Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Services
The Changing Market in Financial
Services
Proceedings of the Fifteenth Annual Economic Policy
Conference of the Federal Reserve 8ank of St. Louis.
edited by
R. Alton Gilbert
The Federal Reserve Bank of St. Louis
"
~.
Preface ix
1
The Opening of New Markets for Bank Assets 3
Gary Gorton and George G. Pennacchi
II 39
2
Interstate Banking, Bank Expansion and Valuation 41
Gerald A. Hanweck
3
The Market for Home Mortgage Credit: Recent Changes and Future
Prospects 99
Patrie H. Hendershott
4
Equity Underwriting Risk 129
J. Nellie Liang and James M. O'Brien
v
VI THE CHANGING MARKET IN FINANCIAL SERVICES
III 159
5
The Competitive Impact of Foreign Commercial Banks in the United
States 161
Lawrence G. Goldberg
6
The Competitive Impact of Foreign Underwriters in the United States 211
Robert Nachtmann and Frederick J. Phillips-Patrick
Index 247
Contributing Authors
Gary Gorton
The Wharton School Herbert M. Kaufman
University of Pennsylvania Financial Systems Research
Philadelphia, Pennsylvania 19104 Arizona State University
Bac 519
Stuart I. Greenbaum Tempe, Arizona 85287
J.L. Kellogg Graduate School of
Management
Northwestern University J. Nellie Liang
Leverone Hall Division of Research and Statistics
2001 Sheridan Road Board of Governors of the Federal
Evanston, Illinois 60208 Reserve System
Washington, D.C. 20551
Gerald A. Hanweck
Department of Finance
School of Business Administration Robert Nachtmann
George Mason University KATZ Graduate School of Business
4400 University Drive University of Pittsburgh
Fairfax, Virginia 22030 352 Mervis Hall
Pittsburgh, Pennsylvania 15260
Samuel L. Hayes III
Graduate School of Business
Administration James M. O'Brien
Harvard University Division of Research and Statistics
Baker 333 Board of Governors of the Federal
Soldier Field Reserve System
Boston, Massachusetts 02163 Washington, D.C. 20551
vii
viii THE CHANGING MARKET IN FINANCIAL SERVICES
Gary C. Zimmerman
Frederick J. Phillips-Patrick Economic Research Department
Office of Economic Research Federal Reserve Bank of San
Office of Thrift Supervision Francisco
1700 G Street NW. P.O. Box 7702
Washington, D.C. 20552 San Francisco, California 94105
Preface
ix
x THE CHANGING MARKET IN FINANCIAL SERVICES
reports evidence that securitization has affected both the level and pat-
tern of fixed-rate mortgage interest rates, making these rates move more
closely with other long-term rates. Another major change has been the
growth of adjustable-rate mortgages. The continued growth of these
mortgages is likely to be limited by the extent to which these loans are
securitized. A third major development in the 1980s was the rapidly
declining role of savings and loan associations as providers of residential
mortgage credit. Hendershott concludes that the past shrinkage in the
share of mortgage credit provided by savings and loan associations does
not appear to have raised home mortgage rates, nor does he believe that
the future shrinkage will affect mortgage rates.
Herbert M. Kaufman focuses on some of the policy issues raised by
Hendershott's article. One such issue is the status of government agencies
that have developed the market for mortgage-backed securities. Kaufman
concludes that the integration of the mortgage market with other capital
markets is now so well developed that it is time to eliminate the subsidies
to these government-sponsored agencies. The other policy issue that
Kaufman addresses is the rationale for specialized mortgage lenders.
Kaufman argues that because of the integration of the mortgage and
capital markets, reductions in the share of mortgage credit have not raised
mortgage interest rates. This integration is so complete that the rapid
reduction in credit provided by a sector that was important in the past
does not affect market rates.
J. Nellie Liang and James M. O'Brien estimate the distribution on
returns to underwriters on individual issues of corporate equity securities.
Their estimates indicate that the frequency of losses is small, with the
underwriting spread more than covering declines in share prices around
the time new issues come to market. The variability of equity underwriting
returns primarily reflects a variability in the number of offerings, not in
the returns on individual issues.
Liang and O'Brien conclude that large commercial banks are relatively
well positioned to gain entry into corporate securities underwriting, and
that, once established, the variability of returns would primarily reflect
the variability in the frequency of underwriting.
Lawrence G. Goldberg describes the growth of foreign banks in the
United States and summarizes the results of studies of the motivations for
this growth. Goldberg concludes that several of the determinants of foreign
bank presence indicate continuing future growth of foreign banks in this
country. He finds no support for arguments that foreign banks have com-
petitive advantages over domestic banks.
In his comment on Goldberg's paper, Gary C. Zimmerman points out
xii THE CHANGING MARKET IN FINANCIAL SERVICES
that all of the recent foreign bank expansion in the United States has
been by Japanese owned banks. He argues that any analysis of growth
in foreign banks in the United States should focus on the growth of
Japanese banks and the reasons behind that growth.
Zimmerman provides additional perspective on the issue of competi-
tive advantage of foreign banks. Some have argued that foreign banks
have lower cost of funds than domestic banks and that foreign banks are
able to exploit this advantage by lending in the United States at rates
below those offered by domestic banks. Zimmerman's analysis indicates
that, rather than using relatively low cost funds raised in Japan to lend in
the United States, the Japanese banks, on net, are raising funds in the
United States and lending them in Japan.
Robert Nachtmann and Frederick J. Phillips-Patrick study the market
shares of domestic and international securities firms in underwriting
corporate securities. They characterize the domestic underwriting business
as very competitive, with falling underwriting spreads throughout most of
the decade and entry by foreign securities firms providing competitive
pressure. The share of securities underwritten in the United States by
foreign securities firms had risen in the 1980s, but the share of securities
underwritten in other countries by firms headquartered in the United
States had also risen. On net, the global market share of firms headquartered
in the United States had remained relatively constant in the 1980s. These
changes in shares reflect the integration of world financial markets with
no evidence of net disadvantage for domestic securities firms. Statistical
tests indicate no significantly adverse effects of the entry by foreign securities
firms into this field on the profits or stock prices of domestic securities
firms.
Samuel L. Hayes, in his comments on the article by Nachtmann and
Phillips-Patrick, focuses on the composition of equity underwriting that is
subject to foreign competition. Hayes assumes that, in the United States
as in other countries, equity underwriting is done primarily by domestic
underwriters. Thus, Hayes assumes that securities firms primarily underwrite
debt issues outside their home markets. Hayes expresses doubts about the
ability of empirical tests to indicate the competitive effects of entry by
foreign securities on domestic firms.
R. Alton Gilbert
Assistant Vice President
I
1 THE OPENING OF NEW
MARKETS FOR BANK ASSETS
Gary Gorton
George G. Pennacchi
Introduction
3
4 THE CHANGING MARKET IN FINANCIAL SERVICES
that banks provide special services that cannot be obtained when securities
are offered to the public. One service deals with the production of infor-
mation regarding the credit risk of a potential borrower. Another service
refers to the monitoring of borrowers so that loan covenants can be enforced
while loans are outstanding. Individual investors may have a very difficult
time undertaking these activities because of "free riding" by investors.
Free riding prevents individual investors from recovering the costs of
producing information or monitoring. Realizing that this problem exists,
no individual would be willing to pay the costs. Even if individuals were
willing to bear the costs, a superior arrangement would be to delegate the
tasks to another agent, namely, a bank, rather than duplicate each others'
costly efforts. 6
The difficulty with delegating these services to banks is that depositors
are unlikely to have the ability to directly observe banks' performance. In
other words, an information asymmetry exists between banks and depositors
(and between banks and other outsiders such as potential loan buyers).
In such an environment what guarantees that banks will actually choose
to produce information and monitor borrowers? Incentive compatibility
between the bank (equity holders) and the depositors requires that the
bank be a residual claimant on the loan's return, such as the case when
the loan is held on the bank's balance sheet or sold with recourse or a
guarantee. This insures that the bank equity holders will see to it that the
bank will, in fact, perform these services (since otherwise, they will be the
first to suffer losses).7 If loans were sold without recourse, the bank would
have no incentive to perform these services, that is, equity holders would
no longer be at risk for nonperformance. These loans would decline in
value because potential loan buyers would realize this incentive problem
exists and offer less to buy the loans. This would have the effect of inducing
banks to hold the loans rather than sell them. Thus, the existence of
banks is synonymous with the creation of assets which must be held until
maturity. This is the basis for the term intermediary.
The analysis equates the existence of banks with the nonmarketability
of their loans. But, this argument appears to be directly contradicted by
the recently observed experience of banks selling loans without recourse
or guarantee. The argument is also contradicted by the willingness of
third parties to credit enhance asset-backed securities. Apparently, banks
can be induced, at least partially, to perform the services associated with
loans without being required to hold a claim on the loans until maturity.
What is this new mechanism for insuring incentive compatibility? In other
words, why are third party buyers, or third party guarantors in the case
of asset-backed securities, willing to rely on the bank to continue to perform
6 THE CHANGING MARKET IN FINANCIAL SERVICES
on the original loan contract after the loan has been sold? This question
is the focus of this article.
A number of hypotheses may explain how markets for bank assets
could develop.8 This paper considers the development of the market in
which banks sell their loans. We explore three hypotheses about mechanisms
that could make loan sales incentive compatible. First, there is the possi-
bility that the bank offers an implicit (unobservable) guarantee or insur-
ance on the value of the loan. Perhaps the bank is willing to buy the loan
back, bearing the losses, if the credit risk of the underlying borrower
increases. Such a promise would have to be implicit since an explicit
promise would not allow the bank to remove the loan from its balance
sheet. Since this arrangement would be implicit, it would likely be hard
to detect though Gorton and Pennacchi (1989) claim to have found some
evidence for this contract feature.
A second possible mechanism for making loan sales incentive compat-
ible involves the selling bank holding a fraction of each loan sold. If the
selling bank sells only a part of the loan, retaining the rest, then the bank
will continue to have at least a partial incentive to perform on the loan.
Either of these two possibilities, or the two features combined, could
explain the apparent paradox of loan sales. However, it is not clear, as we
shall see, how these two contract features could be enforced. Nothing in
written loan sales contracts, at least for commercial and industrial loan
sales, requires selling banks to hold a fraction of the loan, and nothing
requires the bank to guarantee the value of loans sold. If these implicit
contracts exist, then market forces (as opposed to legal forces) must enforce
them.
There is a third possibility that could help explain the feasibility of
loan sales. It may be that there has been a significant reduction of the
agency problem between loan buyers or third party loan guarantors and
loan selling banks. Perhaps, in contradiction of the assumptions of aca-
demic researchers and the implicit assumptions of policymakers, loan buyers
or guarantors can observe whether or not banks produce information and
monitoring. Reductions in the cost of information production and trans-
mission would be at the root of such a change. If the banks' actions,
information production and monitoring, are observable (unlike in the
past), then it may be possible to enforce bank performance without
directly requiring them to risk their equity. Determination of the cause of
the opening of bank asset markets is a necessary part of addressing the
question of whether there are limits to the securitization trend.
A number of articles have addressed banks' motivation for loan sales
and have pointed out the incentive compatibility problems associated with
THE OPENING OF NEW MARKETS FOR BANK ASSETS 7
selling bank assets;9 however, there has been little empirical work inves-
tigating the possibly implicit contractual features that banks might use to
alleviate these incentive problems. One reason for this is a lack of data
to analyze. In this article we review and summarize our attempts to dis-
criminate between the above hypotheses regarding the mechanisms whereby
loan sales could be feasible. While the focus of the article is on loan sales,
the same issues arise for asset-backed securities. So, we begin in the second
section by briefly explaining asset-backed security contracts and by re-
viewing recent developments in asset-backed security markets. Section
three is devoted to briefly tracing and explaining the growth of the (in-
dividual) loan sales market. Section four begins the discussion of the
relative merits of the first two hypotheses explaining the existence of
these markets. The first tests aimed at testing these hypotheses are reviewed.
Section five introduces a simple model that provides the basis for more
sophisticated tests. We then review some earlier work that has tested the
various hypotheses. In Section six we analyze the hypothesis that tech-
nological change has resulted in a significant reduction of the information
asymmetry between banks and outsiders. The opening of new markets for
bank assets has implications for the regulation of banks. These are briefly
discussed in the concluding comments of Section seven.
Asset-backed Securities
* Principal Amount
** As of August 30th.
NA: Not Available
Source: Goldman, Sachs & Co.
total issuance was over $50 billion. Table 1-1 provides the data for the
recent history of the asset-backed market. Table 1-2 provides information
about the types of issuers of these securities.
One important feature of asset-backed securities, mentioned above,
is the presence of credit enhancement devices in the contract. Credit
enhancement is provided by a third party credit facility (such as a letter
of credit or surety policy) as well as over-collateralization. Banks must rely
on third party credit enhancement (rather than issuing their own guarantees
as is sometimes the case with non-bank issuers) because a guarantee by
the issuing bank would not allow the deal to qualify as a sale of assets
according to Regulatory Accounting Principles (RAP). Credit protection
provided by a third party usually covers all losses up to a fixed percentage
of the principal. This percentage is usually significantly in excess of the
loans' historic default rates. In some issues of asset-backed securities, a
straightforward recourse provision is included under which the third party
THE OPENING OF NEW MARKETS FOR BANK ASSETS 9
Loan buyers must be assured that the bank originating the loan will continue
to behave as if the loan had not been sold. If incentives for the selling
bank are not maintained, the value of the loan will decline. In the past it
appears that it was not technologically possible to provide such assurances
without the bank directly risking its equity. To analyze the existence of
markets for bank assets, we will first proceed by attempting to identify the
mechanism that makes loan sales incentive compatible. If we can identify
this mechanism, we will then ask how it is enforced. As mentioned in
the Introduction, three possible mechanisms are: (1) implicit insurance by
the selling bank against possible losses to the buyer due to failure of the
underlying borrower, (2) maintenance by the selling bank of a stake in
Table 1-3. Quarterly Outstanding Loan Sales of Commercial Banks* ($ billions).
Date Loan Sales Loan Purchases
198302 26.7
03 26.8
04 29.1
198401 32.8
02 33.3
03 35.5
04 50.2
198501 54.6
02 59.9
03 77.5
04 75.7
198601 65.4
02 81.2
03 91.3
Q4 111.8
198701 162.9
02 195.2
03 188.9
Q4 198.0
198801 236.3 16.64
02 248.4 16.22
03 263.0 17.65
04 286.8 19.29
198901 272.7 16.16
02 276.5 18.20
03 290.9 17.82
04 258.7 19.89
199001 228.3 16.07
02 190.2 15.94
* Sales reported are gross and exclude sales of consumer loans and mortgage loans. Also
excluded are loans subject to repurchase agreements or with recourse to the seller.
Source: FDIC Call Reports, Schedule L.
Table 1-4. Top 25 Loan Selling Banks in 199002.
DSB Number Name Loans Sold Loans Purchased
12060730 Security Pacific National Bank 53,602,449 o
2364840 Bankers Trust Company 34,602,000 419,000
2364900 Citibank, N.A. 14,526,000 768,000
2364985 Chemical Bank 10,461,000 62,000
2364965 Chase Manhattan Bank, N.A. 8,835,613 226,602
4426723 Mellon Bank, N.A. 8,038,004 80,343
2365328 Morgan Guaranty Trust Company of New York 5,591,960 o
2365290 Manufacturers Hanover Trust Company 5,507,000 4,000
2365441 Security Pacific National Trust Company 4,520,819 o
7171630 First National Bank of Chicago 4,369,881 72,632
7171560 Continental Bank, N.A. 3,915,535 96,244
12061400 Bank of America, NT & SA 3,458,000 150,000
3420204 Philadelphia National Bank 2,696,399 57,091
5512345 Signet Bank-Virginia 2,425,409 132,137
2364880 Bank of New York 2,018,508 62,131
2361310 Marine Midland Bank, N.A. 1,817,920 36,500
12060755 First Interstate Bank of California 1,217,951 41,442
1250487 Bank of New England, N.A. 1,116,253
5512430 Crestar Bank 1,104,857 24,197
°
2364893 Bank of Tokyo Trust Company 1,048,930 113,259
1250370 First National Bank of Boston 816,384 9,236
9274037 First Bank, N.A. 590,536 52,280
7171650 Harris Trust & Savings Bank 581,274 94,685
2342200 Midlantic National Bank 531,009 o
5370355 NCNB National Bank of North Carolina 530,072 267,167
Source: FDIC, Call Reports, Schedule L.
THE OPENING OF NEW MARKETS FOR BANK ASSETS 15
any loan sold, and (3) a lack of information asymmetry between loans
selling banks and loan buyers. We first consider implicit insurance or
guarantees.
The motivation for the possible existence of the second incentive com-
patibility mechanism is empirical. It turns out that selling banks often
retain a fraction of the loan sold. In a sample of 872 loans sold by a major
money center bank during the period from January 20,1987 to September
1,1988, the selling bank held a portion of 360 of these loans (41 percent).
Of those loans for which the bank held a positive share, its share aver-
aged approximately 59 percent.27 Table 1-4 provides more detailed infor-
mation on the fraction retained.
Causal observation of Table 1-4 suggests that the fraction retained
increases as maturity increases, holding the rating of the underlying
borrower constant. Perhaps less obvious is whether the fraction retained
increases as the underlying borrower's rating worsens, holding maturity
constant.28
The behavior of the fraction retained is difficult to understand without
consideration of the prices at which loans were sold, and possibly without
consideration of whether or not an implicit guarantee is offered. Certainly,
an implicit guarantee and maintenance of a fraction of the loan by the
selling bank are not mutually exclusive possibilities. Also, the extent of
the fraction sold and the extent of the guarantee are not independent of
the price of the loan sale.
The fraction of the loan sold by the selling bank was not a variable that
was considered by Gorton and Pennacchi (1989). The behavior of this
variable could explain the results. To see how the fraction of the loan sold
affects the results, Gorton and Pennacchi (1990A) rerun regression (1)
with the new data set, adding the fraction of the loan sold, s, as an ex-
planatory variable. Also, with the new data set the risk of the underlying
borrower defaulting can be directly measured as the spread of the loan
rate over the riskless rate (proxied by LIBOR) of the same maturity,
'b - 't· The results of that regression are as follows:
'1 - 't = .00098 + .2388 ('b - 't) + .0079 b -
.00051 s
(.00009) (.0118) (.0506) (.00010)
No. of obs. = 872. R2 = .329. (Standard errors in parentheses).
Note that the loan sale spread, '1 - 't, is significantly related to the spread
paid by the borrowing firm. It is also positively related to the risk of bank
failure, but this coefficient is not significantly different than zero.29 The
fraction of the loan sold is, paradoxically, negatively related to the spread
THE OPENING OF NEW MARKETS FOR BANK ASSETS 19
Al
Average Fraction Sold .917 .867 .746 .455 1
Number of Observations 8 34 27 20 0 3
A2
Average Fraction Sold .733 .826 .810 .746 .600 .608
Number of Observations 3 41 73 64 18 9
A3
Average Fraction Sold .778 .771 .625 1
Number of Observations 0 3 8 4 1 0
NR
Average Fraction Sold .889 .784 .738 .703 .707 .750
Number of Observations 9 210 206 88 15 10
on the loan sale. This is, however, not necessarily, a puzzle. It may be the
case that banks choose to sell larger fractions of less risky loans. Perhaps
some loans require less monitoring by the selling bank. Selling a larger
fraction of these loans does not increase the yield paid to loan buyers
as much as more risky loans.
Clearly, introduction of the new variable, the fraction sold, s, so suf-
ficiently complicates the analysis that a model is required to be clear
about the issues. Gorton and Pennacchi (1990A) introduce a model aimed
at more rigorously addressing the question of the existence and possible
interaction between the fraction sold, an unobservable guarantee, and the
price of a loan sold.
20 THE CHANGING MARKET IN FINANCIAL SERVICES
In this model, banks have an incentive to sell loans in order to avoid the
costs of required reserves and required capital. Banks can improve the
expected return on loans by monitoring borrowers, but bank monitoring
of borrowers is assumed to be unobservable so that banks and loan buy-
ers cannot write contracts contingent on the level of monitoring. Thus, as
discussed above, since banks provide a special service, namely, monitor-
ing, a moral hazard problem exists. The bank may not monitor at the
most efficient level after having sold its loans. Since loan buyers are
rational, and understand the moral hazard problem, they will only buy
loans if there is some way of forcing banks to monitor.
The bank's problem is to maximize expected profits from the sale of a
particular loan. 31 A bank loan is assumed to require "one dollar of initial
financing. It produces a stochastic return, x, at the end of 'l" periods, where
x e [0, L], and where L is the promised end-of-period repayment on the
loan. The return, x, has a cumulative distribution function of F(x, a), where
a is the level of monitoring by the bank. The bank has a constant returns
to scale technology for monitoring loans. The cost function is given by
c(a) = c*a.
The bank can sell a portion, s, of the loan where s e [0, 1]; the bank
retains the portion (1- S).32 Risk neutral loan buyers require an expected
return on loans purchased of rf' The bank finances its portion with a
weighted average cost of deposit and equity financing given by rio Also,
the bank can offer an implicit (partial) guarantee against default of a loan
that it sells. The proportion of a loan sale that the bank promises to
guarantee is r e [0, 1]. Such implicit guarantees are costly because regu-
lators do not approve of them,33 The cost to the bank is assumed to be
given by k(r), where k' > 0 and kIf ~ O. The bank can fulfill this guarantee
only if it is solvent at the time the loan matures. The probability that the
bank is solvent, p, is assumed to be uncorrelated with the return on the
loan.
The optimal loan sale contract requires the bank to choose a level of
THE OPENING OF NEW MARKETS FOR BANK ASSETS 21
max fOL (1- s)xdF(x, a) - srp fOL (L - x)dF(x, a) - c(a) - k(r) - er/t ]
where:
f:
ing and the expected return on the loan. The form we assume is 34
Using this form, the first order conditions for problem (3) also yield
another expression:
(Je-(r,,-rf)T - J.LIL
s = -------:-'----:----
(1- yp)[1 + () - e-(r,,-rf)'r - J.LIL]
s= '[-'f-J.LI(-rL) (6)
(1- rp)[,[ -'f + 'Is -'f - J.LI( -rL)]
22 THE CHANGING MARKET IN FINANCIAL SERVICES
In expression (6) note that when s is less than one, so that f.l = 0, the
fraction of the loan sold is approximately proportional to the ratio of the
excess cost of internal financing (over the risk free rate) to the excess cost
of internal financing plus the excess cost of funds received from the loan
sale.
Equations (4) and (6) provide the basis for the first empirical tests in
Gorton and Pennacchi (1990A). The two relations define the links between
the bank's optimal choice of sand y. By choosing a parametric form for
the guarantee cost function k(r), the equilibrium values of sand rcan be
determined. However, note that if one assumes that the bank is constrained
to give the same partial guarantee, r, on all the loans that it sells, then
equation (6), alone, is sufficient to determine the bank's optimal share of
each loan sold given the guarantee.
We now turn to summarizing the results of estimating the model. Details
are provided in Gorton and Pennacchi (1990A).
The model is tested using the sample of 872 individual loan sales made by
a large money center bank during the period January 20,1987 to September
1, 1988. If the loan selling bank's partial guarantee, r, is the same for each
loan, then we can treat it as a parameter to be estimated using equation
(6).35
The empirical results in this case do not support this hypothesis. The
basic result obtained is that the term (1- rp) in equation (6) appears with
the wrong sign or that the point estimate of ris negative and significantly
different from zero. This result is nonsensical, and so, offers no support
for the hypothesis. However, we do find empirical support for the propo-
sition that the bank's choice of the fraction sold is a decreasing function
of the loan sale spread, indicating that the loan sales buyers are aware of
the implications of the fraction sold for the bank's incentives.
Summary
P= c 1- rp (7)
L(1_e-(rl,-rf)~) [1-s(1- yp)]
The left hand side of equation (7) equals the loan specific parameter
Pwhich is a measure of a given loan's benefit from greater monitoring.
On the right-hand side are two multiplicative ratios. The first, which is
independent of the fraction of the loan sold, s, and the fraction of the loan
guaranteed, r, is the relationship between the loan sale yield, 'Is> and P
that would hold if monitoring were observable. The second term on the
right hand side denotes the effect of unobservability. We can then test the
relationship given in equation (7) using the borrower's commercial paper
rating as a proxy for Pi' This was done using 360 of the 872 observations
on loan sales for which the borrower reported a commercial paper rating.
Equation (7) was estimated in log form as a probit model, where a
borrower's commercial paper rating was assumed to be a discrete measure
of its benefit from bank monitoring. It was assumed that the fraction
of the loan guaranteed (if any) was the same for all loan sales, so that r
was treated as a parameter.
If monitoring is observable, there should be a positive and significant
relationship between P and the first term on the right-hand side of (7),
while there should be an insignificant relationship between Pand the second
term. Both terms should show a significant relationship.to Pif monitoring
is unobservable.
The empirical results support observability to the extent that the first
term on the right hand side of (7) was significantly related to Pwhile the
second was not. However, the effect of the first term was not as strong as
the theoretical model would suggest. 40 We interpret these results as sug-
gesting that loan buyers have the ability to verify, at least partially, the
performance of loan selling banks.
THE OPENING OF NEW MARKETS FOR BANK ASSETS 27
Notes
1. Evidence that bank loans are unique is provided by James (1987), Lummer and
McConnell (1989), and Fama (1985). Using event study methodology, James shows that
there is an (abnormal) positive return on stocks when nonfinancial firms announce that they
have obtained a bank loan. The stock market reaction (in the sense of abnormal return)
upon the announcement of offerings of common stock, preferred stock, convertible preferred
stock, convertible bonds, and straight bonds is negative, though not always statistically
significant. (See Smith (1986).) A similar study by Lummer and McConnell discriminates
between new and revised bank loan announcements. Their results show that only revised
loan announcements have a positive effect on stock returns, suggesting that banks learn
about borrowers after an initial loan is made. Also, both Fama and James point out that
bank borrowers, not bank depositors, bear the cost of bank reserve requirements. Yield
spreads between bank CDs, which are subject to reserve requirements and other money
market instruments, are not significantly different. This implies that firms are paying more
to borrow from a bank than they would if they borrowed the same dollar amount by issuing
securities on the open market. Presumably, firms would only do this if bank loans involved
the production of some services not obtainable by issuing securities in the capital market.
2. All theories of banking panics crucially depend on the illiquidity of bank assets either
by assuming that there is an information asymmetry between banks and depositors or by
assuming that there is a cost to liquidating bank assets. See Calomiris and Gorton (1990) for
a discussion of the causes of banking panics.
3. Throughout this essay we will ignore mortgage-backed securities, at least those secur-
ities that are backed by mortgages carrying government (e.g., VA and FHA) guarantees
against default. This government credit enhancement appears to be the primary determinant
of their marketability.
4. Apparently, the logic behind this regulation is that if a bank sells a loan with recourse
and thus retains the loan's risk, it should be required to hold the same amount of capital and
reserves as if it had retained the loan's risk by not selling it.
5. Since nearly all loans are sold without recourse or guarantee, this supports the hypo-
thesis that bank loan sales are motivated, at least in part, by a desire to avoid the regulatory
costs associated with required reserves and required capital. Note that this hypothesis is
consistent with the observation that most loan buyers are other banks. Pennacchi (1988)
shows that banks possessing considerable loan making opportunities but competitive deposit
markets (money center banks) will find it profitable to sell loans while banks possessing
limited loan making opportunities but having market power in their local deposit markets
(foreign banks and smaller domestic banks) will find it profitable to buy loans. It follows
that the recent increase in aggregate loan sales volume can be explained by the recent
increase in deposit market competition faced by many banks.
6. The formal arguments, and details are provided by Campbell and Kracaw (1980),
Boyd and Prescott (1986), Diamond (1984), and Gorton and Haubrich (1987).
7. Presumably this is part of the logic behind capital requirements for commercial banks.
8. Some hypotheses explaining the existence of markets for bank assets we do not ex-
plore. For example, one possibility is that bank loan sales and asset-backed securities simply
represent bank underwriting and are strictly limited to assets which do not involve any
information production or monitoring. This hypothesis argues that asset sales represent an
attempt by commercial banks to compete with investment banks. We believe this explanation
THE OPENING OF NEW MARKETS FOR BANK ASSETS 29
cannot be totally satisfactory. First, most loans that are sold are the obligations of borrowers
who do not have commercial paper ratings, suggesting that they do not have ready access
to public securities markets. (See Gorton and Haubrich (1989).) Secondly, it seems unlikely
that, lacking any special services by the bank, this would be a successful way to compete.
In a loan sale, if the underlying borrower fails, the holder of the secondary participation has
no legal connection with the failing firm and hence cannot be represented in bankruptcy
court. Holders of marketable securities, thus, have important rights which holders of sec-
ondary loan participations do not have.
9. Flannery (1989), Benveniste and Berger (1987), Cumming (1987), Greenbaum and
Thakor (1987), Kareken (1987), James (1988), Pennacchi (1988), and Boyd and Smith (1989)
suggest a variety of possible motivations for loan sales.
10. There are a large number of other types of loans that have formed the basis for
asset-backed securities. Examples include light truck loans, motorcycle loans, boat loans,
trade receivables, and equipment leases.
11. Other offerings of automobile securities by commercial banks in 1988 included: two
issues totaling $480 million by Marine Midland Bank, two issues totaling $470 million by
Chemical Bank, a $250 million issue by Huntington National Bank, and a $181 million issue
by Signet Bank (Virginia). Thrifts also issued automobile securities during 1988. Empire of
America Federal Savings Bank made two offerings of automobile securities totaling $631
million. Western Financial Savings Bank came to market four times in 1988 for a total
issuance of $460 million. Rochester Community Savings Bank offered $125 million of these
securities which featured a floating rate coupon.
12. In other words, suppose the initial pool was $100 million of principal with $75 million
sold to the investor class and $25 million sold to the seller class. Then if, at the end of a
month, the principal balance of the pool falls to $95 million, the investor class remains at
$75 million, while the seller class would drop to $20 million.
13. Details and terminology vary with the issuer. For further details see "Credit Card
Backed Securities: An Introduction," Goldman, Sachs & Co., January 1987; "Securitization
of Credit Card Receivables Using a Senior/Subordinated Structure," Goldman, Sachs &
Co., December 1988; "Rating of Credit Card Receivables" Duff & Phelps Inc., (no date);
"Credit Card Receivables: Moody's Examines the Risks," Moody's Investors Service, January
1987.
14. Of the $80.2 billion dollars of outstanding loans sold as of June 30, 1990 by banks
responding to the Federal Reserve System's Senior Loan Officer Opinion Survey of Lending
Practices, 3.5 percent were nonperforming loans. It is not clear that this entire amount
represents defaults since some loans may have been nonperforming when they were sold.
One year earlier 1 percent of loans were nonperforming out of a total outstanding amount
of $72.2 billion for survey respondents.
15. Some loan sales take the form of assignments, rather than participations. Assign-
ments are quite strong contracts because the assignees have all the rights and responsibilities
of the lead or originating bank. In practice these rights vary a great deal depending on the
complexity of the contract and the sophistication of the parties involved. Importantly, many
assignments specify a minimum amount which must be held by the originating bank. See
Gorton and Haubrich (1989) for further discussion.
16. This is based on the sample of banks surveyed in the Federal Reserve System's
Senior Loan Officer Opinion Survey on Bank Lending Practices for various years. See also
Gorton and Haubrich (1989).
17. About 37.5 percent of the outstanding loans sold as of June 30, 1990 were merger and
30 THE CHANGING MARKET IN FINANCIAL SERVICES
acquisition related. (See the Federal Reserve Senior Loan Officer Opinion Survey of Bank
Lending Practices, August 23, 1990.)
18. According to the Federal Reserve's August 1990 Senior Loan Officer Opinion Survey
on Bank Lending Practices, 11.8 percent of loans were purchased by domestic banks with
assets under $2 billion while 25.8 percent of loans were purchased by domestic banks with
assets over $2 billion. Nonbank buyers accounted for 24.1 percent of loan sales, with 7.9
percent of this being purchases by nonfinancial corporations and 16.2 percent being purchases
by other nonbank buyers.
19. No bank we contacted was willing to provide any hard evidence on the amounts of
loan repurchases or the prices of repurchases, though banks were often willing to admit that
repurchases of loans were not infrequent.
20. By the G1ass-Steagal Act banks are prohibited from issuing securities. In order to
avoid blatant violations of Glass-Steagal secondary loan participation contracts explicitly
forbid secondary market sales. The existence of such sales plays a role in the Icgal deter-
mination of what constitutes a security. See Gorton and Haubrich (1989).
21. Investment banks have recently started to make a secondary market in some
participations. See Wall Street Journal, January 4, 1991, p. 1.
22. Investment banks that underwrite commercial paper will usually agree to repurchase
this paper prior to maturity for the expressed purpose of providing liquidity.
23. See Gorton and Pennacchi (1989) for further discussion.
24. The linear specification is ad hoc, but dictated by data limitations as will become
apparent.
25. The data on loan sales yields and LIBOR was collected by a survey of a few money
center banks by an industry publication called Asset Sales Report. Loan sales yields were
classified by maturity and commercial paper rating of the underlying borrower. For details
see Gorton and Pennacchi (1989). The premium for bank risk, b" was calculated using bank
stock prices as described in Gorton and Pennacchi (1989).
26. The guarantees are options, but they are vulnerable options, that is, the bank writing
the option may default. Vulnerable options have different comparative statics than stand-
ard, exchange traded, options. The main point is that an American option can be valued
significantly higher than a European option because the right to exercise early increases in
value when the likelihood of default by the option writer increases. The holder of commercial
paper can only call on the guarantor at the maturity of the paper, whereas the loan sale
holder may be able to exercise at any time. This means that the effective premium that loan
sales buyers place on bank risk, the b, variable, will be quite small relative to commercial
paper buyers. This can explain the negative coefficient obtained on that variable. See Johnson
and Stulz (1987) on vulnerable options.
27. Hence, the overall average of the fraction of loans sold that were retained by the
bank was approximately 24 percent. This data set was obtained somewhat subsequent to the
study by Gorton and Pennacchi (1989). Interestingly, the bank involved was motivated to
provide the data mostly because it was convinced that there was no implicit guarantee. In
the data set, the average maturity of the loans sold was 28.04 days. The average maturity
of the loan sales contracts was 27.63 days. The average loan interest rate was 7.53, while the
average loan sale interest rate was 7.41. Further details on the sample can be found in
Gorton and Pennacchi (1990A).
28. The following OLS regression supports these observations. Letting fr = the fraction
of the loan sale retained by the bank, maturity = the maturity (in days) of the loan, and rate
= {4 if no rating, 3 if A3, 2 if A2, I if AI, and 0 if AI+J, we have:
THE OPENING OF NEW MARKETS FOR BANK ASSETS 31
39. See Gorton and Pennacchi (1990B,C) for the theoretical rationale and empirical
evidence of a shift from bank financing to direct financing.
40. Since estimation was in log form, the right-hand side of (7) was transformed into two
additive terms. The second term's coefficient estimate was of the wrong sign (negative) but
insignificantly different from zero (its theoretical value under the hypothesis of full ob-
servability). The first term had a coefficient estimate that was of the correct sign (positive)
and significantly different from zero (.37 with a standard error of approximately .06) but
also significantly different from its theoretical value of unity.
41. There is, however, abundant evidence that the demand for bank provision of these
services has fallen. The rise of the commercial paper market and the medium term note
market suggest that the same technological forces which make loan sales feasible have
allowed directly marketable instruments to compete more effectively with bank loans.
References
Becketti, Sean. 1990. "The Truth about Junk Bonds," Federal Reserve Bank of
Kansas City Economic Review 75(4), (July/August): 45-54.
Benveniste, Lawrence and Allen Berger. 1987. "Securitization With Recourse: An
Investment that Offers Uninsured Bank Depositors Sequential Claims," Journal
of Banking and Finance 11, (September): 403-24.
Boyd, John and Edward Prescott. 1986. "Financial Intermediary-Coalitions," Journal
of Economic Theory 38, 211-32.
Calomiris, Charles and Gary Gorton. 1990. "The Origins of Banking Panics: Models,
Facts, and Bank Regulation," in Financial Markets and Financial Crises, ed. by
Glenn Hubbard (University of Chicago Press).
Campbell, Tim and William Kracaw. 1980. "Information Production, Market
Signalling, and the Theory of Financial Intermediation," Journal of Finance 35(4)
(September), 863-82.
Cumming, Christine. 1987. "The Economics of Securitization," Quarterly Review
(Autumn), Federal Reserve Bank of New York, 11-23.
Diamond, Douglas. 1984. "Financial Intermediation and Delegated Monitoring,"
Review of Economic Studies LI (July): 393-414.
Fama, Eugene. 1985. "What's Different About Banks?," Journal of Monetary
Economics 15 (January): 29-39.
Flannery, Mark. 1989. "Capital Regulation and Insured Banks' Choice Individual
Loan Default Risks," Journal of Monetary Economics 24, 235-58.
Gorton, Gary and Joseph Haubrich. 1990. "The Loan Sales Market," Research in
Financial Services, ed. by George Kaufman, Vol. 2, JAI Press Inc: Greenwich
CT.
- - . 1987. "Bank Deregulation, Credit Markets, and the Control of Capital,"
Carnegie-Rochester Conference Series on Public Policy 26 (Spring): 289-333.
Gorton, Gary and George Pennacchio 1989. "Are Loan Sales Really Off-Balance
Sheet?," Journal of Accounting, Auditing, and Finance 4(2), 125-45.
THE OPENING OF NEW MARKETS FOR BANK ASSETS 33
35
36 THE CHANGING MARKET IN FINANCIAL SERVICES
the purchase of debt by investors from the high savings country is their
lack of asset-originating capability in the low savings country. It would be
plausible for the investors from the high savings country to engage agents
(brokers) whose role it would be to find attractive assets for the high
savers to purchase. To be sure, the high savings country investors will not
be oblivious to the possibility of being exploited by myopic or cynical
brokers. The investors will, therefore, be attentive to the brokers' repu-
tations, and they may seek assurances, both written and implied. However,
the larger the spread between the home- and foreign-country interest
rates, the more compelling the foreign investment, despite the inhibiting
effects of private information. It is then possible that nothing changed in
the informational sphere and loan sales might have exploded (owing to
nothing more profound than changing population age profiles and savings
proclivities) !
Likewise, consider two countries that have chosen to subsidize their
banks by restricting entry, suppressing deposit interest rates, and under-
pricing governmental deposit insurance. It does not really matter why the
two countries might have chosen to adopt such a policy; only the fact that
deposits are subsidized matters. The deposit subsidy will prompt banks to
accept all proffered deposits, and we may even observe banks competing
in oblique ways to attract deposits, quite irrespective of investment op-
portunities. With sufficiently generous subsidies, the banks could profit-
ably hold risk-free assets, exclusively. In such an environment we would
expect the banks to develop asset origination skills in order to deploy the
proceeds of deposit liabilities.
Now, suppose that, for whatever reason, the deposit subsidies in one of
the countries disappears. Instead of being able to borrow at some sub-
risk-free interest rate, the impacted banks would now face market interest
rates. To make matters really dreadful, suppose that these suffering banks
are downgraded by rating agencies in light of a perceived excess supply
of banks and bankers. After all, the size of the industry was previously
bloated by the subsidies directed to banks.
Would it then seem reasonable for the impacted banks to sense an
excess capacity to originate assets relative to their funding capabilities?
Would it not seem reasonable then to originate assets for sale to others,
perhaps to banks in the country still enjoying deposit subsidies? What, if
anything, would such sales have to do with informational asymmetries?
To be sure, informational asymmetries will impede the asset sales, but in
what sense do the asymmetries explain the growth of loan sales?
I have one more question regarding the Gorton/Pennacchi paper. It is
correctly noted that public regulation of banks is related to the traditional
38 THE CHANGING MARKET IN FINANCIAL SERVICES
Introduction
41
42 THE CHANGING MARKET IN FINANCIAL SERVICES
more than forty-seven states and the District of Columbia have permitted
bank holding companies headquartered in other states to own banks in
those states. The remaining states with no interstate banking statutes are
expected to enact such legislation soon. Since geographic banking oppor-
tunities have been altered, the recent changes in state laws regarding
interstate banking should be expected to affect the valuation of commercial
banks. To the extent that such dramatic changes in permissible bank
expansion affect valuation, it is likely that such effects would vary with
the location, nature of banks, and the type of state law change. As this
article states, these changes in state law are not uniformly reflected in
bank valuation changes.
Even though the expansion by the major banking companies has been
constrained, much interstate expansion of banking services has taken place
since the 1970 amendments to the Bank Holding Company (BHC) Act
via nonbank subsidiaries of bank holding companies (section 4(c)(8) of
the Bank Holding Company Act as amended) and loan production offices
of banks. By the end of 1988 there were 6,778 interstate nonbank offices
of banking companies and 7,492 domestic and foreign interstate banking
offices.' In many respects, the presence of a nonbank office, such as a
mortgage banking company, can be just as effective as the presence of a
full-service bank, since many wholesale and retail customers are shopping
for a particular type of credit arrangement or other financial service. The
major handicap to the lack of a full-service banking office is the solicitation
and provision of many deposit services. The proximity of a bank for cus-
tomers to make deposits and receive cash is important for retailers and
households, and special service offices and automated teller machines
(ATMs) are not convenient substitutes. Thus, even though there has been
considerable geographic nonbank expansion by BHCs, this is not an
effective substitute for interstate banking. It is likely, then, that the trend
will be Lor banking companies to expand geographically through full-
service banking company acquisitions rather than nonbank subsidiaries
of the holding company or bank. Also likely, is that the trend will accel-
erate by early 1991 when approximately thirty-two states adopt some
form of national interstate banking, with twelve of these having no recip-
rocal restrictions. 2
As will be developed in a discussion of the history and state of inter-
state banking, the movement toward full-scale interstate banking has been
hampered by state legislation that has attempted to protect the market
value of existing banking companies in the states. By early 1991, thirty-
two states will have some form of national interstate banking with twelve
of these having no reciprocal restrictions. As a result, almost any large
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 43
Prior to the mid-1960s, banks had little incentive to expand beyond state
boundaries. However, changing economic and technological conditions
have stimulated commercial banks to reconsider their perceived optimal
structural organizations, and many of the larger banks have attempted to
circumvent the legal restrictions on interstate banking by way of nonbanking
affiliates of bank holding companies. 4 State banks are chartered to oper-
ate within a single state. The introduction of national banks by the Na-
tional Banking Act of 1863, and subsequent legislation and interpretations,
prohibited branching by national banks. The McFadden Act of 1927 and
the Glass-Steagall Act of 1933 allowed national banks branching oppor-
tunities equal to that permitted state-chartered banks. This legislation
eliminated, in effect, branching across state lines.
Through the bank holding company organizational form, commercial
banks have been able to circumvent the legal restrictions on interstate
branching. Prior to 1956, bank holding companies were unregulated and
some banks used this vehicle to establish a multistate presence. The Glass-
Steagall Act of 1933 provided for limited regulation of bank holding
companies by the Federal Reserve System, but did not limit interstate
expansion.
The Bank Holding Act of 1956, however, prohibited further establish-
ment of interstate banking offices but grandfathered the seven existing
domestic interstate banking organizations at that time. Foreign banking
organizations were permitted to establish full banking operations across
state lines until the International Banking Act of 1978 attempted to equalize
opportunities available for foreign and domestic banks. Though prohib-
ited from expanding full-service offices across state lines, this legislation
permitted interstate foreign banks to retain existing networks. In addition
to banking companies, there are several types of limited service facilities,
such as Edge Act Corporations, loan production offices, foreign bank
branches and agencies, non bank banks, and savings and loans,S which have
been permitted on an interstate basis.
Section 3( d), the Douglas Amendment, of the Bank Holding Company
Act of 1956 permitted states the power to legislate out-of-state bank holding
companies to acquire, but not charter, and operate banks within each
state. Seven domestic bank holding companies and five foreign banking
companies were permitted to continue their interstate banking activities
through a "grandfather" provision of this Act. Maine was the first state
to pass a law permitting interstate acquisitions in 1975, later liberalized
in 1978 to national reciprocal, and in 1984 to open entry by banks
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 45
Note: Several states prohibit acquisition of banks in operation for less than a specified
number of years.
Source: Financial Structure Section, Board of Governors of the Federal Reserve System.
service banks in multiple states, but note that there are numerous other
banks which control various types of entities in multiple states.
The final group of banks in table 2-3 are those that had expanded
interstate as state laws had begun to change. Since 1982, many states for
example, South Dakota and Maryland, have permitted out-of-state BHCs
to establish banks with limited powers. These banks with limited powers
serve customers of their parent organizations in many states, but the
limitations on their powers keep them from competing with local banks
for a full range of banking services. The primary objective of the states in
permitting entry by the out-of-state BHCs was to attract capital and jobs
to the states. Some states have actively competed to attract large banks.
Money-center banks have pushed for further relaxation of these restric-
tions, and have succeeded in getting reciprocity legislation passed in
several states.
The expansion of banks interstate via chartered banks has been rapid
since 1982. In 1983 there were twenty-six domestic banking companies
operating i full-service banks interstate. Four years later, the Federal
Reserve reported that fifty-one banking organizations had subsidiary
banks in one or more states beside their home state (Savage 1987). Today
this number is approximately 180 banking companies after accounting for
interstate acquisitions since 1987. This number includes expansion by
acquisition of failing banks. This rapid expansion is also reflected in the
number of banking offices controlled by interstate banking companies.
Domestic banking companies with interstate banking operations have
increased the number of banking offices outside their home states under
their control from 1,258 in 1983 to 7,364 in 1988 (see table 2-4). In terms
of banking assets, these banking companies had $442.4 billion in interstate
bank assets as of March 31, 1990, which represents an increase from $148.4
billion as of June 30,1986. These data indicate nearly a tripling of banking
assets associated with interstate banking within the past four years and
nearly a six-fold increase in banking offices since 1983. Both these figures
suggest a very rapid expansion of interstate banking via full-service banks.
VI
......
VI
IV
VI
W
Ul
.j::o.
Table 2-3. Bank Holding Companies with Subsidiary Banks in More Than One State (District of Columbia and Puerto Rico
included as states for this table).
Vl
Vl
56 THE CHANGING MARKET IN FINANCIAL SERVICES
Banks have been able to offer limited interstate banking services through
various means since the passage of the so-called Edge Act in 1919. An
Edge Act corporation can be established outside a bank's home state to
deal primarily with loans and deposits related to international trade. The
International Banking Act in 1978 liberalized the extent to which Edge
Act corporations of banks could offer loans and deposits to customers,
but the international trade motivation of the relationship was supposed
to be maintained. These institutions grew rapidly in the 1970s and early
1980s, but have recently shown a considerable slow down. Since 1983, the
number of domestic Edge Act corporations has declined from 143 in
eighteen states to 79 in sixteen states, as of 1988 (see table 2-4). This
decline is attributed to both an expansion in full-service banking offices in
many areas via interstate banking and a reduction in the attractiveness of
international trade (see King, et ai., 1989).
58 THE CHANGING MARKET IN FINANCIAL SERVICES
1987 1990
Total Total
Interstate Interstate Total %
State Assets Assets Assets Interstate
Oregon 5.8 9.5 22.3 42.5
Pennsylvania 0.0 21.0 160.6 13.1
Rhode Island 3.6 4.7 17.5 27.0
South Carolina 6.9 11.4 24.0 47.7
South Dakota 2.9 2.9 18.0 16.1
Tennessee 0.8 14.2 45.9 31.0
Texas 0.0 68.9 167.2 41.2
Utah 1.3 3.6 12.2 29.8
Vennont 0.0 1.0 5.8 18.1
Virginia 2.6 4.6 67.4 6.9
Washington 12.2 29.8 37.8 78.7
West Virginia 0.0 0.2 17.0 1.2
Wisconsin 1.3 8.4 44.2 19.1
Wyoming 0.5 1.4 4.3 33.0
Total U.S. 148.4 442.4 2,915.9 15.2
Source: Federal Reserve Bulletin, February 1987, Table 2A, and Financial Studies Section,
Board of Governors of the Federal Reserve System, September 1990.
Dating back to the previous century, interstate banking has been con-
sidered by many banking company managements as a preferred means
of conducting banking business. This is evidenced not only by the current
rush by banks to form interstate relationships, but by the number of
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 61
Source: Federal Reserve Bulletin, February 1987, Table 3, and Board of Governors of the
Federal Reserve System, Financial Studies Section, September 1990.
Interstate banking is, as the evidence presented in this study shows, giving
rise to greater consolidation among banking companies and to more large-
size banking firms nationwide. This result has not been unanticipated.
Many observers have predicted that small local banks will not be able to
successfully operate in direct competition with large superregional and
money-center banks. Their ability to survive such competition under
interstate banking will depend on their cost efficiency compared to much
66 THE CHANGING MARKET IN FINANCIAL SERVICES
larger banks, their ability to generate profits, and their efficiency in adopting
new technology. Employing a simple rule of inference, small banks are
likely to survive interstate banking if they have been able to survive the
intrastate expansion that has taken place in a number of states over the
past thirty years. In this regard, the evidence since 1985 (Amel and Jacowski,
1985, 125) clearly shows that the small bank is far from becoming extinct
(see Rhoades, 1985, 1130, for a similar pre-1985 assessment).
Of the studies that have recently addressed interstate banking com-
petitive effects directly, all have found no meaningful anticompetitive or
procompetitive effects. The study by Phillis and Pavel (1986), mentioned
earlier, found that banks expanding interstate preferred to acquire banks
with larger branching networks, strong local deposits bases, and a greater
concentration in consumer lending. A study by Goldberg and Hanweck
(1988) compared the performance of those domestic BHCs grandfathered
by the 1956 Bank Holding Company Act with those banks and BHCs
operating in the same states but without interstate banking privileges.
The study confirms that the grandfathered BHCs experienced a statisti-
cally significant decline in the share of state deposits and a homogeniza-
tion in their profitability performance and portfolio composition compared
with their peers in the same states over the 1960 to 1983 period. By 1970
the extent of the homogenization is nearly complete, and it is difficult to
find much difference between grandfathered BHC affiliates and banks of
the same size in the same state; by 1983 the homogenization process
appears to be complete. These results are similar to those of other related
studies (Rhoades 1984, and Golembe 1979) that support the view that it
is unlikely that interstate banking will result in the dominance of local or
regional banking markets by large interstate banking organizations
(Goldberg and Hanweck, 1988, 67).
Gilligan and Smirlock (1984) found little to suggest that scope economies
are important in banking. Most of these studies do suggest that unit banking
is inefficient, resulting in significant diseconomies of scale at larger sizes
and which were not exhibited by branching banks (Berger, Hanweck and
Humphrey 1987). Several recent studies have focused on scale economies
using only the largest banks in their estimates. Their results suggest
economies of scale among the largest banking companies at about $20
billion in real assets. The problem with these studies is that there is little
factual rationale for such economies particularly in light of the fact that
automation costs have plummeted relative to computing power or service
over the past fifteen years.
The safety and soundness of the banking system may be enhanced
through interstate banking expansion. Banks can now become more
geographically diversified in both their deposit liabilities and lending in
commercial and retail markets. Because of the larger available markets,
interstate banking companies may be able to rely more upon less volatile
core deposits related directly to their banking business in local markets.
Interstate banking provides a stop-gap avenue for acquisition of failing
banks that, when unavailable, raise the cost to taxpayers and the FDIC of
failed bank resolutions. Finally, on a less positive note, interstate banking
is creating many more superregional and large size banking companies
with wide-ranging networks of bank and non-bank affiliation throughout
the nation. The failure of anyone of these will tax the FDIC bank reserve
fund (BIF) and require the Federal Reserve to provide the support to
avoid liquidation until a merger partner can be secured. With the continued
increase of larger banking companies, the FDIC will become increasingly
vulnerable to such failures resulting in greater uncertainty and instability
within the banking system. lO
While these 'mergers of equals' may thus be expected to have relatively few
benefits from a diversification standpoint, recent interstate acquisitions may
nevertheless represent a managerial strategy to change the structure of bank-
ing markets served in order to create an environment more conducive to higher
returns on equity capital, or reduced risk exposure, or greater price discrimi-
nation, etc., for acquisition-bound banking companies.
It may well be that full-service interstate bank expansion is less a vehicle for
the aggressive pursuit of profit and value maximization through expanded
revenues and greater cost economies and more a defensive reaction to in-
creased risk occasioned by deregulation and the emergence of new domestic
and foreign competitors ....
These results suggest that both the motivation and effects on banking
company valuation from interstate expansion will result in a deterioration
in banking company valuation. Consequently, the markets at the time of
the announcements of liberalized interstate banking law changes should
have taken these factors into account. Rose's results suggest that there
should have been either no change in bank valuation or a decline in value
when these laws were first passed and when they became effective.
The effect on banking company valuation from changes in interstate
banking laws is evaluated in a recent study by Goldberg, Hanweck and
Sugrue (1990). They hypothesize that announcement of a liberalization in
interstate banking laws would positively affect the stock prices of banks
in that state, especially for those banks most likely to be acquired. An
increase in the number of potential bidders for a bank would likely increase
the value of a bank in a liberalizing state. In addition, a new law might
trigger reciprocity provisions in other states and thus increase expansion
opportunities to lucrative out-of-state markets. From another perspective,
the liberalization of interstate banking might be regarded as a stimulant
for economic development in a state. Maine, for example, the first state
to adopt interstate banking legislation, was anxious to encourage the New
England banks to bring resources to the state. Alternatively, the value of
established firms in a state could be reduced by increased competition
from out-of-state entrants.
The study by De and Millon (1988) on interstate banking acquisitions
indicates that acquiring firms have increased in value. Therefore, banks
outside of states with changing laws would increase in value if interstate
expansion were deemed to be a valuable option and would not change if
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 71
NY 820628 820628
AK 820701 820701
MA 821230 830701 X X
RI 830517 840701 X X
870701
cr 830608 830608 X
ME 840207 840207
GA 840405 850101 X X X X
UT 840406 840415
KY 840407 840715 X X
860715
SC 840521 860101 X X X X X X X X
FL 840522 850701 X X X X X X
NC 840707 850101 X X X X X X X X
OR 850312 860701 X X X X X
ID 850312 850701
VA 850324 850701 X X X X X X X X
AZ 850418 861001
IN 850418 860101 X X
TN 850501 850701 X X X X X X X
WA 850516 870701
NV 850613 850701 X X X X X
MD 850621 850701 X X
870701
OH 850718 851017 X X X X X X
DC 851008 851122 X X X X X
IL 851125 860701
MI 851205 860101 X X
MO 860218 860813 X X X X
AL 860221 870701 X X X X X X X
MN 860318 860701
W1 860331 870101 X X X X
PA 860626 860825 X X X
LA 860709 870701 X X X X X X X
TX 860924 870101
CA 860929 870701 X X X X X
WY 870316 870523
NH 870514 870901 X X
MMMMMMNNNNNNNNOOOPRSSTTUVVWWWW
A INS 0 T E V H J M Y C D H K R A leD N X T T A A V I Y
x x x
x x x
x x x x
x x x x
x x x x x
x x x x x
x x x x x x
x x x x
x x x
x x x x x
x x
x x x x x x
x x x x x
x x
x x x x x x x x
x x x x x x
X X X
X X
X X X X X X
X X X X
X X X X
X X X X X X X X
X X X X X
X X X
74 THE CHANGING MARKET IN FINANCIAL SERVICES
This study documented in sections two and three the rapid expansion of
interstate banking and the substantial consolidation of banking resources
that has been taking place since about 1982 in the largest fifty and one
hundred banking organizations. From 1980 to 1989 there have been over
ninety-five Federal Reserve approved bank mergers and holding com-
pany acquisitions among banking companies, with each having in excess
of $1 billion in domestic deposits. This total excludes acquisitions by foreign
banks. Most of these domestic banking acquisitions have taken place since
1982 and about thirty-two are acquisitions of banks interstate. The char-
acter of this expansion has been such that regional banks have been at its
forefront of this expansion. In view of adverse developments in profitabil-
ity, market valuation, and capitalization over the past year affecting many
of the "superregional" banking companies, the sustainability of this rapid
expansion is in question and with it the future growth in interstate bank-
ing. Coupling this with the forecast recession in the general economy, and
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 75
Percent
ROA
1.20
ROA-Small ROABanks
1.00 Banks $300M to $5B in
-l Assets
0.80 --- t
--------~~o;--------
.
ROA - All Banks -~_ -----....
- - - l . _ _-7'-~",,--\ ................
0.60 /
/ ......-.:-...... "" ,""
,,,.,,.~
-------------~-- ------------\
0.40
----_J /
/
,\
0.20 ROA Other Large \
Banks ,,
\
,,
0.00
,,
, ,,
-0.20 ,,
,,
,,
-0.40
,,
,
-0.60 \ I
ROA 9 Money \ :
Center Banks - \ :
-0.80
'J
-1.00
During the 1980s, overall bank profitability has both declined, in terms of
either rates of return on assets (ROA) or on equity (ROE), and demon-
strated a high degree of variability (see figures 2-1 and 2-2). Over the
period from 1983 to 1989, for which consistent data are available for
banks by size classes, the ROA for all banks varied from a high of 0.84
76 THE CHANGING MARKET IN FINANCIAL SERVICES
Percent
ROE
25.00
20.00
ROE Banks
15.00 ROE - All Banks
..
$3OOM to $5B in
.. :....
." .-----------,1'-
/ ----
................... ::\... Assets
10.00
-........
/
/ .~
... ---------,;...
....
, ....... ~
----- ..
-:,......... t
----
.... ~~~~
J
---- ./ T ....-,--- - •
5.00 • - ROE-Small \ •
',~ 'I :
ROE Other Large Banks
'/:
.
\
Banks \ \/'
0.00 ,, ".,:
\ \ /
,, ,
,, ,,
,
,, ,
,,,
-5.00
,,
,, ,,
-10.00 ,, ,,
, ,
-15.00
ROE 9 Money \
Center Banks \ :
'
., '
\ :
"'I
-20.00 +,---+-----+---4-----+----+----f...-----i
1983 1984 1985 1986 1987 1988 1989
Source: Board of Governors of the Federal Reserve System
percent to a low of 0.11 percent Likewise, ROE for all banks over the
same period ranged from 13.7 percent to 1.94 percent These values are
in sharp contrast with bank earnings performance even during some of
the turbulent periods of the 1970s.
Much of this variation and general decline is due to the performance
of large banks. As figures 2-1 and 2-2 demonstrate, ROA and ROE for
smaller banks, those with assets less than $5 billion, may have declined
over the period, but even facing difficult earnings pressures from agriculture
and energy price declines, their variability was not excessive. Although
it is not the purpose of this study to dissect the banking industry, and
particularly larger banks, to determine the causes of the present profitability
problems, it is sufficient to point out that these problems have persisted
for much of the 1980s. For larger banks, their sources are well known and
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 77
Percent
NIM to Assets
4.50
NIM-Small
......__ ...,................................ Banks ....... _
4.00 ... -............ __ i ____--------......
Banks $3OOM to
$5B in Assets
t ~~####~---------------------- ... ---- --_ .. -----------
3.50
----------............. NIM . All Banks
--
t
/'
~------...--...-
3.00 ,-}
/' T ~
...-..-"- Other large " ..................
...--- Banks /' --........._
~~
2.00
---------------------
.
----------... ............
9 Money Centet----_ .."
Banks
"'
I" ""
1.50
1.00
stem largely from the overhang of less developed country (LDC) loans
originated in the late 1970s and regional real estate recessions.
In order to emphasize this point, the net interest margin to total assets
(NIM) is presented in figure 2-3 for all banks and those of the four size
classes. Without exception, there has been a slight upward trend in NIM
over the 1983 to 1989 period. Only for the nine money-center banks has
there been any significant variability. These data suggest that banks of all
sizes have been able to maintain a consistent spread between interest
income and interest expense despite deregulation and greater competi-
tion from nonbank and foreign banking sources.
The principal cause of the variation and downward trend in bank
78 THE CHANGING MARKET IN FINANCIAL SERVICES
Percent
LLP to Assets
2.50
•
2.00
" I,,
I
I '- 9 Money Center
I
I \, Banks
I ,
I ,
I ,
I ,
I ,
I ,
1.50 I 1'\ \
I 1\\ I
I
I I \' I
I
I I \ \ I
I
I I \ \ I
I
I I \' I
I I ' I
llt \
I
1.00 LLP-SmaU
II
I
II \'\
\ '
I
I
/
:~~Large '~
I /
l/
........:: .. '/
t ,..,.. . . . . . . . . . . . . . . . . . . . . . ,,-';:.----- ~................... \\ l/
. . . . . .". ,,*' ,'. . . . . /~ . . . . . . . . . - ". . . . . . . . ~ /~"",-
LLP to Assets - .............
0.50
"
....~",,','
..
... '
"",
All Banks ' ,
.....t1---__
____~-- • ,I
....... Banks $3OOM to 'I
$5B in Assets
0.00
profitability is due to loan losses. Even though banks have been able to
maintain consistently strong net interest margins, as demonstrated by loan
loss provisions as a proportion to assets, loan losses have cut deeply into
this basic source of bank earnings. With the exception of the smallest
banks, those with assets less than $300 million, which have shown a re-
versal since 1986, provisioning for loan losses remains on an upward trend
(see figure 2-4). Taking into consideration the possibility of a general
economic recession in 1990 and the spreading regional real estate recessions,
there is little likelihood that this trend will be reversed in the next several
years.
The variability and downward trend in bank earnings has limited bank-
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 79
Percent
Equity to Assets
9.00
Equity to Assets
Small Banks _------
8.00
---
~~~-----
j
------- -_...............
'_"_"_'_"-'~":::"'::--::-~-----------------------------~-!~~7----------~
-.
.
6.00
Equity to Assets
All Banks ~-----~~- ~~~~
.
5.00
" ~/ --------~
"" Other Large
--_...............-
,"'" _--------------............. Banks ..... _
--.".-~--- ..................
............ 9 Money Center .. ~------
Banks
4.00
3.00
ers' abilities to add to capital from internal funds. Despite this, banks
have managed to add to equity in sufficient amount to generally show an
upward trend in equity to assets on a book value basis. As figure 2-5
portrays, the smaller banks have the highest ratios of equity to assets,
nearly double that of the nine-money-center banks. With the high degree
of earnings variability demonstrated by the largest banks in recent years,
investors in these banking firms and their large uninsured depositors will
demand more protection in the form of greater equity. In addition, bank
regulators have adopted more stringent capital requirements and en-
forcement procedures. All of these factors contribute to pressure on banks,
especially the largest banks, to add to equity capital.
80 THE CHANGING MARKET IN FINANCIAL SERVICES
Percent
Cash Dividends to Assets
0.55
0.50
0.45
0.40
/
r ,I'.............. ........
0.35 ,,/ /
0.30 Dividends to
• / / Other Large
/ . . . f"" """l/'
Assets - All // Banks
.... Banks / "
0.25 ~,;:;::::::.. ,,','
.... ----....
....t. . . . .,,'' "
0.20 9 Money Center
Banks
0.15
0.10
Despite the obvious need for equity by banks of all sizes, there has
been an alarming upward trend in cash dividends paid by banks (see
figure 2-6). For example, in 1989 small banks paid out 58 percent of after-
tax earnings as cash dividends and the nine money-center banks, exper-
iencing an after-tax loss, paid 0.36 percent of their assets as cash dividends
or about $2.3 billion. For many banks, cash dividends are paid to the
BHC parent holding company. For many of these holding companies,
their banking subsidiaries are their largest holding and cash dividends
their predominant source of revenue. Although these dividends may be
earnings retained by BHCs, it is likely that they are used to meet holding
company expenses associated with nonbank enterprises and payouts to
shareholders of the BHC. Regardless of their use, they are a source of
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 81
cash flow to banks that is not being returned to them in the face of
growing demands for bank equity capital.
For the banking industry and particularly the largest banking companies,
market valuation has not kept pace with other industries in the United
States. At about the time of the start of the bull stock market in 1983,
bank stock values began to lag behind those in the general market.
In figure 2-7, the S&P 500 stock price index is compared with S&P
stock price indices for thirteen major regional banking companies (regional
banks) and seven money-center banking companies (money-center banks).
These data clearly show the divergence in valuation that began to take
place in the early 1980s. To emphasize the divergence and clarify the
timing, figure 2-8 shows the ratio of the S&P 500 index to the two
banking company indices. These data show that, for the regional
banks, the divergence from the S&P 500 began about 1980 or 1981.
For the money-center banks, this divergence began much later in 1985 or
1986. Based on these data, the market valuation and capitalization of major
U.S. banking companies have not kept pace with a broad range of firms in
other industries.
There are certainly individual exceptions to this generalization, but
data for a broader spectrum of banking companies support this same
conclusion. As reported in Business Week (April 13, 1990, 219-222),
ninety-two major banking companies lost 9 percent in market value in
1989, while all industries showed a gain of 15 percent and general nonbank
financial services (insurance, brokerage, etc.) had a 14 percent increase.
Of the ninety-two banking companies reported, thirty-three showed market
value declines. Some of the major gainers were BankAmerica (California),
Crestar Financial (Virginia), Bank of New York, NCNB (North Carolina),
and Banc One (Ohio), while some of the larger losses were experienced
by Bank of Boston, Shawmut National (Massachusetts), Citicorp, Southeast
Banking (Florida), and First Interstate (California). By way of contrast,
1988 was an improvement for most banking companies, while in 1987
banking company market value over a broad range of companies showed
a decline of about 11 percent (Business Week, April 15, 1988,234).
Smaller regional banking companies have fared no better than major
regional and money-center banks since July 1989. Figure 2-9 compares
the monthly S&P's stock indices for the money-center and major regional
banks with the National Association of Securities Dealers Automated
82 THE CHANGING MARKET IN FINANCIAL SERVICES
300
250
S&P500-
200
150
Regional
1\
100
.j
"
V, ~ Banks A
1V
\;,J,' '\ l
,~\
',_\
I " , \
I \
I '
,'\ ,,--...;
50 ,. ____"J
,I' '.,.... __......
........ _ _ _ _ , '
,....,,1
,_""_-""--............ '.,iJ' Money Center
,---' Banks
o
1958 1962 1966 1970 1974 1978 1982 1986 1990
Source: Standard and Poor's Analyst's Handbook: OffIcial Series, 1989.
Latest data plotted: August 29, 1990
Percent
S & P to Bank Indices
4.5
3.5
1.5
0.5
o
1~ 1~ 1~ 1~ 1m 1m 1m 1m 1~ 1~ 1~ 1~
500 1\
,I I,A.
, \ II ~,
470 I \-J I II
440 IA'I
/Vil
II \I\ '('Jt\l\
I
, 1
\
410 J\(J
'I
II
I I
\I
I I: II
380 I I!
I " \
350
)
I
\,
320 \1\
NASDAQ Regional Banks,J
290
t,N
, I
/ I
260
r/ \V
I
230
50
204--+--+--r--~~-+--+--+--~~-+--+--+--~4--+--+--r~~4--+--
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
Source: National Association of Securities Dealers and 51
law in most states has made the prohibitions of the Douglas Amendment
to the Bank Holding Company Act of 1956 virtually obsolete. In order to
get the greatest competitive benefit from interstate banking, banks must
be free to open offices and banks de novo within any state, much like most
other businesses (including bank loan production offices and finance
company subsidiaries of bank holding companies). Such legislation has
been introduced by Congress in the past and debated to some degree.
Some of this legislation prescribes that there be safeguards, such as the
limitation of a banking company gaining excessive share of any single
local market or state. These issues need to be considered and dealt with
within the broader spectrum of antitrust legislation and enforcement. The
financial condition of the banking system may not hinge on whether there
is more liberal interstate banking legislation, but as the profitability and
valuation picture painted in this study suggests, banking companies will
not have the resources to maintain the interstate expansion pace of the
past eight years. Without an alternative means of expansion that requires
less resources than under the present system and allows smaller, more
profitable banking companies to become active in interstate expansion,
further benefits from interstate banking will be long in coming.
Notes
* The author wishes to give special thanks to Don Savage of the Financial Structures
Section of the Board of Governors of the Federal Reserve System for thoughtful discussion
and valuable data.
1. King, Tschinkel, and Whitehead (1989), Table 2, 49.
2. Clair and Tucker (1989), 14.
3. On September 21,1990, Chase Manhattan, the second-largest U.S. banking company,
announced a severe retrenchment by a cut of 5000 staff or 12 percent of its work force by
December 1990. This announcement was part of a major restructuring that includes a larger-
than-expected $600 million addition to loan loss reserves and a cut in dividend. This is
expected to reduce the losses of $665 million taken in 1989. (See the Washington Post,
September 22, 1990, C1).
4. See Goldberg and Hanweck (1988) for a discussion of the legal framework of inter-
state banking and a detailed description of various exceptions to, and means of, circumventing
interstate banking laws.
5. However, the expansion of powers granted S&Ls in the 1980s allowed them powers
similar to banks until they were restricted in 1989 under FIRREA. Nonbank office expan-
sion by bank holding companies will be discussed below.
6. See King, Tschinkel, and Whitehead (1989, 48-50) for a listing of the approved
activities and those prohibited under section 4(c)(8).
7. It is important to note that statewide concentration, as measured by the average five-
firm ratio, increased from 54.7 percent in 1976 to 60.9 percent in 1987. These results support
the evidence for local market concentration increases.
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 87
References
Adkisson, A. and D. Fraser. 1990. "Competitive Effects of Interstate Banking:
The Impact on Bank Acquisition Markets." Presented at the Bank Structure
Conference, Federal Reserve Bank of Chicago. (May): Mimeograph.
Amel, D. F. 1988. "State Laws Affecting Commercial Bank Branching, Multibank
Holding Company Expansion, and Interstate Banking." Board of Governors of
the Federal Reserve System. (May): Mimeograph.
Arne!, D.F., and Michael Jacowski. 1989. "Trends in Banking Structure since the
Mid-1970s." Federal Reserve Bulletin. Board of Governors of the Federal Reserve
System (March): 120-133.
Ahrony, Joseph, Anthony Saunders, and Ithzak Swary. 1988. "The Effects of
DIDMCA on Bank Stockholders' Returns and Risk." Journal of Banking and
Finance 12: 317-331.
- - . 1986. "The Effects of a Shift In Monetary Policy Regime of the Profitability
and Risk of Commercial Banks." Journal of Monetary Economics 17 (May)
363-377.
Beatty, Randolph P., John F. Reim, and Robert F. Schapperle. 1985. "The Effects
of Barriers to Entry on Shareholder Interstate Wealth: Implications for Banking."
Journal of Bank Research (Spring): 8-15.
Benston, G.J., Gerald A. Hanweck, and David B. Humphrey. 1982. "Scale
Economies in Banking: A Restructuring and Reassessment." Journal of Money,
Credit, and Banking XIII (November Part I): 435-56.
Berger, A., D. Humphrey, and J. Frodin. September 1986. "Interstate Banking:
Impacts on the Payments System." Research Papers in Banking and Financial
Economics. Board of Governors of the Federal Reserve System, Washington,
D.C. #90.
88 THE CHANGING MARKET IN FINANCIAL SERVICES
Trifts, lW., and K.P. Scanlon. 1987. "Interstate Bank Mergers: The Early Evidence."
Journal of Financial Research X (Winter): 305-31l.
Thompson, R. 1985. "Conditioning the Return-Generating Process on Firm-Specific
Events: A Discussion of Event Study Methods." Journal of Financial and
Quantitative Analysis 20 (June): 151-168.
Whitehead, D. 1983. A Guide to Interstate Banking. Federal Reserve Bank of
Atlanta.
Zellner, A. 1962. "An Efficient Method of Estimating Seemingly Unrelated
Regression and Tests for Aggregation Bias." Journal of the American Statistical
Association 57 (June): 348-368.
COMMENTARY
Peter S. Rose
93
94 THE CHANGING MARKET IN FINANCIAL SERVICES
decade of the 1980s-the period over which most of the interstate bank-
ing legislation has appeared. Perhaps, as Hanweck suggests, we need to
do more research focusing on the possible application of the dominant-
firm and linked-oligopoly hypotheses specifically to interstate banking to
determine if the apparent trend toward concentration is damaging to the
public welfare.
I share Hanweck's concern for this apparent rise in banking market
concentration, but I'm not yet convinced that this trend is happening to
the degree that his analysis implies or primarily because of interstate
banking. For example, if you look at the article's numbers and citations
(particularly pages 62-65 and in table 2-6), you discover quickly that
the concentration figures are for domestic commercial banking firms only.
There is no adjustment for foreign-bank penetration of United States
markets, which in selected states has become a potent competitive force,
not only in corporate loans and credit guarantees, but also in consumer
and small business credit. Nor is there any adjustment for nonbank financial-
service competitors, ranging from thrifts, finance companies, credit -card
firms (including most recently AT&T), and insurance companies to security
dealers. We must also hasten to point out that the total assets of banks
that Hanweck focuses upon do not describe the structure of any particular
financial-services market.
There may be another explanation besides interstate banking for the
apparent trend toward greater banking concentration. The 1970s and 1980s
represent a period of sweeping changes in state branching laws. At the
beginning of this period the fifty states were roughly equally divided between
states sanctioning statewide branching activity, those allowing limited
branching within city, county, or district boundaries, and unit banking
states where full-service branch offices are prohibited. By the end of the
period, however, the majority of states had adopted statewide branching
regimes, either de novo or by acquisition, and the number of strict unit
banking states had dwindled to less than five. Historically, bank concen-
tration ratios have always averaged highest in statewide branching states,
followed by states with limited branching rules, and lowest in unit banking
states. Therefore, the trend toward national banking concentration that
Hanweck cites may reflect, in large measure, the shift in state banking
rules applying to intrastate expansion, and not primarily to interstate bank
acquisitions.
I'm also not as eager to discard the research evidence on economies of
superscale as Hanweck may be, particularly for those financial services
that appear to be uniquely important among the very largest banks,
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 95
Recent changes in state laws with respect to interstate banking should be expected
to affect the valuation of commercial banks ... (and) it is likely that such
effects should vary with the location, nature of banks, and the type of state law
change ... As this paper reports, these changes in state law are not uniformly
reflected in bank valuation changes.
Certainly the article reports such a result but it does not rigorously test
that result. There is no model per se and there are no statistical tests (though
the time-series graphs on bank financial performance are both interesting
and well-presented). The closest that we get to a rigorous statement of
hypotheses about how interstate banking laws might impact the valuation
of banks and rigorous testing of those hypotheses is a summary of an
unpublished article prepared earlier by Goldberg, Hanweck, and Sugrue
(1990). This recent unpublished work is deemed to support the view that
interstate banking legislation has mostly benefitted those banks situated
in states that have recently changed their interstate banking laws and
banks located in the same region as the state that made the change. The
rationale given for this finding is that new interstate banking laws in-
crease the number of acquiring firms and broaden geographic expansion
opportunities sufficiently to outweigh the value-reducing impact of greater
competition once interstate entry is allowed. From this we are led to the
general conclusion that "the liberalization of interstate banking will have
a positive impact on the future ability of banking companies to increase
their profitability and value."
These hypotheses may be true (and there is a substantial body of
financial theory with which they appear to be consistent), but there is no
way of evaluating the quality and power of the author's test of them
without seeing the seemingly unrelated regressions model and the sampling
scheme to which that methodology is applied. Moreover, as plausible as
Hanweck's hypotheses are, not all the evidence that we currently have is
totally consistent with the conclusions reached.
There is, for example, the argument that interstate bank expansion is
a selective process; only certain kinds of banking firms are likely to par-
ticipate in it. Indeed, this paper provides us with one of the reasons for
this phenomenon: the peculiar nature of interstate banking laws passed
during the 1980s. The banks involved thus far in interstate acquisitions
have apparently not been "average" in a number of key structural
characteristics or performance dimensions (such as asset returns, loan-
loss risk exposure, and operating efficiency). A recent study by this au-
thor (May 1990) finds that, on average, both interstate acquirers and their
targets have tended to perform less well than comparably sized firms. It
is not intuitively obvious that the announcement of an interstate acquisition
INTERSTATE BANKING, BANK EXPANSION AND VALUATION 97
References
Goldberg, L.G. and G.A. Hanweck. 1988. "What Can We Expect from Interstate
Banking?" Journal of Banking and Finance 12: 51-67.
Hanweck, Gerald A. 1990. "Interstate Banking, Bank Expansion and Valuation."
Proceedings of the Fifteenth Annual Economic Policy Conference, Federal
Reserve Bank of St. Louis. (October 18)
Rose, Peter S. 1989. The Interstate Banking Revolution: Benefits, Risks, and Tradeoffs
for Bankers and Consumers. Westport, Conn.: Quorum Books.
--.1990. "Bidding Theory and Bank Merger Premiums: The Impact of Structural
and Regulatory Factors." Unpublished Paper, Department of Finance, Texas
A&M University (February).
- - . 1990. "The Consequences of Interstate Banking Deregulation for Competition,
the Structure of Service Markets, and the Performance of Interstate Financial-
Service Firms." Proceedings of the Conference on Bank Structure and
Competition. Federal Reserve Bank of Chicago (May).
3 THE MARKET FOR HOME
MORTGAGE CREDIT: RECENT
CHANGES AND FUTURE
PROSPECTS
Patrie H. Hendershott
Three major changes occurred during the 1980s in the market for home
mortgage credit: the securitization of fixed-rate mortgages, the develop-
ment of a national primary market for adjustable-rate mortgages, and the
decimation of the savings and loan industry. These changes and their im-
pacts on various financial industries and homebuyers are the subjects of this
article. The three major changes are summarized before a detailed analy-
sis is presented.
The Federal Housing AdministrationlVeterans Administration (FHA/
VA) fixed-rate mortgage (FRM) market was integrated with the capital
market gradually throughout the 1970s via increased usage of the Ginnie
Mae pass-through program, and the conventional FRM market was in-
tegrated in the first half of the 1980s with the development of active
markets for the mortgage pass-through securities of Freddie Mac and
Fannie Mae. Integration was stimulated in the 1980s by the deregulation
of deposit rate ceilings and the erosion of thrift tax subsidies, two de-
velopments that switched thrifts from low cost funders of mortgages to
high cost funders. As a result, coupons on new issue FRMs now swiftly
adjust to changes in other capital market rates, and mortgage funds are
readily available at all times and places at going market rates.
99
100 THE CHANGING MARKET IN FINANCIAL SERVICES
Table 3-1. The Growth in the Securitization of FHAIVA 1-4 Family Loans.
1 2 3 = 2/1 4
FHA/VA GNMA Share of Mortgage
Originations Issues Originations Banker Share of
($ bi/.) ($ bi/.) Securitized Originations
1971-73 15.6 2.7 .17 .70
1974-75 13.5 5.8 .43 .75
1976-79 28.3 17.6 .62 .78
1980-82 21.5 16.6 .77 .81
1983-86 55.0 55.6 1.01 .78
1987-89 54.9" 70.3 1.28" .71"
• Mortgage banker issues are likely understated. Thus originations and the mortgage
banker share are too low, and the share of originations securitized is too high.
Source: Hendershott and Van Order (1989), updated from DataBase, Secondary Mort-
gage Markets, FHLMC.
As the data in table 3-2 indicate, the pass-through market for conforming
conventional loans developed less rapidly. The best measure of the agencies'
presence in this market is the share of new fixed-rate conventional FRMs
(generally defined as less than one year since origination) eligible for
agency securitization (under the conforming limit) that is, in fact, securitized
by Freddie Mac and Fannie Mae. The upper part of table 3-2 lists total
agency pass-through issues, those backed by seasoned FRMs, new ARMs
and multifamily loans, and, by subtraction, new FRMs. The lower part
puts the agency issues backed by new FRMs into a market perspective.
Total single family conventional originations are multiplied by an esti-
mate of the fraction that had a fixed rate, and the result is divided by
agency issues backed by new FRMs to obtain the agency share of the
total FRM market. Of course, the agencies can only participate in the
conforming end of the market. Assuming this end to be 75 percent of
total FRM volume, the last column is an estimate of the percentage of
new FRM conforming loan volume that is sold directly to Fannie and
Freddie. As can be seen, this estimate rose from 4 percent in the 1977 to
1981 period, to almost 25 percent in the 1982 to 1985 period, and to over
50 percent since 1986, including 69 percent in 1989.
The difference in the development of FHAIVA and conventional pass-
throughs in the 1970s and early 1980s stems largely from the historical
differences in the origination of FHAIV A and conventional mortgages.
Mortgage bankers have tended to dominate the FHA/V A market,
Table 3-2a. The Agencies Increased Role in the Conforming FRM Market.
Pass-Throughs Issued by FHLMC and FNMA ($ bU.)
Pass- Throughs Backed-By
1 2 3 4=1-2-3
Total Seasoned New ARMs, Multis, New
Issues FRMs andFHAIVAs FRMs
The perfect mortgage market model says that the coupon rate for a near-
par mortgage depends on a small number of general capital market
variables, such as the six-month and seven-year Treasury rates and the
volatility of these rates (Hendershott and Van Order, 1989). Moreover,
the response to changes in those variables is predictable and fast. In con-
trast, twenty years ago mortgage lending was tied to the thrifts, which
were heavily regulated and also tax-advantaged to invest in mortgages,
and connections of mortgage and capital markets were tenuous and gradual.
Hence, most researchers focused on things peculiar to the thrift industry,
such as deposit rates and deposit flows, rather than general capital market
conditions as determinants of mortgage rates. If one regressed actual
mortgage rates during such a: period on fictional mortgage rates predicted
by the perfect market model, one would expect to obtain a poor fit.
Moreover, to the extent that the predicted rate had any effect, it would
be a lagged one.
Roth (1988) analyzed the integration of mortgage and capital markets
by looking at trends in the month-to-month correlation of changes in
coupon rates on conventional mortgage commitments and ten-year Treas-
uries annually from 1972 to 1987. His results are reproduced and extended
to include 1988 and 1989 in table 3-3. Prior to 1982, the correlation of the
changes ranged from -0.5 to +0.5 and was never statistically different from
zero. After 1981, though, the correlation was never less than 0.58 and was
always statistically positive. Moreover, in the last three years the correla-
tion has averaged 0.90.
A potential problem with Roth's analysis is that the mortgage rate
THE MARKET FOR HOME MORTGAGE CREDIT 107
Table 3-3. Correlation Between Mortgage Rates and Capital Market Rates.
Year Correlation*
1972 -0.22
1973 0.19
1974 0.46
1975 -0.18
1976 0.16
1977 -0.49
1978 0.42
1979 0.34
1980 0.33
1981 0.42
1982 0.80**
1983 0.81**
1984 0.65**
1985 0.76**
1986 0.58**
1987 0.91**
1988 0.87**
1989 0.92**
* Correlations are between month-to-month changes in the FHLMC commitment rate
and the 10-year Treasury rate.
** Significantly different from zero at a 5 percent confidence level.
Source: Roth (1988, Table 1) for 1972-87; 1988 and 1989 computed by the author.
Table 3-5. Actual and Perfect Market Effective Conventional Commitment Rates
(%).
Table 3-6. Effective Loan Rates for California FRMs with Loan-to-Value Ratios
of 75 and 80 Percent by Loan Size, Selected Years.
(those over the conforming limit). The number of loans in the samples
and the percent of the eligible dollar volume securitized by FNMA and
the FHLMC are reported at the bottom of the table (see Hendershott and
Shilling, 1989, for more detail on the loan samples).
In general, one would expect the loan rate to decline with loan size
because the costs of originating and servicing loans decrease per dollar of
loan as the loan size increases. This is clearly the case for loans below the
conforming limit in all years except 1985, where the limited number of
observations results in much noise in the averages. Of most interest, though,
is what happens to the loan rate when the loan size increases above the
conforming limit. Prior to 1986, the loan rate is either flat (1978) or still
decreasing (1985). But in 1986, the loan rate jumps at, and in 1987 just
above, the loan limit; that is, rates on loans below the loan limit are
noticeably lower than those above the limit. The most likely cause is, of
course, the expanded activities of Fannie Mae and Freddie Mac. These
expanded activities probably also influence rates on loans just above the
limit because such loans are likely to be conforming within a year
(Woodward, 1988). Thus, the low rate for loans that are 100 percent to
115 percent of the limit in 1987 may not be as anomalous as it first appears.
Hendershott and Shilling (1989) estimated, separately on data for 1978
and 1986, the relationship between rates on loans closed and the follow-
ing factors: loan-to-value ratio, loan size, precise month the loan was
closed, dummy variables for geographic regions in the state, and whether
the loan was on a new property, was under the conforming limit, or was
just above the limit. The loan-to-value ratio had the expected positive
impact; the loan size and the new property dummies had the expected
negative impacts; and the responses in the two years were remarkably
similar. For those two years, however, the effects of the conforming limit
differed markedly. In 1986, conforming loans had a thirty basis point
lower rate than well-above-the-limit loans, and soon-to-be-conforming loans
had a fifteen basis point lower rate (standard errors were only five basis
points). In 1978, however, the point estimate for the conforming loan
coefficient was only three basis points.
It should be emphasized that the perfect-market rate listed in table
3-5 is computed from a GNMA price equation, not an equation explain-
ing prices on seven- or ten-year Treasury bonds. The working assumption
of Hendershott and Van Order (1989) was that the GNMA market has
been integrated with capital markets since 1981. This assumption seems
plausible because GNMAs have full faith and credit guarantees and have
traded like Treasuries, with comparably low transactions costs and high
volume, at least since 1981. However, some would argue that the
112 THE CHANGING MARKET IN FINANCIAL SERVICES
Prior to the 1980s, there existed one mortgage type-the standard fixed-
payment mortgage (FRM). Periodically in the 1960s and 1970s, increases
in interest rates reminded thrifts of the problems of borrowing short and
lending long. At these points (1971 and 1974 specifically), federally char-
tered thrifts lobbied for permission to offer borrowers an alternative choice,
adjustable rate mortgages (ARMs), that would reprice more in line with
thrift deposits. Congress made clear to the regulatory body (then the
Federal Home Loan Bank Board) that it did not want borrowers to have
that choice (Cassidy, 1984). In December 1978, an exception was made
for federally chartered S&Ls in California, allowing them to compete
with state chartered S&Ls, and in July 1979, nationwide authority to
invest in ARMs was first granted. However, annual and life-of-Ioan rate
increases were restricted to one-half and two-and-a-half percent, respec-
tively, and the choice of rate index was greatly restricted.
Ironically, about the same time that interest rates were peaking in the
early 1980s, congress gave thrifts the opportunity to invest in ARMs that
might be attractive to both lenders and borrowers. In April 1981, fairly
liberal regulations were implemented for federally chartered S&Ls and
THE MARKET FOR HOME MORTGAGE CREDIT 113
savings banks, and in August 1982, these were loosened further and ex-
tended via the Deposit Institutions Act to all state chartered institutions
(although individual states could override the regulations during a three
year period). Thrifts, indeed, took advantage of this opportunity. In the
middle of 1982, ARMs were only 10 percent of the single family mortgage
portfolio of FSLlC-insured S&Ls. By March 1989,48 percent of the thrift
single family loan portfolio (including mortgage-backed securities) was in
ARMs (Hendershott and Shilling, 1990). Moreover, over the 1984 to 1989
period, ARMs accounted for 43 percent of the conventional single family
loan volume originated by all lenders (computable from table 3-2).
The coupon on a FRM reflects the pure cost of "intermediate" term
(say seven-year) money plus the cost of a call or prepayment option.5 Altern-
atively, the FRM coupon can be thought of as the pure cost of short-term
money (say one-year) plus the cost of an intermediate term (say seven-year)
rate cap that will pay the borrower the difference between the actual short-
term rate and the initial one should interest rates rise. 6 The sole existence
of the FRM caused cross-subsidization among borrowers because all
borrowers pay virtually iq,entical costs for the call option or rate cap even
though the option/cap is certainly more valuable to less mobile households
than to more mobile households because longer lived options/caps have
greater value than shorter lived options/caps. That is, with the FRM, more
mobile households, who pay more for the call option/rate cap than it is
worth to them, subsidize less mobile households who pay less than it is
worth to them. The existence of adjustable-rate mortgages (ARMs) would
give mobile households an alternative to overpaying for the FRM and
would reduce the subsidy to less mobile households who choose the FRM.
A second potential advantage of ARMs for at least some borrowers
stems from the facts that initial coupon rates on ARMs are less than those
on FRMs and lenders qualify borrowers based on the lower rates. Thus,
borrowers financing with ARMs can obtain larger loans and purchase
larger houses. The latter suggests that the ARM share of conventional
single-family mortgage originations will vary cyclically as the FRM-ARM
rate spread varies, and figure 3-1 shows that this variation has occurred.
Large rate spreads (two-and-a-half percentage points) in 1984 to 1985 and
mid-1987 to the end of 1988 were associated with 40 percent to 60 percent
ARM shares, while small spreads (one-and-a-half points) in 1986 to mid-
1987 and late 1989 were associated with 20 percent to 25 percent shares.7
Another important determinant of the ARM/FRM choice is the level
of the FRM rate (Brueckner and Follain, 1989); the higher the FRM rate,
the more desperately do households need to obtain a lower initial rate on
which they can qualify for a reasonably sized loan. The sharply reduced
114 THE CHANGING MARKET IN FINANCIAL SERVICES
0.6
,,
Rate Lagged I
, I
,
I
One Quarter I
I
I 2.5
I
I
I
0.5 -- I
I
I
I
I
I
I
I
I
I
I
I
0.4 I
I
2.0
I
I
,.', I
I
I
.
l " I
,
I
I
I
0.3 ,
', I
\I
"•
1.5
0.2
0.1 L-l---1.....J.......J........L....L-..L..-.L....JL....l---1.....J.......L....L....L-..L..-J......JL....l---L.-L....J......L-...1-..L..J 1 .0
841 843 851 853 861 863 871 873 881 883 891 893 901
842 844 852 854 862 864 872 874 882 884 892 894
Source: FHLMC
Figure 3-1. ARM share and FRM minus ARM rate spread quarterly, 1984-1989.
ARM share in 1986 as related earlier years reflected, in part, the marked
decline in FRM yields in 1985 and 1986.
The successful securitization of the FRM market, but not the ARM
market, has likely reduced the FRM rate as related to the ARM rate. This
can significantly alter the ARM/FRM mix. For example, using an FRM
rate of 10.50, an ARM rate of 8.50 and the mean values of the other
demographic variables relevant to the ARM/FRM choice, Brueckner and
Follain's (1989) estimation equation implies a 23 percent probability of
this household choosing an ARM. However, a thirty basis point higher
FRM rate raises the ARM probability to 32 percent.
THE MARKET FOR HOME MORTGAGE CREDIT 115
0.8r---------.--------r-------,--------~------__,
0.7
0.6
0.5 .....................,
,---,' ' ..
" ' ...
"" .........,
.......... ----........... "S&L Share of Mortgage Market"------"
M ,,
......
--...,,,,
......
,,
0.3 ,
\,
0.2~ __ __
~ ~~~~ __ __ ~ __ __
L_~~_L __ ~ ~~ _L~
1965 67 69 71 73 75 77 79 81 83 85 87 1989
Source: Flow of Funds Accounts. First quarter of 1990 annualized.
Figure 3-2. Share of S&L TFA in home mortgages, S&L share of total home
mortgage market, and ratio of S&L TFA to total home mortgages outstanding.
The institutions that have been most affected by, or causally linked to,
changes in the home mortgage market in the 1980s have been savings and
loans (and Fannie Mae and Freddie Mac). A shrinkage in the savings
and loan industry in the very early 1980s and a relative shift of S&Ls out
of home mortgages in the 1982 to 1984 period raised conventional FRM
rates from one-half percentage point below perfect-market levels in the
1970s to one-half point above perfect-market levels in the early 1980s. This
induced securitization, which successfully lowered conforming FRM rates
back to perfect market levels.
Figure 3-2 provides data on both savings and loan behavior and the
relative role of S&Ls as home mortgage investors over the past quarter
116 THE CHANGING MARKET IN FINANCIAL SERVICES
9.5
7 Year Treasury Rate //"'\
,..._./ \
/
A\ /
/ \
/ \. / \
/
r' . . . / ,v/ \\ /
//...... ---."\
/ \ / - -
8.5 / \ /
\ / \ /
\ / \ /', /
\ // \y/.... " /
~ ~~
7.5 I I I I
0188 0388 0588 0788 0988 1188 0189 0389 0589 0789 0989 1189 0190 0390 0590 0790
Future Prospects
The decline in the S&L industry owing to FIRREA has just begun. A
further 30 percent to 50 percent decline over the next five years should
be expected. While some of the S&Ls will be sold to commercial banks,
many banks also face increased capital requirements, and all face higher
deposit insurance premiums. The depository institutions sector generally
is likely to shrink over the next five years. On the other hand, there is no
evidence that the relative decline in the S&L industry since 1984 or the
absolute decline since early 1989 has raised home mortgage rates relative
to capital market rates.
Conclusions
The market for fixed-rate mortgages is now fully integrated with capital
markets. Since 1986 the share of newly originated conventional conform-
ing fixed-rate mortgages securitized by the Fannie Mae and Freddie Mac
has ranged from 50 percent to 70 percent, and virtually all FHAIVA
mortgages go directly into GNMA mortgage pools. Moreover, empirical
estimation implies that conventional yields adjust fully to changes in capital
market rates within two weeks.
Mortgage rates are currently about what one would expect given capital
market rates. In contrast, conventional rates were a half percentage point
"too low" in the 1970s, owing to thrift tax advantages and portfolio re-
strictions, and a half point "too high" in the 1982 to 1986 period because
thrift profits and portfolio restrictions had effectively disappeared. This
half point "excess" return on mortgages stimulated development and use
of the Freddie Mac and Fannie Mae pass-through programs. Since the
middle of 1987, rates on the conforming FRMs that these agencies can
buy have been "about right", while rates on nonconforming or jumbo
FRMs are about thirty basis points higher. Enactment of user fees on agency
mortgage-backed securities or higher capital requirements for the agen-
cies would likely raise coupons on conforming FRMs.
ARMs have become a major factor in the conventional mortgage market,
with the ARM share of total originations periodically swinging between
a quarter-and-a-half, depending largely on the level and term structure of
interest rates. Borrowers are more likely to choose ARMs the less affordable
is housing (the higher are interest rates) and the more ARMs allow
borrowers to solve the affordability problem (the lower is the ARM rate
relative to the FRM rate). ARM securitization, however, has lagged behind
FRM securitization owing to the lack of both an attractive FHA-ARM
and standardization of conventional ARMs. Introduction of a more
122 THE CHANGING MARKET IN FINANCIAL SERVICES
Notes
1. The agencies also securitize multifamily mortgages. Between 1975 and 1982, 8 percent
to 16 percent of FHA multifamily mortgages were securitized (Seiders, 1983,278).
2. The retail commitment rate and points are those obtained by the Federal Home Loan
Mortgage Corporation in a weekly survey of 125 major lenders conducted since the spring
of 1971.
3. The rates in this table are not adjusted for points, that is, they are the coupon rates
consistent with whatever points were charged. The adjustment would not affect the differ-
ences between actual and perfect rates because the adjustment to both rates would be
identical.
4. The one-year Treasury rate was down to 9 percent in early 1983, versus 8 percent in
the summer of 1990, and the Fannie Mae share price was up from a low of two-and-one-half
in 1981-82. Freddie Mac stock was not freely traded until around year end 1988. Since
then, the share price has moved much like Fannie Mae's.
5. The FRM also reflects the cost of the borrower's default option to the lender.
6. This overstates the cost to the borrower because the actual FRM coupon rate does
not costlessly decline when interest rates fall and this gain to the lender should be priced
in a lower coupon. The same gain means that the cost of the borrower's call option is less
than if the borrower could costlessly refinance.
7. Changes in FRM-ARM rate spreads reflect changes in the Treasury term structure
(seven-year rate less one-year), in the values of the FRM call option and the ARM rate
caps, and in the initial ARM discount. Thrifts increased the attractiveness of ARMs throughout
1987 and 1988 by raising the average initial discount from one-half to three-and-one-half
percentage points (Peek, 1990). With thrifts no longer able to qualify borrowers for larger
loans based upon deep discounts, the initial discount plummeted in 1989 to less than one
percentage point.
8. The 1986 Tax Law has also contributed to these losses, but overbuilding is the pri-
mary culprit.
References
Brueckner, Jan K. and James R. Follain. 1989. "ARMs and the Demand for
Housing." Regional Science and Urban Economics 19(2) (May): 163-187.
Cassidy, Henry J. 1984. "A Review of the Federal Home Loan Bank Board's
Adjustable-Rate Mortgage Regulations and the Current ARM Proposal."
Research Working Paper No. 113. Federal Home Loan Bank Board. (August).
THE MARKET FOR HOME MORTGAGE CREDIT 123
125
126 THE CHANGING MARKET IN FINANCIAL SERVICES
which because of his previous work and that of others has become well
known) but rather where we go from here with regard to the particu-
lar status that the government-sponsored agencies now have in this
environment.
It is unarguable that the success of mortgage market securitization,
especially in the decade of the 1980s, was due to the special status of
sponsored agencies. The view of their obligations as implicitly government
guaranteed, their specialist status in the market, and their expertise in
handling securitization all were critical to the rapid development and
expansion of securitization. These advantages are no longer necessary, or
certainly no longer necessary in the same form for continued development.
Hendershott alludes to the proposals that have been made for capital
requirements and user fees. He quite correctly points out the benefits that
have accrued to stock holders from the existence of the subsidies that are
implicit in agency activity from the government. Additional arguments
for benefits accruing or stemming from the existence of these implicit
subsidies can be made for other players, including bond holders, but also
including managers of these companies. This should not be a goal of
public policy. It strikes me that the time has come to suggest a winnowing,
and significant winnowing away, of the government-sponsored agencies
(Fannie, Freddie) from the kinds of explicit subsidies and implicit guar-
antees that they have available to them. There is no pressing reason to
believe that these institutions could not operate effectively on their own,
after some transitional period, nor is there any reason to believe that the
gains that have been made, which are significant and positive from their
role in integrating the capital and mortgage markets, will be reversed.
Therefore, meeting private capital requirements at a minimum is neces-
sary as a first step. The current proposal of the General Accounting Office
moves in the right direction and logically occurs prior to their being
subjected to bond rating requirements. Hendershott's work here and
elsewhere with various colleagues suggests that integration has reached,
at least with regard to the fixed rate mortgage market, a substantial level,
and will not likely be reversed. I will have more to say about adjustable
rate mortgages and securitization further into my commentary.
Furthermore, I would argue that continuing sponsorship of agencies
has had a negative impact in artificially segmenting the mortgage market
somewhat because of the pressure for conforming limits on conventional
loans. As Hendershott's work at least suggests, although the evidence is
mixed, nonconforming conventional loans seem to bear an interest rate
penalty for exceeding conforming limits. This is not sensible or defensi-
ble, in my judgment, and is primarily because of the political pressures
THE MARKET FOR HOME MORTGAGE CREDIT 127
Introduction
129
130 THE CHANGING MARKET IN FINANCIAL SERVICES
offers it at a pre-set price to the public. A principal result was that, de-
spite evidence of substantial price uncertainty when the offer price is set,
underwriters are able to keep the relative frequency and magnitude of
losses to very low levels. He provided further evidence to suggest that
equity underwriting may not be any riskier than bond underwriting. Armed
with these findings and an analogy between underwriting and bank lending,
Giddy concluded that underwriting stock may not be any riskier than
lending. His conclusions also raise a question about the traditional view
that a principal service provided by the underwriter in a firm-commitment
underwriting is to bear price risk, as developed, for example, in Mandelker
and Raviv (1977). If this were the case, one would expect a large part of
the price risk associated with issuing stock to fall on the underwriter.
However, if other services are principal, such as marketing and distribution
services (Benveniste and Spindt, 1990) or "certification" services (Booth
and Smith, 1986), underwriters may obtain substantial protection against
price risk, thus minimizing the risk-bearing function.
While Giddy is one of the few studies to have estimated equity under-
writing risk per se, his results are not without some conflicting evidence.
Boyd and Graham (1988) have argued from a time series perspective that
investment banking is one of the more risky enterprises in the financial
services industry.! The variance of the stock returns and market f3s of
investment banks are near the high end of the range for financial firms
and much higher than for commercial banks. Because investment banks
are also less highly leveraged than are commercial banks, the stock market
evidence suggests that the variability of the return to investment bank
assets also may be greater. 2
Whether Boyd and Graham's results on the risks of investment banking
actually conflict with Giddy's is unclear. Equity underwriting is a relatively
small part of major investment bank activities and most investment banking
risk may be coming from other activities. 3 However, Giddy's results apply
only to the price risk on individual underwritings and this risk may under-
state the risk associated with equity underwriting over time. We first
reexamine and extend Giddy's results on equity underwriting price risk.
We then consider equity underwriting risk from a temporal perspective
and examine the return variability to equity underwriting over time.
In our reexamination and extension of Giddy, the sample period is
increased and the effects of price risk on underwriting returns are tested
over several subperiods. Tests of the sensitivity of underpricing (setting
the offer price at less than the expected market price) and the underwriter's
spread (the difference between the offer price and the price paid to the
issuer) to the price risk of the issue are also expanded. As a further test
EQUITY UNDERWRITING RISK 131
Table 4-2. Characteristics of Samples of Seasoned and IPO Issues for 1977 to
1979 and 1984 to 1986.
1977-1979 1984-1986
Characteristics Seasoned 1PO Seasoned 1PO
Number of Issues 321 46 694 683
Total Value of Iss. ($ mil.) 9,889.1 263.6 25,887.7 11,694.9
Mean offer size ($ mil.) 30.8 5.7 37.3 17.1
Mean public offer price ($) 20.2 12.6 20.1 9.3
Mean firm size ($ mil.) 1,346.0 17.3 1,957.7 103.6
Mean underwriting spread (% ) 4.8 7.8 5.6 8.0
Std. dev. of underwriting spread 1.9 0.7 2.1 1.4
Source: Securities and Exchange Commission's Registration Offerings Statistics.
offerings, of which 2,580 were seasoned, and 1,641 were IPOs.5 On aver-
age, IPOs are smaller and have lower public offering prices than seasoned
issues. IPOs also tend to be made by smaller firms.6 In addition, under-
writing spreads as a percentage of the offer price are significantly larger
for IPOs than seasoned issues and have less variability.
To investigate the sensitivity of underwriting returns to different mar-
ket conditions, subsamples of common stock issues are also constructed
for the periods 1977 to 1979 and 1984 to 1986. During 1977 to 1979, the
stock market was relatively fiat, with the S&P price index of common
stocks growing by less than 4 percent over the three year period. The
period was also one of low stock price variability. The 1984 to 1986 period
was one of substantial growth, with the S&P 500 stock market index in-
creasing by about 50 percent, and with much greater price volatility than
in the earlier period.
Characteristics of the underwriting data for the two subperiods are
shown in table 4-2. The table shows that during 1977 to 1979 relatively
few issues were made. Our sample consists of only forty-six IPOs and 321
seasoned issues over the three year period. In contrast, 683 IPOs and 694
seasoned issues were made over the three-year period from 1984 to 1986.
The difference in the number of issues and amounts offered between the
two periods may be explained by the incentives of firms to issue equity in
a rising market. Additionally, the average offer size and average firm size
are greater in 1984 to 1986 than in 1977 to 1979, particularly for IPO
issues. Average underwriting spreads and their standard deviations are
also somewhat larger in the latter period.
EQUITY UNDERWRITING RISK 133
Table 4-3a. Percentage Differences between Day-After Market and Offer Prices
for Seasoned and IPO Issues. 1
pected deviation between the market price and its expected value, an
(anticipated) underpricing component. As described above, this compo-
nent is also likely to vary with the issue's risk, ceteris paribus. We do not
model the joint determination of the spread and underpricing. For this
analysis, differences between the realized and the expected market price
are assumed to be the dominant component explaining variation between
the market price and the offer price.
In expanding on Giddy's test, we first regress the underwriting spread(s)
on the absolute value of the percentage difference between the offer price
and the day after market price (AP+ 1), controlling for issue size (IS), firm
size (FS), and (through separate regressions) whether the offering is IPQ
or seasoned. Issue size and firm size are control variables because they
may be inversely correlated with price uncertainty and positively related
to underwriting costs (Pugel and White 1985, 119). Thus, any relation
between spreads and our measure of price uncertainty cannot be interpreted
unambiguously if firm or issue size are omitted from the regression.
The regression results presented in table 4-4a provide strong evidence
that underwriting spreads contain a risk premium that increases with the
price uncertainty of the issue. For both the seasoned and IPQ equations,
the coefficients on the absolute value of the percentage change in price
are positive and highly significant. The negative coefficients on issue size
suggest that there are fixed costs, and hence scale economies, associated
with equity underwriting.
More directly related to the underwriter's price risk is whether the
underwriting spread tends to be larger when the market price is below the
offer price. Table 4-4b presents regression results based on the subsample
of issues for which the day after market price fell below the offer price.
Again, for both the seasoned issues and IPQs, the coefficients on price
uncertainty, measured by the absolute percentage price difference, are
highly significant. These results indicate that underwriting spreads reduce
the downside price risk of underwriting and help to stabilize the under-
writer's return.
Risk of Loss on Equity Underwriting
All Issues
IPO: s = 8.10 + .016*AP+1 - .024*IS + 9.84E-5*FS
(160.3) (7.46) (-13.07) (1.66)
Seasoned: s = 5.58 + .069*AP+1 - .011 *IS + 9.30E-6*FS R2 = .20
(121.3) (14.78) (-15.59) (1.64)
Issues with Negative Day-After Market and Offer Price Differences
IPO: s = 8.06 + .085* AP+1 - .040*IS - .002*FS R2 = .24
(57.7) (7.36) (-5.80) (-1.84)
Seasoned: s = 5.33 + .134*AP+1 - .010*IS - 1.45E-6*FS R2 = .22
(65.8) (10.73) (-9.56) (-0.21)
1 t-statistics in parentheses. Spread (s) and absolute price differences (AP+l) are per-
centages. Issue size (IS) and firm size (FS) are in millions of dollars.
= Q if p < 0). Since part of an issue is usually sold at the offer price even
when the market price is below the offer price, this assumption may
substantially understate the underwriter's returns on poorly received
issues. On the other side, the use of the gross underwriting spread, which
does not net out the underwriter's expenses, overstates the net return on
an issue.
Giddy made three hypothetical calculations of the underwriter's gross
dollar return to each underwritten issue, depending on market conditions
one, five, and ten days after the offer date. Our calculations are returns
per dollar of the amount offered. Using the return notation in equation
(1) and the assumption that Qp = Q if p < 0, our three calculations are as
follows:
1 P+1 ~ 0 s
otherwise P+1 + s
2 P+l or P+s ~ 0 s
otherwise P+s + s
3 P+l' P+s, or P+lO ~ 0 s
otherwise P+lO + s
EQUITY UNDERWRITING RISK 139
Table 4-5. Hypothetical Underwriter Returns for Seasoned and IPO Common
Stock Issues for 1977 to 1987 (percent).
where P+i is the value for P using the closing market price i days after the
offering date.
These three calculations are substitute estimates of an underwriter's
return on an issue. The first calculation assumes that, if the market price
one day after the offer date is less than the offer price, the underwriter
sells the entire issue at the market price one day after the offering. The
next two calculations assume that, if the market price initially falls below
the offer price, the underwriter will hold onto the issue for a limited
number of days before selling the issue at the lower of the offer or market
price. The assumption of a five or ten day holding period is arbitrary and
simply a rough approximation to the gradual unloading of an issue that
is getting a poor reception.
Underwriter returns calculated for seasoned issues and IPOs separately
for our full sample are presented in table 4-5. The mean returns, reported
in the first row, equal the mean underwriting spreads (table 4-1) plus the
mean unconditional expected losses from selling at a market price below
the offer price (table 4-3b). The calculations show that the mean under-
writing returns are positive and significant for both seasoned issues and
IPOs, with somewhat higher average returns on IPOs. The standard de-
viation of returns for IPOs is only moderately higher than that for
seasoned issues, even though the standard deviation of the differences
between market and offer prices, shown in table 4-3a, is much higher for
IPOs. When estimated returns are negative using day-after market prices,
the average return is moderately lower for IPOs, -9.07 percent, as com-
pared to -6.26 percent for seasoned issues (not shown in the table).
In the lower half of table 4-5, the relative frequency of estimated losses
140 THE CHANGING MARKET IN FINANCIAL SERVICES
using next day market prices is seen to be about 6 percent for seasoned
issues and 8 percent for IPOs. These estimated loss probabilities are about
double those reported by Giddy (tables 6.4 and 6.6). Nonetheless, the
probabilities are still low and contrast sharply with the 40 percent and 30
percent probabilities for market prices falling below offering prices for
seasoned issues and IPOs, respectively, shown in table 4-3a. These loss
probabilities based on next day prices indicate that, for most issues, market
price declines are within the range of the underwriting spread, despite the
presence of very substantial price uncertainty. Further, the probability of
loss is only moderately greater for IPOs than that for seasoned issues,
despite the much greater price uncertainty of IPOs.
As the length of the holding period increases, average underwriting
returns for both types of issues decline, the standard deviations increase,
and the loss probabilities increase. These results may simply reflect the
random nature of stock price changes and the upper bound on the under-
writer's return. However, the results may also be influenced by under-
writer attempts to stabilize prices early in the offering period by holding
on to part of an issue getting a poor reception. To the extent that this
occurs, the estimated returns using the day-after prices are upward biased
and the risk of loss is understated. On the other side, the estimated returns
tend to be overstated by not recognizing the partial selling of an issue at
the offer price when the market price is less than the offer price.
Estimated returns are presented separately for the sub-periods 1977 to
1979 and 1984 to 1986 in table 4-6. There are no substantial differences
in returns across the periods, although the number of offerings differ
substantially. The results show, however, slightly greater return variability
and loss frequencies in the latter period as compared to the former. Also,
the mean returns, conditioned on a negative return, were lower (more
negative) in the latter period for both IPOs and seasoned issues (not
shown). This is consistent with the greater market price uncertainty in the
latter period shown in table 4-3a. On the whole, however, these results do
not indicate a lot of sensitivity of underwriting returns to general market
conditions. For both periods, the estimated loss frequencies are moderate
and are in the same general range as the frequencies previously reported
for the longer sample period.
The previous tests provide evidence that price risk is modest for firm-
commitment equity underwriting. These tests required assumptions about
EQUITY UNDERWRITING RISK 141
Table 4-6a, b. Hypothetical Underwriter Returns for Seasoned and IPO Issues
1977 to 1979 and 1984 to 1986 (percent).
1977-79
Seasoned Issues (N = 321) IPO Issues (N = 46)
1 day 5 days 10 days 1 day 5 days 10 days
Mean 3.8* 3.2* 2.9* 6.3* 5.6* 5.5*
Std. dey. 2.3 3.2 3.9 3.3 5.3 6.1
Range -15,10 -15,10 -16,10 -8,10 -14,10 -19,10
Percent of
Gains 97.2 88.5 84.7 93.5 87.0 91.3
Losses 2.8 11.2 14.3 4.3 13.0 8.7
Breakevens 0.0 0.3 0.9 2.2 0.0 0.0
1984-86
Seasoned Issues (N = 694) IPO Issues (N = 683)
1 day 5 days 10 days 1 day 5 days 10 days
Mean 4.5* 4.2* 3.9* 5.8* 5.4* 5.0*
Std. dey. 3.1 3.8 4.8 4.9 5.7 6.5
Range -26,18 -26,18 -31,18 -28,16 -28,16 -32,16
Percent of
Gains 96.0 90.8 88.3 91.5 87.6 84.9
Losses 3.9 8.9 11.5 7.8 11.9 14.2
Breakevens 0.1 0.3 0.1 0.7 0.6 0.9
* Significantly different from zero at 1 percent.
the amount sold at the market price and how quickly the issue was dis-
posed of when the market price was less than the offer price. An alter-
native test of the importance of price risk is to test whether the occurrence
of such events has a significantly unfavorable effect on the investment
bank's stock return. Such an effect may occur if losses on average tend to
be much larger than the calculations in the last section suggest.
For the purpose of this test, daily stock returns were taken for four
investment banks, E.F. Hutton, Merrill Lynch, and Paine Webber over
the eleven-year period and First Boston since mid-1983P For each of the
individual banks, the daily returns were regressed on a market return
index and "low return" variables for seasoned and IPO underwritings in
which the bank was the lead underwriter (information where the bank
142 THE CHANGING MARKET IN FINANCIAL SERVICES
The results presented here support the findings and conclusions of Giddy
concerning the significance of price risk on stock offerings. We find strong
evidence that underpricing and underwriting spreads are sensitive to price
risk which enable underwriters to translate substantial price uncertainty
into small loss probabilities. Also, estimated loss probabilities do not appear
to be very sensitive to general market conditions, as indicated by the
results from our subsamples. Furthermore, in the infrequent instances
when they occur, low underwriting returns or losses do not appear, on
average, to have very significant effects on the profits of large investment
banks.
These results do not imply that there is no potential for large losses.
Underwriters cannot entirely eliminate their exposure to sharp declines
in market prices and there is the element of choice in determining risk
exposure. Vulnerability to market declines is illustrated by a number of
estimated underwriting losses in the week just prior to the October 19,
1987 market crashP These estimated losses produced an estimated nega-
tive dollar return on seasoned equity underwriting for the fourth quarter
Table 4-7. Effects of Low Returns to Equity Underwritings on Investment Bank
Stock Returns. 1
Coefficients
Number of
Seas IPO Low Returns
Market Low Low
Firm Constant Return 2 Return Return R2 Seas IPO
a. Day-of-Offering Stock Returns 3
E.F. .0005 1.009 -.88E-5 .12E-5 .11 28 6
Hutton (1.03) (18.83) (-.85) (.09)
First -.0004 Ll42 .02E-5 1.70E-5 .34 13 3
Boston (-.81) (24.52) (.09) (2.25)
Merrill .0001 Ll29 .02E-5 -.06E-5 .17 110 15
Lynch (.26) (23.92) (1.60) (-1.30)
Paine- .0003 1.098 .02E-5 .04E-5 .12 13 4
Webber (.67) (19.82) (.59) (.74)
b. Day-After Stock Returns 4
E. F. .0005 1.010 -.05E-5 .06E-5 .11
Hutton (1.06) (18.85) (-.55) (.43)
First -.0004 Ll42 .07E-5 .08E-5 .34
Boston (-.89) (24.45) (.28) (.11)
Merrill .0001 Ll31 .01£-5 .05E-5 .17
Lynch (.26) (23.92) (.97) (LlO)
Paine .0003 1.099 -.38E-5 .57E-5 .12
Webber (.64) (19.86) (-.12) (1.20)
1 t-statistics in parentheses.
2 The market return index is the CRSP value weighted index of returns from the NYSE
and AMEX exchanges.
3 Sample size is 2,780 for all investment banks except First Boston which has a sample
size of 1,174.
4 Sample size is 2,779 for all investment banks except First Boston which has a sample
size of 1,173.
144 THE CHANGING MARKET IN FINANCIAL SERVICES
R= "",N R (2)
~n=l n
where (J2 = V(Rn), r = E(Rn), and the last expectation on the rhs of (3)
is taken over the distribution of N.
Three sources of variability in the returns to underwriting over time
appear in equation (3). One is the variance of the return to individual
underwritings, (J2. This variance will depend on the occurrence of losses or
low returns on individual underwritings due to market price uncertainty,
given the dollar size of the underwriting and the underwriting spread. It
will also depend on (unexpected) variation in the dollar size of underwritings
and underwriting spreads. Implicit in the risk analysis of the last section
was the assumption that the size of the underwriting and underwriting
spread were nonstochastic so that their variances did not contribute to
underwriting return uncertainty.
A second source of variance in underwriting returns over time is the
variance in the number of underwritings, V(N). This source also is not
captured when studying returns to individual underwritings. The third
source, which is also not recognized, is the (expected) sum of the covariances
between returns on individual underwritings within the time unit. These
covariances depend on covariances across offerings between the market
prices of individual issues, the size of offerings, and spreads. The under-
writing return covariances may be positive or negative. 20
Firm-commitment
Best Best Efforts plus
Common Preferred Total Efforts Firm-commitment
75 58 74 67 72
and seasoned issues. The time series variances are decomposed into the
three components identified in equation (3). Results are presented in
table 4-8a.
The column in table 4-8a labeled Number of Offerings is the percent-
age of the quarterly dollar underwriting return variance attributable to
the variance in the number of offerings as indicated in equation (3). For
the one-day-after market prices, 70 percent or more of the variance comes
from the variance in the number of offerings per quarter, and slightly less
for the five-day-after results. As shown in the next column, the variance
in returns to individual underwritings contributes less than 15 percent to
EQUITY UNDERWRITING RISK 147
The results just presented suggest that factors other than price risk,
particularly the number of offerings, are the major determinants of return
variability to stock underwriting. Previously, we considered the sensitivity
of investment bank stock returns to only price risk. Here we consider the
sensitivity of investment bank stock returns to variability in underwriting
returns over time.
If unanticipated movements in the bank's underwriting returns over
time have an important effect on the bank's profits, they should be signifi-
cantly related to the bank's stock returns. However, variation in equity
underwriting returns may not be important because the variability is small
in relation to other activities, or because of diversification effects. These
effects would be enhanced if investment banks were able to shift resources
between equity underwriting and other activities in response to exogenous
influences on the returns to anyone activity.
Table 4-9a presents two sets of correlations between quarterly equity
underwriting returns and quarterly returns to the stocks of our sample of
four investment banks. In one set of correlations, the equity underwriting
returns (underwriting-specific) are estimated from issues (seasoned, IPO,
and preferred) for which the respective banks were lead underwriters. 25
We have no data on the returns to underwriting in which the investment
banks participated but were not lead underwriters. For the other set,
equity underwriting returns (aggregate) are estimated from underwritten
issues (seasoned, IPO, and preferred) for all investment banks. These
aggregate returns would proxy for the bank's equity underwriting returns
if, as lead and participating underwriter, the bank earned a pro rata share
of the aggregate returns.
For the four investment banks, the correlation coefficients in table 4-
9a are mostly small and negative. Except for Merrill Lynch, the correla-
EQUITY UNDERWRITING RISK 149
* Significantly different from zero at 5 percent. Significance not computed for averages.
1 Sample size is 40 for all investment banks except First Boston which has a sample size
of 18.
2 Underwriting returns are estimated quarterly returns for which investment bank was
lead underwriter.
Coefficients
Constant Market Return Underwriting Return
Underwriter-specific Common Stock Underwriting Returns
-.021 2.019 .205E-8 .65
~~~ ~n) (~~
Underwriter-specific Common + Preferred Stock Underwriting Returns
-.023 2.058 .307E-9 .64
(-.93) (7.66) (.14)
Aggregate Common Stock Underwriting Returns
-.023 2.060 .365E-1O .64
(-.92) (7.15) (.14)
Aggregate Common + Preferred Stock Underwriting Returns
-.021 2.009 .195E-9 .64
~~~ ~~ (~)
to explain variation in the investment banks' stock returns, our tests can-
not separate any such effects from general market effects.
Conclusions
The results of our study reinforce the evidence and conclusions of Giddy
(1985) concerning price risk in equity underwriting. We find strong evi-
dence that underpricing and underwriting spreads are sensitive to price
risk. This evidence is consistent with further results that the risk of loss
on firm-commitment underwritings is small despite the presence of sub-
stantial market price uncertainty when the offer price is set. These results
are not sensitive to the sample period. In reformulating the underwriter's
risk in terms of a quarterly time series variance of equity underwriting
returns, we find that the major determinant of this time series variance is
the variance in the number of offerings. The return variance on individual
issues (which depends on uncertainty about the market price, the amount
to be underwritten, and the underwriting spread) and return covariances
between issues account for only a minor share of the time series return
variance.
The time series results could help to reconcile Giddy's conclusions that
equity underwriting has relatively low risk with the time series evidence
on the relatively high risk of investment banking. That is, "business risk,"
such as the frequency with which underwritings are done, may have a
much greater effect on the variance of underwriting returns than price or
"market risk", although the two types of risk need not be independent. In
this case, the variance of returns to equity underwriting over time, which
depends on both types of risk, might still be important in accounting for
variation in investment bank profits and hence stock returns. However,
we did not find consistent evidence that variability in equity underwriting
returns over time had a significant effect on the investment bank's stock
returns.
Giddy's and our results, which suggest that underwriting equity may
involve relatively small price risk, raises doubts about the importance
of risk-bearing by the underwriter as a motive for firm-commitment
underwritings. Recent literature has emphasized other motives for stock
underwriting. Baron and Holmstrom (1980), Baron (1982), and Benveniste
and Spindt (1990) explain underwriting and firm-commitment contracts in
terms of marketing and distribution services where investors have infor-
mation advantages over issuers on the demand for new issues. Booth and
Smith (1986) also suggest that underwriters provide a certification service
152 THE CHANGING MARKET IN FINANCIAL SERVICES
Appendix A
R ==
E(Rn) == r
r:=l Rn
V(Rn) == (J2
The variance of R appearing in equation (3) of the text is determined
as follows:
V(R) = EN[V(R)IN] + VN[E(R)IN]
Appendix B
Under the assumption that the number of underwritings and the return
per underwriting are independent, the residual component in table 4-8, as
shown in equation (3), is
Notes
1. Kwast (1989) also reported that the standard deviation of asset returns across trading
departments of commercial banks, which would include all bank securities activities, was
much higher than that for the banks' return on assets excluding the trading department.
Benston (1990) provides a different view of the relative risks of commercial and investment
banking.
2. Allowance has to be made for leverage because of its positive effect on the firm's
equity return variance and market /3. The asset return standard deviation is important to
creditors (or a deposit insurer) concerned with the risk of default, as well as to stockholders.
In option based models of corporate debt yields, such as Merton (1984), the default premium is
an increasing function of the firm's asset return standard deviation.
3. Giddy (1985, pp. 166- 6) reports that underwriting revenue for NYSE firms doing a public
business averaged about 10 percent of total revenue between 1974 and 1982. Equity under-
writing revenue would be a smaller percentage.
EQUITY UNDERWRITING RISK 155
4. The issuer, date of issue, public offering price, gross proceeds, underwriter spread,
managing underwriter, and total assets of the issuing firm are from the SEC Registration
Offerings Statistics (ROS) data. Market prices for the issues subsequent to the issue date are
from the CRSP data.
5. These issues account for 85 percent of the total amount of seasoned issues and 68
percent of IPOs registered with the SEC. All the issues could not be included because of
either the unavailability of CUSIP numbers or stock market price data for one, five, and ten
days after the issue date.
6. The mean firm size reported is based on the 86 percent of firms that reported total
assets.
7. Benveniste and Spindt (1990) also show that underpricing will increase the amount of
the issue that underwriters are able to informally pre-sell to "regular" customers.
8. Tinic (1988) has emphasized that underpricing provides protection against litigation
risk from dissatisfied investors. Litigation risks are not considered here.
9. While there is no allowance for a normal return between the time of the offer price
and when the subsequent market price is observed, using returns in excess of a normal
return would have no appreciable effect on the results. The one-day return for the CRSP
value-weighted market return index over our sample period is .05 percent.
10. The effects of underpricing will not be fully captured by the average underpricing
if underpricing on individual issues is positively related to the uncertainty of the market
price of the issue. Possibly greater price stabilization for IPOs poorly received by the market
also could be a factor reducing the lower end of the range for IPOs as compared to that for
seasoned issues (see Giddy, p. 155).
11. Logue and Lindvall (1974) found the underwriting spread was negatively related to
the issuing firm's sales which they interpreted as proxying for price uncertainty. Pugel and
White (1985) did not find a significant relation between the underwriting spread and the
mean square change in an over-the-counter stock index for the days preceding the offering,
which was their measure of price uncertainty. Booth and Smith (1986) did not find a relation
between underwriting spreads and the variance of the systematic component of the stock's
monthly returns prior to the offering. However, they did get some indication of a positive
relation between the spread and the ratio of the idiosyncratic and systematic variances.
12. Our sample was limited by the small number of investment banks who are frequent
underwriters of equity and met our criterion of having publicly traded stock over a substan-
tial part of our eleven-year sample period.
13. This is an alternative form of the "event study" model. It is recommended by Acharya
(1989) when the firm experiences repetitive events. For another application, see Malatesta
and Thompson (1985).
14. Even though the price risk variable uses the day-after market price, the market may
have information on the market's reception of the offering on the day of the offering.
15. The estimated market f3s using daily returns are considerably lower than those re-
ported for investment banks by Boyd and Graham (1988) (an average of 1.69 for securities
firms using annual return data). Estimated market f3s using daily returns tend to be biased
because of non-synchronous trading that arise when stocks are not traded every day (see
Brown and Warner (1985». When 5-day returns are used, the estimated market f3s are: First
Boston 1.67, Hutton 1.57, Merrill Lynch 1.46, and Paine Webber 1.72.
16. Correlations between "low return" variables and the market return are low, suggest-
ing that low returns are not explained primarily by the overall market return. Thus,
multicollinearity is not an issue in interpreting the insignificant coefficients on the low return
variables.
156 THE CHANGING MARKET IN FINANCIAL SERVICES
17. Since our underwriting data includes only issues of domestic corporations, our esti-
mated losses do not include those experienced by U.S. underwriters of British Petroleum
stock in late October 1987 which were reported to be in the millions of dollars. While these
losses are directly attributable to the stock market crash, they reflect extenuating cir-
cumstances concerning differences in underwriting practices between the United King-
dom and the United States and the willingness of the U.S. underwriters to take a large risk
exposure. Of particular importance was the setting of a firm-commitment offer price more
than two weeks before the offering date (and before the market crash).
18. The moments of R, N, and Rn should be conditioned on information available at the
time the expectations are formed. In general, the conditional expectations will be time
variant. In this presentation, we ignore explicit conditioning and time variation in the mo-
ments of the distributions.
19. This assumption is considered below.
20. For example, returns to underwritings offered on the same day may have a positive
covariance if returns to the underlying stocks have positive market f3s or if the amounts being
underwritten on individual issues have positive serial correlation. Alternatively, if the size
distribution of underwritings in a given period was fixed (e.g., there were a fixed number of
small underwritings and a fixed number of large underwritings) but the sequence of
underwritings of different amounts was stochastic, the sum of the covariances would be
negative, ceteris paribus.
21. The estimated returns are for firm-commitment underwritings of common stock.
Results are reported below for underwritings that include preferred stock and best efforts
offerings.
22. Variances were also calculated using annual returns. As would be expected, the
annual return variances were much higher than the quarterly return variances. However, the
relative importance of the different sources of the return variance was similar to that pre-
sented below for the quarterly returns.
23. Returns to firm-commitment preferred stock offerings and to best efforts offerings
contain only the dollar underwriting spreads earned on the offerings. Thus, for firm-
commitment preferred underwriting, there is no allowance for losses due to unfavorable
market prices. The evidence for common stock underwritings that losses due to unfavorable
market prices are infrequent and relatively small can be expected to carryover to preferred
stock offerings as well.
24. For an individual underwriter, uncertainty about the volume of underwriting may be
even greater than indicated by the variance in the quarterly aggregate underwriting volume
because of unanticipated changes in its market share.
25. Returns to best efforts offerings were not included because the four underwriting
firms were agents for only a few of these offerings.
26. Over the 11 year sample period, Merrill Lynch was the largest underwriter of sea-
soned issues of stock. While its presence in this market was greater than that of other
underwriters, the fraction of its annual revenues from all types of underwriting appeared to
be commensurate with those for other large investment banks. Annual reports show that
investment banking income for Merrill Lynch, which included merger activity as well as
underwriting, accounted for 10.4 percent of its gross revenues over the period 1983-87.
27. Underwriting returns were regressed on the returns over the previous four quarters.
Significant coefficients were almost always confined to the first and second lags.
28. There was substantial serial correlation in the aggregate underwriting returns but
very little in the bank -specific returns.
29. Because of multiple underwritings on some days and variation in the number in days
EQUITY UNDERWRITING RISK 157
between other underwritings, j - i is not a constant time interval. It serves only as a rough
measure of the temporal "closeness" of individual underwritings.
30. As the "distance" was expanded further, the average correlations and covariances
declined, indicating that the correlations and covariances tended to zero for large "dis-
tances." This result held for returns to both IPO and seasoned offerings.
References
Merton, Robert C. 1974. "On the Pricing of Corporate Debt: The Risk Structure
of Interest Rates," Journal of Finance 29 (May): 449-70.
Pugel, Thomas A. and Lawrence J. White. 1985. "An Analysis of the Competitive
Effects of Bank Affiliates to Underwrite Corporate Securities." In Deregulating
Wall Street, Ingo Walter, Editor. New York: John Wiley & Sons: 93-139.
Smith, Clifford W. 1986. "Investment Banking and the Capital Acquisition Process."
Journal of Financial Economics 15: 3-29.
Tinie, Seha M. 1988. "Anatomy of Initial Public Offerings of Common Stock."
Journal of Finance 43 (September): 789-822.
III
5 THE COMPETITIVE IMPACT OF
FOREIGN COMMERCIAL BANKS IN
THE UNITED STATES
Lawrence G. Goldberg
Foreign commercial banks have grown rapidly in the United States in the
1980s and have had a significant competitive impact. The growing role of
foreign banks in domestic markets has stimulated claims of unfair com-
petition. Banking is experiencing the same apprehension that has arisen
from foreign entry into other industries. This article reviews the history
and status of foreign banks in the United States. Reasons for the growth
of foreign banks are closely related to their competitive impact; therefore,
the paper summarizes studies that have analyzed the motivation for growth
of foreign banks in the United States. Because foreign banks can employ
various organizational forms, have entered different geographic areas
unevenly, have originated from different countries at different rates, and
have faced a changing legal and economic environment, an extensive dis-
cussion of the institutional framework within which the foreign banks
operate is provided. Through the use of descriptive material, the review
of past empirical studies and the development of an original empirical
analysis, this article will attempt to assess the competitive impact to date
and the likely future role of foreign banks in the American markets.
Foreign banks have grown much more rapidly in the United States
than American banks have grown abroad. In earlier years, U.S. banks had
far greater assets abroad than foreign banks had in the United States.
In 1955 the gross assets of U.S. bank branches abroad were only $2 billion,
161
162 THE CHANGING MARKET IN FINANCIAL SERVICES
Even though this article is unable to answer all of these questions com-
pletely, it does provide a useful start for the competitive analysis of foreign
banks.
The first section discusses the institutional setting of foreign banks in
the United States, including the types of foreign banking organizations,
the legal framework, and the extent of the foreign bank presence in the
United States. The second section reviews empirical studies of the factors
that have affected the overall growth of foreign banks, the growth by
country, and growth into various states. The growth of Japanese banks is
also reviewed. In the third section the descriptive evidence of competitive
effects is presented and the allegations of unfair competitive advantages
evaluated. Balance sheet ratios of foreign banks are compared to domestic
banks in the fourth section in order to determine in which activities the
foreign banks are concentrating their efforts. The final section summarizes
the evidence with respect to the competitive impact of foreign banks and
assesses the prospects for the future for foreign banks, including an
evaluation of the changes proposed for Europe in 1992.
The first foreign bank to establish an office in the United States was the
Bank of Montreal, which set up an agency in New York in 1818. Other
well known banks that established offices early on were: Hong Kong and
Shanghai Banking Corporation (1875), Lloyds Bank International (1856),
Barclay's Bank International (1890), and Mitsubishi Bank (1920). These
banks engaged primarily in trade finance and funds transfers and par-
ticipated in the New York stock and bond markets. One study has esti-
mated that through 1985, "of foreign commercial banks currently operating
in the United States 74 per cent (224 banks from sixty countries) first
established their U.S. presence after 1970."4 From table 5-1 it can be
seen that only twenty-seven banks from twelve countries were operat-
ing in the United States prior to 1946; of these, seven were British,
five Canadian, three Italian, and three Swiss. 5 Furthermore, the data in
table 5-2 indicate that most of the growth of foreign banks in the United
States has occurred since 1972. The ratio of foreign controlled to do-
mestically controlled assets has risen dramatically from 3.6 percent in
1972 to 21.4 percent in 1989. Every year has witnessed an increase in the
share of foreign banks. These data by themselves would suggest that for-
eign banks are attracting relatively more business and placing competitive
164 THE CHANGING MARKET IN FINANCIAL SERVICES
Number of Number of
Period Foreign Banks Home Countries
Before 1946 27 12
1946-1970 48 18
1971-1985 229 60
Unclassified entry year 2 2
Source: Cho, Krishnan, and Nigh (1987) 61.
pressure upon domestic banks. The areas where the competitive pressure
has been most intense are examined later in the article.
Foreign banks can utilize several different organizational forms to operate
in the United States. There have been changes in the relative advantages
and disadvantages in each form over time, and consequently, the rate of
growth of the usage of each form has varied. Since different forms permit
different types of activities to be performed and require different levels
of commitment by the parent organization, the choices enable foreign
banks to tailor their American operations to their business desires.
The most limited, but the easiest to establish of the organizational
forms, is the representative office. These offices neither take deposits nor
make loans, but can act as agents for the foreign bank and forward payments
or loan documents to the home office. Representative offices are often
established as a precursor for further involvement, and one study has
calculated that 60 percent of foreign banks through 1985 used the rep-
resentative office as their initial entree into the American market. 6
Agencies are an integral part of foreign banks and may make commercial
and industrial loans; however, they cannot make consumer loans nor accept
deposits. They do maintain credit balances that are very similar to deposits,
but most payments cannot be made from these accounts. Funding is from
the parent bank or by borrowing in the Federal Funds or interbank markets.
The branch is the most important organizational form and is an integral
part of the parent bank. It can offer a full range of services and, like the
agency, it has traditionally been engaged mostly in wholesale operations.
The final major form is the subsidiary or commercial bank. Subsidiar-
ies have identical powers as domestic banks and are regulated in the same
manner. Many are oriented toward retail business. Foreign banks can
establish subsidiaries either through acquisition or de novo entry. The
acquisition by foreign banks of several large domestic banks, such as
Marine Midland and Crocker in 1979 and 1980, alarmed many people and
Table 5-2. Total U.S. Assets of Foreign-Controlled U.S. Banking Offices, 1972 to 1989 (Billions of Dollars).
u.s. Banking Offices
Owned by Foreign Banks Memo: Ratio of
u.s. Banks Foreign-Controlled
U.S. Branches Commercial Foreign to Total Domestic
Year and Agencies Banks Other * Individuals Total Banking Assests**
1972 22.2 4.4 1.1 0.6 28.3 3.6
1973 25.2 5.4 1.5 1.0 33.1 3.7
1974 34.0 10.8 1.9 0.6 47.3 4.8
1975 38.2 11.8 2.0 2.1 54.1 5.3
1976 45.7 13.8 1.5 2.7 63.7 5.9
1977 59.1 16.2 1.6 4.9 81.8 6.7
1978 86.8 20.7 2.0 6.4 115.9 8.4
1979 113.5 34.6 2.4 7.7 158.2 10.3
1980 148.3 68.1 2.8 10.2 229.4 13.5
1981 172.6 78.5 3.2 11.5 265.8 14.2
1982 208.2 90.9 3.9 17.5 320.5 15.3
1983 229.0 78.5 4.2 19.7 352.2 15.4
1984 273.2 103.1 4.5 19.8 400.6 16.1
1985 312.4 111.3 5.4 23.0 452.1 16.5
1986 398.1 109.5 5.3 27.0 539.9 17.9
1987 462.7 112.8 6.1 28.7 610.3 19.8
1988 515.3 123.7 6.6 28.3 673.9 20.6
1989 581.3 134.1 6.5 29.1 751.0 21.4
* Includes N.Y. Investment corporations and directly owned Edge corporations.
** Total domestic banking assets include the assets of domestic offices of insured commercial banks plus those of U.S. branches and
agencies of foreign banks. Consequently, in this table they also include balances booked in IBFs.
Source: Houpt (1988), p. 25, and additional information from J. Houpt. Original data were obtained from call reports.
166 THE CHANGING MARKET IN FINANCIAL SERVICES
.....
$
......
-..l
0
Table 5-4. Total Assets by Country of the U.S. Offices of Foreign Banks Listed by Country All Countries (1988). (Assets
in Thousands of Dollars).
Country Assets Country Assets Country Assets
Austria 1,197,628 Canada 39,252,400 Iran 17,266
Belgium 3,131,391 Argentina 864,630 Israel 9,249,139
Denmark 2,047,792 Brazil 4,282,035 Japan 360,853,869
Finland 1,419,688 Chile 85,969 Jordan 316,177
France 29,255,588 Colombia 825,477 Korea, South 4,616,861
Germany (FDR) 13,024,980 Dominican Rep. 29,154 Kuwait 993,404
Greece 1,224,959 Ecuador 187,652 Malaysia 525,101
Ireland 6,425,407 EI Salvador 874 Pakistan 669,442
Italy 29,859,425 Mexico 4,431,142 Phillipines 321,381
Luxembourg 756,076 Panama 136,424 Qatar 116,600
Netherlands 10,126,495 Peru 30,756 Saudi Arabia 743,798
Norway 1,153,410 Uruguay 214,022 Singapore 380,246
Portugal 658,407 Venezuela 2,602,703 China, Rep. of 2,744,694
Spain 6,085,742 Bermuda 137,868 Thailand 824,274
Sweden 1,210,392 Cayman Islands 130,387 United Arab Emir. 257,735
Switzerland 23,879,249 Bahrain 1,027,948 Egypt 174,952
Turkey 310,421 China, Peoples Rep. 964,112 Australia 3,930,744
U.K. 32,708,605 Hong Kong 25,732,010 New Zealand 892,436
Yugoslavia 655,590 India 750,346 Multiple Countr. 936,426
Other W. Europe 5,044,752 Indonesia 2,665,731
Grand TotaL for aLL Countries: $653,092,182
Ten Largest Countries 1980 1981 1982 1983 1984 1985 1986 1987 1988
Canada 15,718,428 21,503,472 22,148,288 25,838,300 38,101,415 39,553,299 42,431,042 44,244,905 39,252,400
France 12,925,772 16,914,185 15,736,972 16,101,115 18,384,141 20,653,757 20,653,757 24,568,815 29,255,588
Gennany, West
(Federal Republic) 7,253,180 7,379,767 8,880,866 7,382,739 7,565,605 8,801,212 11,049,175 13,511,074 13,024,980
Hong Kong 11,920,682 12,983,516 16,707,658 19,705,400 17,288,573 23,376,817 24,920,412 25,561,199 25,732,010
Israel 4,097,400 4,239,426 6,084,051 7,096,780 7,863,308 7,812,780 8,074,245 8,349,632 9,249,139
Italy 9,216,825 10,891,512 14,718,455 17,523,890 23,931,971 29,090,071 36,445,780 41,014,763 39,859,425
Japan 72,484,137 88,646,854 113,005,182 125,982,961 151,259,221 178,761,678 245,571,205 294,499,442 360,853,869
Netherlands 36,681,150 4,813,409 5,284,834 4,894,644 5,335,755 7,134,185 8,539,678 8,684,287 10,126,495
Switzerland 11,312,568 11,225,989 12,929,098 13,215,445 15,280,357 18,338,243 24,518,535 27,957,911 23,879,249
United Kingdom 25,136,319 46,445,237 52,171,252 53,058,135 51,443,593 57,170,899 40,631,292 43,685,868 32,708,605
Totals:
Ten Countries 173,733,461 225,043,367 267,666,656 290,799,409 336,453,939 390,692,941 422,835,121 532,077,896 583,941,760
All Countries 198,115,287 251,217,538 301,021,581 333,336,188 378,313,617 441,525,633 526,589,545 594,077,896 653,092,182
......
-..J
......
......
-.I
N
Bank of Tokyo Ltd. (a) $12,821 $47,210 $25,544 11 +41.67 $9.050 $32,215 $18,070 11 +37.2
2 4 National Westminster Bank Pic,
London 9,595 22,189 17,281 6 +16.66 7,523 17,742 14,799 7 +17.7
3 3 Mitsubishi Bank Ltd., Tokyo 8,357 31,742 15,681 5 +5.9 8,071 32,881 15,689 9 +52.2
4 2 Sumitomo Bank Ltd., Osaka 7,914 24,472 16,551 7 -1.45 8,251 22,722 14,040 8 +64.8
5 7 Industrial Bank of Japan Ltd., Tokyo 7,886 28,710 17,700 7 +31.69 5,988 25,625 12,024 7 +34.8
6 6 Sanwa Bank Ltd., Osaka 7,859 28,763 15,190 8 +13.98 6,895 25,228 13,628 8 +63.8
7 10 Fuji Bank Ltd., Tokyo 7,838 30,372 13,009 8 +40.07 5,595 29,124 12,514 8 +14.2
8 9 Dai-ichi Kangyo Bank Ltd., Tokyo 6,960 33,913 16,070 5 +20.83 5,761 29,361 13,588 5 +23.2
9 5 Hongkong & Shanghai Banking Corp. 6,457 25,380 18,490 18 +9.03 7,104 26,762 19,266 21 -4.9
10 11 Bank of Montreal 5,409 18,936 9,667 16 +0.64 5,374 18,725 10,881 14 +25.5
11 12 Long-Term Credit Bank of Japan,
Ltd., Tokyo 5,223 10,543 3,664 4 +29.62 4,030 8,514 2,616 3 +60.5
12 15 Tokai Bank, Ltd., Nagoya 4,066 17,663 7,729 4 +27.64 3,185 16,684 6,390 4 +35.0
13 13 Mitsui Bank Ltd., Tokyo 3,971 11,209 6,254 4 +6.68 3,722 9,680 2,648 4 +14.0
14 8 Swiss Bank Corp., Basle 3,661 11.398 6,444 7 -38.81 5,979 14,414 8,011 7 +36.3
15 22 Mitsui Trust & Banking Co. Ltd.,
Tokyo 3,647 12,081 7,177 4 +68.35 2,167 10,893 4,889 4 +19.3
16 19 Nippon Credit Bank Ltd., Tokyo 3,572 6,499 2,046 2 +47.41 2,424 5,155 1,575 2 +14.3
17 17 Taiyo Kobe Bank Ltd., Kobe, Japan 2,939 8,453 3,929 5 +8.03 2,720 7,747 3,235 4 +74.3
18 21 Mitsubishi Trust & Banking Corp.,
Tokyo 2,705 10,180 4,768 3 +16.44 2,323 11,797 5,716 3 +12.1
19 20 Banco di Nipoli, Naples 2,635 7,653 4,076 +13.06 2,330 5,136 2,925 -12.4
20 14 Bank of Nova Scotia, Toronto,
Canada 2,621 8,123 3,218 9 -28.03 3,642 10,370 2,853 9 +2.9
21 18 Daiwa Bank, Ltd., Osaka 2,378 11,671 6,296 4 -12.13 2,707 9,650 4,918 4 +23.6
22 25 Banco di Roma. Rome 2,102 8,786 3,468 5 +13.65 1,849 7,677 3,355 4 +65.7
23 26 Banque Nationale de Paris 2,012 9,261 5,230 8 +17.72 1,709 8,482 4,584 9 -23.9
24 33 Sumitomo Trust & Banking Co. Ltd.,
Osaka 1,926 9,830 3,218 3 +35.57 1,420 8,024 4,676 3 +24.9
25 23 Algemene Bank Nederland, Amsterdam 1,830 7,124 4,779 19 -12.71 2,096 5,133 2,648 14 +30.0
Totals for the top 25 foreign banks in the U.S. 126,394 442,661 237,479 173 +12.94% 111,915 399,741 205,538 173 +22.0%
Totals for all foreign banks in the U.S. 178,900 695,600 370,400 697 +8.90% 164,300 638,300 331,200 666 +18.6%
Top 25 banks' share of total for all foreign banks 70.6% 63.6% 64.1% 24.8% 68.1% 62.6% 62.1% 25.9%
Sources: (1) American Banker survey questionnaires and Federal Reserve Board data. The totals for each bank are based on the combined business loans (C&! loans), deposits, and assets
for all of the bank's U.S. banking offices, for which data were available. U.S. banking offices include: agency offices, branches, commercial bank subsidiaries, New York State investment companies,
and Edge banks. % Chg. C&I is the annual growth in commercial and industrial loans over the year ended 6130/89 and 6130/88, respectively. Note: percent changes based on unrounded figures
and may differ slightly from dollar figures shown (a)-The data for 6130/88 does not include the acquisition of Union Bank, Los Angeles, which was completed on 10/31188.
(2) American Banker, February 27, 1990, p. 17A.
>-'
-l
tN
.....
-.l
+:>.
Table 5-7. Total Assets of the U.S. Offices of Foreign Banks, 1980 to 1988. (Thousands of Dollars)
State 1980 1981 1982 1983 1984 1985 1986 1987 1988
New York 132,185,600 149,759,632 190,238,583 214,902,610 243,689,943 293,966,846 367,366,289 412,144,700 453,101,639
California 50,051,420 79,694,302 86,269,854 88,555,387 93,341,093 100,852,400 99,066,950 106,682,400 112,895,868
Illinois 8,323,716 9,514,448 11,100,842 11,905,420 22,814,261 26,126,252 32,816,602 40,387,480 48,124,821
Florida 671,301 1,358,633 3,064,671 3,995,317 4,668,136 6,040,459 7,662,917 9,387,411 9,306,516
Georgia 1,809,330 2,015,763 2,401,384 2,579,083 3,531,717 3,904,019 4,593,574 5,617,684 6,043,273
Maryland 0 0 0 2,980,514 3,393,462 3,819,452 4,539,586 4,709,410 5,485,122
Oregon 3,731,203 5,996,888 4,819,417 4,745,088 3,359,525 2,001,934 2,615,469 2,970,110 3,849,959
Texas 48,216 59,936 95,776 103,085 67,432 196,431 2,079,911 2,427,095 3,055,482
Delaware 0 0 0 124,166 161,327 776,348 2,007,401 2,648,545 3,250,668
Washington 1,630,005 1,571,729 1,662,378 1,883,312 1,806,100 1,650,241 2,005,073 2,355,573 2,093,898
Pennsylvania 360,299 698,339 709,686 883,355 797,578 974,223 822,471 1,052,836 384,189
Massachusetts 278,170 280,435 340,099 335,573 343,980 405,656 511,127 821,641 1,005;386
District of Columbia 5,355 173,175 178,062 263,632 299,137 418,055 435,804 0 368,434
New Mexico 0 0 0 0 0 360,603 362,273 339,012 284,376
Hawaii 22,699 24,707 28,099 36,750 37,926 30,714 48,174 48,782 117,283
Louisiana 0 69,551 113,530 142,896 2,000 2,000 2,000 2,000 0
Arizona 0 0 0 0 0 0 0 0 25,740
North Carolina 0 0 0 0 0 0 0 0 249,128
Totals 199,115,314 251,217,538 301,022,381 333,436,188 378,313,617 441,525,633 526,935,621 591,204,885 649,641,782
Results using other dependent variables were very similar. The study
concluded "that the most important factors determining foreign bank growth
over the period of the study were (i) the size of interest differentials
between United States and foreign deposits and loans, (ii) the falling PI
E ratios for United States bank stocks, (iii) the increased size of (net)
foreign domestic investment in the United States, (iv) the persistent de-
preciation in the dollar, and (v) expectations that the IBA would have a
restrictive effect on foreign bank activity in the United States.,,7
A second study by Goldberg and Saunders (1981b) examined the growth
of foreign banks by agency, branch, and subsidiary with a similar model
of four explanatory factors. "The empirical results suggested that these
factors tended to impact on agencies in a different manner from branches
and subsidiaries, with agencies more affected by international business
considerations and more concerned with current, or short-run, profitabil-
ity. Also, the IBA appeared to have a negative affect on the underlying
growth of agencies and branches leaving subsidiaries relatively unaffected.,,8
Recently, Hultman and McGee (1989) updated these studies employ-
ing a model with three independent variables. They examined foreign
bank activity in the United States between 1973 and 1986 and found the
growth of foreign branches and agencies positively related to changes in
foreign direct investment in the United States, the value of the dollar, and
the International Banking Act of 1978. The growth of subsidiaries was
positively related to the first two of these variables and negatively related
to the bank price earnings ratio.
Useful insights can be obtained by examining the factors that have
resulted in differential growth across countries of origin of the foreign
banks. Grosse and Goldberg (1991) use pooled time-series cross-section
data from 1980 to 1988 to determine the factors affecting growth of foreign
banks from source countries. "The results indicate that foreign banks are
THE COMPETITIVE IMPACf OF FOREIGN COMMERCIAL BANKS 177
drawn to the United States to service customers from the home country,
since bank presence is positively related to foreign direct and portfolio
investment and foreign trade with the United States. Larger banking sectors
in the foreign countries encourage greater investment in banking operations
in the United States. Less stable countries are more likely to have greater
presences in the U.S. banking sector, holding other factors constant.
Finally, the greater the geographic or cultural distance of the source
country from the United States, the more likely are banks to build their
asset bases in the U.S. market.,,9
Until recently, foreign bank data by country were only available for
Japan, Canada, and all of Europe. Hultman and McGee (1989) have used
the data for Japan to identify factors affecting the growth of Japanese
banks in the United States. They find that the share of total banking
assets of Japanese banks in the United States is positively related to
Japanese foreign direct investment in the United States, and the Japanese
Banking Act of 1982 that opened Japan somewhat to foreign banks, and
negatively related to the value of the yen in relation to the dollar. Terrell,
Dohner, and Lowrey (1990) find that "U.S. activities of Japanese banks
during this period appeared strongly related to Japanese domestic financial
variables as well as to conditions in the U.S. market. Commercial and
industrial loans at these offices responded both to expansions in Japanese
trade and to restraints on domestic Japanese interest rates, while interbank
trading at U.S. offices of Japanese banks responded to both price and
quantity restraints on domestic Japanese banking activity.,,10
It is also useful to examine the factors causing differential foreign bank
growth across states. Goldberg and Grosse (1990) find that the foreign
bank presence, whether measured by assets or number of offices, is drawn
to states with large bank market sizes and low levels of regulation of
foreign banks. The level of international trade of the state was also im-
portant in explaining foreign bank assets. New York was entered as a
dummy variable because many foreign banks are primarily interested in
wholesale business and much of the foreign exchange and money market
business is concentrated in New York. The coefficient was the expected
positive sign.
These studies suggest that a number of economic factors and an im-
portant regulatory change affect the growth of foreign banks. The im-
portance of foreign trade and foreign direct investment suggest that foreign
banks frequently are serving customers from their home countries. Rela-
tive profitability, the cost of acquiring banks relative to earnings, and the
value of the dollar appear to motivate banks in the expected directions.
Since past regulatory changes have affected foreign bank growth, it might
178 THE CHANGING MARKET IN FINANCIAL SERVICES
u.s. bank activity in the Eurodollar markets l4 and Japanese bank activity
in Eurodollar and European markets,IS but the evidence is less clear re-
garding competitive relationships in the United States. Baer (1990) claims
that "fears regarding the competitive advantages conveyed by lax regu-
lation at home may be justified, particularly with respect to banks owned
by foreign governments. And although no objective rankings exist, this
concern would also appear to be valid where privately-owned foreign
banks enjoy stronger guarantees from their governments than U.S. banks
enjoy from the U.S. government. Whatever the particulars of the com-
plaint, it ultimately boils down to the assertion that foreign banks are able
to hold less capital per dollar of risk or pay less for the capital that they
raise.,,16
The evidence on capital levels is mixed. Book values of Japanese banks,
for example, are very low compared to American banks. The Japanese
banks, however, have large gains on equity holdings that are not recog-
nized on their balance sheets. When market capitalization is considered,
the Japanese banks have more capital than do banks from any of the
major countries. Recent international efforts have been made to equalize
capital requirements of banks across countries, but implementation of
equalized risk-based capital requirements will be difficult because of dif-
fering accounting practices across countries.
The major competitive impact of foreign banks has been in commercial
and industrial loans. From table 5-3 it can be seen that foreign banks have
a disproportionate amount of their assets concentrated in commercial and
industrial loans. Moreover, the foreign bank share of these loans has risen
steadily since 1982. In order to attract new business in these loans, foreign
banks, and especially the Japanese, are offering lower interest rates. "Where
spreads of seventy-five basis points over prime were once considered the
absolute minimum, some banks, notably the Japanese again, are earning
a reputation for ruthlessness. Comments John Miller, vice president in
the New York office of San Francisco-based Wells Fargo Bank: 'The
foreign banks are cutting their spreads down as low as ten basis points
over prime just to get the business.' ,,17 This has presented domestic banks
with a major competitive challenge.
Much of the increase in C&I loans held by foreign banks has resulted
from the sale of domestic C&I loans by U.S. commercial banks. These
purchases of existing loans have apparently become important since the
mid-1980s and in 1988 accounted for 2.5 percent of total C&I loans. "Thus,
U.S. banks have been directly responsible for over two-fifths of the 5.8
percentage point increase in the market share of U.S. branches of foreign
banks that occurred between 1980 and 1988."18 Generally, smaller foreign
180 THE CHANGING MARKET IN FINANCIAL SERVICES
banks have been more interested in loan participations because they have
less access to major corporate borrowers.
Increasingly, the larger foreign banks are moving toward competing
with domestic banks in a larger range of financial services. They have
moved away from serving only trade finance and large corporations
toward serving small and mid-sized companies. Cho, Krishnan, and Nigh
(1987) have surveyed 271 foreign banks in the United States: " ... 152
banks (56 percent) pointed out trade financing as one of the main areas
of specialization in their U.S. offices. One hundred and nineteen banks
(44 percent) listed corporate banking and 81 banks (31 percent) listed
foreign exchange trading as one of their main areas of specialization. ,,19
Other areas mentioned were money market activities, retail banking, and
business development.
There is much anecdotal evidence of foreign bank participation in a
number of banking activities. Foreign banks participate heavily in finan-
cial loans such as loans to depository institutions and real estate loans.
Increasing participation during the 1980s in real estate lending has not
gone unnoticed. Foreign banks have been able to obtain some corre-
spondent banking business because small banks consider them lesser threats
for stealing customers than large domestic banks. The foreign banks have
engaged in off-balance sheet activities such as support arrangements for
commercial paper. They have had some success in the mergers and ac-
quisitions advisory business, mostly with companies from their home
country. On the deposit side, they have appeared to concentrate on
nontransactions deposits. The only way to assess the competitive impact
of foreign banks in these areas is to examine empirical data. The next
section includes some necessary analysis and suggests other types of
empirical examinations to ascertain competitive impact.
The International Banking Act of 1978 was intended to equalize the
treatment of foreign banks and domestic banks in the United States;
however, because of varying state laws, foreign banks have differing
opportunities in different states. "Many states employ reciprocity provi-
sions, asset maintenance or deposit requirements, and geographic re-
strictions to influence the activities of foreign banks within their borders.
Tax laws may also have some influence on foreign banks' decisions regarding
where to locate and what type of office to establish. Few states pursue an
open policy that could be considered equivalent to a national treatment
approach.,,20 As can be seen in table 5-7, only a limited number of states
have foreign banking offices, and the activity is heavily concentrated in
the top three states. This uneven distribution is due largely to the moti-
vation of foreign banks to concentrate in the types of business done in the
THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 181
Several studies in the past have compared financial ratios of foreign banks
to domestic banks. These studies provide some information on the activities
engaged in by foreign banks. In the earliest study, Terrell and Key (1977)
compare a number of balance sheet items for three dates (November
1972, November 1974, and May 1977) between U.S. offices of foreign
banks and the weekly reporting banks. They also divide foreign banks by
organizational form, parent country (Japan, Canada, Europe, and rest of
the world) and state (New York, California, and Illinois). They note a
growing importance of C&I loans for foreign banks and find that some
offices offer a wide range of both wholesale and retail services.
A report by the General Accounting Office (1979) examines foreign
bank growth up to 1979. The study finds that foreign banks have become
more competitive with domestic banks in C&I loans from 1972 to 1979.
Trends in the growth of several balance sheet items are presented.
A third study by Houpt (1980) examines the characteristics and per-
formance of U.S. banks acquired by foreign banks and U.s. banks estab-
lished de novo by foreign banks. Note that agencies and branches are not
treated in the study. Using a paired-bank approach he finds that foreign-
owned banks held considerably less state and municipal debt, depended
more on purchased funds, and were less profitable. Houpt qualifies his
results because the number of banks was small and most of the acquisi-
tions occurred since 1975.
In order to assess the competitive impact of foreign banks in the United
States, it is first necessary to identify the activities in which foreign banks
have participated. This article does this directly by comparing financial
ratios from the most recent available balance sheets of both foreign- and
domestic-owned banks participating in the American market. From tables
5-2 and 5-3 it has already been indicated that foreign banks do playa
significant role in the American banking market. Comparisons of finan-
cial ratios by type of financial institution indicate in which activities the
foreign banks could be expected to have the greatest relative competitive
import.
182 THE CHANGING MARKET IN FINANCIAL SERVICES
Sixteen financial ratios are chosen for analysis and are presented in
tables 5-8 through 5-15. The first thirteen measures represent particular
assets divided by total assets, while the last three represent individual
liabilities divided by total assets. Data in tables 5-8 through 5-14 are from
the December 1989 Call Reports, while table 5-15 contains data from
December 1980 for foreign banks in order to ascertain changes in foreign
bank activity over the last nine years. Changes in balance sheet defini-
tions and data presentation make it difficult to obtain comparable data
for early periods. Note that since the agency and branch data were obtained
from a different data file than were the full commercial banks, some of
the items on the balance sheet differed and judgment had to be used to
make ratio comparisons meaningful.
The three main types of foreign banking organizations in the United
States follow different regulations and differ in purpose; therefore, the
foreign banks have been separated into three categories in each table.
Since the larger American banks differ from the smaller American banks
in their business activities, the banks with over $1 billion in assets have
been separated from other domestic banks in the analysis. Consequently,
each table (except for table 5-15) contains five columns: agencies; branches;
subsidiaries (commercial banks with greater than 50 percent foreign
ownership); all domestic banks; and large domestic banks (only domestic
banks with greater than $1 billion in assets). All banking organizations
that have reported the relevant data to the bank regulatory agencies are
included in the calculation of the ratios for each type of organization.
The financial ratios for all banks in each category of banking organ-
ization are calculated in two different ways. The first method adds the
particular asset or liability value for all banking organizations in a category
and then divides by the sum of the assets for all organizations. The results
are then multiplied by 100 so as to represent percent of assets. This method
is used in table 5-8 and tables 5-10 through 5-15. The second method
calculates the financial ratio for each institution and then finds the aver-
age of these ratios. Once again, the ratios are multiplied by one hundred
so that the results appear as percentages. Table 5-9 presents these results
for all geographic regions for December 1989 so that the two methods of
calculating ratios can be compared. Ratios using this method have been
calculated for all divisions of banks but because of space constraints these
results are not presented here.
The first method gives greater weight to larger organizations, while the
second method gives equal weight to every organization. This, in effect,
could mean that Citibank is given the same weight as a single office $10
million bank. Examination of the individual ratios reveals many extreme
Table 5-8. All States-Ratio of Sums-December 1989 (Values are percent of total assets).
Agencies Branches Subsidiaries All Domestic Large Domestic
Cash 9.48 28.69 18.54 9.94 11.17
Total Loans 62.32 41.95 57.73 62.51 64.69
US Treas & Govt 0.31 1.63 7.25 11.51 8.56
Other Securities 3.61 4.81 3.38 3.91 3.89
Federal Funds Sold 3.45 3.46 2.65 3.90 2.81
Other Loans 0.27 0.50 2.56 2.66 3.65
Comm & Indus Loans 33.20 22.17 23.99 16.87 20.80
Bankers Acceptances 0.34 0.18 0.12 0.05 0.05
Commercial Paper 0.17 0.16 0.06 0.08 0.07
Real Estate Loans 8.64 4.73 15.91 23.55 21.41
Financial Institutions Loans 14.82 11.36 3.04 1.61 2.22
Foreign Govt Loans 4.21 2.69 1.62 0.82 1.20
Loans for Purchase of
Securities 0.88 0.36 0.73 0.55 0.74
Deposits 21.82 51.90 69.63 77.30 72.76
Federal Funds Purchased 10.83 8.20 5.51 5.69 7.73
Other Liabilities 39.83 19.18 6.98 3.53 4.92
......
00
w
Table 5-9. All States-Average of Ratios-December 1989.
Agencies Branches Subsidiaries All Domestic Large Domestic
Cash 9.08 19.62 16.56 8.11 9.24
Total Loans 60.97 50.26 52.32 53.07 65.06
US Treas & Govt 1.55 2.01 8.78 21.29 11.51
Other Securities 3.05 4.94 4.16 6.88 5.18
Federal Funds Sold 4.82 5.00 7.74 7.24 3.38
Other Loans 1.32 1.15 1.28 0.92 2.11
Comm & Indus Loans 34.84 28.07 21.53 13.88 17.55
Bankers Acceptances 0.26 0.37 0.33 0.13 0.04
Commercial Paper 0.38 0.19 0.06 0.17 0.16
Real Estate Loans 9.87 8.35 23.13 24.93 24.58
Financial Institutions Loans 9.54 9.96 4.24 0.45 1.10
Foreign Govt Loans 5.05 2.29 1.79 0.03 0.20
Loans for Purchase of
Securities 0.16 0.17 0.31 0.18 0.43
Deposits 27.62 42.15 71.41 87.53 76.25
Federal Funds Purchased 5.47 7.27 3.21 0.57 6.60
Other Liabilities 25.43 18.84 7.18 0.37 3.48
Table 5-10. New York-Ratio of Sums-December 1989.
Agencies Branches Subsidiaries All Domestic Large Domestic
Cash 7.65 29.59 21.99 13.29 13.49
Total Loans 55.56 39.16 51.80 61.24 61.31
US Treas & Govt 0.65 1.86 6.78 4.09 3.73
Other Securities 3.37 5.07 3.56 3.23 3.09
Federal Funds Sold 24.92 3.71 2.73 2.25 2.11
Other Loans 1.17 0.52 2.15 5.88 6.02
Comm & Indus Loans 20.05 19.59 24.14 22.42 22.76
Bankers Acceptances 0.00 0.16 0.16 0.06 0.06
Commercial Paper 0.00 0.17 0.09 0.12 0.11
Real Estate Loans 4.72 4.02 9.72 15.91 15.46
Financial Institutions Loans 10.50 11.46 3.79 4.94 5.08
Foreign Govt Loans 19.17 3.04 2.28 3.09 3.21
Loans for Purchase of
Securities 0.00 0.42 0.92 1.01 1.00
Deposits 22.86 56.80 65.44 69.95 69.37
Federal Funds Purchased 25.49 7.90 5.55 6.48 6.68
Other Liabilities 29.40 17.20 8.27 7.32 7.59
.....
00
Vl
......
00
0\
.......
00
-..l
>-'
00
00
.....
00
\0
,....
~
values, particularly for smaller banks. Consequently, the use of the sec-
ond method allows these extreme values to playa larger role. If we view
these ratios as being a sample from a time continuum of different ratios,
then statistical tests of differences between categories can be conducted
using the second method; however, the standard deviations for most of
the ratios are high as compared to the means. Here it was decided to
compare population means rather than to do statistical testing.
Tables 5-8 and 5-9 present data from all states for December 1989.
Tables 5-10 through 5-13 present data from the four most important states
for foreign banks in the United States. Comparisons across states indicate
that foreign banks do not perform identically in different states. Table
5-14 indicates results for all states, excluding the four largest. The only
data presented for 1980 are for all states, but individual state data have
been obtained with similar differences among states for 1980 as in 1989.
The data in table 5-8 of the ratio of sums for all states indicate substantial
portfolio differences for agencies and branches as compared to the three
categories of full commercial banks. In addition, the foreign subsidiaries
also differ from the two categories of domestic banks in a number of
financial ratios.
Commercial and industrial loans is the activity where it has generally
been acknowledged that foreign banks have had the greatest competitive
impact. The results in table 5-8 confirm that all types of foreign banks
have emphasized C&I loans to a greater extent than have their domestic
counterparts. This appears to be especially true for agencies that have
33.20 percent of assets in C&I loans. Branches with 22.17 percent and
subsidiaries with 23.99 percent of assets in C&I loans also exceed the
proportions in C&I loans of all domestic banks of 16.87 percent. As ex-
pected, the larger domestic banks with 20.80 percent in C&I loans con-
centrate on these loans more than the average bank. Agencies and branches,
and to a lesser extent subsidiaries, have placed greater emphasis in banker's
acceptances, commercial paper, loans to financial institutions, and loans
to foreign governments and institutions. There is no clear pattern for
differences between domestic and foreign institutions for loans for the
purchase of securities. Though some observers have indicated that foreign
banks are playing a greater role in real estate loans, the data indicate that
at this time domestic banks are much more involved with these types of
loans. Branches hold substantially more cash and less loans than the other
categories. Subsidiaries, though not as cash-laden as branches, also hold
more cash and less loans than domestic banks, whereas agencies are similar
to domestic banks. Foreign banks, and particularly agencies and branches,
hold substantially lower portfolios of U.S. government securities than
192 THE CHANGING MARKET IN FINANCIAL SERVICES
domestic banks. Results for holding other securities and lending in the
federal funds market vary by type of foreign institutions. Agencies and
branches make fewer other loans, many of which are retail oriented, than
domestic banks. Subsidiaries are close to domestic bank levels in this
activity.
There are also some interesting comparative results on the liability side
of the balance sheet. As expected, because of legal restrictions, agencies
obtain far fewer of their funds from deposits and credit balances than
other categories of institutions. Branches are also considerably less de-
pendent on deposits, and subsidiaries slightly less dependent. Agencies
and branches are more active, however, in borrowing in the federal funds
market and in obtaining funds from other liabilities (borrowing in money
markets). Subsidiaries are more active (but to a lesser extent) in money
market borrowing. Note, that agencies and branches obtain much funding
from their parent organizations and that this is not reported in the tables.
Most of the above-mentioned relationships hold when the second
measure of financial ratios is used in table 5-9. However, there are some
important differences due to the heavier weighting of smaller institutions
in the second method. The major differences are highlighted here and the
analysis by state is only reported using the first method. Results by state
using the second method have been obtained but are not presented here.
These results are, nevertheless, very similar to those presented in the
article. The major result with respect to C&I loans is even more pronounced
using this method because the smaller domestic banks with fewer C&I
loans are given more weight. The foreign banks still have less in real
estate loans and more in loans to financial institutions and foreign govern-
ments. The results for cash are not as distinctive for branches using this
methodology. Because smaller domestic banks have fewer loans and more
government securities, in table 5-9 the domestic banks have a lower ratio
of total loans and higher U.S. government securities than in table 5-8. The
differences in ratios of total loans are reduced among categories of banks
using this methodology, but results for U.S. government securities are
similar to before. Foreign banks hold fewer other securities under the
second methodology. Results for other loans lose their distinctive pattern.
On the liability side there are fewer differences in results by using alter-
native methodologies. Note, however, that the ratio for domestic bank
borrowing in the federal funds market is dramatically reduced because
smaller banks do not borrow heavily in this market.
Comparing ratios across the four most important foreign banking states
and the residual states provides some interesting information about the
nature of foreign bank competition. There are distinct differences among
THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 193
geographic areas. Foreign banks have grown rapidly both in absolute size
and in relation to domestic banks; therefore, an activity showing a de-
crease in its relative importance to foreign banks might still be an activity
in which foreign banks are having an increased competitive impact.
The studies analyzing the factors affecting the growth of foreign banks
in the United States find a number of economic and regulatory factors
that have been important in motivating foreign bank growth. Several of
these variables are likely to change in the future in a direction that will
further encourage the growth of foreign banks. Though foreign banks in
the United States should be playing a greater role in the future, there are
certain factors that could retard this development. Accusations that for-
eign banks have competitive advantages over domestic banks have been
made but, as discussed earlier, there does not appear to be solid evidence
to support these suggestions. Nevertheless, legislation may be proposed
to restrict foreign banks. In addition, proposed legal changes in Europe
may affect American banks in Europe, and this in turn could lead to
changes in the treatment of foreign banks in the United States.
The European Community (EC) is scheduled in 1992 to remove many
commercial barriers between countries, including restrictions on commercial
banks operating in other EC countries. Banks with charters in one EC
country will be allowed to operate in other EC countries and acquisitions
of existing banks by banks in other countries are likely to occur. The
structure of banking markets in Europe should, therefore, change dra-
matically. The resulting structural changes could impact significantly upon
the availability and nature of banking services in these countries. There
could also be important effects upon the ability of banks from non-EC
countries, particularly American and Japanese banks, to operate in these
markets.
Since the Treaty of Rome in 1957, the EC has been moving to eliminate
commercial restrictions among countries. In 1960 the EC adopted the
First Banking Directive that argued for financial integration but retained
national control. In 1988 the EC adopted the Second Banking Directive
that permits any banking organization licensed in any EC country to open
offices in other EC countries and to engage in certain acceptable lines of
business. The directive is intended to be implemented before January 1,
1993.
The directive deals with the treatment of non-EC banking organiza-
tions. Subsidiaries of non-EC banking organizations chartered in an EC
country would be considered to be EC companies regardless of ownership
and, thus, would be allowed to branch throughout the Community without
obtaining additional licenses. Future entry, though, would be subject to
THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 197
Notes
1. Houpt (1988, 7, 11) and additional updated tables provided to the author by James
Houpt.
2. Ibid., (25) and updated material supplied by James Houpt.
3. American Banker (February 27,1990, 18A). See Table 3, footnote (c), in this article.
4. Cho, Krishnan, and Nigh (1987, 60).
5. Ibid., (60).
6. Ibid., (64).
7. Goldberg and Saunders (1981a, 32).
8. Goldberg and Saunders (1981b, 372).
9. Grosse and Goldberg (1991,1109-110).
10. Terrell, Dohner, and Lowrey (1990, 48-49).
11. This is argued by Nadler (1989, 5).
12. Ibid.
198 THE CHANGING MARKET IN FINANCIAL SERVICES
13. Nadler (1989) provides an extensive discussion of how this practice works and the
implications for domestic banks.
14. See Baer and Pavel (1987).
15. See Terrell, Dohner, and Lowrey (1989).
16. Baer (1990, 25).
17. Jedlicka and Tobin (1988, 89).
18. Baer (1990, 25).
19. Cho, Krishnan, and Nigh (1987, 69).
20. Hultman (1987, 348).
References
Baer, Herbert L. 1990. "Foreign Competition in U.S. Banking Markets," Federal
Reserve Bank of Chicago. Economics Perspectives 14 (3) (May/June): 22-
29.
Baer, Herbert L., and Christine A. Pavel. 1988. "Does Regulation Drive
Innovation?" Federal Reserve Bank of Chicago. Economics Perspectives 12, 2
(March/April): 3-15.
Cho, Kang Rae, Suresh Krishnan, and Douglas Nigh. 1987. "The State of Foreign
Banking Presence in the United States." International Journal of Bank Marketing
5,2: 59-75.
General Accounting Office. 1979. "Considerable Increase in Foreign Banking in
the United States Since 1972." Report by the Comptroller General of the
United States. (August).
Goldberg, Lawrence G., and Anthony Saunders. 1981a. "The Determinants of
Foreign Banking Activity in the United States." Journal of Banking and Finance
5, 1 (March): 17-32.
- - . 1981b. "The Growth of Organizational Forms of Foreign Banks in the United
States." Journal of Money, Credit and Banking 13, 3 (August): 365-374.
Goldberg, Lawrence G. and Robert Grosse. 1990. "Distribution by State of Foreign
Bank Activity in the United States." (March) Mimeographed.
Grosse, Robert, and Lawrence G. Goldberg. 1991. "Foreign Bank Activity in the
United States: An Analysis by Country of Origin." Journal of Banking and
Finance 15, 6 (December): 1093-112.
Houpt, James V. 1980. "Foreign Ownership and the Performance of U.S.
Banks." Staff Study 109. Board of Governors of the Federal Reserve System.
Washington, D.C. (July).
--.1988. "International Trends for United States Banks and Banking Markets."
Staff Study 156. Board of Governors of the Federal Reserve System. Washington,
D.C. (May).
Hultman, Charles W. 1987. "The Foreign Banking Presence: Some Cost-Benefit
Factors." Banking Law Journal 104, 4 (July/August): 339-49.
Hultman, Charles W., and Randolph McGee. 1989. "Factors Affecting the Foreign
Banking Presence in the United States," Journal of Banking and Finance, 13,
3, (July): 383-96.
THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 199
Jedlicka, John, and Mary Tobin. 1988. "Foreigners Ferocious Financial Threat."
Euromoney (November): 89-93.
Nadler, Paul. 1989. "Balances and Buggy Whips in Loan Pricing." Journal of
Commercial Bank Lending 72, 6 (February): 4-9.
Terrell, Henry S., Robert S. Dohner, and Barbara R. Lowrey. 1989. "The United
States and U.K. Activities of Japanese Banks: 1980-1988." Board of Governors
of the Federal Reserve System. International Finance Discussion Papers 361
(September).
- - . 1990. "The Activities of Japanese Banks in the United Kingdom and in the
United States, 1980-1988." Federal Reserve Bulletin 76(2) (February): 39-50.
Terrell, Henry S., and Sydney J. Key. 1977. "The United States Activities of
Foreign Banks: An Analytical Survey." Board of Governors of the Federal
Reserve System. International Finance Discussion Papers, No. 113. (November).
Zimmerman, Gary C. 1989. "The Growing Presence of Japanese Banks in
California." Federal Reserve Bank of San Francisco. Economic Review
(Summer): 3-17.
COMMENTARY
Gary C. Zimmerman
Goldberg's article is a timely one. It deals with important issues facing the
U.S. economy: foreign investment, foreign bank control, foreign bank
advantages, foreign bank funding-and examines them in the context
of the rapid growth of foreign banks in the u.s. market. He uses these
issues to analyze the competitiveness of domestic banks and foreign banks,
especially the major Japanese-owned banks operating in the United States.
Clearly, this is a subject that is of interest to many of us. Foreign bank
presence is frequently discussed in the press. And fortunately, we have
some measurable ways of examining their presence using asset size,
commercial lending or activity in the "guarantee" markets. These measures
highlight foreign bank penetration of the U.S. banking markets. For
example, in California $1 out of every $3 in banks' outstanding business
loans are held by Japanese-owned banking institutions. In the guarantee
market, foreign bank presence is even greater. Again, in California, foreign-
owned banks issue about 60 percent of all standby letters of credit issued
by banking institutions in the state.
201
202 THE CHANGING MARKET IN FINANCIAL SERVICES
Review
Rapid Growth
The article cites some dramatic figures on foreign bank growth in the
United States. However, the figures are somewhat misleading because
virtually all of the foreign bank expansion in the United States has been
by Japanese-owned banks. Thus, the article would be more useful if it
reported bank growth for three categories of banks, based on country of
ownership: United States-owned, Japanese-owned, and other foreign-owned
(excluding the Japanese).
THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 203
Growth rates for the various groups illustrate this point. Over the
period from year-end 1985 to mid-year 1990, non-Japanese foreign-owned
banks, branches and agencies as a group have grown at about the same
average rate as domestic banks, about 5 percent per year. In contrast,
Japanese-owned banks, branches and agencies have expanded at close
to a 25 percent annual rate. All three types of Japanese-owned banking
institutions grew rapidly during the period. Clearly, the growth pattern
indicates that the focus should be on the exceptional growth of J apanese-
owned banking institutions in the United States, rather than the broad
grouping of all foreign banks.
After describing the rapid foreign bank growth and the reasons behind
it, Goldberg presents a descriptive section on the evidence of competitive
effects and on the allegations of "unfair" competition from foreign banks.
In this section he makes his point using developments in the commercial
lending market, which is a good way to handle the analysis. Commercial
lending is an important area for foreign-owned banks, and it is a market
in which they playa major role.
One criticism of this section is that it again needs to focus on the
Japanese-owned banking institutions. They account for the exceptional
growth of foreign banks' commercial lending in the United States.
Commercial loans have grown at about a 1 percent annual rate since 1985
for both domestic banks and non-Japanese foreign-owned banks. Yet,
business loans at Japanese-owned banks have grown at over a 30 percent
annual rate over this period. Moreover, they captured nearly a 9 percent
increase in market share over the same period (from 6.4 percent to 15.1
percent).
Competitiveness
Until recently, restrictions on the ability of Japanese banks to offer their domestic
Japanese customers market-determined interest rates on deposits appear to
have had the effect of inducing Japanese banks in the aggregate to shift some
domestically oriented business to their U.S. offices because the low regulated
interest rates in Japan caused funding difficulties and Japanese banks actually
relied on their overseas branches for funds for use at their domestic offices.
This fact suggests that Japanese banks were not on balance borrowing low-cost
funds in Japan and relending them in the United States at low rates.
Thus, the situation facing Japanese banks in their home market ap-
pears to be much like the situation U.S. banks faced when they were
constrained by Regulation Q interest rate ceilings on deposits. For
example, before deregulation in the United States, banks substituted
wholesale borrowings (increasing their reliance on large certificates of
deposit, other money market borrowings, and borrowing from their overseas
affiliates) to make up for retail deposit shortfalls caused by below-market
rates on retail accounts. Faced with a similar constraint on retail deposits,
Terrell indicates that Japanese banks have responded by booking and
funding some loans in the United States and by increasing their borrowing
in offshore markets without interest rate controls.
Terrell also shows that agencies and branches of Japanese banks oper-
ating in the United States are primarily funded by interbank borrowing at
open-market rates from U.S. banks and from affiliates in London and
other offshore markets. Thus, he presents evidence that in the aggregate,
over the last five years Japanese banks overseas have been net lenders to
Japan.
Terrell also indicates that the amount of borrowing to fund their do-
mestic activities has fallen as the share of Japanese deposits subject to
liberalized interest rates has risen. Of course, this trend could have im-
portant implications for the continued growth of Japanese-owned banks
in the United States.
THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 205
California
Financial Ratios
Market Share
Summary
Where does all this take us? Goldberg has answered the questions he
posed earlier in his article: What factors affect foreign bank growth? In what
areas are foreign banks competitive? Do foreign banks have unfair com-
petitive advantages? Are there competitive effects? Thus, he takes us to
what I consider to be the "big" question: What will the future bring for
foreign-owned banks, that is, will their expansion continue or not?
His article provides a good perspective on the inroads of foreign banks,
especially the Japanese-owned banks, and on the importance they now
playas competitors in the U.S. banking system. I would suggest extending
the research and the outlook to include a number of changes currently
going on in the world, in Europe and in Japan, that may impact foreign
bank activity in the United States.
Again, Goldberg's article points us in the right direction, even though
it would benefit from focusing more on the Japanese-banks and whether
they can sustain their rapid growth trend of the past decade. That is
important, since Japanese banks currently hold over 10 percent of the
U.S. bank market share (over 25 percent in California), and yet their
market share has not increased significantly since 1988. Many of the fac-
tors noted in his article and my comments, as well as trends in the Japanese
banking system: from liberalization of interest rates in Japan reducing
pressure to shift funding to the United States and to borrow from the
United States; to uniform international capital standards for banks; to the
significant decline in the Japanese stock market and its reduction of
Japanese banks' "hidden reserves" and capital positions; to banks in Japan
losing their traditional customers to the capital markets (much as it has
THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 209
Introduction
211
212 THE CHANGING MARKET IN FINANCIAL SERVICES
The financial innovations and the rapid domestic growth of the 1980s
expanded domestic markets. Similarly, expansion of business activity abroad
brought additional demands for financial services, an incentive for invest-
ment banks to expand beyond their own domestic horizons. Financial
engineering and new product development required both specialization
and market size to provide an initial market and the liquidity necessary
to launch a new product. These competing demands appear to have led
to substantial growth among a few of the largest firms. For example,
according to Investment Dealer's Digest, in 1980,80 percent of new issue
debt was syndicated; in 1989, only 20 percent was syndicated.
Likewise, changes in the regulatory environment have altered the nature
of the underwriting business. One of the largest changes was the adoption
of Rule 415 in 1982. Rule 415 allows firms to register their offerings with
the Securities and Exchange Commission before they offer them to the
market. The effect of the rule is to permit firms more flexibility in timing
214 THE CHANGING MARKET IN FINANCIAL SERVICES
the market for their new issues. But once the decision has been made to
bring the issue to market, underwriters are under pressure to do so expedi-
tiously. One effect of Rule 415 has been to reduce the role of regional
underwriters in the new issue market, perhaps because they lack sufficient
size to bring a full issue to market promptly.
In April 1990, the SEC adopted rule 144a, which permits the trading
of unregistered securities among "sophisticated" investors, that is, those
with more than $100 million in securities under management. One an-
ticipated result of Rule 144a is the increased trading of foreign securities
within the United States. Likewise, the liquidity provided by the new
trading opportunity is likely to make private placements a more attractive
alternative to normal underwritten issues and again place emphasis on
market size as a driving concern.
Foreign Competition
The Market
Market Identification
The target market for underwriting services is defined for the purposes of
this analysis to be corporate security distributions managed by broker/
dealers that are registered to do business in the United States, acquired
by them through negotiated deliberations, and publicly offered on a firm-
commitment basis.3 A brief explanation of these defining characteristics
follows.
First, we are only concerned with those financial contracts that identify
the transfer of capital to corporations. We do not therefore address
government or municipality financing for their own account. We do,
however, include the taxable issues of corporate agencies that are often
described as quasi-governmental organizations.
Corporations source external capital through the sale of financial
contracts. The sale may be conducted as either a public or private
transaction. Representative private transactions are the private placement
of securities and loan contracts. Private transactions are not considered.
Public offerings are brought to the public market either directly by the
issuer or indirectly through the services of broker/dealers. Examples of
direct offerings such as shareholder dividend reinvestment or stock purchase
plans and the direct sale of rights or warrants typify the direct issue from
the firm to the investor. Indirect public offerings engage the services of
broker/dealers who bring the issue to market at a publicly announced
offer price. Indirect public offerings form the base set of the target market
for underwriting services.
Indirect offerings employ the services of broker/dealers on either a
"best efforts" or a "firm-commitment" basis. Firm-commitment under-
writing involves the direct purchase of the offer from the issuer by one or
more broker/dealers who agree to resell the issue at a specified public
offering price. Firm-commitment underwriting, therefore, incorporates a
distinct element of insurability whose value would be incorporated in the
discounted price to the issuer from the public offer price. Best efforts
underwriting does not entail the offering's purchase by the broker/dealer.
Instead, the broker/dealer simple agrees to extend its "best effort" to the
task of selling the offered securities at the public offer price. Only firm-
commitment offerings are analyzed.
Issuers identify firm-commitment underwriters either through a nego-
tiated or a competitive bid process. 4 Competitive deals result from sealed
216 THE CHANGING MARKET IN FINANCIAL SERVICES
bid auctions in which the syndicate is completely specified and the firm
hires the syndicate with the highest bid. The typical firm commitment
offering is serviced by an underwriting syndicate under the control of a
lead manager.s Underwriting services that are provided through negoti-
ated dealings between the corporation and the broker/dealer syndicate
offer the managing underwriter degrees of freedom in distribution that
may not obtain through the competitive bid process. The negotiated deal
is the primary focus of this paper.6
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
(1) Common Stock· 13.47 14.44 14.18 37.40 14.20 29.33 63.37 76.39 42.90 38.08
(2) Domestic Debt b 37.09 35.87 40.74 46.05 65.32 97.75 219.76 210.48 234.54 272.83
(3) Preferred Stocke 1.90 1.18 4.52 5.14 3.12 6.48 9.89 7.64 7.47 7.61
(4) Convertible Debt d 4.35 4.65 6.00 11.61 12.01 15.01 19.16 27.37 15.91 20.79
(5) Convertible Prefd 1.29 0.46 0.51 3.15 0.82 2.38 4.82 4.30 1.07 1.58
(A) Total (1-5) 58.10 56.60 65.95 103.35 95.47 150.95 317.00 326.18 301.89 340.89
(6) Int'l Debt" n.a. n.a. 62.31 62.47 90.41 151.13 205.30 155.43 207.93 231.14
(B) Total (1-6) 58.10 56.60 128.26 165.83 185.89 302.13 522.38 481.68 509.91 571.52
• Common stock offerings both domestic and foreign excluding debt for equity swaps and rights.
b Domestic, straight debt valued at offer price.
d Convertible offerings include both domestic and foreign issues valued at offer price.
e International debt includes Eurobond and non-Yankee foreign bond issues, expressed in U.S. dollar equivalents at the time of
issuance.
I Public underwritten offerings includes firm commitments only and excludes private placements.
218 THE CHANGING MARKET IN FINANCIAL SERVICES
The Industry
Number
Total 5,714 9,021 1.58
Foreign 37 121 3.27
Assets ($ Bn)
Total 150.5 649.9 4.31
Foreign 12.4 143.5 11.56
% Foreign 8.2 22.1
Equity ($Bn)
Total 10.0 36.0
Foreign 0.5 4.4
% Foreign 5.0 11.4
Equity/Asset Ratio
Total .066 .055
Foreign .040 .031
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
Volume ($Bn)
Top 25 57.54 56.04 95.35 138.28 152.87 254.43 446.32 417.49 442.96 505.93
Domestic Mgr 51.07 51.08 54.27 89.14 92.37 150.32 275.04 253.76 273.65 305.06
Foreign Mgr 6.47 4.96 41.08 49.14 60.50 104.11 171.28 163.73 169.31 200.87
Issues (#)
Top 25 986 1009 1478 2303 1700 2393 4052 3546 4657 5372
Domestic Mgr 908 940 958 1705 1130 1746 2645 2241 2281 3889
Foreign Mgr 78 69 525 598 670 1047 1407 1305 1376 1483
Average Volume ($ Mn)
Domestic Mgr 56.24 54.34 56.65 52.28 81.74 86.09 103.98 113.24 119.97 78.44
Foreign Mgr 82.95 71.88 78.25 82.17 90.30 99.44 121.73 125.46 123.05 135.45
Gross Spread ('Yo),
Domestic Mgr: Mean 3.646 4.138 3.552 4.483 3.319 3.004 2.938 2.661 1.921 1.599
Foreign Mgr: Mean 2.223 2.472 2.103 2.250 1.834 1.848 2.036 2.018 1.979 1.939
I Full credit for issue allocated to book manager. Top 25 Underwriter status satisfied at least once. Appendix A provides the list of
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
(A) Total (1-5)' 58.10 56.60 65.95 103.35 95.47 150.95 317.00 326.18 301.89 340.89
Underwriter Totals
Top 25 57.54 56.04 50.08 94.61 86.42 143.29 296.83 301.42 282.41 317.07
Domestic Mgr 51.07 51.08 43.62 82.75 72.38 115.99 241.75 238.38 242.62 265.52
Foreign Mgr 6.47 4.96 6.46 11.86 14.04 27.30 55.08 63.04 39.79 51.55
Market Share of (A)
Top 25 0.99 0.99 0.76 0.92 0.91 0.95 0.94 0.92 0.94 0.93
Domestic Mgr 0.88 0.90 0.66 0.80 0.76 0.77 0.76 0.73 0.80 0.78
Foreign Mgr 0.11 0.09 0.10 0.11 0.15 0.18 0.17 0.19 0.13 0.15
(B) Total (1-6)' 58.10 56.60 128.26 165.83 185.89 302.13 522.38 481.68 509.91 571.52
Underwriter Totals
Top 25 57.54 56.04 95.35 138.28 152.87 254.43 446.32 417.49 442.96 505.93
Domestic Mgr 51.07 51.08 54.27 89.14 92.37 150.32 275.04 253.76 273.65 305.06
Foreign Mgr 6.47 4.96 41.08 49.14 60.50 104.11 171.28 163.73 169.31 200.87
Global Market Share
Top 25 0.99 0.99 0.74 0.83 0.82 0.84 0.85 0.87 0.87 0.89
Domestic Mgr 0.88 0.90 0.42 0.54 0.50 0.50 0.53 0.53 0.54 0.53
Foreign Mgr 0.11 0.09 0.32 0.30 0.33 0.34 0.33 0.34 0.33 0.35
1 Full credit for issue allocated to book manager. Top 25 Underwriter status satisfied at least once .
• (A) and (B) Market Totals repeated here from table 6-1.
Table 6-5. Dollar Volume Concentration By Security Type and Manager Class. Volume Totals-U.S.$ Billions.
Pricing
of the spread. The most general case is that for common stock. We interpret
the intercept term as the average gross spread on a seasoned issue for a
domestic public utility managed by a domestic firm prior to the initiation
of shelf registration. The spread for all other combinations are condi-
tional on active qualitative variables and additive in them. A summary of
the results follows.
First, the fees represented by the spread are inversely related to issue
volume for each issue type. This general result is consistent with previous
research on underwriter fees. Industry type does not enter into the modeled
spread in a systematic way across security types. The incremental fee
charged for common stock issues is positive and statistically significant for
both industrial and financial firms. The level of the coefficient is higher
for financial sector firms, perhaps indicating the increase of risk in their
operations during the decade of the 1980s. Domestic debt demonstrates
a different pattern for industry type. All three groups have significant
coefficients, but financial firms and corporate agencies enjoy a reduction
in costs that may be attributable to governmental guarantees of some
form. The industry dummy on international debt is negative and only
significant for financial firms.
Shelf registration is not modeled for international debt since the data
series begins in 1982, the year that Rule 415 becomes effective. Shelf
THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 227
Data
According to the data in the previous sections, about fifty firms accounted
for more than 95 percent of the underwriting activity during the 1980s. Of
these fifty large underwriters, we found ten firms that had publicly traded
equity in the United States and whose prices were available on the Center
for Research in Security Prices daily data tapes during some part of the
period from 1976 to 1989. We assembled these firms' daily stock returns.
Table 6-7a lists the firms, their initial listing date, the market where they
traded, the date they stopped trading, and the reason they stopped trading.
We also gathered several measures of accounting profitability for these
firms over the ten-year span, 1980 to 1989, from Compustat. Table 6-7b
lists the averages (in percent) of the firms' yearly return on investment,
return on assets, return on equity, year-to-year percentage changes in
sales, cash flow, and assets. It also lists the ten year geometric averages.
Most of the averages suggest that the mid-years of the 1980s were the
most successful for the firms. However, the trend since then has generally
been down; most of the recent observations are below their averages. For
example, while sales continue to grow, internal cash flows have not kept
pace.
There were at least 550 completed deals during the 1980s involving
domestic broker/dealers and investment banks. Of these 550 deals, ninety-
four involved foreign acquire. Figure 6-1 shows that the number of for-
eign acquire has been growing over the decade, with 28 completed deals
in 1989 alone. We identified 47 announcement dates for deals of $25
million or more involving foreign acquire/investors in domestic broker/
dealer firms. These deals included partial investments, such as Nippon
Life Insurance Company's purchase of a 13 percent stake in Shearson-
Lehman for an estimated $535 million, and Yasuda Mutual Life Insur-
ance Company's purchase of an 18 percent voting stake in Paine-Webber
for an estimated $300 million.
Foreign firms may also enter the U.S. market directly by registering
with the SEC as a broker/dealer. We have identified several registration
dates for foreign firms from confidential internal sources.
Table 6-7a. Domestic Broker/Dealers with Publicly Traded Equity in the 1980s.
Table 6-7b. Various Accounting Rates of Return and Growth Rates. Average of
Ten Publicly Held Domestic Broker/Dealers.
140
1 :::r:::'::::::~ I 1
1201~------------------------------------------------~:
100r---------------------------------------------~~~.
M ~~
60 ~~~~~~r~~~t ::::::::
40 I f:-:':':':':':' x x
1980 81 82 83 84 85 86 87 88 1989
Methodology
Results
0.6T'~~~~~-~'r'------~--------~
0.5 I I 1\ I
0.4 I II ~l l
0.1--+1+1-+--
-0.1
- 0.2 I , \ Af ,"", I
-0.3~--__~------~----~------~----~------~----~------~----~----~
81 82 83 84 85 86 87 88 1989
consolidation within the industry. Figure 6-2 shows that the cumulative
abnormal return peaked late in 1982 at around 60 percent. The early part
of 1983 was disastrous for the underwriters as a group, which lost not only
the cumulated 60 percent but also an additional 10 percent by the end of
the year. The years 1984 and 1985 generally were good ones for investors,
only to be followed by a significant downturn early in 1987. Since the
October 1987 market break, the abnormal returns have been generally
positive, moving from a negative 20 percent cumulative abnormal return
to a positive position of about 15 percent at the end of the decade.
Accounting rates of return may provide additional insights into the
competitive position of domestic underwriters during the 1980s. Table
6-7b presents yearly average accounting rates of return (on investment
(ROI), on assets (ROA), and on equity (ROE» for the ten firms in our
sample. Rates of growth in assets and sales are also presented.
The stock market is forward looking, whereas the accounting data are
historically oriented. The CARs are likely to be leading indicators of
subsequent accounting results. Figure 6-3 shows the yearly change in the
cumulative returns (lead one year); the change in average rate of return
on equity; the average yearly growth rate in sales; and finally, the number
of completed acquisitions (both foreign and domestic) in the industry for
the year. In all years, increases (decreases) in CARs over the past year
were associated with increases (decreases) in sales growth in the current
year. In six of the eight years covered, increases (decreases) in last year's
CARs were associated with increases (decreases) in this year's return on
equity. Finally, in five of the seven years in which there were appreciable
mergers and acquisition activity, increases (decreases) in last year's CARs
were associated with decreases (increases) in this year's mergers and
acquisitions (M&A) activity.
The association of merger and acquisition activity with poor perform-
ance is consistent with the notion that mergers consolidate an industry's
assets, particularly an industry that faces significant fixed costs. Thus, this
evidence indicates that individual merger and acquisition in this industry
is likely to have a mixed impact on the rest of the firms in the industry.
In the next section we look at the impact of this activity on the equity
values of the industry's publicly traded firms.
Changes in Rates
100
I] ROE §
I Sales
E
80
~CAR §
60
I3M&A ~
§
§
40
§
•
§ ~
Jl
20
j §
o m
r
~ AI
•
I
~ f;i;J
-20
~
-40
-60 I I I I I I I
1982 83 84 85 86 87 88 1989
Figure 6-3. CARs as leading indicators of changes in growth rates and M&A
activity.
(lower product prices) and/or in the input market (higher input prices).
These changes are likely to lower the market values of the entrant's
domestic rivals. We attempt to test for this effect by examining the abnormal
returns to a stock portfolio that includes the major publicly traded domestic
underwriters around the announcements of foreign entry into the domestic
market.
Of the 550 completed acquisitions in this industry, we found forty-
seven that involved foreign purchasers of domestic firms or significant
stakes in domestic firms, according to IDD Information Services. These
forty-seven deals were $25 million or larger. We identified the announce-
ment dates of these foreign acquisitions of domestic dealer/brokers and
isolated the abnormal returns (described above) to the portfolio of domestic
underwriters on those particular dates. A two-day window was used, the
day of the announcement and the next day, in case the information arrived
after the close of business on the announcement date. These forty-seven
two-day CARs were then averaged to measure the average impact of an
announcement of a foreign acquisition of a domestic dealerlbroker. Overall,
THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 235
The overall picture one can take from the data presented here is that the
domestic underwriting business is very competitive, with average spreads
falling through most of the decade and with foreign entry through acqui-
sition or registration providing additional competitive pressure. However,
it also appears that the domestic underwriters have been successful in
expanding their business overseas. So, whereas the domestic market share
of the major domestic underwriters has fallen, their global market share
has expanded such that their total market share has remained relatively
constant.
Also salient is that changing nature of the underwriting business. Debt,
especially asset-backed debt, has become the dominant issue in the do-
mestic market. Financial innovations, such as mortgage-backed securities,
236 THE CHANGING MARKET IN FINANCIAL SERVICES
Domestic Foreign
Alex Brown & Sons Kidder, Peabody Algemene Bank Nederland Lloyds Merchant Bank
Allen Laidlaw Amsterdam-Rotterdam Bank Natwest Capital Markets
Bankers Trust Lehman Brothers Banca Commerciale Italiano Nikko Securities
Bateman Eichler L.F. Rothschild Banque Nationale de Paris Nomura Securities
Bear, Stearns Merrill Lynch Banque Paribas Capital Markets N.M. Rothschild & Sons
Citicorp Montgomery Barclays Merchant Bank RBC Dominion Securities
D.H. Blair Morgan Stanley Canadian Imperial Bank of Commerce Societe Generale
Dain Bosworth Moseley Securities Commerzbank Soditic
Dean Witter Paine Webber Credit Commercial de France Swiss Bank
Dillon, Read Piper Credit Lyonaise S.G. Warburg
Donaldson Prudential-Bache Credit Suisse Union Bank of Switzerland
Drexel Robertson Credit Suisse/First Boston Yamaichi International
E.F. Hutton Salomon Brothers Daiwa Securities
Faherty & Faherty Smith Barney Deutsche Bank
First Boston Stephens Dresdner Bank
First Jersey Thomson McKinnon Hambros
Goldman, Sachs Wertheim Schroder Indosuez
Hambrecht & Quist Wheat, First Industrial Bank of Japan
John Muir Wheat-Butcher Kleinwort, Benson
J.P. Morgan William Blair Lazard Freres
238 THE CHANGING MARKET IN FINANCIAL SERVICES
Notes
1. Worldwide offerings by underwriters registered to do business in the United States are all
publicly offered corporate securities that are distributed in the United States, as well as all
offerings issued internationally. We do not have access to strictly domestic, foreign issues, i.e.,
foreign securities only issued in their home country.
2. The source for most of the data in this section is various issues of Investment Dealer's
Digest and IDD Information Services.
3. See Hayes, Spence, and Marks (1983) for an in-depth analysis of the market for underwrit-
ing services during the decade of the seventies as well as for an excellent historical review of the
market and the investment banking sector. Market identifying characteristics applied here are
essentially the same as those described in Hayes, et.al. (1983).
4. Hansen and Khanna (1990) document that competitive bid deals registered with the U.S.
Securities and Exchange Commission for the period 1972-1986 accounted for no industrial firm
offerings (6807 deals registered) and 28.9 percent of public utility offerings (3589 deals regis-
tered). For the period studied here, offerings representing competitive bids account for 2.8
percent of the business done by the managing firms analyzed.
5. The influence and size of syndicates are decreasing over the period of this study. The
importance of the lead manager in negotiated transactions should not be underestimated. The
managing underwriter establishes the underwriting agreement with the issuer and identifies
the terms of the offering as well as each member's commitment. The lead manager controls the
"pot". Typically, each syndicate member controls only a portion of the securities which it has
agreed to underwrite. This portion is referred to as "retention". The balance is placed in a general
syndicate account-the pot-which can be reallocated by the lead manager to syndicate
members on the basis of their ability to sell. An aspect of the "pot" which has taken on increasing
importance in the eighties is that it provides institutional buyers with distributional conveniences
such as central billing and delivery. These institutional orders are typically placed with the
managing underwriter but may be earmarked by the purchaser for the account of designated
selling agents.
6. A form of negotiated, firm commitment underwriting which is not a part of the market
studied involves "swapping". Swap transactions utilize securities as the medium of exchange in
the purchase of underwritten securities. They provide the dealer with the ability to diversify
security holdings during the period of distribution and give institutions an alternative means of
payment if it faces cash flow constraints.
7. Confidentiality of foreign-controlled registrants provided by the NYSE was requested.
Only aggregated information is therefore presented for foreign-controlled firms. Foreign-
controlled firms which appear in Appendix A were identified through public sources independ-
ently of information provided by the NYSE.
8. The annual reports are referred to as FOCUS Reports, the Financial and Operational
Combined Uniform Single Report: SEC Form X-17A-5.
9. This has been documented primarily for the issues of public utilities.
References
Brown, Stephen J. and Jerold B. Warner. 1985. "Using Daily Stock Returns: The
Case of Event Studies." Journal of Financial Economics 14 (March): 3-31.
THE COMPETITIVE IMPACf OF FOREIGN UNDERWRITERS 239
I want to give particular attention to the findings that the biggest change
in the market shares have accrued in the equity area. It looks more dra-
matic in the article than its actual market impact. The increase in percentage
of equity financings done by foreign vendors has got to be offshore
equity. The authors' statistics, as I understand it, lump international and
domestic equity financings together and they therefore were not in a position
to distinguish between those offerings that were done offshore and those
241
242 THE CHANGING MARKET IN FINANCIAL SERVICES
that were done in New York. I believe that disaggregated statistics would
show that foreign vendors have a negligible share of the U.S. equity market.
The weak position of foreign vendors in the New York market is mir-
rored in other national markets around the world. In those markets, home-
based vendors have, for the most part, a very firm grip on the equity
offering business undertaken there.
Among preferred offerings in the United States, I don't believe that
foreign vendors have a significant position, either. Thus, the explanation
for the authors' preferred stock market share figures must be due to the
inclusion of the Euromarket financings in their data.
Vendor Profiles
ject to much greater fluctuations in value. Thus, a larger equity base would
be required by these vendors to ensure their viability. Although the New
York Stock Exchange (NYSE) has "haircut" requirements that require
the setting aside of certain amounts of equity for specific types of assets,
it is my own personal opinion that these minimum requirements have in
some cases proved to be inadequate.
Gross Spreads
The authors chronicle the squeeze in gross spreads over time. There is no
doubt that the marketplace is becoming increasingly competitive. In the
United States we used to have the standard seven-eights of 1 percent gross
spread for "plain vanilla" bond underwritings. That has been drastically
reduced over the past fifteen years, particularly with the advent of shelf
registrations. If we take the total mix of financings done both in the
United States and in the international markets (the authors' frame of ref-
erence), there has been an impressive squeezing down of the spreads.
The authors' statistical results are intuitively appealing. They are what
I would have expected to find with respect to differences between "plain
vanilla" bonds and such things as initial public offerings. I think that
Nachtmann appropriately cautioned that in looking at these gross spreads
and their changes over time, it is important to carefully note alterations
in the financing mix, particularly on the debt side. The heavy volume of
asset-backed securities and so-called "junk bonds," which carry very high
gross spreads, are bound to influence the overall aggregate results.
I have two further comments on gross spreads. With respect to the
shelf registrations, a recent study provides some evidence that gross spreads
are being influenced by the exposure of the vendor to legal risk. Vendors
don't often have the time for careful examination of the issuer's affairs
before bidding on the financing and, therefore, may be vulnerable to a
charge of lack of "due diligence" in the event that the issuer's affairs
subsequently take an unexpected tum for the worse. This phenomenon
might help to account for the authors' finding that debt gross spreads
didn't shrink quite as much as one might have expected.
The other comment relates to a statement in the article about equity
gross spreads that is probably a [misprint.] There is no evidence in the
authors' paper to suggest that the foreign vendors have cut gross spreads
on equity financings in the United States. The data is not there to reach
that conclusion.
244 THE CHANGING MARKET IN FINANCIAL SERVICES
Stockholder Wealth
overcome. And so I think it's a real tribute to the U.S. investment banks
that, despite these offshore barriers to entry, this paper suggests that they
have nonetheless been able to hold their own in the market environment.
Morgan Stanley, which had only several hundred employees in one New
York location in the early 1970s, now has more than 6,000 employees in
locations all over the world. Given these changed requirements, the
authors' comments about increased concentration in the U.S. securities
industry aren't surprising. There are only a half dozen or so firms that
really count in terms of competitive interactions with foreign owned
securities firms. And that handful of dominant U.S. firms continue to be
world-class competitors.
References
Hayes, Samuel L., and Philip Hubbard. Investment Banking: A Tale of Three Cities.
Harvard Business School Press, Cambridge: 1990.