize a number of quali
fications about these data and the delicate task of
drawing reasonable and reasoned conclusions from the
data:
• First, due to substantial differences in regulatory,
accounting, and tax rules in the respective coun
tries, many of the cross-border comparisons are
seriously distorted if taken at face value. Indeed,
to properly interpret the data, one must have at
least a general sense of these accounting and
related differences in order to have the proper per
spective on the various statistics. To cite just an
example or two:
— I know that the underlying capital position
of one or more groups of foreign banks is a
good deal stronger than the raw statistics
would suggest, in part because accounting
rules and tax rules allow some groups of
banks a great deal of flexibility with regard to
the accounting for, and accumulation of, socalled hidden reserves.
— I know that the statistics on profitability are
seriously distorted — perhaps more so than
any other grouping of these d a ta - in a way
that tends to understate the “ core” profit
ability of U.S. banks relative to one or more
6
FRBNY Quarterly Review/Spring 1990
of the other groupings of national banks.
• Second, even aside from data problems, the infor
mation contained in this appendix is limited in the
extent to which it provides decisive insights into
the international competitiveness of one national
group of banks versus others. In part this is true
because of certain individual institutions in this
country and abroad that stand out to such an
extent from their domestic peers that they seem to
defy these international comparisons.
• Third, under the best of conditions, aggregate per
formance data at the level of the individual firm tell
only a part of the story. Accordingly, my associates
are also seeking to investigate international com
petitiveness of U.S. firms in a number of specific
but d is c re te m arkets, ranging from fo re ig n
exchange and interest rate swap markets to retail
banking. This is of course a very difficult undertak
ing but the initial impression one gets from this line
of approach is that U.S. financial institutions con
tinue to be seen in the markets as strong and
imaginative competitors in many individual product
and service lines, especially in the more sophisti
cated and innovative areas. This work has a long
way to go, but if it is successful, we will find an
appropriate vehicle to make the results public
sometime late this year or early next year.
With those qualifications in mind, my personal inter
pretation of the data and inform ation contained in
A p p e n d ix III le a d s me to th e fo llo w in g m ain
conclusions:
• First, on balance, I would place the U.S. banks
somewhere roughly in the middle of the pack of the
national groups of banks studied in terms of all the
performance measures studied.
• Second, looking at U.S. banks and securities firms
combined, relative to the German-style or Britishstyle universal bank or relative to the combination
of Japanese banks and securities companies, I
would be inclined to a similar “ middle-of-the-pack”
or perhaps slightly weaker relative ranking of U.S.
institutions.
• Third, while hindsight in this regard is far from
twenty-twenty, my strong hunch is that a similar
exercise performed ten or twenty years ago would
have provided a result in which the rankings of
U.S. firms would have been higher and perhaps
materially higher. In other words, while the data
may suggest that U.S. firms are still quite capable
of holding their own in an international context,
there is no doubt in my mind that as a group their
position has slipped.
Before I turn to the final two sections of this state
ment, allow me to summarize its main points thus far.
First, the basic structural and supervisory framework
governing the operations of U.S. banking and financial
firms is materially different from that in all other coun
tries studied and, as things now stand, will become
increasingly so in the foreseeable future. Second,
The basic structural and supervisory framework
governing the operations of U.S. banking and
financial firms is materially different from that in
all other countries studied and, as things now
stand, will become increasingly so in the
foreseeable future.
despite sharp differences in key elements associated
with the structure and design of the so-called safety
net, the de facto operation of the safety net —including
the extent of depositor protection — appears quite simi
lar across the industrial countries. If anything, market
discipline may, on balance, play a larger role in this
country. Yet over the past fifteen years, the United
States has had more than its share of banking and
financial disruptions, therefore casting some doubt on
the oft-cited proposition that these problems have as
their basic cause the tendency for authorities in this
country to bail out troubled financial institutions.
Finally, the analysis of the international competitive
ness of U.S. financial firms places such firms some
where in the middle of the pack, a position that almost
surely reflects a deterioration in standing from ten or
twenty years ago.
The analysis of the international competitiveness
of U.S. financial firms places such firms some
where in the middle of the pack, a position that
almost surely reflects a deterioration in standing
from ten or twenty years ago.
Of the many questions raised by these interim con
clusions, two stand out: first, if the de facto operation
of the safety net is quite similar across countries, why
have we in the United States witnessed what strikes
me as a disproportionate number of financial disrup
tions in recent years; and second, what accounts for
the less than strong overall performance of U.S. firms
in an international setting, especially if, as I believe,
that performance has slipped in the last decade or
more?
Underlying factors influencing the performance of
banking and finance
At the risk of great oversimplification, it seems to me
that there are five major reasons that we see a U.S.
banking and financial system characterized by the dual
conditions of recurring bouts of instability and competi
tive slippage both at home and abroad:
• The first major factor that is helping to shape these
trends in the banking and financial sector —espe
cially in a comparative international context —is
macroeconomic performance and policies. Over
the last decade and a half, volatility in GNP, high
and volatile rates of inflation, low savings, and our
weakened external position have all contributed to
a financial environment that breeds difficulties at
home and contributes to slippage abroad. With
regard to the latter, there is simply no question in
my mind that one of the key reasons for the emer
gence of the Japanese financial institutions as so
large a force in global markets is rooted in Japan’s
strong overall economic and financial performance
over much of that period.
One of the key reasons for the emergence of the
Japanese financial institutions as so large a force
in global markets is rooted in Japan’s strong
overall economic and financial performance over
much of that period.
• Second, for a variety of reasons —many rooted in
technological advances in telecommunications and
in the information sciences —the historic value of
the banking franchise is under great pressure. The
institutionalization of savings, the securitization of
financial assets and liabilities, the easy access to
inform ation about creditw orthiness of individual
borrowers, and even the “ 800” telephone number
are all symptomatic of a rapidly changing banking
and financial environment that has unquestionably
undercut the once considerable value of the bank
ing franchise. This tendency is reinforced by the
fact that due to these same technological and
informational factors, the “ shelf life” of most inno
vations in banking and finance is very short in
duration.
As one reflection of this, the most creditworthy
corporate borrowers can now fully bypass the
entire banking and financial system for many of
their day-to-day credit needs. For example, we now
have instances in which firm s with particularly
strong credit ratings are able to place their own
FRBNY Quarterly Review/Spring 1990
7
commercial paper directly with institutional and
other investors, thereby bypassing not only the
commercial banking system —once the exclusive
source of such short-term credit —but also the
underw riting and placing capabilities of the in
vestm ent banking industry. Now th a t is d is
intermediation!
As another, more recent example, AT&T —with
its vast financial and technological resources — has
recently entered the credit card business and in
the process is offering consumers very attractive
terms on such cards in a context in which there
would appear to be potentially very considerable
The institutionalization of savings, the securitiza
tion of financial assets and liabilities, the easy
access to information about creditworthiness of
individual borrowers, and even the “800” tele
phone number are all symptomatic of a rapidly
changing banking and financial environment that
has unquestionably undercut the once consider
able value of the banking franchise.
synergies between this line of business ana
AT&T’s traditional lines of business.
The diminished value of the banking franchise in
this country appears, for a variety of reasons, to
be somewhat more advanced than is the case in
other countries —although that may not last. For
example, capital markets are not nearly as well
developed in Japan as they are in the United
States, the result being that the role of the deposit
intermediary is still more central in Japan than it is
in the United States. Some would also suggest that
the very close relationships between banking and
large industrial firms in other countries work in the
direction of helping to preserve the banking fran
chise. Finally, some would also suggest that the
authorities in other countries may tilt a bit in the
direction of trying to preserve the value of the fran
chise where that is possible.
In suggesting that the value of the banking fran
chise has declined, I am not suggesting that the
developments that have given rise to this situation
are all bad. To the contrary, most of them are very
good in terms of efficiency, reduced costs, greater
com petition, and vastly increased choices for
savers and investors. But taken as a whole, these
d e v e lo p m e n ts have c le a rly and irre v e rs ib ly
changed the rules of the game in a manner that, at
the very least, makes for difficult transitional prob
lems for the affected institutions and markets.
FRBNY Quarterly Review/Spring 1990
Digitized 8
for FRASER
• Third, partly because of the competitive implica
tions of the technolo gical and market forces
described above and partly because we have so
many financial institutions and so many classes of
financial institutions that compete with each other,
we now have, in my view, excess capacity in large
segments of banking and finance. This same con
dition appears to exist internationally, at least in
some segments of wholesale markets. The symp
toms of this condition abound in razor-thin spreads
and pinched margins, and perhaps especially in
the troublesome manner in which we see vast
amounts of very short-term churning and trading in
so many segments of the financial markets. As I
have said on other occasions, this situation seems
at times to create a vested interest in volatility,
since opportunities for trading profits at the level of
the individual firm or individual trader seem great
est when swings in interest rates and exchange
rates are also the greatest. Whatever else one can
say about this, it reinforces the unrelenting preoc
cupation with the short run we see in financial mar
kets and in corporate America more broadly.
Partly because of the competitive implications of
...technological and market forces...and partly
because we have so many financial institutions
and so many classes of financial institutions that
compete with each other, we now have, in my
view, excess capacity in large segments of
banking and finance.
What may be even more important, however, is
that if my conjecture about excess capacity in
financial services is correct, it would clearly imply
that we will have to go through a period of at least
some consolidation in banking and finance. In say
ing this, I am not suggesting for one minute that
we will end up with a highly concentrated banking
and financial system along the lines of what we
see in many other countries. The trick, of course,
is to shape public policies in a manner that pro
vides the highest assurance that the process of
financial market consolidation in the United States
occurs in an orderly and equitable manner consis
tent with broad national goals and priorities.
• The fourth factor I would cite in this regard is the
direct subject of these hearings: namely, the out
dated legal and in stitu tio n a l fram ew ork w ithin
which the U.S. banking and financial institutions
operate. As the first section of this statement
makes clear, the U.S. banking system is simply out
of step with the rest of the world, and more impor
tant, it is out of step with the realities of the mar
ketplace. Even more important, the system as now
configured may be risk and accident prone rather
than risk adverse. Fragmentation alone may pro
duce that result since fragmentation can inhibit
diversification of risks on both sides of the balance
sheet. Similarly, the inescapable tendency of firms
and market participants to push the spirit and the
letter of law and regulation to the limit in a frag
mented system brings with it its own elements of
risk and perversity. Finally, fragmentation inevita
bly brings with it the tendency to shift activities off
shore, which in turn entails loss of income, loss of
jobs, and in some cases the loss of some element
of managerial and supervisory control.
• A final factor I would cite that bears particularly on
the relatively high incidence of financial disruptions
in this country is what are often called gaps or
lapses in the su p ervisory process. The most
important illustration of this, by far, is to be found
in the thrift industry situation, which was far, far
more a fatal flaw in the legal and supervisory pro
cess than a flaw in the architecture of the deposit
insurance system. More generally, I believe at
least something of a case can be made that the
highly fragmented nature of the U.S. banking and
financial system may, in its own right, contribute to
gaps in the supervisory process.
I believe at least something of a case can be
made that the highly fragmented nature of the U.S.
banking and financial system may, in its own right,
contribute to gaps in the supervisory process.
In considering all the factors cited above, note that
for the most part they are secular in nature and, as
such, reflect “ outside” forces impacting on the environ
ment within which banking and financial institutions
operate. However, one should not conclude from this
that all of the elements that have produced strains in
banking and finance are of this nature. For example,
there are internal forces —such as the current prob
lems in real estate markets or the overhang in the mar
ket for high-yield securities —that importantly reflect a
lack of prior restraint on the part of some market par
ticip a n ts. S im ilarly, we have seen all too many
instances of poor management, overly aggressive
strategies, and unfortunately, elements of highly ques
tionable behavior, including some outrageous instances
of outright criminal activity.
But problems growing out of poor management, reck
lessness, or even illegal activities — outside of the thrift
in d u stry problem — are re la tiv e ly few in number,
although at times highly visible in nature. Moreover,
existing regulatory and legal sanctions can probably
cope with these problems, especially with the strength
ened provisions in the Financial Institutions Reform,
Recovery and Enforcement Act. The more generalized
situation having to do with the competitive well-being
and the underlying stability of the banking and financial
system must be looked at in the broader light of the
other issues raised in the e a rlie r p a rts of this
statement.
Reform and modernization of the U.S. banking and
financial systems
In approaching the reform and modernization of the
U.S. banking and financial system, this Committee and
the Congress as a whole can, as I see it, choose
among four basic alternatives or models: (1) the current
fragmented system, (2) the German-style universal
bank, (3) the British-style universal bank, or (4) the
financial services holding company approach along the
broad lines suggested in my 1987 essay, Financial Mar
in approaching the reform and modernization of
the U.S. banking and financial system, this
Committee and the Congress as a whole can, as I
see it, choose among four basic alternatives or
models: (1) the current fragmented system, (2) the
German-style universal bank, (3) the British-style
universal bank, or (4) the financial services
holding company approach along the broad lines
suggested in my 1987 essay.
ket Structure: A Longer View. For my part, I would
quickly rule out the status quo and the German-style
universal bank. In the latter case, I would come to that
conclusion because of my discomfort about the nature
and extent of the “ banking-commerce” linkages that
the German-style universal bank seems to imply.
In a consideration of the relative merits of the finan
cial services holding company versus the British-style
universal bank, there are three major areas of potential
difference — aside from su pervisory arrangem ents,
which will be considered later. Those areas of differ
ence are (1) the definition and scope of financial activ
ities that can be conducted w ithin the group or
company; (2) the presence or absence of the holding
company itsel', keeping in mind that these holding
companies are, in a U.S. context, the financial and
managerial nerve center of the company as a whole;
and (3) the nature and extent of the rules and regula
FRBNY Quarterly Review/Spring 1990
9
tions limiting or preventing various classes of transac
tions or other relationships between the various com
ponent parts of the entity as a whole —that is, Chinese
walls and/or firewalls. While it does not necessarily fol
low, there may also be differences as they relate to
whether and under what conditions the entity or its
parts may have access to the account, payment, and
liquidity facilities of the central bank.
In choosing between these two alternatives, I would
still have a preference for the financial services holding
company, particularly given where we are as a nation
in terms of the evolution of our attitudes on these mat
ters. Having said that, if I had the liberty of starting
with a clean slate, I might well opt for the British-style
universal bank. In either case, I strongly believe that
the approach must be reciprocal along the broad lines
of the philosophy of the Proxmire-Garn bill. That is, for
example, if banks can get into the securities business
as a general matter, securities companies —with some
possible exceptions —can get into the banking busi
ness. In the case of either model, I also believe that
There must be strong Chinese walls that provide
protections against conflicts of interest, unfair
competition, and certain kinds of “tie-ins.”
Similarly, I believe there should be reasonable
protections against undue concentrations.
there must be strong Chinese walls that provide protec
tions against conflicts of interest, unfair competition,
and certain kinds of “ tie-ins.” Similarly, I believe that
there should be reasonable protections against undue
concentrations, recognizing, of course, that some con
solidation in banking and finance will occur under any
circumstances. Finally, I believe that there must be an
instrumentality that would be responsible for the ongo
ing task of defining and limiting the scope of activities
that can be conducted in the group as a whole.
I believe with more conviction than ever that we
must have a system of consolidated supervisory
oversight of any company that has direct or
indirect access to the “safety net.”
There are, however, three areas in which my per
sonal views may place me somewhere between a dis
tinct m inority and a voice in the wilderness — and
probably closer to the latter. They are:
• First, I believe with more conviction than ever that
10
FRBNY Quarterly Review/Spring 1990
we must have a system of consolidated super
visory oversight of any company that has direct or
indirect access to the “ safety net.” To me, the mini
mum that this entails includes (1) system atic
reporting of financials for both on- and off-balance
sheet activities at the level of the holding company
and all of its subsidiaries and affiliates; (2) mini
mum capital standards, including those for the
holding company level; and (3) standby authority
for inspectors or examiners to do on-site reviews
in any entity within the group, including the parent
holding company.
Where the dom inant company is a banking
entity, the Federal Reserve would be responsible
for consolidated oversight. Where the dominant
firm is a securities entity, that task could be given
to the Securities and Exchange Commission. Func
tional supervision, much as we have it today,
would apply to the component parts of the group
as a whole.
To me, the world we live in, together with the
nature of the potential systemic risks we face in
banking and finance, demands that we move in this
direction. Some two and a half years ago, I sum
marized for this Committee the thinking that led
me to this conclusion. I have appended to this
statement, as Appendix IV, an excerpt from that
earlier testimony on this subject. At this juncture,
all I would add is that events since then, including
the Drexel episode, have strengthened my views in
this regard.
I remain strongly opposed to the merging of
banking and commerce and to any arrangements
that would even remotely contemplate the
ownership and control of bank holding companies
or financial services holding companies containing
depository institutions by commercial concerns.
• Second, I remain strongly opposed to the merging
of banking and commerce and to any arrange
ments that would even remotely contemplate the
ownership and control of bank holding companies
or financial services holding companies containing
depository institutions by commercial concerns.
Here too, on an earlier occasion, I spelled out
before this Committee the reasons for my con
cerns in this regard, and I have provided an extract
of that earlier testimony as Appendix V.
Notwithstanding the observations I made earlier
about the diminished value of the banking fran
chise and the inroads of commercial firms into
financial businesses, I still look with concern, if not
alarm, at the economic, financial — and perhaps
even social — implications of Exxon owning Chase
Manhattan, Ford owning Citicorp, or RJR Nabisco
owning J.P. Morgan. Obviously, those examples
draw on more than a little hyperbole in order to
stress the point. But once that door is opened,
there is absolutely no way to anticipate how events
will shake out over time. Therefore, and absent
that compelling public policy reason I spoke of in
my earlier testimony, I would strongly urge that we
m aintain a s tric t se pa ra tion of banking and
commerce.
It may be, in appropriate circumstances, that a
case could be made that a margin of added flex
ibility could be provided whereby a bank or finan
c ia l s e rv ic e s h o ld in g c o m p a n y c o u ld own
somewhat more than 4.9 percent of the equity of a
nonfinancial concern and vice versa. However,
even this would have to be approached with care
in view of the often razor-thin distinctions that now
exist between various classes of “ equity” and
"debt” securities, keeping in mind that the issue
here is not arithmetic but influence and control.
Similarly, some added flexibility might be consid
ered where a bank or financial services holding
company owns even a “ controlling” interest in a
nonfinancial firm so long as that latter firm is, in
some sense, de minimis relative to the bank or
financial services holding company as a whole.
However, this too would have to be approached
with great care, keeping in mind the extent of the
problems that can arise, for example, with a seem
ingly de minimis real estate development company.
The need for great care in this regard is strongly
reinforced by case after case that illustrates that
The well-being of the company as a whole cannot
be safely disentangled from problems or
adversities affecting an affiliated company, no
matter how thick the firewalls nor how well
constructed the legal separation.
the well-being of the company as a whole cannot
be safely disentangled from problems or adver
sities affecting an affiliated company, no matter
how thick the firewalls nor how well constructed
the legal separation. Indeed, in times of stress, not
only does the marketplace fail generally to accept
these distinctions, but the directors and managers
of the firms under stress do not accept them either.
• Third, as suggested above, I have real worries
about “ firew alls” becoming “ walls of fire.” For
these purposes, I want to distinguish between (1)
“ Chinese walls” —which, like regulations 23-A and
23-B, seek to protect against conflicts and unfair
competition — and/or other such regulations gov
erning “ normal” business relations among affili
ated companies, and (2) “firewalls,” which strictly
limit or prevent the mobility of funds and capital
among affiliates. My problem here is not with wellconceived “ Chinese walls” but rather with ill-con
ceived “ firewalls.”
From a broad public policy perspective, the case
for very thick and very high firewalls rests heavily
on concerns about overextension of the safety net,
threats to the deposit insurance fund and ulti
mately to the taxpayer, and the more subtle, but
very important, moral hazard dilemma. Taken indi
vidually or collectively, these issues cannot be dis
missed lightly.
Just as we cannot dism iss these concerns
lightly, neither can we dism iss what the mar
ketplace tells us both here and abroad. And what
the marketplace tells us with almost unfailing regu
larity is that in times of stress, some parts of a
financial entity cannot safely be insulated from the
problems of affiliated entities. Investors, creditors,
and even managers and directors simply do not
generally behave in that fashion, and the larger the
problem the less likely they are to do so. Because
this pattern of behavior seems so dominant and
There seems to me little doubt that taken to an
extreme, absolute firewalls can aggravate
problems and instabilities rather than contain or
limit them.
because the authorities throughout the rest of the
industrial world generally frame their policies with
this in mind, there seems to me little doubt that
taken to an extreme, absolute firewalls can aggra
vate problems and instabilities rather than contain
or limit them. Indeed, I do not have to stretch my
imagination or my memory very far to find exam
ples in which a heavy-handed approach to firewalls
could easily have been the source of significant
problems.
There is also a matter of logic here: That is, if
we are prepared to accept the proposition that
greater flexibility in allowing combinations of enti
ties providing financial services makes sense, we
must be saying, at least implicitly, that such combi
nations make sense on economic grounds. Other
FRBNY Quarterly Review/Spring 1990
11
wise the exercise is sterile. On the other hand, if
we say such combinations are permissible but then
insist on firewalls that are so thick and so high as
to negate the economics of the combination in the
first place, the net economic result will also be
sterile.
As with most things, the whole subject of fire
walls has to be viewed in context. For example, in
the context of an individual firm with very strong
capital resources, presumably the case for fire
walls is greatly diminished, if not eliminated. On
the other hand, during a transition period in which
financial structure is changing —especially if that
process of change is accompanied by some con
s o lid a tio n — a conservative interim approach to
firewalls may be quite appropriate. But even in
those circumstances, I believe care is needed to
ensure that we provide enough flexibility so that
firewalls do not, in fact, become walls of fire.
Safety net arrangements
As I see it, any discussion of the federal safety net
associated with the banking system and individual
banking institutions must start with the fact that such
institutions are subject to a higher degree of regulation
and supervision than is the case for most other kinds
of private enterprise. While the specific points of
emphasis of such regulation will vary from time to time
and place to place, the basic rationale for such
arrangements rests on two elements: first, the unique
nature of the fiduciary responsibilities of such institu
tions, and second, the unique elements of systemic risk
present in banking and finance. While some bankers
are not shy to complain about the burdens of some
forms of regulation, all accept the premise as to why
banks are regulated in the first instance. And most rec
ognize that in exchange for carrying the burden of reg
ulation, banking institutions enjoy certain benefits not
normally accorded by markets or society to other
classes of institutions. For example, the mere presence
of the supervisory apparatus is one of the reasons that
the marketplace allows banking and financial institu
tions to operate with a higher degree of leverage than
most other classes of institutions —a result that is seen
as economically and socially desirable because of its
capacity to help mobilize savings and investment and
thereby foster economic growth and rising standards of
living.
In addition, and in further exchange for accepting the
burden of regulation, society conveys to banking
organizations certain other direct benefits: deposit
insurance, access to the liquidity facilities of the central
bank, and not least, access to the account and pay
ment facilities of the central bank. The safety net must,
12
FRBNY Quarterly Review/Spring 1990
therefore, be viewed as a package deal but one in
which it is explicitly recognized that bankers — knowing
that their business is essentially the business of public
confidence —will conduct their affairs in a safe and
prudent manner consistent with their fiduciary and
societal responsibilities.
Looked at in that light, officially imposed prudential
standards in such areas as capital adequacy, liquidity
management, lending limits, and so forth —as well as
the official examination process itself —are aimed in
The safety net must, therefore, be viewed as a
package deal but one in which it is explicitly
recognized that bankers — knowing that their
business is essentially the business of public
confidence —will conduct their affairs in a safe
and prudent manner consistent with their fiduciary
and societal responsibilities.
part at helping to establish an overall framework within
which such institutions can compete and flourish but do
so in a context that protects the safety and stability of
the system as a whole. But —and this is a very large
but —the first and foremost responsibility for the safe
and prudent operation of individual institutions rests
with the directors and management of those institutions
— not with the authorities.
Because the safety net by its very nature is a pack
age deal, possible approaches aimed at improving the
manner in which it functions must be viewed in that
overall context. For that reason, we must be careful
about approaches that focus largely or exclusively on
any one aspect or feature of the safety net to the
exclusion of others. For example, while there are
opportunities to improve the workings of the deposit
insurance system, the deposit insurance system can
only be as effective — and as cost effective —as the
safety net as a whole, especially its supervisory com
ponents. Indeed, at the end of the day, I would argue
that the broad approach to supervisory policy —includ
ing the examination process itself —is the foundation
upon which an effective deposit insurance system must
rest.
In the current setting, much of the debate about the
safety net in general, and the deposit insurance system
in particular, centers on wholly understandable con
cerns about the cost of bailing out troubled depository
institutions. Within that context, there is a particularly
sharp edge of debate about the school of thought that
focuses on the suggestion that some institutions are
too big to fail and the implications of that for the so-
called moral hazard problem. In other words, how do
we secure the right balance between market discipline
on the one hand and protections against severe —if not
systemic —disruption and dangers on the other, espe
cially in a setting in which the business of banking and
finance is subject to the enormous competitive and
external challenges described earlier in this statement?
To my way of thinking, the most essential part of the
answer to that question lies in the combination of pri
vate a ctio n s and s u p e rv is o ry p o lic ie s th a t w ill
strengthen the financial and capital positions of individ
ual institutions, perhaps especially those institutions
that by their size or character present the greatest risks
to the stability and well-being of the system as a whole.
In this regard, it is perhaps worth noting that over the
past decade we have, in fact, seen a material strength
ening of the financial position of the largest banking
organizations here in the United States.
To illustrate this, I have included in the appendixes to
this statement a series of charts depicting key indica
tors of the performance of the ten largest banking
organizations in the United States over the past
decade. In providing these data, I am mindful of the
problem of having picked the ten largest as opposed to
the twelve largest or the five largest or the twenty-five
largest. I assure you, Mr. Chairman, the number ten
was chosen only because it is a nice round number.
Beyond that, it has no significance whatsoever. And if a
different number were used, the results would not be
affected in any material way.
Taken as a whole, these charts capture rather well
both the problems and the progress these institutions
and the industry at large have experienced over the
last ten years. They also capture the radically changed
character of the banking business over the decade. But
perhaps more than anything else, they capture the very
s iz a b le and very n e c e s s a ry b u ild u p in c a p ita l
resources over the period. To cite just two examples:
I would argue that in the current environment, in
which the domestic and international marketplace
rewards strength, the competitive position of
internationally active U.S. banking organizations
would be improved as they move toward, and
hopefully to the top of, the ladder in terms of their
comparative capital strength.
— Since 1979, the absolute level of primary capi
tal of these institutions has about quadrupled,
reaching almost $80 billion at year-end 1989,
while the prim ary capital ratios have about
doubled.
— As of year-end 1989, the BIS tier-one riskbased capital ratios of these institutions are
already well above the 1992 minimums, using the
more stringent 1992 definitions of capital.
In pointing to these data, I do not want to leave the
impression that I am satisfied that all that needs to be
done in strengthening the financial position of these
and other institutions has been done, for it has not.
Indeed, in the current environment, all institutions
should be working toward overall capital positions that
are considerably in excess of regulatory minimums.
Indeed, I would argue that in the current environment,
in which the domestic and international marketplace
rewards strength, the competitive position of interna
tionally active U.S. banking organizations would be
improved as they move toward, and hopefully to the top
of, the ladder in terms of their comparative capital
strength.
Within the context of public and private initiatives
that will continue to improve balance sheet and capital
strength, the task of possible reforms of the deposit
insurance system becomes far less formidable. Since
the Treasury and the Federal Reserve are both looking
into the subject of deposit insurance reform, I do not
want to muddy the waters by getting into a bill of par
ticulars on deposit insurance reform on this occasion.
However, I will say that in my estimation the single
most serious abuse of the deposit insurance system
has been the misuse of brokered deposits. As a practi
cal matter, fixing this problem without destroying the
legitimate business of money and deposit brokerage
will not be easy. In principle, however, what we should
be striving for is a system in which the $100,000
The first line of defense regarding the workings
and integrity of the deposit insurance system lies
in strong, well-diversified, competitive, and
capital-rich depository institutions and in a strong,
professionally staffed, and politically independent
supervisory apparatus.
deposit insurance limit should apply per individual or
per entity. In mentioning this particular area of con
cern, I do not want to leave the impression that I am of
the view that there may not be other constructive areas
in which reforms might be considered. That is not my
view. But as I have stressed earlier, the first line of
defense regarding the workings and integrity of the
deposit insurance system lies in strong, w ell-diver
sified, competitive, and capital-rich depository institu
tions and in a strong, professionally staffed, and
politically independent supervisory apparatus.
FRBNY Quarterly Review/Spring 1990
13
In this connection, we must also guard against the
seductive appeal of “ cookbook” approaches to prob
lem institutions. With any troubled financial institution,
but especially in the case of large institutions, I believe
that the workings of both the safety net and market
discipline will be better served in a context in which
the authorities maintain a policy of what I like to call
“constructive ambiguity” as to what they will do, how
they will do it, and when they will do it. In saying this, I
recognize that financial market participants do not like
uncertainty, but that is just the point! Moreover, while I
fully understand the yearning in some quarters for the
cookbook approach to problems in financial markets or
institutions — large institutions especially — I regret to
say that in my judgment such a cookbook does not and
never will exist. The circumstances associated with a
particular case, the setting in which it occurs, and the
assessment of the relative costs and benefits of alter
native courses of action will always have to be looked
at case by case. But in no case should it be prudent for
market participants to take for granted what actions the
authorities will take and certainly in no case should
owners and managers of troubled institutions — large or
sm all— conclude that they will be protected from loss
or failure.
Conclusions
Mr. Chairman, my statement and its appendixes have
covered an enormous amount of ground. In the interest
of your patience I will not attempt to summarize at this
time. But in conclusion, allow me to briefly stress three
With or without progressive legislation, the period
ahead in banking and finance will not be easy. But
with progressive legislation, our prospects are so
much better for consumers, for businesses, for
competitiveness, and perhaps most of all for the
stability and soundness of our financial markets
and institutions.
points. First, while I can readily understand why the
thrift industry problem may have dampened the enthu
siasm of the Congress to tackle the issues I have dis
cussed today, it seems to me that the thrift industry
problem tells us rather clearly that the longer a prob
lem festers, the worse it becomes. With or without pro
gressive legislation, the period ahead in banking and
14
FRBNY Quarterly Review/Spring 1990
finance will not be easy. But with progressive legisla
tion, our prospects are so much better for consumers,
for businesses, for competitiveness, and perhaps most
of all for the stability and soundness of our financial
markets and institutions.
Second, over the course of my statement, I have
deliberately stayed away from the subject of possible
reforms in the structural arrangements associated with
the supervisory system. I have done that because I
firmly believe that such reforms should follow from
reform of the banking and financial system —not pre
cede it. But at the risk of appearing self-serving, I do
want to repeat my utter conviction that when reform of
the supervisory structure does occur, it should proceed
in a manner that preserves a central —but by no
means exclusive — role for the Federal Reserve. It
seems to me that experience here and throughout
much of the world tells us in rather certain terms that
helping to ensure the safety and soundness of banking
and financial markets and institutions and helping to
stabilize such markets and institutions in the face of
adversity are functions that relate directly to the very
essence of central banks.
Finally, over the course of this statement, I have
drawn heavily on experience and conditions in other
countries. Having done that, I do not want to leave the
impression that I feel any compelling case to duplicate
precise arrangements in any other country or group of
Our financial markets are still the bellwether of
world financial markets; our banks, investment
banks, and insurance companies are still the
leaders in constructive innovation; and our
financial markets and institutions are still the
world’s safe harbor. Let us keep it that way!
countries, because I do not. To be sure, we have some
work to do in this country in adapting arrangements in
a changing global setting; to be sure, we face some
difficult transition problems in the period ahead; to be
very sure, none of this will be easy. But as we face
those challenges, let us not lose sight of our strengths:
our financial markets are still the bellwether of world
financial markets; our banks, investment banks, and
insurance companies are still the leaders in constructive
innovation; and our financial markets and institutions are
still the world’s safe harbor. Let us keep it that way!
Event Risk Premia and
Bond Market Incentives for
Corporate Leverage
The growth of highly leveraged transactions during the
1980s had a profound effect on corporate shareholders
and bondholders. Studies exploring the effects of the
restructurings, mergers, leveraged buyouts, and recap
italizations of the last decade have typically focused on
the welfare of firms and agents directly involved in the
transactions. Evidence suggests, however, that the
repercussions of these activities have extended to a
much broader circle of corporations and individuals.
This article introduces and explores the hypothesis
that the threat of leveraging from such events has
worked to raise the risk premia paid on the debt of all
but the most leveraged firms. In addition, it argues that
the increase in the cost of debt has been greater for
less leveraged firms. The penalty for low leverage cre
ates incentives for debt financing, thereby compound
ing existing distortions created by tax laws and limited
shareholder liability. In effect, the growth of an active
market for corporate control has embedded in bond
prices the possibility that firms will either leverage
themselves or be leveraged in a change of ownership.
As a resu lt, the m arginal incen tive to leverage
increases. A related implication is that it no longer
appears as feasible as it once was for a firm to hold
down its marginal cost of debt capital by keeping a
very large equity cushion.
The article begins with a more precise exposition of
the hypothesis. Subsequent sections provide empirical
support and estimate the extent to which the leverage
threat affects the schedule of borrowing rates faced by
U.S. firms. The relevance of the hypothesis to the
observed increase in U.S. corporations’ dependence on
debt is also considered. In the closing sections,
the article reviews the implications of the empirical
results for the cost of capital to U.S. corporations in the
1980s and the relationship of these findings to policy
measures aimed at limiting corporate debt.
Effects of event risk premia
Our hypothesis posits that during the 1980s the threat
of unanticipated leverage increased the risk premia
paid by U.S. corporations on their debt and that this
increase was greater for the debt of less leveraged
firms. The hypothesis is based on two observations.
First, a firm ’s existing bondholders face substantial
losses from unanticipated increases in the firm ’s
leverage. Second, the risk is greater for less leveraged
firms because they are better candidates for increased
debt and because their bondholders face greater
potential losses from the increase in debt. A fuller
understanding of this hypothesis requires fam iliarity
with the concepts of the risk premium curve and the
event risk premium.
The risk premium curve
The risk premium is the sum a corporation must pay on
its debt beyond the riskless interest rate in order to
compensate bondholders for the possibility of default.
The most obvious determinant of the firm’s risk pre
mium is its leverage, although factors such as volatility
of cash flow, access to credit markets, asset liquidity,
firm size, and diversity of revenue sources also play a
role.
The risk premium increases with leverage. The lower
curve in Chart 1 depicts the risk premium as a function
of a measure of leverage —the debt ratio. The debt
FRBNY Quarterly Review/Spring 1990
15
ratio is defined as the market value of a firm ’s debt
divided by the total value (market value of equity plus
market value of debt) of the firm. The chart shows the
risk premium curve as rising more than proportionally
with the debt ratio. This representation of the curve is
in the literature and, as demonstrated below, em piri
cally verifiable.
Event risk
Event risk will be defined as the risk of any significant
unanticipated increase in the debt share of a firm. Such
increases may result from leveraged buyouts, share
re p u rc h a s e s , e x tra o rd in a ry d iv id e n d s , and o th e r
leveraging tactics. The tactics include leverage both by
outsiders who gain control and by current managers
who adopt defensive measures.
The existence of event risk will prompt bondholders
to require a risk premium in excess of that which would
be demanded of a firm without such risk. The size of
this additional premium will largely be a function of the
debt ratio, although factors such as industry cyclicality
and cash flow volatility may also play a part. As noted
earlier, a firm with a low debt ratio will tend to pay a
higher event risk premium because it is a more likely
cand id ate fo r a dd itio na l leverage and because the
bondholders have more to lose from the additional
FRBNY Quarterly Review/Spring 1990
Digitized 16
for FRASER
leveraging. The event risk premium thus acts as a pen
alty that decreases with leverage, as shown by the two
curves in Chart 1. This chart illustrates the two basic
effects of event risk. First, event risk raises the risk
premium schedule, ratcheting up borrowing costs over
a wide range of leverage. This effect may be termed
“ raising the intercept of the risk premium curve.” Sec
ond, event risk flattens the risk premium curve. This
lowers the marginal cost of additional leveraging: for a
given increase in leverage, the firm is facing a sm aller
increase in the risk premium. In effect, bond buyers
respond to event risk by making firms pay for leverage
whether they are leveraged or not, thus making it desir
able for firms to leverage up and reap the benefits of
additio na l debt. Even firm s th at are not o th e rw ise
inclined to take on new debt may be encouraged, if not
compelled, to do so by event risk premia.
Role of bond covenants
Event risk premia may not capture the total event risk
penalty paid by firm s since they do not reflect the
implicit costs of bond covenants. By issuing debt with
strong covenants, a firm may eliminate a great deal of
event risk and its accompanying premia. But while the
firm is no longer paying an explicit event risk premium,
it nevertheless may bear other costs in honoring the
covenant. Because covenants typically entail substan
tial indirect costs,1 including loss of flexibility in deci
sion making, they may not provide a way to reduce
overall costs.
Em pirical estim ation of event risk
This section assesses the extent to which event risk
premia have grown. The hypothesis that leverage risk
premia were sm aller before the mid-1980s implies that
differences in risk premia between high-leverage and
low -leverage firm s should have narrow ed over the
course of the decade.
Chart 2 provides casual empirical evidence for the
hypothesis. Note that since 1984 (roughly the start of
the takeover boom), the spread between Aaa corporate
and Treasury rates has grown relative to the spread
between Baa corporate and Treasury rates. This result
suggests a change in the perception of Aaa corporate
debt: once regarded as a near substitute for riskless
debt, Aaa c o rp o ra te d eb t is now an in te rm e d ia te
between Baa and riskless debt. This shift is consistent
with the flattening of the risk premium curve caused by
a rise in the risk premium on high-grade debt.
It should be remembered that both rate spreads tend
to widen during downturns, as Chart 2 suggests for the
"•Clifford W. Smith, Jr., and Jerold B. Warner, “ On Financial
Contracting: An Analysis of Bond Covenants,” Journal of Financial
Economics, June 1979.
period corresponding to the 1981-82 recession. During
the prolonged recovery period, however, the Aaa/Treasury spread has remained high, generally exceeding
1982 levels. In contrast, during the 1975-79 expansion,
Aaa/Treasury spreads were smaller, absolutely as well
as relative to Baa/Treasury spreads. The persistence of
the increased Aaa/Treasury spread during the 1980s
makes it unlikely that this change reflects fluctuations
in market liquidity conditions. Indeed, the massive sup
ply increase of Treasury debt over this period, com
bined w ith weak issuance of Aaa co rp orate d e b t,2
m akes the grow ing d is p a rity e sp e cia lly im pressive.
These considerations fu rthe r support the contention
that since the mid-1980s, event risk premia have risen
for high-grade debt in particular.
Analysis of individual firms
Chart 2 gives only partial evidence of the flattening of
the risk premium curve since it ignores changes in the
com position of the firm s in each rating category. A
regression analysis of individual firms will provide more
complete evidence of this development.
To this end, a sample of forty-seven firms (Table 1) is
selected to cover all industry classifications except
2Edward I. Altman, "Measuring Corporate Bond Mortality and
Performance," working paper, February 1988, Table 6, p. 16.
banking, financial services, and certain utilitie s (see
Table A1 in Appendix A for characteristics of sample
firm s). E le ctron ics, ve hicles, oil, and ch em ica ls are
overrepresented in the sample because of data avail
ability. All firms in the sample pool have been in contin
uous operation since at least 1976, and most have a
large share of publicly traded debt, particularly multiple
issues. Firms whose bonds contain call provisions that
are especially difficult to price are eliminated from the
sam ple. To be com parable, bond yield s should be
adjusted to remove the effects of such call options, but
the valuation can be problematic (see Appendix A).
Although representative size distribution is sought in
the sam ple, the se le ctio n c rite ria bias the sam ple
heavily toward large firm s. If event risk prem ia are
sm aller for very large firms because such firms make
more difficult takeover targets, then the analysis actu
ally underestimates the effects of event risk premia.
The sample dates chosen for the analysis are the
last business day in June for the years 1980, 1988, and
1989.3 The dates have the advantage of spanning the
leverage boom and of representing periods in which
3Data from 1982 gave results similar to those from 1980 data, while
data from 1984 gave results that were between 1980 and 1988
results. The data from 1982 and 1984 may not be directly
comparable with those from other years, however, because of
significantly higher nominal interest rates in these periods.
Chart 2
Credit Quality Spreads
Percentage points
3.5
Annual Averages
I
I Baa corporate rate
I___ I less Treasury rate
3.0
□
2.5
Aaa corporate rate
less Treasury rate
2.0
1.5
1.0
0.5
0
1975-79 Average
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
Source: Moody’s Bond Survey.
Note: Chart shows yield to maturity on seasoned corporate and Treasury debt with ten years to maturity. The 1990 yield is a first quarter average.
FRBNY Quarterly Review/Spring 1990
17
nominal interest rates were roughly equal (see Appen
dix A). In addition, interest rate vo la tility around the
o b s e rv a tio n p e rio d s w as low, a c o n d itio n w hich
ensures more reliable estimates of risk premia.
The firs t step in the a na lysis is to com pare the
Table 1
Firms in Sample by Industry C lassification
Firm
Aluminum Company of
America
American Telephone and
Telegraph
Amoco
Armco
Ashland Oil
Brunswick Corporation
Cabot Corporation
CBS
Centel
Chevron
Chrysler Corporation
Coastal Corporation
Colt Industries
Combustion Engineering
Contel
Corning
CSX Corporation
Cummins Engine
Delta Air Lines
Dow Chemical
Dupont
Exxon
Fairchild
Ford Motor Corporation
G e n e ra l E le c tric
Goodyear Tire and Rubber
GTE
Honeywell
International Business
Machines
ITEL
Kroger
Martin Marietta
McDonnell Douglas
Minnesota Mining and
Manufacturing
Panhandle East
PepsiCo Incorporated
Pitney Bowes
Procter and Gamble
Sara Lee
Teledyne
Texaco
Tosco Corporation
Trinova
United Telecom
Weyerhauser
Whirlpool
Williams Company
:'
Industry Classificationf
Metal refining
Communications
Petroleum refining
Steel refining
Petroleum refining
Miscellaneous manufacturing
Chemicals and pharmaceuticals
Communications
Communications
Petroleum refining
Vehicles
Fuel exploration
Vehicles
Machinery
Communications
Glass and concrete
Transportation
Vehicles
Transportation
Chemicals and pharmaceuticals
Chemicals and pharmaceuticals
Petroleum refining
Vehicles and electronics
Vehicles
Electronics
Rubber and plastics
Communications
Laboratory equipment
Machinery
Wholesale
Retail
Vehicles
Vehicles
Miscellaneous manufacturing
Natural gas and coal
Food and tobacco
Machinery
Chemicals and pharmaceuticals
Food and tobacco
Vehicles
Petroleum refining
Petroleum refining
Vehicles
Communications
Wood products
Electronics
Fertilizer, energy, and materials
flndustry classifications are based on a scheme developed in
William Lee, “ Corporate Leverage and the Consequences of
Macroeconomic Instability,’’ Federal Reserve Bank of New
York, unpublished working paper, 1989.
FRBNY Quarterly Review/Spring 1990
Digitized18
for FRASER
leverage of sample firms with the observed spread of
their bond yields over Treasury yields. The measure of
leverage used is market value of debt as a share of
total firm market value — the debt ratio defined above.
To calculate the market value of debt, the firm ’s publicly
traded debt is repriced at the prevailing market price.
All other debt is left at book value. The total value of
the firm is the sum of this market value of debt and the
market value of outstanding equity.4
The risk premium used is the average of the d iffer
ences in yield to m aturity between the firm ’s publicly
traded bonds and riskless debt of comparable m aturity
and coupon ch aracte ristics.5 One problem with this
measure is the timing of the yield quotes.6 The prevail
ing yield on a corporate bond reflects the most recent
transaction, which may have occurred several days ear
lier. Although such lags do not system atically bias the
slope of the estim ated curve relating risk prem ia to
leverage, they can system atically bias the intercept.7
Nevertheless, steady interest rates in the period imme
diately preceding the sample dates make substantial
bias unlikely, as do the active markets in high-grade
bonds, the debt most affected by event risk premia.
Other potential problems associated with the mea
sure include intertem poral biases induced by system a
tic changes in the m aturity structure of outstanding
corporate debt or by changes in the relation of the
m aturity and seniority of traded and nontraded debt.
These problems do not appear serious, however (see
Appendix A).
4Book value of debt and market value of equity are taken from the
COMPUSTAT data base. Market price of outstanding bonds is taken
from Standard and Poor's Bond Guide (July 1980, July 1988, and July
1989).
5The risk premium is mathematically defined as follows:
k
2
Wj
* (
rf
-
r < ),
i= 1
where k = number of publicly traded bond issues
W; = par value bond /' / total par value of publicly traded debt
r f = yield to maturity of corporate bond /, adjusted for
callability
fif = yield to maturity on riskless debt corresponding to
bond /'.
6Corporate bond yields are taken from Standard and Poor's Bond
Guide (July 1980, July 1988, and July 1989) and represent closing
quotes from the last business day in June. Government bond yields
are from the Wall Street Journal for corresponding days.
7Consider the following example: Interest rates are rising in a market
where the riskless debt is more heavily traded than the corporate
debt. Since the price quotes on the corporate debt are necessarily
older, their yields will be biased downward relative to the riskless
debt. The risk premia on the corporate debt will not be biased
relative to each other, however, although there will be more noise
around the “ true" risk premia.
In calculations of risk premia, an effort is made to minimize the
quote lag problem by taking quotes during weeks in which there was
little interest rate movement and by relying on the more heavily
traded corporate bonds.
slopes of the lines in 1988 and 1989, particularly when
cash flow volatility is included as an independent vari
able in the regression.9
The flattening can be seen more clearly in Chart 3,
which plots the curves generated by the regression
results from Table 2. The risk premium curves for 1988
and 1989 are actually below the 1980 curve for high
debt ratios. There are three possible explanations for
this finding. First, as explained in greater detail below,
changes in general liquidity conditions in fixed income
markets contributed to an upward bias in the 1980 risk
premium curve. Second, because of the low number of
observations for debt ratios above 0.80, the 1988 and
1989 curves are probably estimated as being more lin
ear than they actually are.10 The true curve for these
The next step is to calculate the vo la tility of total
cash flow (pretax profit plus depreciation plus interest
paid) over book value for each of the firms.8 All other
things equal, the firm with lower cash flow volatility will
be less likely to default and should therefore pay a
lower risk premium. Cash flow volatility will serve as an
additional independent variable in some of the regres
sions perform ed below.
The risk premium is then modeled as a function of
the debt ratio. As noted earlier, risk premia increase
with leverage, although the exact relation is an em piri
cal question. Risk premia are first modeled as a linear
function of leverage: risk premium = a + b * debt
ratio. T his re la tio n w orks, but b e tte r re s u lts are
achieved with nonlinear models (see Appendix B for a
complete list of test results). The best overall results
are obtained with the following relation: risk premium
= a + b * (debt ratio)2-6.
The first part of Table 2 shows the results obtained
by regressing risk premia solely on debt ratios. Note
that the zero-leverage intercept has risen in each of
the periods while the slope of the line has fallen. Sta
tistical tests on the regressions indicate that the slope
of the line in 1980 is sign ifican tly different from the
9F-tests are performed to test the significance of the difference in the
coefficient on the debt ratios between the 1980 and 1988 regressions
and the 1980 and 1989 regressions. The results are as follows:
Regression: PRE = a + b * LEV
FDegrees of Freedom
Comparison
Statistic
Numerator Denominator Significance
1980 vs. 1988
4.76
1
77
0.95 +
1980 vs. 1989
3.83
1
77
0 .9 5 Regression: PRE = a + fc>, * LEV + b2 * VOL
FDegrees of Freedom
Comparison
Statistic
Numerator Denominator Significance
1980 vs. 1988
9.10
1
76
0.995 +
1980 vs. 1989
7.00
1
76
0.99 +
8Cash flow volatility is measured as a twelve-quarter rolling variance
of quarterly cash flow rates:
Volatility = i ( Q - Ct )2 / 12,
where
PRE = required risk premium
LEV = debt ratio
VOL = variance of cash earnings (see footnote 8).
i=t -11
where
^ _ pretax profits + depreciation + interest paid in quarter /
book value of firm at end of quarter i
10Firms with debt ratios over 0.90 are excluded from the sample,
partially because the exponential curve underestimates these values
Ct = 1/12 * 2 Q.
i-t -11
Table 2
Required Risk Premium as a Function of Leverage
Regression Equation: PRE = a0 + 6*LEV26
1980
1988
1989
ao
b
Standard Error
of b
Adjusted
R2
Degrees
of Freedom
0.032
0.366
0.579
5.046
2.710
2.288
0.996
0.566
0.650
0.428
0.323
0.198
32
45
45
Required Risk Premium as Function of Leverage and Cash Flow V olatility
Regression Equation: PRE = a0 + 5 ,‘ LEV2-6 + £>2*V0L
ao
1980
1988
1989
-0 .2 3
0.13
0.17
b1
b2
Standard Error
of b.
Standard Error
of b2
Adjusted
R2
Degrees
of Freedom
4.65
2.26
2.17
31.30
18.01
22.99
1.01
0.53
0.61
21.27
5.41
8.11
0.448
0.447
0.306
31
44
44
Notes: PRE = risk premium over riskless rate for debt issues of individual corporations; LEV = debt ratio = market value of debt over
total (debt plus equity) firm market value; and VOL = variance of earnings before interest and tax payments.
FRBNY Q uarterly Review/Spring 1990
19
years should probably be flatter for debt ratios below
0.70 and steeper for debt ratios beyond that. Third, as
m arkets for lower quality debt have improved in the
1980s, the yield on debt of more leveraged firms has
actually fallen (although recent experience indicates
that this trend may be reversing).
The second part of Table 2 presents the regression
results obtained with the addition of cash flow volatility
as an independent variable. In each case the regres
sion fit is improved without a substantial change in the
coefficient on leverage.
The co efficie nt on cash flow vo la tility falls sharply
between the 1980 sample and the 1988 and 1989 sam
ples, dropping from 31.3 in 1980 to 18.0 in 1988 and
23.0 in 1989. This effect is consistent with the results
for the coefficient on leverage: low cash flow volatility
no longer guarantees the firm cheap debt financing; the
market assumes that firms with low cash flow volatility
will be leveraged up.
Treasury debt was higher in 1980 than in 1988 or 1989,
biasing the 1980 risk premium curve upward relative to
the 1988 and 1989 curves.
To obtain this finding, a proxy for the fixed-income
liquidity premium in each of the years was estimated by
comparing Treasury debt yields with yields on com pa
rable World Bank debt and Federal National Mortgage
Association debt (the procedure is described in Appen
dix A). The liquidity premium paid over Treasury debt
by such issuers was about 25 basis points higher in
1980 than in 1988 and 1989.
Adjusting for the change in liqu id ity prem ia would
shift the 1988 and 1989 curves in Chart 3 up by 25
basis points relative to the 1980 curve. Such an adjust
ment would suggest, for example, that the required pre
mium on a firm with a 40 percent debt ratio is about 45
basis points higher in 1988 or 1989 than in 1980 owing
to event risk.
Liquidity conditions and the risk premium curve
The difference between corporate and Treasury yields
reflects not only the risklessness of government bonds,
but also their higher liquidity. Evidence suggests that
the liquidity premium paid by corporate issuers over
The long-run increase in debt costs from event risk can
be quantified. If the upshift in the risk premium curve is
about 15 basis points on average,11 then U.S. corpora
tions eventually stand to pay an additional $1.33 billion
(15 basis points on $885 billion corporate bonds out
standing in 198812) annually on their notes and bonds
alone. Initially, bond investors take the loss, but corpo
rations must pay as maturing bonds are refinanced.
Most studies of event risk proceed by measuring the
losses incurred by bondholders of buyout targets and
weighing the losses against the gains realized by the
target’s stockholders. One of the largest estimates of
lost bondholder wealth has been advanced by Asquith
and Wizman, who calculate that, on average, bond
holders lose 2.5 percent of bond value from a buyout,
with greater losses for bondholders lacking covenant
protection.13 Asquith and Wizman note that even if the
losses are applied to all target firm s’ debt, bondholder
Q uantifying the debt costs of event risk premia
Footnote 10 continued
and partially because the illiquidity of high-yield debt — particularly in
1980 — makes comparisons unreliable.
Chart 3
Required Risk Premia on Corporate Bonds as a
Function of Debt Ratio
Percentage points
2.5
2.0
1.5
11The model gives this figure as the direct increase to an average
(debt ratio = 0.50) corporate bond issuer over the period 1980 to
1989 without an adjustment for liquidity. Such an adjustment would
add another 25 basis points.
1.0
12Board of Governors of the Federal Reserve System, Flow of Funds
Accounts: Financial Assets and Liabilities, Year-End, 1965-1988,
September 1989.
0.5
0
0.1
0.2
0.3
0.4
Debt ratio
0.5
0.6
Note: PRE=required risk premium; LEV=debt ratio.
20for FRASER
FRBNY Quarterly Review/Spring 1990
Digitized
0.7
13Paul Asquith and Thierry A. Wizman, “ Event Risk, Wealth
Redistribution and the Return to Existing Bondholders in Corporate
Buyouts,” Journal of Financial Economics, forthcoming. The losses
reported represent abnormal returns over the period from month-end
two months before the determination of buyout until month-end two
months after the announcement of outcome. The sample is taken
from sixty-five large completed buyouts.
Asquith and Wizman calculate losses of 5.4 percent for
bondholders with no covenant protection and losses of 2.8 percent
for bondholders with weak covenant protection, while bondholders
with strong covenant protection experience gains of 2.3 percent.
MBNHRHMBI
UK
Box: Event Risk Premia and Additional incentives fo r Leveraging
The reduction in leverage disincentives brought about
by event risk can be measured more precisely by calcu
lating the marginal increase in debt cost resulting from
additional leverage. Specifically, one can calculate the
increase in total risk premiums paid as a result of a 1
percentage point increase in the debt ratio (total risk
premiums paid at [x + 1] percent debt ratio less total
risk premiums paid at x percent debt ratio). Mathe
matically, this can be expressed as follows:
TRP = LEV * PRE = LEV * { a + b * ( LEV2-6)},
where TRP = total risk premium as percentage of
firm value
LEV = debt ratio
PRE = required risk premium.
The relation can be expressed as follows:
Marginal cost of leveraging =
= a + ( 3.6b * LEV2 6 ).
The marginal increase in total risk premia associated
with a percentage point increase in the debt ratio is
shown as the marginal cost curve in the chart. Note
that this marginal increase is on average about 40 per
cent lower for 1988 than for 1980 (the 1989 curve is
omitted for readability). Put another way, the marginal
penalty for leveraging exacted by the bond market has,
on average, fallen by about 40 percent because of
event risk.
Let us now compare the marginal risk premia cost of
leverage with the marginal tax benefits of leverage. The
marginal tax benefit of leverage is defined as the
increase in the total tax shield resulting from a 1 per
cent increase in the debt ratio.
The total tax shield can be expressed as:
tc * { LEV * [ r, + PRE ]} = tc * { LEV * [ rf
+ ( a + b * LEV2-6]},
where tc = top bracket corporate tax rate
r, = risk-free interest rate.
The marginal change in the total tax shield with respect
to leverage is then:
tc * { rf + a + ( 3.6b * LEV2 6)}.
If there were no risk premia, then the marginal tax
benefit function would be a flat line equal to the riskless
rate multiplied by the corporate tax rate. Since risk pre
mia are increasing, however, marginal tax benefits are
an increasing function of leverage.
The marginal tax benefit curve is calculated using
1989 tax rates (34 percent, state and local rates not
considered) in each of the years. This step is taken in
order to isolate the event risk premia effects from tax
effects that occurred over the same period, since a shift
in corporate tax rates can move the marginal benefit
curve in a way that amplifies or dampens the effect of
the event risk-induced shift of the marginal cost curve.
The chart shows the marginal risk premia cost curves
against marginal tax benefit curves for the 1980 and
1988 samples. Note that the marginal tax benefits of
leveraging are slightly higher in the 1980 case because
of the steeper risk premium schedule. For the compari
son we use a riskless interest rate of 8 percent, roughly
the rate prevailing in the three time periods.
The “ optimal” debt ratio is reached at the point where
the marginal cost and benefit curves intersect, although
factors such as strategic incentives for debt are not
reflected in the ratio. It is still worth noting, however,
that the “equilibrium” debt ratio suggested by the model
rises from 0.56 in 1980 to 0.69 in 1988, and to 0.72 in
1989 (not shown on chart). The 1980 value is a refer
ence point, not an actual historical estimate, because it
is calculated using 1989 tax rates.
Note that the optimal debt ratio also depends on the
interaction of equity costs and the debt ratio. Appendix
B develops a model of this relation that suggests that
the optimal debt ratio is almost completely determined
by the equation of the marginal costs and benefits of
debt.
Marginal Tax Benefits versus Marginal increases in
Total Risk Premia
Percentage of total firm value
1980 marginal
/
increase in risk premia I
/
1980 marginal
tax benefit
**■
/
I
/
■L
tax benefit
*-
|
1988 marginal
increase in risk premia
I
0.1
-------- - l
0.2
0.3
l
I
0.4
0.5
Debt ratio
I
0.6
I
0.7
I I
0.8
Note: Calculations assume an 8 percent riskless interest rate.
FRBNY Q uarterly Review/Spring 1990
21
losses are less than 8 percent of stockholder gains.
Event studies, however, measure only part of the
bondholder loss from buyouts since they only consider
the impact upon directly affected bondholders. Costs of
event risk that follow the upshift in the risk premium
curve are incurred by firms not involved in buyouts —to
the extent that they have to refinance at higher interest
rates — and by the bondholders of these firms —to the
extent that the value of their longer term debt holdings
is eroded by unexpected rises in event risk. These
losses should be added to the losses of bondholders of
the involved firms. Our conservative estimate of a 15
basis point upshift in the risk premium curve repre
sents (at a 10 percent capitalization rate) a 1.5 percent
discount in the value of bonds issued by uninvolved
firms. Again, this loss applies to all bond-issuing firms,
not just those involved in buyouts. It appears, there
fore, that indirect bondholder and issuer losses from
event risk are far greater than the direct losses borne
by bondholders of buyout targets.
Im plications of event risk premia
The advent of event risk premia has had several im pli
cations for corporations. First, firms now. find it much
less attractive to finance them selves largely through
equity. T hus fin a n c ia l m arke t d e ve lo p m e n ts w hich
would appear to have increased the range of options
for corporate treasurers by making high leverage more
feasible may have actually narrowed the range by mak
ing the choice of low to m oderate leverage more
expensive. Second, the increase in the cost of debt for
firms of low to moderate leverage raises the possibility
that the corporate cost of capital has risen. Third, event
risk premia tend to amplify the effects of tax measures
aimed at decreasing corporate leverage. Each of these
implications is examined in greater detail below.
Incentives for leveraging
Having learned the hard lesson that the strong can
abruptly become the weak, investors in the bond mar
ket tend to discount the bonds of financially strong
firms. As a result, firms of low and moderate leverage
find them selves already paying for a portion of any
increase in leverage they may be contem plating. In
other words, the change in the market environment has
moderated the increase in the interest rate associated
with an increase in leverage.
The Box calculates a measure of the incentives to
shift to debt finance created by event risk premia. The
estim ated m arginal costs of leverage resulting from
increased risk premia are compared to an estimate of
the marginal tax benefits of leverage resulting from the
deductibility of interest. The calculation in the example
implies that, when all other factors are held constant,
FRBNY Quarterly Review/Spring 1990
Digitized 22
for FRASER
the debt ratio balancing tax benefits and escalating
debt costs rises by 16 p ercentage points over the
period 1980-89. Although the impact of this change
must be considered in conjunction with other in flu
ences on capital structure, these calculations suggest
that event risk premia may have been responsible for
inducing sizable increases in debt during the 1980s.
Event risk premia and the cost of capital
The risk that w ell-capitalized firms might be leveraged
up raises the cost of debt by increasing the risk pre
mium paid by all but the most leveraged borrowers. In
turn, the rising cost of debt could conceivably entail an
increase in the overall cost of capital.14
One can argue to the contrary that if share prices
rise to reflect the potential for substantial gains to
shareholders from leveraging transactions, the higher
cost of debt might be offset. But while this argument is
a ttra c tiv e in the a b s tra c t, p ra c tic a l c o n s id e ra tio n s
14Albert Ando and Alan L. Auerbach, “ The Cost of Capital in the U.S.
and Japan: A Comparison,” Journal of Japanese and International
Economics, vol. 2 (1988), pp. 134-58; and Robert N. McCauley and
Steven A. Zimmer, "Explaining International Differences in the Cost of
Capital,” this Quarterly Review, Summer 1989, pp. 7-28.
Chart 4
Incentives for Debt Financing under an Interest
Deduction Cap
Marginal Change in Total Interest Deduction:
Ratio of Change with Interest Deduction Cap to Change
without Cap
Percent
pL-l
0.25
I
l
0.35
I
J—
0.45
1 .. 1
.1.
1
0.55
0.65
Debt ratio
I
1
0.75
I
1
0.85
I
I I
0.95
weigh against it. Consider the position of a treasurer in
a well-capitalized firm who watched in the 1980s as
debt-financed takeovers drove up the yields on prime
corporate debt relative to Treasury yields. Would that
treasurer be easily persuaded that the measurable
costs of event risk premia could be offset by the much
less tangible benefits of lower equity costs?
Tax policy
Tax policies to limit corporate leverage are more effec
tive in the presence of event risk premia. Any weaken
ing in the tax incentives for debt financing as opposed
to equity financing will tend to mitigate event risk.
Therefore, in addition to its direct effect on leverage
incentives, a reduced tax incentive for debt finance will
work to steepen the risk premium schedule. Even poli
cies aimed at reducing extreme leverage will raise the
marginal cost of debt for firms with low or moderate
leverage. As an example, consider the elimination of
deductibility for corporate interest payments paid at a
rate above some specified spread over Treasury yields.
A cap on interest deductibility reduces the incentive for
leverage over a surprisingly wide range of leverage.
The effect of such a cap may be measured as the ratio
of the marginal increase in the value of the interest
deduction with the cap to the value of the interest
deduction without the cap (Chart 4; Appendix 2 pro
vides the complete derivation). For instance, even a
4 percent cap reduces the marginal tax benefit by over
one-half for a low-volatility firm with a debt ratio of 80
percent. A cap becomes effective at lower leverage for
a firm with a more volatile cash flow because bond
holders demand compensation for the greater risk of
debt service difficulties entailed by such a cash flow.
By sharply reducing the worst risks to bondholders
and thereby steepening the risk premium schedule,
even caps set quite high can work indirectly to discour
age additional debt over the entire range of leverage.
This observation does not constitute a case for such a
policy, which would affect firms that are risky for rea
sons other than high leverage and which might pose
adm inistrative challenges. But factoring individual
changes in risk premia into the assessment of tax pol
icy can show it to have greater potency than has been
acknowledged in the past.
Conclusion
Analysis of U.S. corporate bond yields demonstrates
that investors learned to demand a higher premium
over U.S. Treasury yields in the course of the 1980s.
This rise in the risk premium schedule resulted from
mergers and acquisitions activity that made maintaining
a given firm ’s credit standing more uncertain. Debtfinanced changes in corporate control changed the
bond market environment for less leveraged firms and
made it more difficult for those firms to keep debt costs
low by maintaining a comfortable equity cushion.
The least leveraged firms experienced the largest
increase in borrowing costs relative to Treasury yields.
Consequently, the marginal cost of leverage declined
enough to induce firms, according to one calculation, to
increase the debt fraction of total financing by 16 per
centage points.
The rise in the cost of debt has been significant. The
fear of debt-financed takeover activity has caused a
percentage reduction in the value of all U.S. corporate
bonds that is comparable to the percentage reduction
in the value of bonds immediately affected by takeover
events. The effect of event risk premia on the cost of
capital, however, is ambiguous. The demonstrated rise
in debt costs for less leveraged firms must be balanced
against the effect of takeover premia in cheapening the
cost of equity.
The analysis suggests that tax policy can exert a
particularly powerful effect in the presence of event
risk premia. Any reduction in the tax benefits of
leveraging, even narrowly, drawn reductions, can indi
rectly discourage leveraging across the spectrum of
firms.
Steven A. Zimmer
FRBNY Quarterly Review/Spring 1990
23
mmmm
tBmmmKammmmmmmmmmmm
Appendix A: C haracteristics of Sample Firms
Callable bonds
The yields of the corporate bonds in the sample are not
fully comparable with Treasury yields because most of
them reflect some type of call provision. To make the
yields on different corporate bonds comparable, it is
necessary to make price adjustments for the callability
of certain issues.
The call option on corporate bonds can be priced
using contingent claims theory if certain assumptions
are made about adjustment to a “ natural” interest rate.
We elect not to use this method, however, because it
generally yields poor re s u lts .t The m ethod’s basic
weaknesses are compounded by heterogeneous per
ceptions of “ natural” interest rates and debt issuance
costs, which drive a wedge between holder and issuer
call valuation.
To get around the call valuation problem, we use noncallable debt whenever possible. If forced to value a
call option, we try to use debt with call valuations that
are deeply out of the money, although in some cases
we use yield to call on debt that is clearly going to be
called.
In valuing calls, we follow a conservative application
of the call value techniques of Kim, Ramaswamy, and
Sundaresan and of Gastineau.t
fThe best attempt to value debt through a contingent claims
analysis is probably E. Phillip Jones, Scott P. Mason, and Eric
Rosenfeld, "Contingent Claims Valuation of Corporate
Liabilities: Theory and Empirical Tests," National Bureau of
Economic Research, Working Paper no. 1143, June 1983.
The debt pricing in this study returns an average absolute
error of 6.05 percent on market value of debt, although the
estimates with an average prediction error of 0.64 percent,
are relatively unbiased.
tJoon Kim, Krishna Ramaswamy, and Suresh Sundaresan,
"Valuation of Corporate Fixed Income Securities," working
paper, December 1986; and Gary L. Gastineau, The Options
Manual (New York: McGraw Hill, 1988), pp. 347-64.
Table A1
§Robert C. Merton, “ On the Pricing of Corporate Debt: The
Risk Structure of Interest Rates,” Journal of Finance, vol. 29
(1974).
Table A2
C haracteristics of Sample Groups
;
Average debt ratio
Average risk premium
Average earnings
rate variance
Average time to
maturity of debt
Maturity of bonds in sample
The regression results are potentially sensitive to the
maturity of the bonds used to determine the risk pre
mium. In particular, the maturity of the bonds in the
sample can affect the slope of the curve relating risk
premia to leverage.
If the risk premium is a function of the time to matu
rity, then any systematic relationship between time to
maturity and leverage will bias the slope of the risk pre
mium curve. For example, consider a situation in which
low-debt firms in one sample year tend to have shorter
maturity debt than their more leveraged counterparts. If
risk premia are an increasing function of time to matu
rity, then the risk premia of the less leveraged firms will
be biased downward relative to the risk premia of the
more leveraged firms, steepening the apparent slope of
the risk premium schedule.
Ideally, the risk premia in our sample should be inde
pendent of time to maturity. Table A1 shows the results
obtained by regressing time to maturity on leverage in
each of the sample years. The coefficient on time to
maturity is negative in each of the regressions. The
coefficients for the 1988 and 1989 cases are of much
greater magnitude than the coefficient for the 1980
case, which is not significantly different from 0.
The importance of these results depends upon the
relationship between time to maturity and the required
risk premium. Theoretical work by Merton and by Pitts
and Selby concludes that for investment grade bonds,
the required risk premium rises with time to maturity out
to one year and then declines very slowly afterward.§
The required risk premium on lower grade debt declines
asymptotically with time to maturity. Later work by Sarig
and Warga provides strong empirical support for these
Regressions of Leverage on Time to M aturity
1980
1988
1989
0.540
1.220
0.529
0.983
0.562
1.190
0.012
0.019
0.019
14.25
11.36
FRBNY Quarterly Review/Spring 1990
Digitized24
for FRASER
9.62
Debt ratio = a + b * (time to maturity of debt)
1980
1988
1989
a
b
error of b
0.5551
-0.00003
0.00004
0.0267
0.6556
-0.01113
0.00456
0.1171
0.6307
-0.00714
0.00481
0.0466
Appendix A: C haracteristics of Sample Firms (continued)
theories about the time pattern of the premia.||
Since the average time to maturity of bonds in our
sample is roughly ten years (Table A2), we can treat the
risk premium as a downward sloping function of time to
maturity. Since the coefficient on time to maturity is
negative in each sample period, we know from our ear
lier argument that the slope of the risk premia curves in
each of the sample periods is biased upward. Further,
this bias is stronger for the 1988 and 1989 samples,
suggesting that we underestimate the actual extent of
the flattening of the risk premium curve.
Timing of rate quotations
As noted in the article, shifts in interest rates around
the sampling period can bias the estimated intercept of
the risk premium curve. This occurs because Treasury
debt is more heavily traded than corporate debt, with
the result that the observed yields on corporate securi
ties are less cu rre n t than those on governm ent
securities.
Examination of interest rate movements immediately
before the observation period helps to determine the
HC.G.C. Pitts and M.J.P. Selby, "The Pricing of Corporate Debt:
A Further Note," Journal of Finance, vol. 38 (1983); and Oded
Sarig and Arthur Warga, "Some Empirical Estimates of the
Risk Structure of Interest Rates," Journal of Finance,
December 1989.
direction of the bias — if interest rates are falling in the
period before the observation, then the intercept will be
biased upward. The patterns of interest rate movements
in the period preceding the sample date have additional
significance because the spread errors induced by them
will depend upon how recently the issue was traded and
will therefore vary across debt issues. This does not
bias the slope of the risk premia curve, but it does
increase the variance of the individual observations
around the curve and is a potential source of estimation
error.
Chart A1 traces the path of changes in the yield to
maturity of the benchmark five-year government bond
(noncallable, five years to maturity, closest to par cou
pon) over the week ending in the observation date.
Note that in 1980, 1988, and 1989 the Treasury yields
are similar and are relatively steady for the week pre
ceding the sample period. The rate rises somewhat in
1980, suggesting an underestimation of the risk inter
cept, and falls in 1989, suggesting an overestimation.
The rate is steady in 1988. As a rough estimation of the
intercept bias introduced by the rate movement, we
take the average of the rate observations prevailing in
the week before the observation date and compare it to
the rate prevailing on the observation date:
Year
1980
1988
1989
Chart A1
Five-Year Rate
Last Week Average
Observation Date
9.19
8.47
8.24
9.32
8.46
8.13
Five-Year Treasury Rates
The difference between the average of the week’s
observations and the final observation may be taken as
a rough measure of the intercept bias on the risk pre
mium curve. The 1980 curve is therefore biased down
ward roughly 14 basis points relative to the 1988 curve,
and 24 basis points relative to the 1989 curve.
Percentage points
1982
— -
_ ____
1984
— ————
-man
----------------------------I
23
~ 1988
_
]
---------ft---——o------------Q
• ---------•------------ •----------,_______ _____________ _— m
1986
/
I
I
I 1989 I
.1..
1
1
24
25
26
27
28
29
Jun
Note: Rates are yield to maturity on par coupon Treasury debt
with five years to maturity.
30
Liquidity premia
The yield difference between Treasury debt and certain
government agency debt can be used as a measure of
liquidity since some agency debt is essentially riskless,
but less liquid than Treasury debt. The best agency for
this calculation is probably the Federal National Mort
gage Association (FNMA), which has large issues of
nearly riskless straight debt with coupon values similar
to Treasury debt. FNMA debt does pose two problems,
however: First, its credit quality is thought to have
improved slightly between 1980 and 1988, biasing the
observed liquidity spread down slightly. Second, FNMA
debt, unlike Treasury debt, is not deductible at the state
FRBNY Quarterly Review/Spring 1990
25
Appendix A: C haracteristics of Sample Firms (continued)
and local level. Because federal personal income tax
rates have fallen since the early 1980s, the effective tax
on FNMA debt has risen relative to Treasury debt, bias
ing the observed liquidity premium upward.* Neverthe#The tax rate on FNMA debt exceeds the tax paid on Treasury
debt by the effective state and local tax rates. This difference
varies with the tax status of the holder. We can express the
effective state and local rate for a given tax bracket as
fs * (1 - tf), where ts is the combined state and local tax rate
and tf is the federal tax rate. The subtraction of the federal
rate reflects the deductibility of federal taxes from state and
local taxes. Since the top bracket personal federal tax rate
Chart A2
Liquidity Premia
Yield on Ten-Year Federal National Mortgage Association
Debt less Yield on Ten-Year Treasury Debt
Basis points
6 0 --------------------------------------------------------------------1980
less, since the two potential sources of error appear
small and work in opposite directions, we will assume
that they can be ignored.
Chart A2 shows the FNMA/Treasury spread for com
parable ten-year debt* over the last two weeks of June
in each of the years. The observed liquidity premium is
clearly higher for 1980 than for 1988 or 1989. Over the
final week of June, the spread averaged 52 basis points
in 1980 as against 33 basis points in 1988 and 26 basis
points in 1989.
The liquidity spreads on FNMA debt are consistent
with our observations of liquidity spreads on World
Bank debt. World Bank debt is comparable to FNMA
debt but is characterized by lower issue volume and
liquidity, particularly for longer maturities.
At the end of June 1980, five-year World Bank bonds
were yielding 81 basis points more than comparable
Treasuries.ft This difference fell to 47 basis points in
1988 and 63 basis points in 1989. This drop again rep
resents a downward shift of about 25 basis points in the
liquidity premium between the earlier and later periods.
* continued
has fallen from 70 percent to 34 percent in the period
between the early and late sample periods, we may assume
that federal rates have fallen more than statutory state and
local rates and that the effective state and local rate has
therefore risen.
•The yield to maturity is for debt with similar coupons that
matures in ten years from the observation date.
,ol I I. I. I. I I I I I I I I I I . I
15 16 17 18
19 20 21 22 23 24 25 26 27 28 29 30
Jun
Note: Rates are yields on comparable bonds with ten years
to maturity.
tfR a te comparisons use yields on World Bank debt for the last
day in June on which the issue traded. The Treasury rate is
taken as the yield to maturity on comparable (five years to
maturity, similar coupon characteristics) Treasury debt using
the average of several observations from the day of World
Bank debt observation:
Date
Treasury Yield
World Bank Yield
Spread
6-24-80
6-29-88
6-28-89
8.91
8.44
8.25
9.72
8.91
8.88
0.81
0.47
0.63
Appendix B: The Risk Premium Curve and the Optimal Debt Ratio
The cost of equity and the optimal debt ratio
We define the cost of capital as the weighted average of
debt and equity costs. The firm determines the optimal
debt ratio as the debt ratio that minimizes CK, the cost
of capital (more properly, the “ cost of funds” ):
CK = (LEV * D) + {(1 - LEV) * £},
where LEV is the debt ratio, D is the real after-tax cost
of debt, and E is the average cost of equity.
26for FRASER
FRBNY Q uarterly Review/Spring 1990
Digitized
The firm solves:
(
+ (<1 - LEV> - s r b )
le v
- s ib )
= °
The first term represents the simple substitution of a
unit of equity for a unit of debt. The second term repre
sents the resulting change in average debt costs multi
Appendix B: The Risk Premium Curve and the Optimal Debt Ratio (continued)
plied by the size of the debt share (the term
is
simply the slope of the risk premium curve on an after
tax basis), and the third term represents the resulting
change in average equity costs multiplied by the size of
the equity share.
We can prove that the equilibrium debt ratio used in
the text satisfies the above equation if we make certain
assumptions about a CAPM model. The first step is to
rewrite D, the real after-tax cost of debt, as D = rr - (tc
* ir), where rr is the real interest rate facing the firm, tc is
the corporate tax rate, and it is the nominal interest rate
facing the firm . We can then rew rite our optim al
leverage equation:
d
L E V
+
( rr - E
)
( < 1 ' LEV» ‘ ^
+
(
)
‘ (*C
*
>r)
+
L E V
*
d
L E V ^
-°-
The second term in the equation is now simply the
difference between the marginal cost of leveraging and
the marginal tax benefit. At the equilibrium point calcu
lated in the text, these two values are equal, and con
sequently the second term disappears. We now have to
show that the first and third terms are equal.
If we accept CAPM, we can write E = rf + (3 * ( Em rf ), where rt is the risk-free rate of interest and Em is the
Table B1
Required Risk Premium as a Function of Leverage
Regression
Equation
b
Standard Error
of b
Adjusted
R2
Degrees
of Freedom
-1.281
-0 .4 4 8
-0 .3 2 5
4.700
2.701
2.700
1.193
0.580
0.785
0.306
0.311
0.190
32
45
45
PRE = a0 + £>*LEV15
1980
-0 .5 8 5
1988
-0 .0 2 5
0.149
1989
4.478
2.528
2.389
1.045
0.528
0.682
0.345
0.322
0.197
32
45
45
PRE = a0 + /b*LEV20
-0 .2 2 7
1980
0.199
1988
0.397
1989
4.635
2.561
2.299
1.000
0.530
0.651
0.383
0.327
0.200
32
45
45
PRE = a0 + b*LEV2 2
-1 .1 2 6
1980
0.263
1988
0.468
1989
4.752
2.601
2.288
0.994
0.539
0.648
0.398
0.326
0.199
32
45
45
PRE = a0 + £>*LEV24
-0.041
1980
1988
0.318
0.527
1989
4.890
2.652
2.285
0.994
0.551
0.648
0.413
0.325
0.199
32
45
45
PRE = a0 + £>*LEV26
1980
1988
1989
5.046
2.710
2.288
0.996
0.566
0.650
0.428
0.323
0.198
32
45
45
PRE = a0 + b*LEV
1980
1988
1989
ao
0.032
0.366
0.579
log(PRE) = a0 + 5*log(LEV)
1.326
1980
0.625
1988
1989
0.459
2.482
0.506
0.411
32
1.262
0.225
0.399
45
0.895
0.310
0.137
45
J ~ J f ¥* J
-'
. ::::Z
Notes: PRE = risk premium over riskless rate for debt issues of individual corporations; LEV = debt ratio * market value of
debt over total (debt plus equity) firm market value.
FRBNY Quarterly Review/Spring 1990
27
Appendix B: The Risk Premium Curve and the Optimal Debt Ratio (continued)
market equity return. The cost of equity then varies with
leverage through the beta coefficient. The beta coeffi
cient varies with the debt ratio as fo llo w s : ^ ^ = frc fv Here we assume that the firm has systematic asset risk
that is completely absorbed by the equity; hence the
equity beta varies with the inverse of the equity ratio. In
reality, we would expect the beta coefficient for a firm’s
debt to increase with leverage, accounting for some of
the systematic asset risk, but bond betas tend to be
very low and can be ignored for our purposes.
We can now write
• The third term of
the optimum debt equation then becomes (£m - rf).
Setting the second term of the optimum leverage equa
tion equal to zero, we obtain
= ( Em - E ) + ( rr
- rf ). For a market representation of firms we have £m
= E. The value then becomes ( rT - rf ), a positive but
very small number because it is taken over a differential
change in leverage. Recall that our equation for
overlooked the fact that not all of the systematic asset
v o la tility would remain in equity as we leveraged.
Adjusting for this omission would lower the value of the
third term and move the sum of the first and third terms
toward zero. In addition, the extent to which systematic
asset volatility is not absorbed by equity is a positive
function of leverage, just as (rf - rm) is. This suggests
that the relation between the first and third terms is
Table B2
Required Risk Premium as a Function of Leverage and Cash Flow V olatility
Regression
Equation
Standard Error Standard Error
of
of b2
Adjusted
R2
Degrees
of Freedom
23.05
5.50
8.25
0.341
0.431
0.287
31
44
44
1.05
0.50
0.64
22.51
5.46
8.19
0.375
0.441
0.296
31
44
44
33.75
17.75
22.53
1.01
0.50
0.61
21.96
5.42
8.14
0.408
0.446
0.303
31
44
44
PRE = ao + b-i*LEV22 + 62*VOL
1980
-0 .3 9
4.36
1988
0.05
2.16
1989
2.16
0.08
32.92
17.82
22.67
1.00
0.51
0.61
21.73
5.42
8.13
0.422
0.447
0.304
31
44
44
PRE = ao + £>i,*LEV24 + £>2*VOL
-0.3 1
4.50
1980
1988
0.10
2.20
2.16
1989
0.13
32.11
17.90
22.83
1.01
0.52
0.60
21.50
5.41
8.12
0.435
0.447
0.306
31
44
44
PRE - ao + b.,*LEV2 6 + £>2*VOL
1980
4.65
-0 .2 3
1988
0.13
2.26
0.17
2.17
1989
31.30
18.01
22.99
1.01
0.53
0.61
21.27
5.41
8.11
0.448
0.447
0.306
31
44
44
0.21
0.28
0.25
0.51
0.22
0.31
0.24
0.11
0.15
0.407
0.462
0.166
31
44
44
6,
b2
4.24
2.22
2.49
38.00
17.86
21.99
1.19
0.54
0.74
PRE = ao + bi.♦LEV1-5 + 62*VOL
1980
4.07
-0 .8 3
-0 .1 7
1988
2.09
1989
2.22
-0 .2 0
35.85
17.71
22.21
PRE = a0 + b-i*LEV20 + £>2*VOL
-0 .4 9
4.24
1980
0.00
2.12
1988
1989
0.01
2.16
ao
PRE = ao + b.,*LEV + b,,*VOL
1980
-1 .4 7
-0 .5 2
1988
1989
-0 .6 3
log(PRE) =-■ a!o + ivlo g (L E V ) + £>a*VOL
1980
2.31
2.48
1.70
1.13
1988
1.47
1989
0.86
Note: VOL = Variance of earnings before interest and tax payments.
Digitized for FRBNY
FRASER Quarterly Review/Spring 1990
Appendix B: The Risk Premium Curve and the Optimal Debt Ratio (continued)
stable.
Even if one does not accept all of the assumptions of
the CAPM model, it should be clear that the methods
employed yield a fairly consistent estimate of the opti
mal debt ratio. Nevertheless, modelling equity costs as
a function of leverage is a subject requiring further
empirical work.
Generating a risk premium schedule
A schedule of risk premia is generated for help in deter
mining the correct functional form to use in the risk
premia regressions and for use in evaluating the effec
tiveness of an interest deduction cap.
The risk premium curve is calculated for a high-volatility firm (annualized standard deviation of total firm
value = 0.310), a iow-volatility firm (annualized stan
dard deviation of total firm value = 0.220) and a very
low (takeover target) volatility firm (annualized standard
deviation of total firm value = 0.175).t
The calculations assume the following: 1) the stan
dard deviation of total firm market value (debt plus
equity) is constant over the life of the debt; 2) the risk
less interest rate is 7 percent; and 3) the debt ratio is
determined after solving for the required risk premium.}:
The model uses seven-year zero coupon debt to
approximate the duration of ten-year coupon debt. The
risk premia on the medium-term debt for the high- and
Iow-volatility firms form risk premia curves similar to
those estimated in the text.
Estimating the functional form of the risk premium
curve
We perform the following regressions on the generated
points: log (risk premium) = a + b * log (debt ratio).
Some results are given below:
■(•Estimates of volatility of total firm value may be found in
E. Phillip Jones, Scott P. Mason, and Eric Rosenfeld,
"Contingent Claims Valuation of Corporate Liabilities: Theory
and Empirical Tests," National Bureau of Economic Research,
Working Paper no. 1143, June 1983.
^Arbitrage pricing is done under the assumption that the
current value of the firm equals 100. The value of equity is
calculated as a call on total firm value; the value of debt is
calculated as a residual. Different debt ratios are obtained by
varying nominal redemption prices of debt. The risk premium
is calculated as the yield required on the market value of
debt in order for debt to reach redemption price, less the
riskless rate. Mathematically, the risk premium must satisfy:
(1 + if + PRE)' * LEV = K,
where /f is the riskless interest rate, PRE is the required risk
premium, t is the life of debt, and K is the face value
redemption price of debt.
Solving for PRE in continuous time, we obtain:
PRE = t--> * (log K - log LEV) - /{.
Volatility
Time to
Maturity
a
b
R2
0.175
0.220
0.310
3 years
7 years
7 years
-3 3 .0 4
-1 4 .7 9
- 8.80
7.718
3.711
2.490
0.9954
0.9991
0.9993
We then perform additional regressions on the gener
ated points: risk premium = a + b * (debt ratio)k, letting k take on different values. Some results for a firm
with low volatility (0.220) and seven-year zero-coupon
debt follow:
k
a
b
b error
R2
1
1.5
2
2.4
2.6
-5 .3 3
-3 .0 0
-1 .8 2
-1 .2 1
-0 .9 7
12.21
10.46
9.86
9.74
9.75
0.684
0.467
0.340
0.265
0.232
0.8886
0.9263
0.9546
0.9712
0.9780
The values generated by the Black-Scholes model
are considered only over the debt ratio interval of 0.10
to 0.90. The exponential relationship with the generated
points begins to break down with debt ratios greater
than 0.90.
The results obtained from regressions on curves gen
erated from the Black-Scholes model serve as a guide
for regressions on the sample data. The shape of the
risk premium curve is empirically estimated using var
ious functional forms. Table B1 presents the results
obtained by regressing the risk premium solely on the
debt ratio, and Table B2 gives the results obtained by
regressing the risk premium on the debt ratio and cash
flow variance.
Estimating the effects of an interest deduction cap
To determine the effect of an interest deduction cap on
the marginal tax incentive to leverage, we begin by esti
mating the risk premium schedule facing firms. Although
a schedule of this kind was developed in the text, the
estimation of the risk premium curves excluded the use
of firms with debt ratios in excess of 90 percent. For
this reason, and because the exponential form in the
text tends to underestimate risk premia for debt ratios
in excess of 90 percent, we will use the risk premium
curve generated by a log/log regression on a curve
generated by a Black-Scholes model, as calculated ear
lier in this appendix.
Next, we define the benchmark rate (/b) as the high
est coupon rate that the firm is allowed to deduct for
interest expense. The nominal interest rate paid by the
firm is ir The deductible interest rate (/d) of a firm under
the cap can be defined as:
id = 'r if ir < ib
i<s ~ >b
if
ir
>
ib-
FRBNY Quarterly Review/Spring 1990
wammtm
Appendix B: The Risk Premium Curve and the Optimal Debt Ratio (co n tin u e d )
' •
Without the cap, we always have /d = ir
The annual tax savings from the debt shield is simply
the deductible rate times the debt ratio of the firm, LEV:
La * LEV * tc.
c
The marginal change in the total interest deduction with
respect to a change in the debt ratio is:
(B' 1)
{ ( j Ce v ‘ l e v )
FRBNY Quarterly Review/Spring 1990
Without the interest cap (the unrestricted case) we
can substitute /r for /'d, a n d ^ ^ in (B-1) will always be
positive. With the interest cap (the restricted case), the
value of (B-1) will equal that of the unrestricted case
when ir < /b. If ir > /b, then in the restricted case the
value for (B-1) becomes (/b * tc). The values shown in
Chart 4 in the text are simply the ratio of (/b * fc) to the
value calculated in (B-1) for the unrestricted case over
the range for which /r > /'b.
Understanding International
Differences in Leverage Ttends
After a remarkable period of stability in the ratio of
aggregate debt to econom ic activity in the United
States, this ratio and various other m easures of
leverage rose in the 1980s. In earlier decades, trends
in public and private sector debt had tended to offset
each other; over the past several years, however, both
form s of debt have in c re a s e d .1 O bse rvers have
responded to these developments with apprehension.
Some worry that high leverage in the corporate sector
would restrict the ability of firms to adjust to adverse
developments and thus would heighten macroeconomic
instability.2 Others suggest that in a sharp downturn
the higher levels of debt could lead first to a wave of
bankruptcies and then to a general liquidity crisis.3
Researchers who follow international developments
have noted that other major market economies have
not experienced a similar rise in leverage.4 Yet these
economies have been riding the current business cycle
roughly in tandem with the United States. Moreover,
other factors that might influence leverage, such as
1Benjamin M. Friedman, “ Increasing Indebtedness and Financial
Stability in the United States,” in Defy, Financial Stability and Public
Policy, Federal Reserve Bank of Kansas City, 1986; E.R Davis, Rising
Sectoral Debt/Income Ratios: A Cause for Concern? BIS Economic
Papers, no. 20, June 1987.
2See, for example, Henry Kaufman, “ Debt: The Threat to Economic
and Financial Stability,'' in Debt, Financial Stability and Public Policy.
3See, for example, Ben S. Bernanke and John Y. Campbell, “Is There a
Corporate Debt Crisis?" Brookings Papers on Economic Activity, 1: 1968.
4See Claudio E.V. Borio, "Leverage and Financing of Nonfinancial
Companies: An International Perspective,” Paper presented at the
Eighteenth MSG Conference on Financial Markets and Policy,
Brasenose College, Oxford, September 19-21. 1989.
interest rates and stock prices, have also been corre
lated across countries.
This article investigates why leverage trends in the
United States have differed from those in other coun
tries. Distinguishing the developments underlying the
U.S. experience from developments abroad may help
clarify the extent to which high leverage is a problem. If
leverage has risen in the United States only because
investment has so exceeded internal funds that much
debt financing has been required, then the situation
might not be cause for concern. As long as the funds
have been invested well and no adverse shocks arise,
the investment should generate cash flows in the future
to bring leverage back down.
The data presented here show that declines in
leverage abroad have tended to be associated with
reductions of short-term debt. An analysis of firm-level
data suggests that this pattern is consistent with the
so-called pecking-order hypothesis, which links a
decline in leverage to strong internal cash flows rela
tive to investment. In Germany and France, leverage
among large firms has fallen sharply, precisely because
they have had very favorable cash flows. In Japan,
aggregate leverage has declined because cash-rich
firms have been scrupulously retiring debt. The puzzle
is why leverage has risen for large U.S. firms, which
have had reasonably strong cash flows relative to
investment.
This article stresses the finding that much of the rise
in U.S. leverage has been due to a buildup of long-term
debt, particularly by firms that have been borrowing
heavily in order to buy back their own common stock.
FRBNY Quarterly Review/Spring 1990
31
Reacting, perhaps, to perceived threats of takeover,
some firm s have been raising their leverage sharply
through stock buybacks.
The article begins with a fuller description of recent
international trends in leverage. The next section pro
vides a brief discussion of existing theories of leverage.
The third section presents estimates of leverage-target
behavior to assess the degree to which taxes, interest
rates, or stock prices can explain the recent trends. A
direct test of the pecking-order hypothesis follows; the
object of this section is to evaluate whether strength of
cash flows can account for the differences in leverage
trends. The article concludes with a brief interpretation
of the results.
die. These differences appear to be due to differences
in financial practices, not to differences in the mix of
industries within a country.8
Of g re a te r in te re s t, however, are the tre n d s in
leverage and, in p articula r, the em ergence of the
United States as the only country with consistently ris
ing leverage. Leverage in France was initially rising but
has been declining since 1984. Leverage in the United
Kingdom has stayed within a fairly narrow range. In
Germany, Italy, Japan, and the Netherlands, leverage
has clearly been falling. The trends indicate some con
vergence betw een h ig h-leve ra g e and low -leverage
countries. However, while Japan, Italy, and France have
had the highest leverage ratios, Germ any’s ratio has
declined the fastest.
Global patterns of leverage
The fam iliar ratios
Table 1 reports book-value debt-to-asset ratios fam iliar
to researchers who have tried to compare leverage in
different countries.5 The ratios reported here are based
on data from the Banque de Comptes H arm onisees
(BACH) database maintained by the European Com
mission in Brussels, but they tend to be very close to
ratios based on the usual OECD financial statistics and
the various official flow-of-funds statistics.6 The com
putation of these ratios follow s OECD convention in
including accounts payable in the definition of debt,
along with short-term and long-term debt. The argu
ment for this inclusion is the importance of accounts
p aya ble on b a la n ce s h e e ts in such c o u n trie s as
France, Italy, and Japan.7
The ratios confirm the common distinction between
low-leverage and high-leverage countries. France, Ger
many, Italy, and Japan have higher leverage, while the
U nited Kingdom and the U nited S tates have lower
leverage. The Netherlands seems to belong in the mid5The analysis here explains leverage in terms of the book values of
debt and assets instead of market values. This approach is justified
on two counts. First, survey evidence shows that most corporate
financial executives use book values for setting leverage targets: see
A. Stonehill and others, “ Financial Goals and Debt Ratio
Determinants: A Survey of Practice in Five Countries," Financial
Management, Autumn 1977, pp. 27-41. Second, other researchers
have found that it makes little difference whether market or book
values are used: see Paul Marsh, "The Choice Between Equity and
Debt: An Empirical Study,” Journal of Finance, March 1982,
pp. 121-44; and Robert A. Taggart, “A Model of Corporate Financing
Decisions,” Journal of Finance, December 1977, pp. 1467-84.
6See, for example, Janette Rutterford, "An International Perspective on
the Capital Structure Puzzle,” in J.M. Stern and D.H. Chew, eds., New
Developments in Corporate Finance (Oxford: Basil Blackwell, 1988),
pp. 194-207.
7For this sample of companies, accounts payable are on the order of
30 percent of assets for France, Italy, and Japan. In the sample of
larger companies represented in Chart 1, accounts payable on
average amount to 15 percent of assets for France and 20 percent
for Japan.
32for FRBNY
Digitized
FRASER Quarterly Review/Spring 1990
Publicly traded industrial companies
The representation of leverage trends in C hart 1 is
based on G lobal Vantage data on p u b licly traded
8For a discussion of differences in financial systems, see Robert N.
McCauley and Steven Zimmer, “ Explaining International Differences in
the Cost of Capital," Federal Reserve Bank of New York Quarterly
Review, Summer 1989, pp. 7-28. For a study showing that differences
in leverage are not due to differences in industrial mix, see Joelle
Laudy and Daniel Szpiro, "Des Entreprises Industrielles Plus
Endett6es en France que dans les Autres Pays Europ6ens,”
Economie et Statistique, no. 217/218 (1989).
Table 1
Book-Value Debt-Asset Ratios
of Nonfinancial Firms
(Percent)
France
Germany
Italy
Japan
Netherlands
United Kingdom
United States
1982
1983
1984
1985
1986
1987
69.2
62.0
70.4
73.4
55.4
47.9
70.7
60.9
68.2
73.2
55.2
47.6
39.4
76.6
59.7
67.5
72.6
53.9
48.1
40.6
73.3
58.6
67.0
71.5
54.3
47.6
42.2
71.0
57.1
67.3
70.5
52.4
46.6
43.3
70.2
66.9
70.1
52.3
44.3
Source: Banque de Comptes Harmonisees (data collected by
the European Commission from official sources and
"harmonized” to correct for differences in data collection and
to make comparisons possible).
Note: The composition of each country's sample is as follows:
France — Firms with 500 or more employees,
unconsolidated.
Germany — Over 70,000 corporations, sole proprietorships,
and partnerships; unconsolidated.
Italy — Mostly industrial firms and a few construction firms,
unconsolidated.
Japan —All nonfinancial firms, unconsolidated.
Netherlands — Nonfinancial firms, consolidated.
United Kingdom — Large firms, consolidated.
United States — All nonfinancial firms, excluding construction
and services, consolidated.
in d u s tria l co m p an ies.9 The analyses below rely on
these data for inform ation at the firm level. Use of
these data, instead of the official flow-of-funds statis
tics, makes it possible to focus on larger firms, which
can be m ore re a d ily c o m p a re d a c ro s s c o u n trie s
because of th eir sim ilar degree of access to capital
markets. The leverage ratios shown in the chart are
computed with accounts payable included in the defini
tion of debt, as in Table 1, but also with provisions for
pension lia b ilitie s su btracted from assets, because
9Global Vantage (Standard and Poor’s international version of
Compustat) draws data from financial statements of publicly traded
industrial companies.
these provisions are not reported on Dalance sheets in
the United States.10 The firms selected for analysis are
limited to those for which a complete set of observa
tions is available to construct the cash flow estimates
used later, although use of larger samples from Global
Vantage will produce the same leverage patterns.
A comparison of the figures in Table 1 with those in
Chart 1 shows substantial differences in leverage levels
between samples, but the trends in leverage remain
basically the same. The generally lower leverage ratios
in C hart 1 suggest that large firm s are much less
10Provisions for pension liabilities are most significant for German firms:
the item amounts to an average of 15 percent of firm assets.
Chart 1
Ratios of Debt to Assets for Publicly Traded Industrial Companies
Short-term debt has accounted for much of the decline in leverage in some countries,. . .
Percent
70
Japan
35 companies
Short-term debt
60
France
16 companies
liili iiliililii iili
50
Long-term debt
40
Germany
31 companies
'
30
■Bill
i i f i
: 1 1111 m m
20
10
0I
1987
1983
1988
1983
1988
1983
---------'
while long-term debt has accounted for much of the rise in leverage in Australia and the United States.
Percent
60
Australia
88 companies
50
40
United States
958 companies
United Kingdom
139 companies
_______
30
20
Ml
•
■
-----
10
0
;
1983
1988
1983
1987
1983
1988
Source: Basic data are drawn from Standard and Poor’s Global Vantage.
FRBNY Q uarterly Review/Spring 1990
33
leveraged than small firms in a given country. Indeed,
the Global Vantage data now show large German firms
to have lower leverage ratios than large U.S. firms. The
chart includes leverage ratios for Australia, and these
show higher leverage than the ratios for France or Ger
many. In spite of the differences between the two data
sets, the trends persist and in fact become more strik
ing with the Global Vantage data. The United Kingdom
and Australia both show a slight rise in leverage, but
the United States still stands apart because of the
marked rise in its leverage. Note that this rise in U.S.
leverage cannot be attributed directly to leveraged
buyouts (LBOs), since Global Vantage data include
only firms that have remained public. France and Ger
many show sharper declines in leverage than before.
Short-term and long-term debt
Chart 1 also divides the leverage ratios into short-term
debt and long-term debt components. In France and
Germany, large firms have achieved a sharp decline in
total leverage mainly by reducing short-term debt as a
ratio to assets. In Japan, the decline in total leverage
has been more modest, but the reduction in short-term
debt has been just as apparent. In the United Kingdom,
total leverage has remained essentially unchanged,
while the ratio of short-term debt to assets has risen
slightly. In Australia and the United States, a rise in
leverage has been associated with a rise in long-term
debt.
The broad pattern appears to be that in countries
with falling leverage, the decline can be attributed to
short-term debt, while in countries with rising leverage,
the increase can be traced to long-term debt.
Theories of capital structure
Can existing theories of capital structure explain differ
ences in leverage behavior across countries? Can the
theories explain the association of falling leverage with
short-term debt and rising leverage with long-term
debt? It is useful to distinguish two approaches to
analyzing how firms determine leverage. One approach
sees firms as trying to achieve a leverage target or an
optimal capital structure. The other approach sees a
firm’s capital structure as a byproduct of a history of
financing decisions, in which the firm has in every
period matched its uses of funds with the cheapest
sources it could find. The two approaches are dis
cussed in greater detail below.
The leverage-target approach
Under the first approach, the optimal capital structure
or leverage target depends on such factors as condi
tions in capital and credit markets, the tax treatment of
returns on different assets, the riskiness of the firm ’s
34
FRBNY Quarterly Review/Spring 1990
earnings, the costs of financial distress, and various
agency problems associated with debt and equity.11
The costs of financial distress include the loss of flex
ibility experienced by a firm having difficulty servicing
its debt, as well as the trustee and legal fees and reor
ganization costs incurred if the situation deteriorates
into bankruptcy. Agency problems with debt arise when
firms have an incentive to choose riskier projects
against the interest of creditors, while agency problems
with equity occur when firm s have an incentive to
spend on managerial perquisites against the interest of
shareholders.
Indirect evidence for the existence of leverage tar
gets is provided by the finding that average leverage
ratios for broad industry groups tend to be consistent
over tim e.12 Direct evidence for particular models
remains hard to find, however, perhaps because unob
servable agency costs are critical explanatory vari
ables.13 For example, efforts to explain differences in
leverage across countries in terms of tax differences
alone have largely failed. In general, there is little cor
relation between the ranking of countries according to
the relative tax advantage of debt over equity and their
ranking according to leverage.14 Hence explanations
often turn to special institutional factors —for example,
the system of universal banking in Germany or the
organization of keiretsus (groups of companies with
cross-holding of shares) in Japan —which somehow
provide more effective ways of dealing with financial
distress or agency problems of debt.
The pecking-order approach
Under the second approach, the determ ination of
leverage hinges on the strength of cash flows. The for
mal statement of this approach is the so-called peck
ing-order hypothesis developed by Myers and Majluf.15
The theory assumes that managers know more about
the firm than do outside investors. Since managers are
less likely to issue new stock if they regard existing
shares as undervalued than if they regard the shares
11For a summary of this literature, see Colin Mayer, “ New Issues in
Corporate Finance," European Economic Review, vol. 32 (1988),
pp. 1167-89.
12Ezra Solomon, The Theory of Financial Management (New York:
Columbia University Press, 1963), pp. 91-106.
13This lack of evidence led Stewart Myers to the subject of his
presidential address to the American Finance Association, “The
Capital Structure Puzzle," Journal of Finance, vol. 39 (July 1984),
pp. 575-92.
14See Mayer, “ New Issues"; Rutterford, “An International Perspective";
and Borio, “ Leverage and Financing.”
15Stewart C. Myers and Nicholas S. Majluf, "Corporate Financing and
Investment Decisions When Firms Have Information Investors Do Not
Have,” Journal of Financial Economics, vol. 13 (1984), pp. 187-221.
as overpriced, investors will regard a decision to issue
stock as a sign of possible “ bad” news. Thus, firms can
only issue equity at a discount, and cash flows will nor
m ally be a chea pe r source of funds than e xterna l
Box 1: A G raphical E xposition
The chart below illustrates pecking-order behavior for
different levels of investment demand. Abstracting from
dividend payments, a firm with the weak investment
demand /, would finance investment entirely with inter
nal funds and use the remaining cash flow to retire
debt. With the stronger investment demand l2, the firm
would invest all its cash flow and then turn to debt
financing. With the still stronger investment demand l3,
the firm would use internal funds, debt financing, and
outside equity, although the cases in which firms actu
ally turn to outside equity are considered relatively
uncommon. The interesting case occurs when invest
ment demand happens to be l4, falling in the gap
between the costs of internal and external finance. In
this case investment would be constrained by cash flow.
While the return to investment would be high enough to
justify further cash flow financing, it would fall short of
the hurdle rate for any borrowing.
Pecking-Order Behavior
The Cost of Funds for Varying Investment Demands
Cost of funds
equity. The asymm etry of information will also make
debt financing cheaper than external equity, sim ply
because debt contracts are safer in that they limit the
possible ways by which holders could lose. Hence, to
finance investment, firms will first use cash flows as the
cheapest source, then debt financing, and finally out
side equity financing (see Box 1 for an illustration). In
short, the stronger the cash flows relative to invest
ment, the less likely the firms will turn to debt and the
more likely leverage will fall.
Thus far the evidence for the pecking-order theory,
like the evidence for leverage targets, has been indi
rect only. For example, the finding that cash flow is a
s ig n ific a n t d e te rm in a n t of inve stm en t in d ica te s that
there is indeed an im portant distinction between inter
nal and external finance.16 Another form of indirect evi
d ence is the stron g in v e rs e c o rre la tio n b etw e en
profitability and leverage.17 If profitability is correlated
with cash flow, then this result indicates that firms with
strong cash flow avoid borrowing, consistent with pecking-order behavior.
The two approaches together
In practice, firms must decide on capital structure and
financing at the same time. One can imagine firms to
be constantly seeking their optimal capital structure but
often finding themselves bumped away by shocks to
cash flow. U nexpectedly good earnings would put a
firm below its leverage target, and the firm would then
try to raise its debt ratio over time, perhaps by making
investment plans that would require much debt financ
ing. An adverse shock to cash flow would put the firm
above its target; to reduce leverage, the firm might
then postpone large investments.
The empirical results reported in the next two sec
tions suggest somewhat more subtle behavior. Analysis
of Global Vantage data for four major countries in the
1980s in d ic a te s th at le ve ra g e -ta rg e t co n sid e ra tio n s
determine the ratio of long-term debt to assets, while
pecking-order considerations determ ine the ratio of
total debt to assets.18 Together, these results mean that
firms manage their long-term debt to achieve an opti16See Steven M. Fazzari, R. Glenn Hubbard, and Bruce C. Petersen,
“ Financing Constraints and Corporate Investment,’’ Brookings Papers
on Economic Activity, 1: 1988, pp. 141-206. See also Richard Cantor,
“A Panel Study of the Effects of Leverage on Investment and
Employment,” in Studies on Financial Changes and the Transmission
of Monetary Policy, Federal Reserve Bank of New York, May 1990.
17W. Carl Kester, “ Capital and Ownership Structure: A Comparison of
United States and Japanese Corporations,” Financial Management,
Spring 1986, pp. 5-16; and John Baskin, “An Empirical Investigation
of the Pecking Order Hypothesis,” Financial Management, Spring
1989, pp. 26-35.
18Firms may have leverage targets for total debt as well, but the
adjustment may happen too slowly to be detected.
FRBNY Quarterly Review/Spring 1990
35
mal capital structure while they adjust short-term debt
to accommodate cash-flow shocks. The link between
cash flows and short-term debt makes sense to the
extent that shocks to cash flow are transitory, since the
transactions costs for both issuing and retiring debt are
much lower for short-term debt than for long-term debt.
This pattern of behavior implies that in France, Ger
many, and Japan, where major firms have been reduc
ing primarily short-term debt, strong cash flows would
explain the decline in leverage. In Australia and the
United States, where corporations have been taking on
largely long-term debt, something other than weak
cash flows would be needed to explain the rise in
leverage.
Leverage-target behavior
To evaluate the extent to which interest rates and stock
prices can explain recent trends in leverage, this sec
tion presents estimates of leverage-target behavior.
Leverage is specified in terms of both total debt and
long-term debt. The leverage target is specified to
depend on the observable costs of debt and equity and
on a proxy for the costs of financial distress. Overall,
the empirical results for leverage-target behavior are
less than im pressive for explaining the trends in
leverage. To the extent that observable cost factors can
explain such behavior in the 1980s, however, they
appear to work better when leverage is measured in
terms of long-term debt than when leverage is mea
sured in terms of total debt.19
The estimating equations assume gradual adjustment
by firms to the leverage target. Formally, the adjust
ment is described by A d = k(d* - d.J, where d is the
ratio of debt to assets, d* is the desired ratio, d.-, is the
previous period’s debt-asset ratio, and A. is the adjust
ment coefficient, with X = 1 implying complete adjust
ment in a fiscal year. Here the amount of adjustment is
proportional to the difference between desired and
actual leverage. The desired ratio is then specified as a
function, d* = f(cD, cE, cFD), where cD, cE, and cFD are
the costs of debt, equity, and financial distress, respec
tively. Since Ad is d - d.1t the estimating equation can
be written as d = kf(cD, cE, cFD) + (1 - k)dm1. The
cost of debt would have a negative effect on leverage,
the cost of equity a positive effect, and the costs of
financial distress a negative effect.
The equations are estimated for Germany, Japan, the
United Kingdom, and the United States for the period
covering the fiscal years from 1983 to 1988. The esti
19This result is consistent with Donaldson’s finding that the most
common measure of debt capacity used by U.S. corporations is the
ratio of long-term debt to capitalization. See Gordon Donaldson,
Corporate Debt Capacity, Harvard Business School Division of
Research, 1961.
36 forFRBNY
Digitized
FRASERQuarterly Review/Spring 1990
mates are based on panel data from Global Vantage
and use the same sam ple of firm s used for the
leverage ratios in Chart 1. The cost of debt is mea
sured as the real interest rate on corporate bonds
minus the product of the corporate tax rate and the
nominal interest rate.20 The cost of equity is measured
as the ratio of cash flow per share to the stock price for
each firm. The results tend not to be sensitive to the
way in which the costs of debt and equity are meas
ured. For the costs of financial distress, the equations
use the ratio of fixed assets to total assets as a proxy
negatively related to these costs.21
Total debt
Table 2A reports the leverage-target equations for total
debt. The costs of debt and equity for Germany, Japan,
and the United Kingdom show signs contrary to those
predicted by theory. Given that the cost of equity is
measured as the ratio of cash flow to stock price, the
minus sign on this variable may in fact be picking up
pecking-order behavior, with strong cash flows tending
to reduce leverage. The equation for the United States
gives the theoretically correct signs, but the coefficient
is significant only for the cost of equity. In all cases, the
equations show the wrong sign for the ratio of fixed
assets to total assets, a variable which should have a
positive effect on leverage if it is an inverse proxy for
the costs of financial distress. The final disappointment
is that the coefficients on the lagged debt-asset ratio
for German, Japanese, and U.K. firms are not signifi
cantly d ifferent from one, im plying no discernible
adjustment towards a leverage target.
Long-term debt
On the whole, estimates of leverage-target behavior in
terms of long-term debt yield modestly more favorable
results than estimates in terms of total debt. For the
United States, however, it seems to make a difference
whether firms buying back their stock are included in
the sample. Table 2B reports a second set of estimates
for the United States that excludes cash-rich U.S. firms
using long-term debt to finance stock buybacks.22
“ For a fuller explanation of the variables, see Eli M. Remolona, "Why
International Trends in Leverage Have Been So Different," Federal
Reserve Bank of New York, Working Paper no. 9002, February 1990.
^ In justifying this proxy as a measure of the value of tangible assets
available for liquidation in case of bankruptcy, Long and Malitz argue
that the costs of financial distress are associated largely with the
loss of intangible assets. See Michael S. Long and Keen B. Malitz,
“ Investment Patterns and Financial Leverage," in Benjamin Friedman,
ed., Corporate Capital Structure in the United States (Chicago:
University of Chicago Press, 1985), pp. 325-48.
“ This study defines “ cash-rich” firms as those whose measured cash
flows exceed predicted investment and dividends, with the prediction
based on sales and lagged dividends.
U.S. sam ple reduces the e x p la n a to ry power of the
c o s ts of d e b t and e q u ity in th e le v e ra g e -ta rg e t
equations.
With leverage measured in terms of long-term debt,
the estimated effects of the cost of debt for Japan, the
United Kingdom, and both U.S. samples now have the
correct sign, although in no case is the effect statis
tically significant. The estimated effects of the cost of
equity have the right sign for Japan and the U.S. full
sample (with statistical significance in the case of the
latter), but the sign is reversed when the buyback firms
Stock buybacks have become an im portant U.S. phe
nomenon, and it would be of interest to know whether
they represent a separate kind of leveraging behavior.
W hen even ca sh -rich firm s go deeper into debt to
finance buybacks, such behavior may reflect an effort
to leverage up and not ju s t to su b stitu te fo r d iv i
dends.23 The exclusion of these buyback firms from the
“ Bagwell and Shoven view stock buybacks as a substitute for
dividend payments. See Laurie S. Bagwell and John B. Shoven,
"Cash Distributions to Shareholders," Journal of Economic
Perspectives, vol. 3 (Summer 1989), pp. 129-40.
Table 2A
Leverage-Target Equations fo r Total Debt
Dependent Variable Is the Ratio of Total Debt to Assets
Explanatory Variable
Constant
Cost of debt
Cost of equity
Lagged fixed
to total assets
Lagged long-term
debt to assets
F
Adjusted R2
n
Japan
Germany
0.0254
(1.6076)
1.3425
(1.6388)
-0.0021
(-0 .8 7 5 0)
-0.0323
(-0 .7 3 4 1 )
0.9524
(-1 .4 7 8 3 )
1.547
0.012
185
United Kingdom
0.0658
(2.0184)
0.4784
(0.3489)
-0.1183
(-3 .4 7 9 4 )
-0.1 4 4 8
(-2 .5 1 3 4 )
1.0517
(1.5029)
4.874
0.082
173
United States
0.1045
(3.2453)
1.4295
(1.9813)
-0.1 3 0 8
(-6 .5 7 2 9 )
-0 .5 2 2
(-1 .2 4 2 9 )
0.9712
(-0 .5 1 2 5 )
13.622
0.022
2,291
0.1993
(12.7756)
-0.2964
(-0 .4 9 8 4 )
0.0371*
(3.0917)
-0.0 3 7 7
(-1 .7 7 0 0 )
0.6458*
(-1 6 .6 2 91 )
79.579
0.052
5,777
Notes: The t-values for the null hypotheses are in parentheses. An asterisk indicates correct sign and significance at the 5 percent level.
Table 2B
Leverage-Target Equations fo r Long-Term Debt
Dependent Variable Is the Ratio of Long-Term Debt to Assets
United States
Explanatory Variable
Constant
Cost of debt
Cost of equity
Lagged fixed
to total assets
Lagged long-term
debt to assets
F
Adjusted R2
n
Germany
0.0072
(1.2632)
0.1677
(0,4712)
-0.0 0 0 4
(-0 .4 0 0 0 )
0.0060
(0.2985)
0.8806*
(-2 .9 3 3 7 )
2.382
0.029
185
Japan
0.0287
(2.0648)
-0.2 1 9 6
(-0 .3 4 0 2 )
0.0078
(0.4937)
-0.0 2 6 7
(-0 .9 7 4 4 )
0.8836*
(-3 .1 2 9 0 )
3.200
0.048
173
United Kingdom
0.0127
(4.2333)
-0.0 8 3 0
(-0 .8 7 5 5 )
- 0.0003
(-0 .1 1 5 4 )
0.0184*
(3.4074)
0.7912*
(16.5714)
69.044
0.106
2,291
Full Sample
0.0548
(10.1481)
-0.2188
(-1 .0 0 6 9 )
0.0103*
(2.3409)
0.1481*
(18.5125)
0.4436*
(-6 5 .4 5 8 8 )
1,061.232
0.423
5,777
Without
Buyback Firms
0.0297
(6.1875)
-0.1 8 9 9
(-1 .0 5 4 4 )
-0.0 1 7 8
(-4 .3 4 1 5 )
0.0564*
(7.7260)
0.7943*
(-2 1 .2 0 6 2 )
124.294
0.101
5,485
Notes: The t-values for the null hypotheses are in parentheses. An asterisk indicates correct sign and significance at the 5 percent level.
FRBNY Quarterly Review/Spring 1990
37
are excluded for the United States. This time the ratio
of fixed assets to total assets works as a negative
proxy for the costs of financial distress for the United
Kingdom and both U.S. samples.24 The coefficients on
the lagged debt-asset ratio now indicate significant
adjustment towards a leverage target, whether or not
the sample shows rising leverage over the period. The
coefficients suggest that German and Japanese firms
adjust their capital structures about 12 percent a year
towards their leverage targets, British firms adjust 21
percent a year, and American firm s adjust about 56
percent a year.
The sample period may have been too short to allow
enough variation in interest rates and stock prices to
show m arked effects on leverage trends.25 N onethe
less, the fact that the long-term debt equations work
better than the total debt equations suggests that long
term debt is probably what is determined by leveragetarget behavior.
Pecking-order behavior
The pecking-order hypothesis explains why firms might
rationally let cash flow s determ ine leverage. If m an
agers are seen as having an informational advantage,
outside investors will demand a return premium that
w ill make inte rn a l cash flow s a cheaper source of
financing than external funds. Thus firms will use up
their cash flow before they turn to debt, so that strong
cash flows relative to investment will tend to lead to a
decline in leverage.
Aggregate data on cash flows and financing
Table 3 shows how large corporations in six countries
actually matched their sources and uses of funds in the
1980s. The sources are internal cash flow and external
finance. The uses are investm ent and dividends. A
negative residual suggests that total identified financ
ing is short of total known uses.26 The negative resid
uals indicate that it is easier to underestim ate cash
flow than to overestimate it (see Box 2 for the m ea
surement of cash flow).
In most cases, relatively strong cash flows in the
aggregate seem to accompany declines in leverage, a
fin ding w hich is c o n s is te n t w ith the p e c k in g -o rd e r
hypothesis. In the case of Australian firms, even count
ing the large negative residual as unidentified a d d i
tio n a l cash flo w w ould re s u lt in cash flo w s w eak
enough to be som ewhat consistent with the m odest
rise in their total leverage during the period. By con
trast, cash flow s were so strong among French and
German firms that it is easy to see why they had a
sharp decline in leverage.27 Significantly, much of this
cash flow w ent into reducing s h o rt-te rm debt. One
24ln Corporate Debt Capacity, Donaldson finds that debt contracts
often limit new long-term borrowing to a percentage of tangible
assets.
26Sources and uses do not always balance because most of the
figures are constructed from income statements and balance sheets,
not from flow of funds statements.
“ To reduce multicollinearity, the costs of debt and equity were
combined in a single ratio, but the results were not substantially
better. Logarithmic transformations also failed to improve results.
27ln France, the large societies apparently enjoyed strong cash flows in
part because of government policies to restrain wages in the early
1980s.
Table 3
Sources and Uses of Funds in Six Countries
(Averages over the Period, Percent of Total Sources)
Sources
Internal cash flow
External finance
Short-term debt
Long-term debt
Equity
Uses
Investment
Dividends
Residual
Australia
1984-87
France
1983-87
Germany
1983-87
Japan
1984-88
United Kingdom
1983-88
United States
1983-88
34.6
65.4
16.5
39.0
9.8
88.7
11.3
6.6
1.4
3.3
104.6
- 4 .6
-9 .4
2.3
2.6
63.2
36.8
22.2
10.7
3.9
82.5
17.5
- 9 .3
9.4
17.4
77.7
22.3
5.3
15.4
1.9
121.4
13.1
-3 4 .4
105.2
9.1
-1 4 .2
92.6
9.5
-2 .1
99.3
6.2
- 5 .5
83.3
15.3
1.4
73.8
23.5
2.7
Source of basic data: Standard and Poor's Global Vantage.
Notes: The sample of companies corresponds to that in Chart 1. Short-term debt includes accounts payable. Investment includes fixed
capital, inventory stocks, acquisitions, and financial assets. See Box 2 for the components of cash flow.
38 forFRBNY
Digitized
FRASERQuarterly Review/Spring 1990
seeming inconsistency is that Japanese firms appar
ently suffered weaker cash flows than U.S. firms w ith
out seeing the rise in leverage that characterized U.S.
firms during the period.28 But Japanese firms also paid
much sm aller dividends than American firms.
Although these aggregate ex post financing patterns
are suggestive, they do not provide a convincing test of
pecking-order behavior. The patterns do not reveal
whether the firms reducing debt were also the ones
with excess cash flow or whether the firms borrowing
heavily were also the ones w ith strong investm ent
demands. Moreover, if firm s with strong cash flows
were investing the excess cash instead of retiring debt,
then investm ent would appear strong ex post and it
28lf the figures are accurate, financing patterns would necessarily be
consistent with leverage trends. When the ratio of new debt to net
investment is lower than the initial leverage ratio, leverage must
decline. In the case of Japan, for example,, investment net of
depreciation and amortization was 61 percent of funds from all
sources. Given that new debt was 33 percent of all funding,
leveraging at the margin was 54 percent (33 divided by 61). With an
initial leverage ratio of 60 percent at end-1983, the 54 percent debt
financing over the period reduced the leverage ratio to 56 percent at
end-1988.
would be difficult to verify pecking-order behavior.
Testing for p ecking-order behavior
Estimates of cash flow and investment demand at the
firm level allow a more direct test of the pecking-order
hypothesis. The test compares borrowing behavior of
firms with different strengths of cash flow relative to
investment demand to determ ine whether differences
in such behavior seem to reflect the existence of a gap
between the costs of internal and external finance.
The strength of investm ent demand relative to cash
flow is measured in terms of the p re d icte d external
financing need (PEF), which in turn is specified as the
d iffe re n ce betw een p re d ic te d inve stm en t and d iv i
dends (PID) and actual measured cash flow (CF), that
is, PEF = PID - CF.29 For PID, an ex ante concept of
^Clearly, it will not do to use actual external financing or actual
investment and dividends. Given the amount of cash flow, the
accounting balance between sources and uses of funds will ensure
that changes in debt and equity always match actual investment and
dividends. Such data will preclude detection of any gap between the
costs of internal and external financing.
Box 2: Measurement o f Cash Flows
Financial statement data from Global Vantage allow the
estimation of cash flows by a procedure suggested by
Cottle, Murray, and Block.f This procedure adds back
to reported after-tax earnings those reported expenses
that drain no cash —expenses such as depreciation of
fixed assets, amortization of intangibles, increases in
deferred taxes, and additions to provisions and
reserves. The reason for this adjustment is that
tSidney Cottle, Roger F. Murray, and Frank E. Block, Security
Analysis, 5th ed. (New York: McGraw-Hill, 1988), pp. 237-62.
reported earnings alone are not always an adequate
measure of cash flow. As the table below shows, for
most of the countries, and especially for France and
Germany, charges to depreciation of fixed assets and
amortization of intangibles are a more important source
of cash flow than earnings. In the case of Germany,
additions to provisions and reserves represent nearly
twice as much cash flow as earnings. Moreover,
increases in deferred taxes, although not nearly as
important as earnings, can also be a significant compo
nent of cash flow.
Com ponents of Measured Cash Flow
(Percent of Measured Cash Flow)
Earnings after taxesf
Depreciation and amortization
Provisions and reserves
Deferred taxes
Accounts receivable
Australia
1984-87
France
1983-87
Germany
1983-87
Japan
1984-88
United Kingdom
1983-88
United States
1983-88
79.2
51.2
2.3
9.8
-4 2 .5
37.8
59.5
9.7
7.9
-1 4 .9
16.6
59.8
31.0
0.8
- 8 .3
48.7
62.3
4.0
0.8
-1 5 .8
80.8
32.5
0.6
1.0
-1 5 .0
47.6
52.5
—
7.2
- 7 .3
fEam ings include extraordinary items.
FRBNY Q uarterly Review/Spring 1990
39
investment is obtained by taking fitted values from a
regression of the sum of investment and dividends on
the change in sales and the first lag in dividends (all
variables normalized by firm asset size), since acceler
ator models of investment tend to work well and divi
dends tend to adjust slowly.
Subtracting actual cash flow from PID yields the PEF
for each firm in a given year. The observations are then
divided into three groups:
(1) cash-rich firms with negative PEFs (correspond
ing to firms with investment demands such as /,
in Box 1);
(2) a middle group with positive but relatively small
PEFs (corresponding to cash-constrained firm s
with investment demands such as l4, as well as
firms with demands such as l2 in Box 1); and
(3) cash-poor firms with relatively large PEFs (corre
s p o n d in g m a in ly to fir m s w ith in v e s tm e n t
demands such as l2 in Box 1).
The precise separation of firms into the middle and
cash-poor groups is determ ined through an iterative
maxim um -likelihood procedure, which finds the division
that produces the best combined fit for regression esti
mates.30 The idea of the test is to try to detect in the
“ Steve Peristiani wrote an algorithm for seeking a maximum for the
concentrated log-likelihood of a bivariate switching regression model.
The log-likelihood is from Stephen M. Goldfeld and Richard E.
middle group borrowing behavior that reflects a cash
constraint.
For each group, we estimate the equation AD/A = a
+ b (PEF/A), where AD is net borrowing in term s of
the change in total debt, A is total firm assets, a is the
constant term, and b is the pecking-order coefficient.
The variables are divided by assets to avoid problems
of h eteroskedasticity related to firm size. Under the
pecking-order hypothesis, the pecking-order coefficient
for the cash-rich group would be close to one. This
result would reflect the use of excess cash flow to
retire debt, with the amount retired varying one-to-one
with investment demand across firms. Within the middle
group, the coefficient would be less than one, reflecting
the presence in the group of at least some cash-con
strained firms for which debt would be unaffected by
the strength of investment demand. Finally, within the
cash-poor group, the coefficient would again be close
to one, reflecting amounts of borrowing that varied oneto-one with investment demand. (The presumption is
that re la tiv e ly few firm s a c tu a lly re so rt to e xte rn a l
equity financing.)
Footnote 30 continued
Quandt, “The Estimation of Structural Shifts by Switching
Regressions,” Annals of Economic and Social Measurement, vol. 2
(1973), pp. 475-85.
Table 4
Pecking-Order Equations
Dependent Variable Is ^
United States
Explanatory Variable
Cash-rich group
Constant
PEF/A
Middle group
Constant
PEF/A
Cash-poor group
Constant
PEF/A
Germany
n = 74
-0 .0 1 6 6
0.2252
(-9 .3 4 6 2 )
n = 54
0.0028
0.1384**
(0.4878)
n = 27
-0.0 4 2 4
0.8515**
(-0 .4 9 1 4 )
Japan
n —11
0.0285
1.0657**
(0.0674)
n = 66
-0.0062
0.6132**
(1.4819)
n = 68
-0.0267
0.8204**
(-0 .5 3 9 0 )
United Kingdom
n = 229
-0.0 1 0 9
-0.0382
(-1 5 .2 0 06 )
n = 1,190
0.0030
0.4149**
(-8 .9 8 7 7 )
n = 491
-0.1 8 9 7
1.0461 **
(1.3599)
Full Sample
n = 918
0.0577
0.7855
(-3 .8 7 1 8 )
Without
Buyback Firms
n = 626
-0.0 2 0 9
0.9238*
(-1 .6 7 1 0 )
n = 3,375
0.0082
0.4472**
(-1 1 .7 6 17 )
n = 522
-0.2 3 0 6
1,2776
(4.7130)
Notes: The t-values under the null hypotheses are in parentheses. One asterisk indicates failure to reject the pecking-order hypotheses at
the 5 percent level, and two asterisks, failure to reject at the 10 percent level. The null hypotheses are b = 1 for the cash-rich and cashpoor groups and b = 0 for the middle group, where b is the coefficient on PEF/A. In the case of the middle groups for the United Kingdom
and the United States, however, the asterisks indicate rejection of the null hypothesis that b = 1.
40 forFRBNY
Digitized
FRASERQuarterly Review/Spring 1990
Test results
Table 4 reports the regression estimates for Germany,
Japan, the United Kingdom, and the United States for
the same large firms as before. The change in total
debt is used fo r the d ep en de nt va ria b le .31 For the
cash-rich groups in the United States, two sets of esti
mates are reported, one for the full sample and one for
the sample excluding firms that were buying back stock
and engaging in net long-term borrowing at the same
time.
The results are broadly in accord with pecking-order
behavior. Nine of the thirteen estimated pecking-order
co efficients are consistent with the hypothesis. The
most telling result in favor of the pecking-order hypoth
esis is the difference between the. estim ated c o e ffi
c ie n ts fo r the c a s h -p o o r and the m iddle groups,
particularly in the case of German and British firms.
The much lower propensity for debt financing by firms
in the middle groups suggests the presence among
them of c a s h -c o n s tra in e d firm s w ith in v e s tm e n t
demands caught in a gap between the costs of internal
and external finance.32 In the case of the full sample of
U.S. firms, only the coefficient for the middle group is
consistent, but when the buyback firms are excluded,
the coefficient for the cash-rich group becomes consis
tent with pecking-order behavior.
Aberrant behavior
The deviations from pecking-order behavior are note
worthy. It is the cash-rich firms that tend to depart from
pecking-order behavior. In theory, these firms should
have been using all their excess cash flow to retire
debt in order to create slack for possible future borrow
ing needs. It appears, however, that only the cash-rich
firms in Japan were behaving according to the peckingorder hypothesis, conscientiously using excess cash
flow to retire debt 33 In Germany and the United King
dom, the firms were putting their excess cash flow into
financial assets, including stock to acquire other firms;
a few British firms were even borrowing for this pur
pose. The German firm s may have been averse to
retiring debt because of their close relationship to their
banks. In spite of this reluctance to retire debt, many
German firms were so cash-rich that aggregate debt
retirem en t was s u ffic ie n t to cause leverage to fa ll
sharply. The British firms appear to have been simply
taking advantage of unusually favorable credit condi
tions in the United Kingdom at a time of government
budget surpluses. Indeed some large British firms were
31Similar equations were estimated for long-term debt by combining
net short-term borrowing with cash flow, but the results were not
nearly as favorable.
32The large relative sizes of the middle groups, especially for the
United Kingdom and the United States, indicate that these groups
probably include many firms that were not cash-constrained. The
algorithm for dividing the sample between the middle and cash-poor
groups tended not to be very effective because the shape of the
concentrated log-likelihood function was quite flat for wide ranges.
Chart 2
Effect of Stock Repurchases on the Ratio of
Long-Term Debt to Assets in the United States
Percent
Table 5
Stock Buybacks and Long-Term Borrowing by
Cash-Rich U.S. Firms
(Millions of Dollars)
Year
Number of
Firms
Amount of
Buyback
(Net)
Long-Term
Borrowing
(Net)
1983
1984
1985
1986
1987
1988
20
37
64
51
62
58
628.2
2,462.8
11,491.2
3,275.2
4,791.0
5,638.3
557.4
2,019.6
14,474.8
6,907.4
5,542.4
9,028.8
Total
292
28,286.7
38,530.4
Source: Global Vantage
Notes: Cash-rich firms are selected on the basis of the
difference between predicted investment and dividends, on
the one hand, and on the other, actual measured cash flow.
1982
1983
1984
1985
1986
1987
1988
Source: Standard and Poor’s Global Vantage.
Note: Interest expense due to cumulative stock repurchases
since 1982 is calculated using LIBOR.
FRBNY Quarterly Review/Spring 1990
41
also providing trade credit liberally to other firms while
paying off their own trade debt. In effect, these British
and German industrial firms were engaging in some
financial intermediation of their own.
U.S. stock buybacks
Unlike firms in other countries, firms in the United
States may buy back their stock with little restriction .34
indeed, while many cash-rich U.S. firms were exhibiting
good pecking-order behavior by retiring debt, nearly a
third of their number at a given time were returning
cash to their stockholders not only by paying generous
dividends but also by borrowing heavily to buy back
their own common stock. For the sample used here,
292 of 918 observations on cash-rich U.S. firms were
on firms engaged in stock buybacks financed with long
term debt. Table 5 shows the amounts of buybacks and
long-term borrowing by these firms. In the 1980s, stock
repurchases were the d iffe re n c e betw een rising
leverage and falling leverage in terms of long-term debt
in the United States. Chart 2 shows that in the absence
of the repurchases, the ratio of long-term debt to
assets for the sample would have been about 18 per
cent in 1988 instead of 27 percent.
Significantly, nearly all the new debt used to finance
the U.S. buybacks was long term, and in some cases
short-term debt was reduced. If the firms were seeking
m erely to put cash in the hands of stockholders
through some method other than dividends, then they
could have accomplished their goal using more modest
amounts drawn from the available cash flow or financed
with short-term debt that could be promptly repaid. The
fact that cash-rich firms resorted to long-term debt sug
gests that a lasting change in leverage was an impor
tant motive. Some firms may have raised their leverage
ratios to defend themselves against perceived takeover
threats, others to support the market value of their
shares .35 But whatever the motive, it appears that the
capital restructuring in the form it took would be diffi
cult to reverse in the short run .36
^ In the United Kingdom, the power of a company to buy back its own
stock must be granted by a shareholder vote, but even the firms
already granted the power have rarely exercised it.
“ Financial innovations such as junk bonds, strip financing, and blindpool buyout funds may have made takeover threats more credible in
the 1980s than before. Bagwell finds that share repurchases make a
potential target more expensive to acquire. See Laurie S. Bagwell,
“ Share Repurchase and Takeover Deterrence,” Northwestern
University Department of Finance, Working Paper no. 53, 1989.
M ln Australia, where leverage in terms of long-term debt has also
risen, developments similar to those in the United States may be
taking place.
42for FRASER
FRBNY Quarterly Review/Spring 1990
Digitized
Conclusion
The international evidence examined here suggests
that two distinct types of company behavior have
accounted for the differences in leverage trends across
countries. For the most part, declines in leverage
abroad and among many U.S. corporations have
resulted from a type of behavior in which firms enjoying
strong cash flows have retired debt or reduced their
use of debt for financing investment. In contrast, much
of the rise in leverage in ths United States has been a
consequence of a strikingly different type of behavior,
in which some cash-rich U.S. firms have actually bor
rowed heavily, not to invest, but to repurchase their
stocks.
Although both types of behavior can be found among
U.S. firms, the U.S. trend in the aggregate differs from
trends elsewhere because of the large number of firms
buying back their stock. The precise reasons for the
buybacks are not well understood, but the use of long
term debt to finance the buybacks suggests a lasting
change in attitudes towards leverage.
Where aggregate trends show that declining leverage
accompanied strong cash flows, disaggregate data
reveal the firm-level behavior underlying the trends:
companies were acting as if external funds were sub
stantially more costly than internal funds, a behavior
consistent with the so-called pecking-order hypothesis.
Moreover, the finding that the declines in leverage were
achieved largely through reductions in short-term debt
indicates that firms were treating favorable cash flows
as transitory shocks and that the declines in leverage
were a cyclical phenomenon.
Some countries, of course, have had stronger cash
flows than others. The large German companies and
French societies appear to have enjoyed the strongest
cash flows and thus have seen the sharpest declines in
leverage. British firms have also had impressive cash
flows, but their investment in financial assets and
acquisitions has kept their leverage from falling. Japa
nese corporations as a group have had relatively mod
est cash flows, but those firms with excess funds have
been scrupulously retiring their debt, so that leverage
on the whole has declined. American corporations have
actually had somewhat strong cash flows, but the debtfinanced stock buybacks have caused aggregate
leverage to rise.
Eli M. Remolona
Monetary Policy and Open
Market Operations during 1989
Overview
Monetary policy in 1989 sought to sustain the ongoing
expansion of the economy at a moderate pace while at
the same time fostering price stability. Concerned by
signs of escalating inflation, the Federal Open Market
Com m ittee (FOMC) pursued a gradual firm ing of
reserve pressures in the early months of 1989, as it
had during most of 1988. By late spring, however, indi
cations of a slowdown in the economic expansion led
the Committee to move gradually to a more accom
modative posture.
Following its December 1988 meeting, the Committee
directed the Desk to institute a two-stage firming of
reserve pressures in light of evidence indicating that
the economy was expanding at a vigorous pace and
that inflation might intensify. The initial move was
implemented on December 15, and the second step
was taken in early January. As incoming data signaled
mounting inflationary pressures, another tightening
move was made in February. Moreover, the Board of
Governors approved a 1/2 percentage point increase in
the discount rate, to 7 percent, on February 24.
By May, however, the FOMC saw the risks to the
economy of higher inflation and a substantial shortfall
Adapted from a report submitted to the Federal Open Market
Committee by Peter D. Sternlight, Executive Vice President of the
Bank and Manager for Domestic Operations of the System Open
Market Account. Cheryl Edwards, Senior Economist, Open Market
Analysis Division, and R. Spence Hilton, Senior Economist, Open
Market Analysis Division, were primarily responsible for the
preparation of this report under the guidance of Ann-Marie
Meulendyke, Manager, Open Market Operations Department. Other
members of the Open Market Analysis Division assisting in the
preparation were Robert Van Wicklen, John Krafcheck, Theodore
Tulpan, and Geraldine Velazquez.
in economic growth as being more evenly weighted.
Then, in early June, with new evidence pointing to a
slowdown in economic activity and with some indica
tors suggesting that a gradual reduction of inflation was
likely, the FOMC began moving toward a more accom
modative reserve posture. In July, additional data rein
fo rc e d p e rc e p tio n s th a t e co n o m ic a c tiv ity was
moderating, and reserve pressures were reduced twice
in that month. Amid further signs of weakening in the
expansion during the final months of the year, reserve
pressures were eased again in October, November, and
December.
Although the longest recorded economic expansion
in U.S. peacetime history continued in 1989, the pace
of that expansion slowed considerably. Real GNP
advanced 2.6 percent (fourth quarter over fourth quar
ter), or 2.0 percent after adjusting for the impact of the
1988 drought. Consumer spending, investment in pro
ducers’ durable equipment, and net exports accounted
for most of the expansion in real GNP, although growth
of all three components was more subdued than in the
previous year. The reduced pace of economic activity
was reflected in smaller job gains in 1989. Nonethe
less, the civilian unemployment rate in the fourth quar
te r was unchan ged from its y e a r-e a rlie r level.
Meanwhile, most broad inflation measures advanced at
roughly the same pace as in 1988, although pressures
abated somewhat in the second half of the year.
Yields on investment-grade fixed-income securities
fell on balance in 1989. They rose over the first three
months of the year amid indications of econom ic
strength and rising inflation. Yields fell considerably
from late March to early August as the market sensed
FRBNY Quarterly Review/Spring 1990
43
a softening economy and the Federal Reserve shift to
accommodation. Over the balance of the year, yields
backed and filled but showed no trend. Yields backed
up in August and September in response to stronger
than anticipated economic activity and uncertainties
about how much further the Fed would ease. Later,
yields fell a bit as new data suggested weaker eco
nomic expansion and market participants came to
expect that the Federal Reserve would continue easing
its policy.
In contrast, yields on below-investment-grade bonds,
known as “ high-yield” or “junk” bonds, rose sharply.
This sector was buffeted by large defaults and bank
ruptcy threats, especially during the latter part of the
year. These events focused attention on the risks asso
ciated with highly leveraged com panies, causing
spreads to widen between the companies’ debt and
investment-quality instruments. Trading and issuance
thinned, and investors became increasingly sensitive to
the characteristics of specific issues.
Credit worries also remained a problem in the thrift
industry, where large losses and insolvencies at a
number of institutions continued tc place strains on the
financial system. The need to finance the restructuring
and rescue operations was addressed by legislation
passed in August. The Financial Institutions Reform,
Recovery and Enforcement Act of 1989 provided for
$18.8 billion of “ on-budget” federal financing in fiscal
1989. In addition, the act established a new agency, the
Resolution Funding Corporation, with the authority to
borrow $30 billion before October 1991. The agency
auctioned its first offering of bonds late in the year.
Money and debt growth decelerated in 1989. M2
advanced at a 4.6 percent rate (fourth quarter over
fourth quarter) and finished the year well within its tar
get range, while M3 expanded at a 3.2 percent rate
and ended just below the lower bound of its growth
cone .1 For the year as a whole, M1 grew a meager 0.6
percent. Total nonfinancial debt expanded at an 8.0
percent rate, which placed debt below the midpoint of
its monitoring range. M2 and M3 grew slowly over the
first half of the year, while M1 fell .2 In contrast, M1 and
■•All money and debt growth rates cited in this report are based on
the data available on March 15, 1990. The money data incorporate
the February 1990 benchmark and seasonal revisions, subsequent
revisions, and the redefinition of M2. Under the new definition, M2
incorporates thrift overnight repurchase agreements. Over the four
quarters of 1989, these revisions increased the growth rates of M1
and M2 by 0.1 percentage point and lowered the growth rate of M3
by 0.1 percentage point.
2February and March 1990 revisions elevated money growth in the first
half of the year (H1) and lowered growth in the second (H2). The
growth of M2 was raised by 0.5 percentage point in H1 and lowered
by 0.3 percentage point in H2. M3 growth was increased by 0.5
percentage point in H1 and decreased by 0.5 percentage point
44
FRBNY Quarterly Review/Spring 1990
M2 growth accelerated sharply over the second half of
the year as the opportunity cost of holding money fell.
M3 growth initially picked up a bit, along with growth in
the narrower measures, but then weakened when man
aged liabilities at th rifts contracted as part of the
restructuring of the thrift industry.
The Trading D esk’s reserve m anagem ent p ro
cedures, which depend upon a reasonably predictable
re la tio n sh ip betw een borrow ing and the spread
between the federal funds rate and the discount rate,
were again complicated by shifts — mostly downward —
in the willingness of depository institutions to borrow
from the discount window under the adjustment credit
program. As a result, the relationship between the
amount of borrowing undertaken for both adjustment
and seasonal purposes and the degree of money mar
ket firmness was somewhat uncertain. The Desk there
fore pursued the borrowing objective flexibly in order to
achieve the degree of restraint desired by the FOMC.
With adjustment credit running light in 1989, the behav
ior of seasonal borrowing dominated the movements in
the series “adjustment plus seasonal borrowing.” Sea
sonal borrowing tends to be high in the summer and
low in the winter; a number of technical adjustments
were made to the borrowing allowance during the year
in order to accommodate this tendency and leave
reserve pressures unaffected.
Record purchases of foreign currency by U.S. mone
tary authorities altered the nature and timing of the
Desk’s open market operations in 1989. As a conse
quence, the growth of the System’s holdings of foreign
currency provided more than enough reserves to cover
the drain on reserves from the rise in currency —a rise
that was in itself below average. Furthermore, in the
face of weakness in reservable deposits that held down
required reserves, nonborrowed reserves were permit
ted to grow only modestly. The Desk reduced the size
of the System portfolio (on a year-over-year basis) for
the first time since 1957. This reduction was accom
plished through redemptions of maturing Treasury
securities and through sales of Treasury issues in the
market and to foreign customer accounts.
The economy and domestic financial markets
The economy
The economy expanded less vigorously in its seventh
consecutive year of growth. Real GNP grew 2.6 per
cent in 1989, down from 3.4 percent in the preceding
year.3 The U.S. Department of Commerce estimates
Footnote 2 continued
in H2. M1 fell 0.7 percentage point less than originally reported in
H1, and its growth was 0.6 percentage point lower in H2.
3AII references to annual growth rates in this section are on a fourth
that real GNP growth, excluding the effects of the 1988
drought, was 2.0 percent in 1989, about half of the
1988 drought-adjusted rate of expansion. Slower
growth in consumer spending and exports as well as a
sharp drop in residential construction more than
accounted for the deceleration in economic activity.
The pace of nonfarm business inventory accumulation
fell for a second consecutive year in 1989, but not as
much as in 1988. Real final sales increased 2.5 per
cent, compared with 4.4 percent in 1988.4 Employment
gains in 1989 were also below the previous year’s
pace; nonfarm payroll employment was up 2.4 percent,
compared with 3.2 percent in 1988. The civilian unem
ployment rate was mostly steady during the year and
stood at 5.3 percent in the final quarter of 1989,
unchanged from its year-earlier level.
Over the year as a whole, growth was primarily sus
tained by consumer and investment expenditures. Con
sum er spendin g grew 2.5 perce'nt over the four
quarters of 1989, an increase considerably below the
nearly 4 percent advance of 1988. Most of this slippage
reflected some retrenchment in purchases of motor
vehicles .5 Supporting the growth in consumer spending
over the year was a 3.6 percent pickup in real dispos
able income, which was only moderately below its 1988
rate of increase. Heavy purchases of computer-related
equipment encouraged the healthy growth of business
investment in producers’ durable equipment. In con
trast, housing construction slid under the weight of
weak real estate markets, and nonresidential construc
tion remained sluggish in the face of high vacancy
rates.
Economic activity showed signs of losing strength as
the year progressed. Real fixed investment in the sec
ond half of the year was nearly unchanged from its
average level in the first half. Following strong gains in
the first quarter, real net exports only improved a bit,
on balance, over the remainder of the year as slower
export growth was accompanied by an upswing in
imports. In the final quarter, total GNP growth fell to a
1.1 percent annual rate, although the slowdown was
partly a result of the California earthquake and the
strike at the Boeing Company. Meanwhile, employment
Footnote 3 continued
quarter over fourth quarter basis unless specified otherwise.
Quarterly rates are seasonally annualized changes from the
preceding quarter.
4These increases are not drought-adjusted. The slowdown in final
sales growth would be even more pronounced if the impact of the
drought were excluded.
5ln addition, growth in 1988 had been boosted by a low level of
consumer outlays at the end of 1987. Late 1987 consumption was
dampened by the expiration of auto sales incentives and by some
consumer caution in the aftermath of the October 1987 stock market
break.
growth declined in each quarter from a peak rate of
over 3 percent in the first quarter to under 2 percent in
the fourth quarter of 1989.
The slowing pace of economic activity was most evi
dent in the m anufacturin g sector. M anufacturin g
employment edged a bit lower in 1989, after having
risen almost 2 percent in 1988. Sizable manufacturing
job losses occurred in each of the last four months of
1989. These losses stemmed in part from the slacken
ing pace of industrial production over the second half
of the year. Meantime, the capacity utilization rate also
declined modestly over the final two quarters. It began
the year at its 1989 high of 84.3 percent —the peak
level for the current expansion— and closed the year at
83.0 percent.
By most broad measures, prices in 1989 continued to
rise at roughly the pace set in 1988. Led by surging
food and energy costs, price pressures appeared to be
mounting in the first half of the year, but inflation sub
sided later when energy costs declined. The consumer
price index rose 4.6 percent in 1989 (December over
December), or 4.4 percent when the index’s volatile
food and energy components are excluded. These
rates of increase are roughly the same as those
recorded in 1988. The fixed-w eighted price index
advanced 4.1 percent, down from 4.5 percent in 1988.
Price pressures were somewhat stronger at the early
stages of production: the producer price index (PPI),
la rgely re fle c tin g higher food and energy costs,
increased 4.8 percent, up sharply from 4.0 percent in
1988. (Excluding these costs, the PPI advanced at
about its 1988 pace.) Wage pressures showed no signs
of abating. The employment cost index in December
1989 was 4.8 percent above its year-earlier level, a rate
of increase virtually identical to that in 1988, indicating
little change in underlying wage pressures. Indeed, unit
labor costs rose 5 percent in 1989, compared with 3
percent in the previous year, reflecting higher compen
sation costs and a decline in productivity growth.
Solid gains were made in reducing the merchandise
trade deficit early in 1989, but progress stalled around
midyear. Measured in current dollars, the average
annual trade deficit for the year narrowed by $16 billion
to $111 billion; the real trade deficit diminished by a
sim ilar amount and averaged $108 billion. By both
measures, the reduction in the trade deficit was about
half the improvement achieved in 1988. A strong export
performance was again registered in the first half of
1989, extending the pattern set in the preceding two
years, but export growth decelerated markedly in the
final two quarters of the year. Meanwhile, im port
growth continued at its 1988 rate. The slowing pace of
improvement in the trade balance largely reflected the
waning impact of the dollar’s steep 1985-87 decline. In
FRBNY Quarterly Review/Spring 1990
45
1989, the tra d e -w e ig h te d va lu e of the d o lla r rose
sharply in the first half of the year but then skidded to
finish the year close to its year-end 1988 level.6
Fiscal restraint at the federal level left total govern
ment purchases of goods and services, measured in
real terms, virtually unchanged in 1989. Purchases by
the federal government fell for a second consecutive
year, while growth in state and local government pur
chases eased s lig h tly . At the fe d e ra l level, both
defense and n ondefense spending d eclined (eith er
including or excluding purchases by the Com m odity
Credit Corporation). The federal budget deficit in fiscal
year 1989 was $152 billion on a unified basis, close to
its level in each of the preceding two fiscal years. Con
tinued econom ic expansion lifted revenues during the
fiscal year, but sizable increases in net interest pay
m ents and sp e n d in g to liq u id a te in s o lv e n t th rifts
boosted growth in total outlays, despite restraint exer
cised in other spending categories.7
Domestic financial markets
Yields on investm ent-grade fixed-income securities fell
in 1989 (Chart 1). In sharp contrast, yields on many
below -investm ent-grade corporate securities finished
the year markedly higher because major defaults and
bankruptcies in the latter half of the year upset investor
confidence in this sector. In areas not plagued by credit
quality worries, shorter dated issues led the move to
higher yields over the first three months of the year.
A fter peaking in late March, yields fell considerably
through early August. Over the rest of the year, yields
moved in a narrow range and finished modestly above
their midsummer lows.
The principal influences on financial markets in 1989
were the prospects for real economic growth and infla
tion and the outlook for Federal Reserve policy. A
number of econom ic releases, believed to offer insight
into the underlying strength of economic activity and
price pressures, were routinely monitored. They helped
to shape in v e s to rs ’ e x p e c ta tio n s abo ut e co n o m ic
growth, inflation, and the direction of System policy.
M arket p a rtic ip a n ts paid p a rtic u la r a tte ntio n to the
monthly nonfarm payroll employment data, a tim ely and
relatively com prehensive measure of econom ic per6The dollar fell 4.6 percent against the West German mark over the
year, while it rose 15.3 percent against the Japanese yen.
7ln fiscal year 1989, net budget outlays aimed at resolving the thrift
crisis more than doubled, rising from $8 billion in 1988 to $18 billion.
In 1989, roughly half of the net outlays were made by the nowdefunct Federal Savings and Loan Insurance Corporation (FSLIC),
while the remainder were made by the Resolution Trust Corporation —
created by legislation passed in August. Previously, almost all such
outlays had been undertaken by the FSLIC. Expenditures for this
purpose are widely seen as having a minimal impact on economic
activity.
46
FRBNY Quarterly Review/Spring 1990
formance. The monthly national purchasing managers’
report was also closely scrutinized for early signs of
developm ents in the m an ufa ctu rin g sector. Several
price series were watched to keep abreast of the latest
inflation trends; forem ost among these was the PPL
The behavior of the dollar on foreign exchange markets
a lso in flu e n c e d y ie ld s at tim e s . T h is e ffe c t w as
achieved partly through the dollar’s impact on expected
future inflation rates: a strong dollar placed downward
pressure on im port prices and thereby lessened fears
of higher inflation. In addition, a strengthening dollar
was seen as enco urag ing inve stm en t inflow s from
abroad, inflows which would tend to boost the value of
dollar-denom inated instrum ents. T hroughout the year,
y ie ld s o fte n m oved w h e n e ve r m a rke t p a rtic ip a n ts
thought that an imminent change in System policy was
likely. At these tim es, p a rtic ip a n ts close ly follow ed
m ovem ents in the fe de ra l funds rate to gauge the
stance of policy.
Yields on investm ent-grade securities rose over the
first three months of the year, in part reflecting System
moves to increase reserve pressures. Short-term yields
moved up early in January following the System ’s move
to firm reserve pressures, but long-term yields declined
modestly as inflation fears eased. Along with the firm
ing action, a strong dollar in foreign exchange markets
and Chairman Greenspan’s mid-January congressional
testimony reiterating the System ’s commitment to con
trolling inflation dampened inflation expectations. The
m arket’s inflation psychology shifted sharply in early
February, however, and remained pessimistic through
March because economic statistics pointed to a pattern
of robust economic growth coupled with accelerating
inflation. Payroll employment data for January and Feb
ruary showed strong job gains (Chart 2), while the pro
ducer price indexes for these months recorded sharp
advances. Rising oil prices also aggravated the nega
tive inflation prospects. Yields on shorter term issues
Chart 2
Changes in Nonfarm Payroll Employment
Thousands of jobs
400
Total
350
300
250
200
150
100
50
1_lil.li LJ i L.i —11L I 1—1—111—111—Ij
0 i .iL..J
100
Ma nufacturing
50
„ n
0
-50
-100
J
f1
1 U |
1 1 1 1 1 1 1 1—i .j—.i. i
F
M
A
M
J
J
1989
A
S
O
N
D
rose m ore th a n th o s e on lo n g e r te rm is s u e s in
response to p ro spe ctive and actual policy actions
aimed at combating these price trends, including the
discount rate hike in February.8
Evidence that the economy was losing some momen
tum while inflation was stabilizing led to a period of
declining interest rates that lasted from early April until
midsummer. R eports that the purchasing m anagers’
index had tum bled and nonfarm payrolls had shown
only a small gain for March supported some earlier
signs of a slowdown, such as a decline in February
retail sales. M eanw hile, p ro du cer prices fo r March
advanced more m odestly than in the previous two
months. Together, these developments helped to dispel
e xpe ctatio n s th at m onetary p olicy would be firm ed
again, and yields edged off the levels reached late in
March. As May progressed and incom ing data sug
gested a further slowing in economic activity, the mar
ket began to anticipate an easing in the policy stance.
A dollar that showed strength against m ajor foreign
currencies also exerted downward pressure on yields.
Yields tumbled in mid-May after the release of the April
PPI, which showed a slight decline when the volatile
food and energy com ponents were excluded. These
developm ents were reinforced in early June by the
report of weak job gains in May. Moreover, the pur
chasing m anagers’ index dropped to 49.7 percent, the
first reading below 50 percent in thirty-three months. (A
reading below 50 percent implies that activity in the
manufacturing sector is contracting.) Chairman Green
span’s co nce rn s about w eakness in the econom y,
expressed during his July 20 Humphrey-Hawkins testi
mony, b riefly added support to the markets. Also in
July, the yield on the two-year note fell below that on
the thirty-year bond, and the yield curve took on a pos
itive slope for m aturities between two and th irty years.
In August and September, economic activity showed
some signs of vigor, but growth was not expected to
exacerbate inflationary pressures. In this environment,
p olicy was expected to rem ain steady, and yields
moved slightly higher because several easing moves
had already been incorporated into the yield structure.
News of sizable job gains in July, along with a substan
tial upward revision to June’s employment rise, pres
sured yield s higher in early A ugust. U n ce rta in tie s
about financing provisions of the th rift legislation and
about the Treasury’s debt ceiling added briefly to the
pressures, particularly in the Treasury sector. (In early
August, the Treasury obtained a tem porary increase in
the ceiling that lasted until October 31.) There followed
8One outgrowth of the higher yields on shorter dated Treasury issues
in the early months of the year was a surge in noncompetitive
tenders, a measure of individual investor interest, at auctions of
Treasury bills and short-dated notes.
FRBNY Quarterly Review/Spring 1990
47
a series of mixed economic reports that, on balance,
supported the perception of a moderate pace of eco
nom ic a c tiv ity . The p ro d u c e r and co n su m e r p rice
indexes reported during this time generally suggested
lower inflation than earlier in the year.
Yields moved lower on balance over the final three
months of the year, based in part on expectations that
the signs of sluggish economic activity would lead to
additional moves to ease policy. Market participants
increasingly focused on the perform ance of the manu
facturing sector, which appeared to be contracting at
the same time that other sectors of the economy were
showing signs of continued growth. Each of the final
th re e e m p lo y m e n t re p o rts re le a s e d d u rin g 1989
showed a marked fall in manufacturing employment.
Further evidence of a manufacturing slowdown was
found in the purchasing m anagers’ index and the indus
trial production index. Meantime, prices seemed to be
rising at a slower pace than in the early months of the
year. Yield declines, especially on short-term issues,
were fostered by prospective and actual System moves
to ease policy. Indeed, the System reduced reserve
pressures on three more occasions before year-end.
(However, yields responded only briefly to the Decem
ber easing move because the easing had been antici
pated and was already almost fully reflected in yields.)
U.S. Treasury securities
The Treasury yield curve was hump-shaped from the
beginning of the year until early July and again from
m id-August to mid-October (Chart 3). Yields on Treas
ury bills were generally below those on short-dated
coupon issues, w hich in turn m ostly exceeded the
Chart 3
Yield Curves for Selected
U.S. Treasury Securities
Percent
10.50
10.00
-I
-IV arch 22, 198
9.50
9
i
9.00 \ f - \ ------- December 28, 1986
8.50 •
/ I T
|
8.00
|
i
___
r i .. i
1
,i
5
I
|
■”
I
..1,, _ L
10
tem ber27, 1989
—
__Auaust 2. 1989___
1 i 1 I..1.1.1
15
20
Years to maturity
| -L
25
Note: Treasury bill yields are on a bond-equivalent basis.
Coupon yields are constant maturity values.
48for FRBNY
Digitized
FRASER Quarterly Review/Spring 1990
1
n
30
yields on the thirty-year bond. During the remainder of
the year, the yield curve was relatively flat, although bill
yield s fre q u e n tly exceeded those on sh o rt coupon
issues. On balance, yields on Treasury coupon securi
ties, as m easured by the co nsta nt m a tu rity series,
declined between 110 and 140 basis points in 1989,
with sm aller reductions on the longer maturities. Treas
ury bill rates fell 60 to 120 basis points, with the largest
decline recorded for fifty-tw o week bills.
From time to time during the year, yields on Treasury
issues were pushed lower when market disturbances
elsewhere set off “ flight-to-quality” demand. The most
dramatic example occurred in mid-October. Yields fell
on October 13 in response to the late-afternoon, 190point plunge in the Dow Jones industrial average. The
sell-off in stocks was sparked by the failure of a bid
ding group to arrange financing for its proposed take
over of UAL Corporation. The stock market sell-off led
investors to seek the safe haven of Treasury issues.
The yield declines were partially retraced the next trad
ing day as stock prices recovered, but yields remained
below their prior levels, partly because of the soft fed
eral funds rate.
Debt ceiling limitations complicated Treasury financ
ing toward the end of O ctober and briefly affected
yields. Bill rates jumped when the Treasury announced
an earlier than usual settlem ent date for its October 30
bill auctions. The Treasury adopted the earlier settle
ment in order to raise as much cash as possible under
the enlarged tem porary debt ceiling before the ceiling
expired on October 31. The start of the Treasury’s m id
quarter refunding auctions and a regular w eekly bill
auction were postponed until after a new $3.12 trillion
debt ceiling was enacted on November 8. Potential
upward pressures on coupon yield s from the co m
pressed financing schedule were offset by expectations
of a falling rate pattern.
T h rift le g is la tio n a n d its im p a c t on Treasury a nd
agency borrowing
The federal government’s efforts to raise cash to man
age the closing or merger of insolvent th rift institutions
had a significant impact on borrowing by the Treasury
and by U.S. government-sponsored agencies in 1989.
The Financial Institutions Reform, Recovery and En
forcem ent Act of 1989 (FIRREA), originally proposed by
President Bush in February and enacted on August 9,
set forth the framework within which the th rift industry
problems were to be resolved. The legislation was also
aimed at overhauling the institutional structure and the
rules for supervising and regulating the entire industry.
One provision created the Resolution Trust Corporation
(RTC), which was empowered to take possession and
dispose of the assets of failed thrifts over the next sev
eral years. It inherited this role from the Federal Sav
ings and Loan Insurance Corporation (FSLIC), which
discontinued its operations.
The RTC was authorized to spend a net total of $50
billion to resolve the problems of insolvent thrifts. The
legislation stipulated that $18.8 billion of the outlays
were to be financed out of general revenues, and Con
gress appropriated the funds in fiscal 1989. About half
of the appropriated funds had been spent by the end of
the 1989 fiscal year, and it was expected that the
remaining portion would be used over the following two
years. The RTC was to acquire the other $31.2 billion
through the sale of capital certificates to the Resolution
Funding Corporation (REFCORP), a new agency estab
lished by FIRREA .9 To finance its purchase of RTC
capital certificates, REFCORP was authorized to sell
$30 billion of long-term bonds in fiscal years 1990 and
1991, while the Federal Home Loan Banks contributed
another $1.2 billion in fiscal year'1989. Although
REFCORP bonds are not obligations of, nor is their
principal guaranteed by, the U.S. government, they
have strong federal backing. Before each bond issue,
REFCORP, using th rift industry funds, purchases
directly from the Treasury zero-coupon securities with
a principal amount and maturity date that match the
REFCORP obligation, thus defeasing the principal. Fur
thermore, interest on REFCORP borrowing is to be
paid out of Treasury and thrift industry funds, with the
Treasury guaranteeing all interest payments.
The Treasury’s borrowing operations during the year
were affected by these efforts to meet the U.S. govern
ment’s liabilities to th rift depositors. The Treasury
raised part of the $18.8 billion appropriated by the
Congress by increasing Treasury bill issuance. In antic
ipation, bill rates moved higher as the passage of
FIRREA neared. The Treasury expanded the sizes of
the regular weekly bill auctions and of the fifty-two
week bill auctioned on August 24 and raised an addi
tional $5 billion through a 247-day cash management
bill auctioned on August 10. Subsequently, the prospect
of the sale of REFCORP bonds placed some upward
pressure on yields of longer dated Treasury securities.
Nonetheless, the added borrowing undertaken to fund
RTC’s expenditures appeared to have little lasting
impact on interest rates in the Treasury market in 1989.
REFCORP entered the public debt market for the
first time on October 25 and auctioned $4.52 billion of
thirty-year bonds —the agency’s only offering in 1989.
Dealers approached the issue cautiously. Having no
experience with such issues, they were uncertain how
actively the bonds would trade in the secondary mar
9Receipt of these funds by the RTC is scored as a negative outlay in
the federal budget accounts, thereby offsetting positive outlays of an
equivalent amount.
ket. The auction went well, with the average yield about
28 basis points above the yield on the Treasury’s thirtyyear bond. The spread remained near this level in sub
sequent trading during the balance of the year,
although actual trading was generally light. Through
the end of 1989, just over one-quarter of the issue was
stripped to satisfy demand for zero-coupon instru
ments. As required by FIRREA, before the settlement
of the issue, REFCORP purchased the zero-coupon
Treasury bonds needed to ensure repayment of the
principal, at a cost of about $400 million.
In related agency borrowing, the Financing Corpora
tion (FICO) issued a total of $2.3 billion of thirty-year
bonds during the year and used up much of its remain
ing borrowing authority. FICO was created in 1987 as a
subsidiary of the Federal Home Loan Bank Board
(FHLBB) and was authorized to borrow up to $10.8 bil
lion to help recapitalize FSLIC, which at the time was
under the supervision of the FHLBB.
Other U.S. government agency securities
The Tennessee Valley Authority (TVA) returned to the
public debt market for the first time in fifteen years by
selling $4 billion of bonds in October and again in
November. The proceeds of these sales were used pri
marily to refinance (through defeasance) roughly $7
billion in high-coupon debt held by the Federal Financ
ing Bank (FFB), the agency through which the TVA had
previously arranged its financing. TVA officials esti
mated that the refinancing could save TVA as much as
$100 million per year in interest expenses. Typically, an
agency that borrows directly in the public market can
not borrow from the FFB; however, TVA obtained an
alternative credit facility for $2 billion from the FFB for
the next two years. Strong investor demand for the
offerings materialized and their sizes were increased
from their originally planned levels. The November
issue included $2.5 billion of forty-year bonds, callable
after ten years. These bonds were unusual because of
their long maturity .10 They were initially priced to yield
110 basis points over the thirty-year Treasury bond,
which is fully protected against an early call, and the
spread had narrowed somewhat by year-end.
Corporate bonds
Public debt issued by U.S. corporations in the domestic
bond market declined for the third consecutive year in
1989; such issuance fell by 12 percent to $177.4 bil
lion .11 Total issuance was heaviest in the spring and
10Several telephone companies and foreign entities have offered
callable forty-year debt in recent years.
11Data oh corporate and municipal debt issuance were supplied by the
Board of Governors of the Federal Reserve System.
FRBNY Quarterly Review/Spring 1990
49
fall, when borrow ers sought to take advantage of
ebbing interest rate levels. The dropoff in total new
offerings stemmed from a sharp cutback in issuance of
m ortgage-backed securities and a decline in issuance
of b e lo w -in v e s tm e n t-g ra d e s e c u ritie s .12 These d e
creases were partially offset by a modest increase in
inve stm en t-grad e o ffe rin g s and by another jum p in
asset-backed securities that was likely prom pted by
continued efforts to restrain asset growth to meet capi
tal standards. Mortgage-backed issuance fell because
of slow activity in the housing market and because the
relatively flat Treasury yield curve limited profit poten
tia l from the issu a n ce of c o lla te ra liz e d m ortg a g e
obligations.
Yields on highly rated corporate bonds fell about 75
to 85 basis points, but Treasury yields dropped even
more, so that spreads between yields on investmentgrade corporate issues and yields on Treasury securi
ties widened throughout the year (Chart 4). The wider
spreads in part may have reflected investor concern
about holding corporate bonds in a weakening econ
omy. Spreads on debt of individual com panies also
depended on the com panies’ “event-risk” covenants. In
1988, the leveraged buyout of RJR Nabisco made clear
that all but a few firms were subject to mergers, take
overs, or recapitalizations that could cause their out
standing bonds to lose their investment-grade status.
Consequently, bondholders demanded higher yields to
hold bonds that did not have protection against such
occurrences. In 1989, more new issues carried eventrisk protection. One such device, the “ poison put,” per
mits bondholders to resell their bonds to the issuer at a
set price if specified events cause the bonds to lose
their investm ent-grade status. Bonds with event-risk
protection generally had lower yields than sim ilarly
rated issues lacking such protection. To address eventrisk concerns, Standard and Poor’s introduced in July a
rating system that evaluates event-risk covenants. The
covenant rankings assess the degree of protection pro
vided in bond indentures against a sudden and dra
matic decline in credit quality resulting from a takeover
bid, recapitalization, or sim ilar restructuring. E-1 repre
sents the highest level of protection, and E-5 indicates
the lowest level.
Som e b an k h o ld in g co m p a n y (BH C ) d e b t w as
affected by problem loans to domestic real estate ven
tures late in the year. As real estate markets weak
ened, especially in the Northeast, some BHCs had to
increase their loan-loss reserves to account for prob
lems with th eir real estate portfolios, a move which
resulted in depressed earnings. Yield spreads on BHC
12Below-investment-grade bonds are those rated lower than Baa by
Moody's or, if not rated by Moody's, below BBB by Standard and
Poor’s.
50for FRASER
FRBNY Quarterly Review/Spring 1990
Digitized
debt over Treasury issues widened, and the ratings of
some BHC debt were lowered.
In other developm ents affecting the operations of
BHCs, the Federal Reserve Board in January granted
approval to five BHCs to underw rite corporate debt,
contingent upon the Board’s acceptance of the individ
ual BHC’s plan to capitalize its debt-underw riting oper
ations. The Board ruled that such underwriting must be
conducted by a separate subsidiary that does not gen
erate more than 5 percent of its total gross revenue
from underw riting corporate debt and certain other
securities. (This limit was raised to 10 percent in Sep
tember.) In addition, with limited exceptions, federally
insured banks and thrifts cannot provide loans to their
affiliated underwriting subsidiaries. In July, J.R Morgan
Securities, a subsidiary of J.R Morgan Bank C orpora
tion, became the first subsidiary of a BHC to partici
pate in a s y n d ic a te d u n d e rw ritin g of c o rp o ra te
securities since the passage of the Glass-Steagall Act
in 1933, and later becam e the first bank subsidiary
since that time to act as the lead underwriter for a cor-
Chart 4
Yield Spreads
Basis points
800
600
____ Difference between Donaldson, Lufkin & ____ / W
Jenrette index of high-yield active issues
r'
and seven-year Treasury index*
/
v s
-----
500
400
300 l l M l M l l l M l l l J j l l l M l l i l l l l l J l l l l L l l J . i l
Im
i i
Im i!
. Difference between Moody's Aaa-rated .
corporate bond index and
ten-year Treasury index_______
-50
Difference between Moody’s Aaa-rated
_^q q _______ municipal bond index and
ten-year Treasury index
f\
-150 —
f
_______________________ rJ_
. A A
A
-200 - v —---- ^ v ' N /
/ v/
.250 11 I 11 I I 11 I I 11l I I I I 11I I l l I I l 111111 I il l I l I 11 I I I 11 l I i I l
J
F
M
A
M
J
J
A
S
O
N
D
1989
r High-yield index provided by Donaldson, Lufkin & Jenrette.
porate bond offering.
Yields on below-investment-grade or “ junk” bonds
rose sharply during 1989 as investor wariness about
holding such securities intensified in the face of a slow
down in economic activity and the financial difficulties
of several major issuers. The spread between yields on
junk bonds and those on Treasury securities began to
widen in the spring and summer as market expecta
tions of an economic slowdown took hold and raised
doubts about the ability of many issuers of junk bonds
to repay their debts. These doubts were underscored in
mid-June when Integrated Resources, a real estate
partnership syndicate, declared its inability to make a
pending interest payment because of short-term fund
ing problems .13
Yields on junk bonds were boosted even further over
the second half of the year. In mid-September, Campeau Corporation, the Canadian-based owner of Allied
Stores and Federated Department Stores, announced
that it did not have funds to make interest payments on
outstanding bonds of Allied Stores. The value of bonds
sold by both Campeau units tumbled, as did prices on
outstanding issues of other retail establishments. Even
though Campeau received a cash infusion from Olym
pia and York that enabled it to meet its immediate inter
est obligations, prices on Allied and Federated debt
remained depressed as the company’s funding prob
lems persisted .14 The episode increased sensitivity to
the characteristics of specific issues in the junk bond
market. Over the remainder of the year, a nervous
underton e lingered in the m arket, sustained by
rumored or actual adverse developments at many com
panies. “ High-quality” junk bonds held their value bet
te r than “ lo w -q u a lity ” ju n k bonds. Trading was
periodically volatile, and it ground to a virtual halt for a
few days after the stock market declined precipitously
on October 13. By year-end, the spread between the
Donaldson, Lufkin and Jenrette index of yields on
actively traded junk bonds and their index of yields on
Treasury securities with seven years to maturity had
almost doubled from its level at the start of the year
(Chart 4).
Because of the growing problems experienced in this
sector, total issuance of junk bonds during the year fell
to $28.7 billion, about 8 percent below the previous
year’s level. The pace of new offerings dropped off
considerably in the second half of the year in light of
the unsettled market conditions. Included in the year’s
total issuance was an offering of $4 billion of RJR
13lntegrated Resources adopted a restructuring plan later in 1989 but
was forced into bankruptcy in February 1990.
14Allied Stores and Federated Department Stores ultimately filed for
protection under Chapter 11 of the bankruptcy code in January 1990.
Holdings Capital Corporation securities in May —the
largest corporate offering ever. The proceeds were
used to repay short-term loans arranged as part of the
$25 billion leveraged buyout of RJR Nabisco by
Kohlberg Kravis Roberts and Company that was com
pleted in February.
Several other developments during the year also
affected the demand for junk bonds. The August thrift
rescue legislation required savings and loans institu
tions to divest their holdings of low-rated bonds by
1994, although separately capitalized affiliates were
still permitted to invest in such debt; over the remain
der of the year, sizable thrift selling was noted at times.
In November, as part of its budget legislation, Con
gress imposed limits on the deductibility of interest
payments on certain securities that have a maturity
greater than five years, that defer interest payments,
and that have a yield to maturity more than 5 percent
age points above the Applicable Federal Rate, as
defined by the Internal Revenue Service. Both legisla
tive changes had been widely anticipated and had little
immediate impact on the market for low-rated securi
ties, but they underscored growing congressional con
cern about the issuance of such debt, especially to
finance corporate takeovers.
Municipal bonds
The municipal bond market remained relatively quiet in
1989. Total issuance for the year was $113.6 billion,
close to the $114.5 billion issued in 1988. New-money
issues posted a 5.5 percent increase, rising to $84 bil
lion, while refunding issues declined 15 percent to
$29.6 billion. The pace of new issuance was somewhat
faster over the second half of the year, when munici
palities took advantage of lower interest rates.
Yields on highly rated municipal bonds declined 55
to 65 basis points. Movements in municipal bond yields
roughly followed those on Treasury securities, although
the spread between yields on municipal bonds and
those on Treasury securities narrowed somewhat over
the year (Chart 4). The smaller spread over the second
half of the year in part reflected the increased pace of
new issuance at that time. Two other factors also con
tributed. Sizable additions to loan-loss reserves during
the second half of the year reduced many commercial
banks’ needs for tax-exempt income and decreased
their demand for municipals. In addition, some tax ben
efits of holding municipal issues expired at the end of
the year, thus prompting some institutional selling.
A notable development in the municipal bond market
during the year was the reentry of the Washington Pub
lic Power Supply System (WPPSS) in September, when
it sold $721 million of refunding revenue bonds backed
by projects 1, 2, and 3. The bonds were rated A by
FRBNY Quarterly Review/Spring 1990
51
Moody’s and A A - by Standard and Poor’s. This offer
ing marked the first time that WPPSS issued municipal
bonds since it defaulted on $2.25 billion of projects 4
and 5 bonds in 1983 —the largest default in the munici
pal market to date. After some delay because of legal
com plications, the offering went smoothly. Strong
investor demand enabled WPPSS to increase the size
of the new issue from its originally planned level of
$450 million, although yields were about 25 basis
points above those on similarly rated long revenue
bonds. WPPSS sold an additional $738 million of
bonds in December.
Monetary aggregates
Growth of all three m onetary aggregates and total
domestic nonfinancial debt decelerated in 1989 (Chart
5). After having slowed in the latter half of 1988, M2
and M3 growth rates were even more sluggish over the
first half of 1989, while M1 actually contracted. Growth
of M1 and M2 rebounded sharply over the final two
quarters of the year. Despite this rebound in M2 growth
and a modest pickup in bank credit expansion, M3
growth decelerated further because of factors asso
ciated with the restructuring of the thrift industry. Debt
expansion was a bit greater in the first half of the year
than in the second. Overall, M2 and M3 grew 4.6 and
3.2 percent, respectively, from the fourth quarter of
1988 to the fourth quarter of 1989. M1 eked out a gain
of 0.6 percent; total nonfinancial debt expanded at an
8.0 percent rate. These rates of expansion placed
fourth-quarter M2 slightly below the midpoint of the
FOMC’s growth cone and placed M3 just below its
cone. The debt measure finished the year slightly
below the midpoint of its monitoring range.
In February, the FOMC reaffirmed the 1989 growth
ranges for M2 and M3 that it had tentatively estab
lished the preceding June. These ranges called for
growth of 3 to 7 percent for M2 and 31A> to 71/2 percent
for M3, compared with a range of 4 to 8 percent for
both M2 and M3 in 1988. The reduction of the growth
ranges for 1989 was considered to be consistent with
progress towards price level stability and underscored
the Committee’s commitment to an anti-inflationary pol
icy. The width of these ranges was maintained at 4 per
centage points in recognition of the degree to which
the relationship between the monetary aggregates and
economic performance has varied in recent years. M2
in particular has become very sensitive to fluctuations
in interest rates. Consequently, the Committee agreed
to evaluate money growth in light of other indicators,
including inflationary pressures, the strength of the
business expansion, and developments in domestic
financial and foreign exchange markets.
The FOMC also reaffirmed the tentative monitoring
52for FRBNY
Digitized
FRASER Quarterly Review/Spring 1990
range for total domestic nonfinancial debt that it had
established in June 1988, and again decided not to
specify a target range for M1 growth. It adopted a mon
itoring range for debt growth of 6V2 to IOV2 percent,
compared with the range of 7 to 11 percent for 1988. In
deciding not to set a target range for M1 growth, the
Committee continued to view the relationship between
M1 and economic activity as too unpredictable to war
rant reliance on this measure as a guide for the con
duct of monetary policy.
Most of the short-term variation in the demand for
M2 around its trend can be explained by the movement
of the spread between a market interest rate, such as
the three-month Treasury bill rate, and the average rate
paid on M2 deposits; this spread can be interpreted as
the opportunity cost of holding M2 deposits .15 The rate
of growth of M2 is likely to decline, usually with a lag,
as the opportunity cost of holding M2 assets rises.
Short-term variations in M2’s opportunity cost arise
because the rates o ffered on m ost M2 d e p o sits
respond sluggishly to movements in market rates.
When holders of M2 deposits observe that the rates
paid on these deposits are not keeping pace with the
increases in market rates, they will redeploy some of
their M2 holdings into higher yielding money market
instruments and thus depress M2 growth. Gradually, as
market rates stabilize, rates offered on most M2
deposits tend to catch up with the adjustment in market
rates, and the opportunity cost of holding M2 moves
back toward its usual level. As this happens, people
readjust the proportion of their financial assets in M2
toward the earlier ratio, speeding up the growth of M2
in the process.
The impact of a change in market interest rates on
the growth of individual components of M2 depends on
the speed at which the average offering rate for that
component is adjusted. Banks typically adjust the offer
ing rates on NOW accounts, money market deposit
accounts (MMDAs), and savings accounts relatively
slowly. Demand deposits pay no explicit interest by law,
and im plicit returns are altered gradually through
adjustments to charges and services associated with
the account. Rates on money market mutual funds and
small time deposits respond much more quickly to
changes in market rates.
The deceleration of M2 growth over the first two
quarters of 1989 largely resulted from changes in the
opportunity cost of holding money and from the unex
pectedly large tax liabilities faced by individuals in
April. The average spread between market rates and
those on M2 deposits widened in the first quarter; how15See David H. Small and Richard D. Porter, “ Understanding the
Behavior of M2 and V2,” Federal R eserve Bulletin, April 1989,
pp. 244-54.
Chart 5A
Chart 5B
M2: Levels and Target Ranges
M3: Levels and Target Ranges
Cones and Tunnels
Cones and Tunnels
Billions of dollars
4300
Billions of dollars
3300
3200
3100
3900
3000
3800
3700
2900
3600
3500
2800
ONDJ FMAMJ J AS OND J FMAMJ J ASOND
1987
1988
1989
Chart 5D
Chart 5C
Total Domestic Nonfinancial Debt Levels and
Monitoring Ranges
M1 Levels and Growth Rates
Billions of dollars
Cones and Tunnels
Billions of dollars
10200
10000
9800
9600
9400
9200
9000
8800
8600
8400
8200
8000
1987
1988
1989
1987
1988
1989
FRBNY Quarterly Review/Spring 1990
53
ever, it began to narrow in the second quarter as mar
ket rates fell from their highs and deposit rates lagged
behind. Funds may have been funneled into taxes or
nonmonetary assets rather than into M2 deposits —
noncompetitive tenders at Treasury security auctions
were already quite large during the firs t quarter.
Deposits whose rates adjust slowly contracted mark
edly during the first two quarters, with especially pro
nounced outflows in April and May, when individuals
appear to have drawn down their existing balances in
these accounts to meet unanticipated tax obligations.
The sizable declines in demand and other checkable
deposits over the first half of the year caused M1 to fall
sharply. W ithin M2, however, the co n traction of
deposits with relatively unresponsive rates was offset
by gains in small time deposits and money market
mutual funds, especially in the second quarter, when
the average rates on small time deposits and money
funds exceeded those on six-month Treasury bills. On
balance, M2 expanded at an anemic 2.0 percent rate
over the first two quarters, while M1 fell at a 2.3 per
cent annual rate.
The weak expansion of M2 depressed M3 growth.
The non-M2 component of M3 grew briskly in the first
quarter as banks stepped up their issuance of large
time deposits to help fund the modest pace of loan
expansion. The growth of these managed liabilities
moderated in the second quarter because banks were
able to fund credit expansion, which remained modest,
with tax-swollen Treasury tax and loan account bal
ances. Thrift issuance of managed liabilities slowed
from its pace in the latter half of 1988, perhaps reflect
ing heavier reliance on Federal Home Loan Bank
advances to fund credit expansion. On net, M3 grew at
a 3.6 percent rate over the first two quarters of the
year.
At the time of the FOMC’s midyear review of the
growth of the aggregates, M2 was about 1 percentage
point below the lower bound of its growth cone, while
M3 was at its lower bound. Total financial debt stood in
the middle of its monitoring range. M1, meanwhile, was
considerably below the level it had attained on average
during the fourth quarter of 1988. M2 and M3 were
expected to show stronger growth in the second half of
the year in light of the recent declines in market inter
est rates. F u rth e rm o re , th e se a g g re g a te s w ere
expected to finish the year well within their target
ranges. Against this background, the Committee re
affirmed the 1989 target and monitoring ranges.
Over the second half of the year, M2 growth acceler
ated m arkedly as the o p p o rtu n ity cost of holding
deposits narrowed. Deposits with relatively unrespon
sive rates expanded considerably and nearly recovered
the outflows of the first half of the year. Money market
54
FRBNY Quarterly Review/Spring 1990
mutual funds showed sizable m onthly increases,
despite the narrowing spread of their offering rates
over market rates. The strong inflows into these funds
likely reflected the fact that their rates exceeded those
on other monetary instruments. In addition, because
money market funds are perceived as a means of
avoiding the volatility of bond and equity funds, they
may also have benefited from the mounting losses on
junk bond funds and the sharp drop of stock prices on
October 13. The growth of small time deposits slowed,
in part because their rate advantage over some market
rates eroded markedly. On balance, M1 and M2 grew at
rates of 3.5 and 7.1 percent, respectively, over the final
two quarters.
The troubles of the thrift industry appear to have
affected the composition of M2 but not its overall
growth. Thrift small time deposits declined from Sep
tember through December, while other retail th rift
deposits grew slowly. The fall in th rift small time
deposits probably reflected the shrinking spread
between thrift and commercial bank rates on these
deposits. With regulators actively discouraging thrifts
from offering unduly high rates and with troubled insti
tutions (which generally offered the highest rates)
being seized, th rift rates on sm all tim e deposits
declined more than those offered by fcommercial banks.
The shrinkage in thrift small time deposits, together
with the modest growth of other thrift M2 deposits,
appears to have been offset by flows into commercial
bank deposits and money market mutual funds. Conse
quently, commercial banks held a greater share of M2
deposits at the end of the year than at the beginning.
Unlike M2 growth, the growth of M3 in the second
half of the year was significantly restrained by the
restructuring of the thrift industry. FIRREA imposed
strict capital requirements on thrifts, placed limitations
on the structure of their portfolios, and provided for the
use of RTC funds to pay off depositors at liquidated
institutions. FIRREA had its most pronounced impact
on M3 through its effect on the funding practices of
inadequately capitalized thrifts. These thrifts were
required to reduce their balance sheets, and they did
so by restricting their issuance of term repurchase
agreements and large time deposits over the second
half of the year.16 Together, these liabilities fell at a 34
percent annual rate over the final two quarters. Mean
time, banks funded the modest pickup in credit expan
sion w ith M2 d e posits so that th e ir issuance of
managed liabilities was weak. On net, M3 expanded at
16Thrifts also reduced their issuance of overnight RPs, which were
added to M2 in the 1990 redefinition of that aggregate. From June to
December 1989, overnight thrift RPs shrank by $1.1 billion; they
stood at $2.5 billion in December. Although the decline was sharp,
the RPs represent such a small share of the broader aggregates that
the impact on M2 and M3 growth was minor.
a modest 2.9 percent rate over the final half of the
year.
The income velocities of the monetary aggregates all
grew faster than their 1982-88 average rates of growth
(Chart 6).17 The velocity of M2 increased at a 1.8 per
cent rate in 1989, compared with a 2.1 percent rate in
1988. The velocities of M3 and M1 advanced far more
quickly than in 1988. M3 velocity grew 3 percent, while
M1 velocity grew 5.8 percent. They had advanced 1.2
and 3.1 percent, respectively, in 1988. The velocity of
nonfinancial debt fell 1.5 percent, a slightly greater rate
of decline than in the previous year.
Policy implementation
In 1989, the FOMC expressed its desired policy stance
in terms of the degree of reserve pressure, a practice it
has followed, with some modifications, since 1983. The
intended degree of reserve pressure is described as a
designated amount of adjustment and seasonal bor
rowing at the discount window. The Trading Desk uses
this indicated amount of borrowing to derive the objec
tive for nonborrowed reserves for the two-week reserve
maintenance period. The nonborrowed reserve objec
tive is obtained by estimating the demand for total
reserves, constructed by projecting required reserves
and desired excess reserves, and then subtracting
from that estimate the intended level of discount win
dow borrowing. Revisions are made to the objective
during the maintenance period when new information
suggests modifications to the estimated demand for
total reserves. To achieve the nonborrowed reserve
objective, the Desk conducts open market operations
to increase or decrease the supply of nonborrowed
reserves; however, the supply of nonborrowed reserves
in the banking system is also influenced by the move
ments of various “ operating factors” over which the
Desk has little control. As a result, when the Desk
undertakes its operations, it faces uncertainties both
about reserve demand and about the amount of
reserves supplied by the operating factors.
For a given level of the discount rate, higher levels of
borrowing have typically been associated with firmer
money market rates because limitations are placed on
access to the discount window. When higher amounts
of borrowing are desired, fewer nonborrowed reserves
are supplied for a given level of demand for total
reserves. With nonborrowed reserves less plentiful and
with frequent or heavy use of the discount window dis
couraged, depository institutions bid more aggressively
for reserves in the money market and ultimately cut
back on their lending and investing. In this process,
17The income velocity of an aggregate is the ratio of nominal GNP to
the level of the aggregate.
short-term interest rates rise .18
During 1989, however, as in some previous years, the
relationship between the amount of borrowing and the
degree of money market firmness, as measured by the
spread between the federal funds rate and the discount
rate, was somewhat unreliable. For the most part,
banks appeared less inclined to borrow adjustment
credit than in earlier years. The unusual reluctance of
banks to borrow from the discount window complicated
the Desk’s implementation of policy through use of the
borrowed reserve procedure throughout 1989 and
encouraged a flexible interpretation of the objectives
for nonborrowed and borrowed reserves.
As the 1988 report on open market operations
related more fully, banks have shown particular reluc
tance to borrow on a number of other occasions in the
1980s.19 In late 1988, the relationship between borrow
ing and the fe d e ra l fu n d s -d is c o u n t rate spread
appeared to shift once more. Banks became even less
disposed to borrow adjustment credit than they had
been earlier in the year; thus, a much larger spread
between the federal funds rate and the discount rate
was needed in order to induce banks (in the aggre
gate) to borrow the same amount that they would have
before the shift. As a consequence, strict adherence to
the nonborrowed reserve objective implied by a given
level of assumed borrowing would often have forced
federal funds to trade persistently at rates that were
higher than those anticipated by the FOMC. In both
1988 and 1989, the Committee responded to these
shifts by taking account of the observed degree of
reluctance to borrow when it chose the borrowing
allowances. However, it recognized the persisting
uncertainty about the relationship between borrowing
and the federal funds rate and thus encouraged the
Desk to view the assumed levels of borrowing flexibly
in order to achieve the desired degree of restraint.
(Notes on the FOMC directives and the borrowing
assumptions used to construct the reserve paths are in
Table 1.) The Desk exhibited flexibility by accepting
deviations of borrowing from its assumed level when
the deviations were consistent with holding to the
money market conditions anticipated by the FOMC.
Adjustment and seasonal borrowing fell short of the
desired level in four of the first five maintenance
periods of the year. (Actual reserve data appear in
Table 2 .) Consequently, a decision was made to allow
18For a more detailed description of the borrowed reserve procedure,
see Brian F. Madigan and Warren T. Trepeta, “ Implementation of U.S.
Monetary Policy," in Changes in Money-Market Instruments and
Procedures: Objectives and Implications, Bank for International
Settlements, March 1986.
■••"Monetary Policy and Open Market Operations during 1988," Federal
Reserve Bank of New York Quarterly Review, Winter-Spring 1989,
pp. 83-102.
FRBNY Quarterly Review/Spring 1990
55
Chart 6A
Chart 6B
M2 Velocity Growth
M3 Velocity Growth
Percent
Percent
8 -----------------------
1972
74
76
78
80
82
84
86
88 89
Notes: Velocity growth is measured from four quarters earlier.
Shaded areas represent periods of recession as defined by the
National Bureau of Economic Research.
Notes: Velocity growth is measured from four quarters earlier.
Shaded areas represent periods of recession as defined by the
National Bureau of Economic Research.
Chart 6D
Chart 6C
Total Domestic Nonfinancial Debt Velocity Growth
M1 Velocity Growth
Percent
Percent
Notes: Velocity growth is measured from four quarters earlier.
Shaded areas represent periods of recession as defined by the
National Bureau of Economic Research.
Digitized
56for FRASER
FRBNY Quarterly Review/Spring 1990
Notes: Velocity growth is measured from four quarters earlier.
Shaded areas represent periods of recession as defined by the
National Bureau of Economic Research.
for the increased reluctance of banks to approach the
discount window by reducing the borrowing allowance,
on March 9, to a level that was in line with actual expe
rience and th at w ould m aintain the e xisting policy
stance. (Policy had been firmed in January and Febru
ary.) This diminished desire by banks for adjustment
credit persisted for the rem ainder of the year. With
adjustment borrowing generally running at low levels,
swings in seasonal credit tended to dominate move
m e n ts in th e s e rie s “ a d ju s tm e n t p lu s s e a s o n a l
borrowing.”
Adjustm ent borrowing was particularly light over the
last half of the year, when the funds rate generally
exceeded the discount rate by sm aller amounts than in
the firs t half of the year. Adjustm ent credit was fre
quently quite low until the final day of a maintenance
period, when borrowing sometimes rose in the face of
settlem ent-day pressures. As the FOMC eased reserve
pressures over the second half of the year, adjustment
borrowing tailed off to average about $165 million over
the final thirteen maintenance periods of the year, and
even this average was lifted by intervals of somewhat
heavier borrowing associated with natural disasters
and year-end pressures. Adjustm ent credit averaged
less than $50 million during the September 6, Novem
ber 1, and December 13 periods. In the September 6
Table 1
S pecification s fo r D irectives o f the Federal Open Market Com m ittee and Related Inform ation
Prospective Reserve Restraint M odifications
Date of
M eeting
S pecified Short-Term
G rowth Rates
M2
M3
Borrowing
Assum ption
for D eriving
NBR Path
C om m ittee
Preference
G uidelines for
M odifying
Reserve
Pressure
Sought to
increase
som ewhat the
deg ree of
pressure on
reserve
positions
A somewhat
greater degree
w ould be
accep table. A
slightly lesser
degree m ight
be accep table
Indications
of
inflationary
pressure
Strength of
the business
expansion
Behavior of
the monetary
aggregates
Developm ents
in foreign
exchange and
dom estic
financial
m arkets
6.50
Sought to
7.00 on
m aintain the
2/24 existing degree
of pressure on
reserve
positions
A somewhat
greater degree
would be
a ccep table. A
slightly lesser
degree m ight
be a ccep table
Indications
of
inflationary
pressure
Strength of
the business
expansion
Behavior of
the monetary
aggregates
Developm ents
in foreign
exchange and
dom estic
financial
m arkets
Discount
Rate
Factors to C o nsider for M odifications
(In O rder Listed)
1
2
3
4
(Percent)
(M illions of
Dollars)
(Percent)
12/13 to
12/14/88
Novem ber to M arch
3
6Vz
400
500 on 12/15
600 on 1/5
6.50
2/7 to
2/8/89
D ecem ber to M arch
2
31/j
600
700 on 2/14 f
500 on 3 /9 *
M arch to June
5
500
7.00
3
Sought to
m aintain the
existing deg ree
of pressure on
reserve
positions
A somewhat
greater degree
w ould be
accep table. A
slightly lesser
degree m ight
be acce p ta b le
Indications
of
inflationary
pressure
Strength of
the business
expansion
B ehavior of
the monetary
aggregates
D evelopments
in foreign
exchange and
dom estic
financial
m arkets
M arch to June
4
500
600 on 5 /1 1%
500 on 6/6
7.00
1Va
Sought to
m aintain the
existing deg ree
of pressure on
reserve
positions
A som ewhat
g reater or
som ewhat
le sser degree
w ould be
accep table
Indications
of
inflationary
pressure
Strength of
the business
expansion
B ehavior of
the m onetary
a ggregates
Developm ents
in foreign
exchange and
dom estic
financial
m arkets
June to Septem ber
7
7
500
600 on 7/7§
550 on 7/27
7.00
Sought to
d ecrease
slig h tly the
degree of
pressure on
reserve
positions
A som ewhat
greater or
somewhat
le sser degree
would be
accep table
Indications
of
inflationary
pressure
Strength of
the business
expansion
Behavior of
the monetary
a ggregates
Developm ents
in foreign
exchange and
dom estic
financial
m arkets
3/28/89
5/16/89
7/5 to
7/6/89
fO n February 23, the borrowing assum ption was increased to $800 million, but it was returned to $700 m illion on the next day when the discount rate was
raised.
^Borrowing assum ption chan ged for technical reasons.
§C hange in borrowing assum ption reflected a technical adjustm ent and a chan ge in reserve pressures.
FRBNY Quarterly Review/Spring 1990
57
period, when the spread between the funds and dis
count rates was 193 basis points, adjustment borrowing
averaged a skimpy $31 million. This level was the low
est since July 1980, when the funds rate was below the
discount rate. For the year, adjustment credit averaged
$243 m illion per day, w hile the fu n d s-d isco u n t rate
spread averaged 228 basis points (Chart 7). Compara
ble figures for 1988 and 1987 were $293 million per day
at an average spread of 137 basis points, and $286
million with an average spread of 100 basis points.
The rise and fall of seasonal borrowing more or less
followed its normal cyclical pattern over the year (Chart
8), although record high levels were attained during the
summer, somewhat earlier than in 1988. These move
m ents were accom m odated through eight te ch n ica l
adjustments to the borrowing allowance between May
and the year-end, two of which were accompanied by
policy-induced changes. With seasonal credit climbing
in the late spring and early summer, the assumed level
of borrowing was raised in the May 17 and July 12
maintenance periods. While the May move was purely
te chn ical, the July increase was associated w ith a
reduction of reserve pressures. This seem ingly contra
dictory step reflected the preceding surge in seasonal
borrowing, which necessitated an upward adjustment
to the assumed level in order to leave reserve pres
sures unchanged. Since only a portion of the technical
a djustm ent was o ffs e t by the FO M C ’s decision to
reduce reserve pressures, the assum ed borrow ing
level was higher following the easing move. After sea
sonal borrowing peaked in the July 26 m aintenance
period at an average $509 million per day, an all-time
high, it fluctuated in a range of $485 million to $500
million over the three succeeding periods. The peakperiod average in 1988 was $433 million (October 5
period), a previous record high. In 1987, when spreads
Table 1
S pecification s fo r D irectives of the Federal Open Market Com m ittee and Related inform ation
(Continued)
P rospective Reserve Restraint M odifications
D ate of
M eeting
S pecified Short-Term
Growth Rates
M2
M3
Borrowing
A ssum ption
for Deriving
NBR Path
D iscount
Rate
C om m ittee
Preference
G uidelines for
M odifying
Reserve
Pressure
Factors to C o nsider for M odifications
(In O rder Listed)
1
2
3
4
(Percent)
(M illions of
Dollars)
(Percent)
8/22/89
June to Septem ber
9
7
550
7.00
Sought to
m aintain the
existing deg ree
of pressure on
reserve
positions
A slig h tly
greater degree
m ight be
accep table. A
slightly lesser
degree would
be accep table
Progress
toward
price
s tability
S trength of
the business
expansion
Behavior of
the m onetary
aggregates
Developm ents
in foreign
exchange and
dom estic
financial
m arkets
10/3/89
September to December
6 V2
4 V2
550
500 on 10/5*
400 on 10/19§
350 on 11/2*
300 on 11/6
250 on 11/9*
7.00
Sought to
m aintain
existing degree
of pressure on
reserve
positions
A slightly
greater degree
m ight be
accep table. A
slig h tly lesser
degree would
be acce p ta b le
Progress
toward
price
stability
Strength of
the business
expansion
Behavior of
the m onetary
aggregates
Developm ents
in foreign
exchange and
dom estic
financial
m arkets
11/14/89
September to December
71/2
41/ z
250
200 on 11/15*
150 on 12/11*
7.00
Sought to
m aintain
existing degree
of pressure on
reserve
positions
A slightly
greater degree
m ight be
accep table. A
slig h tly lesser
degree would
be a ccep table
Progress
toward
price
stability
Strength of
the business
expansion
Behavior of
the monetary
aggregates
Developm ents
in foreign
exchange and
dom estic
financial
m arkets
12/18 to
12/19/89
Novem ber to M arch
8'/2
51/2
150
125 on 12/20
7.00
Sought to
decrease
slig h tly the
existing degree
of pressure on
reserve positions
A slig h tly
greater or
slightly lesser
degree would
be a ccep table
Progress
toward
price
s tability
Strength of
the business
expansion
Behavior of
the m onetary
aggregates
Developm ents
in foreign
e xcha nge and
dom estic
financial
m arkets
fO rt February 23, the borrowing assum ption was increased to $800 m illion, but it was returned to $700 m illion on the next day when the d iscount rate was
raised.
^B orrow ing assum ption ch an ged tor te chnical reasons.
§C hange in borrowing assum ption reflected a technical adjustm ent and a change in reserve pressures.
Digitized58
for FRASER
FRBNY Quarterly Review/Spring 1990
between the federal funds rate and the discount rate
were lower, the peak-period figure was $298 m illion
(July 1 period).20 As seasonal credit declined in the
early fall, downward technical adjustments were made
at the October 3 meeting and during the November 15
maintenance period (three times) and the December 13
maintenance period. In the November 1 period, the bor
rowing allowance was reduced both to lower reserve
pressures and to account for the decline in the use of
the seasonal borrowing privilege. For the year as a
whole, seasonal borrowing averaged $275 million per
day, compared with $235 million in 1988 and $164 m il
lion in 1987.
20Seasonal borrowing tends to increase as the federal funds-discount
rate spread rises, although traditionally it has not been as responsive
to spread changes as adjustment borrowing.
Open market operations and reserve management
In seeking to bring nonborrowed reserves into line with
the o bjective, the Desk takes account of both the
expected duration and day-to-day pattern of reserve
needs (or surpluses) in determining the timing and size
of its open market operations. Projected reserve sup
plies are com pared with the projected nonborrowed
reserve objectives for the current maintenance period
and a few subsequent periods. In choosing between
permanent and tem porary operations, the Desk con
siders w hether the projected need to add (or drain)
reserves is expected to persist for several consecutive
maintenance periods. If so, the Desk typically opts to
address a portion of the need (or surplus) with outright
purchases (or sales) of securities.
The D esk’s 1989 open m arket operations, in both
their nature and their timing, differed substantially from
Table 2
1989 Reserve Levels
(In Millions of Dollars, Not Seasonally Adjusted)
Period
Ended
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
11
25
8
22
8
22
5
19
3
17
31
14
28
12
26
9
23
6
20
4
18
1
15
29
13
27
Required
Reserves
(Current)
Required
Reserves
(First
Published)
64,256
61,786
60,035
59,278
59,490
59,299
58,977
61,190
60,345
58,357
56,877
59,012
58,154
60,067
58,807
58,766
58,859
58,247
60,195
58,343
60,186
58,827
60,139
59,958
61,149
62,015
64,397
61,735
60,138
59,269
59,533
59,305
58,924
61,107
60,339
58,382
56,923
59,187
58,069
60,060
58,883
58,659
58,737
58,153
60,000
58,117
60,110
58,857
60,279
60,073
61,253
62,019
Excess
Reserves
(Current)
Excess
Reserves
(First
Published)
1,147
972
1,543
1,016
957
735
1,305
223
1,241
859
1,158
897
901
990
1,035
715
951
959
888
996
926
1,128
881
1,009
759
1,018
991
1,070
1,504
1,036
915
805
1,550
289
1,301
960
1,139
817
976
953
915
812
1,104
1,051
1,079
1,160
1,045
1,166
763
868
666
1,022
Total
Reserves
Adjustment
and
Seasonal
Borrowed
Reserves
Nonborrowed
Reserves
plus
Extended
Credit
Borrowed
Reserves
(Current)
65,403
62,757
61,578
60,293
60,446
60,034
60,282
61,413
61,586
59,216
58,034
59,909
59,054
61,057
59,842
59,481
59,810
59,206
61,083
59,338
61,112
59,955
61,020
60,968
61,908
63,033
840
499
478
366
550
422
502
612
581
533
501
469
678
571
591
621
709
516
593
873
634
322
252
418
129
332
64,563
62,258
61,100
59,928
59,897
59,612
59,781
60,801
61,005
58,683
57,534
59,440
58.376
60,486
59,251
58,860
59,102
58,691
60,491
58,466
60,478
59,633
60,768
60,550
61,779
62,701
Nonborrowed
Reserves
plus
Extended
Credit
Borrowed Non borrowed
Reserves
Reserves
(First
Interim
Published)
O bjective!
64,548
62,307
61,162
59,939
59,898
59,689
59,973
60,785
61,059
58,809
57,563
59,537
58,366
60,442
59,206
58,851
59,132
58,689
60,487
58,404
60,521
59,701
60,790
60,523
61,789
62,708
64,793
62,116
60,743
59,464
59,774
59,754
59,376
61,549
60,742
58,677
57,269
59,670
58,548
60,409
59,232
59,058
59,137
58,725
60,400
58,518
60,560
59,447
61,029
60,823
62,024
62,708
Extended
Credit
Borrowed
Reserves
1,208
1,028
792
1,111
1,250
1,164
1,675
1,970
1,387
1,206
1,148
1,657
287
146
90
55
44
22
21
25
19
23
20
23
22
19
fA s of final Wednesday of reserve period.
FRBNY Q uarterly Review/Spring 1990
59
those of earlier years. Heavy purchases of foreign cur
rencies in foreign exchange markets by U.S. monetary
a u th o r itie s a d d e d c o n s id e r a b ly to n o n b o rro w e d
reserves. All intervention took the form of dollar sales
(that is, purchases of foreign currency) and totaled an
unprecedented $22 billion on behalf of both the Fed
eral Reserve and the Treasury. The intervention was
most heavily concentrated in the M ay-to-July period,
when these sales totaled $11.9 billion —the largest U.S.
in te rve n tio n fo r any th re e-m on th rep orting period.
Another $5.9 billion was sold in the August-to-October
interval.
The re s e rv e im p a c t of th e 1989 d o lla r s a le s
depended on how they were financed. In accord with
typical practice, official U.S. intervention generally was
shared equally by the U.S. Treasury, acting through the
Exchange S tabilization Fund (ESF), and the Federal
Reserve System. The Federal Reserve’s share of the
1989 intervention created reserves because the inter
vention took the form of foreign currency purchases,
paid for with reserve-creating dollars. In early 1989, as
in most other years, the ESF’s share of dollar sales had
no re se rve im p act. The U.S. T re a su ry o ffs e t the
reserve impact of the intervention by adjusting its bal
ance at the Federal Reserve; it called in funds from its
tax and loan accounts at d e p o s ito ry in s titu tio n s or
reduced the size of a d ire c t investm ent into those
accounts. By March, however, the ESF had exhausted
its supply of dollars to sell. Between mid-March and
late May, it raised dollars by selling International Mone
tary Fund Special Drawing Rights (SDRs) to the Fed
e ra l R e s e rv e . B e c a u s e th e p r o c e e d s o f th e
monetization were held in the ESF’s account at the Fed
until the funds were used by the ESF, the intervention
Chart 7
Borrowing and the Spread between Federal Funds and the Discount Rate
Millions of dollars
1600Borrowing
Adjustment and
seasonal borrowing
1400
1200
1000
'N ii
v
I I I 1 L.1J...I I I I I I I I I I I L-L L I
V v / \
Adjustment
borrowing
,
\ —
V
A —
W
a
- i- \ - / V
' J \
l i I i i i i i i i i i i i i i i i i iv i i I h i n
Percent
3.0
Federal Funds Rate minus Discount Rate
2.5
2.01.5
1.0
0.5
V
J
......................... .............. .............................. ..
1987
Digitized 60
for FRASER
FRBNY Quarterly Review/Spring 1990
I N
................ .. ............................I I I I I I
1988
1III
1 M il
1 II II 11 II 1 1 1 II
1989
11 1
financed by this method added reserves to the banking
system at the time that the intervention settled. From
mid-June to the end of the year, the Treasury funded its
intervention operations by w arehousing foreign cur
rency with the Federal Reserve. Under this technique,
the System buys foreign currency in a spot purchase
from the ESF and simultaneously agrees to sell it back
to the ESF at the same exchange rate at a future date.
(Such w arehousing o pe ra tio n s have been executed
from tim e to tim e since 1963.) A reserve in je ctio n
occurs at the tim e that the warehousing transaction
settles because the ESF invests the proceeds with the
Treasury, which in turn deposits them into its tax and
loan accounts at com m ercial banks or reduces the
a m o u n t it o th e r w is e w o u ld c a ll in fro m th e s e
accounts.21
The rise in the System ’s holdings of foreign currency
and SDRs provided a to ta l of about $23 billion of
reserves during 1989 (December over December). The
increase in the System ’s foreign currency assets added
$19.7 billion of reserves in 1989— compared with $2.1
billion in 1988 —while the ESF’s monetization of SDRs
more extensive discussion of Treasury tax and loan accounts
appears below.
added $3.5 billion of reserves. The System ’s share of
intervention operations accounted for about $11 billion
of the total increase in its foreign currency holdings,
w hile w arehousing of fo re ign cu rren cy fo r the ESF
totaled $7 billion. Of the remaining rise in the System ’s
fo re ig n c u rre n c y h o ld in g s , ro u g h ly $750 m illio n
stemmed from its portion of a swap arrangement with
the Bank of Mexico and about $1 billion from interest
earned on its foreign currency holdings.
The rese rve p rovision from foreign cu rren cy p u r
chases and monetization of SDRs more than met the
need for reserves for the year. The need to replenish
the supply of nonborrowed reserves prim arily arose
from the $12.3 billion increase in currency outstanding.
(This increase was only about three-quarters of the
1988 rise.) Reserves were also drained by the $1.2 bil
lion decline in extended credit borrowing (ECB).22 The
major user of the program was taken over by the Fed
eral D eposit Insurance C orporation (FDIC), and the
FDIC paid off the user’s borrowing in mid-June. On net,
other operating factors added a m odest amount of
reserves. Meanwhile, required reserves showed their
first decline since 1983, and excess reserves dropped
modestly. Because the supply of nonborrowed reserves
from operating factors (including foreign currency hold
ings and ECB) exceeded demand, the size of the Sys
tem ’s portfolio was reduced over 1989 for the first time
since 1957. The $10.2 billion decline in the portfolio left
its year-end level at $235.6 billion 23
The reduction of the System ’s portfolio in 1989 was
accomplished through redemptions of Treasury bills at
auctions and through sales of Treasury securities in the
market and to foreign customer accounts. Typically, the
Desk exchanges its m aturing h oldings of Treasury
securities for new securities at auction time. However,
the Desk may choose to roll over only a portion of its
holdings, as it did frequently in 1989, and thus drain
reserves. The Desk redeemed a total of $13.2 billion of
Treasury securities in 1989. (The figure includes a $3.5
billion forced redem ption on November 2, discussed
below.) The total includes $500 m illion of Treasury
notes redeemed in late September — only the second
time that the System has chosen to redeem coupon
22ECB is viewed by the Desk as nonborrowed reserves because
institutions using the ECB program cannot easily replace funds
obtained through the ECB facility with other types of funding; these
institutions are under pressure to achieve improvements in their
troubled funding situations.
23The total reflects the commitment to purchase $200 million of
Treasury securities from customer accounts made on the last
business day of 1989, for delivery on January 2, 1990. It excludes
the temporary reduction of the portfolio from that day’s matched salepurchase transaction with foreign accounts; the sale included a
commitment to repurchase the securities on January 2.
FRBNY Quarterly Review/Spring 1990
61
issues .24 The redemptions were heaviest in the May-toJuly period, reflecting the need to offset foreign
exchange intervention. This intervention also prompted
the Desk to sell a record volume of bills in the market
on July 12, an unusual action for that time of year. The
$4.6 billion sale was the Desk’s largest outright sale,
exceeding the previous record by $1.5 billion. The
Desk also sold Treasury bills in February, when the
seasonal drop in currency and in required reserves
produced a sizable need to drain reserves. Finally, the
Desk drained $1.3 billion of reserves in 1989 through
net sales of Treasury securities to foreign customer
accounts. In 1988, it had made net purchases from
these accounts that added $4.3 billion of reserves.
Nevertheless, the Desk at times arranged outright
purchases of securities to address seasonal reserve
needs, such as those that arose around tax dates and
around year-end. The Desk favored Treasury bill pur
chases on these occasions to offset part of the decline
in its bill holdings from redemptions and sales. The
Desk purchased both coupon issues and bills in April
and bills on two occasions in November. The April pur
chases were smaller than those of 1988 because for
eign exchange intervention reduced projected reserve
needs below the norm for late May. The Desk’s pur
chase of bills in early November was prompted by its
forced redemption of bills at the October 30 auction.
The Treasury announced a settlement date for that
auction of Tuesday, October 31, rather than Thursday,
November 2, when the outstanding bills were to
mature, because the debt ceiling was scheduled to
drop on November 1. The Desk is not permitted to buy
s e c u ritie s d ire c tly from the T re asu ry e xce p t in
exchange for maturing issues. Consequently, the timing
disparity forced the desk to redeem its $3.5 billion of
maturing bills.
The net shrinkage in the System portfolio occurred in
its bill holdings, which fell by $11.3 billion, in contrast
with a rise of $5.4 billion in 1988. The Desk increased
the System’s holdings of coupon issues by $1.3 billion
in 1989, compared with $9.7 billion in 1988. As a result,
the average maturity of the System portfolio length
ened a bit in 1989. Redemptions reduced System hold
ings of federally sponsored agency securities by about
$440 million, a decrease just slightly less than in the
previous year.25
The Desk also met reserve needs through temporary
“ The Desk redeemed a very modest amount of coupon issues in 1987
because it purchased some maturing notes from foreign accounts
between the time of the auction for the replacement issue and the
settlement day for that auction.
“ The Desk normally rolls over maturing federally sponsored agency
issues. Its holdings decline when issues are called or when issues
mature and no eligible replacement is available.
Digitized62
for FRASER
FRBNY Quarterly Review/Spring 1990
transactions. When determining the timing of these
operations, it took into account the intraperiod distribu
tion of reserve needs (surpluses). The Desk sought to
avoid extraordinary reserve deficiencies or surfeits on
individual days because both could induce movements
in the federal funds rate that might give misleading sig
nals about the intent of policy. Moreover, a sizable daily
reserve deficiency might leave the banking system with
inadequate reserves for the purpose of clearing trans
actions, lead to extraordinary pressures in the reserve
market, and force spikes in discount window borrowing
that could preclude achieving the path level.
The Desk arranged about the same volume of tem
porary transactions in the market in 1989 as in 1988.
Because of reserve injections associated with foreign
exchange intervention, the Desk made much greater
use of temporary transactions to withdraw reserves in
1989. The volume of matched sale-purchase transac
tions represented just over one-third of total temporary
market transactions in 1989, in contrast to the smaller
shares of previous years. The Desk arranged 69
rounds of matched sale-purchase agreements in the
market for a total of $151 billion, compared with the 22
rounds for $63 billion that it had executed in 1988.
Nearly two-thirds of these reserve draining operations
spanned more than one business day.
A smaller volume of repurchase agreements (RPs)
was executed in 1989 than in previous years because
of the substantial reserve injections associated with
foreign currency intervention. Over the year, the Desk
arranged 28 rounds of System RP transactions for a
total of $168 billion, and 61 rounds of customer-related
RPs for a total of $108 billion. Comparable figures for
1988 were 51 rounds of System RPs for $210 billion,
and 85 rounds of custom er RPs for $143 billion.
Although the Desk conducted fewer rounds of System
RPs, the average daily volume of those RPs was $7.7
billion, or $3.7 billion greater than in 1988. The higher
average volume stemmed partly from the decision to
undertake a smaller volume of outright purchases of
securities to meet the reserve needs arising around
the April tax date. The Desk met these needs primarily
through temporary operations rather than through its
usual outright operations since the reserve shortages
were not expected to extend over several periods (and
since actual reserve needs exceeded projections). In
the May 3 and 17 maintenance periods, the Desk pre
announced term System RPs on three occasions to
ensure adequate propositions. On May 4, the Desk
arranged a record $15.8 billion of System RPs to meet
part of the reserve needs.
The Desk frequently conducted temporary operations
in response to large day-to-day variations in reserve
availability. It also recognized that short-term transac
tions might at times help provide clearer policy guid
ance to financial market participants. Market partici
pants often judged whether the policy stance had
changed by observing the Desk’s use or eschewal of
short-term transactions. However, they did not always
interpret Desk actions correctly.
A technical reserve injection on the day before
Thanksgiving was m isinterpreted by market p a rtici
pants, and subsequent efforts by the Desk to correct
the m isimpression caused heavy borrowing in the
November 29 maintenance period. On November 22,
the Desk faced a fair-sized need to add reserves for
the maintenance period then in progress, and large
daily reserve deficiencies were projected for that day
and for the remaining days of the period. During most
of the morning of November 22, federal funds were
trading at 87/ie percent, just slightly below the 8V2 per
cent rate that participants perceived to be consistent
with the FOMC’s desired degree of reserve restraint. It
was anticipated that many market participants would
be on vacation on Friday, the day after Thanksgiving,
making for relatively inactive securities trading and
financing activity. In these circumstances, the Desk was
concerned that a delay in addressing the estimated
reserve shortage could leave very large reserve needs
toward the end of the period that might be difficult to
meet. Hence, it decided to arrange five-day System
RPs to meet the projected reserve shortage. Shortly
before the Desk’s regular time to enter the market, the
funds rate slipped to 8% percent. Nonetheless, the
Desk felt that its absence that day could lead to strains
in the reserve market. When the Desk announced its
operation, some market participants thought the action
might be signaling a move to ease policy.
On the Friday after the holiday, these misimpressions
were reinforced by an erroneous newspaper article that
cited “ government officials” as confirming an easing
step. The Desk attempted to dispel these notions by
temporarily draining reserves from the banking system
that morning even though a reserve need remained.
Federal funds were trading at 81/4 percent during most
of the morning; however, the funds rate dipped to 83/ie
percent just before the Desk acted. In that circum
stance, many observers interpreted the operation as
signaling the extent of the downward adjustment to the
funds rate and as indicating the Committee’s support
for an 8V4 percent funds rate. The funds rate retained a
soft tone over the afternoon (although it firmed a bit at
the close), and the reserve data released that after
noon were not interpreted by participants as showing
an insurmountable reserve need. The misperception
persisted into the following Monday morning, Novem
ber 27. After discussion at an FOMC conference call
on Monday morning, the Desk entered the market
before its customary time and drained reserves even
though a large deficiency was anticipated. The drain
corrected the m arket’s misimpression of the policy
stance but left very large reserve needs, which were
met with heavy borrowing that evening and with large
RP operations over the next two days.
The miscommunication resulted from a confluence of
factors. The FOMC’s previous decision to reduce
reserve pressures, made in early November, had come
as a surprise to market participants, who had not been
expecting such a move until later in the month or at the
tim e of the C o m m itte e ’s D ecem ber m eeting. On
November 22 there was some speculation that another
step might be in the offing, but discussions between
Desk personnel and market participants did not indi
cate a widespread expectation of an imminent easing,
even though the durable goods report released that
morning had been weaker than anticipated. Moreover,
analysts generally viewed the reserve need as being
smaller than suggested by the Desk’s projections, so
they did not anticipate that a System operation would
be necessary. Finally, the newspaper article seemed to
confirm the view, which had previously been just a sus
picion, that an easing had occurred.
Forecasting reserves and operating factors
When the Desk form ulates a strategy for meeting
reserve needs, it takes account of potential revisions to
the estimated demand for and supply of reserves. On
the demand side, these revisions can take the form of
changes in estimated required reserve levels or in the
banking system’s desired excess reserve balances. On
the supply side, revisions to estimated operating fac
tors, or sources and uses of nonborrowed reserves
other than open market operations, can change the
reserve outlook. In both cases, revisions late in the
maintenance period are especially difficult to deal with
sin ce th e y may n e c e s s ita te ve ry la rg e re se rve
operations.
The accuracy of required reserve forecasts was
about unchanged in 1989 relative to the previous year.
The mean absolute error in forecasting required
reserves on the first day of the period was around $325
million in 1989, compared with about $300 million in
1988.26 This steady forecasting performance came
despite an increase of $125 million in the mean abso
lute period-to-period change in required reserves.
Forecasts became more accurate as the maintenance
period progressed; the mean absolute prediction error
“ The Trading Desk uses forecasts of required reserves, excess
reserves, and operating factors made by both the Federal Reserve
Bank of New York and Board staffs. When a range of forecast errors
is given in the following discussion, it reflects the two staffs’ varying
degrees of success in forecasting reserve measures.
FRBNY Quarterly Review/Spring 1990
63
fell to roughly $200 million at midperiod and to about
$70 million to $90 million on the final day. These errors
are a bit larger than their 1988 counterparts, in addi
tion, some sizable revisions took place after the main
tenance period ended, especially late in the year.
Excess reserves were somewhat more predictable in
1989 than in 1988. The beginning-of-period mean abso
lute forecast errors were about $135 million to $150
million, compared with $160 million in 1988.27 The
mean absolute p eriod -to -p e rio d change in excess
reserves was about the same as in 1988. The largest
forecast errors occurred in the Aprij 19 maintenance
period, when excess reserves averaged $223 million,
the lowest level since contemporaneous reserve ac
counting was introduced in February 1984.
The average level of excess reserves held by the
banking system shrank to $970 million in 1989 from
just over $1 billion in 1988. Excess reserves had risen
each year from 1979 through 1987, and then had stabi
lized in 1988. Provisions of the Monetary Control Act of
1980 that were phased in between 1980 and 1987
expanded the number of institutions subject to reserve
requirements and resulted in increased excess reserve
holdings. In addition, rising Fedwire activity increased
the need fo r re s e rv e b a la n ce s at the F ederal
Reserve .28 Since large banks tend to monitor their
reserve balances closely to avoid holding non-interestbearing excess reserves, their average holdings of
excess reserves over a year are typically close to zero.
These banks generally make use of the carryover privi
lege, under which banks can apply a portion of the
excess reserves held in one period to their require
ments in the following period. Carryovers tend to pro
duce a sawtooth pattern of excess reserve holdings at
large banks, and during 1989 this pattern at times
showed through to aggregate excess reserve holdings.
Smaller banks, however, generally lack the resources
to monitor their reserve positions accurately, and they
tend to hold positive levels of excess reserves.
Despite a marked jump in the variability of operating
factors from period to period in 1989, the accuracy of
operating factor forecasts was about the same as in
1988. The mean absolute error of first-day forecasts
was about $900 million to $1.1 billion in 1989, com
pared with $900 million to $1 billion in the previous
year. Although projections of reserves supplied by
operating factors improved as the period progressed,
the average absolute errors increased relative to their
1988 levels. The mean absolute forecast error around
midperiod was about $450 million, and that for the final
day of the period was roughly $70 million to $90 mil
lion. In 1988, these errors were $325 million to $470
million and about $50 million, respectively.
The 1989 forecasting performance looks better when
compared with the mean absolute period-to-period
change in operating factors. The mean absolute
change surged to $3.4 billion per period, up sharply
from $2.0 billion in the previous year. As a proportion
of the average absolute change, mean absolute errors
in forecasting operating factors on the first day of the
period were only about half as much as their 1988
counterparts.
Much of the increase in the average period-to-period
change of operating factors reflected the behavior of
the Treasury’s balance at the Federal Reserve. The
Treasury tries to maintain a $5 billion balance in this
account .29 Additional funds are held in Treasury tax
and loan (TT&L) accounts at participating depository
institutions .30 If the Treasury anticipates that its bal
ance will fall below the $5 billion target level, it may
“ call” funds from its TT&L accounts to bring its bal
ance up to the target level. Similarly, if the Treasury’s
balance at the Federal Reserve is expected to exceed
$5 billion, the Treasury can directly place funds into
these TT&L accounts. However, since depository insti
tutions must fully collateralize and pay interest on
TT&L funds, the institutions set lim its on the total
amount of funds they will accept based on their ability
to make profitable use of these funds and on the avail
ability of collateral. Treasury funds in excess of TT&L
capacity must be held in the Treasury’s Federal
Reserve balance. Typically, around major tax dates, the
Treasury’s cash holdings substantia lly exceed the
capacity of the TT&L accounts. In 1989, capacity limita
tions forced the Treasury’s Fed balance to exceed its
target level on about fifty-five business days, compared
with about forty days in 1988.
The mean absolute period-to-period change in the
Treasury’s balance rose to $2.8 billion in 1989 from
$1.5 billion in 1988. The increased variability of the bal
ance stemmed in part from an increase in tax receipts
in 1989 relative to 1988, while the aggregate capacity
of TT&L accounts remained about unchanged. With
^These reported forecast errors overstate the degree of uncertainty
about excess reserves. The Desk supplements beginning-of-period
and midperiod forecasts with informal adjustments that are based on
the observed pattern of estimated excess reserve holdings as each
maintenance period unfolds.
M ln late 1988, the Treasury raised this target level to $5 billion from $3
billion in order to reduce the likelihood of an inadvertent overdraft.
“ See discussion in 1988 report, p. 101. In 1989, the turnover rate of
reserve accounts resumed its upward movement after having stalled
in the previous year.
“ Individual nonwithheld taxes are paid directly into the Treasury's
account at the Federal Reserve. Most other Treasury tax receipts are
initially deposited into TT&L accounts.
Digitized 64
for FRASER
FRBNY Quarterly Review/Spring 1990
TT&L capacity at roughly $30 billion, the substantially
higher volume of tax payments received by the Treas
ury in 1989, especially in April and June, caused its
balance at the Federal Reserve to surge to levels sig
nificantly above those in 1988. For example, the Treas
ury’s balance at the Fed averaged $15.1 billion and
$19.7 billion in the May 3 and 17 maintenance periods,
respectively, but averaged only $9.2 billion and $9.6
billion for the corresponding periods in 1988.31 The
buildup and reduction of the Treasury’s balance pro
duced large changes from one period to the next,
resulting in the 1989 rise of the absolute period-toperiod change in this balance.
The forecast errors for the Treasury balance were a
bit larger than in 1988. The mean absolute errors of the
first-day forecasts were about $725 million to $800 mil
lion in 1989, compared with $700 million to $750 million
in 1988. These errors were elevated somewhat by large
forecast errors in the October 4 period. During this
period, RTC payments fell well short of expectations.
On September 29, the Treasury’s Fed balance ex
ceeded expectations by about $61/2 billion to $7V2 bil
lion and thus contributed to a large forecast miss for
the period-average Treasury balance.
Initial forecasts of U.S. currency, the foreign RP pool,
^ In 1989, the Treasury’s Fed balance averaged $14.9 billion per day
on those days when TT&L accounts were at capacity, compared with
$10.7 billion in 1988.
float, and foreign currency were subject to sizable revi
sions as the maintenance period progressed. U.S. cur
rency was difficult to predict in 1989, in part because it
grew considerably more slowly than expected during
most of the year but then experienced a year-end rise
that was somewhat larger than usual. The beginningof-period mean absolute forecasting errors were about
$350 million to $400 million, somewhat above their
1988 levels. Forecasting the foreign RP pool on a twoweek average basis was also harder since the level of
the pool was also somewhat more variable in 1989 than
in 1988; the first-day average absolute forecast error
was about $275 million. First-day forecasts of Federal
Reserve float, including the so-called as-of adjust
ments that correct various reserve transfer errors, had
mean absolute errors of about $200 million to $225 mil
lion. Forecasts of foreign currency had a beginning-ofperiod mean absolute error of about $200 million; how
ever, this error overstates the uncertainty the Desk
faced. The reserve effect of foreign currency interven
tion occurs two days after the transaction. The Desk
was informed about the size of the intervention on the
day before the transaction settled, so that it knew one
day in advance what the reserve impact would be.
Because the Desk was also informed about warehous
ing transactions before they occurred, the deterioration
in forecast accuracy did not pose significant day-to-day
difficulties in implementing policy.
FRBNY Quarterly Review/Spring 1990
65
Treasury and Federal Reserve
Foreign Exchange Operations
February-April 1990
The dollar gained support during the February-April
reporting period as the prospects for U.S. growth came
to be viewed as somewhat better and earlier expecta
tions that monetary policy would ease were replaced
by consideration of a possible tightening. At the same
time, factors abroad strongly influenced individual
exchange rates. Accordingly, the dollar moved up the
most against the Japanese yen, which was affected by
political uncertainties and weakness in Japan’s stock
market. The dollar also stopped its decline against the
German mark amid concern about the potential infla
tionary im plications for Germany of econom ic and
monetary union between East and West.
During the three-month period, the dollar declined
less than V2 percent against the mark and more against
some other European currencies, ending the period
about 18 percent below its mid-June 1989 highs against
the mark. In contrast, the dollar rose 10 percent
against the yen to trade nearly 5 percent above its midJune 1989 levels. On a trade-weighted basis as meas
ured by the staff of the Federal Reserve Board of Gov
ernors, the dollar rose more than 1 percent to end the
period 121A> percent below its highs of mid-June 1989.
Intervention operations carried out by the U.S. mone
tary authorities during the period were aimed primarily
at moderating the rise of the dollar against the yen.
During the three-month period, the Desk sold a total of
$1,780 million, of which $1,580 million was sold against
yen and the remainder against marks. Of these totals,
A report presented by Sam Y. Cross, Executive Vice President in charge of
the Foreign Group at the Federal Reserve Bank of New York and Manager
of Foreign Operations for the System Open Market Account. Christopher B.
Steward was primarily responsible for preparation of the report.
Digitized 66
for FRASER
FRBNY Quarterly Review/Spring 1990
$375 m illion of the dollar sales against yen was
financed by the Federal Reserve System. The remain
ing $1,205 million sold against yen, together with the
entire $200 million sold against marks, was financed by
the U.S. Treasury through the Exchange Stabilization
Fund (ESF).
February through the beginning of April
Reports indicating that U.S. economic growth was
quickening helped improve sentiment toward the dollar
as the period opened, tending to dispel the prevailing
impression that U.S. m onetary policy was likely to
ease. In the firs t w eeks of February, data were
released showing strong increases in employment, a
strengthening of retail sales, and a sharp rise in pro
ducer prices for January. About the same time, com
ments from some Federal Reserve officials suggested
that the likelihood of a recession in the United States
had diminished. Subsequent data and testimony by
U.S. monetary officials reinforced the view that the bal
ance of policy concerns had shifted from supporting
growth to restraining inflation.
This shift in assessments of economic and policy
conditions in the United States provided a foundation
for a firmer tone to the dollar for much of the period
under review. W ithin that framework, the extent to
which the dollar advanced against another country’s
currency depended very much on developments in that
country.
Thus, as the period began, the European currencies
continued to gain support from the enthusiasm engen
dered late last year by reforms in Eastern Europe. With
institutional investors, especially Japanese investors,
re p o rte d ly expanding th e ir p o rtfo lio investm ents in
European currencies, the dollar extended the decline
against the mark that had begun when the Berlin Wall
was opened in O ctober. On February 7 the d ollar
reached DM 1.6490 — its lowest level against the mark
in almost two years.
Then on February 7, the West German government
announced plans fo r im m ediate talks on m onetary
union between East and West Germany, and the dollar
began to firm against the mark as attention quickly
focused on the possible inflationary consequences of
such a move. Market participants feared that the con
version of East Germ an m arks into W est Germ an
marks would result in a worrisome increase in German
monetary aggregates or unleash pent-up demand for
German products. They noted that the new source of
demand would m aterialize at the same time that West
German residents would be feeling the effects of a tax
cut. A new round of wage negotiations by the country’s
largest trade union was an additional source of con
cern. As inflation anxieties mounted, German bond
prices softened, and yields hit levels above comparable
Chart 1
The dollar again came under strong upward pressure against the Japanese yen during the period . . .
Percentage points
20
May
Jun
Jul
Aug
1989
Sep
Oct
Nov
Dec
Jan
while remaining well below its year-earlier highs against most other currencies.
Percentage points
15
■20
F M A M J J A S O N D J
1987
F M A M J
J A S O N D J
1988
F M A M J J
1989
A S O N D J
F M A
1990
The top chart shows the percent change of weekly average rates for the dollar from May 1989. The bottom chart shows the percent change of
weekly average rates for the dollar from February 23, 1987, the date of the Louvre Accord. All rates are calculated from New York
closing quotations.
FRBNY Quarterly Review/Spring 1990
67
U.S. bond yields for the first time in ten years.
Against the yen, the dollar was trading in a fairly nar
row range when the period opened, as market partici
pants awaited the results of parliam entary elections in
Japan on February 18 to see whether the ruling Liberal
Democratic party (LDP) could maintain control in the
lower house. On news that the LDP had won a larger
than expected majority, the dollar declined to touch its
period low against the yen of ¥143.60 on February 19.
But soon thereafter, the dollar began to firm as market
p a rtic ip a n ts q u e stio n e d w h e th e r the L D P ’s stron g
showing would be sufficient to ward off pressures from
the opposition-dominated upper house and address in
a meaningful way the challenging economic and trade
issues confronting Japan. Before the elections, Japa
nese short- and long-term money m arket yields had
been rising slowly in the expectation that an increase
in th e B ank of J a p a n ’s d is c o u n t ra te w o u ld be
announced soon afterward. In the event, concerns over
escalating interest rates abroad and accelerating mon
etary expansion at home combined to trigger a sharp
drop in Japanese stock and bond prices in the weeks
Chart 2
The Japanese stock market came under intense
downward pressure during the period, stabilizing
in mid-April.
Percentage points
following the elections. The Nikkei Dow index of the
Tokyo stock market fell 10 percent and Japanese gov
ernm ent bond yields clim bed well above 7 percent.
Market participants came to believe that the Bank of
Japan would be reluctant to tighten monetary policy in
these circu m stan ces, but they rem ained co nce rn ed
whether adequate measures would be taken to break
the rapidly deteriorating sentiment surrounding Japan’s
financial markets.
Against this background, upward pressure on the dol
lar against the yen intensified. On February 23, the
U.S. m onetary authorities, in keeping with Group of
Seven (G-7) understandings on exchange rate cooper
ation, intervened to resist the dollar’s rise against the
Japanese currency. As upward pressure on the dollar
against the yen continued, the U.S. authorities inter
vened in moderate amounts through March 2, selling a
total of $650 million against yen. Financing of these
operations was shared equally by the Federal Reserve
and the U.S. Treasury.
At that time, the dollar also came under more intense
upward pressure against the mark as fears of German
Chart 3
Long-term interest rates rose nearly
1 percentage point in Germany and Japan
and about 1/2 of 1 percentage point in
the United States.
Percentage points
Oct
The chart shows the percentage changes for the weekly average
stock indexes from September 31, 1989. All figures are based
on closing rates.
Digitized for
68 FRASER
FRBNY Quarterly Review/Spring 1990
Nov
1989
Dec
Jan
Feb
Mar
Apr
1990
The chart shows daily U.S., German, and Japanese government
long-bond yields.
in fla tio n in te n s ifie d . As th e d o lla r a d v a n c e d to
DM 1.7237 on March 1, its highest level against that
currency since January, widespread reports circulated
through the m arkets that the Bundesbank and other
European central banks were intervening in the foreign
exchange markets to support the mark and resist the
dollar’s rise.
On March 5, with the dollar rising against both cur
rencies, U.S. Treasury officials decided to reinforce
their intervention in yen with sales of dollars against
marks. Prior to those mark purchases, several officials
within the Federal Reserve had expressed concern that
the size of the intervention operations might contribute
to u n c e rta in ty about the Federal R eserve’s p rio rity
toward achieving price stability. Concerns were also
voiced that expanding the operations to include other
currencies might be seen as an attempt to promote a
broad-based decline in dollar exchange rates. At the
time, Federal Reserve holdings of foreign currencies,
taking into account anticipated fu rthe r interest e arn
ings, were approaching the limit of $21 billion autho
rized by the Federal Open Market Committee (FOMC).
Under these circumstances, the decision was made not
to seek authorization from the FOMC for continued
Table 1
Federal Reserve
Reciprocal Currency Arrangem ents
In Millions of Dollars
Amount of Facility
Institution
April 30, 1990
Austrian National Bank
National Bank of Belgium
Bank of Canada
National Bank of Denmark
Bank of England
Bank of France
Deutsche Bundesbank
Bank of Italy
Bank of Japan
Bank of Mexico
Netherlands Bank
Bank of Norway
Bank of Sweden
Swiss National Bank
Bank for International Settlements:
Dollars against Swiss francs
Dollars against other
authorized European currencies
250
1,000
2,000
250
3,000
2,000
6,000
3,000
5,000
700
500
250
300
4,000
600
1,250
Total
30,100
Chart 4
Data released during the period were seen as indications of stronger growth and price pressures in
the U.S. economy.
Percent
16
Consumer Price Index
Index 1987=100
109
Industrial Production
14
108 -
12
10
107 -
8
6
106 -
-
4 -
2
May Jun
A V \V
Jul
V\
Aug Sep Oct
1989
m
A
Nov Dec
Jan
*
A
Feb Mar
1990
2
-
May Jun
Jul
Aug Sep
1989
Oct
Nov Dec Jan
Feb Mar
1990
The left chart shows monthly industrial production figures for January, February, and March, seasonally adjusted. The right chart shows the
monthly rise in the seasonally adjusted U.S. consumer price index, annualized. Consumer price data for January, February, and March were
released on February 21, March 20, and April 17, respectively. Industrial production figures for January, February, and March were released
on February 16, March 16, and April 17, respectively.
FRBNY Q uarterly Review/Spring 1990
69
Federal Reserve operations pending a review of Fed
e ra l R e s e rv e fo re ig n c u rre n c y o p e ra tio n s at the
F O M C ’s M arch 27 m e e tin g . T hus, from M arch 5
through March 27, all U.S. in te rve n tio n o perations,
to ta lin g $830 m illion a ga in st yen and $200 m illion
against marks, were financed solely by the U.S. Treas
ury through the ESF.
At the M arch 27 m e e tin g , th e FOMC v o te d to
approve an increase in the authorized limit on Federal
Reserve holdings of foreign currencies from $21 billion
to $25 b illion . W hen the d o lla r again cam e under
upward pressure against the yen on March 28, the
Table 2
Drawings and Repayments by Foreign Central Banks under Reciprocal Currency Arrangem ents w ith
the Federal Reserve System
tn Millions of Dollars; Drawings ( + ) or Repayments ( - )
Central Bank Drawing on the
Federal Reserve System
Bank of Mexicot
Bank of M exicof
Amount
of Facility
Outstanding as of
January 31, 1990
February
March
April
700.0
700.0
700.0
—
-7 0 0 .0
—
+ 700.0
-1 5 8 .2
Outstanding as of
April 30, 1990
0
541.8
_
Data are on a value-date basis.
fDrawn as a part of the $2,000 million near-term credit facility established on September 14, 1989. Facility expired on February 15, 1990.
^Represents the FOMC portion of a $1,300 million short-term credit facility established on March 23, 1990.
Table 3
Drawings and Repayments by Foreign Central Banks under Special Swap Arrangem ents w ith the
Federal Reserve System
In Millions of Dollars; Drawings ( + ) or Repayments ( - )
Central Bank Drawing on the
Federal Reserve System
Bank of Mexicot
Amount
of Facility
Outstanding as of
January 31, 1990
February
125.0
34.1
-34 .1
'
March
April
Outstanding as of
April 30, 1990
-
-
-
Data are on a value-date basis.
fDrawn as a part of the $2,000 million near-term credit facility established on September 14, 1989. Facility expired on February 15, 1990.
Table 4
Drawings and Repayments by Foreign Central Banks under Special Swap Arrangem ents w ith
the U.S. Treasury
In Millions of Dollars; Drawings ( + ) or Repayments ( - )
Central Bank Drawing on the
U.S. Treasury
Bank of Mexicof
Bank of Mexicof
Central Bank of Venezuela§
National Bank of Poland||
Amount
of Facility
Outstanding as of
January 31, 1990
February
425.0
600.0
104.0
200.0
334.1
-334.1
—
—
—
_
86.0
-8 6 .0
March
_
+ 600.0
+ 25.0
0.0
Data are on a value-date basis.
fRepresents the ESF portion of a $2,000 million near-term credit facility that expired on February 15, 1990.
^Represents the ESF portion of a $1,300 million short-term credit facility established on March 23, 1990.
§Represents the ESF portion of a $400 million near-term support facility that expired on April 30, 1990.
j|Represents the ESF portion of a $500 million short-term credit facility established on December 27, 1989.
The ESF facility will expire on May 31, 1990.
70FRASER
FRBNY Quarterly Review/Spring 1990
Digitized for
April
_
-1 3 5 .6
-2 5 .0
0.0
Outstanding as of
April 30, 1990
_
464.4
—
0.0
Desk in te rv e n e d to se ll $50 m illio n a g a in s t yen,
financed once more by the Federal Reserve and the
U.S. Treasury equally.
As April began, sentim ent toward the yen was still
n e g a tiv e and J a p a n e s e s to c k and b o n d p ric e s
remained under downward pressure amid market con
cerns about political leadership and economic policy in
Japan. A March 20 increase of 1 percentage point in
the Bank of Japan’s discount rate had not immediately
relieved the pressures against the yen. An important
round of negotiations under the Structural Impediments
Initiative was getting under way between the United
States and Japan in an atmosphere of growing U.S.
concern and considerable u ncertainty in Japan over
their possible outcome. On the first trading day of the
new Japanese fiscal year, the Nikkei Dow index closed
nearly 25 percent below — and yields on Japanese gov
ernm ent bonds about 75 basis points above — levels
prevailing at the beginning of the period. Against that
background, the dollar reached ¥160.35 on April 2, the
highest level for the dollar against that currency since
December 1986.
Up to that point, most market participants had not
expected the upcoming G-7 meeting to focus strongly
on cu rren cy issues. But as developm ents in Japan
unfolded, reports began to circulate that the G-7 might
implement a massive yen support package. Market par
ticipants therefore adopted a more cautious attitude as
the date of the April 7 meeting approached, and the
dollar began to ease somewhat against the yen.
A pril developments follow ing the Paris G-7 meeting
The statement released following the G-7 meeting on
April 7 reported that the G-7 “ Ministers and Governors
discussed developm ents in global financial markets,
especially the decline of the yen against other curren
cies, and its undesirable consequences for the global
adjustment process a nd ...reaffirm ed their commitment
to econom ic policy coordination, including cooperation
in the exchange markets.”
On the Monday following the G-7 meeting, the dollar
moved erratically as market participants assessed the
implications of the statement. In Far Eastern trading
before the opening of the Tokyo market, the dollar rose
sharply against the yen as market participants inter
preted the statement as indicating a lack of G-7 com
m itm ent to aggressive intervention in support of the
yen. In Tokyo, how ever, m a rke t re p o rts p re d ic tin g
stepped-up concerted intervention by the G-7 coun
tries brought the dollar under strong downward pres
sure, and the dollar moved down from ¥160 to ¥155.45
within a few hours. During the day, the dollar began to
rise once again on reports of coordinated, but not
large-scale, intervention. As the dollar continued to rise
against the yen in New York, the U.S. monetary author
ities, in their only intervention in the month of April,
sold $50 million against yen, shared equally between
the Federal Reserve and the U.S. Treasury.
In the weeks that follow ed, a num ber of develop
ments combined to lessen negative sentim ent towards
the yen. First, market participants came to view the
Bank of Japan’s March discount rate increase as suffi
cient, at least for the time being, and expectations of a
further rise in interest rates dissipated. Furthermore, in
retrospect, market participants came to see the variety
of international negotiations leading up to the Struc
tural Im pediments Initiative talks and the G-7 meeting
as being rather successful and dim inishing the strains
within the Japanese leadership. Finally, Japanese stock
and bond markets became more settled in early April,
recouping some of the losses sustained earlier in the
period.
In Europe, the focus of market attention during April
remained on Germany and German monetary union.
Market participants began to feel that the inflationary
risks of m onetary union had been exaggerated. As
confidence grew that Bundesbank policy would be suf
fic ie n tly re s tric tiv e to co nta in pressures th at m ight
em erge, se ntim en t tow ard the m ark becam e more
favorable.
U nder th ese c irc u m s ta n c e s , the d o lla r s to p p e d
advancing against the mark and the yen during April.
But at the same tim e that sentim ent towards these
other currencies was improving, the dollar continued to
draw s u p p o rt from U.S. d ata re le a se s su g g e s tin g
strong economic activity and from the belief that U.S.
monetary policy would remain directed at dealing with
price pressures. The dollar therefore closed the period
near its three-month high against the yen at ¥158.90
Table 5
Net Profits ( + ) o r Losses ( - ) on
United States Treasury and Federal Reserve
Foreign Exchange Operations
In Millions of Dollars
February 1, 1990 to
April 30, 1990
Realized
Valuation profits and losses on
outstanding assets and
liabilities as of
April 30, 1990
Federal
Reserve
U.S. Treasury
Exchange
Stabilization
Fund
0
+ 292.4
+ 1,996.9
+ 881.8
Data are on a value-date basis.
FRBNY Quarterly Review/Spring 1990
71
and around the midpoint of its three-month trading
range against the mark at DM 1.6790.
For the period as a whole, the U.S. monetary authori
ties sold a total of $1,780 m illio n — $1,580 million
against Japanese yen and $200 million against Ger
man marks. The U.S. Treasury, through the ESF, sold
$1,205 million against yen and $200 million against
marks, while the Federal Reserve sold $375 million
against yen. During March, the ESF “ w arehoused"
$ 2,000 million equivalent of foreign currencies with the
Federal Reserve, bringing the total of warehoused
funds to $9,000 million equivalent. The warehousing
transactions resulted in realized profits of $292.4 mil
lion for the ESF, reflecting the difference between the
rate at which the warehoused funds had originally been
acquired in the market and the rate at which they were
exchanged with the Federal Reserve.
*
*
*
In other operations during the period, on February 9,
1990, Poland repaid in full its outstanding commitment
to the ESF of $86.0 million. The commitment was from
a drawing made at the end of 1989 under a $500 mil
lion short-term support package established with the
ESF and the Bank for International Settlements (acting
for certain participating member banks).
Also in February, Mexico repaid in full its outstanding
commitments under the $ 2,000 million multilateral facil
ity established on September 14, 1989. Mexico made
partial repayments of $6.1 million each to the Federal
Reserve and ESF on February 2. Final repayments of
$728.0 million and $328.0 million were made to the
Federal Reserve and ESF, respectively, on February 15.
On March 23, the Federal Reserve and Treasury
agreed to establish a $1,300 million short-term credit
facility with Mexico. The Federal Reserve’s share of
72
FRBNY Quarterly Review/Spring 1990
$700 million was provided under an existing reciprocal
swap line, while the ESF’s share of $600 million was
provided under a special swap arrangement. On March
28, Mexico drew the entire amount of the facility. Sub
sequently, Mexico made partial repayments on two
occasions in April, reducing its outstanding commit
ments to $541.8 million for the Federal Reserve and
$464.4 million for the ESF by the end of the period.
On March 16, 1990, the U.S. Treasury and the Bank
for International Settlements (acting for certain partici
pating member banks) agreed to provide the Republic
of Venezuela with short-term support of $400 million
for economic adjustment efforts. The ESF’s share in
the facility was $104 million, of which $25.0 million was
drawn on March 30. Venezuela repaid $15.3 million to
the ESF on April 6 and the balance on April 30,
thereby liquidating the facility.
As of the end of April, cumulative bookkeeping or
valuation gains on outstanding foreign currency bal
ances were $1,996.9 million for the Federal Reserve
and $881.8 m illion fo r the ESF (the la tte r figure
includes valuation gains on warehoused funds). These
valuation gains represent the increase in dollar value of
outstanding currency assets valued at end-of-period
exchange rates, compared with the rates prevailing at
the time the foreign currencies were acquired.
The Federal Reserve and the ESF regularly invest
their foreign currency balances in a variety of instru
ments that yield market-related rates of return and that
have a high degree of quality and liquidity. A portion of
the balances is invested in securities issued by foreign
governments. As of the end of April, holdings of such
s e c u ritie s by the Federal R eserve am ounted to
$6,631.3 million equivalent, and holdings by the Treas
ury amounted to the equivalent of $6,977.3 million val
ued at end-of-period exchange rates.
A New Publication of the Federal Reserve Bank of New York
U.S. MONETARY POLICY AND FINANCIAL MARKETS
Ann-M arie Meulendyke
U.S. M onetary Policy and Financial Markets describes
the developm ent of m onetary policy by the Federal
Open Market Com m ittee and its implem entation at the
Open Market trading desk. Author Ann-M arie M eulen
dyke, a senior econom ist and m anager in the Open
Market Function at this Bank, offers a detailed exam
ination of the tools and procedures used to achieve
policy goals. She takes the reader through a typical
day at the tra d in g desk, e x p la in in g how the s ta ff
gathers and analyzes information, decides on a course
of action, and executes an open m arket operation.
The book also places monetary policy in broader his
torical and operational contexts. It traces the evolution
of Federal Reserve m onetary policy procedures from
their introduction in 1914 to the end of the 1980s. It
describes how policy operates through the banking
system and the financial markets. Finally, it considers
the transm ission of monetary policy to the U.S. econ
omy and the effects of policy on econom ic develop
ments abroad.
This book should be of particular interest to students
of money and banking and to people working in the
financial industry who seek a fuller understanding of
the Federal Reserve’s role.
1990, paper, 231 pages
Orders should be sent to the Public Inform ation
Department, Federal Reserve Bank of New York, 33
L ib e rty Street, New York, N.Y. 10045. Checks should
be made payable to the Federal Reserve Bank of
New York. Postpaid: $5.00 U.S., $10.00 foreign.
FRBNY Q uarterly Review/Spring 1990
73
Recent FRBNY Unpublished Research P apersf
9006. Peristiani, Steve, and Gikas A. Hardouvelis. “ Mar
gin Requirements, Speculative Trading and Stock
Price Fluctuations: The Case of Japan.” May 1990.
9007. B e n n e tt, Paul. “ E v id e n ce on the In flu e n c e of
Financial Changes in Interest Rates and Monetary
Policy.” May 1990.
9008. Ryding, John. “ Housing Finance and the Transmis
sion Mechanism of M onetary Policy.” May 1990.
9009. Hirtle, Beverly. “ Bank Loan Commitments and the
Transmission of Monetary Policy.” May 1990.
9010. H irtle , Beverly. “A Sim ple M odel of Bank Loan
Commitments and Monetary Policy.” May 1990.
9011. Cantor, Richard. “A Panel Study of the Effects of
Leverage on Investm ent and Em ploym ent.” May
1990.
9012. Lee, W illiam. “ Corporate Leverage and the Conse
quences of M acroeconom ic Instability.” May 1990.
9013. H irtle, Beverly, and Jeanette Kelleher. “ Financial
M arket Evolution and the Interest S e nsitivity of
Output.” May 1990.
9014. Chan, Anthony, Carl R. Chen, Cheng F. Lee, and
Shafiqur Rahman. “A Cross-Sectional Analysis of
Mutual Funds’ M arket Timing and Security Selec
tion Skill.” June 1990.
9015. W o rthin gto n , Paula R. “ In vestm e nt and M arket
Power: Evidence from the U.S. Manufacturing Sec
tor.” June 1990.
9016. McCauley, Robert N., and Dan P. Eldridge. “ The
British Invasion: Explaining the Strength of U.K.
Acquisitions of U.S. Firms in the Late 1980s.” June
1990.
tS in g le co p ie s of th e s e p a p e rs are a v a ila b le upon
request. Write to Research Papers, Room 901, Research
Function, Federal Reserve Bank of New York, 33 Liberty
Street, New York, N.Y. 10045.
Digitized 74
for FRASER
FRBNY Q uarterly Review/Spring 1990
Single-copy subscriptions to the Quarterly Review (ISSN 0147-6580) are free. Multiple
copies are available for an annual cost of $12 for each additional subscription.
Checks should be made payable in U.S. dollars to the Federal Reserve Bank of New
York and sent to the Public Information Department, 33 Liberty Street, New York, N.Y.
10045 (212-720-6134). Single and multiple copies for U.S. and for other Western
Hemisphere subscribers are sent via third- and fourth-class mail, respectively. All
copies for Eastern Hemisphere subscribers are airlifted to Amsterdam and then for
warded via surface mail. Multiple-copy subscriptions are packaged in envelopes con
taining 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,
providing they are reprinted in full and include credit to the author, the publication,
and the Bank.
Library of Congress Catalog Card Number: 77-646559
FRBNY Quarterly Review/Spring 1990
J3
CD
C
3
u
03
(Q
CD
Q
c
03
s s
5 u.
2 S'
z £5
■< 5
. d
>
@FedFRASER