You are on page 1of 246

The Changing Market in Financial

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

"
~.

Springer Science+Business Media, LLC

Library 01 Congress Cataloging-in-Publication Data

Economic Policy Conference of the Federal Reserve Bank of
SI. Louis (15th: 1992: Federal Reserve Bank of SI. Louis)
The changing market in financial services: proceedings of the
Fifteenth Annual Economic Policy Conference of the Federal
Reserve Bank of SI. Louis/edited by R. Alton Gilberl.
p. cm.
Includes bibliographical relerences.
ISBN 978-94-010-5322-8 ISBN 978-94-011-2976-3 (eBook)
DOI 10.1007/978-94-011-2976-3
1. Financial services industry-United States-Congresses.
1. Gilbert, R. Alton. II. Federal Reserve Bank of SI. Lauis. III. Title.
HG1818.E36 1992
332.1 '0973-dc20 91-40394
CIP

Copyright © 1992 by SpringerScience+Business Media NewYark
Originally published by Kluwer Academic Publishers in 1992
Softcover reprint of the hardcover 1st edition 1992

AII rights reserved. No part of this publication may be reproduced,
stored in a retrieval system or transmitted in any form or by any
means, mechanical, photo-copying, recording, or otherwise, without
the prior written permission of the publisher, Springer Science+Business
Media, LLC.

Printed an acid-tree paper.

Contents

Contributing Authors vii

Preface ix

1
The Opening of New Markets for Bank Assets 3
Gary Gorton and George G. Pennacchi

Commentary by Stuart I. Greenbaum 35

II 39

2
Interstate Banking, Bank Expansion and Valuation 41
Gerald A. Hanweck

Commentary by Peter S. Rose 93

3
The Market for Home Mortgage Credit: Recent Changes and Future
Prospects 99
Patrie H. Hendershott

Commentary by Herbert M. Kaufman 125

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 Commentary by Gary C. Hayes 241 Index 247 . Zimmerman 201 6 The Competitive Impact of Foreign Underwriters in the United States 211 Robert Nachtmann and Frederick J. Phillips-Patrick Commentary by Samuel L.

Kaufman University of Pennsylvania Financial Systems Research Philadelphia. Nellie Liang Leverone Hall Division of Research and Statistics 2001 Sheridan Road Board of Governors of the Federal Evanston. Box 248094 Ohio State University University of Miami 1775 College Road Coral Gables. Hendershott Department of Finance Academic Faculty of Finance P. Virginia 22030 352 Mervis Hall Pittsburgh. 20551 vii . Hayes III Graduate School of Business Administration James M.Contributing Authors Lawrence G. Massachusetts 02163 Washington. Greenbaum Tempe. 20551 Gerald A. Florida 33124 Columbus. Arizona 85287 J. Pennsylvania 19104 Arizona State University Bac 519 Stuart I. O'Brien Harvard University Division of Research and Statistics Baker 333 Board of Governors of the Federal Soldier Field Reserve System Boston. 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. Goldberg Patrie H. Kellogg Graduate School of Management Northwestern University J. D. D.C. Ohio 43210-1309 Gary Gorton The Wharton School Herbert M.C.O. Illinois 60208 Reserve System Washington.L. Pennsylvania 15260 Samuel L.

California 94105 . P. Pennacchi Peter S.C. Rose University of Illinois Finance Department 304G David Kinley Hall Blocker Building. Illinois 61801 College Station. 20552 San Francisco. Texas 77843 Gary C. D. Box 7702 Washington.O. Zimmerman Frederick J.viii THE CHANGING MARKET IN FINANCIAL SERVICES George G. Room 340 1206 South Sixth Street Texas A&M University Champaign. Phillips-Patrick Economic Research Department Office of Economic Research Federal Reserve Bank of San Office of Thrift Supervision Francisco 1700 G Street NW.

and foreign financial firms have taken a rising share of the market financial services. banks gain valuable information about the financial condition of their depositors that is not available to other investors. If banks have special information. The theory of financial intermediation devel- oped in recent years is based on an information asymmetry between banks and other investors: By providing payments services. banks have been permitted to participate in a limited form of interstate banking. their customers. The article by Gary Gorton and George Pennacchi examines why loans sold by commercial banks in recent years have grown rapidly. This observation raises ques- tions about the nature of relationships among banks. This market has been changing rapidly in recent years: Business loans have become more liquid. as the market for loan sales grows. why would others buy loans from them.Preface The articles and commentaries included in The Changing Market in Financial Services were presented at the Fifteenth Annual Economic Policy Conference of the Federal Reserve Bank of St. commercial banks have been given permission to offer additional underwriting services. and those who buy the loans. The art- icles describe these changes and examine their implications for financial institutions and their customers. since banks are likely to sell only the poorest quality loans? Gorton and Pennacchi explore the empirical evidence on several competing hypotheses about why investors would buy loans from ix . about $240 million of commercial and industrial loans were sold. In 1989. The conference focused on the effects of a variety of recent changes in the market for financial services in the United States. Louis. the market for resi- dential mortgage credit has been transformed through securitization and the declining role of savings and loan associations. compared to insignificant amounts five years earlier.

or safety and soundness. Rose also cites the results of his own work as providing one explanation for the lack of positive stock price response to the announcements of plans for interstate mergers reported in several studies. which is the basis for much of the theory of financial intermediation and for bank regulation. banks to attract deposits in relation to their capacity to originate loans. including the downgrading of bank debt by rating agencies. Stuart I. Hanweck finds little evi- dence that regional interstate banking has stimulated competition among banks or increased their operating efficiency.S. Patric H. Hanweck's data are for domestic commercial banking firms only. One important reason for the increase in banking concentration at the national level has been that states currently do not allow banks in other states to enter by establishing new offices. Greenbaum points out that other fac- tors reduce the capacity of major U.S. The authors conclude that technological progress has drastically reduced the information asymmetry between banks and many other investors. changes in state laws to permit intrastate branching. Hendershott . Peter S. tend to underperform other banks.S. compensating them for the moral hazard risk which continues to be important. banking assets of foreign banks or nonbank financial firms. faced with such changes. profitability.x THE CHANGING MARKET IN FINANCIAL SERVICES commercial banks. but raises questions about the nature of the concentration data and the reasons for the rise in concentration. Greenbaum suggests in his commentary that Gorton and Pennacchi may be exaggerating the role of technological change in ex- plaining the growth of loan sales. Gerald A. banks. instead. Interstate banking would induce greater com- petition if banks could expand their operations across state lines by estab- lishing new offices. Rose shares Gerald Hanweck's concern with rising banking concentration at the national level. and those to be acquired in interstate mergers. The primary impact has been a substantial consolidation of banking resources within the largest fifty or one hundred bank- ing companies. Rose finds that both the acquiring firms. One of the major developments in this market has been the securitization of fixed-rate mortgages. to a large extent. they must enter by acquiring existing banks. The rise in concentration also reflects. made loan purchases more attractive to potential investors. He suggests that U. they do not reflect the impact of the growth in U. Hendershott takes a look at changes in the market for home mortgage credit in the 1980s. Hanweck analyzes the impact of regional interstate banking on banking competition and bank performance.

O'Brien estimate the distribution on returns to underwriters on individual issues of corporate equity securities. Another major change has been the growth of adjustable-rate mortgages. Goldberg describes the growth of foreign banks in the United States and summarizes the results of studies of the motivations for this growth. Lawrence G. Nellie Liang and James M. with the underwriting spread more than covering declines in share prices around the time new issues come to market. Gary C. One such issue is the status of government agencies that have developed the market for mortgage-backed securities. The other policy issue that Kaufman addresses is the rationale for specialized mortgage lenders.PREFACE xi reports evidence that securitization has affected both the level and pat- tern of fixed-rate mortgage interest rates. and that. He finds no support for arguments that foreign banks have com- petitive advantages over domestic banks. the variability of returns would primarily reflect the variability in the frequency of underwriting. Zimmerman points out . Kaufman focuses on some of the policy issues raised by Hendershott's article. once established. 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. 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. J. Their estimates indicate that the frequency of losses is small. making these rates move more closely with other long-term rates. In his comment on Goldberg's paper. Goldberg concludes that several of the determinants of foreign bank presence indicate continuing future growth of foreign banks in this country. Liang and O'Brien conclude that large commercial banks are relatively well positioned to gain entry into corporate securities underwriting. A third major development in the 1980s was the rapidly declining role of savings and loan associations as providers of residential mortgage credit. not in the returns on individual issues. The continued growth of these mortgages is likely to be limited by the extent to which these loans are securitized. Herbert M. reductions in the share of mortgage credit have not raised mortgage interest rates. 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. Kaufman argues that because of the integration of the mortgage and capital markets. The variability of equity underwriting returns primarily reflects a variability in the number of offerings.

They characterize the domestic underwriting business as very competitive. Hayes. Hayes assumes that securities firms primarily underwrite debt issues outside their home markets. Alton Gilbert Assistant Vice President . equity underwriting is done primarily by domestic underwriters. focuses on the composition of equity underwriting that is subject to foreign competition. Thus. The share of securities underwritten in the United States by foreign securities firms had risen in the 1980s. with falling underwriting spreads throughout most of the decade and entry by foreign securities firms providing competitive pressure. 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. Samuel L. the global market share of firms headquartered in the United States had remained relatively constant in the 1980s. are raising funds in the United States and lending them in Japan. These changes in shares reflect the integration of world financial markets with no evidence of net disadvantage for domestic securities firms. On net. the Japanese banks.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. 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. Hayes expresses doubts about the ability of empirical tests to indicate the competitive effects of entry by foreign securities on domestic firms. R. Zimmerman's analysis indicates that. but the share of securities underwritten in other countries by firms headquartered in the United States had also risen. in the United States as in other countries. Robert Nachtmann and Frederick J. Phillips-Patrick study the market shares of domestic and international securities firms in underwriting corporate securities. Hayes assumes that. in his comments on the article by Nachtmann and Phillips-Patrick. Zimmerman provides additional perspective on the issue of competi- tive advantage of foreign banks. 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. rather than using relatively low cost funds raised in Japan to lend in the United States. on net.

I .

These two methods of raising capital are not identical. The illiquidity of bank loans underlies the rationale for bank regulation and deposit insurance because. should be nonmarketable. the absence of information-revealing markets for these securities led to banking panics. Pennacchi Introduction When a corporation seeks to raise resources. 1 THE OPENING OF NEW MARKETS FOR BANK ASSETS Gary Gorton George G.l A key implication of theories explaining the uniqueness of bank lending is that bank loans should not be resold once originated by banks. at least it is changing significantly. Empirical evidence sug- gests that a bank loan is significantly different from a marketable secur- ity. 2 In fact. 3 . In the last few years many banks have been able to sell substantial amounts of loans. historically. it can borrow from a bank. or alternatively. (Possibly bank regulation would also end. it can issue securities such as bonds or equities.) If banking is not coming to an end. and there- fore. one might say that if banks were able to sell loans. it would be the end of banking. Bank loans are assumed to be difficult to value by investors. The growth of markets for bank assets has been remarkable.

and wealthy individuals. although recently. The primary buyers include professional money managers. No recourse.4 THE CHANGING MARKET IN FINANCIAL SERVICES In a few short years the outstanding principal of repackaged automobile loans and credit card receivables sold has grown from nothing to about $50 billion. suggesting that certain kinds of loans are more attractive candidates for resale. as well as differences in the types of investors who buy them. In contrast to mortgages and consumer loans. At the same time. the loans that are sold must be sold without recourse to the selling bank. guarantee. These pools of loans are "credit-enhanced" by being guaranteed or insured by a third party guarantor and almost always receive triple A ratings. These loan sales.4 Hence. if a bank is motivated to sell loans as a way of avoiding the financing costs associated with required reserves and capital. However. or sec- ondary loan participations. and the beginnings of secondary markets for the claims on loans. bank trust departments.3 Asset-backed securities typically represent senior claims on a pool of homogeneous consumer bank loans such as automobile loans or credit card receivables. 5 Recent information-based theories of financial intermediation argue . in fact. commercial and indus- trial loans are almost never pooled when sold. the volume of commercial and industrial loan sales has grown from insignificant amounts to $240 million in 1989. The single loan underlying a loan sale is often the unsecured obligation of a non-investment grade firm. Loans such as mortgages and consumer re- ceivables are typically pooled and then sold as mortgage. There also appears to be important differences in the types of contracts used to sell loans. Growth in volume has been accompanied by a deepening of the new markets as maturities of securitization contracts have lengthened. pension funds. There is usually no rating. the ratio of loans originated to loans sold varies significantly by loan type. insurance. are proportional (equity) claims to the cash flow from a single commercial or industrial loan. or credit enhancement is included. The typical buyers of commercial and industrial loans are foreign banks and smaller domestic banks.or asset-backed securities. an increase in the average risk of the underlying loans. Since nearly all loans sales are. it must not provide any guarantee of a loan's quality to either a third party buyer (in the case of commercial and industrial loan sales) or a third party guaran- tor (in the case of asset-backed securities). about one quarter of the buyers have been nonfinancial firms. so that the proceeds obtained from loan buyers are not classified as deposits subject to required reserves and required capital. sold without recourse to the originating bank. Banking regulations require that for loan selling banks to remove loans from their balance sheet. they seem to contradict the nonmarketability of bank loans.

One service deals with the production of infor- mation regarding the credit risk of a potential borrower. they will be the first to suffer losses). to perform the services associated with loans without being required to hold a claim on the loans until maturity. This is the basis for the term intermediary. no individual would be willing to pay the costs. 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.7 If loans were sold without recourse. This would have the effect of inducing banks to hold the loans rather than sell them. that is. such as the case when the loan is held on the bank's balance sheet or sold with recourse or a guarantee. Thus. Realizing that this problem exists. willing to rely on the bank to continue to perform . the bank would have no incentive to perform these services. a bank. perform these services (since otherwise. at least partially. equity holders would no longer be at risk for nonperformance.THE OPENING OF NEW MARKETS FOR BANK ASSETS 5 that banks provide special services that cannot be obtained when securities are offered to the public. The analysis equates the existence of banks with the nonmarketability of their loans. or third party guarantors in the case of asset-backed securities. In other words. Apparently. an information asymmetry exists between banks and depositors (and between banks and other outsiders such as potential loan buyers). Individual investors may have a very difficult time undertaking these activities because of "free riding" by investors. These loans would decline in value because potential loan buyers would realize this incentive problem exists and offer less to buy the loans. rather than duplicate each others' costly efforts. What is this new mechanism for insuring incentive compatibility? In other words. 6 The difficulty with delegating these services to banks is that depositors are unlikely to have the ability to directly observe banks' performance. the existence of banks is synonymous with the creation of assets which must be held until maturity. this argument appears to be directly contradicted by the recently observed experience of banks selling loans without recourse or guarantee. Free riding prevents individual investors from recovering the costs of producing information or monitoring. in fact. a superior arrangement would be to delegate the tasks to another agent. namely. But. The argument is also contradicted by the willingness of third parties to credit enhance asset-backed securities. Even if individuals were willing to bear the costs. why are third party buyers. Another service refers to the monitoring of borrowers so that loan covenants can be enforced while loans are outstanding. banks can be induced. This insures that the bank equity holders will see to it that the bank will.

if the credit risk of the underlying borrower increases. A number of articles have addressed banks' motivation for loan sales and have pointed out the incentive compatibility problems associated with .8 This paper considers the development of the market in which banks sell their loans. requires selling banks to hold a fraction of the loan. as we shall see. There is a third possibility that could help explain the feasibility of loan sales. If these implicit contracts exist. If the banks' actions. in contradiction of the assumptions of aca- demic researchers and the implicit assumptions of policymakers. and nothing requires the bank to guarantee the value of loans sold. information production and monitoring. then the bank will continue to have at least a partial incentive to perform on the loan. A number of hypotheses may explain how markets for bank assets could develop.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. could explain the apparent paradox of loan sales. Nothing in written loan sales contracts. bearing the losses. 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. However. there is the possi- bility that the bank offers an implicit (unobservable) guarantee or insur- ance on the value of the loan. Reductions in the cost of information production and trans- mission would be at the root of such a change. how these two contract features could be enforced. Since this arrangement would be implicit. retaining the rest. A second possible mechanism for making loan sales incentive compat- ible involves the selling bank holding a fraction of each loan sold. 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. it is not clear. First. or the two features combined. it would likely be hard to detect though Gorton and Pennacchi (1989) claim to have found some evidence for this contract feature. then it may be possible to enforce bank performance without directly requiring them to risk their equity. at least for commercial and industrial loan sales. If the selling bank sells only a part of the loan. Perhaps. loan buyers or guarantors can observe whether or not banks produce information and monitoring. are observable (unlike in the past). Either of these two possibilities. We explore three hypotheses about mechanisms that could make loan sales incentive compatible. Perhaps the bank is willing to buy the loan back. 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. then market forces (as opposed to legal forces) must enforce them.

the underlying loans may all be automobile loans or credit card accounts. 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. A pass-through structure allows for a sale of the receivables or loans to the investors. Section five introduces a simple model that provides the basis for more sophisticated tests. the total issuance of asset-backed securities during the years 1985 to 1988 was $36. for example. Asset-backed securities. we begin in the second section by briefly explaining asset-backed security contracts and by re- viewing recent developments in asset-backed security markets. $23.3 billion (9 percent) was backed by other types of loans. In 1989 the . Asset-backed Securities An asset-backed security is a claim on a portfolio of underlying bank loans that have some common features. These are briefly discussed in the concluding comments of Section seven. as with mortgage-backed securities. while a pay-through security is a debt of the issuer backed by the receivables. One reason for this is a lack of data to analyze. are generally issued in one of two forms. Excluding mortgage-backed securities.lO The underlying loans have the feature that they are fairly standardized loan contracts and have well known risk- return characteristics (including historic default experience and prepayment propensities). While the focus of the article is on loan sales. there has been little empirical work inves- tigating the possibly implicit contractual features that banks might use to alleviate these incentive problems. Section three is devoted to briefly tracing and explaining the growth of the (in- dividual) loan sales market.2 billion.1 billion (64 percent) was automobile loan-backed securities. So.THE OPENING OF NEW MARKETS FOR BANK ASSETS 7 selling bank assets. We then review some earlier work that has tested the various hypotheses. $9. In fact. the same issues arise for asset-backed securities.8 billion (27 percent) was credit card receivable-backed securities and the remainder of $3. Section four begins the discussion of the relative merits of the first two hypotheses explaining the existence of these markets. 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.9 however. car loans and credit card receivables are the two leading examples of loan types used to construct asset-backed securities. Of this. either pass-through or pay- through. The first tests aimed at testing these hypotheses are reviewed.

8 THE CHANGING MARKET IN FINANCIAL SERVICES Table 1-1.9 0. 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). total issuance was over $50 billion. This percentage is usually significantly in excess of the loans' historic default rates. In some issues of asset-backed securities.4 31 1988 14. Asset-backed Securities Issuance by Type of Issuer (1988) ($ billions).7 2.0 Credit Card Receivables 5. Commercial Dealer Manufacturing Backing Loans Banks Thrifts Affiliates Subsidiaries Retailers Car Loans 3. Credit protection provided by a third party usually covers all losses up to a fixed percentage of the principal.041 9.0 16 1987 9.237 0.3 0. a straightforward recourse provision is included under which the third party . NA: Not Available Source: Goldman.0 0.134 6.) Backed Securities Backed Securities of Issues 1985 1.2 0. Sachs & Co. mentioned above.0 0.939 5.5 0.3 0.54** NA NA 27 * Principal Amount ** As of August 30th.5 7. Table 1-2 provides information about the types of issuers of these securities.0 2.1 Other Loans 0.0 Source: Goldman. Total Car Loan Credit Card Total Number Year Amount* ($ bit.3 0. Sachs & Co.0 7 1986 10. Credit enhancement is provided by a third party credit facility (such as a letter of credit or surety policy) as well as over-collateralization. Table 1-2.0 0. Asset-backed Securities Volume (Issued Annually). Table 1-1 provides the data for the recent history of the asset-backed market. One important feature of asset-backed securities.0 1.9 0.1 1. is the presence of credit enhancement devices in the contract.6 0.4 59 1989 9.

The largest of these were by Citicorp which (through its subsidiary. and legal purposes.078 billion.THE OPENING OF NEW MARKETS FOR BANK ASSETS 9 guarantor looks to the issuer to back up the guarantee. Recent innovations in automobile-backed securities include the use of a senior/subordinated structure and reductions in the extent of the credit . Initial agree- ments can be made such that loans added to the pool are required to surpass a certain quality based on analysis from credit scoring models. General Motors Acceptance Corporation had 16 offerings during this period. however. and other expenses. The trust is administered by an independent trustee. ll For automobile-backed securities. Inc. Banks. utilizing a grantor trust. 19 of these offerings were by financial subsidiaries of major automobile companies and not by commercial banks. Sometimes the spread is used to support the credit enhancer. Citicorp Acceptance Company. However. Principal and interest are passed through monthly as received (with a lag of 24 days). this problem is likely to be minimized due to the ability to objectively classify the credit risk of these rather homogeneous loans. Since then. a potential incentive compatibility problem exists between the issuing bank and the third party guarantor. the pass-through. The simplest pass-through structure has certificates paying a lower coupon than the rate paid on the receivables. regulatory accounting practice (RAP). Another important feature of asset-backed securities is the existence of secondary markets. In the period from 1985 to 1987 there were 27 offerings. the issuing bank can be forced to buy back these below standard loans. If loans are later found to have not met these initial requirements. the market has grown rapidly.) issued two automobile pass-through securities totaling $1. These markets appear to be increasingly liquid. hence. In particu- lar. which then issues certifi- cates representing undivided interests in the trust assets. auto- mobile receivables are sold to the grantor trust. In a typical pass-through. In that year commercial banks were the major issuers. tax. During 1988 there were 27 new offerings. cannot use such a mechanism since it is inconsistent with the regulatory definition of an asset sale. guarantee fees. is the most common format because it achieves sales treatment for generally accepted accounting practice (GAAP). Automobile Receivable-backed Securities The automobile receivable-backed securities market began in May 1985 with a $60 million offering by Marine Midland Bank. The spread between the average loan rate and the certificate coupon rate is used to pay servicing fees. Importantly.

so that there is less need for pooling to obtain economies of scale in selling them. being revolving credits. or secondary loan participations. are quite distinct from asset-backed securities because they typically do not involve the cash flows from a pool of loans. unlike most other loans. and (2) principal pay-down (or amortization).13 The Loan Sales Market Sales of commercial and industrial loans. Each month's activity in each account can vary. The seller class must absorb variations in the size of pool balances. These developments seem consistent with a growing maturity of this market since these changes parallel the previous transformation of the mortgage-backed securities market. Variations in the principal amounts are handled via a two-class payment structure: an investor class and a seller class. followed by an amortization period. so that the principal and interest cash flow from the total pool will also change. usually representing 60 percent to 80 percent of the original pool principal.10 THE CHANGING MARKET IN FINANCIAL SERVICES enhancement. The investor class has a fixed principal amount. During the interest only period. have no natural maturity. commercial and industrial loans tend to be of much larger denomination than most consumer loans and mortgages. Therefore. since all the risks of the pool are shared with the investor class pro rata. First. Second. Investors purchase a fixed amount of principal with a fixed coupon rate. the seller class is not subordinate. There are generally two forms of distribution to investors: (1) interest only (or nonamortization). all principal payments are allocated to the seller class and are used to purchase all the additional receivables which have been added to the designated accounts. It appears that these loans are not pooled for at least two reasons. their lack of homogeneity makes them less easily classified by credit scoring methods . The cash flows from these accounts consist of principal and interest. Interest is paid monthly at a fixed coupon rate. credit card securities have the dis- tinguishing feature of a period during which only interest is paid (usually eighteen months to three years). Credit Card Receivables Credit card receivable-backed securities are claims on a designated (and selected) pool of outstanding credit card accounts. However. Credit card accounts. but from a single loan. 12 The seller also receives any increases in receivable balances as well.

execution or delivery by any Borrower of any Participated Loan or Loan Documents. and by 1989 only 35 percent were of investment grade borrowers. any representation or warranty or information made or furnished by any Borrower. Also. sufficiency. that is. validity. they come without explicit issuer or third party guarantees. that is. the performance or observance by any Borrower of any of the provisions of any Loan Docu- ments. indicating a lengthening of loan maturities. if the underlying buyer of the loan fails (as has occurred in several cases) the loan buyer has no recourse to the selling bank. Initially. however. 16 While this fraction rose somewhat in 1990 to 44 percent. the value. the legality.THE OPENING OF NEW MARKETS FOR BANK ASSETS 11 than consumer loans. en- forceability or collectibility of any Participated Loan or Loan Documents. the overall trend has been toward the sale of higher risk 10ansY . and shall have no responsibility for. This would appear to make it more difficult and costly for a third party guarantor to analyze the overall credit risk of a pool of these loans. validity. 80 percent of the loan sales contracts had maturities of ninety days or less. the due authorization. 14 Moreover. They do not necessarily have standardized contracts with readily known risk-return characteristics. the loans that were sold were very short-term obligations of banks' best corporate customers. Loan or Loan Documents. Today. and the Bank makes no representation or warranty as to. Background on the development of this market can be found in Gorton and Haubrich (1988). less than half of these loans are the obligations of publicly rated investment-grade firms. or any other matter concerning. the financial condition of any Borrower or any other matter relating to any Borrower. The important feature of loan sales. the loan buyer has no legal relationship with the underlying borrower since the loan sale contract is strictly between the bank and the loan buyerY In the early 1980s the incentive problems associated with loan sales did not seem as great as they do now. The key provision of the secondary loan participation contract is the following. taken from a typical contract: Each Participation purchased by the Participant hereunder will be without recourse to the Bank and for the Participant's own account and risk. By mid-1987 over half of the loan sales contracts had maturities exceeding one year. that distinguishes them from asset-backed securities is the fact that loan sales offer no explicit mechanism for reducing the risk to the loan buyer. In other words. any collateral or other support for any Participated Loan. by mid-1987 less than half the loans sold were the obligations of investment-grade borrowers. perfection or priority of. very well-known firms with triple A credit ratings. In 1985.

Paradoxically. To date. we will then ask how it is enforced. loan sales are not obviously an innovation since the loan sales contract itself has been in use for decades. Over three- fourths of loan sales are purchased by foreign banks (38. 18 Thus. we will first proceed by attempting to identify the mechanism that makes loan sales incentive compatible. loan sales do not appear to represent a simple growth of the correspondent banking network.) Because loan sales seem to be such a clear contradiction of the nonmarketability of bank assets. Thus. (Table 1-4 provides the amounts of loans purchased by the top twenty-five loan selling banks in the second quarter of 1990. although the volume sold under the contract was not particularly import- ant.3 percent) or domestic banks (37. it is the historical illiquidity of these bank assets that has been the basis of deposit insurance. to over $200 million. The outstanding volume of commercial and industrial loan sales has grown from minuscule amounts of loans. As mentioned in the Introduction. nine money center banks accounted for approximately 69 percent of loan sales.6 percent). If incentives for the selling bank are not maintained. the change allowing for the sudden spurt in sales of loans cannot be accounted for by any particular change in the contract per se. the value of the loan will decline. In addition. To analyze the existence of markets for bank assets. Implicit Loan Sales Contracts: The First Tests Loan buyers must be assured that the bank originating the loan will continue to behave as if the loan had not been sold. In 1990. most buyers of loans have been other banks. three possible mechanisms are: (1) implicit insurance by the selling bank against possible losses to the buyer due to failure of the underlying borrower. essentially sold through the traditional correspondent network. (2) maintenance by the selling bank of a stake in . however. In the past it appears that it was not technologically possible to provide such assurances without the bank directly risking its equity. This growth is shown in table 1-3. Table 1-4 provides informa- tion about the loan sales volume of the top twenty-five loan selling banks in the second quarter of 1990. Something in the environment must have changed to allow this existing contract to support the market. though the proportion purchased by nonbanks is becoming significant.12 THE CHANGING MARKET IN FINANCIAL SERVICES The ability of the loan sales market to incorporate increasingly risky loans has been accompanied by an enormous growth in volume over a very short period of time. If we can identify this mechanism. they are of most interest to us.

07 02 190.1 198401 32.6 02 59.94 * Sales reported are gross and exclude sales of consumer loans and mortgage loans.2 03 91.29 198901 272.5 04 50.2 15.3 16.9 02 195.3 Q4 111. .8 19.82 04 258. Date Loan Sales Loan Purchases 198302 26.7 16.9 03 77.5 18.2 198501 54.4 02 81.7 198601 65.0 198801 236.3 03 35.4 16.65 04 286.3 16. Schedule L.Table 1-3. Quarterly Outstanding Loan Sales of Commercial Banks* ($ billions).7 19.20 03 290.89 199001 228.9 Q4 198.2 03 188.8 02 33.0 17.16 02 276.8 198701 162. Also excluded are loans subject to repurchase agreements or with recourse to the seller.8 04 29. Source: FDIC Call Reports.22 03 263.5 04 75.9 17.7 03 26.64 02 248.

A.500 12060755 First Interstate Bank of California 1. 1. N.857 24.920 36.072 267.000 2364900 Citibank.137 2364880 Bank of New York 2. Call Reports.930 113.236 9274037 First Bank.280 7171650 Harris Trust & Savings Bank 581.A.960 o 2365290 Manufacturers Hanover Trust Company 5.048. DSB Number Name Loans Sold Loans Purchased 12060730 Security Pacific National Bank 53.881 72.A.536 52.384 9.038.A.632 7171560 Continental Bank.217.018.602 4426723 Mellon Bank.819 o 7171630 First National Bank of Chicago 4.602.369.461.915. N.591. 8.535 96.000 3420204 Philadelphia National Bank 2.000 150.449 o 2364840 Bankers Trust Company 34.508 62. N. Top 25 Loan Selling Banks in 199002.000 2364985 Chemical Bank 10.116.197 ° 2364893 Bank of Tokyo Trust Company 1.817.004 80. Schedule L.835.000 2364965 Chase Manhattan Bank.000 419. 590.399 57. 3.442 1250487 Bank of New England.000 62. .091 5512345 Signet Bank-Virginia 2.009 o 5370355 NCNB National Bank of North Carolina 530.951 41.000 4.A.000 2365441 Security Pacific National Trust Company 4.253 5512430 Crestar Bank 1.131 2361310 Marine Midland Bank.425.343 2365328 Morgan Guaranty Trust Company of New York 5.613 226.685 2342200 Midlantic National Bank 531.000 768.602. N.259 1250370 First National Bank of Boston 816. 1. 8.Table 1-4. 14.244 12061400 Bank of America. N.507.458.167 Source: FDIC.526. NT & SA 3.696. N.520.409 132.A.104.274 94. N.A.

causing the value of the buyer's participation to unexpectedly decline. loan buyers must be concerned with the possibility that they may wish to sell a loan they have purchased. claiming that there is a need for cash. Suppose a loan buyer discovers that the underlying borrower's credit risk has deteriorated. because of regulatory restrictions. 19 Two factors combine to suggest the possibility of an implicit guarantee on loans sold. Second. The question of the existence of secondary markets is usually important in any security market. Again. can appear in the following way. The bank. The intuition for their procedure was as follows: If loan buyers believe the participation contract literally. Anecdotal evidence suggests that banks sometimes buy back loans they have sold. then. and (3) a lack of information asymmetry between loans selling banks and loan buyers. because of securities-law restrictions. This buyer may approach the selling bank and ask the bank to buy back the loan. though we cannot document the observations. but this is not possible if the bank wants to remove the loans from the balance sheet and thus avoid reserve and capital requirements.22 The bank also knows that the borrower is more risky now than originally. At what price does the bank buy the participation back. if the buyer needs cash. However. It is. If there . banks. only the selling bank is likely to buy the loan back. however. cannot be explicit about guarantees. 20 Thus. motivated by some empirical observa- tions. Implicit Guarantees By the Selling Bank The possibility of implicit insurance is clearly not motivated by any ob- served feature of the contract.THE OPENING OF NEW MARKETS FOR BANK ASSETS 15 any loan sold. the contract explicitly denies the existence of such recourse. First. We first consider implicit insurance or guarantees. 21 The issue of implicit insurance or a guarantee. wanting to provide some liquidity for its participations has already (informally) said that it will repurchase its participations (even though the participation contract is explicit about the bank not being required to repurchase loans). once having bought the loan. Gorton and Pennacchi (1989) indirectly tested for the presence of such an implicit guarantee. so that there is no implicit guarantee or insurance of the loan. then loan sales prices should contain no component which values the default risk of the selling bank. if it does buy it back? There are no data to directly address this question. Both selling banks and loan buyers might prefer that guarantees be explicit. loan sales contracts explicitly forbid resale of the participation.

16 THE CHANGING MARKET IN FINANCIAL SERVICES is an implicit guarantee. but insures that the debt is riskless. then the value of this guarantee would be related to the default risk of the selling bank since the guarantee would be worth less if the bank defaulted. In equation (1) the coefficient a jl would be less than 1 if the bank partially guarantees the debt. a direct pay commitment. The bank's guarantee. f. consider the following model: (1) Yjtis the yield at time t on a debt security of type j.. A risk premium corresponding to this probability of failure is measured by the variable bt • Unfortunately. will be limited by the likelihood of the bank defaulting and. by substitution: . (2) the risk of the firm issuing the paper or receiving the loan. is a measure of the default risk of the borrowing firm. Gorton and Pennacchi (1989). then a jl would equal zero. the variable f. j = 1 in the case of a loan sale and j = c in the case of commercial paper. how- ever. being unable to fully honor the guarantee. 23 Commercial paper yields should contain a component which essentially prices the default likelihood of the bank that provides this backup credit. r t is the opportunity cost of an equivalent maturity riskless investment at time t. however. b t is a measure of the default risk of the bank providing the guarantee on the security. Almost all com- mercial paper is "backed" by a line of credit. equation (1) cannot be directly estimated for either commercial paper or loan sales because there is no data available for calculating the underlying firms' default likelihoods.24 The risk premium attached to the borrowing firm is the coefficient on the measure of the firm's likelihood of default. Gorton and Pennacchi (1989) posited that the yields on loans sold and on commercial paper should be a function of: (1) the opportunity cost of funds to loan buyers and commercial paper holders. In particular. This same logic clearly applies to commercial paper. an irrevocable standby letter of credit. loan sales prices should be related to the selling bank's default risk. In this latter case. The aj/s are assumed to be constant and E jt is an independently distributed error term. If the bank's guarantee is not only credible. were able to estimate the following equation which eliminates the unknown variable f. thus. The guarantee is explicit in the case of commercial paper since commercial paper typically requires some kind of credit enhancement from a bank. or an insurance company indemnity bond. and (3) the quality of the guarantee given by the bank to buy back the loan sold or provide a backup credit line in the case of commercial paper.

Not only did an indirect approach have to be taken to address the question of interest. averaged loan sales yields collected from a survey by Asset Sales Report. Empirical tests of the model. The estimated coefficient on the bank default risk variable. even though there is nothing in the contract about this. b l . both coefficients were significantly different from zero at the five percent confidence level. in contradiction to the first set of estimates. Omitting the details. 2S The coefficient all/ac! measures the relative strength of the weight placed on the borrower's default risk. while the guarantee on loan sales may be an American option. however. but the data set used did not reveal the second possible mechanism for insuring incentive compatibility. only became clear when more data became available. were carried out using weekly data from July 1987 to March 1988. the data set. This result suggests that inves- tors place less weight on borrowers' default risk in the case of loans sold relative to commercial paper. a total of 32 observations. . The more complicated coefficient on b l measures the relative strength of the guar- antees provided by banks on the two instruments. how- ever. this would. The details are provided by Gorton and Pennacchi (1989). One point raised concerns the fact that the guarantee on commercial paper is a European option.8 for each quality borrower (that is. the results were as follows: The estimated coefficient all/ac! is about 0. seem to imply that a stronger guarantee is given on commercial paper that for a loan sold. was not fine enough to unambiguously answer the question of whether implicit guarantees are at the root of the loan sales market. Al\Pl and A2\P2). was significantly negative in both cases. The possibility that the fraction of a loan sold which was retained might playa significant role.26 Basically.THE OPENING OF NEW MARKETS FOR BANK ASSETS 17 Equation (2) is designed to test the relative strengths of guarantees given by banks on selling loans versus guarantees in the form of backup credit lines for commercial paper. Gorton and Pennacchi (1989) were limited by the available data. Gorton and Pennacchi (1989) offered several reasons why these appar- ently contradictory results might be consistent with the hypothesis of the existence of implicit guarantees. Briefly. The model was estimated for ninety day loan sales for the cases of Al \Pl rated borrowers and separately for A2\P2 rated borrowers.

and possibly without consideration of whether or not an implicit guarantee is offered. but this coefficient is not significantly different than zero. 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. its share aver- aged approximately 59 percent. Also.00009) (.0506) (.'t = . = 872. adding the fraction of the loan sold.00098 + . Also.1988. Gorton and Pennacchi (1990A) rerun regression (1) with the new data set.2388 ('b . To see how the fraction of the loan sold affects the results. the selling bank held a portion of 360 of these loans (41 percent).29 The fraction of the loan sold is. as an ex- planatory variable.0079 b - . Perhaps less obvious is whether the fraction retained increases as the underlying borrower's rating worsens. holding the rating of the underlying borrower constant.00051 s (.'t· The results of that regression are as follows: '1 . The fraction of the loan sold by the selling bank was not a variable that was considered by Gorton and Pennacchi (1989).329.27 Table 1-4 provides more detailed infor- mation on the fraction retained. It turns out that selling banks often retain a fraction of the loan sold. of obs.'t.1987 to September 1. s. is significantly related to the spread paid by the borrowing firm. '1 . It is also positively related to the risk of bank failure.'t) + . Note that the loan sale spread. the extent of the fraction sold and the extent of the guarantee are not independent of the price of the loan sale. an implicit guarantee and maintenance of a fraction of the loan by the selling bank are not mutually exclusive possibilities. Certainly. paradoxically.18 THE CHANGING MARKET IN FINANCIAL SERVICES Maintenance of a Stake in the Loan Sold The motivation for the possible existence of the second incentive com- patibility mechanism is empirical. In a sample of 872 loans sold by a major money center bank during the period from January 20. holding maturity constant.28 The behavior of the fraction retained is difficult to understand without consideration of the prices at which loans were sold.0118) (.00010) No. 'b . Of those loans for which the bank held a positive share. R2 = . negatively related to the spread . Causal observation of Table 1-4 suggests that the fraction retained increases as maturity increases. The behavior of this variable could explain the results. (Standard errors in parentheses).

703 .889 . an unobservable guarantee.738 .810 .917 .625 1 Number of Observations 0 3 8 4 1 0 NR Average Fraction Sold . on the loan sale. introduction of the new variable. Clearly.784 . Gorton and Pennacchi (1990A) introduce a model aimed at more rigorously addressing the question of the existence and possible interaction between the fraction sold. not necessarily. so suf- ficiently complicates the analysis that a model is required to be clear about the issues.843 . This is. .THE OPENING OF NEW MARKETS FOR BANK ASSETS 19 Table 1-5. and the price of a loan sold. Perhaps some loans require less monitoring by the selling bank. however.455 1 Number of Observations 8 34 27 20 0 3 A2 Average Fraction Sold . Selling a larger fraction of these loans does not increase the yield paid to loan buyers as much as more risky loans.733 .867 .707 .778 . the fraction sold.746 . a puzzle.771 .746 .608 Number of Observations 3 41 73 64 18 9 A3 Average Fraction Sold . Maturity (days) Rating 0-5 6-15 16-30 31-60 61-90 90+ A1+ Average Fraction Sold 1 1 1 .750 Number of Observations 9 210 206 88 15 10 Source: Money Center Bank.556 Number of Observations 1 1 1 9 3 0 Al Average Fraction Sold . Fraction Sold.826 . s.600 . It may be the case that banks choose to sell larger fractions of less risky loans.

where x e [0. The cost function is given by c(a) = c*a. The bank may not monitor at the most efficient level after having sold its loans. where k' > 0 and kIf ~ O. The bank's problem is to maximize expected profits from the sale of a particular loan.S). Since loan buyers are rational.33 The cost to the bank is assumed to be given by k(r). at the end of 'l" periods. 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. the bank can offer an implicit (partial) guarantee against default of a loan that it sells. of the loan where s e [0.30 The model provides a basis for empirical work. as discussed above. a moral hazard problem exists. and understand the moral hazard problem. The bank has a constant returns to scale technology for monitoring loans. The proportion of a loan sale that the bank promises to guarantee is r e [0. the bank retains the portion (1. s. banks have an incentive to sell loans in order to avoid the costs of required reserves and required capital. is assumed to be uncorrelated with the return on the loan. and where L is the promised end-of-period repayment on the loan. The bank can fulfill this guarantee only if it is solvent at the time the loan matures. they will only buy loans if there is some way of forcing banks to monitor. The bank can sell a portion. The optimal loan sale contract requires the bank to choose a level of . The probability that the bank is solvent. It produces a stochastic return. The return. x. has a cumulative distribution function of F(x. 31 A bank loan is assumed to require "one dollar of initial financing.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. namely. Thus. A Model of Loan Sales In this model. 1]. since banks provide a special service. 1]. Banks can improve the expected return on loans by monitoring borrowers. monitor- ing. the results of which are summarized here.20 THE CHANGING MARKET IN FINANCIAL SERVICES Loan Sale Incentive Compatibility: More Sophisticated Tests This section summarizes the model presented in Gorton and Pennacchi (1990A). Such implicit guarantees are costly because regu- lators do not approve of them. L]. p. where a is the level of monitoring by the bank. a). x.

. r: max fOL (1. the first term of the objective function is the expected return on the portion of the loan return held by the bank. The second term is the expected value of the guarantee the bank gives to the loan buyer.THE OPENING OF NEW MARKETS FOR BANK ASSETS 21 monitoring.e-(r. and e is the Lagrange multiplier on con- straint (iii) of the problem. To obtain a testable implication of the model.L is the Lagrange multiplier associated with constraint (ii) of the problem.1 (3) In the bank's problem.J. s. a) + sr fOL (L . a) .. Omitting the explicit steps needed to solve the bank's problem.LIL] s= '[-'f-J.rfT [f: sxdF(x.-rf)T . s. Constraint (i) is the incentive compatibility constraint.ae-pa ) (5) Using this form. J. a.J. a)] subject to: (i) foL [(1-s)+srp]xdFa (x. r: sp( (JL . 1 (iii) y:5.LI(-rL) (6) (1.c(a) .-rf)'r .[ -'f + 'Is -'f .a)=c'(a) (ii) s :5./ .rp) (4) where (J == exp[(.er/t ] where: ] = 1. a) = L(1. the first order conditions for problem (3) also yield another expression: (Je-(r.e.L) = (k'( r) + e)(1 .rp)[. a) .J. ] is the amount of internal (debt and equity) funding which the bank provides when fraction s of the loan is sold.x)dF(x.)-r]. and the fraction of the loan to be implicitly guaranteed.. The form we assume is 34 xdF(x. it is necessary to adopt an explicit functional form for the relation between the level of monitor- f: ing and the expected return on the loan.yp)[1 + () . (3).srp fOL (L . (3).LIL s = -------:-'----:---- (1. and the size of the implicit guarantee.J.k(r) . the fraction of the loans to be sold.. The relation (4) relates the endogenous vari- ables of the fraction sold and the size of the guarantee.LI( -rL)] .x)dF(x. at the optimum the following relation holds between the choice of the frac- tion to be sold.s)xdF(x.

By choosing a parametric form for the guarantee cost function k(r). 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.22 THE CHANGING MARKET IN FINANCIAL SERVICES In expression (6) note that when s is less than one. In this case. 1988. so that f.1987 to September 1. Empirical Results 2: Different Guarantees on Loan Sales When the bank is assumed to be able to make different levels of partial guarantees on different loans that it sells. r.rp) in equation (6) appears with the wrong sign or that the point estimate of ris negative and significantly different from zero. If the loan selling bank's partial guarantee. is sufficient to determine the bank's optimal share of each loan sold given the guarantee.l = 0. The basic result obtained is that the term (1.35 The empirical results in this case do not support this hypothesis. r. alone. Empirical Results 1: Equal Guarantee on All Loan Sales The model is tested using the sample of 872 individual loan sales made by a large money center bank during the period January 20. the parameter r cannot be directly estimated. is the same for each loan. then it will optimize using equa- tion (4). on all the loans that it sells. Equations (4) and (6) provide the basis for the first empirical tests in Gorton and Pennacchi (1990A). Instead. then we can treat it as a parameter to be estimated using equation (6). note that if one assumes that the bank is constrained to give the same partial guarantee. then equation (6). This result is nonsensical. We now turn to summarizing the results of estimating the model. However. However. indicating that the loan sales buyers are aware of the implications of the fraction sold for the bank's incentives. 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. The two relations define the links between the bank's optimal choice of sand y. We assumed that this function had a simple quadratic form: k(r) = . the equilibrium values of sand rcan be determined. we obtain estimates of the regulatory cost function k(r). Details are provided in Gorton and Pennacchi (1990A). as well as (6). and so. offers no support for the hypothesis.

that the results are not only tests of the hypothesis of immediate interest. 36 The "test" of the hypothesis is essentially to examine whether the estimates of the guarantees on the individual loans are sensible. however. The results again turn out to be unsupportive of the existence of an implicit guarantee hypothesis. It is important to keep in mind. As noted in the Introduction. Thus. the contract suggests that the selling bank is not providing any guarantee. is the fact that it assumes that loan buyers can force the bank to honor the guarantee or enforce the restriction that not all of the loan be sold. Technological Change Creates Symmetric Information The results of estimating the model in search of evidence for an implicit guarantee are distinctly negative. Since any implicit contract feature would have to be enforceable by market forces. but also of joint hypotheses concerning specific functional forms for the monitoring tech- nology and for the function determining the regulatory cost of issuing implicit guarantees. On the contrary. could explain the incentive compatibility of loans sales. Based on the parameter estimates for k1 and k2 the optimal level of the loan sale guarantee can be calculated for each obser- vation. by itself. This observation is true for any implicit contract which is proposed as an explanation of the opening of the loan sales market.THE OPENING OF NEW MARKETS FOR BANK ASSETS 23 ko + k 1r+ (l/2)k2r2. though there is some weak evidence for the use of the fraction sold as a device for insuring incentive compatibility. however. in fact. or any contract provision which restricts the proportion of the loan which the bank may sell. however. it is not clear how either of these two mechanisms would be enforced. An important feature of the model. the question is raised of whether there is some change in the underlying contracting technology that. There is weak evidence that the fraction sold is used by selling banks in a way which is consistent with incentive compatibility. as the implied values for r for the different loans are predominantly negative. Summary There appears to be no evidence of an implicit guarantee or implicit insurance in loan sales contracts. providing a guarantee. there is nothing in the participation contract that specifies that the selling bank is. Various sources of indirect evidence suggest that significant changes .

or pension funds. one would expect that they should hold a large proportion of those assets where information acquisition would be critical to valuation. Computer and telecommunications systems are vital to lowering the cost of banks' and thrifts' transactions services. and settling in the range of $25 to $30 billion through the end of the decade?7 Im- portantly. in part because they were positioned to benefit by economies of scope from investing in computers and telecommunications. there should be less need for banks to provide information production and monitoring services for many bor- . the cost of insurance companies' accounting and analysis of customers' risks. the cost of mutual funds' servicing of shareholder's accounts. High yield bonds or "junk" bonds are assets that would appear to fit into this category.24 THE CHANGING MARKET IN FINANCIAL SERVICES affecting the underlying contracting technology have occurred. One would expect that institutional investors who acquired the necessary computer and telecommunications technologies would have benefitted more than most individual investors. and pen- sion funds were likely to be most affected by this technological progress. mutual funds. mutual funds. as of year end 1988.38 Given that financial institutions have invested relatively greater amounts in information technology and have therefore experienced relatively greater reductions in information costs. the tremendous advances in computer technology and telecommunications during the past two decades have undoubtedly led to a drastic reduction in the cost of gathering. While it is difficult to measure precisely. three-quarters of the stock of junk bonds was held by insurance companies. if other financial institutions can directly acquire infor- mation about borrowers at low cost. analyzing. and the cost of pension funds' servicing their participants' accounts. what effect would this have on bank lending? First. and transmitting information. This is not to say that this lower cost of collecting and disseminating information has reduced information asymmetries between all borrowers and investors in the economy by equal amounts.5 billion. The junk bond market grew dra- matically during the 1980s. Intermediaries. such as commercial banks. annual new issues were less than $1. Prior to 1981. money managers. Individual investors owned only 5 percent of the stock of junk bonds. insurance companies. thrifts. but peaked to over $30 billion new issues in 1986. Informal Evidence of Technological Change If institutional investors have benefitted the most from technologies that have reduced information costs.

if we make the logical assumption that banking institutions. To take this idea a step further. One effect would be that third party guar- antees on asset-backed securities would become feasible if the third party bank or insurance company could verify the accuracy of the originating bank's credit analysis. The ratio of nonbank commercial paper to banks' commercial and industrial loans rose from less than 10 percent in 1959 to over 75 percent in 1989. Therefore. This would explain the ability of banks to sell single commercial and industrial loans to other institutions. This would be expected to produce two effects on the market for loan sales. indicating a migration of large and medium sized corporations from bank financing to publicly issued securities. 39 Second. This would explain the growth of the asset-backed securities market. if regulations such as reserve and capital re- quirements increase the costs of funds of banks who already must pay competitive rates for deposit and equity financing. are able to verify the credit analysis and monitoring of the loan selling bank at lower cost than nonbank financial institutions. they may be able to inexpensively verify the information collected by another bank regarding these borrowers. this would explain why over three-fourths of loan buyers are other banks. these banks would become uncompetitive as a source of financing for many borrowers. being in the same line of business as a loan selling bank. note that the incentive compatibility constraint (3i) makes the assumption . There is also evidence that banks' comparative advantage has been reduced even more at the other end of the debt risk spectrum.THE OPENING OF NEW MARKETS FOR BANK ASSETS 25 rowers. even if many financial institutions could not directly acquire information about certain classes of borrowers at low cost. Empirical Results 3: Observability in the Loan Sales Market A final test is performed in Gorton and Pennacchi (1990A) that provides some evidence of the ability of loan buyers to verify the performance of loan selling banks. Recalling the model described in the previous section. Banks' comparative advantage in eliminating duplication of infor- mation services or free-riding problems by multiple investors is likely to be reduced when these multiple investors have low costs of information acquisition. The previously mentioned growth in the junk bond market might reflect this phenomenon. The other effect would be that individual financial in- stitutions could verify the accurate production of credit information and monitoring by a bank wishing to sell a single loan that it has originated.

The first. the following relationship can be derived: P= c 1. r. . 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. the effect of the first term was not as strong as the theoretical model would suggest. the level of moni- toring the bank would choose if it had not sold the loan. If monitoring were observable. If monitoring is observable. It was assumed that the fraction of the loan guaranteed (if any) was the same for all loan sales. 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.-rf)~) [1-s(1. This suggests that a direct empirical test of the incentive compatibility constraint might be able to shed light on the question of whether or not bank monitoring of borrowers is observable by loan buyers. that is.rp (7) L(1_e-(rl. which is independent of the fraction of the loan sold. while there should be an insignificant relationship between Pand the second term. s. where a borrower's commercial paper rating was assumed to be a discrete measure of its benefit from bank monitoring. On the right-hand side are two multiplicative ratios. and the fraction of the loan guaranteed. However. the bank and loan buyer could contract to set the level of monitoring at its most efficient level. 40 We interpret these results as sug- gesting that loan buyers have the ability to verify. so that r was treated as a parameter. namely the level which would satisfy (3i) where s = 0.to Pif monitoring is unobservable. The second term on the right hand side denotes the effect of unobservability. there should be a positive and significant relationship between P and the first term on the right-hand side of (7). as well as the assumed functional form for the expected return on the loan.26 THE CHANGING MARKET IN FINANCIAL SERVICES that the bank's level of monitoring is not directly observable by the loan buyer. Both terms should show a significant relationship.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. 'Is> and P that would hold if monitoring were observable. Using constraint (3i). at least partially. equa- tion (5). is the relationship between the loan sale yield. the performance of loan selling banks. Equation (7) was estimated in log form as a probit model.

aimed at providing the public with a circulating medium which did not expose people to losses either due to better informed traders or because of banking panics. The regulatory issues which must be confronted in the banking indus- try are larger than can be discussed here in detail. The technology underlying banking is changing more generally. the combina- tion of illiquid assets with demandable debt liabilities appears to be an arrangement which is no longer the most efficient way of providing either a circulating medium or monitoring services. as we have argued elsewhere. like those that appear to be developing.41 The asset sales contracts mean. at least originally. the market incom- pleteness necessitating government intervention would seem to be gone. Suffice it to say that technological forces cannot and should not be regulated away. While the loan sales market is sizeable. The existence of well functioning markets for bank assets. The same technological forces that have allowed the markets for asset sales to open. If bank assets can be sold in fairly liquid markets. that is. The loan sales market has the important feature that bank assets are sold without the creation of a contingent liability. All the explanations for loan sales considered above imply that banks still offer services for certain classes of borrowers that cannot be obtained in capital markets via issuance of open market securities. If markets for bank assets open. . risking their equity during the life of the asset created. have also allowed for a much larger set of nonbank possibilities for the private provision of liquidity. While this development is not yet well understood.THE OPENING OF NEW MARKETS FOR BANK ASSETS 27 Concluding Comments: Marketable Bank Assets and Regulation Technological change appears to have resulted in the opening of markets for bank assets. however. there is no explicit contract feature which eliminates the risk that the selling bank will fail to perform on the original contract with the underlying borrower. it is by no means clear that the requisite volume of bank assets are marketable. Previously nonmarketable assets can now be sold. (These are discussed in Gorton and Pennacchi (1990C). then the rationale for bank regulation is called into question since it is fundamentally based on the illiquidity of bank assets. Technological change is forcing a separation between these two sets of activities. does not mean that intermediation per se is ending. it would seem to be a rather important shift in the way banking is conducted. that it is no longer necessary for banks to hold loans until maturity. Deposit insurance was.) In fact.

firms would only do this if bank loans involved the production of some services not obtainable by issuing securities in the capital market. 4. bear the cost of bank reserve requirements. suggesting that banks learn about borrowers after an initial loan is made. For example.) A similar study by Lummer and McConnell discriminates between new and revised bank loan announcements. preferred stock. Apparently. 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. 7. Since nearly all loans are sold without recourse or guarantee. at least those secur- ities that are backed by mortgages carrying government (e. The stock market reaction (in the sense of abnormal return) upon the announcement of offerings of common stock. James shows that there is an (abnormal) positive return on stocks when nonfinancial firms announce that they have obtained a bank loan. which are subject to reserve requirements and other money market instruments. this supports the hypo- thesis that bank loan sales are motivated. not bank depositors. Their results show that only revised loan announcements have a positive effect on stock returns. Lummer and McConnell (1989). 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. Some hypotheses explaining the existence of markets for bank assets we do not ex- plore. Evidence that bank loans are unique is provided by James (1987). This government credit enhancement appears to be the primary determinant of their marketability. Using event study methodology. 3. 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.28 THE CHANGING MARKET IN FINANCIAL SERVICES Notes 1. and Fama (1985). and Gorton and Haubrich (1987). The formal arguments. convertible preferred stock. and straight bonds is negative. convertible bonds.. 6. at least in part. 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. by a desire to avoid the regulatory costs associated with required reserves and required capital. This hypothesis argues that asset sales represent an attempt by commercial banks to compete with investment banks. Presumably. (See Smith (1986). Boyd and Prescott (1986). We believe this explanation . 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. Yield spreads between bank CDs. are not significantly different. See Calomiris and Gorton (1990) for a discussion of the causes of banking panics. Diamond (1984). and details are provided by Campbell and Kracaw (1980). Presumably this is part of the logic behind capital requirements for commercial banks. the logic behind this regulation is that if a bank sells a loan with recourse and thus retains the loan's risk. Throughout this essay we will ignore mortgage-backed securities. VA and FHA) guarantees against default. 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. Note that this hypothesis is consistent with the observation that most loan buyers are other banks. though not always statistically significant. Also.g. both Fama and James point out that bank borrowers. 8. 5. 2.

Of the $80. Western Financial Savings Bank came to market four times in 1988 for a total issuance of $460 million. In a loan sale. 14.5 percent were nonperforming loans. 15.. 16. the principal balance of the pool falls to $95 million. at the end of a month. Greenbaum and Thakor (1987). Pennacchi (1988). "Credit Card Receivables: Moody's Examines the Risks.. Kareken (1987). trade receivables. 11. suggesting that they do not have ready access to public securities markets. have important rights which holders of sec- ondary loan participations do not have. James (1988). For further details see "Credit Card Backed Securities: An Introduction. "Rating of Credit Card Receivables" Duff & Phelps Inc. (no date). the holder of the secondary participation has no legal connection with the failing firm and hence cannot be represented in bankruptcy court. 12. Flannery (1989). Rochester Community Savings Bank offered $125 million of these securities which featured a floating rate coupon. while the seller class would drop to $20 million. Benveniste and Berger (1987). 17. Empire of America Federal Savings Bank made two offerings of automobile securities totaling $631 million.) Secondly.5 percent of the outstanding loans sold as of June 30. motorcycle loans. and a $181 million issue by Signet Bank (Virginia). 1990 by banks responding to the Federal Reserve System's Senior Loan Officer Opinion Survey of Lending Practices. lacking any special services by the bank. In practice these rights vary a great deal depending on the complexity of the contract and the sophistication of the parties involved. 3. About 37. 9. and equipment leases. Then if. (See Gorton and Haubrich (1989). most loans that are sold are the obligations of borrowers who do not have commercial paper ratings." Goldman. and Boyd and Smith (1989) suggest a variety of possible motivations for loan sales. 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. this would be a successful way to compete. if the underlying borrower fails. Some loan sales take the form of assignments. It is not clear that this entire amount represents defaults since some loans may have been nonperforming when they were sold. There are a large number of other types of loans that have formed the basis for asset-backed securities. a $250 million issue by Huntington National Bank. two issues totaling $470 million by Chemical Bank. many assignments specify a minimum amount which must be held by the originating bank. rather than participations. See also Gorton and Haubrich (1989). "Securitization of Credit Card Receivables Using a Senior/Subordinated Structure. 1990 were merger and . Thrifts also issued automobile securities during 1988. Examples include light truck loans. Sachs & Co. In other words.. thus. Other offerings of automobile securities by commercial banks in 1988 included: two issues totaling $480 million by Marine Midland Bank. See Gorton and Haubrich (1989) for further discussion. First. 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. it seems unlikely that. January 1987." Moody's Investors Service. One year earlier 1 percent of loans were nonperforming out of a total outstanding amount of $72. Assign- ments are quite strong contracts because the assignees have all the rights and responsibilities of the lead or originating bank. Holders of marketable securities. Cumming (1987).2 billion dollars of outstanding loans sold as of June 30. 13. boat loans. Details and terminology vary with the issuer.THE OPENING OF NEW MARKETS FOR BANK ASSETS 29 cannot be totally satisfactory." Goldman. Importantly. 10. December 1988. Sachs & Co.2 billion for survey respondents. January 1987. the investor class remains at $75 million.

variable. 25.04 days. 11. 26. For details see Gorton and Pennacchi (1989).63 days. (See the Federal Reserve Senior Loan Officer Opinion Survey of Bank Lending Practices. This can explain the negative coefficient obtained on that variable. 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. I if AI. and rate = {4 if no rating.1 percent of loan sales. 2 if A2.53. See Wall Street Journal. 22. The guarantees are options. August 23. The average loan interest rate was 7. p. 24. the b. and 0 if AI+J. but dictated by data limitations as will become apparent. Investment banks have recently started to make a secondary market in some participations. 19. In order to avoid blatant violations of Glass-Steagal secondary loan participation contracts explicitly forbid secondary market sales. with 7. whereas the loan sale holder may be able to exercise at any time. In the data set. See Gorton and Pennacchi (1989) for further discussion. The premium for bank risk. 23. 1990.8 percent of loans were purchased by domestic banks with assets over $2 billion. The following OLS regression supports these observations.) 18. This means that the effective premium that loan sales buyers place on bank risk. Interestingly. This data set was obtained somewhat subsequent to the study by Gorton and Pennacchi (1989). will be quite small relative to commercial paper buyers.9 percent of this being purchases by nonfinancial corporations and 16. b" was calculated using bank stock prices as described in Gorton and Pennacchi (1989). The existence of such sales plays a role in the Icgal deter- mination of what constitutes a security. 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. the bank writing the option may default. the bank involved was motivated to provide the data mostly because it was convinced that there was no implicit guarantee. 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. exchange traded. Investment banks that underwrite commercial paper will usually agree to repurchase this paper prior to maturity for the expressed purpose of providing liquidity.2 percent being purchases by other nonbank buyers. The linear specification is ad hoc. 20. 28. By the G1ass-Steagal Act banks are prohibited from issuing securities. 1. The average maturity of the loan sales contracts was 27. Nonbank buyers accounted for 24. Vulnerable options have different comparative statics than stand- ard. Loan sales yields were classified by maturity and commercial paper rating of the underlying borrower.8 percent of loans were purchased by domestic banks with assets under $2 billion while 25. but they are vulnerable options. See Gorton and Haubrich (1989). The holder of commercial paper can only call on the guarantor at the maturity of the paper. the overall average of the fraction of loans sold that were retained by the bank was approximately 24 percent. Letting fr = the fraction of the loan sale retained by the bank. the average maturity of the loans sold was 28. According to the Federal Reserve's August 1990 Senior Loan Officer Opinion Survey on Bank Lending Practices. while the average loan sale interest rate was 7.30 THE CHANGING MARKET IN FINANCIAL SERVICES acquisition related. we have: . 21. Hence. Further details on the sample can be found in Gorton and Pennacchi (1990A). that is. 27. 1991. See Johnson and Stulz (1987) on vulnerable options.41. maturity = the maturity (in days) of the loan. options. January 4. 3 if A3.

00762) No.THE OPENING OF NEW MARKETS FOR BANK ASSETS 31 fr = .a).012. 37. and securities laws. of Obs. Our purpose here is only to lay bare the issues by summarizing the model. See Pennacchi (1988) for an analysis of the initial portfolio choice.2355 (rb . = 872. Note that if no monitoring is done. The model does not explicitly account for the noise term. L. 34. the expected return is L(1 .r.0120) (. s. For example.) Thus. we estimate the model as a non- linear Tobit model since the model is censored (the right-hand side of (6) is observed while the left-hand side is not) when s = 1. 36. 30. 31. Note that since the fraction of the loan sold. The parameters a and /3 are assumed to be positive and loan specific. These figures are from United States Securities and Exchange Commission (1990). as expected.02965) (. Note that the assumed function implies that as the level of monitoring rises towards infinity.) Thus. Presumably. 29. The assumption is consistent with the observed contracts in this market. The parameter a is also assumed to be less than unity. The model is estimated in log form by assuming that the log of the fraction of a loan sold is equal to the log of the right-hand side of equation (6) plus a normally distributed error term.00151 maturity + . but only the maturity variable is significant. the expected return on the loan asymptotically approaches the promised payment. 38. of Obs. This means that the loan sale contract is restricted to a proportional equity split between the bank and the loan buyer.0512) No.311. statistics from the Federal Reserve Bulletin show that individual investors held more that 10 percent of all outstanding Treasury securities as of year-end 1988.17890 + . prohibitions on explicit guarantees. = 872. takes on a value that is at most equal to 1 (and hence the In(s) is at most zero). But. the loan sale spread is.00046) (. The likelihood function is given in Gorton and Pennacchi (1990A). R' = . 32. The parameter /3 is a measure of the marginal increase in expected return on the loan from additional monitoring. The reader is referred to Gorton and Pennacchi (1990A) for the details of the model. The reader is referred to Gorton and Pennacchi (1990A) for details.0068 b (. = . assumed to be uncorrelated with the right-hand side of (6). senior/subordinated positions are not allowed. significantly positively related to the spread paid by the borrowing firm. 35. with the speed determined by the parameter /3. though it is of the expected positive sign.00004) (. The cost of implicit guarantees is assumed to take the form of pressure by regulators. (Standard errors in parentheses. This problem is independent of the bank's choice of the number of loans to orig- inate. regulatory constraints prevent other contract forms.00742 rate (.00061 + . R' = . For purposes of comparison. The idea is that this noise term can capture the influence of missing factors. the coefficient on the probability that the selling bank will fail is insignificantly different from zero. . See Becketti (1990) for more description of recent developments in the junk bond market.) + . This is done by solving for y using equations (4) and (6). When the fraction sold is omitted and equation (1) is estimated with this data set the results are: r" . These constraints include the requirements for taking the loan off the balance sheet. point estimates suggest the fraction retained increases with maturity and the risk of the rating. (Standard errors in parentheses. 33.r.

125-45. "What's Different About Banks?. "Information Production. Credit Markets. Eugene. however. 1987. the right-hand side of (7) was transformed into two additive terms. See Gorton and Pennacchi (1990B.06) but also significantly different from its theoretical value of unity. "Financial Intermediary-Coalitions. 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). Flannery. and the Control of Capital. 1987. by Glenn Hubbard (University of Chicago Press). Gary and Joseph Haubrich. (July/August): 45-54. Diamond." Journal of Economic Theory 38. "Are Loan Sales Really Off-Balance Sheet?." Journal of Finance 35(4) (September). "The Truth about Junk Bonds. Christine. "Capital Regulation and Insured Banks' Choice Individual Loan Default Risks. Campbell. 1980." Federal Reserve Bank of Kansas City Economic Review 75(4). 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.. "Securitization With Recourse: An Investment that Offers Uninsured Bank Depositors Sequential Claims." Journal of Monetary Economics 15 (January): 29-39. 41. Fama. Cumming. and Bank Regulation. 1990. 1990. Federal Reserve Bank of New York. Tim and William Kracaw. 1984. Gary and George Pennacchio 1989." in Financial Markets and Financial Crises. Mark." Journal of Banking and Finance 11.. 2. Lawrence and Allen Berger. Since estimation was in log form. . Boyd. The first term had a coefficient estimate that was of the correct sign (positive) and significantly different from zero (. Facts. by George Kaufman. 235-58. 1986. Sean." Journal of Monetary Economics 24. John and Edward Prescott." Review of Economic Studies LI (July): 393-414. Gorton. "The Origins of Banking Panics: Models. (September): 403-24.C) for the theoretical rationale and empirical evidence of a shift from bank financing to direct financing." Carnegie-Rochester Conference Series on Public Policy 26 (Spring): 289-333. There is. Calomiris. Auditing. "The Economics of Securitization. Market Signalling. Douglas. 11-23. 1987. 211-32. 1989. ed. 40. References Becketti. "Financial Intermediation and Delegated Monitoring. Charles and Gary Gorton. and the Theory of Financial Intermediation. "The Loan Sales Market.32 THE CHANGING MARKET IN FINANCIAL SERVICES 39. . 1985. Gorton. ed." Journal of Accounting." Quarterly Review (Autumn). abundant evidence that the demand for bank provision of these services has fallen.37 with a standard error of approximately . JAI Press Inc: Greenwich CT. Vol. "Bank Deregulation." Research in Financial Services. 863-82. Benveniste. and Finance 4(2). 1990.

Herb and Rene Stulz. 1989.. Forthcoming. (December): 217-35." unpublished working paper.." Journal of Financial Economics 25. U." Journal of Finance 43." Journal of Banking and Finance 11. 1987." Journal of Monetary Economics 22. Greenbaum. Scott and John McConnell. Johnson." Journal of Finance 42(2). 1986. 1987." Journal of Banking and Finance 11. "The Emergence and Regulation of Contingent Commitment Banking." Reform of Deposit Insurance and the Regulation of Depository Institutions in the 1990s. "Further Evidence on the Bank Lending Process and the Capital Market Response to Bank Loan Agreements. Kareken.S. 1990. "Bank Funding Modes: Securitization versus Deposits. Securities and Exchange Commission. 1990.3-29. Christopher. "Some Evidence on the Uniqueness of Bank Loans. John. ed. Lummer. (November): 395- 422. 1990A.359-77. ." Journal of Finance 45(1). March 15. 1988. 1988. Pennacchi. "The Use of Loan Sales and Standby Letters of Credit by Commercial Banks. "Banks and Loan Sales: Marketing Nonmarketable Assets. 1990B. George. 99-122. Clifford.THE OPENING OF NEW MARKETS FOR BANK ASSETS 33 . "The Pricing of Options With Default Risk. 1987. . "Recent Developments in the High Yield Market.." Journal of Financial Economics 15. Stuart and Anjan Thakor. "Financial Intermediation and Liquidity Creation." Staff Report. (June): 267-80. 375-95." Journal of Financial Economics 19. .49-72...1990C. "Investment Banking and the Capital Acquisition Process. --.. "Financial Innovation and the Provision of Liquidity Services. (September): 379-401. by James Barth and Dan Brumbaugh. James. "Loan Sales and the Cost of Bank Capital. Smith. 1987.

commercial paper. capital markets. This has facilitated specialization among providers of financial services with unmistakably positive implications for the efficiency of the intermediation process. there is nothing terribly new about banks selling loans. if there is anything to be explained about banking. this same recent period was characterized by explosive growth in Euromarkets. COMMENTARY Stuart I. the answer will be found in an information 35 . arguably one of the more signifi- cant recent developments in the practice of financial intermediation. it seems to me that the authors fall into a paradigmatic snare. What explains the striking growth of this phenomenon at this par- ticular time and place? In addressing this question. and futures as well. and securitization to a lesser degree. Nevertheless. A compelling way to begin a research agenda is to ask the question why. the documented growth of loan sales. financial options. or financial inter- mediation more generally. Moreover. Of course. junk bonds. particularly in the United States where these innovations were pioneered. is stunning. Greenbaum Gorton and Pennacchi provide a service by directing the attention of academe to loan sales and securitization. it has fostered the decomposition of the credit transaction into its more elemental components. As generally known.

If it is not a moral hazard. however. Three possibilities are considered. However. Second.36 THE CHANGING MARKET IN FINANCIAL SERVICES asymmetry. The obverse would be expected in the low sav- ings country. However. First. For example. One might. After all. then surely it is a self-selection problem. there was nothing new in the idea of re- taining a fraction of the loan or of granting an implicit guarantee. this conclusion should not be considered a criticism of the authors. asset prices in the high savings country would rise and real interest rates would fall. the originator may provide implicit guarantees against losses on loans sold. suppose we have two more or less autarkic countries that could. to their credit. And further suppose that these two countries have approximately the same savings rates. one impediment to . thereby weaken- ing the earlier basis for principal/agent type moral hazard. the Gorton! Pennacchi conclusion could have been inferred without the benefit of the data manipulation. the savings rate in one country triples and in the other it falls in half. and the authors tentatively conclude that the technological advances hypothesis seems worthy of respect because the data fail to support the alternative two hypotheses. by retaining a fraction of the loan. The three hypotheses are lucidly articulated and then confronted with sketchy data relating to loan sales and securitization. What sort of responses would we expect? Presumably. But nothing seems to matter very much. expect investors from the high savings country to seek out assets in the low savings country because of their relatively attractive prices. Hence. under the circumstances. (note 5 notwithstanding). While the latter exercise can only be described as disappointing. if one is precommitted to an information argument. Now for demographic reasons. but for whatever reason choose not to. The more important point. Finally. The only possibility is that advances in technology overwhelmed the inhibiting influences of informational asymmetries. The more recent growth of loan sales must then be attributable to some mitigation of the moral hazard. seem to share that misgiving. the authors consider the possibility that technological advances may have reduced informational asymmetries. is that the growth of loan sales may have had nothing to do with informational asymmetries. From the authors' perspective the earlier absence of loan sales was attributable to an insurmountable moral hazard-the loan originator's failure to exercise the same due diligence and monitoring that it would provide as funder of the assets. the originator may align its own incentives with those of the loan purchaser. and this is a possibility the authors strain to ignore. trade very much. these provide no basis for temporally anchoring the observed loan sales growth. I doubt that this is a very robust conclusion and the authors.

Instead of being able to borrow at some sub- risk-free interest rate. The investors will. and they may seek assurances. The deposit subsidy will prompt banks to accept all proffered deposits. but in what sense do the asymmetries explain the growth of loan sales? I have one more question regarding the Gorton/Pennacchi paper. To be sure. suppressing deposit interest rates. After all. the size of the industry was previously bloated by the subsidies directed to banks. the deposit subsidies in one of the countries disappears. despite the inhibiting effects of private information. the larger the spread between the home. Now.THE OPENING OF NEW MARKETS FOR BANK ASSETS 37 the purchase of debt by investors from the high savings country is their lack of asset-originating capability in the low savings country. It is correctly noted that public regulation of banks is related to the traditional . 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. perhaps to banks in the country still enjoying deposit subsidies? What. only the fact that deposits are subsidized matters. 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. With sufficiently generous subsidies. It does not really matter why the two countries might have chosen to adopt such a policy. for whatever reason. informational asymmetries will impede the asset sales. However. would such sales have to do with informational asymmetries? To be sure. therefore. To make matters really dreadful. quite irrespective of investment op- portunities. the banks could profit- ably hold risk-free assets. suppose that these suffering banks are downgraded by rating agencies in light of a perceived excess supply of banks and bankers. be attentive to the brokers' repu- tations. In such an environment we would expect the banks to develop asset origination skills in order to deploy the proceeds of deposit liabilities. 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. the high savings country investors will not be oblivious to the possibility of being exploited by myopic or cynical brokers. the more compelling the foreign investment. and under- pricing governmental deposit insurance. exclusively. and we may even observe banks competing in oblique ways to attract deposits.and foreign-country interest rates. suppose that. if anything. both written and implied. the impacted banks would now face market interest rates.

as already indicated. The link is more basic than that. this is a defining characteristic of tra- ditional banks: they are repositories of illiquid financial claims. the inevit- able accompaniment of the lender of last resort. and this shifts deposit seigniorage from the public sector to the privately-owned banks. This is presumably how we learn! My major reservation. But. I applaud the Gorton/Pennacchi effort. relates to over-reliance on the information paradigm. Illiquidity of bank assets gives rise to the need for a lender of last resort. In explaining the nexus between illiquidity and public regulation. This moral hazard is solved with the introduction of cash-asset reserve re- quirements. Indeed. the introduction of the lender of last resort weakens the bank's incentive to hold cash assets. the authors stress the role of deposit insurance. To be sure. infor- mation arguments have been illuminating. how- ever. In closing. It directs attention to an important phenomenon and it should excite scholarly interest. and therefore seductive.38 THE CHANGING MARKET IN FINANCIAL SERVICES illiquidity of bank assets. and we consequently have endogenous regulation. espe- cially in banking. . for then the paradigm becomes a toy instead of a tool. Their raison d'etre was to swap liquid for illiquid claims with a clientele displaying a preference for the former. But informational asymmetries should not be expected to explain everything.

II .

which singly pose no anticompetitive threat. but strained. -Ralph Waldo Emerson Introduction After a decade of deregulation in banking. had these companies and the banking industry been more profitable over the past decade to fuel expansion. but collectively may lead to significant entry barriers and threats to competition in the future. Hanweck The field cannot well be seen from within the field. There is in place a haphazardly developed set of state laws regulating interstate banking. there is no public policy addressing Loncentration of bank- ing resources through acquisitions. the number of banks may have declined substantially to the point of concern for regional and local market competition for banking services. and an expansion by banking companies and others via the acquisitions of failed thrifts and banks. With the expansion of interstate banking opportunities in recent years. Federal law that is piecemeal and that is being eroded by the liberalization of banking powers by the courts and Federal Reserve. as this article points out. the field remains unseen. however. that further undermil'es law and regulation. BANK EXPANSION AND VALUATION Gerald A. pace toward geographic expansion by larger banking companies. To date. 2 INTERSTATE BANKING. there appears to be a rapid. 41 . Additionally.

778 interstate nonbank offices of banking companies and 7. even though there has been considerable geographic nonbank expansion by BHCs. The major handicap to the lack of a full-service banking office is the solicitation and provision of many deposit services. As a result. The proximity of a bank for cus- tomers to make deposits and receive cash is important for retailers and households. Thus. the presence of a nonbank office. 2 As will be developed in a discussion of the history and state of inter- state banking. it is likely that such effects would vary with the location. almost any large . Since geographic banking oppor- tunities have been altered. this is not an effective substitute for interstate banking.492 domestic and foreign interstate banking offices. then. thirty- two states will have some form of national interstate banking with twelve of these having no reciprocal restrictions. the recent changes in state laws regarding interstate banking should be expected to affect the valuation of commercial banks. The remaining states with no interstate banking statutes are expected to enact such legislation soon. 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. such as a mortgage banking company. 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. To the extent that such dramatic changes in permissible bank expansion affect valuation.' In many respects. Even though the expansion by the major banking companies has been constrained. these changes in state law are not uniformly reflected in bank valuation changes. Also likely. is that the trend will accel- erate by early 1991 when approximately thirty-two states adopt some form of national interstate banking. By early 1991. 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. nature of banks. with twelve of these having no recip- rocal restrictions. and special service offices and automated teller machines (ATMs) are not convenient substitutes. and the type of state law change.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. It is likely. By the end of 1988 there were 6. As this article states. 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.

is deteriorating.INTERSTATE BANKING. As suggested. the implications of balkanized legislation. This Act also should be thoroughly amended to encourage interstate banking through branching and acquisition and to promote a long-run competitive environment through transition rules limiting local market and state shares and. The second section of this article summarizes and reviews the legal background and the state of interstate banking. the profitability of many of the "superregional" banks. The final section expresses some conclusions and suggestions for further research and public policy toward interstate banking. The fifth section will consider the view that banking profitability will be the constraining factor in the future development of interstate banking. it is imperative that federal legislation be enacted to eliminate McFadden Act provisions to interstate branching and the Douglas Amendment to the BHC Act. only a handful of banking companies. With this as a prospect. like their larger money-center counterparts. excessive expansion. which will only serve to encourage bankers to seek profitable investment opportunities elsewhere such as in the new markets emerging in Europe after 1992. BANK EXPANSION AND VALUATION 43 banking company with sufficient resources can acquire a bank in these states. This section reports on several recent studies that have attempted to evaluate interstate banking law effects on the profitability and valuation of banking companies. in terms of profitability and valuation. and capitalization to expand in the near future? Consequently. The result is that these banks. having recently grown rapidly through interstate expansion. through existing regulation. to changes in interstate banking legislation when they are within the affected state or region and outside a state or region enacting interstate banking legislation. will not have the profitability. This section will assess the proposition that banking resources will be so limited and the overhang of the real estate market recession will be so substantial that bank expansion geographically or into alternate product lines will be significantly hampered for the next several years. resources. and a Congress preoccupied with the growing costs of the savings and loan (S&L) bailout crisis. will be able to expand geographically to any great extent and much of this expansion is likely to be by way of acquisitions of failing banks and thrifts. including some foreign banks. The third section considers the public policy and competitive issues surrounding interstate banking and reviews some of the relevant literature and studies that have addressed these areas. Though the development of interstate banking is poised to accelerate. . continue to hamper the development of interstate banking and expanded product lines and will contribute to the deteriorating profitability of banking companies. In the fourth section is developed hypothesized reactions of banks.

and savings and loans. Though prohib- ited from expanding full-service offices across state lines. changing economic and technological conditions have stimulated commercial banks to reconsider their perceived optimal structural organizations. the Douglas Amendment. However. The Glass- Steagall Act of 1933 provided for limited regulation of bank holding companies by the Federal Reserve System. 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. Section 3( d). and in 1984 to open entry by banks . bank holding companies were unregulated and some banks used this vehicle to establish a multistate presence. this legislation permitted interstate foreign banks to retain existing networks. 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. however. Maine was the first state to pass a law permitting interstate acquisitions in 1975. branching across state lines. commercial banks have been able to circumvent the legal restrictions on interstate branching. Through the bank holding company organizational form. This legislation eliminated. but not charter. later liberalized in 1978 to national reciprocal. foreign bank branches and agencies. In addition to banking companies. prohibited branching by national banks. and subsequent legislation and interpretations.S which have been permitted on an interstate basis. and operate banks within each state. of the Bank Holding Company Act of 1956 permitted states the power to legislate out-of-state bank holding companies to acquire.44 THE CHANGING MARKET IN FINANCIAL SERVICES The History and State of Interstate Banking Prior to the mid-1960s. The Bank Holding Act of 1956. but did not limit interstate expansion. loan production offices. in effect. The introduction of national banks by the Na- tional Banking Act of 1863. such as Edge Act Corporations. 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. banks had little incentive to expand beyond state boundaries. Prior to 1956. non bank 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. prohibited further establish- ment of interstate banking offices but grandfathered the seven existing domestic interstate banking organizations at that time. there are several types of limited service facilities.

Otto Bremer Foundation. such as securities firms (for example. No other states followed Maine's lead until 1982 when New York and Alaska enacted liberalized interstate banking legislation. BANK EXPANSION AND VALUATION 45 headquartered in any state. commercial banks have found a number of ways to maintain operations in more than one state. and transportation have improved within the past thirty years. many of the nation's largest banking companies had established one-bank holding companies that were exempt from cov- erage of the 1956 BHC Act and had entered into nonbanking activities which were of questionable legality if done through the bank and. as discussed above. In rapid succession many states followed so that currently forty-seven states and the District of Columbia have some type of law permitting interstate banking through bank holding companies (see table 2-1). In response to the changing financial marketplace environment and in order to compete more effectively with their nonbank financial rivals. Prior to passage of this law. First Bank System. assets controlled by banks from other states are 15.2 percent of total banking assets nationwide with control as high as over 84 percent in Maine and 79 percent in Washington. of which twelve will have no reciprocal restrictions (see tables 2-1 and 2-2). Other companies that appear to compete directly with commercial banks in some services. it has become more advantageous for banks to conduct their business across state lines. (2) limited service (nonbank bank) banking offices. Commercial banks have not. Two of these ways have been the grandfathered BHCs and the nonbanking BHC subsidiaries. The seven interstate domestic holding companies grandfathered by the Douglas Amendment are First Interstate Bancorp (formerly Western Bancorporation and once part of Bank of America). and (3) nonbanking offices. and retail stores (Sears). more importantly. By early 1991. been able to establish offices that accept deposits and make commercial loans across state boundaries. finance companies (GMAC). communication. thirty-two states will have effective national interstate banking laws.INTERSTATE BANKING. are not limited by state lines and can follow their markets to any geographic location. often crossed state lines. As technology. The current status of all the major exceptions to the interstate banking restriction can be classified into three categories: (1) full-service bank-charter offices. The 1970 Amendments to the Bank Holding Company Act redefined a BHC to include one-bank holding companies and also established procedures to determine permissible nonbanking activities for BHCs. As of August 1990. we consider more prominently the first group of exceptions. Credit . Merrill Lynch's cash management account). Northwest Bancorporation (now Norwest). Since the focus of this paper is on interstate banking in a full-service office context.

Colorado Currently Reciprocal. reciprocal. IL. TN. reciprocal. State Legislation in Effect Area Alabama Currently Reciprocal. Florida Currently Reciprocal. MO. no reciprocity. SC. TX. 1991 National. FL. MO. 13 States and DC (AR. Connecticut Currently National. WI). December 1. WV. WV). GA KY. TN. 12 States (lA. NE. VA. 6 States (lA. MI. MS. LA. WY). FL. 11 States (AL. KY. NC. Idaho Currently National. KS. 11 States (AK. MO. UT. NE. OH. Kentucky Currently National. 11 States and DC (AL. SC. PA. TX. CO. District of Columbia Currently Reciprocal. WI). Illinois Currently Reciprocal. 7 States (AZ. January 1. LA. MD. AR. July 1. NC. NM. SD. KY. VA. Louisiana Currently National. WV. 16 States and DC (AL. Georgia Currently Reciprocal. WA). California Currently Reciprocal. NC. TN. no reciprocity. reciprocal. . no reciprocity. VA. IN. Arkansas Currently Reciprocal. GA. NC. NV. LA. GA. DC). MS. MD. Arizona Currently National. TN. NC. DC). reciprocal. FL. 10 States and DC (AL. Alaska Currently National. reciprocal. MS. AZ. SC. NM. MI. January 1. OR. no reciprocity. Maine Currently National. 1990 National. 6 States (IL. WV). 1990). no reciprocity. 1991 Reciprocal. 1992 National. MS. UT. WI. MO. Delaware Currently National. DC). GA. KS. OK. LA. MD. TX. SC. TN. MD. TN. VA. Interstate Banking Legislation by State (As of September 5. ID. WV. reciprocal. VA. Indiana Currently Reciprocal. HI. FL. MN. DC). MN. MD. Iowa January 1. KY. NE. reciprocal. SC. VA. OK. LA. MS.Table 2-1. 1991 National.

KS. South Carolina Currently Reciprocal. ND. reciprocal.) State Legislation in Effect Area Maryland Currently Reciprocal. MS. reciprocal. Oklahoma Currently National. MS. Oregon Currently National. reciprocal. TN). SC. IL.DC). lA. KY. VA. New Hampshire Currently National. 14 States (CO. TN. PA. FL. reciprocal. LA. KY. WV. reciprocal. GA. 1991 National. TX. no reciprocity. OK. reciprocal. 12 States and DC (AL. New York Currently National. NC. TN. WI. KY. AR. NE. VA. KY. Minnesota Currently Reciprocal. BHC must be from state offering reciprocity or wait 4 years to expand. ND. AR. LA. no reciprocity. VA. lA. DC). FL. reciprocal. DE. no reciprocity. DC). KS. LA. MO. reciprocal. MO. MN. FL. GA. SC. (Cont. Rhode Island Currently National. reciprocal. SD. no reciprocity. MS. TX. MT. Mississippi Currently Reciprocal. NE. MD.Table 2-1. 13 States and DC (AL. Tennessee Currently Reciprocal. Nebraska Currently Reciprocal. GA. Nevada Currently National. 8 States (AR. South Dakota Currently National. WV. WI. WV. WY). New Mexico Currently National. Massachusetts Currently National. TN. MT. 14 States and DC . SC. NC. FL. NC. ID. KS. Michigan Currently National. Missouri Currently Reciprocal. GA. TN. MO. After initial entry. LA. AR. KY. MD. WA. Ohio Currently National. SD. Pennsylvania Currently National. New Jersey Currently National. 13 States (AL. 14 States and DC (AL. North Carolina Currently Reciprocal. January 1. IL. WV). lA. WY). VA. AR. 10 States (CO. IN.

The next group of BHCs in table 2-3 are the foreign banking organ- izations with interstate operations. (Cont. Virginia Currently Reciprocal. KY. AR. TN. no reciprocity. reciprocal. MO. MS. foreign banks were not subject to the interstate restrictions affecting domestic organizations. SC. MN. WV. Source: Financial Structure Section. DC). no reciprocity. NC. OR). FL.48 THE CHANGING MARKET IN FINANCIAL SERVICES Table 2-1. reciprocal. KY. when most states began to change their interstate banking legislation. SC. Utah. In order to equalize competitive advantages. GA. the International Banking Act of 1978 prohibited foreign banks from expanding interstate in ways not permitted for domestic banks. and First Security Corpo- ration (see table 2-3). Washington Currently National. IN. MS. Prior to 1978. the extent of full-service banking geographic opera- tion is shown in table 2-3. GA. Three of these BHCs are headquartered in Minnesota and one each in California. The foreign BHCs listed in table 2-3 are those retaining full- . Utah Currently National. MD. IL. IN. and Commerce American Holdings (formerly Financial General Bancshares). VA. As of 1982. reciprocal. These organizations have full-service banking offices in twenty-four states with a total of over two hundred banks as of 1983 (since then some have consolidated within a state). and the Netherland Antilles (formerly in the District of Columbia). Wisconsin Currently Reciprocal. Vermont Currently National. January 1. no reciprocity. West Virginia Currently National. Wyoming Currently National. LA. Note: Several states prohibit acquisition of banks in operation for less than a specified number of years. 12 States and DC (AL. LA. 1991 National. but allowed them to keep existing interstate networks.) State Legislation in Effect Area (AL. Board of Governors of the Federal Reserve System. MO. This table also indicates all other banking organizations with subsidiary banks in more than one state and those that had been approved. AR. DC). MD. WV. MI. KY. FL. NC. Mis- souri. Texas Currently National. General Bancshares Corporation. 8 States (lA. reciprocal.

South Dakota and Maryland. This number includes expansion by acquisition of failing banks. Some Explanations For the Rapid Spread of Interstate Banking The rapid expansion in interstate activity has explanations in fundamental economic reasons for banking managements' desire to expand their mar- kets. have permitted out-of-state BHCs to establish banks with limited powers. 1990. This rapid expansion is also reflected in the number of banking offices controlled by interstate banking companies. The final group of banks in table 2-3 are those that had expanded interstate as state laws had begun to change. these banking companies had $442. the Federal Reserve reported that fifty-one banking organizations had subsidiary banks in one or more states beside their home state (Savage 1987).364 in 1988 (see table 2-4). BANK EXPANSION AND VALUATION 49 service banks in multiple states. These banks with limited powers serve customers of their parent organizations in many states.INTERSTATE BANKING. The expansion of banks interstate via chartered banks has been rapid since 1982.258 in 1983 to 7.4 billion in interstate bank assets as of March 31. and have succeeded in getting reciprocity legislation passed in several states.1986. Both these figures suggest a very rapid expansion of interstate banking via full-service banks. Four years later. has shown a remarkable increase in interstate . Domestic banking companies with interstate banking operations have increased the number of banking offices outside their home states under their control from 1. In 1983 there were twenty-six domestic banking companies operating i full-service banks interstate. Money-center banks have pushed for further relaxation of these restric- tions. 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. 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. A review of the states with the greatest interstate banking activity in terms of asset or office growth reveals that Florida. Today this number is approximately 180 banking companies after accounting for interstate acquisitions since 1987. a state with a rapid growth in banking assets. Some states have actively competed to attract large banks. In terms of banking assets. many states for example. which represents an increase from $148. Since 1982. but note that there are numerous other banks which control various types of entities in multiple states.4 billion as of June 30.

7 states (AZ. SC. FL. Colorado Currently Reciprocal. TN. NH. 17 MO. 7 Jan. 1991 National. SC. AZ. OH. RI. reciprocal. 16 states and DC (AL. WV). NCNB and Northern Trust Corporation are grandfathered and can make further acquisitions. NC. NJ. AR. Alaska Currently National. VA. 12 NC. 11 states (AK. GA. OR. LA. TX. UT. 6 Special-purpose banks permitted. Florida Currently Reciprocal. MD. NC. WA). Number of State Effective Date Area Partner States Alabama Currently Reciprocal. NM. GA. 11 TX. June 30. FL. 11 SC. MS. KY. 12 states and DC (AR. TN. Connecticut Currently Reciprocal. LA. NM. HI. Reciprocity hinges on commitments to community reinvestment. SO Arkansas Currently Reciprocal. GA. MD. no reciprocity if community development SO commitments are made. 5 Delaware Currently Reciprocal. NC. SC. reciprocal. WY). FL. LA. LA. 1. no reciprocity. CO. MS. Interstate Banking Legislation by State (As of February 1. VA). 1. 1989). District of Columbia Currently Nationwide. no reciprocity. MD. Hawaii None 0 . NE. MS. ID. OK. WV). MD. OK. KS. Georgia Currently Reciprocal. 1990 National. 10 states and DC (AL. WV). 5 states (MA. ME. VT). PA. NV.S states and DC (MD. Under a 1972 law. SO Arizona Currently National. 13 MS. VA. VA. UT. KY.VI 0 Table 2-2. California Currently Reciprocal. reciprocal. VA). Jan. NE. 1991 National. KS. TN. 11 states and DC (AL. TN.

4 states (AL. reciprocal. PA.. AR. Idaho Currently National. AR. MO. VA. WV). ME. 1990 Reciprocal. 11 WA. WV). VA. 20* Minnesota Currently Reciprocal. LA. SD. . OH. V A. GA. PA. RI. SC. IL.. SC. FL. reciprocal. 31* Louisiana Currently National. 50 Maryland Currently Reciprocal. ND. reciprocal. Kansas None 0 Kentucky Currently National. 1992 National. TN. MT. 6 Nationwide. AR. Mississippi Currently Reciprocal. reciprocal. Indiana Currently Reciprocal. 29* Maine Currently National. MI. no reciprocity. KS. NH.. Under a 1981 law. WY). WV) and special-purpose banks. WI). 11 states (CO. TN. KY. KY. FL.. GA. 13 states (AL. VI . TN). 50 Illinois Currently Reciprocal. MO. reciprocal. Dec. MO. 15 MS. 11 states (lA. NC. KY. 6 states (lA. MI. LA. LA. Massachusetts Currently Reciprocal. IA. Iowa 1972 Under a 1972 law. TN. 1. Norwest Corporation is grandfathered and 0 is permitted to acquire banks in Iowa. 4 July 1. IL. 11 WI. IN. 5 states (CT. 5 Michigan Currently National. NC. DE. ID. General Bancshares Corporation is grandfathered and can make further acquisitions in the state. MO. no reciprocity.. VT). TX. 14 states and DC (AL. KY. July 1. organizations may acquire failed institutions if the 6 failed institution is larger than $1 billion in assets. 1990 National.

LA. 13 MD. 23* Oklahoma Currently National. 1989 National. VT). no reciprocity. 19* North Carolina Currently Reciprocal. NV. MA. reciprocal. SC. CA. lA.) Number 0/ State Effective Date Area Partner States Missouri Currently Reciprocal. VA. lA. MN.S states (CT. no reciprocity. Nevada Currently National. reciprocal. 0 Jan. AR. TN). MO. 8 July 1. 1. 10. RI. reciprocal. ME. KS. 5 New Jersey Currently National. . Ohio Currently National. SO. 8 states (AR.VI IV Table 2-2. WY). UT. OK. NO. KS. KY. 50 New Hampshire Currently Reciprocal. W A). no reciprocity. 1990 National. HI. TN. GA. IL. WI. 1991 National. FL. reciprocal. 1. 50 Jan. 50 Oregon Currently 8 states. 12 states and DC (AL. North Dakota Currently A grandfathered interstate banking organization is permitted to 0 sell its North Dakota banks to out-of-state bank holding companies. 8 Montana None o Nebraska Currently Special-purpose banks. KY. New York Currently National. MS. MT. no reciprocity. WV). Jan. 21 * New Mexico Currently Nationwide acquisition of failing banks. (Cont. NE. no reciprocity (AK. 10 states (CO. AZ. 1. 1990 Reciprocal.

reciprocal. NH. VA. AR. March 4. MN. KY. FL. VA. reciprocal and special-purpose banks. AR. 13 MS. MA. RI). SC. 13 MD. KY. 50 * Does not count the two states where nationwide entry by acquisition of failing banks is possible. NC. OH. 1. IN.S states (CT. Pennsylvania Currently Reciprocal. MI. 50 Vermont Currently Reciprocal. no reciprocity. LA.7 states and DC (DE. 1990 National. KY. reciprocal. 8 WV). KY. West Virginia Currently National. GA. no reciprocity. ME.1990 National. VA. no reciprocity. MO. 12 states and DC (AL. Economic Review (May/June): 35-36. 1989. LA. 23* South Carolina Currently Reciprocal. Source: Federal Reserve Bank of Atlanta. Washington Currently National. South Dakota Currently National. Rhode Island Currently National. FL. MS. Virginia Currently Reciprocal. 29* Wisconsin Currently Reciprocal. 21* Tennessee Currently Reciprocal. 12 states and DC (AL. MS. GA. 8 states (lA. IN. SC. NC. Failing institutions may be acquired by 21* organizations from any state. WV). reciprocal. MO. NJ. KY. TN. reciprocal. TN. 13 MD. 8 Wyoming Currently National. 13 states (AL. FL. GA. reciprocal. Texas Currently National. IL. VI W . 5 Feb. NC. LA. OH). MD. WV). 50 Utah Currently National. WV). AR.

Amsterdam Financial General Bankshares. Japan CA. Osaka. SD. ID. NY. Name BHC Location States in Which Banks are Located First Interstate Bancorp. MT. NY DE. MO. MT.NY Carclsys Bank International. (acqd. UT ID. WI Northwest Bancorporation Minneapolis. Minneapolis. MD.NY Canadian Imperial Bank of Commerce Toronto. MN lA. S. NV. Tokyo. Inc. London. WY First Bank System. Paul. England CA. MN MN. Washington. Paul. WA. CA AZ. ND. NY. Los Angeles. PUERTO RICO Citicorp New York. MN MN. PA.j::o. Lts. DE 12/30/81) . NY DE. New York. ND. NA DC. Spain NY. Inc. Ltd. MN Credit and Commerce American Holdings Curacao. SD. OR. PA DE. Table 2-3.Ul . Canada CA. MO IL. England The Sumitomo Bank.NY Barclays Bank Limited London. CA. SD. Madrid. UT Bank of Montreal Montreal. CO. Bank Holding Companies with Subsidiary Banks in More Than One State (District of Columbia and Puerto Rico included as states for this table). Ltd.NY The Bank of Tokyo. Louis. MN. TN First Security Corporation Salt Lake City. NY. UT. NE.P. DE 12/21/81) The Girard Company Bala Cynwyd. (acqd. MT. PUERTO RICO Banco Central. (acqd. NM.ND. TN. DC General Bancshares Corporation St. Japan CA. WI Otto Bremer Company St. DE 9/14/82) J.A. Morgan & Co. Canada CA. WI Otto Bremer Foundation St. VA Credit & Commerce American Invest.HI The Royal Bank of Canada Montreal. Canada NY.

(approved 06/02/83) Source: Board of Governors of the Federal Reserve System. NY. (acqd. MD. DE 05/10/82) Equitable Bancorporation Baltimore. Inc. PA. Bank Holding Companies and Subsidiary Banks as of December 31. (acqd. PA DE. MD DE. PA. NY ME. MD DE. (acqd. NY DE. NY. NY DE. ME 06/01183) First National of Nebraska. NY DE. DE 03110/82) Northern Trust Corporation Chicago. (acqd. DE 06/01182) First Maryland Bancorp. Albany. NY. WA. Inc. (acqd. PA DE. Vl Vl . DE 02/11/82) Provident National Corporation Philadelphia. DE 01103/83) Pittsburgh National Corporation Pittsburgh. Omaha. PA. IL. MD. Baltimore. FL 01108/82) Chase Manhattan Corporation New York. (acqd. (acqd. SD. NC. NC FL. (approved 03/29/83) Norstar Bancorp.IL FL.NE NE. (acqd. PA DE. DE 10/01182) Manufacturers Hanover Corporation New York. CA CA. FL 04/05/82) Maryland National Corporation Baltimore. NCNB Corporation Charlotte. NY. MD DE. (acqd. (acqd. 1982. DE 03115/82) NEWCO Philadelphia. WA AK. PA. WA. (approved 03131183) Philadelphia National Corporation Philadelphia. DE 06111182) Chemical New York Corporation New York. (acqd. PA DE. PA. DE 01120/83) PNC Financial Corp to acquire HC 0465 Mellon National Corporation Pittsburgh. (acqd. MD. (acqd. DE 04/06/83) Mellon to acquire HC 0463 Ranier Bancorporation Seattle. PA DE. (approved 04/14/83) Bank America Corporation San Francisco.

2-2 and 2-5). Total Interstate Offices 16.159 95 Thrift Institutions 1. or states in the Mid-Atlantic or southeastern compacts (Georgia.9 billion as of August 1990 are by far the largest of any state (see table 2-5). Number Number Change in Reported Reported Number Percent Type of Office in 1988 in 1983 Reported Change Bank Offices Controlled by Domestic Bank: Holding Company 7. the more entry activity should be expected. Changes in Interstate Banking Presence (1983-88). States such as Maine and Arizona. and Maryland) have experienced considerable out-of-state entry (see tables 2-1. Accordingly. Federal Reserve Bank of Atlanta (May/June) 1989.492 7.702 1. More liberal interstate banking laws for partner states encourage out- of-state bank entry." Economic Review.086 433 Offices of Foreign Banks 302 241 61 25 Domestic Edge Act Corporations 79 143 -64 -45 Total Offices for Foreign Transactions 381 384 -3 -1 Section 4(c)(8) Offices 6.076 19 Total Offices of Banks 14.258 6.267 N. as more .364 1. Florida. states with potential markets and potential merger partners are the more attractive markets and this should be reflected in entry by out-of-state banks once formidable legal barriers to entry are reduced.A. nonreciprocal entry.446 5. Florida's interstate bank assets at $55.406 6.616 N. Most interstate banking legislation has limited entry by acquisition of an existing bank rather than by de novo bank formation.492 1. The effect has been to encourage out-of-state banks to bid quickly for those banks they wish to acquire in any given state for fear that the price might be bid up by in-state banks desiring to position themselves for an interstate acquisition. Therefore.500 946 17 Loan Production Offices 332 202 130 64 Total Nonbank Offices 6. Source: "Interstate Banking Developments in the 1980s. With the exception of Texas. the lower the barriers to entry. This has stimulated potential acquired banks to choose between being an acquisition target or an acquirer.56 THE CHANGING MARKET IN FINANCIAL SERVICES Table 2-4.106 485 Bank Offices Controlled by Foreign Bank Holding Company 128 148 -20 -14 Total Bank Offices 7.778 5. which allow nation- wide. Thus.A.651 7. activity.

et ai. but with an altered regional focus. Nonbank Companies. federal law has become progressively liberal in allowing interstate acquisitions of failing banks and thrifts as a means of reducing the cost of these failures to the Federal Deposit Insurance Corporation (FDIC). Consequently. As other regional economies begin to experience recessions and as banks in these areas suffer losses from the increasingly severe real estate downturn. Limited-Service Banks. as of 1988 (see table 2-4). and especially Texas. the largest of any state. Germain was passed in 1982.INTERSTATE BANKING. assets held by out-of-state banking companies in Texas increased from zero to $68. and Foreign Bank Interstate Expansion Banks have been able to offer limited interstate banking services through various means since the passage of the so-called Edge Act in 1919. Since 1983. has seen a large number of interstate banking acquisitions of failing or failed banks and thrifts. is substantially due to acquisitions such as those by BancOne (Ohio) and NCNB (North Carolina) of large. 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 Southwest. 1989). . Since Garn-St. The International Banking Act in 1978 liberalized the extent to which Edge Act corporations of banks could offer loans and deposits to customers.9 billion (see table 2-5). the more vigorous should interstate activity become. Given the size of many of these banking companies (those in the Northeast). failing Texas banking companies. but have recently shown a considerable slow down. BANK EXPANSION AND VALUATION 57 states move to nationwide. only acquisitions by an out-of-state superregional bank would be viable. The weakness of banks and thrifts in certain regional economies in the United States has also encouraged interstate acquisitions.. interstate bank expansion is likely to continue at a rapid pace. they will become targets for acquisition because their market values have fallen (making them more attractive takeover candidates) and/or the FDIC will encourage regulatory mergers. nonreciprocal interstate banking. This increase. These institutions grew rapidly in the 1970s and early 1980s. the number of domestic Edge Act corporations has declined from 143 in eighteen states to 79 in sixteen states. Between mid-1986 and mid-1990. 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. but the international trade motivation of the relationship was supposed to be maintained.

therefore. 5). nonbank subsidiaries of the acquired company in the acquired company's state will not be counted as interstate nonbank companies. interstate acquisitions of bank holding companies transfer nonbank ownership to the acquiring company. Since the 1970 Amendments to the Bank Holding Company Act of 1956 were adopted.58 THE CHANGING MARKET IN FINANCIAL SERVICES Many banks maintain loan production offices outside of their home state in order to have a lending sales representation in a local area with- out the need for a large physical presence. nonbank affiliates have the advantage of unrestricted interstate operations. the Federal Reserve Board has permitted BHCs to engage in certain nonbank activities pursuant to section 4(c)(8) of that Act. (1989). Many BHCs have taken advantage of this fact and since 1970 have expanded their operations interstate so that by 1983 there were 5. the activities of bank holding company nonbank affili- ates are in areas traditionally conducted by banks: commercial finance. With these qualifications. consumer finance. The net effect has not been determined.6 The most prominent of these nonbank activities by assets have been commercial finance. consumer finance. they do not serve as an effective means to raise deposit funds and are not superseded as easily as Edge Act corporations when full-service banks are acquired in a local market. one of the dominant means by which banking companies have expanded interstate is through nonbank subsidiaries of bank holding companies. the data in table 2-4 indicate that BHC's use of nonbank . Finally. To a great extent. As sug- gested by King. but it is likely that there may be a net underestimation of interstate 4(c)(8) expansion.500 offices of BHC nonbank affiliates associated with 4(c)(8) activities and 6. Offsetting this to some extent are those nonbank affiliates in the acquired BHC's state that will be counted as interstate companies after an interstate BHC ac- quisition.446 by 1988 (see table 2-4). mortgage banking. One interpretation of this 64 percent increase is that it reflects the continuing need for even the largest banking companies to have a local presence in order to effectively service local loan customers. and leasing. et al. the data may not reflect the full character of interstate nonbank holdings. and securities brokerage. However. A Federal Reserve Bank of Atlanta survey of large banks reported that these offices have increased from 202 in 1983 to 332 in 1988 (see table 2-4). Much of this growth was concentrated in the activities previously mentioned. Even though the recorded expansion of interstate nonbank affiliates has not been as dramatic since 1983 as interstate bank expansion. accounting for $146 billion in nonbanking assets in BHC affiliate nonbank activities in 1987 (see Liang and Savage 1990. mortgage banking. Since deposit solicitation cannot be accom- plished directly from these offices.

7 64.9 53.0 173.0 2.1 57.1 25.8 Massachusetts 2.5 20.2 17.6 North Carolina 0.1 North Dakota 2.5 36.0 1.6 9.2 3.6 .3 Iowa 2.3 Louisiana 0.3 55.2 2.3 Maryland 7.0 Idaho 2.9 5.2 25.6 0.0 Michigan 0.7 8.3 16.0 3.1 0.1 58.0 New Hampshire 0.0 24.0 15.4 87.7 8.9 15.3 Nevada 3.7 0.3 7.0 0.6 4.6 Missouri 0.8 District of Columbia 1.0 55.6 California 0.5 0.7 15.6 New York 1.9 4.0 15.5 3.7 New Mexico 0.3 15.Table 2-5. Assets by State and of Interstate Banks ($ Billions).1 48.1 39.4 92.0 0.4 1.0 94.8 1.1 Kansas 0.4 291.3 Illinois 0.3 Arizona 11.3 3.0 1.8 Minnesota 0.4 6.0 32.0 2.2 420.0 0.0 0.9 0.4 8.3 0.0 50.1 Oklahoma 1.0 84.5 7.4 Connecticut 11.4 38.5 Delaware 0.0 33.9 36.9 Florida 30.0 27.4 46.9 65.0 Kentucky 1.1 19.6 27.9 3.4 24.0 0.5 8.6 106.9 8.4 Alaska 0.9 14.8 14.9 10.1 9.2 Montana 3.4 7.9 New Jersey 2.1 Colorado 2.0 0.5 Arkansas 0.6 0.6 0.1 0.6 27.2 Nebraska 1.5 2.0 15.8 Georgia 10.1 75.8 Ohio 0.7 6.4 2.8 66.3 2.0 5.3 Hawaii 0.9 2.0 15.0 0.6 Maine 4.1 14.8 41.2 36.6 0.5 26.3 36.9 16.4 Mississippi 0.6 Indiana 3.5 18.0 32.2 21.9 8.1 5.0 0.9 133. 1987 1990 Total Total Interstate Interstate Total % State Assets Assets Assets Interstate Alabama 0.0 0.2 22.

6 12.915.0 0.2 29.4 2. there might be a flurry of interstate acquisitions of regional invest- ment banking companies.5 27.4 44.3 42.3 33.4 442. Table 2A. Board of Governors of the Federal Reserve System.0 1.6 67. interstate banking has been con- sidered by many banking company managements as a preferred means of conducting banking business.2 19.8 78.0 21.7 17.9 31.1 Rhode Island 3.S.8 18.8 37.5 1.1 Tennessee 0.0 47. If full investment banking powers were per- mitted.5 22.4 24.8 9. February 1987.60 THE CHANGING MARKET IN FINANCIAL SERVICES Table 2-5.6 13.6 4. (Cont.9 167. affiliates continues to be strong even though banking alternatives have been the dominant means for interstate expansion since 1983.5 Pennsylvania 0.9 11.0 68.2 Wisconsin 1.4 4.8 14. and Financial Studies Section.9 18.2 41.2 Source: Federal Reserve Bulletin.0 160. This is evidenced not only by the current rush by banks to form interstate relationships.0 1.7 West Virginia 0.3 3.1 Virginia 2.2 17.8 Vennont 0. The prospects for bank expansion via nonbank BHC subsidiaries will likely hinge on the willingness of Congress and/or the Federal Reserve Board to permit expanded nonbank powers.4 6. 148.1 Wyoming 0.3 8.2 45. September 1990.0 5.0 Total U.9 Washington 12.2 29.9 15. but by the number of .9 2.7 South Dakota 2.0 16.) 1987 1990 Total Total Interstate Interstate Total % State Assets Assets Assets Interstate Oregon 5. Competition and Other Public Policy Issues Arising From Interstate Banking Dating back to the previous century.0 South Carolina 6.0 Texas 0.6 4.2 Utah 1.

As suggested above. The motivations for banking management to expand interstate are numerous-not the least of which is the desire to secure a monopolistic position in a banking market or many banking markets. Complementary to this motivation is that of establishing a dominant market position through absolute firm size and/or a large market share. Further evidence of the desire for interstate banking rela- tionships is the extent of the complex of correspondent banking relationships that continue even in the era of extensive electronic funds transfer and information media. The effect of this legal structure implies that banking companies must acquire existing banks or bank holding companies in order to expand interstate. The Competitive Issues and the Evidence The interstate banking movement is a transition from a condition of re- stricted (practically blockaded) entry of banking companies headquartered in one state into another. this condition represents a disequilibrium caused by governmental prohibitions to bank expansion. Because of the resources required for such expansion and the managerial experience to coordinate the expanded organization after the merger. many banking company owners and managers found it in their best interest to expand interstate. The dual banking system. Once these restrictions were lifted. only the largest and better capitalized bank holding companies have been the major . BANK EXPANSION AND VALUATION 61 interstate banking companies established before the 1956 Bank Holding Company Act and the one-bank-holding-company attempts to expand interstate and by product line in the 1960s via nonbank holding company subsidiaries. Effective provision of banking services in local communities 3.INTERSTATE BANKING. The issues considered are those raised in the Treasury Report on Geographic Restrictions on Commercial Banking (1981): 1. Competition and concentration of financial resources 2. de novo entry and interstate branching are not possible. Safety and soundness of the banking system 5. Under the present structure of state laws governing interstate banking. In this section the analysis turns to considering the competitive and other public policy issues arising from interstate banking and to public policy recommendations for improving the transition to a desirable competitive environment once interstate expansion has reached a steady state. Viability of small banks 4.

for both SMSA and non-SMSA markets. This type of expansion and the lowering of interstate entry barriers may be thoroughly consistent with fostering better banking services. As measured by the three-firm con- centration ratio. even though some are among banks competing in the same local markets (for example.C. Interstate banking acquisitions or mergers are primarily of the market-extension type. There has not been sufficient time for interstate banking to produce a record of effects on local market and statewide concentration. are mixed.8 percent in 1982 to a value of 67. and Amel and lacowski 1989) have attempted to address this issue. Rhoades compares local mar- ket concentration as measured by the three-firm concentration ratio and the Herfindahl index from 1966 to 1981 and concludes that. Many of these companies such a NCNB (North Carolina). These acquisitions may be considered more favorably by banking authorities if interstate banking were in place and result in an increase in local market and statewide concentrations. metro- politan area). Rhoades concludes from these data that interstate banking will likely have little effect on local and statewide banking market concentration if the same banking company merger standards are enforced. with the exception of states with strict unit banking laws and prohibitions on multibank holding companies. Citicorp (New York). Additional evidence by state reveals that the rise in local . He also finds that there is no difference in the average local market con- centration among states with different interstate banking laws. the average value increased from a low of 65. Some evidence of a reversal of the declining trend of local banking market concentration has been reported by Amel and lacowski (1989. particularly in urban markets. and lessening the concentration of banking resources. The possible competitive effects. mergers between several Virginia bank holding companies and Maryland and District of Columbia banking companies were among banks competing in the Washington. 12) after making a study of geographic restrictions on commercial banking. However. Rhoades does not consider the possibility of banks within states or regions consolidating in order to increase their size and value to ward off potential unwanted acquirers. BancOne (Ohio). The affirmative of this position was taken by the Treasury Department in a Report of the President.5 percent by 1986 (the latest date for which data are available). However. reducing monopolistic profits. (1981. several authors (Rhoades 1985. if this form of expansion continues. local market concentration has declined.62 THE CHANGING MARKET IN FINANCIAL SERVICES participants in interstate expansion. D. 132). First Interstate (California). Savage 1987. and Sovran (Virginia) have expanded into multiple states forming a network of interstate banking offices.

and generally better capitalized. As was suggested above.1 percent in 1985 to 36. These consolidations may have been the cause of the rise in local market concentration beginning in 1982. national financial and. there had been many intrastate bank and banking holding company mergers and acquisitions in preparation for interstate banking. particu- larly. (table 2-6).5 percent in 1990. BANK EXPANSION AND VALUATION 63 banking market concentration has been widespread. these data suggest that the trend reversal in local concentration occurs at the time when many states began to adopt liberalized interstate banking laws. however. banking resources has been of serious political and economic concern in the United States since the beginnings of the nation. From these data. The balance between a concentrated federal system and a more dispersed state-controlled sys- tem is seen nowhere more sharply than in the U. Coupled with interstate banking laws of many states placing limitations on money- center bank entry into these states and large banking company failures in . and for the largest fifty banking companies from 45.2 percent in 1990. despite remaining practically constant over the previous fifteen years. The most recent data show that the share of banking assets of the twenty-five largest domestic banking companies has increased from 33.INTERSTATE BANKING. the non-money-center banking companies have been the most aggressive. Additionally. By themselves. These data indicate that thirty-three states experienced an increase in the average three-firm concentration ratio in urban markets of those states from 1981 to 1986 while only seventeen states recorded a decrease. As was discussed in the previous section. via the BHC Act as amended. These concerns over an undue concentration of resources in banking may presently be misplaced. dual banking system with its federal and state chartering and regulatory control of banks and thrifts. the structure of the Federal Reserve is a compromise between those wanting a strong central bank and those wanting regional control of this power. As the authors remark. has been in the share of the one hundred largest banking companies which rose from 57. However. there has occurred over the past five years a marked increase in the share of banking assets controlled by the largest firms.7 The undue concentration of aggregate. the largest increases have taken place among those banking companies in the superregional group and outside the top ten inasmuch as there has been little change in the share of the top ten.7 percent in 1985 to 51. interstate banking acquisitions and mergers should not have a direct effect on local banking market concentration. over larger banking companies.7 percent in 1990 (see table 2-6).S. Interstate banking does not often change state or federal control with one exception: It gives the Federal Reserve author- ity. The greatest increase.7 percent in 1985 to 64. in interstate banking expansion.

These larger superregional banks with banking offices in a number of states may be more able to compete for customers over a wider region or nationally than they had been able to do with a narrower. In contrast to this sanguine outlook for competition in the face of increasing national and statewide concentration.4 1975 13. regional setting. this hypothesis would suggest that as banks expand interstate they will begin to be competitors or rivals in numerous markets throughout the United States.7 36. these developments in nationwide concentration might be considered healthy. Shares of Domestic Commercial Banking Assets Held by Largest Banking Organizations 1 (Percent).6 33.6 51.to larger-size busi- ness and household customers.3 32. Only insured commercial banks are included. is the view that anti- competitive effects may arise from "linked oligopoly" or multimarket interdependence among competitors.8 1980 13.8 20. attributed to Corwin Edwards (1955) and applied to banking.3 33. Such developments may increase competition for medium.4 1985 12.1 41.7 64.8 41. As the progress of interstate banking proceeds. Table 3. there is little wonder that nationwide concentration has increased in the particular way it has.1 21. Within the context of interstate banking. This hypothesis.1 50.7 21. they will develop a mutual awareness and inter- dependence and avoid aggressive behavior in all mutual markets and retaliation in any market.6 41. Financial Studies Section.1 50. Year Top 5 Top 10 Top 25 Top 50 Top 100 1970 14.4 32.5 21.64 THE CHANGING MARKET IN FINANCIAL SERVICES Table 2-6. These results demonstrate that a group of superregional banks are developing. the Southwest. These results need not be interpreted as alarming or as a substantial degradation of national competition. is that as multimarket firms meet in numerous markets.1 45. more banking companies that have not previously been direct competitors will become . 2 June 1990 Report of Condition Source: Federal Reserve Bulletin.7 1990 2 13. September 1990.5 51.2 1 Banks are ranked by banking assets. willing to take risks to expand nationwide and ready to challenge the traditional money-center banks for banking customers throughout the nation. Indeed.0 21.7 57. February 1987. and Board of Governors of the Federal Reserve System. non- deposit trust companies are excluded.

without clear evidence of anticompetitive effects. 9) finds. No research has directly addressed this competitive issue within an interstate banking framework. This suggests that as interstate expansion proceeds. This conclusion is supported by Mester (1987). BANK EXPANSION AND VALUATION 65 so in many local markets. The com- petitive effect is to blunt the pro-competitive aspects of interstate banking such as new entry by independent banking companies into previously blockaded markets and the potential competition of new firms not presently in these markets. antitrust policy and enforcement are impotent in redressing competitive harm from interstate banking. but at least the study by Hanweck and Rhoades found a significant relationship between firm dominance and rivalry within local banking markets. As Rhoades (1983. This result has not been unanticipated. To a great extent this research has not provided much support for this hypothesis. there has been research in this area using data from intrastate bank expansion into local markets (Rhoades 1983. If the dominant firm-deep pockets- hypothesis is correct. Curry and Rose 1984.INTERSTATE BANKING. so called "deep pockets. Their ability to survive such competition under interstate banking will depend on their cost efficiency compared to much . there may result large. there is no systematic evidence of multimarket linkage among competitors and various measures of local market rivalry such as mobility and rates of turnover. giving rise to greater consolidation among banking companies and to more large- size banking firms nationwide. and Mester 1987). However." of multimarket banking companies is a source of barriers to entry into local markets and may reduce interfirm rivalry (see Rhoades 1983. 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. nationwide banking firms which. Not only does further research need to be done on this and the multimarket interdependence hypothesis. as the evidence presented in this study shows. Research in this area has been mixed. A related hypothesis maintains that the absolute size and financial strength. but. substantial barriers to entry may be erected leading to significantly higher than competitive prices for banking services in these markets. gain dominance within local markets.8 The Survival of Small Banks Interstate banking is. The development of mutual forbearance among these various competitors is then likely to become established. and Hanweck and Rhoades 1984). by acquisition and from a pre-existing market structure.

67). Berger. 9 The achievement of scale and scope economies is of dubious worth inas- much as the overwhelming number of studies have found economies of scale in banking to exist only up to at most $500 million in assets (see Benston. 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. mentioned earlier. the evidence since 1985 (Amel and Jacowski. and it is difficult to find much difference between grandfathered BHC affiliates and banks of the same size in the same state. 1985. Studies by Berger. Of the studies that have recently addressed interstate banking com- petitive effects directly. 1988. by 1983 the homogenization process appears to be complete. strong local deposits bases. all have found no meaningful anticompetitive or procompetitive effects. and their efficiency in adopting new technology. 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. Hanweck and Humphrey 1982. These results are similar to those of other related studies (Rhoades 1984. The study by Phillis and Pavel (1986). 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. 125) clearly shows that the small bank is far from becoming extinct (see Rhoades. Productive Efficiency and Safety and Soundness Issues Bankers have contended that interstate expansion is necessary to achieve economies of scale and scope in the production and distribution of bank- ing services and achieve greater geographic and product diversification. and Clark 1988 for a review of this literature). 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. Hanweck and Humphrey 1987. for a similar pre-1985 assessment).66 THE CHANGING MARKET IN FINANCIAL SERVICES larger banks. In this regard. their ability to generate profits. By 1970 the extent of the homogenization is nearly complete. Hanweck and Humphrey (1987) and . Employing a simple rule of inference. 1130. found that banks expanding interstate preferred to acquire banks with larger branching networks. Economies of scope fare no better. and a greater concentration in consumer lending. 1985.

Interstate banking provides a stop-gap avenue for acquisition of failing banks that. Most of these studies do suggest that unit banking is inefficient. on a less positive note. BANK EXPANSION AND VALUATION 67 Gilligan and Smirlock (1984) found little to suggest that scope economies are important in banking. As discussed above. 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. Their results suggest economies of scale among the largest banking companies at about $20 billion in real assets. when unavailable. Banks can now become more geographically diversified in both their deposit liabilities and lending in commercial and retail markets. Many states have reciprocity provision that prohibit entry by banks from states that prohibit interstate banking. Several recent studies have focused on scale economies using only the largest banks in their estimates. interstate banking companies may be able to rely more upon less volatile core deposits related directly to their banking business in local markets. some states have tried to attract entry by banks throughout the country by not placing restrictions on interstate banking. resulting in significant diseconomies of scale at larger sizes and which were not exhibited by branching banks (Berger. a number of . will be bid up. the FDIC will become increasingly vulnerable to such failures resulting in greater uncertainty and instability within the banking system. Analysis of the effects of state laws upon bank valuation must distinguish among the types of state laws and among the types and loca- tions of banks. Finally. 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. With the continued increase of larger banking companies. Because of the larger available markets. 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. and the market value of those banking companies being potential acquirers. raise the cost to taxpayers and the FDIC of failed bank resolutions. The safety and soundness of the banking system may be enhanced through interstate banking expansion. Hanweck and Humphrey 1987). Finally. lO Interstate Banking and Banking Company Valuation Nationwide interstate banking may increase the number of bidders so that the prices of banks that might be targeted for acquisition.INTERSTATE BANKING.

Adkisson and Fraser (1990) consider the effects of the removal of barriers to entry by liberalized interstate banking laws on premiums (price paid less book value) paid by acquiring banks in banking mergers and acquisitions. 11 The effects on bank valuation is likely then to vary by type of state law. Studies attempting to assess the effect of interstate banking law changes must be able to examine the effects of announcements of these changes not only on banks eventually acquired. target banks in states with less restrictive intrastate banking laws and. This method has not been entirely successful as the money-center banks have obtained entry into some of the restrictive states through other means. In a recent study. but on all banks in all states that might be affected. and particularly advocate entry on a national rather than a regional basis. Based on these findings. The authors' results support the first hypothesis since they find statistically significant higher acquisition . Their choice of time period was made to mainly include acquisitions that occurred after the Supreme Court in June 1985 upheld the constitutionality of regional banking pacts in the Northeast Bancorp decision. The second hypothesis is that the reduction in entry barriers. resulting in higher merger premiums. The effects of state law changes on bank valuation should vary also by type of bank. Additionally. and De and Millon 1988). lowers their merger premiums. They test two separate. as a result of reducing the local market power and potential excess profits of target acquired banks.12 In contrast. A large volume of literature on mergers generally finds that acquired firms experience increases in stock value. De and Millon (1988) find positive announcement effects on returns for both acquired and acquiring firms in interstate banking acquisitions. but the results are mixed and usually insignificant for acquiring firms. This type of rule is intended to deter money-center banks from entry into these regions. Adkisson and Fraser use an analysis of covariance and a sample of 174 bank holding company acquisitions in 1985 and 1986. James and Weir 1987. hypotheses.68 THE CHANGING MARKET IN FINANCIAL SERVICES states have combined into regional compacts. will tend to have lower merger premium offers. permitting entry only by bank holding companies within the region. a number of studies on bank mergers have found positive announcement effects on returns for acquiring firms as well as acquired firms (Desai and Stover 1985. but related. The first is that the removal of interstate banking restrictions increases the number of potential bidders for target acquired banks. thus more market opportunities for expansion by potential acquirers. according to this hypothesis. such as acquisition of failing institutions. they recommend relaxation of interstate restrictions. In one of the studies addressing interstate banking acquisitions.

The . these interactions cannot be as- sessed even though conceptually meaningful in determining bid pricing. in face of substantial evidence of unprofitable and perhaps even negative net present value acquisitions and mergers. BANK EXPANSION AND VALUATION 69 and merger premiums associated with acquisitions in states allowing interstate banking and for those states with unlimited intrastate holding company expansion. from 1960 to 1983 these organizations experienced a statistically signifi- cant decrease in their share of state deposits. made possible by liberalized interstate banking laws. Savage 1987. gain scale or scope economies. As was mentioned.INTERSTATE BANKING. The issue of motivations for interstate expansion. In fact. Evidence supporting the view that even longer term advantages of interstate banking are not necessarily forthcoming to those companies expanding interstate is found in Goldberg and Hanweck (1988). This suggests that acquiring organizations would not be expected to gain significant advantages. Un- less surviving banking companies can operate more efficiently than the acquired banks. the advantage of interstate banking to acquiring banking companies is doubtful or extremely long-term and beyond the horizons of most investors. Since the authors provided no evidence of effects on premiums of acquisitions in restrictive intrastate banking and liberalized interstate banking state or vice versa. Rose (1990. studies have shown that interstate acquirers prefer banking companies with more extensive branching systems in place (see Phillis and Pavel 1986. or reduce risk. and Rose 1990). In this study the author compares factors possibly influencing the risk and expected value of the firm to shareholders in the case of seventy-seven acquired banking companies and ninety-two acquiring companies involved in interstate acquisitions. This study examined the performance of the seven grandfathered interstate banking organizations and found them to differ little from their peers. on average. across the entire sample of banking firms reviewed. is considered in a recent study by Rose (1990). All acquisitions involving a failed or failing banking company were eliminated from the respective samples. as the authors conclude. Potential increases in value may be competed away in the bidding for acquisitions and later by the greater competition for banking services arising from reduced entry barriers. there is unlikely to be an increase in value for surviving banks in the long-run. their results confirm the common finding that excess returns are negative for acquiring companies in bank acquisitions. interstate acquirers generally performed less well than comparably sized and located noninterstate banking companies displaying greater average risk to capital and lower average asset returns and operating efficiency. If. 39) summarizes the results from the study as follows: The Study finds that.

39-40) concludes that: While these 'mergers of equals' may thus be expected to have relatively few benefits from a diversification standpoint. etc. 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 . for example. The study by De and Millon (1988) on interstate banking acquisitions indicates that acquiring firms have increased in value. the first state to adopt interstate banking legislation. 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. the liberalization of interstate banking might be regarded as a stimulant for economic development in a state... Consequently. Alternatively. The effect on banking company valuation from changes in interstate banking laws is evaluated in a recent study by Goldberg.. or reduced risk exposure. Therefore.70 THE CHANGING MARKET IN FINANCIAL SERVICES same deficiencies appeared to characterize the firms targeted for acquisition across state lines relative to peer institutions . or greater price discrimi- nation. Maine. a new law might trigger reciprocity provisions in other states and thus increase expansion opportunities to lucrative out-of-state markets. 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.. was anxious to encourage the New England banks to bring resources to the state. Hanweck and Sugrue (1990). 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 .. From another perspective. especially for those banks most likely to be acquired. 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. An increase in the number of potential bidders for a bank would likely increase the value of a bank in a liberalizing state. the markets at the time of the announcements of liberalized interstate banking law changes should have taken these factors into account... In addition. for acquisition-bound banking companies. the value of established firms in a state could be reduced by increased competition from out-of-state entrants. They hypothesize that announcement of a liberalization in interstate banking laws would positively affect the stock prices of banks in that state. Rose (1990.

Third. The results reported in this study are similar for both sets of dates. Hanweck and Sugrue (1990) . In the empirical analysis. The authors estimate the effect of interstate banking law changes for thirty-five passage events and thirty-five effective events using a sample of 131 banking companies with continuous trading over the period of February 1. Goldberg. In addition. Second. Not only are the events clustered. 1987. the study is able to determine the effects on those banks outside of states and regions for each separate event period and. they have already been able to engage in various interstate operations so the new powers would not provide relatively as much in increased opportunities as for smaller banks. In addition. money-center banks might be expected to be less affected by the opportunity for interstate banking expansion for several reasons. The matrix by state and event date of interstate banking law changes is presented in table 2-7 and demonstrates the complex of events that took place for which effects on bank valuation must be estimated. Banks are identified not only if they are in a state that had an interstate banking law change. Banks in nearby states should be affected to a greater extent. In fact. First. Estimates were made of the fifty-one nonoverlapping passage and effective dates of interstate banking law changes using daily stock return data and the Zellner (1962) seemingly unrelated regression technique. but they are expected to affect some banks differently than others. can measure separate effects for all banks at different times. The results of the study by Goldberg. especially when a regional interstate banking law is the applicable law. Enactment of the law represents the realized effect on bank value. new interstate operations would be a smaller share of a money-center bank's operations and thus would have less of an impact upon firm value. Hanweck and Sugrue (1990) measure the reactions to the new state laws in two ways. but if they are in a state with a regional interstate banking compact. if the regional interstate law is expected to allow smaller regional banks to grow interstate and become stronger competitors with money-center banks.INTERSTATE BANKING. Passage of the law reflects anticipation of the law's effect. This study differs from most other event studies in that it attempts to simultaneously estimate the effects of numerous events affecting many banking companies over an extended period of time. a physical presence could be expected to be relatively more important for smaller banks. the value of money-center banks might very well decrease. to September 1. therefore. the authors separate out-of-state and out-of-region money-center banks from other commercial banks in order to uncover the different effects of the laws. BANK EXPANSION AND VALUATION 71 interstate expansion were considered to have no special advantages. 1982.

Table 2-7. Banking Law Dates and Reciprocal Provisions. A A A A C CCDDFH G I I I I K K L MM State Passage Effective L K Z R A 0 T E C L I A D L N A S Y A E D 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 .

they do not appear to be positively affected by the liberalization of interstate banking. This total excludes acquisitions by foreign banks. Except for money- center banks. The authors find that in-state and in-region banks benefitted most from the liberalization in interstate banking that has taken place since 1982. Because money-center banks already have some interstate operations and the new opportunities. and . In view of adverse developments in profitabil- ity. market valuation. out-of-state and out-of-region banks appear also to have benefitted from passage and enactment of the liberalized interstate banking laws.74 THE CHANGING MARKET IN FINANCIAL SERVICES support those of Adkisson and Fraser (1990) and contradict. Future Interstate Bank Expansion and Bank Profitability and Capitalization 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. Coupling this with the forecast recession in the general economy. This can be due to increased expansion opportunities and to triggering of reciprocity provisions which may increase the availability of potential acquirers of the out-of-state banks. This has apparently outweighed the negative impact on value of increased competition. with each having in excess of $1 billion in domestic deposits. to some extent. Most of these domestic banking acquisitions have taken place since 1982 and about thirty-two are acquisitions of banks interstate. It is in this regard that these results are in direct contrast with those of Rose for acquired and acquiring banks over the same time period. do not represent a relatively large part of their busi- ness. 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. Finally. Both the increase in the number of potential acquiring firms and the added expansion opportunities for these banks would be expected to increase their value. From 1980 to 1989 there have been over ninety-five Federal Reserve approved bank mergers and holding com- pany acquisitions among banking companies. the generaf findings of this study suggest that investors in banking companies believe that the liberalization of interstate banking will have a positive impact on the future ability of banking companies to increase their profitability and value. those of Rose (1990). as a result of the legal changes. The char- acter of this expansion has been such that regional banks have been at its forefront of this expansion.

for which consistent data are available for banks by size classes.84 . BANK EXPANSION AND VALUATION 75 Percent ROA 1. 0..."" .00 ... the ROA for all banks varied from a high of 0.\ : -0. -0.00 1983 1984 1985 1986 1987 1988 1989 Source: Board of Governors of the Federal Reserve System Figure 2-1.60 / / .-------- .. overall bank profitability has both declined.40 ...20 .40 ----_J / / .. and demon- strated a high degree of variability (see figures 2-1 and 2-2)... .. . \ .All Banks -~_ -----...20 ROA-Small ROABanks 1. -0...... t --------~~o. Bank Profitability and Cash Flow During the 1980s.\ 0. and stiffer regulatory capital requirements and oversight of commercial real estate lending.~ -------------~--­ ------------\ 0. 0. "" ...00 Banks $300M to $5B in -l Assets 0.. ...:-.. Over the period from 1983 to 1989... Rate of Return on Assets (ROA) by Bank Size real estate in particular. _ _-7'-~".... .. -0...60 \ I ROA 9 Money \ : Center Banks . . ..l ..80 'J -1.--\ ... many banks will find it almost impossible to expand without greater equity capital and/or regulatory assistance in FDIC-assisted acquisitions..INTERSTATE BANKING.....-.80 --. ROA ...20 ROA Other Large \ Banks .. in terms of either rates of return on assets (ROA) or on equity (ROE)..

but even facing difficult earnings pressures from agriculture and energy price declines....~ ...1'- / ---- .... those with assets less than $5 billion. .. ::\. . ~~~~ J ---.....00 ROE . . Assets 10..11 percent Likewise.. $3OOM to $5B in .. ROA and ROE for smaller banks. -5..00 ..76 THE CHANGING MARKET IN FINANCIAL SERVICES Percent ROE 25. ROE-Small \ • '. .... and particularly larger banks. may have declined over the period. Although it is not the purpose of this study to dissect the banking industry. • 5. ---------. For larger banks. .. \ Banks \ \/' 0.00 ROE Banks 15......00 20... Rate of Return on Equity (ROE) by Bank Size percent to a low of 0.... it is sufficient to point out that these problems have persisted for much of the 1980s. .00 ROE 9 Money \ Center Banks \ : ' ..-----i 1983 1984 1985 1986 1987 1988 1989 Source: Board of Governors of the Federal Reserve System Figure 2-2.. ROE for all banks over the same period ranged from 13... ' \ : "'I -20.. :..... .. .. -15.-----------.. their sources are well known and . .00 • ..." .... / / ..00 +../ T . to determine the causes of the present profitability problems...94 percent These values are in sharp contrast with bank earnings performance even during some of the turbulent periods of the 1970s.... . their variability was not excessive. -10...00 .7 percent to 1... As figures 2-1 and 2-2 demonstrate.. ..00 ... .. t ---- .... ..: \ \ / ...~ 'I : ROE Other Large Banks '/: .--.. -:. .. Much of this variation and general decline is due to the performance of large banks.. ..00 -. . ..All Banks ..-.---+-----+---4-----+----+----f. ... ".. ~ ----.

00 --------------------- .... Other large " .__ .. __ i ____--------.... 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..... .....50 1..... -." Banks "' I" "" 1.........-} /' T ~ .......... ----------....50 ----------....--..- 3._ ~~ 2... the net interest margin to total assets (NIM) is presented in figure 2-3 for all banks and those of the four size classes... NIM ... Only for the nine money-center banks has there been any significant variability... In order to emphasize this point.......--_ .. Banks . Banks $3OOM to $5B in Assets t ~~####~---------------------........INTERSTATE BANKING... Without exception.......... ----------- 3...-.. The principal cause of the variation and downward trend in bank ...... . BANK EXPANSION AND VALUATION 77 Percent NIM to Assets 4....00 1983 1984 1985 1986 1987 1988 1989 Source: Board of Governors of the Federal Reserve System Figure 2-3. _ 4....-"..... All Banks -- t /' ~------.. ---... Banks /' --...00 .... Net Interest Margin (NIM) to Assets by Bank Size stem largely from the overhang of less developed country (LDC) loans originated in the late 1970s and regional real estate recessions......... there has been a slight upward trend in NIM over the 1983 to 1989 period......50 NIM-Small ..... 9 Money Centet----_ ... .....00 ..--..

.. .-'. . I . which have shown a re- versal since 1986... .. Banks I .50 " .. ~.'..... .*' .. I .. . .... ...- LLP to Assets .... .50 • 2. . .t1---__ ____~-. those with assets less than $300 million... . . .. /~ . as demonstrated by loan loss provisions as a proportion to assets. .00 " I. ' "". \\ l/ . . ....'.. .. .:: . .. . Banks $3OOM to 'I $5B in Assets 0.. '/ t .. Even though banks have been able to maintain consistently strong net interest margins. there is little likelihood that this trend will be reversed in the next several years.. .I . All Banks ' . ..78 THE CHANGING MARKET IN FINANCIAL SERVICES Percent LLP to Assets 2. The variability and downward trend in bank earnings has limited bank- .. ...~".. . • ..00 LLP-SmaU II I II \'\ \ ' I I / :~~Large '~ I / l/ . . I I '.. . . 0. loan losses have cut deeply into this basic source of bank earnings. provisioning for loan losses remains on an upward trend (see figure 2-4). I .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. . 1. .9 Money Center I I \. . . .... I ... . Loan Loss Provisions (LLP) to Assets by Bank Size profitability is due to loan losses. With the exception of the smallest banks.... ..... .. ".... ~ /~"".' . . .".. Taking into consideration the possibility of a general economic recession in 1990 and the spreading regional real estate recessions. .:. . ... . .. . I .. .00 1983 1984 1985 1986 1987 1988 1989 Source: Board of Governors of the Federal Reserve System Figure 2-4.. . .. ... . .----. .

. Equity to Assets by Bank Size ers' abilities to add to capital from internal funds..00 Equity to Assets Small Banks _------ 8.... ~~~~ .. ~------ Banks 4..00 3.. to add to equity capital. ...00 --- ~~~----- j ------... 7...... 6.. 9 Money Center .. bank regulators have adopted more stringent capital requirements and en- forcement procedures..... _ --..00 " ~/ --------~ "" Other Large --_... As figure 2-5 portrays...... Banks ......... -_.. the smaller banks have the highest ratios of equity to assets. 5. .... In addition.....00 Banks $300M to '_"_"_'_"-'~":::"'::--::-~-----------------------------~-!~~7----------~ -.."..INTERSTATE BANKING. nearly double that of the nine-money-center banks..00 Equity to Assets All Banks ~-----~~... especially the largest banks.00 1983 1984 1985 1986 1987 1988 1989 Source: Board of Governors of the Federal Reserve System Figure 2-5... ....- .... With the high degree of earnings variability demonstrated by the largest banks in recent years.-~--...... All of these factors contribute to pressure on banks. . investors in these banking firms and their large uninsured depositors will demand more protection in the form of greater equity... BANK EXPANSION AND VALUATION 79 Percent Equity to Assets 9.... 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....

it is likely that they are used to meet holding company expenses associated with nonbank enterprises and payouts to shareholders of the BHC.25 ~..10 1983 1984 1985 1986 1987 1988 1989 Source: Board of Governors of the Federal Reserve System Figure 2-6. cash dividends are paid to the BHC parent holding company.55 0.. ... .. For many banks. ...35 ...'' " 0.....15 0. Regardless of their use..' .20 9 Money Center Banks 0.. Although these dividends may be earnings retained by BHCs.80 THE CHANGING MARKET IN FINANCIAL SERVICES Percent Cash Dividends to Assets 0. . ..45 0.All // Banks .. Cash Dividends to Assets by Bank Size Despite the obvious need for equity by banks of all sizes..'. .... they are a source of ....50 0. ..40 / r .::::::. ----.. in 1989 small banks paid out 58 percent of after- tax earnings as cash dividends and the nine money-center banks. there has been an alarming upward trend in cash dividends paid by banks (see figure 2-6).. . their banking subsidiaries are their largest holding and cash dividends their predominant source of revenue....3 billion. For example../ / 0... For many of these holding companies. . exper- iencing an after-tax loss..t. paid 0.I'. Banks / " 0.36 percent of their assets as cash dividends or about $2.30 Dividends to • / / Other Large / . f"" """l/' Assets ..:.... 0.

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. 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). To emphasize the divergence and clarify the timing. etc. By way of contrast. bank stock values began to lag behind those in the general market. the divergence from the S&P 500 began about 1980 or 1981. These data clearly show the divergence in valuation that began to take place in the early 1980s. Some of the major gainers were BankAmerica (California). while all industries showed a gain of 15 percent and general nonbank financial services (insurance. 1988 was an improvement for most banking companies. Crestar Financial (Virginia). 1990. while in 1987 banking company market value over a broad range of companies showed a decline of about 11 percent (Business Week. the market valuation and capitalization of major U. Shawmut National (Massachusetts). brokerage. As reported in Business Week (April 13.S. These data show that. Smaller regional banking companies have fared no better than major regional and money-center banks since July 1989.234). NCNB (North Carolina). At about the time of the start of the bull stock market in 1983. Banking Company Capitalization and Valuation For the banking industry and particularly the largest banking companies. and First Interstate (California). while some of the larger losses were experienced by Bank of Boston. 219-222). Based on these data.INTERSTATE BANKING. April 15. banking companies have not kept pace with a broad range of firms in other industries. 1988. Southeast Banking (Florida). In figure 2-7. thirty-three showed market value declines. Of the ninety-two banking companies reported. market valuation has not kept pace with other industries in the United States. There are certainly individual exceptions to this generalization.) had a 14 percent increase. Citicorp. for the regional banks. For the money-center banks. this divergence began much later in 1985 or 1986. 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 . and Banc One (Ohio). figure 2-8 shows the ratio of the S&P 500 index to the two banking company indices. ninety-two major banking companies lost 9 percent in market value in 1989. Bank of New York. but data for a broader spectrum of banking companies support this same conclusion.

S._""_-""--.... _ _ _ _ .._\ I " .j " V....' '\ l . These are substantially greater declines than the 7 percent decline in the S&P 500 stock index over the same time period. and the erosion of the profitability of these compa- .... __... industries......82 THE CHANGING MARKET IN FINANCIAL SERVICES Stock Price Index 350 300 250 S&P500- 200 150 Regional 1\ 100 . 1989.1 ... From July 1989 to November 1990..'\ . and the value for the smaller regional banks declined by 47 percent. The marked decline in larger banking company market valuation relative to other U. Bank Stock Indices Value vs S&P 500 Quotations (NASDAQ) Bank stock index composed of 209 regional banking companies (NASDAQ Regional Banks).. '... .. \ I \ I ' .J. 50 .... ' ......iJ' Money Center .. 1990 Figure 2-7..--. Latest data plotted: August 29. ~ Banks A 1V \. the stock value of major regional and money-center banks declined by 45 percent.---' Banks o 1958 1962 1966 1970 1974 1978 1982 1986 1990 Source: Standard and Poor's Analyst's Handbook: OffIcial Series. ____"J .....~\ '..I' '...

will impede the expansion of banking companies interstate. the valuation of smaller banking companies has not fared much better than that of the largest companies. However.5 3 S&P to Regional Banks 2. In view of these data and the serious problems arising in commercial real estate in a number of formerly real-estate-boom regions of the United States. 1989. S&P 500 Index to Bank Stock Indices nies. Figure 2-8.INTERSTATE BANKING.5 o 1~ 1~ 1~ 1~ 1m 1m 1m 1m 1~ 1~ 1~ 1~ Source: Standard and Poor's Analyst's Handbook: Official Series. BANK EXPANSION AND VALUATION 83 Percent S & P to Bank Indices 4.5 t 2 1. the next few years may provide an exceptional . Unfortunately. banks will need to devote their resources to supporting loan losses and employ managerial expertise to survival and consolidation of operations.5 4 3. these smaller companies have not suffered as severe erosion in profitability as the larger companies. For those banking companies that are presently well-capitalized and have a strong cash flow.5 0.

. Bank Stock Indices January 198Q--November 1990 ."' 140 .J .I I. 170 . 320 \1\ NASDAQ Regional Banks. 1 \ 410 J\(J 'I II I I \I I I: II 380 I I! I " \ 350 ) I \. \ II ~.N . 470 I \-J I II 440 IA'I /Vil II \I\ '('Jt\l\ I . Bank Stock Indices 530 500 1\ . I / I 260 r/ \V I 230 200 /"" {' j"' y (' .A.J(~j 110 r~J \J 80 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 Figure 2-9.J 290 t.

Under present Federal law. Current . Of the numerous studies reviewed.INTERSTATE BANKING. All of these may have some benefit to the public in terms of less expensive and more competitive financial services. However. There has been little evidence that the provision of financial services has become any more competitive or that banks have improved their operat- ing efficiency. banking companies cannot open offices outside their home states unless explicitly permitted by state law. The principal policy prescription from this study is that there needs to be a thorough review of federal law dealing with interstate banking. several indicated that banks have incrementally improved their market values as a result of the passage and effective implementation of liberalized interstate banking laws. Consequently. there may also be little reason to expect noticeable im- provements in the efficiency of bank operations. In fact. rapid expansion into markets that are poorly understood is a prescription for failure. The most notable observation that can be made of the expansion of interstate banking is that there has been a substantial consolidation of banking resources within the largest fifty and one hundred banking companies. there may be little reason to expect that entry restricted only to acquisition will result in improved competition in local banking markets. However. Conclusion and Public Policy Recommendations Liberalized interstate banking legislation since 1982 has presented banking in the United States with an opportunity for the expansion of geographic markets. there are those studies that suggest that acquiring banking companies will not fare as well as they expect once expansion is completed. a consolidation of banking resources. or safety and soundness. Few studies have been able to convincingly demonstrate that there are significant scale and/or scope economies in banking to cause an improvement in operating effi- ciency as a result of simply increasing size. many studies suggest that without proper diligence by management. BANK EXPANSION AND VALUATION 85 opportunity for expansion interstate because of more banks being offered for take over and fewer competitors for them. Additionally. profitability. States presently do not allow banks in other states to enter de novo and do not allow branches to be operated in a state unless explicitly provided by state law. and increased competition via entry by acquisition. the data presented in this section suggest that there are not many larger banking companies in such an enviable position.

banks must be free to open offices and banks de novo within any state. Tschinkel. This is expected to reduce the losses of $665 million taken in 1989. announced a severe retrenchment by a cut of 5000 staff or 12 percent of its work force by December 1990. Tschinkel. more profitable banking companies to become active in interstate expansion. C1). further benefits from interstate banking will be long in coming. 7. 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. The financial condition of the banking system may not hinge on whether there is more liberal interstate banking legislation. such as the limitation of a banking company gaining excessive share of any single local market or state. increased from 54. 4.1990. 48-50) for a listing of the approved activities and those prohibited under section 4(c)(8). . It is important to note that statewide concentration. 49. and Whitehead (1989). 1. Chase Manhattan.9 percent in 1987. Some of this legislation prescribes that there be safeguards.86 THE CHANGING MARKET IN FINANCIAL SERVICES law in most states has made the prohibitions of the Douglas Amendment to the Bank Holding Company Act of 1956 virtually obsolete.S. the second-largest U. as measured by the average five- firm ratio.7 percent in 1976 to 60. On September 21. and Whitehead (1989. These results support the evidence for local market concentration increases. See Goldberg and Hanweck (1988) for a discussion of the legal framework of inter- state banking and a detailed description of various exceptions to. much like most other businesses (including bank loan production offices and finance company subsidiaries of bank holding companies). In order to get the greatest competitive benefit from interstate banking. banking company. 5. (See the Washington Post. See King. Without an alternative means of expansion that requires less resources than under the present system and allows smaller. the expansion of powers granted S&Ls in the 1980s allowed them powers similar to banks until they were restricted in 1989 under FIRREA. and means of. circumventing interstate banking laws. Clair and Tucker (1989). Nonbank office expan- sion by bank holding companies will be discussed below. These issues need to be considered and dealt with within the broader spectrum of antitrust legislation and enforcement. 1990. 2. Such legislation has been introduced by Congress in the past and debated to some degree. but as the profitability and valuation picture painted in this study suggests. 3. King. 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. banking companies will not have the resources to maintain the interstate expansion pace of the past eight years. However. 6. 14. Table 2. September 22.

and Robert F. one of the reasons given for the Continentallllinois failure in 1984 was that it was a large unit bank located in a state that did not permit branching or. The use of the potential competition doctrine in its two forms has not been successful in the courts. v. See Amel (1988) for a discussion of the motivations for some of the interstate bank- ing laws. September 1986. BANK EXPANSION AND VALUATION 87 8. 12. Berger. G. "Competitive Effects of Interstate Banking: The Impact on Bank Acquisition Markets. See Jensen and Ruback (1983) and Jarrell. . 1988. 1990. "Interstate Banking: Impacts on the Payments System. Benston. "Scale Economies in Banking: A Restructuring and Reassessment. Brickley and Netter (1988) for summaries of this literature.F. References Adkisson. Multibank Holding Company Expansion. F. Schapperle. Washington.. at the time. Federal Reserve Bank of Chicago. Anthony Saunders. so that it was unable to acquire stable core deposits and needed to rely heavily on volatile purchased funds. Humphrey. Gerald A. multibank holding companies.J. and the Federal Reserve Board virtually abandoned the doctrine after the Mercantile Texas case decision in 1981. 9.. BIF.. "Trends in Banking Structure since the Mid-1970s. and D. "The Effects of DIDMCA on Bank Stockholders' Returns and Risk. For example.C. Arne!.. and Banking XIII (November Part I): 435-56. has an estimated value of $11 billion." Federal Reserve Bulletin. Board of Governors of the Federal Reserve System. A. (Mercantile Texas Corp. D. Joseph. Credit. A failure of the size of the Bank Of New England. D. A. D. Fraser.." Presented at the Bank Structure Conference. Ahrony." Journal of Monetary Economics 17 (May) 363-377. and David B. and Michael Jacowski." Research Papers in Banking and Financial Economics. 1268 (5th Cir. At the present time there is considerable concern in the ability of the FDIC to meet a single failure of any of the large New England banks. "The Effects of a Shift In Monetary Policy Regime of the Profitability and Risk of Commercial Banks. 11. 638 F. Board of Governors. The Bank Insurance Fund." Journal of Money. Rose (1985) addresses the problems of antitrust enforcement in market extension mergers and acquisitions. Reim. Board of Governors of the Federal Reserve System (March): 120-133. 1989. and J.." Board of Governors of the Federal Reserve System." Journal of Banking and Finance 12: 317-331." Journal of Bank Research (Spring): 8-15. with deposits of $9 billion as of year-end 1989. 1985.INTERSTATE BANKING. and Ithzak Swary. "The Effects of Barriers to Entry on Shareholder Interstate Wealth: Implications for Banking. . Randolph P. 1982. Also see Rhoades (1985). (May): Mimeograph. Hanweck. #90. 1981». Beatty. after accounting for estimated FDIC losses of about $3 billion in 1990.2d 1255. D. (May): Mimeograph. "State Laws Affecting Commercial Bank Branching. 1988. 1986. John F. Humphrey. and Interstate Banking. Amel. 10. may result in a cost to the FDIC of $2-$5 billion. Frodin.

Douglas D. "Retail Bank Deposits as Quasi-Fixed Factors of Production..C. 23. Geographic Restrictions on Commercial Banking in the United States." American Economic Review (June): 527-536. Millon. S. Federal Reserve Bank of Dallas (November): 1-20. Scope. L. Stover. A. Clair.H. Federal Reserve Bank of Kansas City (September/October): 16-33. J. Dunham. Clark." 20 December Journal of Monetary Economics 20: 501-520. 1985." Rand Journal of Economics 16 (Summer): 167-183. and Roger D. Credit. Frieder. 1984. 1989. 1989. 1985b. 1988. Demsetz. "An Examination of Stock Market Reactions to Interstate Bank Mergers. Desai. Cheng. 1986. "Bank Holding Company Acquisitions." Journal of Financial Research 8 (Summer): 145-156. De. Curry. Mark J. Syron." Journal of Law and Economics 16 (April): 1-9. A and T. "Interstate Banking and the Outflow of Local Funds. Humphrey. G.. . Robert T. "Economics of Scale and Scope at Depository Financial Institutions: A Review of the Literature. David C.AI987. "The Price-Concentration Relationship in Banking.88 THE CHANGING MARKET IN FINANCIAL SERVICES Berger. and John T. . Wall. Federal Reserve Bank of Boston (May/June): 11-28." Economic Review." Review of Research in Banking and Finance (Fall): 1-17. Flannery. 1973. Evanoff. Policy." Journal of Money. 1986." Journal of Accounting Research 23 (Spring): 370-383." Economic Review." Antitrust Bulletin XXIX (Winter). and Diane Fortier. Federal Reserve Bank of Boston (March/April): 7-19. "The Impact of Geographic Expansion in Banking: Some Axioms to Grind. "Measuring the Effects of Regulation with Stock Price Data. Department of the Treasury. 1982.H. Including the attached compendium of research prepared by the federal banking agencies. Berger. Washington. Stockholder Returns and Regulatory Uncertainty. Tucker. and Product Mix Economies. Gup. and M. and Misconceptions. and Banking XXI 4 (November): 524-536. Timothy J. Jeffrey A 1988.. "Interstate Banking and the Federal Reserve: A Historical Perspective." Finance and Economics Discussion Series. Benton E. The Report of the President. 1984. 1988. Hanweck. Binder. Federal Reserve Bank of Chicago (May/June): 24-38. and D. "Financial Determinants of Bank Takeovers. 759-74. 1987. pp. and Paula K." Economic Perspectives. and Public Policy." New England Economic Review. "On the Use of the Multivariate Regression Model in Event Studies. Jan 1981. Hannan. 1985a. D. Rose." (July): Mimeograph. Federal Reserve Board (April).. Harold. and Larry D. "Diversification and Barriers to Entry: Some Evidence from Banking. No. Constance R. Market Rivalry. "Interstate Banking: The Drive to Consolidate. "Industry Structure. "Competitive Viability in Banking: Scale. Anand S." New England Economic Review. "Interstate Banking: Landscape. --and Richard F.

Rhoades. King.. G. Frank. Credit." Journal of Economics and Business XXXVII (December): 391-402. "Interstate Banking Game Plans: Implications for the Midwest. N." Journal of Money. 1990. "New Data on the Performance of Nonbank Subsidiaries of Bank Holding Companies.A. Savage.. and Banking 16 (November Part 2): 617-644." Journal of Financial Economics 11 (April): 5-50. "What We Can Expect From Interstate Banking. and Stephen G. 1987. and S. 1984. Federal Reserve Bank of Atlanta. and David D. and S. G." Economic Perspectives. . Credit." Journal of Money. Phillis. and Banking XIX.. R. Tschinkel. "Differential Impact on Bank Valuation of Interstate Banking Law Changes.C. Liang. G. Federal Reserve Bank of Chicago. "Returns to Acquirers and Competition in the Acquisition Market: The Case of Banking. "Predicting Takeover Targets. Goldberg. Jensen. 1987. and Richard S. 1979." Midland Corporate Finance Journal (Summer): 6-32. Netter. 1982." Journal of Money. Rhoades. Pavel. 1986. 1. and Peggy Weir. Palepu. Stephen A. and T. Cambridge University Press. Federal Reserve Bulletin (March) Board of Governors of the Federal Reserve System. Sheila L. (February) and Summary. G. "Dominant Firms. MA." Staff Study 159. Whitehead. Loretta J. Christopher M. 4 (November): 538-49. Golembe Associates." Department of Finance. 'Deep Pockets'. D. 1988. Maddala.INTERSTATE BANKING.. "Multiple Market Contact Between Savings and Loans.. Qualitative Variables in Econometrics." Journal of Political Economy 95 (April): 355-370. and J. 1988. Krishna G. 1989." Journal of Accounting and Economics (March): 3-35. Rhoades.A. "Interstate Banking Developments in the 1980s. and Gerald A. and Banking XLIII (May): 152-66. D. 1988. B. Hunter.A. Hanweck. and C. BANK EXPANSION AND VALUATION 89 Gilbert. 1986. "The Takeover Controversy: Analysis and Evidence. Hanweck. William C. "The Market for Corporate Control: The Empirical Evidence Since 1980. "Structure-Performance Studies in Banking: An .A. 1983. For the Association of Bank Holding Companies." Journal of Economic Perspectives 2: 49-68. A Study of Interstate Banking by Bank Holding Companies. Organizational Form. Timme. (May).. and the Structure of Bank Production. "Geographic Diversification and Risk in Banking. Sugrue." Journal of Economics and Business 40: 271-84. .S." Economic Review. 1984. M. Ruback. (September): Mimeograph.. Jarrell. . "Bank Market Structure and Competition: A Survey. Credit. Lawrence G. Washington. George Mason University. 1990. 1986. and Local Market Competition in Banking. "The Market For Corporate Control: The Scientific Evidence. Mester. James. Brickley. 1986. Cambridge. and D. "Technical Change. (May/June): 32-5l. Hanweck. Alton. 1986." Journal of Banking and Finance 12 (March): 51- 67..

1987. Conn. Theil." Loyola of Los Angeles Law Review 18(4): 1115-1164." Research Paper #8612. Greenwood Press. "The Consequences of Interstate Banking Deregulation for Competition. RJ. H. Stigler." Staff Study. March Board of Governors of the Federal Reserve System. "Bank Merger Pricing Premiums and Interstate Bidding. "Can Small Banks Compete?" Bankers Magazine (Jan. --. 1986.S. the Structure of Service Markets. "A Note on the Resource Allocation Efficiency of MBHCs Versus Independent Banks." Staff Study 102. "National and Local Market Banking Concentration in an Era of Interstate Banking. Simonson... John Wiley & Sons. Las Vegas. Syron. 1971. and the Attractiveness of Banking Markets for New Entry.G. John T. October 14-17.. New York. Issues in Bank Regulation (Spring): 29-36.): 59-65. 1987. "Economies of Superscale and Interstate Expansion." Paper Presented at the Financial Management Association Meetings.1989. (4 November): 289-312. Shaffer." Quarterly Review of Economics and Business XXIII (Summer): 112-117. Edmond. 1985." Bell Journal of Economics and Management Science 2 (Spring): 3-21. "Multimarket Interdependence and Per- formance In Banking: Two Tests.. Federal Reserve Bank of Boston. (MarchI April): 5-17. "Geographic Expansion of Banks and Changes in Banking Structure.. Principles of Econometrics.. Inc. 1983. . 1985. Federal Reserve Bank of New York. . 1987. 1986. D." Antitrust Bulletin (Fall): 729-743. 1990. 1987. and D.. "Determinants of Premiums Paid in Bank Acquisitions." Federal Reserve Bulletin (February): 79-92. .lFeb. and the Performance of Interstate Banking Financial-Service Firms. Nevada. .. --. 1971. S.. August.. 119. . 1987." Antitrust Bulletin (Winter): 975-995. Westport." Atlantic Economic Journal." Paper presented at the Bank Structure Conference. "Interstate Banking and Small Business Finance: Implications from Available Evidence.1979. 1985. Rose. Rose. Quorum Books. Federal Reserve Bank of Chicago. .J. XV (March): 20-30.." Journal of Economics and Business XXXIX. and Arnold A." American Journal of Small Business (Fall): 23-39.T. Board of Governors of the Federal Reserve System.. Potential Competition. "Interstate Banking. . Savage. "The Theory of Economic Regulation.. Heggestad. and D. (May). "Interstate Banking Developments. The Interstate Banking Revolution: Benefits and Costs for Business and Consumers.. ." New England Economic Review. Richard F.. 1981.. P. "The 'New England Experiment' in Interstate Banking. . Vol 164 Rogowski..90 THE CHANGING MARKET IN FINANCIAL SERVICES Updated Summary and Evaluation. 1985. "The Impact of Mergers in Banking: Evidence from a Nationwide Sample of Federally Chartered Banks. G. 1984. "Interstate Banking and Product Line Expansion: Implications from Available Evidence.

. Zellner. A. lW. Whitehead. 1987. Thompson. A Guide to Interstate Banking. and K. Scanlon. 1983. R. "Conditioning the Return-Generating Process on Firm-Specific Events: A Discussion of Event Study Methods." Journal of the American Statistical Association 57 (June): 348-368." Journal of Financial and Quantitative Analysis 20 (June): 151-168. Federal Reserve Bank of Atlanta. "Interstate Bank Mergers: The Early Evidence. D." Journal of Financial Research X (Winter): 305-31l.INTERSTATE BANKING. 1985.P. 1962. BANK EXPANSION AND VALUATION 91 Trifts. . "An Efficient Method of Estimating Seemingly Unrelated Regression and Tests for Aggregation Bias.

ten. fifty and one hundred domestic commer- cial banking organizations have all risen since 1985. espe- cially in urban areas. Most of the concentration increases appear to have occurred in the 93 . and addresses a number of emerging trends in local. state. Similarly. Amel and Jacowski calculate that thirty-three states experienced a rise in their mean three-firm urban banking concentration ratios from 1981 to 1986. the aggregate shares of all but one of these groups (the top five banks) have been increasing nationwide since 1970. COMMENTARY Peter S. The article is a combination of thoughtful normative economics and positive economic analysis. Moreover. One of these potentially disturbing trends is an apparent increase in both local and national banking concentration. with only seventeen states experiencing a decline in concentration. the shares of nationwide banking assets held by the top five. Hanweck presents an interesting and thoughtful article on the recent developments in interstate banking and their effects on bank stock values and bank market structure. Rose Gerald A. Hanweck cites a study by Amel and Jacowski (1989) that points to an apparent reversal of the historical trend toward lower local banking market concentration. and national market structure that are at least partly the result of recent interstate banking legislation. twenty-five.

and insurance companies to security dealers. For example. bank concen- tration ratios have always averaged highest in statewide branching states. or district boundaries. in large measure. Historically. those allowing limited branching within city. finance companies. particularly for those financial services that appear to be uniquely important among the very largest banks. By the end of the period. credit -card firms (including most recently AT&T). Therefore. which in selected states has become a potent competitive force. At the beginning of this period the fifty states were roughly equally divided between states sanctioning statewide branching activity. and not primarily to interstate bank acquisitions. and the number of strict unit banking states had dwindled to less than five. and unit banking states where full-service branch offices are prohibited.94 THE CHANGING MARKET IN FINANCIAL SERVICES decade of the 1980s-the period over which most of the interstate bank- ing legislation has appeared. the shift in state banking rules applying to intrastate expansion. I'm also not as eager to discard the research evidence on economies of superscale as Hanweck may be. . but also in consumer and small business credit. you discover quickly that the concentration figures are for domestic commercial banking firms only. the trend toward national banking concentration that Hanweck cites may reflect. There may be another explanation besides interstate banking for the apparent trend toward greater banking concentration. the majority of states had adopted statewide branching regimes. but I'm not yet convinced that this trend is happening to the degree that his analysis implies or primarily because of interstate banking. Nor is there any adjustment for nonbank financial- service competitors. not only in corporate loans and credit guarantees. if you look at the article's numbers and citations (particularly pages 62-65 and in table 2-6). ranging from thrifts. I share Hanweck's concern for this apparent rise in banking market concentration. 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. The 1970s and 1980s represent a period of sweeping changes in state branching laws. either de novo or by acquisition. There is no adjustment for foreign-bank penetration of United States markets. followed by states with limited branching rules. and lowest in unit banking states. county. 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. Perhaps. however.

with our imperfect models and statistical tools. have captured the interstate initiative with at least forty-seven states now permitting some form of interstate expansion. in fact. corporate security underwriting.INTERSTATE BANKING. correspondent banking. interstate full-service expansion is limited essentially to acquisitions. If there is a significant con- centration trend overall in financial services. BANK EXPANSION AND VALUATION 95 including standby credit guarantees. I do share Hanweck's concern with the path the federal government and the states have recently chosen to deregulate geographic expansion in banking. If. His article contains a very competent and well-written literature re- view of the interstate field and it blends in rather neatly the data on recent changes in bank returns and stock values with research evidence from the period of most intense interstate banking activity. As he correctly notes. swap contracts. However. There is. Hanweck is right in his assessment of economies-of-scale studies. due to the greater costs involved. The author begins with the observation that: . simply have not been able to competently measure. giving smaller banks a much better chance at full participation in the interstate movement. for example. and other risk-hedging products- to name just a few. the risk of dominant banking firms stretching across the nation's landscape to confront each other in market after market. Neither de novo charters nor branching across states lines are permitted in most instances and the latter is probably a far less costly vehicle for expansion than acquisitions. practicing mutual forbearance and limiting competitive rivalry among themselves. as Hanweck notes. and may not generate the optimal market structure needed to spur competition. we would perhaps benefit more in a com- petitive sense if we permitted de novo entry and full-service branching across state lines. devolved upon the states because Congress has not yet moved to amend or abrogate the McFadden and Glass-Steagall Acts or the Douglas Amendment to the Bank Holding Company Act which prohibit full-service interstate expansion. that trend may have an operating efficiency basis that we. What is somewhat disappointing about the article lies in what it does not do with its central hypothesis. geographic financial-services deregulation in the United States has. for their part. for all intents and purposes. The states. and not just in traditional banking. Scale economies research to date has several glaring holes and one of these centers upon the often enormous differences in large bank and small bank product lines. Such a restriction tends to limit interstate expansion to the largest and most capital-endowed banking firms.

not all the evidence that we currently have is totally consistent with the conclusions reached. and operating efficiency).. on average. loan- loss risk exposure. as plausible as Hanweck's hypotheses are. Hanweck..96 THE CHANGING MARKET IN FINANCIAL SERVICES Recent changes in state laws with respect to interstate banking should be expected to affect the valuation of commercial banks . and Sugrue (1990). There is. 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). 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. 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. It is not intuitively obvious that the announcement of an interstate acquisition . (and) it is likely that such effects should vary with the location. 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. Indeed. 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. both interstate acquirers and their targets have tended to perform less well than comparably sized firms.. Moreover. A recent study by this au- thor (May 1990) finds that. the argument that interstate bank expansion is a selective process. this paper provides us with one of the reasons for this phenomenon: the peculiar nature of interstate banking laws passed during the 1980s. only certain kinds of banking firms are likely to par- ticipate in it. 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. nature of banks. 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. for example.. and the type of state law change . As this paper reports." These hypotheses may be true (and there is a substantial body of financial theory with which they appear to be consistent). 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.

will yield value effects that could be either positive or negative- again. this study suggested that individual firm profitability and risk are important in explaining a substantial proportion of the variability in observed bank merger premiums. This. the article on interstate banking by Gerald Hanweck makes a . It may be that a change in state banking law that significantly alters both the number of potential acquisition targets. There is a branch of auction and bidding theory that argues that the winning valuation of an asset subject to sale at auction depends not only on how many bidders (acquirers) there are. Or.INTERSTATE BANKING. depending upon the relative change in potential acquirers versus acquisition targets. The study drew upon the developing theory of auctions and bidding to estimate the impact on bank merger pricing from both structural factors (such as changes in state banking laws) and individual bank risk-return measures. I certainly agree with Hanweck's overall observation that the character of a state's interstate banking law can have a significant impact on bank asset valuation. but also upon the supply of alternative assets available for auction. But there was something else-evidence of a statisti- cally significant negative relationship between interstate banking legisla- tion and merger premiums posted. About a year ago this author conducted a study of the premiums paid by about five hundred acquiring banks involved in acquisitions and merg- ers over the 1978 to 1987 period (February 1990). but. too. The other point about that earlier auction and bidding study that seems relevant here is that the merger sample used in that study contained a substantial number of banks from states suffering severe economic dislo- cations in the wake of the energy. and real estate crises of the mid- 1980s. More research is clearly needed here. As expected. Perhaps capital market investors have seen too many acquisitions by major banking firms go awry to believe that positive returns will automatically follow the passage of a new interstate banking bill. Many of these states passed reciprocal interstate banking laws at the time when the market values of their banks were falling sharply. perhaps these investors have read an excellent article by Goldberg and Hanweck (1988) that does not emerge with optimistic findings regarding the performance of grandfathered interstate banking firms. farm. Overall. as well as the number of potential acquirers. BANK EXPANSION AND VALUATION 97 of a below average performing banking firm by another below average performer will generate positive returns. The condition of each state's economy may well be an important valuation factor that could also account for the estimated negative link between merger premiums and the presence of new interstate banking legislation. is an area where more research is clearly warranted. the direction of that impact may be ambivalent.

Texas A&M University (February). Hanweck. 1990. "Bidding Theory and Bank Merger Premiums: The Impact of Structural and Regulatory Factors.A. Conn. and the Performance of Interstate Financial- Service Firms. Risks. .1990. Federal Reserve Bank of St. Gerald A. ." Unpublished Paper. 1989. and Tradeoffs for Bankers and Consumers. (October 18) Rose. Bank Expansion and Valuation. "The Consequences of Interstate Banking Deregulation for Competition.98 THE CHANGING MARKET IN FINANCIAL SERVICES worthwhile contribution to the literature and especially to discussions of appropriate public policy in the banking structure field. 1990. Federal Reserve Bank of Chicago (May). Westport.: Quorum Books. Department of Finance. It forcefully re- minds us that the structure of banking in the United States is changing rapidly without benefit of a cohesive national policy that places the welfare of the consuming public first and without full awareness of the consequences for competition and efficiency that only careful research can provide. The Interstate Banking Revolution: Benefits.G. Hanweck. --. "What Can We Expect from Interstate Banking?" Journal of Banking and Finance 12: 51-67." Proceedings of the Conference on Bank Structure and Competition.." Proceedings of the Fifteenth Annual Economic Policy Conference. 1988. the Structure of Service Markets. References Goldberg. and G. Peter S.. "Interstate Banking. L. Louis.

As a result. 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. 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. the develop- ment of a national primary market for adjustable-rate mortgages. 99 . The three major changes are summarized before a detailed analy- sis is presented. These changes and their im- pacts on various financial industries and homebuyers are the subjects of this article. coupons on new issue FRMs now swiftly adjust to changes in other capital market rates. two de- velopments that switched thrifts from low cost funders of mortgages to high cost funders. and the decimation of the savings and loan industry. Integration was stimulated in the 1980s by the deregulation of deposit rate ceilings and the erosion of thrift tax subsidies. Hendershott Three major changes occurred during the 1980s in the market for home mortgage credit: the securitization of fixed-rate mortgages. 3 THE MARKET FOR HOME MORTGAGE CREDIT: RECENT CHANGES AND FUTURE PROSPECTS Patrie H. and mortgage funds are readily available at all times and places at going market rates.

the industry grew at a relatively slow rate. and by 1989. the S&L problems in the early 1980s substantially raised conventional FRM rates. and ARMs with loose per adjustment period and life-of-Ioan rate caps were not permitted until 1981. Between the end of 1984 and the end of 1990. real estate loans in 1983 and 1984. The widespread existence of ARMs reduced the previous cross subsidy from more mobile to less mobile households and increased the ability of many households to qualify for larger loans. when interest rates skyrocketed. Two years later Ginnie Mae began guaranteeing mortgage-backed pass-through securities. The first two describe the development of mortgage securitization and its impact on mortgage rates. Government National Mortgage Association Securities (GNMAs) representing shares in pools of FHA/V A loans. and then from rapid S&L growth and accelerated investment in nontraditional. and expost nonprofitable. the industry has been in freefall. Section five speculates on future prospects for the market for home mortgage credit. Recovery and Enforcement Act of 1989 (FIRREA). ARMs accounted for two- fifths of the total conventional loan volume originated by all lenders over the 1984 to 1989 period. and since passage of The Financial Institutions Reform. Investors in pass-throughs receive a pro rata share of the payments. In response to this authorization.100 THE CHANGING MARKET IN FINANCIAL SERVICES Nationwide authority to originate adjustable-rate mortgages (ARMs) was not given to federally chartered thrifts until 1979. the Government National Mortgage Association (Ginnie Mae) was formed within the United States Department of Housing and Urban Development to administer government mortgage support programs. As I will discuss. The remainder of this article is divided into six sections. The Development of Mortgage Pass-Through Securities In 1968. The large losses of the savings and loan (S&L) industry in the 1980s stemmed first from badly mismatched asset and liability portfolios in the late 1970s and early 1980s. but the recent decline in the S&L home mortgage share has not raised FRM rates. at least for the 90 percent of FRMs that can be sold to Fannie Mae and Freddie Mac. and Section six sum- marizes the paper. both scheduled and early (in the . the S&L share of home mortgages outstanding has declined from 43 percent (also the average share in the second half of the 1970s) to 25 percent. half of the thrift home loan portfolio was in ARMs. In the 1985 to 1988 period. The next two document the growth in the ARM market and the decline in the S&L industry.

on the underlying mortgages. The size of the FRANA and conventional loan volume that can be securitized by Ginnie Mae and the sponsored agencies (Fannie and Freddie) is restricted by limits on the dollar value of loans that can be pooled into the various pass-through securities. The 1990 limit varies regionally from $67. two-thirds of FRANA originations went into GNMA pools. The markets for fixed rate FRANA and conforming conventional loan pass-throughs developed at different rates. limits were raised in numerous regions many times during the 1980s. investors in GNMAs are guaranteed the full timely payment of principal and interest. The lower limit was not changed in the 1980s. While investors in whole FRANA loans are insured by FRA or VA against loss of principal and interest. and it has an implied guarantee comparable to Freddie Mac's.THE MARKET FOR HOME MORTGAGE CREDIT 101 event of prepayment or default). the Freddie Mac guarantee adds more value to the underlying mortgage than does the Ginnie Mae guarantee. but the upper limit. changes annually with a house price index but does not vary regionally. even though Freddie Mac doesn't have a full faith and credit Federal guarantee.500 to $124. In 1987. The dollar limit on GNMAs follows from the limit on the underlying FRA and VA loans. By the second half of the 1970s. the Federal Rome Loan Mortgage Corporation (Freddie Mac) was chartered to spur the development of a secondary market for con- ventional mortgages. and this percentage has been fairly constant. buying mortgages and issuing its own debt since 1938. was increased in both 1988 and 1989. the "conforming" limit. The dollar limit on conventional loans that Fannie Mae and Freddie Mac can purchase. The 1989 limit was $187.600. up 63 percent since 1985 (the limit was virtually unchanged in 1990). Table 3-1 presents data on the growth in the securitization of fixed-rate FRANA loans. Freddie Mac introduced the first conventional mort- gage pass-through in 1971. the Mortgage Participation Certificate (PC). by the early 1980s four-fifths did. Fannie has intermediated in the more traditional sense.000 between 1980 and 1987. Moreover. the Federal National Mortgage Association (Fannie Mae) initiated a conventional mortgage-backed security (MBS) program similar to Freddie Mac's PC program. In 1970.750. The importance of securitization to the new origination market is measured as the ratio of GNMA issues backed by single family loans to total originations of these loans (Ginnie Mae is prohibited from securitizing FRANAs over eighteen months old). In 1981.! . which was $90. over 90 percent of home mortgage loans (80 percent of dollar volume) were eligible for pooling by the agencies. in contrast to FRANA loans. Because the underlying conventional mortgage is not itself fully insured. and since 1982 all FRANA have gone into GNMAs.

01 . to almost 25 percent in the 1982 to 1985 period.6 . Total single family conventional originations are multiplied by an esti- mate of the fraction that had a fixed rate.17 .77 .) Securitized Originations 1971-73 15.102 THE CHANGING MARKET IN FINANCIAL SERVICES Table 3-1.43 . the last column is an estimate of the percentage of new FRM conforming loan volume that is sold directly to Fannie and Freddie.6 2. Of course.71" • Mortgage banker issues are likely understated. and.28" . Thus originations and the mortgage banker share are too low. The Growth in the Securitization of FHAIVA 1-4 Family Loans. 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. FHLMC.81 1983-86 55.78 1987-89 54.7 . Mortgage bankers have tended to dominate the FHA/V A market. those backed by seasoned FRMs. 1 2 3 = 2/1 4 FHA/VA GNMA Share of Mortgage Originations Issues Originations Banker Share of ($ bi/. 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. the agencies can only participate in the conforming end of the market. securitized by Freddie Mac and Fannie Mae. new ARMs and multifamily loans.5 5. Source: Hendershott and Van Order (1989). Assuming this end to be 75 percent of total FRM volume.3 17. the pass-through market for conforming conventional loans developed less rapidly. As the data in table 3-2 indicate. and to over 50 percent since 1986.62 . The lower part puts the agency issues backed by new FRMs into a market perspective. and the share of originations securitized is too high.0 55.6 . The upper part of table 3-2 lists total agency pass-through issues.5 16. and the result is divided by agency issues backed by new FRMs to obtain the agency share of the total FRM market. .70 1974-75 13. in fact.78 1980-82 21.75 1976-79 28. including 69 percent in 1989.3 1.6 1. As can be seen. updated from DataBase. new FRMs.9" 70.8 . by subtraction. Secondary Mort- gage Markets.) ($ bi/. this estimate rose from 4 percent in the 1977 to 1981 period.

8 60.0 0.47 159.2 17. Percentage of New 1-4 Family Conventional Originations Securitized.00 125.2 0.65 50.6 4.1 29. New Issues FRMs andFHAIVAs FRMs 1977-1981 4.7 3.3 41 55 1988 338.1 38 51 1989 296.55 112.2 21. Pass-Throughs Issued by FHLMC and FNMA ($ bU.Throughs Backed-By 1 2 3 4=1-2-3 Total Seasoned New ARMs.7 36. .77 278.7 1986 160.1 11.) Securitized Securitized 1977-1981 125.6 0.0 0.0 17.0 1982 38.66 248. The Agencies Increased Role in the Conforming FRM Market.7 10.1 31.2 16.1 43 57 1987 376.1 0.8 0.5 29.4 1989 145.3 25.6 0.2 0.0 13 17 1985 204.8 9.1 1984 32.Table 3-2a.7 119.5 0.2 Table 3-2b.0 1.0 3 4 1982 77.7 1987 138.2 97.8 14.7 1988 94. Multis.69 106.5 52 69 Source: Hendershott (1990a.) Pass.2 28.1 1.63 186.7 4.) Rate ($ bil.7 10.75 % New Total Fraction FRM % New Conforming Originations Fixed Originations FRMs FRMS ($ bil.5 28 37 1986 362. Table 4).4 13 17 1984 176.3 100.46 81.6 19 25 1983 154. 5 6 7=5x6 8=417 9 = 8/0.5 5.4 1983 33.8 1985 62.

1980). Thus. 1980.104 THE CHANGING MARKET IN FINANCIAL SERVICES accounting for 70 to 80 percent of originations (see table 3-1) versus only 7 percent of conventional originations prior to 1982 and 15 percent since. In the 1960s when the transfer fraction was 60 percent. depository institutions have dominated conventional originations (80 percent to 90 percent). between 1969 and 1979. In effect. savings and loans would have accepted a three-quarters percentage point lower pretax return on tax preferred housing-related assets than on comparable non-preferred assets. or equity issues) encouraged investment in residential mortgages prior to the 1980s. bonds. these investments were especially profitable to thrifts. the share of savings and loan total assets in home mortgages and agency se- . mortgage bankers quickly took advantage of the opportunity. Most strikingly. In contrast. Appendix). thereby avoiding taxes. and the statutory fraction of income that can be transferred to reserves. The incentive provided by the extraordinary loan-loss provisions for investment in residential mortgages depends on the expected level of thrift taxable profits over the expected life of the investment (with no profits now or in the future. By 1979. Thus. the incentive is zero). By the early 1980s. Two major factors have driven the increase in conventional loan securitization in the 1980s. the fraction was gradually reduced to 40 percent. and they have tended to keep their originations as portfolio investments. Between 1962 and 1969. thrifts were allowed to transfer large portions of their before-tax income to reserves. the income tax rate. the incentive was substantial in the 1960s and 1970s (Hendershott and Villani. The tax preference was the ability of thrifts to compute loan-loss reserves that far exceeded a reasonable provision for normal losses. owing to special tax advantages. an improved selling method alone was not sufficient to stimulate the conven- tional pass-through market. thrifts maintained their share of mort- gage originations but reduced their relative investment in home mort- gages (have sold some of the originated mortgages). the Tax Reform Act of 1986 lowered the fraction to 8 percent. Moreover. when the transfer fraction was down to 40 percent. Assuming a 1 percent net pretax return on assets. they would have accepted a half percentage point less. Portfolio restrictions on savings and loans (no corporate loans. First. and they sell virtually all their originations to other investors. the transfer was limited to 60 percent of taxable income. virtually all of mortgage banker originations were sold into Ginnie Mae securities or to Fannie Mae and Freddie Mac (some conventionals were sold to Fannie Mae for its portfolio). as long as thrifts invested a large fraction of their assets in housing related loans or liquid assets (Hendershott and Villani. when an improved method for selling mortgages became available.

The reduced profitability eroded the tax incentives for residential mortgage investment. while the expansion of powers encouraged thrifts to invest more widely. Second. This portfolio shift reflected the reduced profitability of savings and loans. some loans were simply swapped for pass-throughs. and the pass-throughs were retained in portfolio and "repoed" or used to increase advances. etc. and the expansion of savings and loan asset powers. credit losses. The securitization of conventional conforming ARMs by Fannie Mae and Freddie Mac is less prevalent. that is. though. the slicing up of these securities into different maturity tranches (CMOs or REMICs) or into interest or principal strips (lOs and POs) . The first resecuritization (issuance of a multiclass mortgage security) was a Freddie Mac CMO issue in 1983. 1988). Finally. Fully private mortgage securitizers cannot compete with Fannie Mae and Freddie Mac for two reasons: (1) The agencies do not have to maintain as much capital as fully private institutions. and in the 1980s these became cheaper marginal sources of funds than deposits for many savings and loans. issues were dominated by private sector participants. The latter stems from lower explicit costs (exemption from SEC requirements. Issues have risen from $16 billion in 1986 to $100 billion in 1989. Recent years have also marked a surge in the resecuritization of fixed- rate mortgage-backed securities. It appears that only 2 percent to 3 percent were securitized in 1984 to 1985 and that the percentage is still only 10 percent to 12 percent. Fannie Mae and Freddie Mac accounted for only 2 percent of issues in 1986 to 1987. By 1989. The greater securitization of fixed-rate than adjustable-rate mortgages likely reflects both the greater standardization of FRMs than ARMs and the greater desire of originators to hold ARMs than FRMs in their portfolio. state and local taxation. and (2) The agencies' costs of securitizing are lower. some investment banks and large thrifts also securitize home mortgages. Throughout the 1986 to 1989 period.THE MARKET FOR HOME MORTGAGE CREDIT 105 curities (largely Fannie and Freddie pass-throughs) fell from 72 percent to 59 percent during the 1982 to 1984 period.) and from the economies- . have been the collateral for 96 percent of multiclass issues. savings and loans increased such debt by over $150 billion. pass-through securities. first due to high interest rates and a maturity mismatch and then due to disinflation. Initially. but these institutions largely (possibly exclusively) limit themselves to nonconforming or jumbo loans and they likely securitize only 10 percent to 20 percent of the market (Woodward. these agencies accounted for 83 percent of issues. a Freddie Mac PC or Fannie Mae MBS is excellent collateral for borrowing through FHLB advances and security repurchase agreements. During the 1984 to 1988 period. rather than whole mortgage loans.

such as the six-month and seven-year Treasury rates and the volatility of these rates (Hendershott and Van Order. Moreover. In con- trast. Empirical evidence relating to each of these impacts is discussed below. Moreover. 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. the response to changes in those variables is predictable and fast. Moreover. Hence. most researchers focused on things peculiar to the thrift industry. The expected impact of securitization on the general level of mortgage rates is more complicated. His results are reproduced and extended to include 1988 and 1989 in table 3-3. 1989).5 and was never statistically different from zero. such as deposit rates and deposit flows. The Impact of Securitization on Mortgage Coupons A clear implication of mortgage securitization is that mortgage yields should be more closely connected to capital market rates with securitization than without.58 and was always statistically positive. it would be a lagged one. Timing of Conventional Mortgage Rate Adjustment to Capital Market Rates 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. the correlation of the changes ranged from -0. which were heavily regulated and also tax-advantaged to invest in mortgages.5 to +0. the correlation was never less than 0. Prior to 1982. one would expect to obtain a poor fit. If one regressed actual mortgage rates during such a: period on fictional mortgage rates predicted by the perfect market model. and connections of mortgage and capital markets were tenuous and gradual.90. to the extent that the predicted rate had any effect. though. After 1981. twenty years ago mortgage lending was tied to the thrifts. A potential problem with Roth's analysis is that the mortgage rate .106 THE CHANGING MARKET IN FINANCIAL SERVICES of-scale achieved as a result of being the low-cost securitizer in a large part of the market. rather than general capital market conditions as determinants of mortgage rates. in the last three years the correla- tion has averaged 0.

and in some periods the value of the call-premium may have changed markedly. The analysis consisted of two parts. they regressed conventional commit- ment mortgage coupon rates on current and past values of the estimated perfect-market coupon rate taken from the GNMA equation.81** 1984 0. The price equation was estimated on weekly GNMA price and coupon data from the January 1981 to July 1988 period.16 1977 -0.42 1982 0.80** 1983 0.92** * Correlations are between month-to-month changes in the FHLMC commitment rate and the 10-year Treasury rate. possibly disguising a close relationship between the non-call component of the mortgage coupon and the Treasury coupon.58** 1987 0. incorporates a call-premium while the Treasury rate does not. and solved the equation for the perfect- market retail coupon rate. Table 1) for 1972-87. set the price equal to the new-issue price.THE MARKET FOR HOME MORTGAGE CREDIT 107 Table 3-3.22 1973 0. ** Significantly different from zero at a 5 percent confidence level.65** 1985 0. 1988 and 1989 computed by the author. Source: Roth (1988. In this equation. Second. the . they estimated a price equation for GNMAs. Hendershott and Van Order (1989) attempted to correct for this potential problem by constructing a perfect mortgage-like capital market rate and estimating the adjustment of conventional commitment mortgage rates to this per- fect rate (rather than to a Treasury rate). First. Correlation Between Mortgage Rates and Capital Market Rates.42 1979 0.87** 1989 0.33 1981 0.18 1976 0.49 1978 0.91** 1988 0. Year Correlation* 1972 -0.46 1975 -0.19 1974 0.76** 1986 0.34 1980 0.

108 THE CHANGING MARKET IN FINANCIAL SERVICES

Table 3-4. The Time Response of Conventional Commitment Rates to Fictional
Perfect Market Rates.
Adjustment to One Point Rise in Perfect Rate
Time Period Current 3 weeks 5 weeks 7 weeks 9 weeks
1986-88 .59 .95 .96 .87 .84
1983-85 .16 .55 .68 .83 .88
1980-82 .08 .45 .75 .93 1.05
1976-79 .01 .36 .62 .66 .86
1971-75 .06 .17 .37 .56 .74
Source: Hendershott and Van Order (1989, Table 5).

GNMA price was regressed on the coupon (adjusted to a bond-equivalent
basis), the seven-year Treasury rate, and two determinants of the value of
the borrower's call option-the term structure slope (seven-year rate less
six-month rate) and an estimate of the volatility of the seven-year rate.
Various interactions of these variables were included to allow for nonlinear
price responses.
To obtain the perfect-market rate, the estimated price equation was
solved for the coupon rate after the mortgage price was set equal to one
hundred, less the actual points charged in the conventional market (less
one point presumed to equal origination costs). This coupon was then
converted to a mortgage (rather than bond-equivalent) basis, and fifty
basis points were added for servicing and other costs. As the degree of
integration increased, changes in the perfect-market coupon rate should
have been reflected more quickly in the conventional commitment rate. 2
Retail conventional commitment rates were regressed on the current
and lagged one-to eight-week values of the perfect-market rates for various
parts of the 1971 to 1988 period. Table 3-4 reports the cummulative
adjustment of the commitment rate concurrently and over two, four, six,
and eight week lags. The shift toward integrated markets is striking. The
percentage of the change in the perfect-market rate that is reflected in-
stantaneously in the retail conventional rate rose monotonically from
effectively zero in the 1970s to 8 percent in the 1980 to 1982 period, 16
percent in the 1983 to 1985 period, and 59 percent in the 1986 to 1988
period. The fraction of the change in the perfect-market rate reflected in
the conventional rate within two weeks rose monotonically from a sixth
in the first half of the-1970s, to almost half in the early 1980s, to over half
in the 1983 to 1985 period, and to nearly one in recent years.

THE MARKET FOR HOME MORTGAGE CREDIT 109

Table 3-5. Actual and Perfect Market Effective Conventional Commitment Rates
(%).

Actual Perfect Market Difference
1971 7.54 8.33 -.79
1972 7.38 7.92 -.53
1973 8.04 8.97 -.93
1974 9.19 9.78 -.60
1975 9.05 9.92 -.87
1976 8.86 9.22 -.35
1977 8.84 9.09 -.24
1978 9.64 10.08 -.44
1979 11.20 11.34 -.14
1980 13.76 14.24 -.48
1981 16.69 16.55 .13
1982 15.97 15.24 .73
1983 13.23 12.86 .37
1984 13.89 13.52 .37
1985 12.43 11.95 .48
1986 10.19 9.69 .49
1987 10.21 10.01 .20
1988 10.23 10.21 .02
Source: Hendershott and Van Order (1989, Table 6).

Securitization and the Level of Mortgage Rates

Probably the best starting point is a comparison of the actual conven-
tional commitment rate and the Hendershott-Van Order fictional perfect-
market rate. Table 3-5 lists annual values of these rates and the difference
between them for the 1971 to 1988 period. 3 The precise differences are,
of course, subject to some error: the actual rate is a survey rate, and the
perfect rate is computed from an empirical equation estimated with error.
Nonetheless, the overall pattern of the differences seems both systematic
and plausible. The actual rate was three-quarters of a percentage point
below the perfect-market rate in the 1971 to 1975 period; a third of a
point below in the 1976 to 1980 period; and roughly half a point above the
perfect rate in the 1982 to 1986 period.
As explained above, the low mortgage rates in the 1970s can be attributed
to tax advantages for thrift mortgage investments and portfolio restrictions
against nonmortgage investments, and the switch in the 1980s reflects a

110 THE CHANGING MARKET IN FINANCIAL SERVICES

Table 3-6. Effective Loan Rates for California FRMs with Loan-to-Value Ratios
of 75 and 80 Percent by Loan Size, Selected Years.

Percent of Conforming Loan Limit 1978 1985 1986 1987
0.0-50.0 10.12 11.75 10.65 9.83
50.1-67.0 10.04 11.87 10.53 9.82
67.1-80.0 9.97 11.98 10.51 9.77
80.1-90.0 9.97 11.66 10.40 9.63
90.1-100.0 9.95 12.22 10.36 9.62
100.1-115.0 9.94 11.13 10.62 9.63
115.1-130.0 9.97 11.46 10.65 9.91
130.1-145.0 9.95 11.39 10.68 9.80
Over 145.0 9.94 10.97 10.70 9.83
Number of Loans 3,590 710 1,157 1,706
Percent of Dollar Volume Securitized 4 37 57 55
Source: Hendershott (1990b, p. 156).

sharp relative shift of thrifts out of home mortgage investments owing to
the reduced (non) profitability of savings and loans and the expansion of
savings and loan asset powers. The half percentage point premium in the
early 1980s provided the incentive for the securitization of conforming
conventional FRMs. The premium covered the start-up costs of the
securitizers and the liquidity premium demanded by investors.
Beginning in the middle of 1987, the actual rate is very close to the
perfect-market rate, the conventional conforming mortgage market
seemingly being fully integrated into capital markets. Thus, as the volume
of mortgage pools grew, bid/ask spreads were bid down (and thus the
liquidity premium fell), and the per dollar costs of the securitizers declined.
This suggests that the rates on conforming loans, which are eligible for
purchase by the agencies, should have declined relative to rates on non-
conforming or jumbo loans.
The raw data displayed in table 3-6 suggest that beginning in 1986
yields on conforming loans were lowered relative to those on nonconforming
loans. Average effective rates are listed for loans of increasing size (percent
of the conforming loan limit) with similar loan-to-value ratios (75 percent
to 80 percent) originated in California in the May to June periods of 1978,
1985, 1986, and 1987. Holding the loan-to-value ratio and state of origi-
nation constant controls for default risk. California was chosen because it
accounts for roughly a quarter of the dollar value of all conventional
FRMs closed in the United States and more than half of all jumbo loans

THE MARKET FOR HOME MORTGAGE CREDIT 111

(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

investing and insuring (Fannie and Freddie). On the other hand. increases in interest rates reminded thrifts of the problems of borrowing short and lending long. Ironically. congress gave thrifts the opportunity to invest in ARMs that might be attractive to both lenders and borrowers. The agency securitization explosion and the resultant reduction in FRM yields has affected a number of financial firms importantly. different firms are concentrating on origination. that would reprice more in line with thrift deposits. allowing them to compete with state chartered S&Ls. the shareholders of Fannie Mae and Freddie Mac have benefitted. has lowered GNMA yields. there existed one mortgage type-the standard fixed- payment mortgage (FRM). rela- tive to Treasury yields. 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. particularly the surge in CMOs and REMICs. Most obvi- ously. Finally. and thus conventional mortgage rates. servicing. In December 1978. 1990).112 THE CHANGING MARKET IN FINANCIAL SERVICES resecuritization phenomenon. adjustable rate mortgages (ARMs). 1984). One guide is the movement in Fannie Mae's stock price. The latter group includes both S&Ls and commercial banks. both those investing in FRMs and those wishing to originate more ARMs for their portfolios. nationwide authority to invest in ARMs was first granted. and in July 1979. respec- tively. fairly liberal regulations were implemented for federally chartered S&Ls and . At these points (1971 and 1974 specifically). However. Periodically in the 1960s and 1970s. In April 1981. This might suggest an additional quarter point relative decline in the cost of FRMs since the first half of the 1980s (0 'Keefe and Van Order.4 Also benefitting were those investors who were holding FRMs when the agen- cies were lowering FRM rates and thus mortgage prices were rising. about the same time that interest rates were peaking in the early 1980s. The Growth in Adjustable-rate Mortgages Prior to the 1980s. annual and life-of-Ioan rate increases were restricted to one-half and two-and-a-half percent. and the choice of rate index was greatly restricted. an exception was made for federally chartered S&Ls in California. which quadrupled from 8 percent in early 1983 to 32 percent in the summer of 1990. the thirty basis point reduction in FRM yields hurts ongoing portfolio lenders. federally char- tered thrifts lobbied for permission to offer borrowers an alternative choice. securitization has led to greater specialization in separate phases of the mortgage process.

indeed. ARMs accounted for 43 percent of the conventional single family loan volume originated by all lenders (computable from table 3-2). The latter suggests that the ARM share of conventional single-family mortgage originations will vary cyclically as the FRM-ARM rate spread varies. Thrifts. 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.7 Another important determinant of the ARM/FRM choice is the level of the FRM rate (Brueckner and Follain. over the 1984 to 1989 period. the more desperately do households need to obtain a lower initial rate on which they can qualify for a reasonably sized loan. By March 1989. 1990). took advantage of this opportunity. borrowers financing with ARMs can obtain larger loans and purchase larger houses. with the FRM. 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. more mobile households. 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. subsidize less mobile households who pay less than it is worth to them.5 Altern- atively. 1989).THE MARKET FOR HOME MORTGAGE CREDIT 113 savings banks. The sharply reduced . who pay more for the call option/rate cap than it is worth to them. Moreover. ARMs were only 10 percent of the single family mortgage portfolio of FSLlC-insured S&Ls. the higher the FRM rate. 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. 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. Thus. That is. and in August 1982. In the middle of 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).48 percent of the thrift single family loan portfolio (including mortgage-backed securities) was in ARMs (Hendershott and Shilling. 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. and figure 3-1 shows that this variation has occurred.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.

..L.-L...... using an FRM rate of 10.JL.J 1 ..114 THE CHANGING MARKET IN FINANCIAL SERVICES ARM Share Rate Spread 0..........50 and the mean values of the other demographic variables relevant to the ARM/FRM choice..... I .L....5 0.. Brueckner and Follain's (1989) estimation equation implies a 23 percent probability of this household choosing an ARM.. ARM share and FRM minus ARM rate spread quarterly. a thirty basis point higher FRM rate raises the ARM probability to 32 percent..---~---__. has likely reduced the FRM rate as related to the ARM rate...-....... I I I I I I I I I I I I 0. 3..l---L.L-..L. The successful securitization of the FRM market. ARM share in 1986 as related earlier years reflected.5 -...J.5 I I I 0.0 0.----.JL.6 . I \I "• 1... I I I 0. 1984-1989.2 0. but not the ARM market.. Rate Lagged I ...l---1.. an ARM rate of 8.1 L-l---1. the marked decline in FRM yields in 1985 and 1986. I One Quarter I I I 2...0 I I .3 ..'..1-..L.L.-----r-----. However...L...L-. in part...L.50..L-. This can significantly alter the ARM/FRM mix. I I I .....J.J..4 I I 2......-J.. '. l " I .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..7 r-----. .. For example..J.

...... 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 . . This induced securitization. Share of S&L TFA in home mortgages. 0...8r---------......... "" .. ----.--------r-------.... and ratio of S&L TFA to total home mortgages outstanding.............' ' ... The Savings and Loan Industry The institutions that have been most affected by. --. \.5 .. Figure 3-2........ . 0.... " ' . ....--------~------__... or causally linked to.. which successfully lowered conforming FRM rates back to perfect market levels. "S&L Share of Mortgage Market"------" M .. . S&L Portfolio Share 0. 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...... S&L share of total home mortgage market.2~ __ __ ~ ~~~~ __ __ ~ __ __ L_~~_L __ ~ ~~ _L~ 1965 67 69 71 73 75 77 79 81 83 85 87 1989 Source: Flow of Funds Accounts. ........3 .6 0...7 0.. First quarter of 1990 annualized....---.THE MARKET FOR HOME MORTGAGE CREDIT 115 0.

The sharp decline in 1983 and 1984 reflected accelerated growth in the S&L industry (18 per- cent annual growth rate).25. The ratio slipped further to 57 percent at the end of 1987. and fewer non-housing-related loans are now clas- sified as qualified than was previously the case. Recovery and Enforcement Act of 1989 (FIRREA). The ratio of S&L total financial assets to total home mortgage debt outstanding rose from 0. this proportion varied within a fairly narrow 72 percent to 74 percent range until 1981. The S&Ls presence in the home mortgage market is measured as the ratio of S&L total (direct and indirect) home mortgage holdings to total home mortgage debt outstanding. While S&L total financial assets grew between 1984 and 1988. reaching a peak of 0. of course. The act strongly encourages such investment in two ways. As can be seen. restrictions on nonhousing related loans are substantially increased. For example. Since then. the S&L presence rose gradually throughout most of the 1970s. a strengthened qualified thrift lender (QTL) test directly mandates more investment. with a constant S&L mortgage portfolio share. but has since risen to 61 percent. the S&L share of the home mortgage market fell from 73 percent to 63 percent. S&Ls have liquidated nearly $90 billion in agency securities and over $50 billion in direct home mortgage holdings. tangible personal . the S&L presence has been halved. Second. Beginning at 0. they grew at a slower rate than the home mortgage market. in just a year-and-a-half the ratio of S&L total assets to home mortgage debt has fallen from 0. Since early 1989. FIRREA has offsetting effects on home mortgage investment by S&Ls.36 to 0.43 in 1965. not an actual shift out of mortgages. S&Ls must now keep 70 percent of assets in qualified loans rather than 60 percent. First.70 in 1977. before declining to 0. The recent decline is. related to the Financial Institutions Reform.56 in 1969 to 0. limi- tations are tightened on commercial real estate loans. although that act was itself largely a belated recognition of economic realities. The increase between 1969 and 1977 and subsequent decline through 1981 reflected swings in the size of the S&L industry relative to the size of the home mortgage market (recall that the S&L mortgage portfolio share was constant over this period). This presence is the product of the other two series in the figure. The behavior of S&Ls is reflected in the proportion of S&L assets invested in home mortgages either directly or indirectly through holdings of agency securities.51 in 1977.63 in 1981. before plummeting to 59 percent at the end of 1984. S&L assets have been shrinking rapidly. the fraction of S&L assets invested in home mortgages and the ratio of S&L total assets to the book value of all outstanding home mortgages.116 THE CHANGING MARKET IN FINANCIAL SERVICES century. In this year-and-a-half.

the use of brokered deposits. The commitment rate has continuously tracked the seven-year rate. Figure 3-3 contains monthly average values for the Freddie Mac commitment rate and the seven-year Treasury rate. "good" real estate investments have turned bad. This mismatch was. Second. The latter was made easier by the increase in deposit insurance coverage from $40. although regulators are forcing undercapitalized thrifts to liquidate sooner). While simply comparing these rates is generally inappropriate because the call premium built into mortgage coupon rates can change. that interest rates on conforming home mortgages do not appear to have risen relative to Treasury rates. A quick look at the policy mistakes that caused the crippling of the S&L industry is probably worthwhile. Regulatory forbearance and loose oversight.000 to $100. 1994. sharp changes in the slope of the term structure and the vola- tility of interest rates do not appear to have occurred in the 1988 to 1990 period. The Federal Savings and Loan Insurance Corporation (FSLIC) debacle is generally viewed as occurring in two stages (Kane.000 per account. the downsizing. The negative effect of FIRREA on S&L home mortgage investment comes from the increased capital requirements that are already leading to a downsizing of the S&L industry. aided and abetted by congress. which is what one would expect given the full integration of the market. S&Ls then took substantial risks (doubled their bets) in the 1980s. as one might expect. S&Ls are shrinking and being sold off wholesale to nonS&Ls. and the enactment of new asset powers (including additional flexibility in writing mortgage contracts). this second effect. which prevented the widespread introduction of ARMs by a decade. to substantial overbuilding throughout the United States. and the difference in 1990 is about the same as in 1988 and is less than in 1989. Whether ARM rates or jumbo FRM rates have been affected is unknown. this time aided and abetted by both Congress and the Administration. the decline in the home mortgage portfolio share of S&Ls has already been arrested. and wholly-owned service corporations. To date. 1989). First. and existing junk bond holdings must be liquidated entirely. commercial and corporate debt. It is noteworthy. As a result. encouraged this risk-taking and led. sharply rising interest rates easily eliminated the net worth of most S&Ls owing to their asset liability mismatch (borrowing short and lending long).THE MARKET FOR HOME MORTGAGE CREDIT 117 property holdings. of course. though. While some of these provisions are not yet fully in force Gunk bonds do not need to be fully liquidated until July 1. has dominated and the S&L role in home mortgage financing has plummeted. 8 . and commercial banks and other real estate investors are starting to take large real estate losses. in conjunction with the generous tax provisions of the 1981 Tax Law.

5 FRM Commitment Rate 9...5 7 Year Treasury Rate //"'\ . . .- 8.. FRM and treasury rates... / ._... " / ~ ~~ 7./ \ / A\ / / \ / \...5 I I I I 0188 0388 0588 0788 0988 1188 0189 0389 0589 0789 0989 1189 0190 0390 0590 0790 Figure 3-3.. / \ // \y/. / \ / r' .5 / \ / \ / \ / \ / \ /'.v/ \\ / //."\ / \ / ..5~1--------------------------------------------------------------------------------~ 10. .11. ---.

reregulating (new assets powers are out) and reintroducing strict oversight. untoward events (credit problems. As noted. many of their adjustable-rate mortgages have rate caps that will bind in a period of sustained interest rate increases. With a low profit stream. shortly increase taxpayer liabilities. It is noteworthy. the $400 billion so funded is slightly more than was so funded in 1978. Will there be an increase in ARM securitization. 1990). Will the advantages of Fannie and Freddie be reduced. and. however. User fees have been proposed to compensate the government/taxpayer (the in- surers of the agencies) for the greater risk associated with lower capital .THE MARKET FOR HOME MORTGAGE CREDIT 119 The cost to taxpayers and investors of forebearance is truly going to be enormous. What will happen to the S&L industry and what impact will this have on mortgage rates? These questions are addressed in turn. If interest rates should repeat their 1977 to 1986 pattern. Owing to increases in the costs of deposits and advances relative to Treasury rates. In early 1989. rising inter- est rates) quickly reduce capital. has not been removed. with little capital. leading to lower ARM yields and expanded use of ARMs? 3. the vulnerability of S&Ls to periods of sustained increases in interest rates. the agencies have not had to hold as much capital as fully private institutions and the agencies have lower explicit costs. taxpayers would certainly lose another $50 billion or more in present value dollars (Hendershott and Shilling. that the original source of the problem. S&Ls haven't been the low cost supplier of home mortgage credit for some time (Hendershott. 1990b). Moreover. S&Ls were still using roughly 40 percent of their short-term deposits to fund long-term fixed-rate mortgage investments. possibly leading to an increase in FRM rates and to greater fully-private securitizing? 2. Future Prospects Three questions follow from our earlier analysis: 1. rather than just lowering dividends. We are now supposedly solving the S&L problem by recapitalizing or shrinking (forbearance is out). Complicating matters is the erosion in the basic profitability of S&L mortgage portfolio lending.

FHA-ARMs constitute only about 1 percent of the total ARM market because the FHA's one-fifth program (maximum one per- centage point rate adjustment per period and five percentage points over the life of the loan) is unattractive to both borrowers and lenders. Maybe half of the thirty basis point fee required by the agencies (the differences between the mortgage rate. Currently. the FHA share of the ARM market would rise significantly. A doubling of capital from roughly 0. never should we expect as great ARM as FRM securitization. but other less popular products will always exist and these will not have suf- ficient volume and thus liquidity to be widely securitized. and the primary advantage of an ARM to a borrower is a low coupon that increases affordability. would provide some discipline for yields on other ARMs. would have about the same effect. which would seem to be the minimum required if the Treasury's proposal were adopted. If FHA were given a 2/6 ARM program.75 percent of assets to 1. maybe rates would only rise by ten basis points. Possibly the most popular one or two ARMs will become widely (above 50 percent) securitized. about a ten basis point rise in FRM rates would occur. and thus in FRM rates. Successful securitization of an FHA-ARM would likely lead to securitization of a similar conventional ARM. increases in required capital have been advocated to reduce the insurer's risk. However. net of servicing. in turn. the borrowers pay and the yield investors in MBSs receive) is return on equity. Nonetheless the secondary market would provide a ceiling for yields on the popular ARMs and this. with the rest covering explicit costs and a fair charge for the default guarantee the agencies provide MBS investors.120 THE CHANGING MARKET IN FINANCIAL SERVICES requirements and to recapture for the taxpayers some of the advantages extended. However. Bor- rowers dislike the one-fifth loan because lenders charge a higher initial coupon to compensate for the tight rate caps. . The doubling in capital would then necessitate a fifteen basis point increase in the agency fee. the required return will be less because the agencies would be only half as levered.5 percent. To the extent that the agencies can fully pass through their costs to borrowers. Each of these proposals would tend to raise the coupon rate on new-issue conforming FRMs and thus reduce the tilt of new originations from ARMs to FRMs. Because not all of the cost is necessarily shiftable. Securitization of ARMs would be stimulated by a more attractive FHA- ARM. Thus. the fifteen basis point annual user fee on agency mortgage- backed securities (MBSs) proposed by the Bush Administration would raise FRM rates by fifteen basis points. again. if the required return on capital investment in the agencies did not change. Alternatively.

rates on the conforming FRMs that these agencies can buy have been "about right". owing to thrift tax advantages and portfolio re- strictions. In contrast. has lagged behind FRM securitization owing to the lack of both an attractive FHA-ARM and standardization of conventional ARMs. Introduction of a more . ARM securitization. with the ARM share of total originations periodically swinging between a quarter-and-a-half. however. depending largely on the level and term structure of interest rates. This half point "excess" return on mortgages stimulated development and use of the Freddie Mac and Fannie Mae pass-through programs. and virtually all FHAIVA mortgages go directly into GNMA mortgage pools. ARMs have become a major factor in the conventional mortgage market. 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. Mortgage rates are currently about what one would expect given capital market rates. 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. While some of the S&Ls will be sold to commercial banks. 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. and a half point "too high" in the 1982 to 1986 period because thrift profits and portfolio restrictions had effectively disappeared. conventional rates were a half percentage point "too low" in the 1970s.THE MARKET FOR HOME MORTGAGE CREDIT 121 The decline in the S&L industry owing to FIRREA has just begun. empirical estimation implies that conventional yields adjust fully to changes in capital market rates within two weeks. A further 30 percent to 50 percent decline over the next five years should be expected. many banks also face increased capital requirements. and all face higher deposit insurance premiums. On the other hand. The depository institutions sector generally is likely to shrink over the next five years. 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). Moreover. Conclusions The market for fixed-rate mortgages is now fully integrated with capital markets. Since the middle of 1987.

1989. (August). 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. 7." Regional Science and Urban Economics 19(2) (May): 163-187. 2. 113. Freddie Mac stock was not freely traded until around year end 1988. 1990). as well as the FHA-ARM. 8. and in the initial ARM discount. The S&L industry has been shrinking relatively since 1984 and abso- lutely since passage of FIRREA. 1983. 5. Cassidy. versus 8 percent in the summer of 1990. but overbuilding is the pri- mary culprit." Research Working Paper No. and there is no reason to believe that the future shrinkage will. and James R. the share price has moved much like Fannie Mae's. 3. Between 1975 and 1982. Jan K. . Notes 1. 8 percent to 16 percent of FHA multifamily mortgages were securitized (Seiders. 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 1971.278). The rates in this table are not adjusted for points. The adjustment would not affect the differ- ences between actual and perfect rates because the adjustment to both rates would be identical. With thrifts no longer able to qualify borrowers for larger loans based upon deep discounts. 4. The agencies also securitize multifamily mortgages. 6. they are the coupon rates consistent with whatever points were charged. References Brueckner. The one-year Treasury rate was down to 9 percent in early 1983. in the values of the FRM call option and the ARM rate caps. "A Review of the Federal Home Loan Bank Board's Adjustable-Rate Mortgage Regulations and the Current ARM Proposal. The same gain means that the cost of the borrower's call option is less than if the borrower could costlessly refinance. Follain. "ARMs and the Demand for Housing. Since then. Henry J. The past shrinkage does not appear to have raised home mortgage rates. Federal Home Loan Bank Board. Another 25 percent to 50 percent decline is likely. Changes in FRM-ARM rate spreads reflect changes in the Treasury term structure (seven-year rate less one-year). the initial discount plummeted in 1989 to less than one percentage point. and the Fannie Mae share price was up from a low of two-and-one-half in 1981-82.122 THE CHANGING MARKET IN FINANCIAL SERVICES attractive FHA-ARM would likely lead to greater securitization of the comparable conventional ARM product. 1984. The FRM also reflects the cost of the borrower's default option to the lender. 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 1986 Tax Law has also contributed to these losses. that is.

and J. Proceedings of the 14th Annual Conference. "A Call to ARMs: Adjustable Rate Mortgages in the 1980s. Edward J. "Volatile Mortgage Rates-A New Fact of Life?" Economic Review. The S&L Insurance Mess: How Did It Happen? Washington.H.C. 1990. 8-9): 152-163. 1983. Hendershott. Proceedings of the 13th Annual Conference." Expanded Competitive Markets and the Thrift Industry.S. P. (March). 1990b. 1989.H.3 (March): 16-28.. Roth. D.. "The Future of Thrifts as Home Mortgage Portfolio Lenders.. 1989. "The Continued Interest Rate Vulnerability of Thrifts." The Journal of Real Estate Finance and Economics 2: 101-115." Regional Science and Urban Economics 19. Michael and Robert Van Order. O'Keefe. 1990. Federal Home Loan Bank of San Francisco. "The Impact of the Agencies on Conventional Fixed-Rate Mortgage Yields. Villani. Susan E. ." New England Economic Review (March/April): 47-61. P. Hendershott. (Dec. 10-11): 169-174. 1988. P.H. No. Joe. 1990a. Federal Reserve Bank of Kansas City. Kane. 1990. Hendershott. Seiders. "The Composition of Mortgage Originations in the Year 2000.D. "Secondary Residential Mortgage Markets and the Cost of Mortgage Funds. Van Order. Also forthcoming in NBER volume.THE MARKET FOR HOME MORTGAGE CREDIT 123 Hendershott. Peek. and KE. 1989. "Integration of Mortgage and Capital Markets and the Accummulation of Residential Capital. Shilling.H. "Mortgage Pass-through Securities: Progress and Prospects. Woodward.: The Urban Institute Press." Forthcoming in FNMA publication. "Mortgage Pricing: Some Provisional Empirical Results." AREUEA Journal 11 (Summer): 264-87. D. Roth: Volume 73. P." AREUEA Journal 18 (Fall): 313-322." AREUEA Journal 8 (Spring): 50-76. Federal Home Loan Bank of San Francisco. and R. . 1980. 2 (May): 188-210. "Policy Issues in the Privatization of FNMA and FHLMC. 1989." The Future of the Thrift Industry. H." Presented at NBER Conference on Financial Crisis. (Dec. ..L.

and see if I can carry them somewhat further with respect to the arguments made. (which is handled quite well-but. Hendershott does a very thorough job reviewing the history of securitization and analyzing the impact of securitization. I am in agreement. What I would like to have seen developed further is. Having said that of course. they are always well prepared. it could be for being overly brief in this article and attempting to deal with too many issues at the cost of reduced attention to each. Let me start with securitization. In particular. and with which. and conscientiously argued. This particular article embraces many issues in mortgage credit and is really a summary of sev- eral developments that have taken place in the mortgage market. not the issue of the integration of the mortgage and capital markets as a result of securitization. I want to focus on a couple of issues that he raised. and Freddie Mac played in the development of securitization to its current prominent position is well presented and argued. COMMENTARY Herbert M. 125 . Kaufman It is always a pleasure to read one of Patrie Hendershott's articles. I always learn something from them. The role that Ginnie Mae. I think if Hendershott can be faulted for anything. I wouldn't be a true discussant if I now sat down and left it at that. by the way. Fannie Mae.

meeting private capital requirements at a minimum is neces- sary as a first step. which are significant and positive from their role in integrating the capital and mortgage markets. of the government-sponsored agencies (Fannie. will be reversed. although the evidence is mixed. at least with regard to the fixed rate mortgage market. especially in the decade of the 1980s. and significant winnowing away. This is not sensible or defensi- ble. This should not be a goal of public policy. nonconforming conventional loans seem to bear an interest rate penalty for exceeding conforming limits. and is primarily because of the political pressures . and their expertise in handling securitization all were critical to the rapid development and expansion of securitization. Hendershott's work here and elsewhere with various colleagues suggests that integration has reached. 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. in my judgment. These advantages are no longer necessary. Additional arguments for benefits accruing or stemming from the existence of these implicit subsidies can be made for other players. or certainly no longer necessary in the same form for continued development. I will have more to say about adjustable rate mortgages and securitization further into my commentary. their specialist status in the market. 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. It is unarguable that the success of mortgage market securitization. Therefore. was due to the special status of sponsored agencies. Hendershott alludes to the proposals that have been made for capital requirements and user fees. including bond holders. 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. and will not likely be reversed. It strikes me that the time has come to suggest a winnowing. nor is there any reason to believe that the gains that have been made. There is no pressing reason to believe that these institutions could not operate effectively on their own. Freddie) from the kinds of explicit subsidies and implicit guar- antees that they have available to them. after some transitional period. a substantial level. As Hendershott's work at least suggests. The view of their obligations as implicitly government guaranteed. Furthermore. but also including managers of these companies.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.

I now want to tum to a couple of other issues that Hendershott raises. however. certainly the exist- ence of the more complete integration with general capital markets shown in his work has led to increased production in securitized fixed-rate mort- gages and ceteris paribus lower-rate differentials relative to ARMs because of the reward for standardization. not at all surprising that in an integrated en- vironment the impact on mortgage rates will be minimal from the decline of thrifts. However. One is the relative lack of securitization in the ARMs market and the differences that dictate the choice of ARMs vs. his work itself is an argument for why that has happened. I agree with the thrust of his analysis that rate differentials are driving the choice of ARMs vs. that this erosion will con- tinue. Again. which are artificial and suggest non-market determined differentials. namely we in fact have mortgage credit availability from the capital markets in general. that the demise of savings and loan participation in the mortgage market has resulted in no basic rise in mortgage rates. therefore. fixed-rate mortgages. the development of large-scale securitized ARMs will be forthcoming. However. I think over time. If the market can successfully securitize credit cards and automobile loans. will be removed. the flow of funds into the market should not be materially affected. a way to securitize ARMs effectively can be developed. . It strikes me that he is probably overly pessimistic. fixed-rate mortgages. but not in the conclusion. It is unlikely that until these agencies are fully privatized that constraints such as these. If one player's participation in the mortgage market declines. everything else the same. He expresses apparent surprise in the body of the paper. I want to make one point about the demise of the thrift industry. about the development of securitization for ARMs. Finally.THE MARKET FOR HOME MORTGAGE CREDIT 127 that are on these agencies to keep conforming loans at some reasonable level with which middle America is comfortable. as the interest in doing so becomes more apparent and profitable. It is. There is no question that Hendershott is right.

In an important study of equity underwriting risk. understanding the risks associated with underwriting and dealing can be important in understanding what determines the demand for these services. Nellie Liang James M. As banks expand their securities powers. 4 EQUITY UNDERWRITING RISK J. that is. In this study we examine equity underwriting risk within the context of earlier work on price risk associated with equity offerings and underwriting risks over time. the Federal Reserve permitted commercial bank holding companies to engage in limited amounts of corporate securities underwriting and dealing and has recommended to Congress the removal of the Glass-Steagall separation of commercial and investment banking. In January 1989. Furthermore. underwritings for which the underwriter purchases the issue from a corporation and 129 . O'Brien Introduction Risk in securities activities such as underwriting and dealing has become of increasing interest as the separation between commercial and invest- ment banking erodes. the associ- ated risks have a policy significance because of deposit insurance and the safety net provided to large commercial banks by the federal government. Giddy (1985) estimated returns to firm-commitment equity underwritings.

1990) or "certification" services (Booth and Smith. de- spite evidence of substantial price uncertainty when the offer price is set. thus minimizing the risk-bearing function. In our reexamination and extension of Giddy. We first reexamine and extend Giddy's results on equity underwriting price risk. While Giddy is one of the few studies to have estimated equity under- writing risk per se. underwriters may obtain substantial protection against price risk. 2 Whether Boyd and Graham's results on the risks of investment banking actually conflict with Giddy's is unclear. if other services are principal. his results are not without some conflicting evidence. Equity underwriting is a relatively small part of major investment bank activities and most investment banking risk may be coming from other activities. Armed with these findings and an analogy between underwriting and bank lending. the stock market evidence suggests that the variability of the return to investment bank assets also may be greater. the sample period is increased and the effects of price risk on underwriting returns are tested over several subperiods. Because investment banks are also less highly leveraged than are commercial banks. If this were the case. However. 3 However. 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. 1986). in Mandelker and Raviv (1977).! 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. Giddy concluded that underwriting stock may not be any riskier than lending. 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. for example. We then consider equity underwriting risk from a temporal perspective and examine the return variability to equity underwriting over time. He provided further evidence to suggest that equity underwriting may not be any riskier than bond underwriting. as developed.130 THE CHANGING MARKET IN FINANCIAL SERVICES offers it at a pre-set price to the public. A principal result was that. underwriters are able to keep the relative frequency and magnitude of losses to very low levels. such as marketing and distribution services (Benveniste and Spindt. 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 . 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. one would expect a large part of the price risk associated with issuing stock to fall on the underwriter.

Overall conclusions and some implications for commercial bank expansion into equity underwriting appear in the final section.580 1.78 9.) 1. we constructed a data set of com- mon stock issues underwritten over the period 1977 to 1987. Following the analysis of price risk.EQUITY UNDERWRITING RISK 131 Table 4-1.48 8.78 Mean firm size ($ mil. We then test for the significance of the time series of equity underwriting returns on the stock returns for our sample of investment banks.3 Mean public offering price ($) 20.743.409. are shown in table 4-1.641 Total value of issues ($ mil. and made under a firm-commitment offering during 1977 to 1987. for which data were available.0 54. Characteristics Seasoned IPO Number of issues 2. The sample contains 4.7 13.1 Mean underwriting spread (%) 5. This time series vari- ance is decomposed and the effects and relative importance of price risk and other factors on the variance are analyzed.9 21. Included in our sample are all domestic issues registered with the Securities and Exchange Commission (SEC) with gross proceeds of at least $1. Characteristics of Underwritten Equity Issues To examine equity underwriting risk.13 1. If variability in equity under- writing returns over time contributes significantly to the (high) variance of investment bank stock returns. dev.) 89. of the importance of price risk. Characteristics of Samples of Seasoned and IPO Common Stock Issues for 1977 to 1987.5 million.643.13 Std. negotiated.3 Mean offer size ($ mil. the time series of equity underwriting returns should bear a significant relation to the investment banks' stock returns. equity underwriting risk is refor- mulated as return uncertainty over time and measured as a time series variance of dollar returns from equity underwriting.58 Source: Securities and Exchange Commission's Registration Offerings Statistics. we examine the significance of low returns to equity underwritings on the stock market returns for a small number of investment banks. with issues divided between seasoned and initial public offerings (IPOs).221 .) 34. of underwriting spread 2. 4 Some characteristics of this sample of common stock issues.

132 THE CHANGING MARKET IN FINANCIAL SERVICES

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

Price Uncertainty and Risk on Individual Equity
Underwritings

To analyze the price risk associated with a firm-commitment underwrit-
ing, the gross underwriting return is defined as the offer price times the
number of shares sold at the offer price plus the market price times
the number of shares sold at the market price, minus the issue price times
the number of shares issued. This is a gross return because there is no
deduction of expenses incurred by the underwriter in bringing the issue
to market. Formally, the dollar underwriting return, R, is
R = sQ if the market price is greater than or equal to
the offer price (p 2: 0)
=pQp + sQ if the market clearing price is less than the
offer price (p < 0)
where Q == dollar amount of the offering (offer price times number of
shares issued)
Qp == dollar amount of the offering times fraction sold at the
market price if p < 0
p == the difference between the market and offer price per
dollar of the offering price
s == the underwriting spread between the offer and issue price
per dollar of offering price. (1)
The uncertainty of the return at the time the offer price is set will depend
on the uncertainty of the market price, and hence p, given the offer price,
the underwriting spread, and the amount of the offering.
In this section we examine two factors that can reduce the under-
writer's price risk: (1) underpricing the offering (defined as the expected
value E(p) > 0) and (2) making the underwriting spread(s) sensitive to
price risk. Underpricing reduces the underwriter's price risk by increasing
the probability that the subsequent market price of an issue is greater
than the offer price. 7 ,8 The underwriter then earns the entire spread. Also,
if the underwriting spread includes a cushion for price risk, the market
price can fall below the offer price by the amount of the cushion before
cutting into direct costs.

Underpricing and Market Price Uncertainty

Giddy's results indicated that price uncertainty for underwritten issues is
substantial, with that for IPOs being much greater than for seasoned

134 THE CHANGING MARKET IN FINANCIAL SERVICES

Table 4-3a. Percentage Differences between Day-After Market and Offer Prices
for Seasoned and IPO Issues. 1

Seasoned Issues IPO Issues

1977-87 1977-79 1984-86 1977-87 1977-79 1984-86
Mean 1.0* 0.2 2.0* 8.4* 9.3* 5.9*
Std. dev. 9.7 5.2 11.7 22.3 14.9 20.0
Range -58,163 -24,45 -36,163 -50,212 -15,50 -38,212
Distribution of Differences Between Market and Offer Prices (percent)
positive 39.7 27.4 44.4 53.4 63.0 51.0
negative 40.0 46.1 36.9 30.1 30.4 32.9
same 20.3 26.5 18.7 16.5 6.5 16.1

Table 4-3b. Conditional and Unconditional Negative Percentage Differences
between Day-After Market and Offer Prices.

Seasoned Issues IPO Issues
1977-87 1977-79 1984-86 1977-87 1977-79 1984-86
Conditional: Market Prices Below Offer Prices!
Mean -3.69* -2.24* -3.14* -7.60* -5.03* -6.79*
Std. dev. 5.40 2.62 3.73 7.98 3.86 7.23
Unconditional 2
Mean -1.47* -1.03* -1.16* -2.29* -1.53* -2.24*
Std. dev. 3.83 2.10 2.72 5.59 3.13 5.23
* Significantly different from zero at 1 percent.
1 Market prices: closing prices taken from the CRSP daily stock files.
2 For the unconditional, the percentage difference equals the actual percentage difference
if the market price is below the offer price; otherwise, the percentage difference equals zero.

issues. In this section, we expand the analysis of the importance of
underpricing in reducing the underwriter's exposure to price risk. Per-
centage differences between the day after market and offer price, specified
as p in equation (1), are presented in tables 4-3a and 4-3b for the entire
sample and for the sub-samples. The presentation of table 4-3a follows
that of Giddy (tables 6.3 and 6.6, 156, 162) except for different sample
periods. Table 4-3b provides additional information on the distribution of
differences between market and offer prices when the market price is
below the offer price.
The mean percentage difference between the day-after market and

EQUITY UNDERWRITING RISK 135

offer prices, presented in row one of table 4-3a, is our measure of
underpricing. Underpricing is significant but modest for seasoned issues
over the eleven year period. 9 For IPOs, underpricing is substantial, 8.4
percent of the offer price over the entire sample period. The standard
deviation of the differences between market and offer prices, presented
in row two, is our measure of the underwriter's price uncertainty. As
Giddy found with his sample, price uncertainty appears to be substantial
for both seasoned issues and IPOs, although it is much greater for IPOs.
However, the greater underpricing for IPOs as compared to that for sea-
soned issues, reduces the magnitude of negative differences between the
market and offer prices. These reductions can at least partly explain the
comparability between the lower ends of the ranges for seasoned and IPO
price differences, in the third row of table 4-3a, despite the higher stand-
ard deviations for IPO price differences. lO
Underpricing can also explain differences in the frequency distributions
for IPOs and seasoned issues presented in the lower half of table 4-3a. As-
suming that positive and negative expectations errors on the market prices
are equally likely, actual market prices would be above offer prices with
the same relative frequency as they would be below offer prices in the
absence of underpricing. Thus, the difference between the relative fre-
quencies of market prices being above and below the offer prices is an
indication of the effects of underpricing in reducing the probability of the
market price falling below the offer price. Under this reasoning, the small
amount of underpricing on seasoned issues does not appear to materially
affect the likelihood of the market price falling below the offer price for
seasoned issues. However, underpricing substantially reduces the likeli-
hood of the market price falling below the offer price for IPOs. Indeed,
as shown in the table, the likelihood of a negative difference is smaller for
IPOs than for seasoned issues.
The top half of table 4-3b presents means and standard deviations for
percentage differences between the market and offer prices when the market
price is below the offer price. For all three periods, the conditional mean
differences for IPOs are considerably lower (more negative) than for
seasoned issues, with the differences in means significant at the one percent
level. Thus, when market prices fall below offer prices, the expected
percentage difference is substantially greater for IPOs than for seasoned
issues. However, while the conditional price risk is greater for IPOs than
for seasoned issues, the probability that market prices will be less than
offer prices is smaller for IPOs, as shown in the lower half of table 4-3a.
These counterbalancing effects are reflected in the unconditional mean
negative price differences which are equal to the conditional mean dif-

The average underpricing of seasoned issues in the latter period is significantly greater than in the former period at the one percent level. Following Giddy. That is. the greater is ex ante uncertainty about the market price. however. He concluded from this that risk-sensitive underwriting spreads help to stabilize the underwriter's returns (p. IPOs still have greater downside price risk. our evidence of substantial market price uncertainty on stock offerings and underpricing is consistent with Giddy's results. With respect to the sub-periods. Underwriting Spreads and Market Price Uncertainty In addition to underpricing. However. ceteris paribus. though differences in means are still mostly significant at the one percent level (the exception being the 1977 to 1979 period). If underwrit- ing spreads are higher for issues with greater price risk. shown in the bottom half of table 4-3b. although the difference is significant only at the 10 percent level. and the evi- dence from these studies is inconclusive. the standard deviations and ranges for both IPOs and seasoned issues and the conditional and unconditional means for the negative price differences indicate greater price risk in the latter period. In sum. the differences between seasoned issues and IPOs are more moderate. price variability produces smaller underwriting return variability since issues for which realized market prices are low will tend to have higher underwriting spreads. reported that underwriting spreads for IPOs and seasoned issues combined were posi- tively associated with the absolute differences between offer prices and subsequent market prices. in addition to the unex- . While this makes the price risk for underwriting IPOs more comparable to that for under- writing seasoned issues. 156 and figure 6. the larger is the absolute difference between the realized market price and the offer price likely to be. ll Giddy.8). we use the absolute percentage difference between the offer and (day after) market price as an ex post measure of uncer- tainty about the market price when the offer price is set. the average underpricing of IPOs is smaller in the latter period than the former period. risk-related underwriting spreads will also reduce the sensitivity of underwriting returns to price risk. Based on these unconditional mean values. This price difference will contain. We are aware of only a few studies that have attempted to find evidence of risk premiums in equity underwriting spreads.136 THE CHANGING MARKET IN FINANCIAL SERVICES ferences multiplied by the probability of getting a negative difference. This suggests that underpricing also should be greater in the latter period. Our analysis of the effects of underpricing indicates that underpricing for IPOs is important in modifying these issues' downside price risk.

associated with equity underwriting. 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. As described above. For both the seasoned and IPQ equations. the coefficients on price uncertainty. firm size (FS). 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.EQUITY UNDERWRITING RISK 137 pected deviation between the market price and its expected value. In expanding on Giddy's test. These results indicate that underwriting spreads reduce the downside price risk of underwriting and help to stabilize the under- writer's return. Again. 119). in equation (1). are highly significant. this compo- nent is also likely to vary with the issue's risk. for both the seasoned issues and IPQs. the coefficients on the absolute value of the percentage change in price are positive and highly significant. 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. measured by the absolute percentage price difference. and hence scale economies. Risk of Loss on Equity Underwriting The tendency of underpricing and underwriting spreads to increase with the price uncertainty of the issue being underwritten can be expected to reduce the sensitivity of underwriting returns to price risk. 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). For this analysis. The negative coefficients on issue size suggest that there are fixed costs. Table 4-4b presents regression results based on the subsample of issues for which the day after market price fell below the offer price. 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. we assume that the entire issue is sold at the market price if the price is less than the offer price (that is. Qp . controlling for issue size (IS). an (anticipated) underpricing component. ceteris paribus. We do not model the joint determination of the spread and underpricing. In this section. and (through separate regressions) whether the offering is IPQ or seasoned. Following Giddy. we examine estimated returns on underwritings and their sensitivity to price risk using the procedures developed by Giddy. 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.

84E-5*FS (160.36) (-5. Our calculations are returns per dollar of the amount offered.06 + .73) (-9.56) (-0.22 (65.59) (1.010*IS .8) (10.. Effects of Price Uncertainty on Underwriting Spreads.58 + .134*AP+1 .138 THE CHANGING MARKET IN FINANCIAL SERVICES Table 4-4a..07) (1. this assumption may substantially understate the underwriter's returns on poorly received issues.011 *IS + 9. Since part of an issue is usually sold at the offer price even when the market price is below the offer price.. our three calculations are as follows: Calculation Market Condition Underwriter Return 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.20 (121.016*AP+1 .. Giddy made three hypothetical calculations of the underwriter's gross dollar return to each underwritten issue.040*IS . b.46) (-13.3) (7. 1 All Issues IPO: s = 8. depending on market conditions one.3) (14. five. Spread (s) and absolute price differences (AP+l) are per- centages. or P+lO ~ 0 s otherwise P+lO + s . = Q if p < 0).21) 1 t-statistics in parentheses..78) (-15.1.45E-6*FS R2 = . On the other side.7) (7. and ten days after the offer date.84) Seasoned: s = 5.069*AP+1 .30E-6*FS R2 = .80) (-1. overstates the net return on an issue.66) Seasoned: s = 5.10 + .024*IS + 9.085* AP+1 . Using the return notation in equation (1) and the assumption that Qp = Q if p < 0. the use of the gross underwriting spread.64) Issues with Negative Day-After Market and Offer Price Differences IPO: s = 8.33 + . which does not net out the underwriter's expenses.002*FS R2 = . Issue size (IS) and firm size (FS) are in millions of dollars.24 (57.

20 -46.0 4. The standard de- viation of returns for IPOs is only moderately higher than that for seasoned issues.2 0.1 Breakevens 0.2 0.21 -49.3* 4. These three calculations are substitute estimates of an underwriter's return on an issue. if the market price initially falls below the offer price. 4. dev.5 * Significantly different from zero at 1 percent.5 7. Seasoned Issues IPO Issues 1 day 5 days 10 days 1 day 5 days 10 days Mean 4. The calculations show that the mean under- writing returns are positive and significant for both seasoned issues and IPOs. 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.4* 5. When estimated returns are negative using day-after market prices.6 87.8* Std. The mean returns. The next two calculations assume that.EQUITY UNDERWRITING RISK 139 Table 4-5.2 12. if the market price one day after the offer date is less than the offer price.20 -39.8* 5.7 86. reported in the first row. In the lower half of table 4-5. the relative frequency of estimated losses .20 Percent of Gains 93.6 Range -48. even though the standard deviation of the differences between market and offer prices. as com- pared to -6. 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).5 Losses 6. where P+i is the value for P using the closing market price i days after the offering date.2 87.26 percent for seasoned issues (not shown in the table). is much higher for IPOs. The first calculation assumes that. Hypothetical Underwriter Returns for Seasoned and IPO Common Stock Issues for 1977 to 1987 (percent).0 91.1 13. Underwriter returns calculated for seasoned issues and IPOs separately for our full sample are presented in table 4-5.1 84.5 0. 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.6* 3.4 15.2 0. -9.0* 3.07 percent.7 5. the average return is moderately lower for IPOs. with somewhat higher average returns on IPOs.21 -49.8 8. shown in table 4-3a. the underwriter sells the entire issue at the market price one day after the offering.2 12.4 6.3 5.21 -42.6 0.

the standard deviations increase. despite the much greater price uncertainty of IPOs.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. However. and the loss probabilities increase. although the number of offerings differ substantially. Further. the estimated loss frequencies are moderate and are in the same general range as the frequencies previously reported for the longer sample period. This is consistent with the greater market price uncertainty in the latter period shown in table 4-3a. As the length of the holding period increases. These estimated loss probabilities are about double those reported by Giddy (tables 6. These loss probabilities based on next day prices indicate that. Also. 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. On the whole. the mean returns. average underwriting returns for both types of issues decline. despite the presence of very substantial price uncertainty. To the extent that this occurs. 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. however. market price declines are within the range of the underwriting spread. respectively. These tests required assumptions about . These results may simply reflect the random nature of stock price changes and the upper bound on the under- writer's return. There are no substantial differences in returns across the periods. Investment Bank Stock Returns and Low Returns on Equity Underwritings The previous tests provide evidence that price risk is modest for firm- commitment equity underwriting. 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. For both periods. the estimated returns using the day-after prices are upward biased and the risk of loss is understated. slightly greater return variability and loss frequencies in the latter period as compared to the former. The results show. shown in table 4-3a. conditioned on a negative return. for most issues.4 and 6. these results do not indicate a lot of sensitivity of underwriting returns to general market conditions. Estimated returns are presented separately for the sub-periods 1977 to 1979 and 1984 to 1986 in table 4-6. On the other side. Nonetheless. the probability of loss is only moderately greater for IPOs than that for seasoned issues. were lower (more negative) in the latter period for both IPOs and seasoned issues (not shown).6). however.

5 87.8 4.0 1984-86 Seasoned Issues (N = 694) IPO Issues (N = 683) 1 day 5 days 10 days 1 day 5 days 10 days Mean 4. dey.1 3.9 3.7 6. Hutton.10 -16.9 * Significantly different from zero at 1 percent.9 8.9 5. and Paine Webber over the eleven-year period and First Boston since mid-1983P For each of the individual banks.0 90.3 Losses 2.3 3.4* 5.0 0.3 0.3 5.2 3.7 0. 1977-79 Seasoned Issues (N = 321) IPO Issues (N = 46) 1 day 5 days 10 days 1 day 5 days 10 days Mean 3.3 4.10 -8. Merrill Lynch.10 -15.9 Losses 3.0* Std.9* 5. Such an effect may occur if losses on average tend to be much larger than the calculations in the last section suggest.7 93. 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. 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 .6 84.2* 3.6* 5.2 88.3* 5.5* Std.2 Breakevens 0.9* 6.2 0.5* 4.F.10 -19. Hypothetical Underwriter Returns for Seasoned and IPO Issues 1977 to 1979 and 1984 to 1986 (percent).9 2.6 0.18 -31. 2.5 7.8 11.2* 2. E.5 Range -26. dey.16 Percent of Gains 96.0 0.10 -14. daily stock returns were taken for four investment banks.5 87.8 4.7 Breakevens 0.10 Percent of Gains 97. For the purpose of this test.EQUITY UNDERWRITING RISK 141 Table 4-6a.3 0.1 0.8* 5.8* 3.0 8. 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.18 -26. 3.9 11.8 11.5 84.3 91. b.2 14.18 -28.1 Range -15.3 6.8 88.9 14.16 -32.1 0.0 91.16 -28.3 13.

and zero otherwise. Its advantage is that it gives us a meaningful number of observations of low returns since estimated losses are so infrequent for individual lead underwriting firms. IS These results suggest that the market typically does not regard low returns on equity underwritings as having much significance for the investment bank's profits and reinforces our conclusions from our earlier tests that price risk may not be significant. 16 Conclusions and Limitations The results presented here support the findings and conclusions of Giddy concerning the significance of price risk on stock offerings. in- stead of a low return variable. to have very significant effects on the profits of large investment banks. Also.142 THE CHANGING MARKET IN FINANCIAL SERVICES was only a participating underwriter is not available)Y The "low return" variable equals the number of shares offered multiplied by the difference between the market price and the offer price. These results do not imply that there is no potential for large losses.14 The coefficient on the low return variable is expected to be positive if low or negative returns to equity underwritings have a signifi- cant negative impact on the bank's profits and hence stock returns. Vulnerability to market declines is illustrated by a number of estimated underwriting losses in the week just prior to the October 19. 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. were similar. 1987 market crashP These estimated losses produced an estimated nega- tive dollar return on seasoned equity underwriting for the fourth quarter . The regression results are reported in table 4-7 using alternatively the investment bank and market return on the day of the offering and on the next day. as indicated by the results from our subsamples. While the majority of the low return coefficients have the expected positive sign. on average. the t-statistics are low and only one coefficient is significant at usual test levels. Furthermore. low underwriting returns or losses do not appear. results from using a loss variable. Nonetheless. This is a low return definition and not a "loss" definition because it does not take into account the underwriter's spread. when this difference is negative. Underwriters cannot entirely eliminate their exposure to sharp declines in market prices and there is the element of choice in determining risk exposure. in the infrequent instances when they occur. estimated loss probabilities do not appear to be very sensitive to general market conditions.

85) (-. Effects of Low Returns to Equity Underwritings on Investment Bank Stock Returns.52) (.70E-5 .780 for all investment banks except First Boston which has a sample size of 1.06) (18.0005 1.08E-5 .0003 1.26) (23.174.02E-5 .09) First -.0004 Ll42 .88E-5 .02E-5 1. .12 Webber (.07E-5 . .05E-5 .89) (24.04E-5 .Table 4-7.82) (.099 -.06E-5 .11 28 6 Hutton (1.43) First -.098 .01£-5 .09) (2.0001 Ll29 .34 Boston (-. Day-of-Offering Stock Returns 3 E.55) (.0004 Ll42 .06E-5 .05E-5 .0003 1.173.74) b. .20) 1 t-statistics in parentheses. .12) (1.26) (23. 4 Sample size is 2.38E-5 .F.11 Hutton (1.30) Paine.0005 1.779 for all investment banks except First Boston which has a sample size of 1.83) (-.010 -.86) (-.25) Merrill .28) (.03) (18.81) (24. F.97) (LlO) Paine .57E-5 .92) (.11) Merrill .0001 Ll31 . 2 The market return index is the CRSP value weighted index of returns from the NYSE and AMEX exchanges.64) (19.85) (.17 110 15 Lynch (.92) (1.34 13 3 Boston (-. 3 Sample size is 2.17 Lynch (. Day-After Stock Returns 4 E.02E-5 -.59) (.60) (-1.009 -. 1 Coefficients Number of Seas IPO Low Returns Market Low Low Firm Constant Return 2 Return Return R2 Seas IPO a.12E-5 .12 13 4 Webber (.45) (.67) (19.

we first specify the relation between the vari- ance of dollar returns to equity underwriting over time and the various sources of return uncertainty. This will permit some evaluation of the contribution of the variance in equity underwriting re- turns over time to the variance of investment bank stock returns. There is. This variance is a more inclusive measure of underwriting return uncertainty. Temporal Uncertainty of Underwriting Returns Time Series Variance of Underwriting Returns. are the sum of the returns on individual underwritings . Underwriting Returns Over Time Return uncertainty in an underwriting business comes from more than price risk on individual issues. Other sources of uncertainty include unex- pected variability in the number of stock offerings. this adverse experience is relatively modest when compared to the size of the October stock market decline. Underwriting returns during time period t. but it may also be influenced by a reluctance to underwrite in order to reduce price risk. Following this. There is also the potential for incurring multi- ple losses over a short period. a limitation in just using price risk to measure the risk associated with underwriting. however. R. we examine the relation between the time series measures of equity underwriting returns and the stock returns of the investment banks studied in section three. This and other aspects of uncertainty over time in underwriting returns are taken up in the next section. the number of underwritings varied greatly between the two subsamples. Nonetheless. In this section.144 THE CHANGING MARKET IN FINANCIAL SERVICES of 1987. let Rn be the gross dollar return on the nth underwriting during time period t (t is one quarter). the size of offerings. The combination of these factors will determine the variance of the firm's equity underwriting returns over time. This partly reflects a reduced supply of offerings. and underwriting spreads. While our results on price risk were not very sensitive to the sample period examined. The relative importance of different sources are then estimated. This reduced volume of underwriting also reduces underwriting returns and should be taken into account in measuring underwriting risk. and underwritings fell sharply in the fourth quarter of 1987 following the stock market crash. To define underwriting uncertainty on a temporal basis.

The under- writing return covariances may be positive or negative. given the dollar size of the underwriting and the underwriting spread. To reduce this complex- ity. A second source of variance in underwriting returns over time is the variance in the number of underwritings. r = E(Rn). Let V(R) be the variance of R which is our measure of underwriting uncer- tainty over time.21 One quarter was selected as the time unit and the time-ordered returns were aggregated into quarterly returns for each of the two types of underwritings. 22 Each time series can be viewed as representing the quarterly returns to a firm that gets a pro rata share of the quarterly underwriting income in IPOs . we first had to obtain a time series of underwriting returns. assume that Nand Rn are independent. which is also not recognized. is the (expected) sum of the covariances between returns on individual underwritings within the time unit. This source also is not captured when studying returns to individual underwritings. the size of offerings. (J2. These covariances depend on covariances across offerings between the market prices of individual issues. and spreads. This was done by time-ordering the previously estimated returns on the individual underwritings for seasoned issues and IPOs separately. To estimate the relative importance of the three sources of uncertainty identified in equation (3). 20 Relative Importance of Dift'erent Sources of Uncertainty. One is the variance of the return to individual underwritings. The third source. 19 As shown in Appendix A. the variance of R under this assumption is V(R) = E(N)(J2 + r2V(N)+2E[ L7L~ Cov(Rj. and the last expectation on the rhs of (3) is taken over the distribution of N. Rj )] (3) where (J2 = V(Rn). Three sources of variability in the returns to underwriting over time appear in equation (3).N R (2) ~n=l n where N is the stochastic number of underwritings during the period.EQUITY UNDERWRITING RISK 145 R= "". This variance will depend on the occurrence of losses or low returns on individual underwritings due to market price uncertainty. It will also depend on (unexpected) variation in the dollar size of underwritings and underwriting spreads. 1s The stochastic nature of the number of underwritings per period complicates the determination V(R). V(N). 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.

the variance in returns to individual underwritings contributes less than 15 percent to . Percentage Contribution of Variance of Number of Offerings to Variances of Alternative Quarterly Underwriting Returns. 2 Firm-commitment Best Best Efforts plus Common Preferred Total Efforts Firm-commitment 75 58 74 67 72 r2 == square of the sample average dollar return per underwriting V(N) == sample variance of the quarterly number of underwritings E(N) == sample average of quarterly number of underwritings a 2 == sample variance of the dollar return per underwriting Residual == V(R) . As shown in the next column. Results are presented in table 4-8a.39).E(N)a 2. as in part a of the table. best efforts 2132 (3. 812 were seasoned issues. The number of issues (average size per issue. 70 percent or more of the variance comes from the variance in the number of offerings per quarter. 2020 were IPO issues. $ mil. For preferred firm-commitment. preferred firm- commitment 844 (62.03). for best efforts.36). For the one-day-after market prices.r2V(N) . Variance Decomposition of Quarterly Underwriting Returns. 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). and seasoned issues. Percentage Contribution of Variance from:! Number of Offerings Individual Returns Days After Offering [r2V(N)] [E(N)a 2] Residual Seasoned Underwritings 1 72 13 14 5 63 16 21 Initial Public Offerings 1 70 12 18 5 66 12 22 Table 4-8b. The time series variances are decomposed into the three components identified in equation (3). and slightly less for the five-day-after results.) are: Common firm-commitment 4221 (26.146 THE CHANGING MARKET IN FINANCIAL SERVICES Table 4-8a. 2 The percentage contribution from rV(N).

returns were calculated for individual underwritings and time aggregated for these alternative groups. The aggregation of these (small) covariances can account for a significant portion of the residual variance component. Nor is consideration given to firm-commitment preferred stock underwritings or best efforts offerings. Further calculations suggest a posi- tive covariance between the number of IPO underwritings in the quarter and the return per IPO underwriting. Some additional calculations (reported in Appendix B) examine the residual component further.EQUITY UNDERWRITING RISK 147 the quarterly underwriting return variance. As can be seen. Roughly 15 percent to 20 per- cent of the quarterly return variance comes from the Residual. Although not explicitly allowed for in our specification of the quarterly return variance. The results (not re- ported) were the same as for the longer sample period in that over 70 percent of the return variance was coming from variability in the number of underwritings. However. over 70 percent of the return variance comes from the variance in the number of underwritings. for broader groupings. The results in table 4-8a do not report the variance decomposition for returns from underwriting both IPOs and seasoned common stock. this positive covariance may also be accounting for some part of the residual variance for IPOs. The variance decomposition of returns reported in table 4-8a were calculated using the entire eleven-year sample period. The most important implication of the results of table 4-8b is that broadening the types of stock underwritten does not reduce the relative importance of the number of underwritings in accounting for the variance of underwriting returns. the relative importance of the variability in the number of offerings over such a long period may not be of much significance if it reflects long-term trends in the volume of underwriting to which firms can easily adjust their scale of operations. the variance of the return on individual stock offerings. The underwriting firms in our sample that did firm- commitment underwritings generally did both seasoned offerings and IPOs for both common and preferred stock. which . To determine the importance of the number of offerings in the variance decomposition using these alternative and broader categories of under- writing. We thus calculated for the various underwriting categories the relative importance of the number of underwritings for the return variances over subperiods of three to four years. However. In sum. These calculations suggest that dollar returns on in- dividual underwritings have on average a small positive covariance. the major underwriters were associated with only a very small amount of best efforts offerings.23 The percentages of the under- writing return variance accounted for by the number of offerings (r2V(N)) are presented in table 4-8b.

For the other set. IPO. The major factor accounting for the estimated quarterly underwriting return variability is the number of underwritings. 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. Previously. and preferred) for all investment banks. the bank earned a pro rata share of the aggregate returns. as lead and participating underwriter. IPO. Quarterly Underwriting Returns and Investment Bank Stock Returns The results just presented suggest that factors other than price risk. particularly the number of offerings. 25 We have no data on the returns to underwriting in which the investment banks participated but were not lead underwriters. appears to be a small part of the variance of equity underwriting income over time. the correlation coefficients in table 4- 9a are mostly small and negative. they should be signifi- cantly related to the bank's stock returns. However. Except for Merrill Lynch. 24 This result holds for broader as well as narrower measures of equity underwriting returns and for shorter and longer sam- ple periods. In one set of correlations. and preferred) for which the respective banks were lead underwriters. For the four investment banks. These aggregate returns would proxy for the bank's equity underwriting returns if. equity underwriting returns (aggregate) are estimated from underwritten issues (seasoned. are the major determinants of return variability to stock underwriting. the correla- . or because of diversification effects. the equity underwriting returns (underwriting-specific) are estimated from issues (seasoned. If unanticipated movements in the bank's underwriting returns over time have an important effect on the bank's profits. Here we consider the sensitivity of investment bank stock returns to variability in underwriting returns over time. variation in equity underwriting returns may not be important because the variability is small in relation to other activities. A small positive covariance between returns on individual underwritings may also contribute to the time series return variance. 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. we considered the sensitivity of investment bank stock returns to only price risk.148 THE CHANGING MARKET IN FINANCIAL SERVICES includes losses due to unfavorable market prices.

097 Common + Preferred .082 -.222 -.194 .098 Total Common -.006 . First Merrill Paine Offering Hutton Boston Lynch Webber Average Underwriter-specific Underwriting Returns 2 Seasoned .269 -.257 .112 .485* .119 .008 . 1 Underwriter-specific Underwriting Returns 2 Common -.375* .016 Common + Preferred -.212 -. 26 With the possible ex- ception of Merrill Lynch.091 Total Common . Correlations Between Innovations in Quarterly Equity Underwriting Returns and Investment Bank Stock Returns.090 .076 -.EQUITY UNDERWRITING RISK 149 Table 4-9a.102 IPO -.316* . The unanticipated component is measured as the innovation in the equity underwriting re- turn where the innovation is the residual from an autoregression of equity .434* .138 .059 .103 Aggregate Underwriting Returns Common .207 IPO -.150 .391* .376 .020 -.105 .165 .369* . F.045 . We have no ready explanation for the significance of the results for Merrill Lynch. since it is the unanticipated component that the market would be expected to respond to.008 -. measures of unanticipated bank-specific and aggregate equity underwriting returns are used. 1 Sample size is 40 for all investment banks except First Boston which has a sample size of 18.002 -.062 Aggregate Underwriting Returns Seasoned . Significance not computed for averages. Correlations Between Quarterly Equity Underwriting Returns and Investment Bank Stock Returns. 2 Underwriting returns are estimated quarterly returns for which investment bank was lead underwriter.378 * Significantly different from zero at 5 percent.020 Common + Preferred -.330* -.133 -.320* -.032 .097 -.007 -. In table 4-9b.224 . 1 Type of E.051 -.107 .017 .097 -.199 .204 .095 Table 4-9b.141 .081 -. the results presented in table 4-9a do not indicate that variability in quarterly returns to equity underwriting contributes importantly to variability in investment bank stock returns.128 .305* .209 .308 Common + Preferred .124 .056 . tion coefficients are all insignificant.

64 ~~~ ~~ (~) 1 t -statistics in parentheses.060 .205E-8 .195E-9 . 27 The results reported in table 4- 9b for the underwriter-specific returns show no significant correlations between innovations in bank-specific underwriting returns and the invest- ment banks' stock returns.65 ~~~ ~n) (~~ Underwriter-specific Common + Preferred Stock Underwriting Returns -. Thus. regressions of the stock returns of Merrill Lynch on a market index and innovations in the various aggregate underwriting return measures are presented.64 (-.019 .28 Because the innovations in quarterly equity underwriting returns are significantly correlated with the market return index. the various equity underwriting return measures have no significant effect on Merrill Lynch's stock returns when market return effects are controlled for.15) (.93) (7.' Coefficients Constant Market Return Underwriting Return Underwriter-specific Common Stock Underwriting Returns -.14) Aggregate Common + Preferred Stock Underwriting Returns -. However. underwriting returns on its past values. As can be seen.14) Aggregate Common Stock Underwriting Returns -.365E-1O . we performed a further test to see if innovations not associated with the market index were related to the investment bank's stock returns. are higher and frequently significant. Effects of Innovations in Quarterly Underwriting Returns on Merrill Lynch Stock Returns. since it was Merrill Lynch whose stock returns were most highly correlated with the underwriting returns. For this test we used only the stock returns for Merrill Lynch. In table 4-10.058 .021 2.150 THE CHANGING MARKET IN FINANCIAL SERVICES Table 4-10.009 .023 2. the correlation coefficients using the innovations in the aggregate underwriting returns.021 2.66) (.92) (7.307E-9 . it is possible that the aggregate underwriting return measure is simply proxying for the market return. Sample size is 40. shown in the lower half of table 4-9b.023 2. While it is still possible that variation in equity underwriting returns does help .64 (-.

We find strong evi- dence that underpricing and underwriting spreads are sensitive to price risk. the variance of returns to equity underwriting over time. However. our tests can- not separate any such effects from general market effects. 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. 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. That is. we find that the major determinant of this time series variance is the variance in the number of offerings. These results are not sensitive to the sample period. 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. Giddy's and our results. In reformulating the underwriter's risk in terms of a quarterly time series variance of equity underwriting returns. "business risk. which suggest that underwriting equity may involve relatively small price risk. 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. the amount to be underwritten." 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". and the underwriting spread) and return covariances between issues account for only a minor share of the time series return variance. Recent literature has emphasized other motives for stock underwriting. The return variance on individual issues (which depends on uncertainty about the market price. although the two types of risk need not be independent. Conclusions The results of our study reinforce the evidence and conclusions of Giddy (1985) concerning price risk in equity underwriting. Baron (1982).EQUITY UNDERWRITING RISK 151 to explain variation in the investment banks' stock returns. which depends on both types of risk. Booth and Smith (1986) also suggest that underwriters provide a certification service . Baron and Holmstrom (1980). In this case. raises doubts about the importance of risk-bearing by the underwriter as a motive for firm-commitment underwritings. might still be important in accounting for variation in investment bank profits and hence stock returns.

152 THE CHANGING MARKET IN FINANCIAL SERVICES on security offerings when the issuer has an information advantage over investors on the value of the firm. These arguments bear on the risks and opportunities for commercial banks as they expand into the business of underwriting corporate equity. the potential for profit may lie in having an established network for the marketing and distribution of issues and a reputation for being able to evaluate the issuing firm and the worth of the issue. Both of these "assets" of the underwriting firm can reduce the underpricing of an issue and thus make the firm more attractive to issuers. Appendix A Variance of a Stochastic Sum of Random Variables The following definitions and assumptions are used: Rn == a sequence of random variables with a common marginal distribution N == the number of variables in the sequence and is independent of Rn . given their dealings with institutional investors and their traditional role and customer relationships in the financing of firms. our results are limited to equity underwriting. variability in the frequency of underwriting may be the greatest source of variability in equity under- writing income. large commercial banks would seem to be relatively well positioned to gain entry in corporate securities underwriting. as well as the other terms of underwriting contracts. Going beyond our results and utilizing suggestions from the recent literature on underwriting. While this suggests that underwriting is not a "quick entry" business. These studies emphasize that the effective provision of these services requires an ongoing business so that customer relations and reputation are established. the substantial uncertainty about market clearing prices on stock offerings requires prudence in setting offering prices (underpricing) and underwriting spreads. Although we did not find strong evidence that this in- come variability has been important for investment banks. Once established. Although our results suggest that underwriters are able to obtain strong protection against price risk. Income variance from activities such as brokerage and dealing could be of greater importance for commercial banks expanding their investment banking services.

7(10)10. 2.EQUITY UNDERWRITING RISK 153 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] = EN[ r: V(Rn)+2r~ r:jCOV(RbRj)IN J+ VN [r: E(Rn)INJ = (J2EN[r: IN] + 2EN[ r~ r:jCOV(Rb Rj)INJ + r 2VN [r: IN] = (J2E(N) + r2V(N) + 2E[ r~ r:jCOV(Ri. . 3.. for j > i. as shown in equation (3). the ordering was random. If i is the ith-ordered underwriting and j is the jth-ordered underwriting.018 and the average covariance was $5.29 The average correlation was . 30 Evaluating "Residual" at the mean covariance (with N = 37) yields . Rj )]' Appendix B Residual Variance in Quarterly Underwriting Returns Under the assumption that the number of underwritings and the return per underwriting are independent. A rough attempt was made to determine the importance of underwriting return covariances on the quarterly return variance by examining covariances on returns to "nearby" underwritings. the residual component in table 4-8.. For this purpose. the "distance" between i and j is defined as j . . For multiple underwritings on the same day. seasoned issues and IPOs over our eleven-year sam- ple were each chronologically ordered. Rj )] where E[ ] is taken over N.i. There was no readily apparent pattern in the correlations and covariances as the "distance" increased up to 37. 37 (the average number of IPO underwritings per quarter was 37). For IPOs. is "Residual" = 2E [r~ r:Fov(R i . we calculated correlations and covariances between dollar returns to underwritings i and j for "distances" between i and j of 1.

This compares with the actual residual in table 4-8a of $1841(10)11.>uld also contribute to the quarterly return variance through the residual component. we calculated correlation coefficients between the average dollar return per underwriting each quarter and the number of issues underwritten in the quarter for IPOs and seasoned issues. "Residual" evaluated at the mean covariance suggests that a substantial fraction of the actual IPO residual reported in table 4-8a can be attributed to small positive covariances between returns on nearby underwritings. which is significant at the one percent level. The average correlation and covariance were . which would include all bank securities activities. pp. Kwast (1989) also reported that the standard deviation of asset returns across trading departments of commercial banks.39 for IPOs. In option based models of corporate debt yields. The actual correlations were . Notes 1. 3. The same exercise was repeated for seasoned offerings. If the dollar return on individual underwritings within the quarter is independent of the number of underwritings (as assumed in equation (3». Again. Equity under- writing revenue would be a smaller percentage. this positive correlation w<. The estimated "Residual" with the covariances evaluated at the mean was $1628(10)11. As a rough indication. such as Merton (1984). there was no readily apparent pattern in the correlations and covariances as the "distance" increased up to 59.8(10)10. Giddy (1985. except that the correlations between returns on i and j were calculated for "distances" up to 59 (the average number of seasoned offerings per quarter). as well as to stockholders. Finally.008 and $4. was much higher than that for the banks' return on assets excluding the trading department. and . If so. we would expect these cor- relations to be zero.154 THE CHANGING MARKET IN FINANCIAL SERVICES $756(10)11. The asset return standard deviation is important to creditors (or a deposit insurer) concerned with the risk of default. 2. the default premium is an increasing function of the firm's asset return standard deviation. Benston (1990) provides a different view of the relative risks of commercial and investment banking. These correlations suggest that there may be some positive dependency between dollar returns on individual IPO underwritings and the number of issues being underwritten in the quarter. which compares with the actual residual in the table of $3409(10)11.6) reports that underwriting revenue for NYSE firms doing a public business averaged about 10 percent of total revenue between 1974 and 1982.12 for seasoned offerings which is not significant even at the five percent level. Allowance has to be made for leverage because of its positive effect on the firm's equity return variance and market /3. . 166. respectively.

155).05 percent. and Paine Webber 1. 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». Litigation risks are not considered here.69 for securities firms using annual return data). The mean firm size reported is based on the 86 percent of firms that reported total assets. Tinic (1988) has emphasized that underpricing provides protection against litigation risk from dissatisfied investors. While there is no allowance for a normal return between the time of the offer price and when the subsequent market price is observed. . 15. These issues account for 85 percent of the total amount of seasoned issues and 68 percent of IPOs registered with the SEC. 6. see Malatesta and Thompson (1985). and ten days after the issue date. 14. suggest- ing that low returns are not explained primarily by the overall market return. The one-day return for the CRSP value-weighted market return index over our sample period is . 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. 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. Market prices for the issues subsequent to the issue date are from the CRSP data. Thus. This is an alternative form of the "event study" model. Merrill Lynch 1. and total assets of the issuing firm are from the SEC Registration Offerings Statistics (ROS) data. 16. It is recommended by Acharya (1989) when the firm experiences repetitive events. Correlations between "low return" variables and the market return are low. Hutton 1. gross proceeds.67. When 5-day returns are used. p. five. date of issue. However.57. 11. 9. 5.EQUITY UNDERWRITING RISK 155 4. they did get some indication of a positive relation between the spread and the ratio of the idiosyncratic and systematic variances. using returns in excess of a normal return would have no appreciable effect on the results.72. 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. Even though the price risk variable uses the day-after market price. underwriter spread.46. managing underwriter. 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. 7. which was their measure of price uncertainty. 13. 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. multicollinearity is not an issue in interpreting the insignificant coefficients on the low return variables. All the issues could not be included because of either the unavailability of CUSIP numbers or stock market price data for one. the estimated market f3s are: First Boston 1. 12. For another application. 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. 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. 10. public offering price. the market may have information on the market's reception of the offering on the day of the offering. 8. The issuer. 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.

29. However. 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. 26. While these losses are directly attributable to the stock market crash. which included merger activity as well as underwriting. The estimated returns are for firm-commitment underwritings of common stock. the sum of the covariances would be negative. 24. if the size distribution of underwritings in a given period was fixed (e. accounted for 10. The moments of R. Over the 11 year sample period. While its presence in this market was greater than that of other underwriters. the conditional expectations will be time variant. For an individual underwriter. In this presentation. Significant coefficients were almost always confined to the first and second lags.. 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. 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. There was substantial serial correlation in the aggregate underwriting returns but very little in the bank -specific returns.S. 22. 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). Thus. ceteris paribus. 28. Variances were also calculated using annual returns. 20. N. 18.g. Merrill Lynch was the largest underwriter of sea- soned issues of stock. 21. our esti- mated losses do not include those experienced by U. we ignore explicit conditioning and time variation in the mo- ments of the distributions.4 percent of its gross revenues over the period 1983-87. 27. As would be expected. for firm- commitment preferred underwriting. 23. 19. the annual return variances were much higher than the quarterly return variances. This assumption is considered below. In general. Underwriting returns were regressed on the returns over the previous four quarters. 25. Results are reported below for underwritings that include preferred stock and best efforts offerings. and Rn should be conditioned on information available at the time the expectations are formed. Annual reports show that investment banking income for Merrill Lynch. there is no allowance for losses due to unfavorable market prices. Returns to best efforts offerings were not included because the four underwriting firms were agents for only a few of these offerings. Because of multiple underwritings on some days and variation in the number in days . underwriters of British Petroleum stock in late October 1987 which were reported to be in the millions of dollars.156 THE CHANGING MARKET IN FINANCIAL SERVICES 17. underwriters to take a large risk exposure.S. the relative importance of the different sources of the return variance was similar to that pre- sented below for the quarterly returns. the fraction of its annual revenues from all types of underwriting appeared to be commensurate with those for other large investment banks. 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. Returns to firm-commitment preferred stock offerings and to best efforts offerings contain only the dollar underwriting spreads earned on the offerings. 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. For example. Since our underwriting data includes only issues of domestic corporations.

Smith. 1982. Myron L. "Partially Anticipated Events: A Model of Stock Price Reactions with an Application to Corporate Acquisitions. "Insiders Decision of Corporate Events: A Simple Measure of Stock Price Effect. indicating that the correlations and covariances tended to zero for large "dis- tances. and Stanley L. and Rex Thompson. Conditional on Outsiders Information. "The Investment Banking Contract for New Issues Under Asymmetric Information: Delegation and the Incentive Problem. "The Impact of Underwriting and Dealing on Bank Returns and Risks. Giddy. "Is Equity Underwriting Risky for Commercial Bank Affiliates?" In Deregulating Wall Street. 30." Unpublished paper (May). Ingo Walter. Paul H. 1985. Baron. Kwast." Journal of Finance 32 (June): 683-694.EQUITY UNDERWRITING RISK 157 between other underwritings. Stephen J. 1989. Lindvall. Boyd." Journal of Banking and Finance 13 (March): 101-125. 1985. John H. Benston. and Bengt Holmstrom. 1985." Journal of Financial Economics 14 (June): 237-250. 1974. "The Behavior of Investment Bankers: An Econometric Investigation.i is not a constant time interval. James R. New York: Oxford University Press." Journal of Financial Economics 24 (October): 343-36l." Federal Reserve Bank of Minneapolis Quarterly Review 12(2) (Spring): 3-20." Journal of Financial Economics 14 (March): 3-31. "A Model of the Demand for Investment Banking Advising and Distribution Services for New Issues. . Booth. the average correlations and covariances declined. David P. "The Profitability and Risk Effects of Allowing Bank Holding Companies to Merge with Other Financial Firms: A Simulation Study. 1986. 1989. George J. Mandelker. Warner. New York: John Wiley & Sons: 145-69. Dennis and John R. Logue. References Acharya. 1990. Malatesta. 1977. Lawrence M." Journal of Financial Economics 15 (January/ February): 261-28l. Ian H. and Jerold B. Benveniste. 1988. "How Investment Bankers Determine the Offer Price and Allocation of New Issues. Gershon and Arthur Raviv. Brown. and Richard L. Editor. "Investment Banking: An Economic Analysis of Optimal Underwriting Contracts. j . Graham. 1990." Journal of Finance 29 (March): 203-215." This result held for returns to both IPO and seasoned offerings. "Using Daily Stock Returns: The Case of Event Studies. The Separation of Commercial and Investment Banking." Journal of Finance 35(5) (December): 1115-1138. 1980. "Capital Raising. David P. It serves only as a rough measure of the temporal "closeness" of individual underwritings. and Paul A. Sankarshan. Spindt. Baron. As the "distance" was expanded further. Underwriting and the Certification Hypothesis." Journal of Finance 37(4) (September): 955-976.

158 THE CHANGING MARKET IN FINANCIAL SERVICES

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

but by 1972 had grown to $77.4 billion (representing about 80 percent of
the total American presence abroad), and by 1988 to $318 billion. The
number of foreign branches of U.S. banks grew from seven in 1955 to 627
in 1972 and to 849 in 1988. Foreign subsidiaries of U.S. banks had total
assets of $146.6 billion in 1988. 1 The assets of foreign banks in the United
States grew from only $28.3 billion in 1972 to $751.0 billion in 1989.2 The
total number of foreign bank offices in the United States at year-end 1989
was 704. 3
This pattern of greater foreign direct investment in banking in the
United States, compared to American investment abroad in the 1980s, is
similar to that found in other industries. The increased internationaliza-
tion of business has resulted in increased competition in many markets
that were previously protected from entry by vigorous competitors from
other countries. Whereas in the early part of the 1970s complaints of
unfair competition were leveled against American banks abroad, the latter
part of the 70s and the 80s found more complaints from domestic American
institutions. In fact, the International Banking Act of 1978 was intended
to equalize competitive opportunities between domestic and foreign banks
in the United State and will be discussed later. In 1992 Europe will allow
banks to operate across country borders. Because of reciprocity questions,
this could affect competition within the United States between foreign
and domestic banks and is discussed in the last part of this article.
Foreign banks can operate within the United States by employing various
organizational forms, each with its own purpose and focus. The competitive
analysis must distinguish among the major organizational forms. Foreign
banks are attracted to different states for different reasons and conse-
quently may compete for different types of business by state. The major
states with foreign banks are analyzed separately here. Finally, banks
from different countries might behave differently. Japan, which has by far
the largest number of offices and the greatest value of assets of foreign
banks in the United States, has attracted much attention and is accorded
special consideration here.
This article seeks to answer several questions about foreign banks in
the United States including:

1. What has affected their growth?
2. In what areas do they compete?
3. Do they have unfair competitive advantages?
4. What has been their competitive effect?
5. What will be the role of foreign banks in the United States in the
future?

4 percent in 1989. One study has esti- mated that through 1985.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 163 Even though this article is unable to answer all of these questions com- pletely. "of foreign commercial banks currently operating in the United States 74 per cent (224 banks from sixty countries) first established their U."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. The first section discusses the institutional setting of foreign banks in the United States. Barclay's Bank International (1890). Other well known banks that established offices early on were: Hong Kong and Shanghai Banking Corporation (1875). and three Swiss. the growth by country. The ratio of foreign controlled to do- mestically controlled assets has risen dramatically from 3.S. and Mitsubishi Bank (1920). The second section reviews empirical studies of the factors that have affected the overall growth of foreign banks. and growth into various states. Every year has witnessed an increase in the share of foreign banks. The growth of Japanese banks is also reviewed. the legal framework. which set up an agency in New York in 1818. The Institutional Setting The first foreign bank to establish an office in the United States was the Bank of Montreal. the data in table 5-2 indicate that most of the growth of foreign banks in the United States has occurred since 1972. 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. presence after 1970. including an evaluation of the changes proposed for Europe in 1992. and the extent of the foreign bank presence in the United States. of these. three Italian. seven were British. These banks engaged primarily in trade finance and funds transfers and par- ticipated in the New York stock and bond markets. 5 Furthermore. 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. These data by themselves would suggest that for- eign banks are attracting relatively more business and placing competitive . it does provide a useful start for the competitive analysis of foreign banks. In the third section the descriptive evidence of competitive effects is presented and the allegations of unfair competitive advantages evaluated. Lloyds Bank International (1856).6 percent in 1972 to 21. including the types of foreign banking organizations. five Canadian.

164 THE CHANGING MARKET IN FINANCIAL SERVICES Table 5-1. Foreign banks can establish subsidiaries either through acquisition or de novo entry. Many are oriented toward retail business. They do maintain credit balances that are very similar to deposits. but can act as agents for the foreign bank and forward payments or loan documents to the home office. These offices neither take deposits nor make loans. like the agency. Krishnan. 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. Subsidiar- ies have identical powers as domestic banks and are regulated in the same manner. the choices enable foreign banks to tailor their American operations to their business desires. it has traditionally been engaged mostly in wholesale operations. Year of United States Entry by Foreign Banks. and Nigh (1987) 61. however. but most payments cannot be made from these accounts. Since different forms permit different types of activities to be performed and require different levels of commitment by the parent organization. The final major form is the subsidiary or commercial bank. Representative offices are often established as a precursor for further involvement. The areas where the competitive pressure has been most intense are examined later in the article. The acquisition by foreign banks of several large domestic banks. they cannot make consumer loans nor accept deposits. It can offer a full range of services and. Funding is from the parent bank or by borrowing in the Federal Funds or interbank markets. The most limited. 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. pressure upon domestic banks. and consequently. such as Marine Midland and Crocker in 1979 and 1980. the rate of growth of the usage of each form has varied. 6 Agencies are an integral part of foreign banks and may make commercial and industrial loans. The branch is the most important organizational form and is an integral part of the parent bank. There have been changes in the relative advantages and disadvantages in each form over time. Foreign banks can utilize several different organizational forms to operate in the United States. but the easiest to establish of the organizational forms. is the representative office. 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

stimulated several proposals to limit foreign bank growth in the United
States.
Other less important forms should be mentioned for completeness.
Investment companies are similar to agencies but are the only form
permitted to deal in securities. They are limited to New York state. The
International Banking Act permitted foreign banks to establish Edge Act
Corporations to maintain competitive equality with domestic banks. They
are chartered by the Federal Reserve Board and must engage in only
international banking activities outside the home state of the parent bank.
Table 5-3, taken from a recent survey by the American Banker, indi-
cates the size of foreign banks by type and in relation to domestic banks
for June 1989 and June 1988 as measured in four different ways: number
of offices, C&I loans, deposits, and assets. Clearly, foreign banks now
play an important role in the American banking scene. Significantly, it
should be noted that since 1982, by all four measures, foreign banks have
expanded steadily compared to domestic banks. In terms of assets in
1989, branches were by far the most important form with 63.8 percent of
total foreign assets, followed by commercial banks with 23.2 percent, and
agencies with 12.1 percent. The higher ratios of C&I loans, than other
measures, indicates that foreign banks may be more important in the
market for loans to large companies. This will be further explored when
the financial ratios of foreign banks in relation to domestic banks are
analyzed.
Foreign banks from more than sixty countries have some form of office
in the United States. Table 5-4 lists the assets by country of foreign banks
for 1988. The ten most important countries, however, account for almost
90 percent of the total assets of foreign banks in the United States, and
table 5-5 presents the total assets for banks from each of these countries
from 1980 to 1988. In 1988, the Japanese share was greater than 55 percent
of all foreign banks and was more than five times larger than the next two
most important countries, Canada and Italy. Note that there is quite a bit
of variability among the countries over time. Some, such as Italy, Japan,
and Canada, have grown rapidly, while others, such as West Germany,
France, and the United Kingdom have experienced a more moderate rate
of growth.
The dominance of the Japanese is further illustrated by a list in table
5-6 of the twenty-five top foreign banks in the United States in 1989. Of
the ten largest banks, eight are Japanese, and of the twenty largest, fifteen
are Japanese. Consequently, it is appropriate to examine closely the
Japanese banks in the United States.
Foreign bank activity in the United States varies widely by state. Part

THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 167

of this variance is due to the types of laws erected by states to either
attract or deter foreign banks, but it is also due to the economic attrac-
tiveness of the states to foreign banks. Table 5-7 presents the total assets
of United States offices of foreign banks by state from 1980 to 1988. In
1988, only sixteen states plus the District of Columbia had foreign bank
offices, with New York, by far the most important, with greater than two-
thirds of total assets. California is the second most important state and
has been particularly attractive for Japanese banks. It is not surprising
that the top three states, New York, California, and Illinois, accounted for
94.6 percent of total foreign bank assets in 1988, since these three states
contain the major international financial centers in the United States. On
the other hand, there has been substantial growth of foreign banking
activity in other states. Florida, for example, has attracted almost as many
foreign banks as Illinois. Since Florida serves primarily as a point of entry
for foreign wealth from Latin America and the Caribbean, it has a major
foreign bank presence in deposit-taking, but relatively little in lending. In
the later analysis of foreign bank financial ratios, each of the four largest
states for foreign bank activity is examined separately.
In a discussion of the competitive impact of foreign banks in the United
States, it is important to recognize the legal environment in which the
foreign banks operate. The most important law of recent vintage is the
International Banking Act of 1978. Prior to 1978, many aspects of foreign
bank activity in the United States were unregulated and domestic banks
claimed that foreign banks had unfair competitive advantages. In particular,
foreign banks had significant authority to establish offices across state
lines, while domestic banks were severely limited. The main purpose of
the International Banking Act was to equalize the treatment of foreign
and domestic banks. The act limits the multi-state operations of foreign
banks by prohibiting them from establishing offices outside their own
home state. New branches and agencies established by foreign banks outside
their home states could only accept the same types of deposits accepted
by the Edge Act Corporation, a form available to domestic banks. However,
foreign banks were allowed to retain their existing interstate operations.
This led to a flurry of foreign bank expansion across state lines prior to
the effective date of the law. Federal chartering of branches and agencies
was established. The Federal Reserve was given authority to impose reserve
requirements on both federal and state chartered branches and agencies
and FDIC insurance was required for all branches taking retail deposits.
Foreign banks were made subject to the non-banking provisions of the
Bank Holding Company Act and thus could no longer engage in new
activities not permitted by domestic bank holding companies. Finally, the

.......
0\
00 Table 5-3. Foreign Banks in the United States, summary of Survey Findings. C&I Loans, Deposits, and Assets at U.S.
Offices of Foreign Banks (Dollar Amounts in Billions).
Number of Offices C&I Loans Deposits Assets
1. Agencies and Branches 6/30/89 565 $129.9 $247.1 $527.5
6/30/88 543 118.5 216.0 479.6
+22 +9.6% +14.4% +10.0%
1A. Agencies 6/30/89 201 $27.8 $22.4 $83.9
6/30/88 194 25.2 17.1 86.1
+7 +10.3% +31.0% -2.6%
lB. Branches 6/30/89 364 $102.1 $224.7 $443.5
6/30/88 349 93.3 198.9 393.5
+15 +9.4% +13.0% +12.7%
2. Commercial Banks (a) 6/30/89 100 $46.6 $121.3 $161.3
6/30/88 85 43.9 112.9 152.0
+15 +6.2% +7.4% +6.1%
3. Edge Act Banks (b) 6130/89 26 $0.26 $1.6 $2.0
6/30/88 28 0.29 1.9 2.4
-2 -10.3% -15.8% -16.7%
4. Investment Companies 6/30/89 11 $2.2 $0.4 $4.9
6/30/88 10 1.7 0.3 4.2
+1 +29.4% +33.3% +16.7%
Totals for U.S. Offices of Foreign Banks (c) 6/30/89 697 $178.9 $370.4 $695.6
6/30/88 666 164.3 331.2 638.3
+31 +8.9% +11.8% +9.0%

073. $ .5 -451 +5. Banks (d) 6/30/89 13.9% 6/30/86 4. deposit.4% 6/30/85 4. branches and agencies of foreign banks. (b) Edge bank number count includes Edge branches.4% 24. and assets. 1989 and June 30.S.8% 6/30/87 4. all U. Total for All U.3% 20.5% 16.S.4% 12. offices of foreign banks-including those for which data were not available-divided by all U. data for Edge banks. Source: American Banker.8% 16.639 $628.-chartered. C&I loans.2 2.1% 21. commercial banks.7% 27.6% 20.S.6% 6/30/84 3. deposits and assets for June 30. To avoid double counting.138.5% 14.S. (c) The total number of U.2 $3.2% 18.S.2% 15. . are excluded from the totals for C&I loans.6% 18. The number of U.1% 28. 1990.S.7% 6/30/83 3.0% Ratio of Foreign to All U. FDIC-insured commercial banks. deposits. Banks (e) 6/30/89 5.3% 22. Edge Act bank offices.S. and C&I data for all U.2% 21.. the U. which are subsidiaries of foreign bank-owned U..2% 21.1% 14. and New York State investment companies. These data were provided by the Federal Deposit Insurance Corporation.7% 11.9% 17. banking offices..4% 16.0% 6/30/82 3. (e) The ratio is all U.3 $2. banking offices of foreign banks-including those offices for which data were not available-were 704 on 6/30/89 and 675 on 6/30/88.3% 22.S. Branches of these banks are not included.5% 17. banks includes all U.042. (a) Commercial bank number count includes bank head offices only.6 6/30/88 14.6% 6/30/88 4. 1988 were used for all banking offices. banks are from the Federal Reserve Bulletin.S.S.7% +5.. and Federal Reserve Board.8% 15.S.5 2.0% +4. February 27.5% All data totals and the number of offices shown here-under each office type-are for the banking offices providing data for the survey.090 598. 18A. (d) The asset.927.6% 21.

252.381 Luxembourg 756.224.154 Kuwait 993.861 Germany (FDR) 13.404 Greece 1.392 Cayman Islands 130.588 Colombia 825. Country Assets Country Assets Country Assets Austria 1.616.282.425 Mexico 4..S. 936.756 Saudi Arabia 743.952 Turkey 310.131.085.407 EI Salvador 874 Pakistan 669.266 Belgium 3. 257.425.436 Yugoslavia 655. Peoples Rep.027.421 China.477 Korea.798 Norway 1.210.419. Total Assets by Country of the U.859.969 Jordan 316.139 Denmark 2.744.948 Egypt 174. South 4.044.387 United Arab Emir..400 Iran 17.177 France 29. (Assets in Thousands of Dollars). Offices of Foreign Banks Listed by Country All Countries (1988).600 Netherlands 10.022 Singapore 380.930.182 Source: Federal Reserve System..708.035 Japan 360.630 Israel 9.407 Venezuela 2.869 Finland 1. Europe 5.410 Uruguay 214.732.249.092. .744 U. 29.153.602. 32.112 Australia 3.424 Qatar 116.047.752 Indonesia 2.126.735 Switzerland 23.142 Phillipines 321.688 Chile 85. 964.431.495 Peru 30.l 0 Table 5-4.426 Other W.442 Italy 29..246 Portugal 658.024..346 Multiple Countr.959 Ecuador 187.868 Thailand 824.101 Ireland 6.255.980 Dominican Rep. -.742 Bermuda 137.853.731 Grand TotaL for aLL Countries: $653.076 Panama 136..792 Brazil 4.694 Spain 6.590 India 750.628 Canada 39.274 Sweden 1.K.665.703 China.391 Argentina 864.249 Bahrain 1. Rep. of 2.652 Malaysia 525.010 New Zealand 892..605 Hong Kong 25.197.879.

.042 44.633 526.896 653.134.084.409 336.484.739 7.905 39.658 19.617 441.431.971 29.139 Italy 9..733.853.801.170..126.957.757 20. (Thousands of Dollars) Ten Largest Countries 1980 1981 1982 1983 1984 1985 1986 1987 1988 Canada 15.287 251.090.941 422.605 8.453.736.021.685.400 17.336.767 8.244.835.442 360.631.920.259.205 294.415 39.961 151.838.812.518.287 10.705.S.646.682 12.653.588 Gennany.644 5.732.J .312.300 38.243 24.384. -.815 29.799.511.666.539.914.891.760 All Countries 198.367 267.516 16.412 25.400 4.252 53.925.931.212 11.854 113.425 Japan 72.288.568.101.455 17.071 36.684.115.503.338.185 8.280. Table 5-5.252.349.896 583.382.525.538 301..121 532.989 12..718.772 16.834 4.972 16.894.237 52.656 290.043.535 27.136.175 13.182 125.868 32.097..523.780 8.010 Israel 4.983.426 6.890 23.376.253.216..245 8.678 245.137 88.863.092.215.199 25.284.225.445 15.096.249 United Kingdom 25.461 225.761.428 21.135 51.472 22.678 8.074.074 13.255.817 24.495 Switzerland 11.899 40.553..299 42.171.813.939 390.077.545 594.982.150 4.512 14. 1980 to 1988.005.249.409 5.239..101.185 15.632 9.707.879.755 7. Offices of Foreign Banks.780 7.929. .319 46. Total Assets of the U. Ten Largest Countries (as of 1988).692.077.445.920.708.148.780 41.098 13.859.565.561.681.024.653.911 23.757 24.980 Hong Kong 11.589.313.221 178.568 11.718..866 7.014.115 18.941.869 Netherlands 36. West (Federal Republic) 7.188 378.141 20.049.445.335.571.308 7.058.443.357 18.880.180 7.825 10.217.288 25.499. .763 39.182 Source: Federal Reserve System.379..051 7.581 333.292 43.605 Totals: Ten Countries 173.593 57.400 France 12.573 23.

.5 11 12 Long-Term Credit Bank of Japan.8 6 6 Sanwa Bank Ltd. Tokyo 6.684 6.030 8.457 25. Ltd.710 17.893 4.372 13..071 32. Ltd.472 16.9 8.210 $25.64 3.8 5 7 Industrial Bank of Japan Ltd.859 28.67 $9.762 19.. 1988 (dollar amounts in millions)..380 18.960 33.. 1989.664 4 +29.971 11.104 26.514 8 +14. (a) $12..98 6.62 4. Based on Business Loans booked through U.81 5.3 . -..040 8 +64.. of % Chg.177 4 +68.414 8.190 8 +13. of % Chg. compared with June 30. Osaka 7.011 7 +36.07 5.7 3 3 Mitsubishi Bank Ltd.444 7 -38. Top twenty-five Foreign Banks in the United States.729 4 +27.69 5.680 2.2 8 9 Dai-ichi Kangyo Bank Ltd.881 14 +25.628 8 +63.616 3 +60.167 10.523 17. Basle 3. Totals on June 30. Tokyo 7.647 12.409 18.0 13 13 Mitsui Bank Ltd.881 15.838 30.595 29.625 12.66 7. 1988 Rank Rank No... Off. Tokyo 3. Tokyo 3.124 12.I N Table 5-6.9 10 11 Bank of Montreal 5. Ltd.661 11.979 14.595 22.742 14. Tokyo 8..03 7.281 6 +16.761 29.361 13.081 7..S.2 4 2 Sumitomo Bank Ltd.936 9. banking offices June 30.514 2.266 21 -4..988 25.543 3.763 15..215 $18.070 11 +37.2 9 5 Hongkong & Shanghai Banking Corp.68 3. Tokyo 5.228 13.490 18 +9..5 12 15 Tokai Bank.251 22.390 4 +35.799 7 +17..024 7 +34. Tokyo 7. 6/89 6/88 C&I Loans Assets Deposits U.066 17.254 4 +6.689 9 +52.374 18. 6.663 7.722 14. Osaka 7.185 16.895 25. London 9. Nagoya 4.648 4 +14..45 8. No.S.398 6.357 31.681 5 +5.050 $32.551 7 -1.821 $47.64 5.3 15 22 Mitsui Trust & Banking Co. C&I Bank of Tokyo Ltd.725 10.009 8 +40. 1989 Totals on June 30. C&I C&I Loans Assets Deposits U. Off.35 2.S.700 7 +31.722 9.0 14 8 Swiss Bank Corp.83 5.070 5 +20.913 16.667 16 +0.889 4 +19.544 11 +41.209 6.223 10.8 7 10 Fuji Bank Ltd.914 24.886 28.588 5 +23.742 15.2 2 4 National Westminster Bank Pic.189 17.

650 4.468 5 +13.370 2.133 2.453 3. February 27.830 7.482 4.123 3. Ltd. >-' -l tN .218 3 +35.797 5. branches.330 5.65 1.538 173 +22.786 3.6% 62.918 4 +23. Tokyo 2.621 8. % Chg. commercial bank subsidiaries.235 4 +74.6% 64.661 237.6 22 25 Banco di Roma. 178.0 Totals for the top 25 foreign banks in the U.677 3. Kobe.915 399.7 23 26 Banque Nationale de Paris 2.94% 111. 17A.768 3 +16.S. Toronto.41 2.0% Totals for all foreign banks in the U.853 9 +2.230 8 +17. U.779 19 -12.575 2 +14.71 2.400 697 +8. p. which was completed on 10/31188.72 1.4 20 14 Bank of Nova Scotia.136 2.720 7. Amsterdam 1.180 4.200 666 +18.3 18 21 Mitsubishi Trust & Banking Corp.9% Sources: (1) American Banker survey questionnaires and Federal Reserve Board data.648 14 +30.046 2 +47.6% Top 25 banks' share of total for all foreign banks 70.03 2.9 21 18 Daiwa Bank.1% 25.323 11.102 8.709 8.8% 68.939 8. and assets for all of the bank's U.378 11.355 4 +65. for which data were available. deposits. New York State investment companies. banking offices.394 442.705 10.57 1.707 9.300 331.424 5.572 6.671 6.296 4 -12.479 173 +12.300 638.1% 62. respectively. 126.716 3 +12. Naples 2. and Edge banks. (2) American Banker.926 9.S.642 10. Tokyo 3.3 17 17 Taiyo Kobe Bank Ltd. Ltd.676 3 +24. Osaka 2.024 4..653 4.9 24 33 Sumitomo Trust & Banking Co.S.012 9. 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.1 19 20 Banco di Nipoli.155 1..06 2.849 7.600 370. The totals for each bank are based on the combined business loans (C&! loans).6% 63..90% 164.S.261 5. Canada 2. Rome 2.218 9 -28. Osaka 1.584 9 -23.076 +13.925 -12.499 2.096 5.420 8..900 695.03 3. 16 19 Nippon Credit Bank Ltd.124 4.44 2.1% 24. banking offices include: agency offices.830 3..9 25 23 Algemene Bank Nederland. Japan 2.747 3. Los Angeles.13 2. C&I is the annual growth in commercial and industrial loans over the year ended 6130/89 and 6130/88.741 205. 1990.929 5 +8.635 7.

782 Source: Federal Reserve System .055.355 797.401 2.555.250.740 North Carolina 0 0 0 0 0 0 0 0 249.480 48.814.821 Florida 671.122 Oregon 3.630.144.252 32.545 3.482 Delaware 0 0 0 124.471 1.381 333.996.174 48.387.386 District of Columbia 5.950 106.525.073 2.603 362.551 113.l +:>.970.700 453.621 591.217.022.079.384 2.436.317 4.936 95.926 30.000 2.289 412. -.095 3.959 Texas 48.662.935.417 4..043.905. Offices of Foreign Banks.911 2.648.101.420 79.400 99.888 4.816.600 149.641.836 384.005.578 974.632 299.387 93..241 2.745.301 1.299 698.378 1.093 100.846 367.995.804 0 368.261 26.639 California 50.632 190.896 2.459 7.359.729 1.052.269.189 Massachusetts 278.617 441. Table 5-7.538 301.330 2.175 178.571.694.099 335.593.776 103.355 173.137 418.188 378.327 776.782 117.682.849.617.127 821.124.610 243.366..917 9.573 2.312 1.656 511.854 88.302 86.306.650.943 293.668 Washington 1.699 24.842 11.001.216 59.223 822.420 22.203 5.099 36.005 1.055 435.686 883.323.427.348 2.314 251.401.431 2.273 339.115.586 4.358.966.530 142.868 Illinois 8.759.531.684 6.462 3.750 37.040.689.376 Hawaii 22.019 4.469 2.885 649.093.573 343.980 405. Total Assets of the U.579.452 4.088 3.819.583 214.716 9.432 196.238.355.341.633 526.387.064.411 9.662.136 6.714 48.615. 1980 to 1988..731.448 11.083 3.100 1.051.000 2.012 284.126.895.085 67.717 3.000 2.166 161.485.668.707 28.709.005.062 263.339 709.539.883.902.410 5.015.641 1.602 40.110 3.898 Pennsylvania 360.809.633 3.S.273 Maryland 0 0 0 2.400 112.000 0 Arizona 0 0 0 0 0 0 0 0 25.514 3.525 2.763 2.904.806.283 Louisiana 0 69.516 Georgia 1.128 Totals 199..574 5.007.852.100.934 2.819.185.170 280.393.434 New Mexico 0 0 0 0 0 360.204. (Thousands of Dollars) State 1980 1981 1982 1983 1984 1985 1986 1987 1988 New York 132.980.671 3.313.514.435 340.066.

Several of the major studies are summarized here. including studies of overall foreign bank growth.S. a number of studies have been done to empirically isolate the factors most responsible for foreign bank growth. G. $IDM = the exchange rate between dollars and German marks (dollars per mark). MN = U.S. by country of origin. The Motivation for Foreign Bank Growth in the United States In order to evaluate the competitive situation. imports divided by U. The study employed quarterly-time-series data from 1972 through 1979 to estimate the following model of foreign bank growth: y\ = f(RDIFF1.S. balance of payments (balance on current amount). commercial bank assets. GNP relative to the growth of OECD countries' GNP. personal income.S. This act went a long way toward equalizing the competitive opportunities available to foreign and domestic banks. $IDM. it is important to first understand the factors that have affected the growth of foreign banks in the United States.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 175 power of Edge Corporations was expanded and foreign banks were permitted to own Edge Corporations. we examine Japanese bank growth in the United States because of the import presence of Japanese banks. arguments are being made that foreign banks have competitive advantages. RDIFF1 = difference between the quarterly average Federal funds rate and the 3-month Eurodollar rate. PIE. MN. CA. and by state. FDI.S. DUM77).S. In addi- tion. PIE = the price-earnings ratio for bank stocks. CA = U. where Y\ = foreign banks' share of total U. Since 1981. (1) . DUM77 = dummy variable for anticipation of International Banking Act. but once again. FDI = net foreign domestic investment in the U. G = the growth of U. The first study estimating the factors that determine the level of for- eign bank activity in the United States was Goldberg and Saunders (1981a).

774) +0. the value of the dollar. (ii) the falling PI E ratios for United States bank stocks. 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.. (2) Results using other dependent variables were very similar. Useful insights can be obtained by examining the factors that have resulted in differential growth across countries of origin of the foreign banks. the IBA appeared to have a negative affect on the underlying growth of agencies and branches leaving subsidiaries relatively unaffected.0.00001 b CA + 0.939) (6. branch.509) (0.176 THE CHANGING MARKET IN FINANCIAL SERVICES The regression estimates are presented in equation (2).00198 $IDM . Hultman and McGee (1989) updated these studies employ- ing a model with three independent variables.9332 " Significant at 5 percent level. profitabil- ity.7 A second study by Goldberg and Saunders (1981b) examined the growth of foreign banks by agency. and the International Banking Act of 1978.112) (-2.0.W = 1..690) (-1. (iv) the persistent de- preciation in the dollar.8 Recently. 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. and (v) expectations that the IBA would have a restrictive effect on foreign bank activity in the United States. The growth of subsidiaries was positively related to the first two of these variables and negatively related to the bank price earnings ratio.0023 8 PIE + 0. "The empirical results suggested that these factors tended to impact on agencies in a different manner from branches and subsidiaries.0.1654 MN (0. with agencies more affected by international business considerations and more concerned with current. Also. 0. (iii) the increased size of (net) foreign domestic investment in the United States. "The results indicate that foreign banks are .093) (-5. and subsidiary with a similar model of four explanatory factors.0103" DUM77 (2.0004 G + 0. 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. Y1 = -0.8130 R2 = 0. b Significant at 10 percent level.00006 8 FDI .0030 ..540) D .303) (-2..521) (0.00268 RDIFF1 . or short-run.

Canada. 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.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. The level of international trade of the state was also im- portant in explaining foreign bank assets.S. The im- portance of foreign trade and foreign direct investment suggest that foreign banks frequently are serving customers from their home countries. Hultman and McGee (1989) have used the data for Japan to identify factors affecting the growth of Japanese banks in the United States. and the Japanese Banking Act of 1982 that opened Japan somewhat to foreign banks. activities of Japanese banks during this period appeared strongly related to Japanese domestic financial variables as well as to conditions in the U.. offices of Japanese banks responded to both price and quantity restraints on domestic Japanese banking activity. the more likely are banks to build their asset bases in the U. whether measured by assets or number of offices. Goldberg and Grosse (1990) find that the foreign bank presence..10 It is also useful to examine the factors causing differential foreign bank growth across states. while interbank trading at U. market. Rela- tive profitability. These studies suggest that a number of economic factors and an im- portant regulatory change affect the growth of foreign banks.9 Until recently. Finally.S. market.S. Since past regulatory changes have affected foreign bank growth. and all of Europe. The coefficient was the expected positive sign. the cost of acquiring banks relative to earnings. Terrell. it might .S. and the value of the dollar appear to motivate banks in the expected directions. Less stable countries are more likely to have greater presences in the U. and Lowrey (1990) find that "U... and negatively related to the value of the yen in relation to the dollar. banking sector. Larger banking sectors in the foreign countries encourage greater investment in banking operations in the United States. holding other factors constant. is drawn to states with large bank market sizes and low levels of regulation of foreign banks. the greater the geographic or cultural distance of the source country from the United States. Commercial and industrial loans at these offices responded both to expansions in Japanese trade and to restraints on domestic Japanese interest rates. Dohner.S. 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. foreign bank data by country were only available for Japan.

This also relates to claims that lax regulation of banks by foreign governments provides a competitive advantage over U. and certain countries have earned large trade surpluses and are trying to find places to invest. foreign banks are alleg- edly more flexible in how they are compensated for loans. as well as other arguments. In addition. banks. In their lending activities foreigners are said to be more accommodating on paperwork and not as rigid on documentation as U. banks. The following section provides empirical evidence from the balance sheets of both foreign and domestic banks indicating the areas where foreign banks have the greatest competitive presence.13 It is widely claimed that foreign banks as new entrants have priced their products (particularly commercial and industrial loans) below domestic competitors in order to obtain business. The Competitive Impact of Foreign Banks Many observers have commented on the competitive impact of foreign banks. It has been argued that foreigners have been able to penetrate American banking because of greater funds availability. many have claimed that foreign banks have a number of competitive advantages over domestic banks because of regulatory factors and differences in economic factors in the home countries of the parent organizations. They allow the borrower to determine how the price of the loan will be paid by a com- bination of fees and balances.178 THE CHANGING MARKET IN FINANCIAL SERVICES be expected that future changes in the regulatory environment could also affect their growth. For example. They have been willing to accept smaller profit margins than their domestic competitors. ll Note that these arguments. The section concludes with an analysis of legal differences among states and the relationship to the doctrine of national treatment espoused by the International Banking Act of 1978. Zimmerman (1987) asserts that Japanese banks in California may be able to borrow at lower cost from their parent organizations because of greater funds availability. Excessive regulation in home markets has clearly stimulated . The major area of competitive impact is commercial and industrial loans. This section first assesses the arguments about the com- petitive advantages of foreign banks and then indicates the specific areas in which foreign banks have presumably provided competitive alterna- tives to domestic banks. are most applicable to the case of Japan. Most foreign countries have higher savings rates than the United States.12 Finally. and this activity is highlighted. It is claimed that they can do this because of lower capital requirements and a greater ability to utilize leverage.S.S.

THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 179 u.S. "Where spreads of seventy-five basis points over prime were once considered the absolute minimum. particularly with respect to banks owned by foreign governments. branches of foreign banks that occurred between 1980 and 1988. And although no objective rankings exist. for example. foreign banks.16 The evidence on capital levels is mixed. "Thus. The Japanese banks. 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. These purchases of existing loans have apparently become important since the mid-1980s and in 1988 accounted for 2. 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. government. banks enjoy from the U.' . Book values of Japanese banks. In order to attract new business in these loans.S. are offering lower interest rates. however. From table 5-3 it can be seen that foreign banks have a disproportionate amount of their assets concentrated in commercial and industrial loans. U.8 percentage point increase in the market share of U. the Japanese banks have more capital than do banks from any of the major countries. commercial banks. bank activity in the Eurodollar markets l4 and Japanese bank activity in Eurodollar and European markets. Much of the increase in C&I loans held by foreign banks has resulted from the sale of domestic C&I loans by U. Moreover. are earning a reputation for ruthlessness.. but implementation of equalized risk-based capital requirements will be difficult because of dif- fering accounting practices across countries. smaller foreign .IS but the evidence is less clear re- garding competitive relationships in the United States. Recent international efforts have been made to equalize capital requirements of banks across countries. Comments John Miller.17 This has presented domestic banks with a major competitive challenge. Baer (1990) claims that "fears regarding the competitive advantages conveyed by lax regu- lation at home may be justified.. notably the Japanese again. The major competitive impact of foreign banks has been in commercial and industrial loans."18 Generally. some banks. the foreign bank share of these loans has risen steadily since 1982.S.5 percent of total C&I loans. Whatever the particulars of the com- plaint. banks have been directly responsible for over two-fifths of the 5. and especially the Japanese.S.S. are very low compared to American banks. When market capitalization is considered.s. this concern would also appear to be valid where privately-owned foreign banks enjoy stronger guarantees from their governments than U.. have large gains on equity holdings that are not recog- nized on their balance sheets.

. 152 banks (56 percent) pointed out trade financing as one of the main areas of specialization in their U. They have moved away from serving only trade finance and large corporations toward serving small and mid-sized companies. the larger foreign banks are moving toward competing with domestic banks in a larger range of financial services. Tax laws may also have some influence on foreign banks' decisions regarding where to locate and what type of office to establish. and business development. 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.19 Other areas mentioned were money market activities. and the activity is heavily concentrated in the top three states. they have appeared to concentrate on nontransactions deposits.. only a limited number of states have foreign banking offices. The foreign banks have engaged in off-balance sheet activities such as support arrangements for commercial paper. retail banking. 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. They have had some success in the mergers and ac- quisitions advisory business. Increasing participation during the 1980s in real estate lending has not gone unnoticed. and geographic re- strictions to influence the activities of foreign banks within their borders. mostly with companies from their home country. Krishnan.S.180 THE CHANGING MARKET IN FINANCIAL SERVICES banks have been more interested in loan participations because they have less access to major corporate borrowers.. On the deposit side. The only way to assess the competitive impact of foreign banks in these areas is to examine empirical data.. "Many states employ reciprocity provi- sions. Increasingly. There is much anecdotal evidence of foreign bank participation in a number of banking activities. 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. offices. and Nigh (1987) have surveyed 271 foreign banks in the United States: " . 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. This uneven distribution is due largely to the moti- vation of foreign banks to concentrate in the types of business done in the . however.. Foreign banks participate heavily in finan- cial loans such as loans to depository institutions and real estate loans. Cho. asset maintenance or deposit requirements. foreign banks have differing opportunities in different states. because of varying state laws.

Trends in the growth of several balance sheet items are presented. offices of foreign banks and the weekly reporting banks.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 181 largest cities. 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. and rest of the world) and state (New York. These restrictions may affect the ability of U. Houpt qualifies his results because the number of banks was small and most of the acquisi- tions occurred since 1975. . Terrell and Key (1977) compare a number of balance sheet items for three dates (November 1972. depended more on purchased funds. Europe. but it is probably also somewhat affected by state laws prohibiting some or all types of foreign banking organizational forms. 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. The study finds that foreign banks have become more competitive with domestic banks in C&I loans from 1972 to 1979. A third study by Houpt (1980) examines the characteristics and per- formance of U. These studies provide some information on the activities engaged in by foreign banks. Balance Sheet Differences Between Foreign and Domestic Banks Several studies in the past have compared financial ratios of foreign banks to domestic banks. and Illinois). A report by the General Accounting Office (1979) examines foreign bank growth up to 1979. California. This article does this directly by comparing financial ratios from the most recent available balance sheets of both foreign.S. Note that agencies and branches are not treated in the study. November 1974.S.S. In order to assess the competitive impact of foreign banks in the United States.and domestic-owned banks participating in the American market. They also divide foreign banks by organizational form. From tables 5-2 and 5-3 it has already been indicated that foreign banks do playa significant role in the American banking market. and May 1977) between U. Using a paired-bank approach he finds that foreign- owned banks held considerably less state and municipal debt. it is first necessary to identify the activities in which foreign banks have participated. banks acquired by foreign banks and U. parent country (Japan. In the earliest study.s. banks to expand in Europe in the future as will be discussed in the last section of the paper. banks estab- lished de novo by foreign banks. Canada. and were less profitable.

each table (except for table 5-15) contains five columns: agencies. the ratios are multiplied by one hundred so that the results appear as percentages. The first method gives greater weight to larger organizations. in effect. Examination of the individual ratios reveals many extreme . The second method calculates the financial ratio for each institution and then finds the aver- age of these ratios. 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. The three main types of foreign banking organizations in the United States follow different regulations and differ in purpose. This method is used in table 5-8 and tables 5-10 through 5-15. Once again. some of the items on the balance sheet differed and judgment had to be used to make ratio comparisons meaningful. Changes in balance sheet defini- tions and data presentation make it difficult to obtain comparable data for early periods. Table 5-9 presents these results for all geographic regions for December 1989 so that the two methods of calculating ratios can be compared. 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 thirteen measures represent particular assets divided by total assets. while the last three represent individual liabilities divided by total assets. the banks with over $1 billion in assets have been separated from other domestic banks in the analysis. and large domestic banks (only domestic banks with greater than $1 billion in assets). therefore. the foreign banks have been separated into three categories in each table. This. all domestic banks. Consequently. branches. 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. Since the larger American banks differ from the smaller American banks in their business activities. Ratios using this method have been calculated for all divisions of banks but because of space constraints these results are not presented here. could mean that Citibank is given the same weight as a single office $10 million bank. Note that since the agency and branch data were obtained from a different data file than were the full commercial banks. The results are then multiplied by 100 so as to represent percent of assets.182 THE CHANGING MARKET IN FINANCIAL SERVICES Sixteen financial ratios are chosen for analysis and are presented in tables 5-8 through 5-15. while the second method gives equal weight to every organization. subsidiaries (commercial banks with greater than 50 percent foreign ownership). Data in tables 5-8 through 5-14 are from the December 1989 Call Reports.

..82 51.90 69.56 Other Securities 3.18 6.95 57.66 3.90 2.91 3.69 18.53 4.04 1.65 3.61 2.69 7.22 Foreign Govt Loans 4.88 0.27 0.63 77..94 11.16 0.18 0.45 3.74 Deposits 21.62 0.82 1.05 0.73 0.06 0.36 0. Agencies Branches Subsidiaries All Domestic Large Domestic Cash 9.20 22.12 0.46 2.30 72.55 0.92 .82 11.55 21.73 62.80 Bankers Acceptances 0.81 Other Loans 0.31 1.73 15. All States-Ratio of Sums-December 1989 (Values are percent of total assets).38 3.17 0.63 7.41 Financial Institutions Loans 14.51 64.25 11.81 3.64 4.51 8.65 Comm & Indus Loans 33.91 23.34 0.20 5..17 23.32 41.87 20.05 Commercial Paper 0.69 1.83 19.17 Total Loans 62.07 Real Estate Loans 8. Table 5-8.61 4.69 US Treas & Govt 0.54 9.98 3.51 5.89 Federal Funds Sold 3.08 0.36 3.56 2.20 Loans for Purchase of Securities 0.21 2.50 2.99 16.73 Other Liabilities 39.48 28.83 8..76 Federal Funds Purchased 10. 00 w .

60 Other Liabilities 25.62 16.74 7.06 US Treas & Govt 1.13 24.25 Federal Funds Purchased 5.43 18. Agencies Branches Subsidiaries All Domestic Large Domestic Cash 9.88 5.05 4.88 17.37 3.26 52.48 .15 1.11 9.54 9.53 76.79 0.32 1.03 0.24 Total Loans 60.04 Commercial Paper 0.08 19.18 Federal Funds Sold 4.97 50.07 21.55 Bankers Acceptances 0.20 Loans for Purchase of Securities 0.51 Other Securities 3.55 2.24 3.43 Deposits 27.92 2.16 0.96 4.41 87.56 8.05 2.47 7.15 71.87 8.11 Comm & Indus Loans 34.Table 5-9.10 Foreign Govt Loans 5.45 1.27 3.16 Real Estate Loans 9.31 0.33 0.62 42.28 0.58 Financial Institutions Loans 9.24 0.06 0.35 23.57 6.29 1.21 0.94 4.07 65.38 Other Loans 1.17 0.84 7.84 28.18 0.00 7.37 0.38 0.78 21.32 53.18 0.19 0.13 0.01 8.17 0.16 6.53 13.29 11. All States-Average of Ratios-December 1989.93 24.82 5.26 0.

42 22.09 Federal Funds Sold 24.21 Loans for Purchase of Securities 0.29 13.56 39.17 0.50 11.91 15.27 7.06 0.14 22.99 13.08 Foreign Govt Loans 19.46 3.79 4.12 0..49 7.92 1. Agencies Branches Subsidiaries All Domestic Large Domestic Cash 7.86 6.16 0.05 19.11 Real Estate Loans 4.90 5.16 0.59 21.40 17.00 0.52 2.17 0.94 5.09 0.73 2.00 Deposits 22.02 9.24 61.71 2.25 2.88 6. Table 5-10.15 5.65 29.78 4.31 US Treas & Govt 0.46 Financial Institutions Loans 10.72 4. 00 Vl .49 Total Loans 55.06 Commercial Paper 0.65 1.37 Federal Funds Purchased 25.48 6.86 56..09 3.37 5.28 3.55 6.68 Other Liabilities 29.59 ..00 0.76 Bankers Acceptances 0.80 65.23 3.59 24.02 Comm & Indus Loans 20.07 3.11 Other Loans 1.92 3.09 3.72 15..95 69.44 69.01 1.73 Other Securities 3.42 0.32 7.80 61.04 2.56 3.20 8.17 3.16 51.00 0. New York-Ratio of Sums-December 1989.

.29 0.36 0.35 2.00 0.29 3..50 0.28 2.16 0.02 35.57 .37 Other Loans 0.05 4.89 16.30 3.25 0..26 0.24 4.04 12.70 2.07 Comm & Indus Loans 34.91 Total Loans 63.03 3.43 0.84 Financial Institutions Loans 16.20 22.35 20.54 32.22 81.53 5.17 0.56 25.37 0.70 74.14 2.73 32. California-Ratio of Sums-December 1989.46 Federal Funds Purchased 11.89 72.04 1.05 Real Estate Loans 9.05 0.52 US Treas & Govt 0.51 11..55 72.41 1.25 11. Agencies Branches Subsidiaries All Domestic Large Domestic Cash 10.08 77.56 3.61 Deposits 14.17 1.52 11.33 0.18 Commercial Paper 0.00 5.14 3.58 3.90 Other Liabilities 52.30 Other Securities 4.50 0..74 Loans for Purchase of Securities 1.27 2.84 Federal Funds Sold 0.47 4.00 0.90 80.32 31.15 Bankers Acceptances 0.10 59.64 3.68 34.06 0.45 24.98 3. 00 0\ Table 5-11.69 12.03 0.48 Foreign Govt Loans 1.45 11..06 0.44 1.22 1.89 2.

13 0.62 US Treas & Govt 1.16 0.67 0. Table 5-12.. Agencies Branches Subsidiaries All Domestic Large Domestic Cash 29.77 Financial Institutions Loans 11.66 9.61 16.06 3...04 0..61 0.06 18.13 Commercial Paper 0.50 7.67 Federal Funds Sold 3.36 1.84 75.95 1.48 .05 Loans for Purchase of Securities 0.76 4.19 2.18 0.89 Comm & Indus Loans 22.57 Other Securities 4.05 0.43 1.09 13. 00 -.25 33.l .23 Foreign Govt Loans 2.64 13..63 Other Liabilities 19.28 7.82 Other Loans 0.71 7..85 79.20 Total Loans 41.51 4.25 2..40 10.85 2. Illinois-Ratio of Sums-December 1989.37 15.51 2.93 2.65 12.07 3.02 Real Estate Loans 4.49 3.77 9.30 2.15 0.09 19.94 5.79 5.70 Federal Funds Purchased 8.68 56.63 6.14 54.04 0.36 1.95 64.82 60.00 Deposits 51.03 25.51 Bankers Acceptances 0.

70 6.84 0.58 0.80 0.32 0.14 0.10 1.51 Other Loans 0.001 0.00 0.22 9.26 0.51 9.>-' 00 00 Table 5-13.35 36.64 4.35 4.87 65.19 4.14 1.50 0.31 Other Liabilities 5.06 3.99 8.99 0.02 0.63 .06 0.19 1.76 11.57 13.02 31.35 US Treas & Govt 0.12 1.002 Commercial Paper 0.16 Federal Funds Sold 4. Florida-Ratio of Sums-December 1989.81 11.13 Loans for Purchase of Securities 0.03 0.42 Federal Funds Purchased 1.57 Financial Institutions Loans 8.61 88.01 3.95 80.18 Comm & Indus Loans 10.001 Real Estate Loans 2.79 64.19 Deposits 88.90 1.26 8.00 0.23 0.84 11.11 Total Loans 27.56 Foreign Govt Loans 4.45 0.61 34.88 4.00 4.18 Other Securities 1.08 0.26 4.41 Bankers Acceptances 0.26 37. Agencies Branches Subsidiaries All Domestic Large Domestic Cash 16.98 82.

97 4.26 0.03 72.41 0.00 0.81 2. 00 \0 .10 Other Loans 0.15 9.75 12.13 50.63 1.37 4.64 Financial Institutions Loans 17.83 Federal Funds Sold 0.. Illinois and Florida Ratio of Sums-December 1989.76 1.07 0.24 11.83 11.43 3.95 15.95 Total Loans 96.04 2. Table 5-14.90 31.82 9.70 61.59 Deposits 0.85 26.01 78.81 1.32 0..07 0.84 Bankers Acceptances 0.90 0.98 4.82 64.06 Real Estate Loans 12.92 21.37 8.18 Other Liabilities 7.38 10.16 .55 Federal Funds Purchased 2.47 Other Securities 0.71 Comm & Indus Loans 55.40 3.04 7.57 70.24 Loans for Purchase of Securities 0.11 0. Agencies Branches Subsidiaries All Domestic Large Domestic Cash 0.00 0.44 10.76 56.00 0.19 0.03 Commercial Paper 0.37 0..28 3.04 0.37 0.18 3.15 Foreign Govt Loans 10.01 0.56 0.64 US Treas & Govt 0.13 23.68 6.00 0.15 0.39 4.85 2.17 72..77 4. All States Except New York.16 23.00 0.25 0. California.14 19.25 4.01 0.62 14.66 6.

71 Bankers Acceptances 4..77 Comm & Indus Loans 31..58 Other Securities 0.44 43.20 5.90 6.51 0.32 4.63 0.99 6.63 2.73 Loans for Purchase of Securities 0.46 US Treas & Govt 2.83 54.76 Deposits 7.02 2.66 Commercial Paper 0.00 66.35 1.32 Federal Funds Purchased 8.54 Federal Funds Sold 4.37 0.14 0..35 0.57 1.92 Foreign Govt Loans 4.06 1. ~ Table 5-15.. Agencies Branches Subsidiaries All Domestic Large Domestic Cash 7.57 16. All States-Ratio of Sums-December 1980.29 12.87 Financial Institution Loans 13.86 Total Loans 54.81 Other Liabilities 47.57 17.85 .84 30.42 Other Loans 0.64 2.20 26.74 19..19 1.09 Real Estate Loans 3.62 6.78 22.

The only data presented for 1980 are for all states.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 191 values. Subsidiaries. Comparisons across states indicate that foreign banks do not perform identically in different states.87 percent. Branches with 22. Agencies and branches. also hold more cash and less loans than domestic banks. excluding the four largest. then statistical tests of differences between categories can be conducted using the second method. Tables 5-10 through 5-13 present data from the four most important states for foreign banks in the United States.80 percent in C&I loans con- centrate on these loans more than the average bank. and to a lesser extent subsidiaries. loans to financial institutions.20 percent of assets in C&I loans. Commercial and industrial loans is the activity where it has generally been acknowledged that foreign banks have had the greatest competitive impact.17 percent and subsidiaries with 23. whereas agencies are similar to domestic banks. particularly for smaller banks. the use of the sec- ond method allows these extreme values to playa larger role.S. If we view these ratios as being a sample from a time continuum of different ratios. There is no clear pattern for differences between domestic and foreign institutions for loans for the purchase of securities. This appears to be especially true for agencies that have 33. In addition. however. commercial paper. Foreign banks. government securities than . 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. 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. Tables 5-8 and 5-9 present data from all states for December 1989. the standard deviations for most of the ratios are high as compared to the means. hold substantially lower portfolios of U. Table 5-14 indicates results for all states. the foreign subsidiaries also differ from the two categories of domestic banks in a number of financial ratios. and loans to foreign governments and institutions. Here it was decided to compare population means rather than to do statistical testing. 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.99 percent of assets in C&I loans also exceed the proportions in C&I loans of all domestic banks of 16. and particularly agencies and branches. the larger domestic banks with 20. Consequently. have placed greater emphasis in banker's acceptances. though not as cash-laden as branches. As ex- pected. Branches hold substantially more cash and less loans than the other categories.

than domestic banks. 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. in borrowing in the federal funds market and in obtaining funds from other liabilities (borrowing in money markets). 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. There are also some interesting comparative results on the liability side of the balance sheet. Note. There are distinct differences among . The results for cash are not as distinctive for branches using this methodology. 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. government securities than in table 5-8. Note. Foreign banks hold fewer other securities under the second methodology. and subsidiaries slightly less dependent. government securities are similar to before. however. however. The differences in ratios of total loans are reduced among categories of banks using this methodology. very similar to those presented in the article. that agencies and branches obtain much funding from their parent organizations and that this is not reported in the tables. Results by state using the second method have been obtained but are not presented here. but results for U. Agencies and branches are more active. because of legal restrictions. Most of the above-mentioned relationships hold when the second measure of financial ratios is used in table 5-9. Because smaller domestic banks have fewer loans and more government securities.192 THE CHANGING MARKET IN FINANCIAL SERVICES domestic banks. However. As expected. The major differences are highlighted here and the analysis by state is only reported using the first method. agencies obtain far fewer of their funds from deposits and credit balances than other categories of institutions.S. On the liability side there are fewer differences in results by using alter- native methodologies. These results are. Subsidiaries are close to domestic bank levels in this activity. many of which are retail oriented. The foreign banks still have less in real estate loans and more in loans to financial institutions and foreign govern- ments. Subsidiaries are more active (but to a lesser extent) in money market borrowing. Agencies and branches make fewer other loans. nevertheless. Branches are also considerably less de- pendent on deposits. Results for holding other securities and lending in the federal funds market vary by type of foreign institutions. there are some important differences due to the heavier weighting of smaller institutions in the second method.S. Results for other loans lose their distinctive pattern. in table 5-9 the domestic banks have a lower ratio of total loans and higher U.

The two rapidly growing states. The most noteworthy differences are discussed here. It should be noted that there are no agencies in Illinois nor branches in Florida. but the Florida foreign banks are not. but in Illinois the branches and subsidiaries do more C&I loans than all domestic banks. and Illinois engage in more C&I loans than the average for all states. Apparently. Correspondingly. The foreign banks in the residual states also appear to be making extensive efforts in real estate lending. All domestic Florida banks have higher deposit ratios than the national average. there are noticeable differ- ences among states for domestic banks. have large real estate sectors and their domestic banks participate heavily in real estate loans. but subsidiaries hold slightly more. California and Florida.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 193 states in the activity emphasis of both domestic and foreign institutions. In Florida. Finally. but not more than the large banks. Regarding commercial and industrial loans. On the liability side. In New York the foreign agencies and branches hold less in C&I loans than do their domestic counterparts. they also obtain less from other sources. In New York and Illinois the domestic banks are less involved in real estate. The ratio for Florida banks is substantially below the average for all states and the residual states are slightly below the average. Foreign subsidiaries in Florida have substantially higher cash ratios than any other banks and this may indicate services to customers with relatively high demand for cash. Agencies in the residual states have high loan ratios and extremely low cash ratios. branches. California. but the Florida foreign banks obtain even more funds from deposits than the domestic banks. however. A second activity with interesting differences among states is real estate loans. the Florida banks also have fewer loans. and subsidiaries are heavily concentrated in C&I loans. Agencies in New York have lent proportionately higher sums in the federal funds market. in the residual states the agencies and branches are more heavily invested in C&I loans than the average. the agencies and subsidiaries have emulated the domestic banks with low concentrations in C&I loans. Florida agencies and subsidiaries appear to greatly differ from banks in other states. The California foreign banks are also heavily in- volved in real estate. In California the foreign agencies. All banks and large domestic banks in New York. there is less effort by foreign banks to lend to the large corporations in Florida than in the other three states. The . California agencies and branches and the residual states agencies obtain substantially less funds from deposits. This may indicate that the foreign banks in New York have diver- sified into other activities and may be playing a stronger competitive role in these other areas than they were previously.

It is also illuminating to observe how the balance sheets of foreign banks have changed from 1980 to 1989. On the liabilities side. We have discussed the major differences among states. The proportion of C&I loans increased slightly for agencies and sub- sidiaries but decreased by over 8 percent for branches. It is clear. All categories of foreign banks have substantially increased their proportion of assets in loans backed by property. On the other hand. Although foreign banks have grown so rapidly during this period compared to domestic banks. we have no information on certain financial ratios for subsidi- aries at this time. though they are not discussed here. has increased their competitive impact in certain areas. nevertheless. along with their overall growth. Note that because of data comparability problems. that foreign banks be- have differently in different states and that their competitive impact will differ by state. they have. The main conclusion that can be reached from these two tables is that foreign banks have changed their activity mixes over the nine-year period and this. that the data requirements are very severe for most of these methods and some may be impossible to do at this point in time. there has been a large decrease in holdings of banker's acceptance for agencies and branches and commercial paper for branches. The agencies and branches have raised their holdings of other securities. however. Comparisons of tables 5-8 and 5-15 reveal some significant changes in the balance sheets of the foreign bank over the nine-year period. . It is beyond the capability of this type of analysis. but there are a number of other differences that perhaps deserve consideration. This change plus the overall growth of foreign banks explains the observations mentioned earlier about the increased competitive impact of foreign banks in real estate loans. several ways to measure this impact are suggested. to be able to measure the degree of the competitive impact of foreign banks. been able to provide extra competitive pressure upon domestic banks in the important activity of commercial and industrial loans.194 THE CHANGING MARKET IN FINANCIAL SERVICES California agencies and branches are more heavily involved in money market borrowing. Note. Later. The ratio analysis above identifies activities that foreign banks have pursued and gives some indications about where the competitive impact of foreign banks can be found. The activity that appears to exhibit the greatest change is real estate loans. agencies and branches have increased their reliance on deposits. Loans to financial institutions have decreased for branches and subsidiaries and loans to foreign governments have decreased substantially for all categories. and beyond the scope of the article.

which are difficult to quantify. the pricing of loans and deposits involves other factors besides interest rates. Profitability measures are subject to distortion because of accounting procedures. There is still much work to be done to measure the competitive impact of foreign banks. or C&I loans. They have apparently affected the behavior of domestic banks. due to competitive pressure by different organizational forms of foreign banks. we would expect subsidiaries to have a larger impact on retail banking and agencies a relatively larger impact on wholesale banking. must be taken into account. It is difficult to measure their competitive efforts. and may be changing as the interest rates change. The change in shares can then be related to performance variables. It may be difficult to isolate the effects of foreign bank competition because of the presence of other factors changing in the economy at the same time. in this type of analysis one has to be careful about other factors which are affecting the behavior of banks. Conclusions and Prospects for the Future Foreign banks are playing an increasingly important role in the American financial marketplace. If foreign banks do have a competitive impact. As an alternative to examining the effect on various measures of profitability. They are providing services in many other areas and have expanded their efforts in several areas. deposits. one might expect domestic banks to emulate what foreign banks have done. One could alternatively examine changes in the balance sheet of do- mestic banks after foreign bank entry and growth. There are several serious difficulties with this type of analysis. but the anec- dotal and empirical evidence available indicates a definite competitive impact. The analysis has indicated the balance sheet areas that the different categories of foreign banks have emphasized and shown differences among . For example. the differential effect on domestic banks. and this is not done in the article. In addition. Domestic banks may become less profitable after foreign bank growth due to added competitive pressure. one could look directly at the impact on loan and deposit pricing. The measurement of the competitive impact of foreign banks is important to the government in formulating policies to regulate foreign banks. such as real estate loans. Finally. Pricing data are difficult to obtain. especially in the area of commercial and industrial loans. However.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 195 Market shares of foreign banks in different markets (local markets or states) can be calculated over time using either assets.

The structure of banking markets in Europe should. proposed legal changes in Europe may affect American banks in Europe. There could also be important effects upon the ability of banks from non-EC countries. Accusations that for- eign banks have competitive advantages over domestic banks have been made but. would be subject to . legislation may be proposed to restrict foreign banks. to operate in these markets. In 1960 the EC adopted the First Banking Directive that argued for financial integration but retained national control. The European Community (EC) is scheduled in 1992 to remove many commercial barriers between countries. Nevertheless. and this in turn could lead to changes in the treatment of foreign banks in the United States. 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. there does not appear to be solid evidence to support these suggestions. In addition. including restrictions on commercial banks operating in other EC countries. therefore. as discussed earlier. thus. the EC has been moving to eliminate commercial restrictions among countries. would be allowed to branch throughout the Community without obtaining additional licenses. change dra- matically. though. Subsidiaries of non-EC banking organizations chartered in an EC country would be considered to be EC companies regardless of ownership and. 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. 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. 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 resulting structural changes could impact significantly upon the availability and nature of banking services in these countries. Future entry. particularly American and Japanese banks. 1993. The directive is intended to be implemented before January 1. there are certain factors that could retard this development. Though foreign banks in the United States should be playing a greater role in the future.196 THE CHANGING MARKET IN FINANCIAL SERVICES geographic areas. Several of these variables are likely to change in the future in a direction that will further encourage the growth of foreign banks. Foreign banks have grown rapidly both in absolute size and in relation to domestic banks. therefore. Since the Treaty of Rome in 1957. The directive deals with the treatment of non-EC banking organiza- tions.

THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 197 reciprocity conditions. 3. See Table 3. Ibid. If Europe is treated as one entity-then the treatment appears to be reciprocal. Approval would depend upon whether the other country granted equal opportunities to Ee banks in that country as granted to domestic banks. Ibid. Goldberg and Saunders (1981a.1990. in this article. 60). and Nigh (1987. this still produced some inequities from the European point of view. Notes 1. 6. American Banker (February 27.1109-110). 9. but European banks were restricted to full banking operations in only one state. (64). 11. that is. Retaliatory measures against European restrictions on American banks could reduce the competitive importance of foreign banks. and Lowrey (1990.. Houpt (1988. 18A). The Second Banking Directive explicitly provides for reciprocal treatment of banks from non-Ee coun- tries. 7. Terrell. Ibid. However.. . The United States faced this problem in 1978 and opted for a policy of treating foreign banks on a national treatment basis. 5. footnote (c). 2. However. The major banks in Europe already have nationwide banking organi- zations within their home countries. Ameri- can banks were not restricted within individual European countries geo- graphically in most cases. This might provide a barrier to foreign entry or might serve as a bargaining tool to open other markets to Ee banks. On the other hand.. 48-49). Cho. 12. Grosse and Goldberg (1991. This is argued by Nadler (1989. that is. The Ee must assess whether the American treatment is reciprocal. greater liberalization of the treatment of foreign banks could enhance their ability to provide mean- ingful competition to American banks. except for those already present by 1978. there are still restrictions on interstate banking in the United States. 10. 8. 4. 372). 11) and additional updated tables provided to the author by James Houpt. all foreign banks were accorded equal treatment with domestic banks. 32). though the situation has been greatly liberalized in the last decade. 5). (25) and updated material supplied by James Houpt. Dohner. Krishnan. Ibid. 7. The interaction of legal changes could affect the ability of foreign banks to compete in the United States. though. but if each country is treated separately-it is not. whether genuine national treatment was granted to Ee banks. Goldberg and Saunders (1981b. (60).

. Herbert L. and Robert Grosse. Baer (1990. Banks. Baer. Krishnan. 15. "Foreign Bank Activity in the United States: An Analysis by Country of Origin." Staff Study 156. Nadler (1989) provides an extensive discussion of how this practice works and the implications for domestic banks. 1987. and Douglas Nigh.C. 25). "International Trends for United States Banks and Banking Markets. "The Determinants of Foreign Banking Activity in the United States. 1990. 1991. Lawrence G. Robert." Journal of Banking and Finance 5. (July): 383-96. Credit and Banking 13. Goldberg." Staff Study 109. "Foreign Competition in U. 1987. 1990. General Accounting Office. (May).." Report by the Comptroller General of the United States. 19. Lawrence G. 4 (July/August): 339-49. and Lawrence G. Kang Rae. Goldberg. 16. Grosse.1988. 1980. 13. "Factors Affecting the Foreign Banking Presence in the United States. Economics Perspectives 12. and Lowrey (1989). Baer (1990. Herbert L. (August)." Federal Reserve Bank of Chicago. 89). 69). Suresh Krishnan. Cho." (March) Mimeographed. "The Growth of Organizational Forms of Foreign Banks in the United States. "Distribution by State of Foreign Bank Activity in the United States. . 18. D. "Foreign Ownership and the Performance of U. Banking Markets. and Christine A. Washington.. 1981a. Charles W.S. Board of Governors of the Federal Reserve System. and Anthony Saunders.C. and Randolph McGee. 2 (March/April): 3-15. Hultman (1987." Banking Law Journal 104." International Journal of Bank Marketing 5. "The Foreign Banking Presence: Some Cost-Benefit Factors. 3." Journal of Money. Jedlicka and Tobin (1988. "The State of Foreign Banking Presence in the United States. "Considerable Increase in Foreign Banking in the United States Since 1972. 348).S.198 THE CHANGING MARKET IN FINANCIAL SERVICES 13. 6 (December): 1093-112. and Nigh (1987. "Does Regulation Drive Innovation?" Federal Reserve Bank of Chicago." Journal of Banking and Finance. See Terrell. Houpt. Goldberg. Hultman. 25). Board of Governors of the Federal Reserve System." Journal of Banking and Finance 15. 14. --. References Baer. Cho.. Dohner. 3 (August): 365-374. 1989. Charles W. Washington. (July).. 1981b.2: 59-75. . 1 (March): 17-32. 20. Pavel. 17. See Baer and Pavel (1987). Hultman. D. 1988. 1979. Economics Perspectives 14 (3) (May/June): 22- 29. James V.

1989. (November). Nadler. Paul. No. Terrell. Activities of Japanese Banks: 1980-1988. . "The United States and U. "Balances and Buggy Whips in Loan Pricing. Gary C. International Finance Discussion Papers 361 (September).THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 199 Jedlicka." Board of Governors of the Federal Reserve System. Henry S.. Zimmerman." Federal Reserve Bank of San Francisco.. Henry S. Lowrey. and Mary Tobin. 1980-1988. 1977. John." Euromoney (November): 89-93. "Foreigners Ferocious Financial Threat. 1990. "The Growing Presence of Japanese Banks in California. . 1988. and Barbara R. Robert S. "The Activities of Japanese Banks in the United Kingdom and in the United States.. Terrell. 1989.K. 1989. 6 (February): 4-9. International Finance Discussion Papers. "The United States Activities of Foreign Banks: An Analytical Survey. Economic Review (Summer): 3-17. Dohner. Key. and Sydney J.. 113." Board of Governors of the Federal Reserve System." Journal of Commercial Bank Lending 72." Federal Reserve Bulletin 76(2) (February): 39-50.

He uses these issues to analyze the competitiveness of domestic banks and foreign banks. 201 . And fortunately. economy: foreign investment. foreign bank advantages.S. Clearly. COMMENTARY Gary C. banking markets. foreign- owned banks issue about 60 percent of all standby letters of credit issued by banking institutions in the state. foreign bank presence is even greater. foreign bank control.s. market. in California $1 out of every $3 in banks' outstanding business loans are held by Japanese-owned banking institutions. especially the major Japanese-owned banks operating in the United States. Foreign bank presence is frequently discussed in the press. foreign bank funding-and examines them in the context of the rapid growth of foreign banks in the u.S. Again. It deals with important issues facing the U. this is a subject that is of interest to many of us. These measures highlight foreign bank penetration of the U. commercial lending or activity in the "guarantee" markets. In the guarantee market. For example. in California. Zimmerman Goldberg's article is a timely one. we have some measurable ways of examining their presence using asset size.

Featured prominently on that list are foreign trade activity. and price/earnings ratios across countries. Rapid Growth The article cites some dramatic figures on foreign bank growth in the United States. This is a process that is familiar. they have established banking relationships.202 THE CHANGING MARKET IN FINANCIAL SERVICES Review Goldberg's article provides us with a good background on the growth of foreign banks in the United States. and on their varied organizational forms. direct foreign investment.-owned banks. foreign-owned banks are likely to have a competitive advantage in dealing with these "home country" customers.S. The article then moves into a review of the literature and empirical studies that attempt to explain the growth of foreign banks in the United States. the article would be more useful if it reported bank growth for three categories of banks. This latter information is helpful to our understanding of the scope of their operations. the figures are somewhat misleading because virtually all of the foreign bank expansion in the United States has been by Japanese-owned banks.S. and they speak their language. exchange rates. However. It is not surprising that overseas. to which Goldberg has been a significant contributor. In addition. and other foreign-owned (excluding the Japanese). elaborates on a number of factors influencing foreign bank activities. A generation or more ago many major U. U. corporations often bank with U. market. based on country of ownership: United States-owned. their competitive impacts. . Japanese-owned. Japanese banks in particular have followed Japanese multi- national firms to the U. There is strong evidence that foreign banks follow trade and invest- ment activities. they understand their customs.S. or that Japanese multinational firms' overseas operations bank with Japanese-owned banks. The literature. because at this time foreign-owned banks tend to be most active in several major banking markets and their organizational form and business emphasis can vary dramatically from state to state. on their status and history. which is useful. inter- est rate differentials across countries. domestic economic conditions. the article provides informa- tion on the foreign bank presence by states. regulatory differences. Thus. and it is this economic process that I believe should be emphasized. and their possible advantages. banks expanded their overseas activities and portfolios as their major corporate customers invested abroad.S. Clearly.

which is a good way to handle the analysis. .4 percent to 15. Competitiveness The descriptive section of the article provides a series of brief explana- tions and stylized facts concerning the competitiveness issue. business loans at Japanese-owned banks have grown at over a 30 percent annual rate over this period. branches and agencies have expanded at close to a 25 percent annual rate. and it is a market in which they playa major role. about 5 percent per year. non-Japanese foreign-owned banks.1 percent). All three types of Japanese-owned banking institutions grew rapidly during the period.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 203 Growth rates for the various groups illustrate this point. the growth pattern indicates that the focus should be on the exceptional growth of J apanese- owned banking institutions in the United States. After describing the rapid foreign bank growth and the reasons behind it. Commercial lending is an important area for foreign-owned banks. While it hits on important issues (such as differential capital requirements across countries. less restrictive regulation at home. potential overseas funding sources. lower cost of funds. and willingness to offer lower loan rates and accept smaller margins). One criticism of this section is that it again needs to focus on the Japanese-owned banking institutions. it does not provide the reader with a clear sense of what the competitive advan- tages might be or how significant they are. Yet. Japanese-owned banks. Commercial loans have grown at about a 1 percent annual rate since 1985 for both domestic banks and non-Japanese foreign-owned banks. by focusing on the rapid growth of Japanese-owned banks we may gain more insight into the competitiveness issue. less paperwork and documentation. rather than the broad grouping of all foreign banks. Moreover. Over the period from year-end 1985 to mid-year 1990. Again. In contrast. They account for the exceptional growth of foreign banks' commercial lending in the United States. Clearly. branches and agencies as a group have grown at about the same average rate as domestic banks. Goldberg presents a descriptive section on the evidence of competitive effects and on the allegations of "unfair" competition from foreign banks. they captured nearly a 9 percent increase in market share over the same period (from 6. In this section he makes his point using developments in the commercial lending market.

In fact. this trend could have im- portant implications for the continued growth of Japanese-owned banks in the United States. Henry Terrell (1990) in testimony before Congress provided evidence that the growth of Japanese banks in the United States was NOT funded by raising low cost funds in Japan and lending them through their affiliates in the United States. Thus. Faced with a similar constraint on retail deposits. banks substituted wholesale borrowings (increasing their reliance on large certificates of deposit.S.d the allegation that foreign-owned banks have a lower cost of funds. Thus.S. 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.204 THE CHANGING MARKET IN FINANCIAL SERVICES Moreover. Terrell suggested that net sources of funds data indicate that just the opposite pattern took place. 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. over the last five years Japanese banks overseas have been net lenders to Japan. the situation facing Japanese banks in their home market ap- pears to be much like the situation U. before deregulation in the United States. . 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. he presents evidence that in the aggregate. other money market borrowings. Of course. especially with respect to the funds availability question ap. banks faced when they were constrained by Regulation Q interest rate ceilings on deposits. and borrowing from their overseas affiliates) to make up for retail deposit shortfalls caused by below-market rates on retail accounts.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. 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. 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. Terrell stated: Until recently. banks and from affiliates in London and other offshore markets. there may be some stronger evidence that there are no significant advantages. For example.

Thus. In this case it was possible to examine the cost of funds for interest-bearing deposits and for other borrowings (which would include borrowings from parents). I found that Japanese-owned subsidi- ary banks generally had very small gross borrowings from their parents. the California evidence does not support either the assertion that Japanese-owned banks are funded by their overseas parents and/or that they have a significant cost of funds advantage over domestic banks. Thus. although the sheer volume of information and tables may be a bit overwhelming. Incorporating this work with the research by Terrell would provide empirical evidence that Japanese-owned banks likely do not have a signifi- cant advantage over domestic banks in terms of access to below market- rate funds. The results showed no statistical evidence from the cross-sectional regressions that Japanese-owned banks had a lower cost of funds for either total interest-bearing deposits or for other borrowings. and the parents were willing to provide funds at below-market rates. however. there are some sizeable differences between the ratios of various assets and liabilities to total assets. The next question is whether Japanese-owned bank subsidiaries actually reported a lower cost of funds than domestic banks.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 205 California In my own work of studying the role of Japanese-owned banks in Cali- fornia. . it would still represent only a small share of their total funds. it is hard to judge their importance. and this section highlighted those differences. but we have no idea whether differences between foreign and domestic banks are statistically significant or not. even if they were being funded by the parent organi- zations. For example. The aggregate financial ratios presented in the tables do provide some evidence of differences between foreign-owned banks and domestic banks. Financial Ratios Goldberg's section on balance sheet differences moves on to a compari- son of financial ratios for foreign banks and domestic banks. on average less than 5 percent of their assets were funded by borrowings from parents. There are some major differences in the activity and organization of foreign-owned banking institutions. I also looked at the level of gross borrowing by foreign-owned banks from their overseas parents. I found this section especially useful from a regional perspective.

and these differences also were significant. The other holds the view that . I'd suggest that as an area for additional research. are important competitors for domestic banks. Japanese. controlling for differences in bank size. branching structure and other bank specific factors. Still. On average. at the individual bank level. In spite of the statistically significant differences. while controlling for ownership by groups (do- mestic.206 THE CHANGING MARKET IN FINANCIAL SERVICES A stronger case could be made by examining empirical evidence from a cross-section of banks. and those differ- ences were statistically significant. Japanese-owned banks had lower ratios of both total deposits and retail deposits to assets. other foreign-owned banks. The results suggested that as a group there were some statistically significant differences among Japanese-owned banks. Thus. Moreover. those differences were not statistically significant. however. the California data provide stronger evidence that there are some important portfolio differences between foreign-owned banks (especially Japanese) and domestic banks. and in fact. most foreign-owned banks have a large share of retail deposits. In my own research on the foreign bank presence in California. in California. Conversely. Japanese-owned banks also reported more reliance on wholesale funding. Japanese-owned banks had higher ratios of commercial and industrial loans to assets. and total loans to assets (reflecting their strong commercial presence). Further research is necessary if we are to get beyond the two schools of thought expounded by many California bankers. like other banks in the state. they tend to behave and hold portfolios that resemble other retail banks in the state. or other foreign). One school suggests that foreign-owned banks. Japanese-owned banks in California also have an important presence in the middle market lending area. and domestic institutions. I ran regressions on cross-sections of Cali- fornia banks. they had much higher ratios for both large time deposits and money market borrowings to assets. commercial real estate lending to assets. As well as providing retail banking services at the branch level. and suggests another empirical avenue for examining the financial ratio data. especially the Japanese-owned institutions. it is important to note that Japanese-owned banks and other foreign banks still are primarily funded by deposits raised in domestic markets. In other areas (total real estate lending and consumer lending) there were differences of several percentage points in the aggregate loan-to-asset ratios. including ownership status. there are some statistically significant differences that allow us to make a more forceful statement about the portfolio differences of foreign- owned banks and their effect on competition.

However. Nationally. California. rather than foreign banks in general. over the last nine years average return on assets (ROA) for foreign banks was only about half that for domestic institutions. Nationally.5 percent of their market share as measured by assets and commercial loans. Goldberg also suggested that we may wish to look at profitability to see if profit levels varied with foreign ownership. In this case I did a quick examination of industry profits by the three ownership groups (domestic. One way to measure whether foreign banks are important competitors.S. Japanese-owned. Market Share Herb Baer of the Federal Reserve Bank of Chicago recently made an interesting observation on the foreign bank presence in the United States. Writing in the May/June 1990 edition of Economic Perspectives. in the area of market share it is the Japanese-owned institutions. Japanese-owned banks added 4. including other banks. there was a rapid growth of Japanese-owned banks' market share.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 207 foreign-owned banks are just another group of competitors among many. domestic banks lost 3. that have been building market share. Over the same period. he noted that after leather goods. I followed up on this suggestion. wholesale banking has the next highest foreign- owned market share of major U. lower profits also would be consistent with the suggestion that foreign banks are using "low margins" to build market share at the . and other foreign-owned) for several geographic regions (the United States. Goldberg suggests that we might look at market share data to help evaluate the competitive impact of foreign-owned banks. Between 1985 and 1990.7 percent and 7. and if their presence is becoming more important. and Illinois). Clearly. thrifts. Neither Japanese-owned banks nor other foreign-owned banks ever report a higher ROA than domestic banks. Other foreign-owned banks lost market share in both categories as well. and include here some of my observations. In contrast. California appears to be the exception to the national trend. New York. foreign banks' market share has risen sharply. Of course. it was offset by reductions in other foreign-owned share.1 percent of commercial loans. is to examine industry market shares. nonbank financial intermediaries and out- of-state banks.4 percent to their market share for assets and 9. so that total foreign ownership generally held in the 31 percent range for assets. industries. respectively.

Goldberg's article points us in the right direction. especially if there was evidence that foreign banks offered lower rates on "similar" loans. depending upon ownership.S. that may impact foreign bank activity in the United States. in Europe and in Japan. one could test to see if significant differences in interest rates exist across individual banks. There may be an interesting story here. I would suggest extending the research and the outlook to include a number of changes currently going on in the world. to uniform international capital standards for banks. Using the Federal Reserve's individual bank data on commercial and agricultural loan rates. especially the Japanese-owned banks. it may still be possible to test whether foreign banks price loans differently than domestic banks. 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.S. to banks in Japan losing their traditional customers to the capital markets (much as it has . since Japanese banks currently hold over 10 percent of the U. banking system. will their expansion continue or not? His article provides a good perspective on the inroads of foreign banks. It may be possible to examine this competitive pricing issue. While this research may have a few unavoidable problems. and yet their market share has not increased significantly since 1988. 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. that is. he takes us to what I consider to be the "big" question: What will the future bring for foreign-owned banks. Many of the fac- tors noted in his article and my comments. like a relatively limited sample of foreign- owned banks in the reporting panel.208 THE CHANGING MARKET IN FINANCIAL SERVICES expense of current profits. and on the importance they now playas competitors in the U. although it clearly is beyond the scope of this commentary. bank market share (over 25 percent in California). to the significant decline in the Japanese stock market and its reduction of Japanese banks' "hidden reserves" and capital positions. Again. 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.

or even a pause in the dramatic growth of Japanese banks in the United States.S. A slowdown. market. will clearly slow the pace of foreign bank growth in the U. to a new interest in consumer banking in Japan-all suggest that the growth of Japanese-owned banks in the United States will slow down.THE COMPETITIVE IMPACT OF FOREIGN COMMERCIAL BANKS 209 happened in the United States). .

Phillips-Patrick Introduction Global integration of financial markets and the internationalization of the modern corporation are indicators of the openness of the world economy. Economists argue that an open world economy increases economic effi- ciency by enabling resources to flow to their most productive uses. In the second section. The first section discusses the major developments affecting the domestic underwriting business in the 1980s. we define and document the extent of 211 . 6 THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS IN THE UNITED STATES Robert Nachtmann Frederick J. facilitating the transfer of corporate control. and enabling risk freturn decomposition through financial product innovation. The investment bank not only contributes to the process of international integration. It identifies those firms that service global underwriting activity and attempts to evaluate the impact of foreign underwriters on the domestic underwriting industry in the United States. servicing the rebalancing of investor portfolios. This article investigates one business segment of the investment bank- ing firm-the underwriting of corporate securities. The article is organized as follows. In- vestment banking firms contribute to the enhanced efficiency of a global economy by intermediating the process of capital formation. It is also affected by it.

below investment-grade bonds (junk bonds). Significant Developments in the 1980s Domestic broker/dealers and investment bankers witnessed dramatic changes in both their business and their business environment during the turbulent 1980s. New Product Development Investment banks created or developed several new financial products during the 19808 that have grown exponentially. the new issue volume in mortgage-backed securities was $0. we summarize our results and discuss directions for future research. in section five.5 . so has the competition for such business. In 1980. and interest and currency swaps. For the new issue market. Eight years of economic expansion. so have foreign firms entered the domestic market. Finally. But as domestic firms have expanded overseas. New products were introduced that now represent a sig- nificant portion of the market. produced an unprecedented growth in new issues being brought to the market.1 In the third section. This domestic growth has been matched and exceeded in some foreign capital markets. accom- panied with the dramatic growth in corporate consolidations and the substitution of debt for equity in corporate balance sheets. Thus. Shelf registration and Rule 144a changed the underwriting environment.212 THE CHANGING MARKET IN FINANCIAL SERVICES the market and identify underwriters for worldwide corporate offerings. This section briefly reviews some of the significant developments in the under- writing business in the 1980s. The fourth section measures the impact of foreign entry and acquisitions of domestic rivals on the equity values of publicly traded domestic underwriters. while opportunities for new business have expanded. Both of these developments appear to be a way for assets traditionally held by banks to become securitized. This section also provides evidence concerning performance and pricing for the domestic and foreign segments of the underwriting industry. Among them are mortgage- backed and asset-collateralized securities. we examine underwriting performance as a function of whether the underwriting firm is domestic or foreign. junk bonds and mortgaged-backed securities have played an especially large role.

far below the average of the early years. an incentive for invest- ment banks to expand beyond their own domestic horizons. in 1980. were offered. with growth in IPOs far outdistancing secondary offer- ings. in 1989. Rule 415 allows firms to register their offerings with the Securities and Exchange Commission before they offer them to the market. well over 20 percent of the new issue market. Last year.23 billion in 1985 to $23. Below investment-grade bonds have also experienced dramatic growth during the early and middle parts of the decade. only 20 percent was syndicated. growth in other types of collateralized securities has been equally explosive. In 1986 and 1987. about $25 billion. For example. One of the largest changes was the adoption of Rule 415 in 1982. for 1989. in 1989. Similarly. the new issue market rebounded. equity offerings represented only about 7. it was $115. representing more than a third of the value of all new issues in 1989. 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. the annual new issue volume of junk bonds was about $1 billion.5 billion. according to Investment Dealer's Digest. Likewise. Changes in the Business Environment The financial innovations and the rapid domestic growth of the 1980s expanded domestic markets. equity offerings dropped significantly. a substantial portion of which was just one issue. changes in the regulatory environment have altered the nature of the underwriting business.80 percent of new issue debt was syndicated. only to fall off over the last two years. After the market crash of 1987. The equity underwriting market collapsed in 1984. expansion of business activity abroad brought additional demands for financial services. off by more than two-thirds from the previous year. The impact of these two products can be seen in the new issue market by their effect on the proportion of the equity offerings as a percent of all offerings. equity offerings averaged about $30 billion. equity offerings surged. In 1980. increasing from $1. These competing demands appear to have led to substantial growth among a few of the largest firms. about 18 percent of the total new issue market. In 1983. The effect of the rule is to permit firms more flexibility in timing .3 billion in 1989.3 percent of the new issue market. From 1980 through 1982. It hit a high in 1986 at about $32 billion in new offerings.2 Since 1985.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 213 billion.

One an- ticipated result of Rule 144a is the increased trading of foreign securities within the United States. Whether foreign entry into the U.214 THE CHANGING MARKET IN FINANCIAL SERVICES the market for their new issues. market signals increased com- petition for other broker/dealers or increased expansion opportunities for both domestic and foreign business is an empirical issue. we describe the depth and the extent of the market. or about 17 percent. Throughout the decade of the 1980s there were about 550 completed combinations involving domestic dealer/brokers. perhaps because they lack sufficient size to bring a full issue to market promptly. Foreign Competition In 1978 Credit Suisse first formed a domestic investment banking venture with First Boston. but it marked the beginning of rapid change and consolidation in the domestic underwriting market. the SEC adopted rule 144a. that is. underwriters are under pressure to do so expedi- tiously. In April 1990. which permits the trading of unregistered securities among "sophisticated" investors. some ninety- four.S. Likewise.S. involved foreign acquire. U. In the next section. Mergers and acquisitions were not unidirectional. For example. The effect of foreign acquisitions on the market values of other domestic broker/dealers is considered in section four. Of these 550 deals. The highest percentage concentration of foreign acquisitions took place during the period 1985 to 1987. This was not the beginning of foreign investment in domestic underwriters. repre- senting over $36 billion in reported value. with twenty-eight of the total ninety-four deals occurring in 1989 alone. 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. One effect of Rule 415 has been to reduce the role of regional underwriters in the new issue market. according to IDD Information Services. over the period 1985 to 1989. those with more than $100 million in securities under management. . domestic firms completed forty-seven deals involving foreign target broker/dealers. But once the decision has been made to bring the issue to market. although the highest number of acquisitions has occurred more recently. Recently. firms have started to acquire foreign broker/dealers.

Firm-commitment underwriting. 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. acquired by them through negotiated deliberations. The sale may be conducted as either a public or private transaction. Private transactions are not considered. the broker/dealer simple agrees to extend its "best effort" to the task of selling the offered securities at the public offer price. Best efforts underwriting does not entail the offering's purchase by the broker/dealer. we are only concerned with those financial contracts that identify the transfer of capital to corporations. Indirect public offerings form the base set of the target market for underwriting services. and publicly offered on a firm- commitment basis. 4 Competitive deals result from sealed . Only firm- commitment offerings are analyzed. Representative private transactions are the private placement of securities and loan contracts. Corporations source external capital through the sale of financial contracts. 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.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 215 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. include the taxable issues of corporate agencies that are often described as quasi-governmental organizations. Instead. however. incorporates a distinct element of insurability whose value would be incorporated in the discounted price to the issuer from the public offer price. Indirect offerings employ the services of broker/dealers on either a "best efforts" or a "firm-commitment" basis. We do not therefore address government or municipality financing for their own account. We do. First. therefore. Issuers identify firm-commitment underwriters either through a nego- tiated or a competitive bid process.3 A brief explanation of these defining characteristics follows.

3 trillion. Issue volume increased almost threefold (growth factor 2. All data were collected from IDD Information Services' Do- mestic and International New Issue Database. domestic debt. The typical firm commitment offering is serviced by an underwriting syndicate under the control of a lead manager. Totals (A) and (B) in table 6-1 are provided to account for the staging of international debt. Overall.81). preferred stock. Common. The valuation of issues entering into the volume totals is consistent with the IDD's yearly rankings for underwriting business done by investment banking firms. and international debt. convertible debt. Industrial firms were the second largest issuing group--$0. If the cost to issuers in .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. Table 6-1 lays out the depth of the market over time and across secur- ity types. The primary issuing group was financial firms-$1. Common and preferred issues are valued at share price times the number of shares distributed.9 trillion worth of corporate securities over the period.4 and $0. The most dramatic period of growth was 1983 through 1985. debt offerings are reported separately. Since information on interna- tional debt offerings is not available until 1982. convertible preferred.6 Extent of the Market The market for worldwide corporate offerings grew at an annual rate of approximately 20 percent during the 1980s. Public offerings have increased not only in the aggregate but also by security type. We classify separately the following types of offerings: common stock. re- spectively. Corporate agen- cies and utilities accounted for approximately $0. Only isolated incidents can be identified where the volume of any issue in a given year was not higher than its previous year's level. Debt issues are valued at net proceeds to underwriter com- puted as principal times price. preferred.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.4 trillion or 48 percent of all corporate issues. and convertible issues combine both domestic and international offerings. This growth was fueled primarily by the rate of growth in corporate debt issues.9 trillion. The negotiated deal is the primary focus of this paper. the underwriting market accounted for the distribution of $2.

91 20.01 19.37 15. b Domestic.91 571.54 272.61 12.95 103.41 151. straight debt valued at offer price.89 (6) Int'l Debt" n.68 509.82 4.S.46 0.82 2.10 56.47 150.90 38.64 7.47 7.52 • Common stock offerings both domestic and foreign excluding debt for equity swaps and rights.13 205.20 29.35 95.37 76.38 4. I Public underwritten offerings includes firm commitments only and excludes private placements.39 42. 62.52 5.35 4.a.40 14.48 9.61 (4) Convertible Debt d 4. e International debt includes Eurobond and non-Yankee foreign bond issues.29 0.60 65. dollar equivalents at the time of issuance. expressed in U.95 317.51 3.79 (5) Convertible Prefd 1.14 3.93 231.14 (B) Total (1-6) 58. .16 27.S.32 97.00 326.58 (A) Total (1-5) 58. d Convertible offerings include both domestic and foreign issues valued at offer price.18 37. Trends in Public Offerings of Underwritten Corporate Securities 1 (1980-1989).09 35.75 219.12 6.48 234.$ Billions.26 165.10 56.83 (3) Preferred Stocke 1. 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 (1) Common Stock· 13.18 4.13 522.05 65.76 210.00 11.01 15.Table 6-1. C Preferred stock both domestic and foreign valued at offer price.83 185.44 14.30 1.15 0.30 155.89 302.31 62.87 40. U.08 (2) Domestic Debt b 37.38 481.43 207.18 301.47 90.89 340.60 128.47 14.74 46.a.89 7.65 6. n.33 63.07 1.90 1.

The classification procedure was performed in the following way. counterparts. Table 6-2 provides information about the growth of registered broker/dealers as well as equity and asset . the period beginning in 1985 and proceeding through 1989 should result in a reduction in issuers' cost. (2) perhaps more importantly. The trend data in table 6-1 suggest that this cost effect is most compelling in the debt markets. the insurance premium set by the underwriting should decrease given the larger pool of transactions. Growth of Registered Broker/Dealers. First. domestic or foreign. had to have been represented in the top twenty-five underwriters of world- wide corporate offerings for at least one year during the period 1980 to 1989. identify a registered firm to be foreign-controlled if 25 percent or more of the firm's equity is owned by citizens or firms of foreign countries. Identification as a registered broker/dealer and the foreign/domestic status of the firm was identified as follows. only registered broker/dealers with the U. the Securities Industry Association surveys members of the National Association of Securities Dealers with regards to their ownership structure. These sources were used for the identification of foreign-controlled registrants in the United States and resulted in the identification of thirty-seven firms in 1980 and 121 firms in 1989. Third. to the extent that the increase in volume represents a more diversified set of issuers.218 THE CHANGING MARKET IN FINANCIAL SERVICES the aggregate is related to the volume of business being done in the underwriting industry. each firm.7 Appendix A lists the underwriting firms used in this study.S. Second. The foreign/domestic status of registered broker/dealers is not cur- rently required by the Securities and Exchange Commission.S. Security and Exchange Commission were considered. Top twenty-five underwriter status for worldwide corporate offerings was identified from various issues of IDD. classifica- tion as a foreign firm required an equity ownership by foreign citizens or firms of minimally 25 percent. The Industry Underwriting Firm Identification. however. Additionally. To evaluate the impact of foreign underwriting firms on their U. Increasing volume for the industry could drive this result in two ways: (1) fixed costs for the industry are now spread over a larger volume of business. firms were classified into domestic and foreign classes. The New York Stock Exchange does.

2 percent of the industry's assets in 1980 and 22.3 percent in 1989). Furthermore.9 4.0 Foreign 0. these figures would suggest that. the position of the foreign registrant in terms of firm number and firm size is shown to have increased substantially over the past decade.2 billion or seventeen times the industry average.8 In relation to the total number of firms in the industry.4 143. .1 percent of the industry's assets in 1989. The assets of the average firm in the industry in 1989 is $72 million while foreign firm assets were on average $1.4 Equity/Asset Ratio Total .031 information for firms registered in 1980 and 1989. Broker/Dealer Registration (1980-1989).2 22.714 9.58 Foreign 37 121 3. This small number of firms. 1980 1989 Growth Factor Number Total 5.5 11. Accounting informa- tion was taken from annual reports filed by broker/dealers registered with the Securities and Exchange Commission.0 36. it holds assets that are less risky than those of the average firm in the industry.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 219 Table 6-2.1 Equity ($Bn) Total 10. however.040 . while the average foreign firm operates at a scale commensurate with the larger broker/dealers in the industry. 1. Since capital re- quirements are a function of the riskiness of the business conducted.066 . It does.5 4.27 Assets ($ Bn) Total 150. Equity to assets ratios indicate that the equity position of foreign firms is approximately 60 percent that of the average firm.5 649.31 Foreign 12.0 11. Table 6-2 does not necessarily provide information that is pertinent solely to the business of underwriting corporate securities.56 % Foreign 8.7 percent in 1980. support the premise that foreign entry into the market for financial services in the United States is an important element of the market.021 1. foreign firms represent a small percentage (.4 % Foreign 5. accounted for 8.055 Foreign . however. The assets of broker/dealers support other activities such as retail sales in secondary markets.

the business of foreign under- writers since it does not include strictly foreign domestic business. however.058 percent in . One indication of operational efficiency is the comparison of average volume($) per issue. increasing the profitability of the transaction. Indicators of Operating Performance Table 6-3 catalogues the trends in four measures of operating perform- ance: volume of business. and the domestic/foreign seg- ments are presented in Panel 1. The relative positions of the industry groups segmented by foreign/ domestic ownership is evaluated both by aggregate worldwide business conducted as well as by security type. The number of issues placed and average volume per issue are reported in Panels 2 and 3. The average gross spread of the domestic group is higher than the foreign group in eight of the ten years. The trend in gross spread figures is one indication of increasing competition in the industry. A measure of revenue performance is given by the gross spread. If domestic and foreign underwriters face the same fixed costs. Both industry groups increased the level of business done on an annual basis in all but two years of the sample (1987 and 1988). The volume of underwritten offerings for the top twenty- five (all firms identified in Appendix A). For example. The difference in spreads measured in absolute values ranges from a high of 2. average volume per issue. and gross spreads. This comparison understates. The fixed costs incurred in transacting would therefore be spread over a larger issue volume. Panel 4 provides the average gross spread for all deals conducted for each group. Domestic firms appear to have an advan- tage with contracted premiums for the services that they provide. tend to bring to market larger issue volume. on average. Panel 4 reports the average gross spread in percentage terms. number of issues.220 THE CHANGING MARKET IN FINANCIAL SERVICES Performance of Industry Segments This section provides evidence on performance of the underwriters listed in Appendix A in accordance with their foreign/domestic classification. The domestic industry underwrites a greater volume of new business than does the foreign sector throughout the sam- ple. Panel 3 shows that foreign underwriters. then the foreign underwriters may enjoy a cost side advantage.45 percent in 1982 to a low of 0. issues underwritten by Nomura Securities for a Japanese firm distributed solely in Japan or an issue for a West German firm distributed solely in Germany are not counted.

Table 6-3.37 150.76 273.46 123.65 52.47 4.93 Domestic Mgr 51.472 2.848 2.43 446.223 2.24 54.09 103.14 92.250 1. and selling fees as a percent of offer price.25 82.17 90.54 56.06 Foreign Mgr 6.921 1. Aggregate Offering Production and Associated Fees 1 for Domestic and Foreign Underwriters.28 152.04 95.31 200.004 2.319 3.32 417.49 442.938 2.45 Gross Spread ('Yo).50 104.73 169. 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 Volume ($Bn) Top 25 57.44 Foreign Mgr 82.34 56.14 60.28 163.44 121.30 99.599 Foreign Mgr: Mean 2.07 51.05 135.87 254.98 113. underwriting.32 275. . Top 25 Underwriter status satisfied at least once.65 305.138 3.036 2.96 505.04 253.27 89.661 1.646 4.88 78.96 41.834 1.11 171. Domestic Mgr: Mean 3.28 81.08 54.103 2.73 125. • Gross spread is computed as the sum of managing.979 1.95 71.87 Issues (#) Top 25 986 1009 1478 2303 1700 2393 4052 3546 4657 5372 Domestic Mgr 908 940 958 1705 1130 1746 2645 2241 2281 3889 Foreign Mgr 78 69 525 598 670 1047 1407 1305 1376 1483 Average Volume ($ Mn) Domestic Mgr 56. Appendix A provides the list of underwriters used for each year in the period studied .97 78.74 86.08 49.939 I Full credit for issue allocated to book manager.552 4.24 119.483 3.018 1.35 138.

The market shares of foreign underwriters increased sub- stantially for the domestic market over the period. Common stock under- writing increased from a 7 percent share to a 27 percent share in 1987 and a 26 percent share in 1989. The top twenty-five underwrote more than 90 percent of the market in each year.222 THE CHANGING MARKET IN FINANCIAL SERVICES 1988. Although this use of market share figures presumes that the market has been correctly specified. Panels 3 and 4 provide share information for all offerings that are inclusive of international debt. Domestic debt and preferred stock underwriting have also been successful target markets for foreign underwriters. It also appears to be the case that the market share increase of foreign under- writers represented a transfer of business from the domestic underwriting group. ranging from 50 percent to 54 percent. Trends in Market Concentration One indicator of relative competitiveness is market share. Domestic debt shares ranged from 10 percent in 1982 to 20 . This represents an approximate increase in business of 300 percent over the period. Panels 1 and 2 of Table 6-4 report underwriting totals and market shares for all issues other than international debt. achieving a high of 19 percent in 1987. Market shares for worldwide business appear to be constant. The reversal in average profitability is evidence consistent with in- creasing competition across groups. except for 1982. domestic underwriters achieved a 90 percent market share. This appears to have been reduced and ranged from 73 percent to 80 percent for the second half of the decade. while table 6-5 shows the same data conditional on security type. Foreign firms' market share was approximately 10 percent in 1980. Table 6-5 provides market share information that is conditional on security type. The average market share of foreign underwriters is 33 percent ranging from 30 percent to 34 percent. Table 6-4 shows for each group total production. Table 6-4 and 6-5 provide market share information. The period 1982 through 1989 shows a somewhat startling result. the stability of the market share figures at the international level supports our specification. In the early part of the decade. Surprisingly. the presence of foreign underwriters as an important factor in the domestic market was evident for the entire decade. their share increased over the period. This suggests that the exchange of market share observed in the domestic market is counterbalanced by activity in the international market. Domestic underwriters achieve an average market share of 52 percent.

79 51.86 14.37 150.66 0. Volume Totals and Market Shares for ''Top 25" Underwriters. • (A) and (B) Market Totals repeated here from table 6-1.83 0.14 92.54 56.76 0.08 54.99 0.54 0.76 273.54 0.91 0.07 Domestic Mgr 51.78 Foreign Mgr 0.82 0.13 0.96 41. Top 25 Underwriter status satisfied at least once .34 0.99 0.68 509.35 95.06 Foreign Mgr 6.14 60.09 0.04 27.00 326.49 442.33 0.92 0.87 0.99 0.27 89.87 0.95 103.50 0.35 1 Full credit for issue allocated to book manager.43 446.95 0.99 0.11 0.60 128.50 0.75 238.89 340.80 0.73 169.10 0.60 65.15 (B) Total (1-6)' 58.32 417.26 165. 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 (A) Total (1-5)' 58.47 4.85 0.30 0. 1 Volume Totals-U.54 56.15 0.93 Domestic Mgr 0. .75 72.99 241.33 0.38 242.13 522.04 39.50 104.28 152.90 0.83 301.74 0.89 Domestic Mgr 0.46 11.10 56.09 0.04 253.11 0.47 4.62 265.83 185.08 49.04 95.88 0.87 Global Market Share Top 25 0.31 200.18 301.35 138.90 0.52 Underwriter Totals Top 25 57.96 505.53 0.38 115.91 571.93 Domestic Mgr 51.92 0.65 305.55 Market Share of (A) Top 25 0.08 43.38 481. $ Billions.11 0.Table 6-4.77 0. Market Share--%.41 317.11 171.95 317.76 0.34 0.53 Foreign Mgr 0.10 56.42 282.29 296.08 94.73 0.17 0.S.96 6.94 0.94 0.62 82.52 Foreign Mgr 6.08 63.07 51.32 275.47 150.89 302.61 86.42 143.04 50.18 0.53 0.87 254.42 0.80 0.07 51.84 0.32 0.30 55.19 0.88 0.28 163.89 Underwriter Totals Top 25 57.76 0.33 0.

17 Foreign 0.85 % Mkt 0.92 0.86 9.37 76.23 Market 4.10 0.39 19.60 187.93 231.47 90.39 20.51 0.79 0.85 6.33 33.89 0.50 0.20 0.16 0.10 40.15 0.50 7.89 % Mkt 0.16 6.23 0.33 63.00 10.54 0.75 0.20 29.80 Foreign 8.64 0.13 % Mkt 0.77 0.97 1. Dollar Volume Concentration By Security Type and Manager Class.65 41.71 1.16 0.55 0.28 46.08 0.08 Dom Debt" Domestic 73.10 0.41 151.18 37.08 0.14 a Domestic Debt and Preferred combines straight and convertible issues.00 0.20 100.69 129.83 5.51 0.09 0.10 0.81 116.19 0.06 12.13 Market 81.10 0.68 39.94 8.50 3.91 0.88 0. .81 5.83 188.76 238.31 62.65 Market 0.26 Market 27.85 250.91 0.08 0.39 42.52 44.38 4.29 15.77 % Mkt 0.03 39.51 13.57 0.76 0.94 8.07 0.Table 6-5.17 0.90 0.44 9.30 155.43 234.$ Billions.13 0.38 31.68 23.62 Preferred" Domestic 4.40 0.78 0.75 Foreign 0.27 0.08 0.60 0.46 76.21 25.11 0.13 205.62 1.00 62.19 In1'l Debt Domestic 0.88 0.85 0.93 0.27 31.76 0.71 11.61 32.01 34.55 0.70 0.38 60.54 % Mkt 0.48 4.86 14.85 % Mkt 0.54 9.29 3.43 207.82 40.30 0.92 237.45 293.76 8.49 22.52 149.20 4.82 0.17 0.20 0.68 2.68 22.49 9.56 % Mkt 0.82 0. Volume Totals-U.90 38.57 39.32 % Mkt 0.12 2.99 34.65 0.62 Foreign 2.56 0.15 0.92 0.00 34.03 85.61 0.97 204.10 0.82 1.51 6.62 37.66 77.22 0.05 0.40 14.79 0. 1980-81 1982 1983 1984 1985 1986 1987 1988 1989 Common Domestic 24.08 0.33 112.62 0.14 0.91 14.74 57.74 0.61 8.06 9.21 12.22 0.S.16 0.03 8.00 0.13 0.96 46.

Pricing The cost of underwriting has been documented in the finance literature as systematically related to various attributes of the issue. Additionally we control for the initiation of shelf registration under SEC Rule 415. The domestic underwriters then continued to increase market share throughout the decade as did the foreign underwriters. we model the rela- tionship between the gross spread accruing to the underwriting syndicate and several factors that have been shown to be related to the spread. explanatory variables. This does not appear to be the case for debt. corporation). domestic underwriters appear to have lost 33 percent of their debt market but not to the foreign underwriter group. effective March of 1982. Ordinary least-squares regressions were run separately for com- mon stock. and the nationality of the issuer (foreign issuer or U. the status of the underwriter (domestic or foreign). The international debt figures appear to be relatively constant for for- eign underwriters. Due to the use of qualitative. domestic to foreign. Generally. Additionally. The results are re- ported in Table 6-6. We model the gross spread as a function of the volume of the issue. for equal amounts of capital raised. and the initial public offerings of common stock. Foreign firms' rising share of common stock underwriting appears to have been captured at the expense of domestic underwriters. Domestic underwriters simply failed to recover the position they enjoyed at the beginning of the' decade. the industry of the issuer. we attempt to identify systematic influences peculiar to the market segments studied here. 9 Here. and they tend to be smaller for seasoned equity than for initial public offerings. Issue-specific information was collected from IDD Information Services. The total share of the market in 1980 and 1981 was split 88 percent to 11 percent. and international debt. No apparent trend appears to be in place for the domestic group. the interpretation of regression coefficients on the dummy variables must be performed with care. Their influence on the gross spread is in relation to that identified by the null (off) condition on qualitative determinants . gross spreads tend to be smaller for debt issues rather than for equity. domestic debt.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 225 percent in 1985. Spreads have been found to be inversely related to the volume of the issue and they tend to be smaller for competitive versus negotiated offerings.S. while preferred offerings ranged from 10 percent to 22 percent in 1988. In 1982.

03 0.76 1.74 23.21 Issuer (For = 1. N = 0) -0.02 F-statistic 30. perhaps indicating the increase of risk in their operations during the decade of the 1980s. Industry type does not enter into the modeled spread in a systematic way across security types. Rsq 0. First. The level of the coefficient is higher for financial sector firms. the fees represented by the spread are inversely related to issue volume for each issue type. the year that Rule 415 becomes effective.80 -0. This general result is consistent with previous research on underwriter fees.93 Intercept 6.37 Agency (Y = 1.05 level. 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.91 -0.56 29. N = 0) 1. All three groups have significant coefficients.27 Mgr (For = 1.12 0. The incremental fee charged for common stock issues is positive and statistically significant for both industrial and financial firms. Model Common Domestic Debt Int'l Debt Gross Spread 5. Shelf .20 N 5938 8525 6900 1 Insignificant at p = .99 Industry Industrial (Y = 1.16 0.20 SEC Rule 415 -0. Domestic debt demonstrates a different pattern for industry type.226 THE CHANGING MARKET IN FINANCIAL SERVICES Table 6-6. N = 0) 1. Results of OLS Regressions to Explain Gross Spreads.14 0.31 3. The most general case is that for common stock. N = 0) 1. A summary of the results follows. The spread for all other combinations are condi- tional on active qualitative variables and additive in them.09 -0. of the spread.39 0. Shelf registration is not modeled for international debt since the data series begins in 1982. but financial firms and corporate agencies enjoy a reduction in costs that may be attributable to governmental guarantees of some form.02 1.93 1 -0.37 -0.14 -0.85 0.05 1 Financial (Y = 1.28 1.28 In (Volume) -0.16 IPO (Y = 1. The industry dummy on international debt is negative and only significant for financial firms.56 0. Dom = 0) -0. Dom = 0) -0.04 1 Adj.03 -0.78 1.

uses stock price changes to test whether horizontal mergers increase the probability of successful collu- sion among the remaining rivals. Since most equity issues for foreign firms are in the form of American Depository Receipts (ADR). Issuer status is also significant in each model but has varying signs. This cost reduction is expected to be passed on to issuers and should be independent of issue type. foreign underwriters underprice domestic issues as an entry cost to the domestic market but do not have to system- atically underprice in their established markets. Initial public offerings. Two interpretations are consistent with the signs of these coefficients. Alternatively. is statistically significant and is the largest increment to average gross spreads. Both common and domestic debt have significant and negative effects on fees. This suggests that foreign underwriters are charging lower gross spreads on average for business underwritten in the United States.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 227 registration should reduce the cost of issuance to managing underwriters. such as mergers and acquisitions. a relevant variable for common issues only. or a negative pattern. We do not control for sponsored versus unsponsored ADRs. In this section. this is not the case for international debt. Eckbo (1983). the results could be consistent with the relative risk of the business conducted in each market. a reason for consolidation through mergers. we examine patterns of abnormal rising and falling stock prices for the underwriting industry over the 1980s as an indicator of consolidation or entry. for example. First. Similarly. Foreign Competition and Stockholder Wealth Stockholder wealth has been used in the industrial organization literature to measure the impact of various corporate events. The findings presented here suggest that the expected result is found for common stock issues but not for debt issues. an overall industry pattern of positive abnormal returns can indicate profitable investment opportuni- ties. The coefficients on the status of the managing underwriters are consist- ent with increased competition in the underwriting market. Foreign issuers of common stock incur lower fees on average as indicated by the issuer variable. Surprisingly. This is as expected and consistent with extant research. this lower cost may be capturing the sponsor's layer of intermediation. In the next section we attempt to measure the impact of foreign com- petition on the equity values of domestic rivals. We also examine some accounting measures of profitability to ascertain . We look specifically at the effect of foreign entry. Neither case can be refuted by these results.

year-to-year percentage changes in sales. the trend since then has generally been down. Table 6-7b lists the averages (in percent) of the firms' yearly return on investment. 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. ninety- four involved foreign acquire. return on equity. There were at least 550 completed deals during the 1980s involving domestic broker/dealers and investment banks. Of these fifty large underwriters. with 28 completed deals in 1989 alone. the market where they traded. market directly by registering with the SEC as a broker/dealer. from Compustat. internal cash flows have not kept pace. These deals included partial investments. We also gathered several measures of accounting profitability for these firms over the ten-year span. Most of the averages suggest that the mid-years of the 1980s were the most successful for the firms. such as Nippon Life Insurance Company's purchase of a 13 percent stake in Shearson- Lehman for an estimated $535 million. about fifty firms accounted for more than 95 percent of the underwriting activity during the 1980s. We then look at the effect of horizontal mergers be- tween foreign and domestic firms on the remaining domestic rivals equity value. We assembled these firms' daily stock returns. and Yasuda Mutual Life Insur- ance Company's purchase of an 18 percent voting stake in Paine-Webber for an estimated $300 million. the date they stopped trading. Finally. Data According to the data in the previous sections. For example. while sales continue to grow.S. It also lists the ten year geometric averages. However. we examine the stock price reaction among domestic rivals to the entry of foreign competition into the domestic market through their registration with SEC as a broker/dealer. cash flow. return on assets. We identified 47 announcement dates for deals of $25 million or more involving foreign acquire/investors in domestic broker/ dealer firms. Table 6-7a lists the firms. . and the reason they stopped trading. Of these 550 deals.228 THE CHANGING MARKET IN FINANCIAL SERVICES whether the abnormal stock price behavior is confirmed by subsequent accounting data. Figure 6-1 shows that the number of for- eign acquire has been growing over the decade. most of the recent observations are below their averages. their initial listing date. Foreign firms may also enter the U. We have identified several registration dates for foreign firms from confidential internal sources. 1980 to 1989. and assets.

936 10.394 1985 0.811 1983 1.02 16.426 12. 07127n1 Morgan Stanley Group Inc.416 20.471 13.391 20.039 1982 1.981 27.761 17.379 1988 1.818 18. Account ROA RaJ ROE Sales Cash Total Year Flow Assets 1980 2.382 18.862 17.398 5.537 4.911 16.905 -41.101 9.723 Ten Year 1.63 10.63 25.931 1984 0.005 8.903 14. Inc. Various Accounting Rates of Return and Growth Rates.437 20.175 21.922 1989 0. Name Date Listed Date Delist Reason/or (Per CRSP) Delisting Alex Brown.731 50. Average of Ten Publicly Held Domestic Broker/Dealers.524 Averages .401 12.768 23.44 28.115 14.341 14.755 18.037 21.35 8.736 21.855 34.808 15.Table 6-7a.746 14.399 8.044 12.807 12.365 22.24 26.829 8.27 1987 1.621 -0.992 25.685 1986 2.071 -13.001 13.81 12.43 6.108 24. before 7/62 Shearson Lehman 05/07/87 Table 6-7b. Domestic Broker/Dealers with Publicly Traded Equity in the 1980s. 02/28/86 Bear Stems Company 10/29/85 Citicorp 11101168 First Boston Inc. 12/14n2 12/22/88 Acquired by Credit Suisse Merrill Lynch & Co.507 -0.407 13.224 16.8 1981 1.226 70. 02/03/86 Salomon Inc. 03/21186 Paine Webber Group 04/03n2 Piper Jaffray Inc.939 6.993 2.21 9.

Completed mergers and acquisitions among domestic broker/dealers. M ~~ 60 ~~~~~~r~~~t :::::::: 40 I f:-:':':':':':' x x 2: : 1'1 ~ II. 1980 81 82 83 84 85 86 87 88 1989 Figure 6-1.: =m tJ t'a<l . ..Number of Completed Deals 160r-----------------------------------------------------~====~--~ 140 1 :::r:::'::::::~ I 1 1201~------------------------------------------------~: 100r---------------------------------------------~~~.

estimated in the regression equation above. A pattern of rising cumulative abnormal returns may signal new profit opportunities and thus provide an incentive for entry. A minimum of 250 returns had to be available for the estimation procedure. These averaged abnormal returns were then cumulated across time.t and 'm. The cumulative abnormal return (CAR) is computed in two stages. and captures the effect of industry. and hence may be associated with . The error term measures the firm's nonsystematic (or idiosyncratic) risk component. The a i and ~i are estimated yeady for each firm using ordinary least squares and the stock returns from the prior two years.i. This procedure is similar to cumulating the abnormal return of a portfolio with changing components that is con- tinuously rebalanced to maintain equal weight in each component stock. forming portfolio abnor- mal returns. In the second stage of the cumulative process the abnormal returns were averaged across firms at each point in time. Results Industry Patterns.t. First. Because each firm's systematic risk may be unstable.and firm-specific events on a firm's equity value. Each firm's {3.t are the daily. advancing one year. continuously compounded rates of firm i and the equally weighted market portfolio. a market model is estimated for each firm as. Using these estimated coefficients. captures the firm's systematic movement with changes in the market's return. (2) This same procedure was used repeatedly.THE COMPETITIVE IMPACf OF FOREIGN UNDERWRITERS 231 Methodology Figure 6-2 shows the cumulative average abnormal returns experienced by the sample of underwriting firms during the 1980s. A pattern of falling cumulative abnormal returns can indicate falling demand or rising supply prices.t + E. abnormal returns (forecast errors) were computed for the current year. 'p. forming CARs. as.t (1) where 'i. using the most recent two year span to estimate the coefficients.t = ai + ~l m. a rolling window was used to estimate the market model coefficients. then calculating one year's worth of abnormal returns. 'i.

."".Cumulative Abnormal Return 0.2 I I.)' \ ill' II•. 1 ML.ilF'\.. I -0./\ 0.. . fIJ"fr W. I N' tf k F\ I I 0.. \ Af .3+ .1--+1+1-+-- o I' '" I.' ' 11\. Ill.3~--__~------~----~------~----~------~----~------~----~----~ 81 82 83 84 85 86 87 88 1989 Figure 6-2.5 I I 1\ I 0. Cumulative abnormal returns.2 I . Underwriters during the 1980s.7 I 0. -0.6T'~~~~~-~'r'------~--------~ 0.1 .0.111.4 I II ~l l 0. RWllfl 1\1 'l .

Figure 6-3 shows the yearly change in the cumulative returns (lead one year). The years 1984 and 1985 generally were good ones for investors. The stock market is forward looking. and finally. which lost not only the cumulated 60 percent but also an additional 10 percent by the end of the year. in five of the seven years in which there were appreciable mergers and acquisition activity. only to be followed by a significant downturn early in 1987. the entry of a foreign competitor is likely to signal increased competition in the output market . In six of the eight years covered. increases (decreases) in CARs over the past year were associated with increases (decreases) in sales growth in the current year. In all years. increases (decreases) in last year's CARs were associated with decreases (increases) in this year's mergers and acquisitions (M&A) activity. Table 6-7b presents yearly average accounting rates of return (on investment (ROI). however. and on equity (ROE» for the ten firms in our sample. increases (decreases) in last year's CARs were associated with increases (decreases) in this year's return on equity. on assets (ROA). the abnormal returns have been generally positive. Since the October 1987 market break. Thus. The early part of 1983 was disastrous for the underwriters as a group. and thus increase the equity value of the remaining firms in the industry. Here. whereas the accounting data are historically oriented. the average yearly growth rate in sales. Rates of growth in assets and sales are also presented. Unexpected changes in a firm's operating environment can change the market value of its equity. Finally. moving from a negative 20 percent cumulative abnormal return to a positive position of about 15 percent at the end of the decade. The Impact of Foreign Entry. An anti-competitive merger might raise product prices. the change in average rate of return on equity. particularly an industry that faces significant fixed costs. 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.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 233 consolidation within the industry. The CARs are likely to be leading indicators of subsequent accounting results. In the next section we look at the impact of this activity on the equity values of the industry's publicly traded firms. Figure 6-2 shows that the cumulative abnormal return peaked late in 1982 at around 60 percent. the number of completed acquisitions (both foreign and domestic) in the industry for the year. The association of merger and acquisition activity with poor perform- ance is consistent with the notion that mergers consolidate an industry's assets. Accounting rates of return may provide additional insights into the competitive position of domestic underwriters during the 1980s.

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. CARs as leading indicators of changes in growth rates and M&A activity.i.234 THE CHANGING MARKET IN FINANCIAL SERVICES Changes in Rates 100 I] ROE § I Sales E 80 ~CAR § 60 I3M&A ~ § § 40 § • § ~ Jl 20 j § o m r ~ AI • I ~ f. according to IDD Information Services. Overall.J -20 ~ -40 -60 I I I I I I I 1982 83 84 85 86 87 88 1989 Figure 6-3. These forty-seven deals were $25 million or larger. (lower product prices) and/or in the input market (higher input prices). the day of the announcement and the next day. in case the information arrived after the close of business on the announcement date. we found forty- seven that involved foreign purchasers of domestic firms or significant stakes in domestic firms. Of the 550 completed acquisitions in this industry. A two-day window was used. 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. 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. .

negative 2 percent. these results are weak and not statistically significant. We find no consistent reaction to foreign acquisitions of domestic broker/dealers.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 235 the standard deviation of two-day CARs is 1.5 percent on the an- nouncement date and a two-day average abnormal return of -0. whereas the domestic market share of the major domestic underwriters has fallen. slightly negative. From confidential internal sources we have identified seven registration dates involving major foreign broker/dealers. we find that industry abnormal stock price performance to be a leading indicator of merger and acquisition activity within the group. neither negative abnormal return is statistically significant.23 percent. with average spreads falling through most of the decade and with foreign entry through acqui- sition or registration providing additional competitive pressure. thus. Financial innovations. especially asset-backed debt. consistent with consolidation motive for an industry that faces fixed costs and rising competition.65 percent. [An examination of the two-day cumula- tive returns for individual acquisitions shows significant variation.17 percent.7 percen~. Although we find a negative impact on the equity values of publicly traded domestic broker/dealers when foreign firms enter the domestic market. at the other extreme. In summary. However. we found negative average portfolio abnormal returns of -0. However. it also appears that the domestic underwriters have been successful in expanding their business overseas. Using the same methodology as outlined above. their global market share has expanded such that their total market share has remained relatively constant. the standard deviation of the portfolio's daily abnormal return was 1. such as mortgage-backed securities.] However. Debt. A second avenue of foreign entry into domestic markets is to register directly with the SEC as a broker/dealer. has become the dominant issue in the do- mestic market. So. . but not significantly different from zero. Summary and Direction of Future Research The overall picture one can take from the data presented here is that the domestic underwriting business is very competitive. with some as large as a positive 2 percent but. either through acquisition or direct reg- istration. the net result is approximately zero. Also salient is that changing nature of the underwriting business. The two-day cumulative average abnormal return on the announcement of a domestic acquisition by a foreign firm is -0.

We hope to pursue some of these issues in our future research. now represent a major portion of the underwriting business. Finally. such as exchange offers on existing debt. It is likely that new products and services. may well represent dynamic growth areas for the future. Do underwriters just service their own domestic clients and follow their customers overseas when they expand into foreign market? Or do underwriters enjoy suffi- cient economies of scale to overcome the natural cost disadvantage of operating in a foreign market. For example. we do not know the source of the competitive advantage they might enjoy. . While we have necessarily provided just summary statistics. they do indicate some direction for future research. we have attempted to provide an overall picture of the impact of foreign competition on the domestic underwriting business. while we have documented the extent of foreign presence in the domestic market.236 THE CHANGING MARKET IN FINANCIAL SERVICES which were a negligible portion of the market in the beginning of the decade.

Benson J. 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.F. Read Piper Credit Lyonaise S. Sachs Wertheim Schroder Indosuez Hambrecht & Quist Wheat. Warburg Donaldson Prudential-Bache Credit Suisse Union Bank of Switzerland Drexel Robertson Credit Suisse/First Boston Yamaichi International E.Appendix A Name List of Underwriting Firms 1980-1989. Stearns Merrill Lynch Banque Paribas Capital Markets N. Rothschild & Sons Citicorp Montgomery Barclays Merchant Bank RBC Dominion Securities D.P. Hutton Salomon Brothers Daiwa Securities Faherty & Faherty Smith Barney Deutsche Bank First Boston Stephens Dresdner Bank First Jersey Thomson McKinnon Hambros Goldman. Rothschild Banque Nationale de Paris Nomura Securities Bear.F. 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.G.M.H. Morgan William Blair Lazard Freres . Domestic Foreign Alex Brown & Sons Kidder. First Industrial Bank of Japan John Muir Wheat-Butcher Kleinwort.

Foreign- controlled firms which appear in Appendix A were identified through public sources independ- ently of information provided by the NYSE. See Hayes. foreign securities only issued in their home country. 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. The lead manager controls the "pot". A form of negotiated.e. 2. Market identifying characteristics applied here are essentially the same as those described in Hayes. "Using Daily Stock Returns: The Case of Event Studies. 8. 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. 9. For the period studied here. 3. 1985. Securities and Exchange Commission for the period 1972-1986 accounted for no industrial firm offerings (6807 deals registered) and 28. 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. The influence and size of syndicates are decreasing over the period of this study. The managing underwriter establishes the underwriting agreement with the issuer and identifies the terms of the offering as well as each member's commitment. Confidentiality of foreign-controlled registrants provided by the NYSE was requested.. This portion is referred to as "retention". et.238 THE CHANGING MARKET IN FINANCIAL SERVICES Notes 1. We do not have access to strictly domestic. Spence. Typically. as well as all offerings issued internationally. and Jerold B. 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. i. each syndicate member controls only a portion of the securities which it has agreed to underwrite. 7. Stephen J. the Financial and Operational Combined Uniform Single Report: SEC Form X-17A-5. The importance of the lead manager in negotiated transactions should not be underestimated.al. 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. .9 percent of public utility offerings (3589 deals regis- tered).8 percent of the business done by the managing firms analyzed. offerings representing competitive bids account for 2. 5. Only aggregated information is therefore presented for foreign-controlled firms. 4. References Brown. The annual reports are referred to as FOCUS Reports. 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. (1983). Warner. Hansen and Khanna (1990) document that competitive bid deals registered with the U. These institutional orders are typically placed with the managing underwriter but may be earmarked by the purchaser for the account of designated selling agents. foreign issues. This has been documented primarily for the issues of public utilities.S. The source for most of the data in this section is various issues of Investment Dealer's Digest and IDD Information Services." Journal of Financial Economics 14 (March): 3-31. 6.

) 1985. Cambridge. (May). New York: John Wiley & Sons.THE COMPETITIVE IMPACf OF FOREIGN UNDERWRITERS 239 Eckbo. "Horizontal Mergers." Journal of Economics and Business 28 (Winter): 96-103. "The Competitive Effects of U." Journal of Financial Economics 11 (April): 241-273." Journal of Law and Economics 30 (April): 181-206. 1989. Logue. and Japanese Commercial Bank Participation in Eurobond Underwriting. and Stockholder Wealth. B. Trimble. "Using Financial Data to Measure Effects of Regulation. 1990. Dennis E. (ed. 1983.. 1987.1983." Journal of Financial and Quantitative Analysis 8: 91-103." Working paper. 1973. Walter.S.. H. Kidwell. Rogowski. M. 1981. Stoll. R. Schwert. G. Hayes. David S. "Why Negotiation with a Single Syndicate May Be Preferred to Making Syndicates Compete. and G. William. R." Journal of Finance 29 (December): 1531-1543. Collusion. "On the Pricing of Unseasoned Equity Issues: 1956-1969." Journal of Law and Economics 24 (April): 121-158. 1976. Samuel L. M. and N. A. and David Van Praag Marks. Hansen. "The Yield Spread on New Issues of Corporate Bonds. Competition in the Investment Banking Industry. Michael Spence. Wayne Marr. Espen." Financial Management 18 (Winter): 47-54. Deregulating Wall Street: Commercial Bank Penetration of the Corporate Securities Market. Marr.. Wayne. "Shelf Registration: Competition and Market Flexibility. . Khanna. Louis H. and J. "The Pricing of Underwritten Offerings of Listed Common Stocks and the Compensation of Underwriters. Ederington. I.1974. MA: Harvard University Press. Rodney Thompson.

It looks more dra- matic in the article than its actual market impact. as I understand it. it's the study of global competition. however. COMMENTARY Samuel L. The fact that there is relatively little in the article that deals with the U. competition. Hayes The NachtmannlPhillips-Patrick article makes a useful contribution to understanding what is happening in the global underwriting area. Rather than being a study of domes- tic U. The increase in percentage of equity financings done by foreign vendors has got to be offshore equity. It chronicles the movement of issuer financing offshore. in large part.S. 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 .S. Buyers and sellers of capital are. The authors' statistics. shopping globally. domestic underwriting market doesn't really surprise me. Equity Market Shares I want to give particular attention to the findings that the biggest change in the market shares have accrued in the equity area. I think the article was mislabeled.

-controlled securities firms is probably much more heavily weighed towards principal business.242 THE CHANGING MARKET IN FINANCIAL SERVICES that were done in New York. in many cases. The weak position of foreign vendors in the New York market is mir- rored in other national markets around the world. I don't believe that foreign vendors have a significant position.S. The authors note that the financial leverage of the foreign firms was greater than that of the domestic firms. The business mix of U. that large an equity base is not needed. equity market.S. on balance.S. the explanation for the authors' preferred stock market share figures must be due to the inclusion of the Euromarket financings in their data. They are. then the authors are probably correct in attributing it to their mix of business. this is. after all.S. It also seems likely that foreign controlled vendors operating in the United States would. with assets sub- . If it is largely an agency business without the kinds of assets that are vulnerable to substantial price fluctuation. a very firm grip on the equity offering business undertaken there. Assuming that we are talking about the balance sheets of the securities underwriting subsidiaries of these foreign organizations. my guess is that their financial structures and operating policies have nonetheless been influenced by the deep pockets of their parent organizations. offshoots of much larger financial service conglomerates? I don't know whether or not the data that the authors were able to obtain isolated the underwriting subsidiaries of these foreign parents' firms.S. I wonder if that can't be explained in part by the fact that these foreign controlled organizations are. Thus. securities firms. the ones who can afford to come and establish a foothold position in a foreign market that's as competitive and as expensive to operate in as the U. But even if this was the case. marketplace. Vendor Profiles Characteristics of the vendors is another interesting aspect of this article. for the most part. home- based vendors have. It is certainly noteworthy that the firms that are foreign controlled are a lot larger than those that are U. I think. either. controlled. be larger than the group of twenty-five leading domestic U. The authors distinguish between firms that are owned and controlled within the United States and firms where at least 25 percent of their equity is in the hands of offshore owners. an interesting observation. I believe that disaggregated statistics would show that foreign vendors have a negligible share of the U. In those markets. Among preferred offerings in the United States.

That has been drastically reduced over the past fifteen years. 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. therefore. They are what I would have expected to find with respect to differences between "plain vanilla" bonds and such things as initial public offerings. a larger equity base would be required by these vendors to ensure their viability." which carry very high gross spreads. are bound to influence the overall aggregate results. Thus. particularly on the debt side. With respect to the shelf registrations. I have two further comments on gross spreads. there has been an impressive squeezing down of the spreads. 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.] 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. . it is my own personal opinion that these minimum requirements have in some cases proved to be inadequate.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 243 ject to much greater fluctuations in value. Gross Spreads The authors chronicle the squeeze in gross spreads over time. a recent study provides some evidence that gross spreads are being influenced by the exposure of the vendor to legal risk. The other comment relates to a statement in the article about equity gross spreads that is probably a [misprint. The authors' statistical results are intuitively appealing. The data is not there to reach that conclusion. Vendors don't often have the time for careful examination of the issuer's affairs before bidding on the financing and. 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). 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. particularly with the advent of shelf registrations. 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 heavy volume of asset-backed securities and so-called "junk bonds. 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.

I see a lot of these "event" studies and I'm not at all surprised that the authors came up with an empty bag. The authors are therefore able to confirm that the stock market is efficiently anticipating results not yet formally released. Citicorp's real investment banking activity. But. the article provides some useful perspectives on underwriting competition. That may be true in an idealized world. In closing. The composition of these firms' business is just too diverse. it courses throughout that whole system. The authors conclude that the competitive position of the U. investment banks' aggressive efforts to position themselves in other offshore market centers as they follow their customers. While they are working toward that goal. By contrast. On the one hand. to the U. the vendors serving this interlinked marketplace are not yet global.S. vendors has not been eroded by the increased competition during the 19808. I think it was helpful to pinpoint the special events that are particularly relevant to the underwriting business. when you plug into one part of that grid. is very small.244 THE CHANGING MARKET IN FINANCIAL SERVICES Stockholder Wealth Turning to the stockholder wealth question.S. securities vendors. I also had one other less-weighty question about why Citicorp would be included in this group often U. as opposed to the premium returns earned by this group of public com- panies during each subsequent year. in part. In a recent book (Hayes and Hubbard). It is also useful to note the backward-looking nature of the accounting data issued each year. as a fraction of its total rev- enues. there are a number of barriers to entry into various marketplaces that are difficult to . I believe this has been due. but I'm not too optimistic that it would work that way in the real marketplace. As a sort of satellite photograph. Philip Hubbard and I have concluded that the marketplace for money is already global from the point of view of both buyers and sellers of money. The "electrical grid" of the global capital market is pretty much in place and. I would also take exception to the statement in the article suggesting that one might find that an influx of foreign competitors would cause the prices of the domestic competitors' stocks to go up because the new entrants would invigorate domestic vendors to greater innovative efforts. as opposed to using some broader index measure.S. we can see some features that are important to an understanding of the overall direction of the public marketplace and the operations of vendors within that public marketplace. it does seem to me that it was a long shot on the authors' part to hypothesize that the influx of foreign vendors into the United States would be likely to have much of an impact on the share prices of the ten publicly held financial vendors used in their sample.

Morgan Stanley. Investment Banking: A Tale of Three Cities. despite these offshore barriers to entry.S.000 employees in locations all over the world. And that handful of dominant U. the authors' comments about increased concentration in the U. Harvard Business School Press.. securities industry aren't surprising.S. firms continue to be world-class competitors. this paper suggests that they have nonetheless been able to hold their own in the market environment.S. which had only several hundred employees in one New York location in the early 1970s. References Hayes. and Philip Hubbard. And so I think it's a real tribute to the U.THE COMPETITIVE IMPACT OF FOREIGN UNDERWRITERS 245 overcome. investment banks that. Samuel L. There are only a half dozen or so firms that really count in terms of competitive interactions with foreign owned securities firms. Cambridge: 1990. . now has more than 6. Given these changed requirements.