frb_091998

VOLUME 84 D NUMBER 9 SEPTEMBER 1998

FEDERAL RESERVE

BULLETIN

BOARD OF GOVERNORS OF THE FEDERAL RESERVE SYSTEM, WASHINGTON, D.C.

Digitized for FRASER http://fraser.stlouisfed.org/ Federal Reserve Bank of St. Louis

Table of Contents

703 BANK MERGER POLICY AND THE NEWCRA DATA

Data now being collected under the CommunityReinvestment Act can be used to examine sev-eral issues related to analyses of the competitiveeffects of proposed bank mergers and bank hold-ing company acquisitions. This article looks attwo of those issues: whether locally based com-mercial banks and savings institutions face sig-nificant competition in small business lendingfrom institutions based outside local bankingmarkets, and whether measured levels of marketconcentration differ significantly when calcu-lated on the basis of small business loan origina-tions rather than bank deposits, as has tradition-ally been done.

716 TREASURY AND FEDERAL RESERVEFOREIGN EXCHANGE OPERATIONS

Throughout the second quarter of 1998, theJapanese yen weakened consistently, depreciat-ing 4.1 percent against the U.S. dollar and6.2 percent against the German mark. Continuedpessimism about the outlook for the Japaneseeconomy pressured the yen to eight-year lowsagainst the dollar. This contrasted with the rela-tively positive outlooks for the United Statesand Europe, with the dollar benefiting fromcontinued low inflation and a strong domesticeconomy while the mark benefited from opti-mism toward European monetary union. Duringthe first two weeks of June, the yen's declineaccelerated, a development that put downwardpressure on other regional currencies as mar-ket participants increasingly doubted that effec-tive reforms of the Japanese banking systemand stimulus of the economy would be forth-coming. Then, on June 17, the U.S. monetaryauthorities intervened in the foreign exchangemarkets by selling a total of $833 billion againstthe Japanese yen. The operation was conductedin cooperation with the Japanese monetaryauthorities.

722 INDUSTRIAL PRODUCTION AND CAPACITYUTILIZATION FOR JULY 1998

Industrial production declined 0.6 percent inJuly, to 126.8 percent of its 1992 average, aftera drop of 1.1 percent in June; the June declinewas larger than previously estimated. From Mayto July, capacity utilization in manufacturing,mining, and electric and gas utilities dropped1.9 percentage points, to 80.5 percent.

725 STATEMENTS TO THE CONGRESSLaurence H. Meyer, member, Board of Gover-nors, comments on the "discussion draft" of theregulatory relief bill and highlights those provi-sions of the legislation that the Board supportsand believes are particularly significant in reduc-ing burden and promoting efficient bank regula-tion. Governor Meyer states that the Board haslong endorsed regulatory relief to ensure regu-latory equity for all participants in the financialservices industry, minimize the chances thatfederal safety net subsidies will be expandedinto new activities and beyond the confines ofinsured depository institutions, guarantee ade-quate federal supervision of financial organi-zations, and ensure the continued safety andsoundness of financial organizations, in testi-mony before the Subcommittee on FinancialInstitutions and Consumer Credit of the HouseCommittee on Banking and Financial Services,July 16, 1998.

730 Edward M. Gramlich, member, Board of Gover-nors, discusses key policy questions in regard topossible reforms to the Truth in Lending Act(TILA) and the Real Estate Settlement Proce-dures Act (RESPA) in connection with recom-mendations for new legislation presented in ajoint report by the Federal Reserve Board andthe Department of Housing and Urban Develop-ment to the Congress. Governor Gramlich statesthat the report focuses on four broad policyquestions concerning closed-end home-secured


loans: the effectiveness of the finance chargeand the annual percentage rate (APR), the relia-bility of settlement cost disclosures, the timingof disclosures, and substantive protections totarget abusive lending practices, in testimonybefore the Subcommittee on Financial Institu-tions and Regulatory Relief and the Subcommit-tee on Housing Opportunity and CommunityDevelopment of the Senate Committee on Bank-ing, Housing, and Urban Affairs, July 17, 1998.(Governor Gramlich presented identical testi-mony before the Subcommittee on FinancialInstitutions and Consumer Credit and the Sub-committee on Housing and Community Oppor-tunity of the House Committee on Banking andFinancial Services on July 22, 1998.)

735 Alan Greenspan, Chairman, Board of Gover-nors, presents the Federal Reserve's midyearreport on monetary policy and states that over-all, the performance of the U.S. economy con-tinues to be impressive and that the task is tomaintain disciplined economic policies andthereby contribute to maintaining and extendingthese gains in the years ahead, in testimonybefore the Senate Committee on Banking, Hous-ing, and Urban Affairs, July 21, 1998. (Chair-man Greenspan presented identical testimonybefore the Subcommittee on Domestic and Inter-national Monetary Policy of the House Com-mittee on Banking and Financial Services onJuly 22, 1998.)

742 Chairman Greenspan presents the FederalReserve Board's views on the regulationof over-the-counter (OTC) derivatives andaddresses the fundamental underlying issue, thatis, whether it is appropriate to apply the Com-modity Exchange Act (CEA) to OTC deriva-tives (and, indeed, to financial derivatives gener-ally) in order to achieve the CEA's objectivesdeterring market manipulation and protectinginvestors. In conclusion, Chairman Greenspanstates that the Board continues to believe that,aside from safety and soundness regulation ofderivatives dealers under the banking or securi-ties laws, regulation of derivatives transactionsthat are privately negotiated by professionals isunnecessary, in testimony before the HouseCommittee on Banking and Financial Services,July 24, 1998. (Chairman Greenspan presentedsimilar testimony before the Senate Committee

on Agriculture, Nutrition, and Forestry onJuly 30, 1998.)

746 ANNOUNCEMENTS

Statement by Chairman Greenspan on the deathof William McChesney Martin.

Revision to Regulation H and recission of Regu-lation P.

Amendment to Regulation DD.

Appointment of Governor Ferguson as Chair-man of the Year 2000 Council.

Publication of a joint report on legislative rec-ommendations of ways to simplify and improveconsumer disclosures in home-secured loans.

Availability of data on mortgage lending trans-actions in 1997 at institutions covered by theHome Mortgage Disclosure Act.

Availability of revised lists of over-the-counterstocks and of foreign stocks subject to marginregulations.

Scheduling of an additional day for the publicmeeting on the proposed acquisition of Bank-America by NationsBank.

Accessibility to the Board's World Wide Website through a new address.

Changes in Board staff.

750 MINUTES OF THE FEDERAL OPEN MARKETCOMMITTEE MEETING HELDON MAY 19, 1998

At its meeting on May 19, 1998, the Committeeadopted a directive that called for maintainingconditions in reserve markets that were consis-tent with an unchanged federal funds rate ofabout 5'/2 percent and that contained a biastoward the possible firming of reserve condi-tions and a higher federal funds rate.

757 LEGAL DEVELOPMENTS

Various bank holding company, bank servicecorporation, and bank merger orders; and pend-ing cases.


Al FINANCIAL AND BUSINESS STATISTICS A82 BOARD OF GOVERNORS AND STAFF

These tables reflect data available as of A g 4 fEDERAL OpN MARKET COMMITTEE AND

July I), 19J8. STAFF; ADVISORY COUNCILS

A3 GUIDE TO TABULAR PRESENTATION ^ ^ ^ ^ ^ BQARD pUBUCATl0NS

A4 Domestic Financial StatisticsA42 Domestic Nonfinancial Statistics A88 MAPS OF THE FEDERAL RESERVE SYSTEMA50 International Statistics

A90 FEDERAL RESERVE BANKS, BRANCHES,A63 GUIDE TO STATISTICAL RELEASES AND AND OFFICES

SPECIAL TABLES

A80 INDEX TO STATISTICAL TABLES

PUBLICATIONS COMMITTEELynn S. Fox, Chairman \~\ S. David Frost Donald L. Kohn [ .. J. Virgil Mattingly, Jr. Michael J. Prell Dolores S. Smith Richard Spillenkothen Edwin M. Truman

The Federal Reserve Bulletin is issued monthly under the direction of the staff publications committee. This committee is responsible for opinions expressedexcept in official statements and signed articles. It is assisted by the Economic Editing Section headed by S. Ellen Dykes, the Multimedia Technologies Centerunder the direction of Christine S. Griffith, and Publications Services supervised by Linda C. Kylcs.


Bank Merger Policy and the New CRA Data

Anthony W. Cyrnak, of the Board's Division of Re-search and Statistics, prepared this article. Dennis W.Campbell provided research assistance.

One consequence of the current merger trend in thebanking industry has been heightened interest inthe analytical framework and data used by federalauthorities to assess the competitive effects of bankmergers and bank holding company acquisitions (seebox "Federal Antitrust Review of Bank Mergers").Although the analytical methods used by the bankregulatory agencies and the Department of Justice aresimilar, some differences have emerged over time.Two issues related to these differences can now beexamined using data collected for the first time as aresult of 1995 changes in the regulations implement-ing the Community Reinvestment Act (CRA).

One of the issues is whether locally based commer-cial banks and savings institutions ("in-market"firms) face significant competition in small businesslending from banks and savings institutions basedoutside local banking markets ("out-of-market"firms). This issue is important because, in the view ofsome banking industry observers, the presence andcompetitive influence of out-of-market firms in localbanking markets warrant greater recognition by bank-ing regulators and the Department of Justice. If thecompetitive role of out-of-market institutions weremore fully recognized, they argue, more bankingmarkets would be viewed as structurally competitive,and bank mergers within these markets would raisefewer antitrust concerns.1 The new CRA data canshed some light on this issue by providing informa-tion about the number of, and lending activity of,out-of-market small business lenders in urban andrural banking markets.

The other issue is whether levels of concentrationwithin U.S. banking markets differ significantly whenmeasured by small business loan originations ratherthan by bank deposits. Legal precedent and economictheory have traditionally supported the use of totalbank deposits as the basis on which to calculatemarket concentration. This practice has been reexam-

ined in recent years, however, as industry observershave questioned whether competition in local bank-ing markets might be better measured on the basisof some other variable, notably loans to small busi-nesses.2 Competition in small business lending, somehave argued, is more local than is competition fordeposits and better reflects the competitive realitiesof local banking markets.

2. See, for example, "Are Deposits the Best Antitrust Measure?"Bank Mergers & Acquisitions, vol. 12 (December 1997), pp. 14.

1. To simplify discussion, the term mergers hereafter refers toboth mergers of banks and acquisitions of banks by bank holdingcompanies.

Federal Antitrust Review of Bank Mergers

All proposed mergers of insured banks and proposedacquisitions by bank holding companies must beapproved by a federal banking regulator. The review ofthe proposal is conducted by the regulatory agency thatwould be the primary regulator for the surviving bankinginstitution. If the resulting institution would be a nationalbank, the Office of the Comptroller of the Currencyreviews the proposal. If it would be a state-charteredbank that is not a member of the Federal Reserve System,the Federal Deposit Insurance Corporation reviews theproposal. And if the proposal involves an acquisition by abank holding company or the resulting bank would be astate-chartered member of the Federal Reserve System,the Federal Reserve is the reviewer. The Department ofJustice, which has general enforcement authority underfederal antitrust laws, reviews proposed mergers andacquisitions approved by the banking regulators.

Proposals for banking consolidations are reviewedunder two primary federal statutes, the Bank Merger Actof I960 (as amended in 1966) and the Bank HoldingCompany Act of 1956 (as amended in 1970). The legalstandards applied by the banking regulators under thesestatutes correspond to the standards set forth in the Sher-man Antitrust Act of 1890 and the Clayton Act of 1914.Specifically, no bank merger or bank holding companyacquisition may be approved if the effect of the merger oracquisition would be to create a monopoly (ShermanAct) or may substantially lessen competition or tend tocreate a monopoly in a particular market (Clayton Act).Importantly, the Clayton Act does not define the term"substantially." Court decisions, as well as the legislativehistory of this act, however, make it clear that the termmust be examined in the context of each industry.


704 Federal Reserve Bulletin _l September 1998

Concentration, the HHI, and the Department of Justice Merger Guidelines

Market concentration is an important factor in the competi-tive analyses of mergers and acquisitions conducted byfederal bank regulatory agencies and the Department ofJustice. Both theoretical models and empirical studies sug-gest that higher levels of market concentration are generallyassociated with less competition, higher prices, and greaterprofitability. These relationships are consistent with the"structure-conduct-performance" paradigm, which sug-gests that there is a causal relationship between the structureof a market (the number and size distribution of firmswithin a market and factors affecting the ability of newfirms to enter that market), the behavior of firms within thatmarket (competition, collusion, or other competitive strate-gies), and the performance of firms in that market (typically,profits, prices, or firm growth).

The Department of Justice has for many years publishedformal guidelines that identify structural changes resultingfrom mergers that are likely to cause the department tochallenge a merger. Since 1982, the department has basedits merger guidelines on the Herrindahl-Hirschman index ofconcentration (HHI). This measure, which is also used bythe bank regulatory agencies, is calculated by squaring themarket share of each firm competing in a defined geo-graphic banking market and then summing the squares. TheHHI can range from zero in a market having an infinitenumber of firms to 10,000 in a market having just one firm(with a 100 percent market share).1

The HHI is a static measure and, therefore, gauges mar-ket concentration at a single point in time. Algebraically, itcan be depicted as

HHI= I

where MS is the market share of the ith firm and n is thenumber of firms in the market.

The HHI is a particularly useful tool for bank mergeranalysis because it accounts for the presence of everycompetitor in a market and provides a measure of the

I. A more complete discussion of the HHI h presented in Stephen A.Rhoadcs, "The Herh'ndah!-Hir.vchman Index." Federal Reserve Bulletin.vol. 79 (March 1993). pp.188-81).

structural effect of a merger of any firms in a market. Inaddition, the squaring of market shares gives greater weightto firms that have large market shares. This weighting of thelargest competitors in a market is consistent with economictheories that predict weak competition in markets in whicha few competitors hold a large combined market share.

A merger of two banking competitors will increase theHHI in their shared market. (Mathematically, the increase isequal to the product of the two firms' market shares timestwo.) The amount of the increase and the level to which theHHI would rise after the merger are key elements in thestructural analysis of bank mergers. In the Department ofJustice's guidelines, markets with an HHI of less than 1000are considered "unconcentrated"; those with an HHI of1000 to 1800 are "moderately concentrated"; and thosewith an HHI greater than 1800 are "highly concentrated."

The department's merger guidelines for nonbankingindustries generally indicate that a merger that increases theHHI 50 points or less in a highly concentrated market isunlikely to have adverse competitive effects and ordinarilywould not be challenged on antitrust grounds. The depart-ment's guidelines for the banking industry are more lenient,largely in recognition of the widespread presence of non-banking providers of financial services. In particular, in1985, the department stated that it would not ordinarilychallenge a proposed bank merger unless the rise in the HHIwould be more than 200 points in a highly concentratedmarket. This practice was subsequently adopted by theFederal Reserve System as part of its Rules RegardingDelegation of Authority, an administrative procedure thatallows certain mergers to be approved directly by theReserve Banks. (Proposed bank mergers that raise substan-tive issues must normally be acted upon by the Board ofGovernors following a formal review and voting process.)

Although these structural guidelines constitute a criticalscreening device for the Department of Justice and thebanking regulators, the analysis of other economic factors isalso an integral part of the assessment of the competitiveeffects of consolidation. These factors can include, forexample, market economic conditions, trends in concentra-tion, and the competitiveness of other banking and thriftinstitutions.

The measurement of market concentration plays akey role in the competitive analysis of proposed bankmergers (see box "Concentration, the HHI, and theDepartment of Justice Merger Guidelines"). Gener-ally, banking regulators and the Department of Jus-tice intensively review any proposed merger thatwould add significantly to and result in a high degreeof market concentration. Importantly, concentrationindexes can be calculated using various types of datadepending on the type of product for which compe-

tition is being analyzed. The use of different data,moreover, is likely to result in different calculatedlevels of concentration. Thus, the choice of a productmarket and the type of data used to measure concen-tration could influence the regulatory outcome of amerger proposal.

The new CRA data make it possible to considerthis second issue by examining the level of concentra-tion within U.S. banking markets using a base otherthan bank deposits, namely small business loan origi-


Bank Merger Policy and the New CRA Data 705

nations. The results of this examination may suggestwhether, on average, antitrust standards are likely tobe more or less difficult to meet when concentrationmeasures are based on small business loans ratherthan on deposits.

i > \ ! AM..-// ! C i >i />' \ X K L \ t , ,\ Y / y / A V ' V / / Y . ) / . / < ) '

Antitrust policy in the United States generally recog-nizes that competition results in lower prices andbetter service for consumers. Thus, a primary goal ofbanking antitrust policy is to prevent the creation of,or an increase in, market power such that a firm couldimpose above-competitive prices and earn excessprofits at the expense of consumers.

Mergers in the banking industry are subject tothe same antitrust standards as are mergers in otherindustries. This principle was made clear by the U.S.Supreme Court in its 1963 decision involving theproposed merger of two competing Philadelphiabanking institutions.3 In that case (PhiladelphiaNational Bank) the Court made three findings thathave become fundamental tenets of banking antitrustpolicy in the United States. First, antitrust laws applyto banking mergers, and the legal standards for judg-ing mergers in the banking industry are the same asthose embodied in the Sherman Antitrust Act of 1890and the Clayton Act of 1914. Second, commercialbanking is a distinct line of commerce in which banksprovide a unique "cluster" of products and servicesnot provided by any other type of financial institu-tion.4 Third, commercial banks compete in local mar-kets, primarily because of their need to provide con-venient service to their customers.

In reaching its decision, the Supreme Court vali-dated the "structural" approach to analyzing thecompetitive effects of bank mergers. In this approach,the likely competitive effect of a proposed bankmerger is presumed to be indicated by the merger'seffect on such structural variables as market concen-tration, market shares, and the number of competitorsin the banking market.5 The structural approachrequires that both a product market and a geographic

3. United Slates v. Philadelphia National Bank, 374 U.S. 321(1963).

4. The Court noted that, among other products and services,demand deposits, trust administration, and various types of credit suchas commercial lending are uniquely supplied by commercial banksand constitute key elements of the banking cluster.

5. Other approaches to merger analysis may emphasize the likelyeffects on the behavior of the merger participants and may exam-ine business strategy, activity, and pricing as guides to assessingcompetition.

market be defined in order to determine whetherthe merging banks provide substitutable products toessentially the same group of customers. If they do, itcan be assumed that they operate within the sameproduct and geographic markets and that their mergerwould lessen competition.

For such mergers, regulators must then determinewhether the loss of competition may be substantial.Generally, the expected loss of competition isassumed to be serious if the market share of thesurviving bank would increase significantly, to a highlevel, and if the merger would increase marketconcentration substantially, to a high level. In sucha case, banking regulators would be unlikely toapprove the merger unless significant mitigatingfactors were present. Even if such a merger wereapproved by a bank regulatory agency, it could bechallenged by the Department of Justice. To mini-mize regulatory uncertainty and to provide a moreconsistent interpretation of merger-induced structuralchanges within banking markets, the Department ofJustice has for many years published formal guide-lines that indicate which mergers are likely to exceedits antitrust standards. To a large extent, the bankingregulators have adopted these standards as part oftheir merger analyses.

A significant aspect of Philadelphia National Bankwas that it established legal precedents for defining asingle product market (the "cluster" of products andservices) and for defining geographic banking mar-kets as being local in extent. Since 1963, these defini-tions have been used widely and have been supportedby both agency and court decisions.6 However, thedefinitions have been controversial, particularly inrecent years as the number of nonbank firms thatprovide bank-like financial products on a nationwidebasis has increased. Some observers believe that therelevant product market in banking is no longer asingle cluster of products and services provided onlyby commercial banks. Nor do they believe that geo-graphic banking markets are local in extent. Rather,they argue, banks provide multiple and largely dis-crete financial products in broad geographic marketsthat in some instances are national in extent. Notsurprisingly, some of these observers believe that abanking antitrust policy based on the cluster productmarket and local geographic markets is no longerrelevant and does not reflect competitive realities.Above all, they argue, current policy may be undulyrestrictive.

6. See. for example, United States v. Marine Bancorporcnion, 418U.S. 602 (1974): and United States v. Central Slate Bank, 621 F. Supp.1276, 1292 (W.D. Mich., 1985), Aft"d 817 K2d, 22 (6th Cir., 1987).


706 Federal Reserve Bulletin C September 1998

1)1! n A 7 . W / , V IX Mi.R(,!.R ,\X\[YSIS

Federal banking regulators and the Department ofJustice commonly take the structural approach inassessing the competitive effects of proposed bankmergers, and their methods of analysis are generallysimilar. Some differences exist, however. One impor-tant example is that the Federal Reserve continues todefine the product market as the "cluster" of prod-ucts and services described in Philadelphia NationalBank, whereas the Department of Justice typicallydisaggregates the product market into two or morecustomer classesusually banking services for con-sumers and banking services for small businesses(mainly lending).

The separation of customer classes typically resultsin some differences in the way proposed mergers areanalyzed. Among the aspects of merger analysis thatmay be affected are the definition of geographicmarkets, the competitive role accorded thrift institu-tions and nonbanking firms, and divestiture policy.Perhaps most important, the separation of customerclasses affects the choice of data with which to calcu-late concentration, a key variable in bank mergeranalysis. In particular, whereas the Federal Reserverelies on commercial bank and thrift deposits as themeasure of the cluster of products and services, theJustice Department may use business lending and ser-vices, or it may use commercial bank deposits only.

These differences in analytical methods stemlargely from differing assumptions about bank behav-ior, consumer preferences, and the competitiveness ofbanking markets. The use of deposits as a proxy for abank's ability to provide the cluster of products andservices to customers embodies two key assumptions.One assumption is that the product mix of a bank is amatter of managerial prerogative, and that this mix,which includes business loans, can be changed easily.The other is that all banks within a geographic bank-ing market influence the pricing of products, even ifthey do not offer an identical range of products andservices.

The practice of disaggregating the product marketrests largely on the assumption that retail bankingservices and business banking services (and possiblyothers) are distinct products that face different supplyand demand conditions. For example, competitionto provide the financial services used by small busi-nesses (primarily small business lending) is oftenregarded as weaker and more locally limited thancompetition to provide the retail services used byhouseholds. Thus, mergers that affect competitionwithin the business customer class may require closerantitrust scrutiny.

1 HE CRA DATA

The data on small business loans used in this analysiscame primarily from depository institutions report-ing under the Community Reinvestment Actindependent commercial banks and savings institu-tions (savings banks and savings and loan associa-tions) with total assets of $250 million or more andinstitutions of any size if owned by a holding com-pany with total assets of $1 billion or more.7 TheCRA data cover the 1996 originations of small busi-ness loans by 1,871 institutions1,460 commercialbanks and 411 savings institutions.8 These institu-tions ("CRA reporters"), which constitute 16 percentof all U.S. commercial banks and savings institu-tions, reported having originated 2.4 million smallbusiness loans in 1996 totaling $145 billion. Theseloans accounted for about two-thirds of the totaldollar volume of small business loans made by alldepository institutions in the United States in thatyear.

Included among the CRA reporters are bankswhose primary activity is credit card lending. Twelvecredit card banks reported having originated creditcard small business loans in 1996.9 These insti-tutions, though small in number, can have a largeeffect on analytical results, particularly in urbanmarkets. Overall, they accounted for 30 percent ofall small business loans reported under the CRA,though only 2.9 percent of the dollar volume of suchloans.

The CRA loan data are geocoded, that is, reportedby geographic location of the borrower (census tractor block number area), making it possible to aggre-

7. Some loan data came from other sources; see footnote 11.For additional information on the new CRA data, see Raphael W.

Bostic and Glenn B. Canner, "New Information on Lending to SmallBusinesses and Small Farms: The 1996 CRA Data," Federal ReserveBulletin, vol. 84 (January 1998), pp. 1-21.

8. Small business loans arc defined as commercial and industrialloans having an original amount of $1 million or less. This definitionof a small business loan for CRA purposes is the same as that used forthe Report of Condition and Income (Call Report), the quarterly reportsubmitted by banks to their primary regulator and the FD1C, and forthe Thrift Financial Report, the quarterly report submitted by savingsand loan associations to the Office of Thrift Supervision. By focusingon loan amount, the term refers to a small loan made to a business ofany size; such a loan may or may not be a loan to a small business.See box "When Is a Small Business Loan Not a Loan to a SmallBusiness?"

9. The CRA regulations direct institutions that issue credit cards tothe employees of small businesses to report all of the credit card linesopened on a particular day as a single business loan, with the "amountof the loan" equivalent to the sum of the available credit lines of thosecredit card accounts opened on that day.



When Is a Small Business Loan Not a Loanto a Small Business?

"Small business loans" are often not loans to smallbusinesses but, rather, small loans to businesses that arequite large. This distinction can be important to policy-makers seeking to preserve or create credit opportunitiesfor small firms.

For this article, small business loans are defined asthey are for reporting by lenders under the CommunityReinvestment Act (CRA reporters)commercial andindustrial loans of $1 million or less extended to busi-nesses of any size. However, because CRA reportersmust indicate separately the number and dollar volumeof small business loans made to firms having annualrevenues of $ I million or less, the data can be used todetermine the proportion of small business loans made tosmall businesses (when a small business is defined as onehaving annual revenues of $1 million or less).1

In urban markets, 54.4 percent of small business loansoriginated in 1996 by CRA reporters were made to"small" firms. In-market CRA reporters made 60 percentof their small business loans to small firms, comparedwith only 47 percent for out-of-market CRA lenders.Thus, it appears that out-of-market lenders make a greaterproportion of their small business loans to "large" firms(those having annual revenues of more than $1 million).One possible explanation for this finding is that firmseventually outgrow their local banking institutions andneed to go out of market to obtain financing, particularlywhen they are in small markets. For CRA reporters as agroup, the proportion of small business loans extendedto firms having annual revenues of $1 million or lessincreases as market population decreases. This inverserelationship probably reflects the relative scarcity of la/gebusinesses in smaller markets.

Similar results were found for rural markets; however,the proportion of small business loans made to smallfirms across rural markets of all sizes was somewhathigher. In rural markets. CRA reporters made 68 percentof their small business loans to firms having annualrevenues of $1 million or less. As with urban markets,in-market reporters made a higher proportion of theirsmall business loans to smaller firms (78 percent) thandid out-of-market CRA reporters (45 percent). Also aswith urban markets, the data generally show an inverserelationship between market population and the propor-tion of small business loans extended to small firms.

1. There is no commonly accepted criterion defining a "small busi-ness." The 1993 National Survey of Small Business Finances (sponsoredby the Federal Reserve Board and the U.S. Small Business Administra-tion), which defines small businesses as those having fewer Ihan 500employees, notes that about 84 percent of all small businesses haveannual revenues of less than $ I million. For a discussion of that survey,see Rebel A Cole, John D. Wolkcn, and R. Louise Woodbum, "Bank andNonbank Competition for Small Business Credit: Evidence from the 1987and 1993 National Surveys of Small Business Finances." Federal ReserveBulletin, vol. 82 (November 1996), pp. 983-95.

gate lending data for in-market and out-of-marketCRA reporters by geographic market.10

SMALL 3USIMSS !.I:.\;MN(_; ACTIVITY

The CRA data for urban and rural markets wereanalyzed separately because such markets generallyhave different economic characteristics, structures,and degrees of competition.

Urban Markets

Much of the merger activity in recent years has beenin large metropolitan banking markets. Proponents ofmergers often argue that out-of-market lenders are animportant source of small business credit in theselarge markets but that their influence on competitionis largely overlooked. To assess the competitiveimportance of out-of-market lenders in these markets,the data were aggregated across metropolitan statisti-cal areas (MSAs)the 313 urban population centersin the United States that are frequently used in eco-nomic research and merger analysis as approxima-tions for urban geographic banking markets.

Presence of Small Business CRA Lenders

Every MSA has at least one in-market institutionreporting small business lending (CRA reporter), andthese reporters constitute a sizable proportion of thetotal number of institutions within these markets(table 1)." Perhaps more important, the average num-ber of out-of-market reporters in MSAs of all sizes isquite large relative to the number of all in-marketinstitutions. In fact, for all but the largest MSA sizecategory, the number of out-of-market CRA reportersexceeds the number of in-market firms. This is apotentially important finding because it supports theclaims of some merger advocates that, on average, arelatively large number of out-of-market banks and

10. In addition. CRA reporters must indicate the number and dollarvolume of their small business loans by size class ($100,000 or less,$100,001-$250,000, or $250,001-$ I million) and the number anddollar volume of loans to firms having annual revenues of $1 millionor less.

11. Data for institutions not reporting small business loans under(he Community Reinvestment Act ("nonreporters") were taken fromor derived from the Call Report, the Thrift Financial Report, and theSummary of Deposits report (which is submitted both by banks and bysavings and loan associations to their primary regulator and theFD1C). Population data used to categorize both urban and ruralmarkets by size were derived from 1994 regional economic accounts,Bureau of Economic Analysis, U.S. Department of Commerce.


708 Federal Reserve Bulletin LI September 1998

I . A v e r a g e n u m b e r o f s e l e c t e d l y p t ' s o f l e n d i n g i n s l i l u i i o n s p e r u r b a n m a r k e t , b y t y p e i>f i n s t i t u t i o n a n i l M / L 1 n f m a r k e t .

Type of lending institution

All in-markct institutionsIn-market CRA reportersOut-of-market CRA reporters

Excluding credit card banks

MEMORatio of in-market CRA reporters to all

in-market institutionsRatio of out-of-mnrke! CRA reporters to all

in-market institutionsRatio or out-of-market CRA reporters to all

institutions'Number of markets

Market population (thousands)More than

5.000 1,000-5.000 500-999 250-499Less than

250

All urbanmarkets

203 60 28 19 13 2758 18 11 9 6 10

105 65 39 27 19 32103 63 37 25 17 30

.29 .30 .39 .47 .46 .37

.52 1.08 1.39 1.42 1.46 1.19

.34 .52 .58 .59 .59 .544 51 42 70 146 313

Noi!;. Urban markets are metropolitan statistical areas as defined by the U.S.Department of Commerce. All markets have at least one in-market CRAreporter.

I. Denominator is all in-market institutions plus out-of-market CRAreporters.

savings institutions make small business loans inurban markets.

Of course, the mere presence of out-of-marketsmall business lenderseven in large relativenumbersdoes not necessarily demonstrate that suchfirms are an important competitive force within urbanmarkets. They may have a relatively small presencein terms of number or dollar volume of small busi-ness loans originated.

Number of Loans

Although in-market CRA reporters typically consti-tute less than half of all in-market firms (table 1, firstmemo item), these reporters, in each MSA size cate-gory, account for about two-thirds of the estimatednumber of small business loans made by all in-marketfirms (table 2).12 The dominance of in-market CRAreporters is not particularly surprising: Because ofthe asset-size criteria for reporting under the CRA,reporters tend to be the larger firms within bankingmarketsand in many markets these institutions arethe most active small business lenders.

A more important observation is that the averagenumber of small business loans originated by out-of-market reporters is, for each MSA size category,considerably smaller than the average number ofsmall business loans originated by in-market firms.This result is noteworthy because out-of-market lend-ers outnumber in-market lenders in all but one MSAsize category (table 1). Thus, out-of-market CRA

12. Small business loan originations by each in-market nonreporterwere estimated by (1) multiplying the ratio of its small business loansoutstanding io its deposits by its local, in-market deposits and then(2) multiplying (hat product by 60 percent, the approximate ratio ofsmall business loan originations to small business loans outstandingfor all CRA reporters.

lenders account for a considerably smaller number ofloans than might be suggested by their presencerelative to in-market firms.

Interestingly, not only is the share of small busi-ness loans originated by out-of-market reporters con-siderably less than that originated by in-market firms(in terms of number made), but this market sharedecreases steadily as the population of the MSAdecreases. The data show, therefore, that as urbanmarkets decrease in population, so does the impor-tance of out-of-market lenders, as measured by theirshare of the small business loans originated. As willbe seen later, this relationship also exists for ruralmarkets of different sizes.

The direct relationship between market populationand the market share of out-of-market firms has sev-eral possible explanations. One is that out-of-marketlenders simply view smaller markets as being poten-tially less profitable and, therefore, expend relativelyless effort on developing a lending business in suchmarkets. Another is that businesses in small marketshave strong and long-standing ties to local lendersthat are not easily competed away by out-of-marketfirms. A number of observers have noted that thecredit conditions faced by a small business are likelyto be influenced by the nature of the borrower-lenderrelationship. In particular, when credit relationshipsare of long standing, lenders are likely to have betterinformation about borrowers, and this knowledgemay result in more favorable terms on business loans.Thus, a small business borrower probably has anincentive (in addition to convenience) to develop andmaintain a close relationship with a lender that isbased in its banking market.13

13. This issue was recently discussed in Katherine Samolyk,"Small Business Credit Markets: Why Do We Know So Little aboutThem?" FDIC Banking Review, vol. 10, no. 2 (1997), pp, 14-32.Digitized for FRASER

http://fraser.stlouisfed.org/ Federal Reserve Bank of St. Louis


Average number ul small business loans originati 'il por urban market , by l \ pc of k'luliriLL institiuiiviami si/I ' of markel , I'Wn


All in-market institutions'Ill-market CRA reportersOut-of-maricet CRA reporters

Excluding credit cam banks

Mf.MoRatio of loans by in-iaarket CRA reporters

to loans by all in-markel institutionsRatio of loans by out-of-market CRA reporters

to loans by all in-markel institutionsRatio of loans by out-of-market CRA reporters

to loans by all institutions3Excluding credit card banks

Market population (thousands i

More than5,000 1.000-5.000 500-999 250-499

Less than250

All urbanmarkets

44.291 14.409 5.601 3J38 1,775 5,23929.443 9,646 4.106 2.340 1,140 3.55429.579 7.370 2,297 1,039 466 2.3374,728 1.321 504 244 121 454

.66 .67 .73 .70 .64 .68

.67 .51 .41 .31 .26 .45

.40 .34 .29 .24 .21 .31

.10 .08 .08 ,07 .06 .08

NOTF.. See general note to table 1.I. Covers loans reported by CRA reporters and estimated loans by

nonreporters.

When the small business loan originations of creditcard banks are deducted from the out-of-markettotals, the number of loans made by out-of-marketCRA lenders falls sharply (table 2). For example, inthe largest MSAs, the average number of small busi-ness loans by out-of-market firms drops from 29,579to only 4,728. All other MSA size categories exhibitsimilar proportional declines.

The effect of these declines, not surprisingly, is tosubstantially reduce the market shares held by out-of-market CRA reporters. For example, for the largestMSAs, the share of small business loans originatedby out-of-market firms declines from 40 percent tojust 10 percent when credit card banks are excluded.Similar sharp declines occur for all other MSA sizecategories. For all MSAs combined, the sharedeclines from 31 percent to just 8 percent.

The competitive significance of credit card banksis not clear. The new CRA data do indicate that theyaccount, on average, for a significant proportion ofsmall business loans within urban markets. However,whether a credit card small business loan is thefunctional and competitive equivalent of other smallbusiness loans is arguable. Certainly, national sur-veys of small business lending have demonstratedthat many small businesses regularly use credit cardsas a source of credit. However, several factors sug-gest that credit card loans may differ from other smallbusiness loans. For example, the average credit cardloan is considerably smaller, and the credit standardsassociated with credit card loans may not be as strin-gent as those for conventional small business loans.

Dollar Volume ol' Loans

The average dollar volume of small business loanoriginations within MSAs declines as MSA popu-lation decreases (table 3). This pattern reflects the

2. Denominator is loans made by all in-murket institutions plus loans madeby out-of-market CRA reporters.

lower absolute level of aggregate loan demand typi-cally associated with smaller banking markets.

A large proportion of the dollar volume of smallbusiness lending by in-market firms is accounted forby in-market CRA reporters (first memo item). Thisfinding is noteworthy because it suggests that thesmall business lending activity of all banks and sav-ings associations based within urban markets can beapproximated rather closely by using only data fromCRA reporters. For example, for all MSAs com-bined, in-market CRA reporters accounted for 88 per-cent of the average dollar volume of all small busi-ness loans originated by in-market lenders (thoughthey originated only 68 percent of the number of suchloans across all MSAs).

Perhaps the most important observation from thedata in table 3 is that out-of-market CRA reportersaccount, on average, for a relatively small proportionof the dollar volume of small business lending inurban markets (third memo item). This finding issomewhat surprising, given that out-of-market CRAreporters generally outnumber in-market firms inurban markets and account for a sizable proportionof small business loans in these markets by number.

Across all MSAs, the average volume of smallbusiness lending by all in-market firms in 1996 was$388.5 million, compared with only $32.8 million forout-of-market CRA reporters. Thus, out-of-marketCRA reporters accounted, on average, for only 8 per-cent of the average dollar volume of small businessloan originations across all MSAs (third memo item).This small market share in terms of dollar volumecontrasts sharply with the share held by out-of-market CRA reporters in terms of number of loans(31 percent, table 2) and their relative presence inurban markets (54 percent of lenders are out-of-market CRA reporters, table 1). Similarly small dollarshares are found for each MSA size category.


710 Federal Reserve Bulletin D September 1998

Avi-rauL- i l u l k u v o l u m e ] s m a la n d \ i / c iif miirki- ' t . I l W d

Millions of dollars

loans oriyinuU'il |vr urban market, by lype ul'


All in-market institutions'In-market CRA reportersOut-of-market CRA reporters


MEMORatio of volume by in-market CRA reporters

to volume by all in-market institutions . . .Ratio of volume by out-of-market CRA

reporters to volume by all in-marketinstitutions

Ratio of volume by out-of-market CRAreporters to volume by all institutions2 ..


Market population (thousands)More than

5,000 1,000-5,000 500-999 250-499Less than

250

All urbanmarkets

3573.8 1.119.5 415.7 224.2 105.6 388.53,424.2 985.1 378.0 201.4 90.6 342.4

351.1 97.3 35.3 16.5 8.6 32.8214.9 65.9 26.1 12.5 7.0 23.0

.86 .88 .91 .90 .86 .88

.09 .09 .08 .07 D8 .08

.08 .08 .08 .07 .08 .08

.05 .06 .06 .05 .06 .06

NOTF.. See general note to table 1.I. Covers loans reported by CRA reporters and estimated loans by

nonreporters.

2. Denominator is volume of loans made by all in-market institutions plusvolume made by out-of-market CRA reporters.

The results for urban markets appear to provideimportant evidence regarding the competitive role ofout-of-market firms in such markets. Specifically, the1996 CRA data show that out-of-market lenders are,on average, both numerous and active small businesslenders compared with in-market firms. However,these lenders account, on average, for only a smallproportion (typically 7 percent or 8 percent) of thedollar volume of small business lending in MSAs ofvarious sizes. Moreover, when credit card banks areexcluded, the share of originations by out-of-marketCRA reporters, in terms of dollar volume, is even less(typically 5 percent or 6 percent).

Rural Markets

Proposals for bank mergers in rural markets oftenraise particularly serious antitrust issues. Typically,rural markets are more highly concentrated thanurban markets and have relatively few competitors.Many rural markets are not attractive for new entrybecause of their small population and modest eco-nomic prospects. Thus, any adverse competitiveeffects resulting from a bank merger are likely topersist. As in urban markets, however, parties toproposed bank mergers frequently argue that com-petition in rural markets is understated because theregulatory agencies and the Department of Justice intheir analyses do not take into account the competi-tive influence of out-of-market banks and savingsinstitutions.

The data considered here provide some evidencerelevant to this argument. In general, they are consis-tent with the data for urban markets in that out-

of-market CRA reporters, on average, outnumberin-market firms. In rural markets, however, out-of-market CRA reporters account for a smaller propor-tion of small business loan originations by num-ber of loans, and a somewhat higher proportion bydollar volume of loans. The higher proportion ofdollar lending suggests that out-of-market lendersare a relatively more important source of competitionfor small business lending in rural markets than inurban markets.

Presence ol Siiin.ll Business CRA Lenders

The majority of rural markets (non-MSA countiesthat have at least one banking or savings institution)are quite small and have a population of less than25,000 (table 4). A large proportion of these mar-kets (68 percent) have at least one in-market CRAreporter.

Rural markets, on average, have six in-marketbanking and savings institutions, and two of the sixare CRA reporters. Not surprisingly, the number ofboth in-market institutions and in-market CRAreporters declines with market population.

The average number of out-of-market CRA report-ers in rural markets (as in urban markets) is largerthan the average number of in-market institutions; forexample, for all rural markets, the average ratio ofout-of-market CRA reporters to all in-market firms is1.67. Moreover, the ratio of out-of-market reportersto in-market firms is inversely related to market size.This relationship, which also characterizes urbanmarkets, indicates that out-of-market competitorshave a relatively greater presence in smaller marketsthan in larger markets.



4. Avcr;ii!c in imrvr t' s d e e t e d types of Icndini! insii t iuions per rural iinn kcE. In type of institution :ind s i /c of m;irkei.


In-markci institutionsIn-market CRA reportersOut-of-market CRA reporters


MEMORatio of in-market CRA reporters to all

in-markel institutionsRatio of oul-of-rnarket CRA reporters to all

in-market institutionsRatio of oul-of-market CRA reporters to all

institutions'Number of marketsNumber of markets with an in-market CRA

reporter

Market population (thousands)

More than 100 50-100 25-49 Less than 25All ruralmarkets

12 9 7 4 66 4 3 2 2

18 14 12 8 1016 12 10 7 8

.50 .44 .43 .50 .33

1.50 1.56 1.71 2.00 1.67

.60 .61 .63 .67 .6326 199 485 1,561 2,271

26 195 450 881 1,552

NOTE. Rural markets are non-MSA counties that have al least one banking orsavings institution. Data are for rural markets that have al leas! one CRAreporter.

1. Denominator is ull in-markel institutions plus oul-of-market CRAreporters.

Number of Loans

The relative number of small business loans made byin-market and out-of-market lenders is a potentiallyimportant gauge for determining the level of compe-tition in a banking market. For rural markets, as forurban markets, the number of small business loansoriginated in 1996 diminished with market popula-tion (table 5). This pattern most likely reflects thelower loan demand associated with smaller marketsand fewer business borrowers of all sizes.

The proportion of in-market loans originated byin-market CRA lenders declines as market populationdecreases (first memo item). For example, the propor-tion of in-market loans made by in-market CRAlenders is 0.65 for the largest rural markets, com-pared with 0.44 for the smallest rural markets. For allrural markets, the ratio is 0.51, somewhat smallerthan the 0.68 ratio for urban markets (table 2).

The share of small business loans made by out-of-market CRA reporters generally declines as marketpopulation decreases (table 5, third memo item). Thisfinding is somewhat surprising, given that the num-ber of such firms relative to the total number ofin-market institutions increases as market populationdecreases (table 4, second memo item). This relation-ship, which also exists for urban markets, may reflecta preference by small firms in small, more concen-trated markets to borrow from local institutions.

Dollar Volume ol' Loans

As in urban markets, in-market CRA reportersaccount for most of the dollar volume of small busi-ness lending by in-market institutions in rural mar-kets, averaging 74 percent across all rural markets(table 6). However, the relative volume of lending by

A v e r a g e n u m b e r o f s m a l l b u s i n e s s l o a n s o r i g i n a t e d p e r r u r a l m a r k e l , h y t y p e


-\\LT:L!JL" dollar vo lume nl small IHI\MII>N loans nriymaU'd per rural market , by lypo nl lendinj; iiisiiitumnand s\/c nl' maikt ' t .Millions of dollars


All in-market institutions'In-markel CRA reporters . . - . . . - . . .Out-of-market CRA reporters

Excluding credit card banks ,*

MEMORatio of volume by in-market CRA reporter*

to volume by all in-market institutions . . .Ratio of volume by oui-of-nwket CRA

reporters to volume by alt uvrnaritfitinsfitutions

Ratio of volume by ouHrt'-martaH CRAreporters to volume by all institutions* ..

Excluding credit card banks ,


Mora than 100 50-100 25-4t Less thai 25All ruralmarkets

53.8 43.7 18,5 6 3 15.646.S JS.S 13.0 4.0 11.5

7.? 4.9 3.1 1.S 2 J6.3 4.3 2.9 1.4 2.3

.817 J l .70 .63 .74

,14 .11 .17 .24 .16

.13 .10 .14 .19 .14.10 .09 .14 .18 .13

NOTF. See general note to table 4.I. Covers loans reported by CRA reporters and estimated loans by

nonreporters.

2. Denominator is volume of loans made by all in-market institutions plusvolume made by out-of-market CRA reporters.

out-of-market CRA reporters is somewhat larger inrural markets than in urban markets. Out-of-marketCRA reporters account, overall, for 14 percent of theaverage dollar volume of small business lending inrural markets (third memo item), compared with8 percent in urban markets (table 3). In the smallestrural markets, out-of-market CRA reporters accountfor an average of 19 percent of the volume of smallbusiness loans.

f Oitt-of-Market CRA LendersAn overall judgment as to whether out-of-marketCRA lenders are an important competitive forcewithin local banking markets (that is, are able toaffect prices and services) is difficult to make and is amatter for additional research. The data indicate thatin both urban and rural markets, the average numberof out-of-market CRA reporters, relative to the aver-age number of in-market institutions, is not trivial. Inurban markets, out-of-market CRA lenders constitute54 percent of all institutions that extend small busi-ness loans, in rural markets, 63 percent. Out-of-market CRA lenders also account for a sizable pro-portion of small business loans, by number, in localbanking marketsan average of 31 percent in urbanmarkets and 17 percent in rural marketsthoughthese percentages fall sharply when credit card banksare excluded.

Out-of-market reporters seem to be less significantwhen small business lending is measured by dollarvolume. In urban markets, they account for only8 percent of small business loan volume, includingloans by credit card banks; in rural markets, they

appear to be more important, accounting, on aver-age, for about 14 percent of small business lendingvolume.

SMALL BUSINLSS LOAN CONCENTRATION

Indexes of concentration, as noted earlier, are a keycomponent of the competitive analyses of bank merg-ers conducted by the bank regulatory agencies andthe Department of Justice (see box on concentrationand the Herfindahl-Hirschman index). The new CRAdata enable the construction of concentration indexesbased on small business loan originations and thecomparison of these indexes with more traditionalindexes based on bank deposits. An analysis of theseconcentration measures provides evidence that is rele-vant to the competitive analysis of bank mergers.

Urban Markets

HHIs for urban markets were calculated in four ways.Two of the measures were based on deposits, and twowere based on small business loan originations. Oneof the deposits-based HHIs was calculated using thedeposits of commercial banking institutions onlythe measure often used by the Department of Justiceto approximate concentration in small business loanmarkets. The other deposits-based measure, "stan-dard deposits," was calculated using commercialbank deposits plus 50 percent of thrift institution(savings association) depositsthe measure typicallyused by the Federal Reserve. A portion of thriftinstitution deposits are included in the latter measurebecause these institutions generally provide competi-



tion for a portion of the "cluster" of banking prod-ucts, namely banking services for households; inmany cases they provide little or no competition forsmall business services.

One of the loan-based HHIs was calculated usingthe dollar volume of loans originated by onlyin-market institutions, including CRA reporters andnonreporters. The other was calculated using thedollar volume of loans originated by in-market insti-tutions plus out-of-market CRA reporters, A compari-son of the two indexes provides a measure of theaverage effect of out-of-market lenders on small busi-ness loan concentration.

In all but one instance, concentration rises as urbanmarket population declines (table 7). This is a com-mon finding and reflects the tendency of smallermarkets to have fewer banking 6rms, each holding arelatively large individual market share. And, as dis-cussed in the box on concentration and the HHI, themathematical properties of the HHI are such thatlarge individual market shares have the effect ofrapidly increasing the level of the HHI.

The inclusion of thrift deposits causes the averageHHI to decline substantially from the level of thebanks-only HHI. For example, for all urban markets,the average banks-only HHI is 1991 and the averagestandard-deposits HHI is 1639.

The disparity between the average banks-only andstandard-deposits HHIs across all urban markets callsattention to the analytical differences between theDepartment of Justice and the Federal Reserve Sys-tem. The Department of Justice has often used thebanks-only HHI as a proxy for small businessloan competition but has only occasionally includedthrift institutions in its HHI calculations. The Fed-eral Reserve, in contrast, typically includes 50 per-cent of thrift deposits in its calculations of concentra-tion. Thus, for any bank merger proposal, these twoantitrust authorities can have different initial percep-tions of market concentration.

The data also show that the loan-based HHIs aresmaller than the deposits-based HHIs for the largesturban markets but increase rapidly as populationdecreases and exceed the deposits-based HHIs for thesmaller urban markets. Significantly, the inclusion oforiginations by out-of-market CRA reporters has apronounced effect in lowering the level of loan con-centration for urban markets. For all urban mar-kets, for instance, the average loan-based HHI fallsfrom 2130 to 1816 when out-of-market lenders areincluded. This finding further supports the notion thatout-of-market lenders may influence concentrationand competition in urban banking markets. Theirimportance in reducing loan concentration, moreover,appears to increase as market population decreases.Comparable declines occur for all but the largest sizecategory of urban markets.

Rural Markets

Because very high HHI values are typical for marketswith fewer than four competitors, the HHIs for ruralmarkets were calculated in two waysfor marketswith one or more institutions and for those with fouror more institutions (table 8). The findings are similarto those for urban markets: Higher levels of the HHI,both deposits-based and loan-based, are associatedwith progressively smaller markets; the effect ofincluding out-of-market CRA lenders in the loan-based HHIs is pronounced; and as the population ofthe market decreases, the deconcentrating effect ofout-of-market lenders appears to increase.

Effects of Data Choice on Measuresof Market ConcentrationIn general, the proportion of urban markets that arehighly concentrated (HHI above 1800) or very highly

M i l l in n i l i , in i m i r k c i v h \ t in c u L ' u l a i u i n ,IMBasts for calculating HHI

Dollar volume of deposit.*Banks-only total depositsStandard deposits'

Dollar volume of haltsSmall business loan originations by ali

in-inarket institutions2Small business loan originations by all

institutions *


More than 5,(XX) 1.000-5.000 500-999 250-499 Less than 250All urbanmarkets

1505 1812 1885 1829 2170 19911009 1360 1528 1560 1815 1639

67S 1364 1707 2006 2617 2130

576 1160 1443 1743 2220 1816

1. Commercial bank deposits plus 50 percent of thrift institution deposits.2. Covers loans reported by CRA reporters and estimated loans by

nonreporters.

3. Covers loans made by all in-market institutions and by out-of-market CRAreporters.


714 Federal Reserve Bulletin September 1998

A\LTLIUL' Hil l in rural m a r k e d , hv basis for ca lcula t ion, n u m b e r of irixiiuilinib in market , and si/.o ot marke t .

Basis for calculating HHI andnumber of institutions in market

Dollar volume of depositsBanks-only total deposits

One or more institutions

Standard depositsOne or more institutions

Dollar volume of lamsSmall business loan originations by all

in-mnrket institutions

Small business loan originations by oilinstitutions

Four or more institutions


More than 100

26812660

19681964

25802580

30592059

50-100

25562530

21462113

32693245

25812575

25-49

28882810

25322452

44254306

31093023

Less than 25

47843172

45262867

67685563

44533734

All ruralmarkets

41452955

38442611

58124756

39553305

NOTE. See notes to table 7.

concentrated (HHI above 2200) can vary consider-ably depending on the choice of data used in the HHIcalculation (table 9). In rural markets, however, thechoice of data matters much less because the vastmajority of rural markets are highly concentratedregardless of which of the four HHI measures isconsidered.

In both urban and rural markets, the standard-deposits HHI generates the largest proportion ofunconcentrated or moderately concentrated markets(HHI of 1800 or less). For example, 64 percent ofurban markets are so rated when HHIs are calculatedusing bank deposits plus 50 percent of thrift deposits,compared with only 46 percent when only bankdeposits are used.

Urban markets are most concentrated when cal-culations are based on the small business lendingof in-market lenders. For example, 60 percent of allurban markets are highly or very highly concentratedwhen HHIs are based on the dollar volume of smallbusiness loans originated by in-market institutions.The proportion falls to 43 percent when small busi-ness loans originated by out-of-market CRA report-

ers are included. The same general relationship existsfor rural markets, but the vast majority of thesemarkets are very highly concentrated regardless ofthe way the HHI is measured.

CONCLUSIONS AND LIMITATIONS

Overall, the new CRA data provide support for theview that out-of-market lenders may be an importantsource of small business lending in many markets.The data indicate that out-of-market lenders are rela-tively numerous throughout both urban and ruralbanking markets, and in most markets they outnum-ber in-market institutions. Also, in many markets,out-of-market lenders appear to account for a sizableproportion of small business loans, by number,although their share of such loans declines steadilywith market population. Notably, if the loans of creditcard banks are excluded, the average market share ofloans provided by out-of-market lenders declines tohalf or less, suggesting that credit card loans accountfor a large proportion of the lending activity of out-of-market reporters.

Pcrcentaue Jisinhution of urban and rural markets bv decree at concenlnttiiui, J

Basis for calculating HHI

Dollar volume af depositsBanks-only total depositsStandard deposits

Dollar volume of loansSmall business loan origination!!

by all in-market institutions . . . . .Small business loan originations

by all institutions . .

Unconcemrated(HHIM50)

Urban Rural

Moderatelyconcentrated

ami looo-isoo)Urban Rural

Highlyconcentrated

(HHI 1801-2200)Urban Rural

Very highlyconcentrated

(HHI above 2200)

Urban Rural

4.2 .2 42.2 6.6 23.3 8.7 30.4 84.610.2 .3 54.0 11.0 20.8 11.2 15.0 77.5

4.8 8.5 35.1 .7 21.1 2.2 39.0 88.6

13.4 43.5 5.9 14.4 9.7 28.8 84.3

NOTE. Distributions may not sum to 100 because of rounding. See notes totable 7.

Less than 0.05 percent.



The lending activity of out-of-market CRA report-ers seems least significant when the dollar volume ofloans extended by such firms is considered. In urbanmarkets, these firms account, on average, for about8 percent of the dollar volume of small businesslending (6 percent if credit card banks are excluded).In rural markets, their share is higher (14 percentacross all rural markets), suggesting that out-of-market small business lenders may have a greaterinfluence on competition in rural markets than inurban markets.

The competitive role played by out-of-market lend-ers (as gauged by their presence, number of loans,and dollar lending volume) is, however, quite vari-able across individual markets. Thus, the competitiveimpact of proposed mergers must be assessed on amarket-by-market basis. An important complementto the analysis reported here would be research todetermine whether out-of-market lenders influenceloan rates and whether they extend credit to a broadbase of business borrowers within local markets.

Indexes of concentration based on the new CRAdata suggest that concentration for small businessloans is generally higher than concentration fordeposits, particularly in rural markets. However, HHI

levels can vary widely depending on the data under-lying the index. On average, concentration is low-est when HHIs are calculated using "standarddeposits"commercial bank deposits plus 50 per-cent of thrift depositsand highest when calculatedusing the small business loan originations of onlyin-market lenders.

Finally, it should be noted that the 1996 CRA dataused in this study are the first of this sort to becollected, and that data for later years could yielddifferent results. Also, in this study the small businesslending activity of nonreporting in-market institu-tions, although believed to be only a small proportionof overall lending activity in most markets, had to beestimated. The findings also do not account for thelending activity of out-of-market firms that are notCRA reporters. Although nonreporters are smallerinstitutions and are less likely to make loans in mar-kets in which they do not have offices, the inclusionof loans by these firms would probably have a furthermoderating effect on loan-based HHIs. These limita-tions notwithstanding, the new CRA data provide apotentially valuable source of new information that islikely to be helpful in the competitive analysis ofbank mergers. 1A


716

Treasury and Federal ReserveForeign Exchange Operations

This quarterly report describes U.S. Treasury andSystem foreign exchange operations for the periodfrom April through June 1998. It was presented byPeter R. Fisher, Executive Vice President, FederalReserve Bank of New York, and Manager, SystemOpen Market Account. Daniel Osborne was prima-rily responsible for preparation of the report.

Throughout the second quarter of 1998, the Japaneseyen weakened consistently, depreciating 4.1 percentagainst the U.S. dollar and 6.2 percent against theGerman mark. Continued pessimism about the out-look for the Japanese economy pressured the yen toeight-year lows against the dollar. This pessimismcontrasted with the relatively positive outlooks forthe United States and Europe, with the dollar benefit-ing from continued low inflation and a strong domes-tic economy while the mark benefited from optimismtoward the European Monetary Union (EMU). Dur-ing the first two weeks of June, the yen's declineaccelerated, putting downward pressure on otherregional currencies as market participants increas-ingly doubted that effective reforms of the Japanesebanking system and stimulus of the economy wouldbe forthcoming.

On June 17 the U.S. monetary authorities inter-vened in the foreign exchange markets, selling a totalof $833 million against the Japanese yen. The opera-tion, which was divided evenly between the U.S.Treasury Department's Exchange Stabilization Fundand the Federal Reserve System, was conducted incooperation with the Japanese monetary authorities.

Si I'I'ORI iOR U.S. ASSETS FROM SICKS OFMODERATION IN THE DOMESTIC FXONOMYAND KX/'ECTAHONS OF STEADY MONETARYFOLIO

U.S. economic data during the second quarter contin-ued to show strong economic activity and low infla-tion, although increasing signs of slower growthappeared as the period progressed. This environmentsupported expectations of a steady near-term interestrate policy, although at times this sentiment fluctu-

ated. In April U.S. treasuries traded in a volatilemanner, with the benchmark thirty-year bond yieldrising above 6.00 percent after the April 27 publi-cation of a press release reporting that at its Marchmeeting, the Federal Open Market Committee(FOMC) had shifted toward a tightening bias. Thebenchmark bond yield subsequently traded backbelow 6.00 percent after the April 30 release of theemployment cost index, which was interpreted assuggesting lower-than-expected labor cost growth forthe first quarter.

I . S p u l i w t h . i i u i i . 1 r u l e n t ' I l k ' i l n l k i r . i s . \ i i n M i l u 1 . l . ip .niCM.* \ jn; u u l o p i u m i m p l i e d v n l i i l i l i i v . I W S ' H I

U N

14-1

1411

l ih

I2X

124

| _

t1 > 1*11 JKT I . S . dutlar

One-month volatilities

A"Vr\AArV~V ^j

W Split I'Y

Jan. Feb. Mar.

ihiiiiKi' nilf- 1 i

Apr.1998

Percent per year

A - 1 - 18

JA/ INrr 16

r> \ 14

\J\J 12

1 1May June

S O U R C L . F e d e r a l R e s e r v e B a n k of N e w York; R e u t e r s .

2 . S p i n L ' X L ' I K I U ' J L 1 ntlL1 o l l l i c i l o i l a i ' ; I J ; ; I I I I - . I i l v < i ' j n n a n i m i r kid h i n p l n > n - i n i p i i t ' l l \ n l ; i l i l i l \ . I ' f ' S ' H I

Crnn

2.110

I.Mfc

\.nI.XK

I.N4

\M

1.72

in tntirk> fitT I1 Js. dtillur

Spiil t'vcha

i i iJan. Feb.

n("t rale

[Vw\. A.

One-monlh

1Mar. Apr.

1998

volatilities

1 IMay

Percent per year

13

11

V

June

S O U R C E . Federa l R e s e r v e Bank of N e w York; R e u t e r s .


717

3. U.S. Treasury yields. IWS:Q2

U.S. intervention

6.1

5.9

S.7

5.5

Apr. May June

SOUFUT.. Bloomberg L.P.

This outlook of low inflation and moderatinggrowth appeared to be supported by Federal ReserveChairman Alan Greenspan's June 10 testimony onthe state of the economy to the Joint EconomicCommittee of the Congress. In his testimony, theChairman stated, "the current economic performance,with its combination of strong growth and low infla-tion, is as impressive as any I have witnessed in myhalf-century of daily observation of the Americaneconomy." Further reflecting on inflation, the Chair-man stated that although "consumer price inflationhas moved up in the second quarter . . . the rate ofrise remains quite moderate overall."

Chairman Greenspan's testimony reinforced per-ceptions of a favorable U.S. economic outlook amidstvolatility in the financial markets of emerging coun-tries and amidst the resultant decline in risk appetite.Strong demand for treasuries pushed the nominalbenchmark thirty-year Treasury yield to record lows,while two-year through ten-year Treasury yieldsdeclined below the 5.50 percent federal funds targetrate.

WEAKENING OF THE YEN IN RESPONSE TO ADETERIORATING ECONOMIC OUTLOOK IN ASIA

Throughout the quarter, the yen was pressured byanticipation of renewed capital flows out of Japanafter the March 31 conclusion of the Japanese fiscalyear, the beginning of Japan's liberalization of itsforeign exchange laws, and continued weakness inthe Japanese economy and financial system. Anticipa-tion of falling demand for yen assets from Japaneseinvestors overshadowed concerns about potential for-eign exchange intervention and possible reaction toJapan's growing current account surplus. The April 2release of the weaker-than-expected Tankan survey

of business sentiment and Moody's revision ofJapan's sovereign debt outlook from stable to nega-tive on April 3 contributed to perceptions of a wors-ening Japanese economy. Within the first three trad-ing sessions in April, the Nikkei stock index declined6.1 percent, to 15,518, and the yen depreciated from133.18 to levels above 135. The yen subsequentlyregained all of its losses, appreciating to 131.22 onApril 8, after comments by Japanese politicians indi-cating forthcoming economic stimulus measures.

On April 9 the Japanese government announceddetails of its fiscal stimulus package, including4 trillion in temporary tax cuts. The yen declined to133.64 as market participants concluded that theplanned stimulus measures would not adequatelyrevive the Japanese economy. Later, during the earlyNew York session, the yen rose against the dollaron the basis of reports that the Japanese monetaryauthorities had intervened in the marketbuyingyen and selling dollars. Japanese Finance MinisterMatsunaga confirmed the intervention, announcingthat Japan had taken "decisive action" in the foreignexchange market. U.S. Treasury Secretary RobertRubin stated, "We welcome Prime Minister Hashi-moto's announcement of steps to stimulate the Japa-nese economy." and noted that "what is crucial isthat Japan move quickly to put in place a strongprogram." He also said that "we share the concernexpressed by the Japanese Prime Minister aboutrecent weakness in the yen, and in that context wewelcome the action undertaken by the Japaneseauthorities in the exchange market to support thevalue of the yen." On April 10 the yen appreciated to127.38 amid reports that the Japanese monetaryauthorities were intervening heavily in Asian foreignexchange markets.

Despite the yen's gains, sentiment toward the yencontinued to remain negative, with twelve-month riskreversals continuing to be skewed toward yen puts,reflecting a higher cost of protection against a sharpdepreciation of the yen against the dollar. Subse-quently, the Group of Seven (G-7) finance ministersand central bank governors met on April 15 andissued a statement emphasizing the importance ofavoiding excessive yen depreciation and encouragingsteps to stimulate Japanese domestic demand andreduce external imbalances.

Meanwhile, pressure on other Asian currency andcapital markets re-emerged, as social unrest in Indo-nesia and potential military tensions in the Indiansubcontinent led to a diminishing investor appetitefor risk. In Indonesia, the Jakarta JCI stock index fell17.6 percent during the quarter, and the rupiahdeclined to record lows above IDR 16,500 per dollar.


718 Federal Reserve Bulletin September 1998

4. Asian equity markets, IWS:Q2Index. April 1 100

US. intervention

Japan

Korea

ThailandJ

Apr. May1998

June

SOURCE. Bloomberg L.P.

100

70

xr- 60

(i. Moscow, Times slock index. l'WS:Q2

80

Apr. May1998

June


With the KOSPI stock index declining 38.1 percentduring the quarter, Korean markets weakened oncontinued signs of economic contraction and height-ened labor unrest. In Hong Kong, warnings of thepossibility of recession from Hong Kong officials anddata indicating that first-quarter GDP had contracted2 percent led to a 25.8 percent decline in the HangSeng stock index during the second quarter. HongKong one-month interbank rates climbed as high as20.5 percent, and implied yields on three-month for-ward contracts for Hong Kong dollars rose to quarterhighs near 15 percent in mid-June.

Declines in Asian markets and evidence of eco-nomic contraction in several Asian countries helpedfocus investor attention on the weaknesses in othermarkets. The Russian Moscow Times stock indexdeclined 53.5 percent during the quarter and yields

5. Selected CUMVMCV values a.c.iinsl the U.S. dollar. ll>l).S:(J2

Iraki. April 1 100

U.S. intervention

100Canada

Apr. May1998

on Russian debt instruments rose higher than 80 per-cent on the basis of concerns about the government'sinability to correct its fiscal imbalance and showprogress on structural reforms, and concerns aboutthe effect of declines in commodity prices, partic-ularly oil, on Russian exports. Commodity pricedeclines also pressured the currencies of Canada,Mexico, Australia, and South Africa, with a numberof these currencies reaching new record lows. TheCanadian dollar depreciated nearly 3.5 percent duringthe quarter and reached CAD 1.4735 per U.S. dollaron June 15, because of concerns about the effect onthe Canadian economy of diminishing exports toAsia and higher nominal U.S. interest rates.

In Japan, economic data releases continued to rein-force negative market sentiment. Japan's unemploy-ment rate reached 4.1 percent in April, its highestlevel since World War II. The yen depreciated beyondthe 136 level after the Birmingham Group of Eight

7. Japanese henehnwk bond yield. !WR:Q2

U.S. intervention

- 1.6

1.4

1.2

Apr- May1998

June

SOURCE. Bloomberg L.P. SOURCE. Bloomberg L.P.


Treasury and Federal Reserve Foreign Exchange Operations 719

summit meeting in mid-May. This meeting and theG-7 finance ministers' meeting of the prior weekwere viewed by many market participants as evi-dence that the G-7 had moved toward a tolerance ofyen depreciation that had not been present at theApril 15, G-7 meeting of finance ministers and cen-tral bank governors. In late May, the yen declined toa seven-year low of 139.25, and the Japanese bench-mark bond yield established record lows. Subsequentcommentary from numerous Asian and Europeanmonetary officials stating concern over the weakyen's effect on global stability underscored marketpessimism. Further yen declines stalled ahead of theJune 10, G-7 meeting of deputy finance ministers, butresumed upon news that the meeting did not discussways to address the yen's decline.

The yen's depreciation accelerated after early June.Concern about a rapid move beyond the 140 levelbecame a focal point for further declines in othermarkets, starting in Asia and extending through othermarkets. Market nervousness about the risks of afurther round of currency depreciation in Asiaincreased after certain Chinese officials voiced dis-satisfaction with the effect of yen depreciation onChinese trade and economic conditions. Also, mar-ket participants initially viewed Secretary Rubin'sJune 11 remarks in a congressional hearing on eco-nomic conditions in Japan as stressing the negativeeffects of Japanese economic fundamentals anddownplaying prospects for coordinated interventionin the foreign exchange market. The yen rapidlydeclined to an eight-year low of 146.67 on June 15,and one-month implied volatility on dollar-yenoptions reached multiyear highs. Comments fromsenior Japanese officials that the dollar-yen exchangerate did not reflect economic fundamentals andrepeated warnings of "decisive action" from Japa-nese authorities did little to stem the yen's slide.

PURCHASES OF YEN IN THE MARKET H)(is. MONETARY AUTHORITIES

The June 16 Tokyo trading session was notably vola-tile, with the yen initially declining to a new low of146.78, then abruptly retracing its decline as a Japa-nese wire service indicated that U.S. officials wereconsidering a trip to Tokyo to discuss Japan's eco-nomic situation. As market participants reacted to thepossibility that measures to address yen weaknesswere under consideration, the yen continued tostrengthen during the June 17 Tokyo trading ses-sion, gaining more than one yen from Tuesday'sNew York close of 143.30.

X. IloMui \cn o n e - i m n i l l ) . i iul l r i sk

\A

\

1

SOURCE.

Twelve-month /'A ^ v ^

f Dollar calls favored / 1

\ Yen calls favored /

Apr.

J.P. Morgan.

1 One-month

US. intervention

1

V

/VMay June1998

Perctnl

r 0.5

n' \ - 0-5

1.0

1

During the June 17 morning session in New York,the foreign exchange desk intervened on behalf ofthe U.S. monetary authoritiesselling dollars andbuying yen. The operation began at 7:55 a.m., withthe dollar trading at 142.21, and the desk operatedintermittently until about 9:20 a.m. As the operationbegan, Japanese Prime Minister Hashimoto issued astatement in which he pledged to "make every effortto restore [the Japanese] banking system to health[and] to achieve domestic demand-led growth. . . ."Japanese Finance Minister Matsunaga further elabo-rated on the Prime Minister's statement, promising to(1) dispose of banks' bad assets more aggressively,(2) accelerate the restructuring of financial insti-tutions, (3) improve the transparency of Japanesebanks, and (4) enhance bank supervision standards.

Later, Treasury Secretary Rubin issued the follow-ing statement:

This morning, the Prime Minister of Japan outlined hisGovernment's plans to restore the health of the Japanesefinancial system and to strengthen domestic demand. Welook forward to implementation of a comprehensive actionprogram that will create the conditions that are essential fora healthy and prosperous economy. Japan has the financialresources and the capacity to deal with the challenges itfaces. Asia and the international community as a wholehave a large stake in Japan's success.

In the context of Japan's plans to strengthen its econ-omy, the U.S. monetary authorities operated in theexchange market this morning in cooperation with themonetary authorities of Japan. We are prepared to continueto cooperate in exchange markets, as appropriate.

Over the course of the morning, the U.S. monetaryauthorities sold $833 million against the yen. Theamount was evenly split between the U.S. TreasuryDepartment's Exchange Stabilization Fund and theFederal Reserve System. The intervention was coor-dinated with the operations of the Japanese monetary



authorities. The yen continued to appreciate through-out the remainder of the New York session, climbingmore than five yen to intraday highs near 136 beforeclosing at 136.51.

After the intervention, the yen initially extended itsgains, largely on market expectations of further eco-nomic and financial policy developments in Japan inthe near term. However, in market participants' view,such expectations diminished when the June 20 meet-ing of the Deputies of the Manila Framework Groupand the G-7 Deputies in Tokyo concluded with nonew public commitments from Japanese officials.The dollar ended the quarter at 138.88.

STRENGTHENING OF THE MARK AGAINST BOTHTHE DOLLAR AND THE YEN RESULTING FROMCONFIDENCE IN THE EMU AND SIGNS OFGROWTH IN CORE EUROPE

The mark strengthened against both the dollar and theyen during the second quarter. Against the dollar, themark traded within a DM 1.76-1.86 range throughmost of the period, climbing more than 10 pfennigsfrom early quarter lows but partially retracing thesegains toward the latter half of the quarter. The mark'sinitial appreciation resulted from a removal ofresidual uncertainty regarding the EMU after formalselection of eleven first-round members at theMay 1-3 European Union summit. Confidence in thesingle currency was not significantly diminished bythe protracted debate over the selection of WimDuisenburg as the European Central Bank's firstpresident, nor by the debate over the length of histerm, which some viewed as potentially compromis-ing the central bank's future credibility. Europeanasset markets benefited from the positive EMU out-look as well as from shifts of portfolio preferencesfrom riskier asset markets. Equities in Germany and

'). CJcmiiin bonchniuik bond yiekl, IWS:Q2Percent

Apr. May1998

June


France continued to rise, with the German DAX andthe French CAC-40 stock indexes up 14.6 percentand 8.5 percent for the quarter and 38.3 and 40.2 per-cent for the year respectively. European credit mar-kets rallied, with the yield on the German ten-yearbenchmark bond reaching record lows.

In early May, expectations for tighter monetarypolicy in Germany contributed to further markstrength. After the selection of the founding membersof the EMU, Bundesbank officials alluded to takingEMU considerations into account more when con-ducting monetary policy. These comments height-ened expectations for a German rate increase, asmarket participants speculated that the Bundesbankmight seek to quell inflationary pressures in certainEMU member countries.

Toward the latter half of the quarter, the mark gaveback nearly half of its gains against the dollar asexpectations for tighter monetary policy in Germanydiminished given the benign inflation environmentin core European countries, renewed weakness inemerging markets, and concerns over developmentsin Russia. Reflecting these developments, the impliedyield on the December 1998 Euromark contractdeclined 22 basis points from peak levels in mid-April. The dollar ended the quarter at DM 1.8080,after having declined 2.2 percent.

TREASURY AND FEDERAL RESERVEFOREIGN EXCHA NGF. RESER VFS

At the end of the quarter, the current values of theGerman mark and Japanese yen reserve holdingstotaled $17.3 billion for the Federal Reserve Systemand $13.9 billion for the Exchange StabilizationFund. The U.S. monetary authorities invest all oftheir foreign currency balances in a variety of instru-ments that yield market-related rates of return andthat have a high degree of liquidity and credit quality.A significant portion of these balances is invested inGerman and Japanese government securities helddirectly or under repurchase agreement. As ofJune 30, outright holdings of government securitiesby U.S. monetary authorities totaled $7.1 billion.

Japanese and German government securities heldunder repurchase agreement are arranged eitherthrough transactions executed directly in the marketor through agreements with official institutions. Gov-ernment securities held under repurchase agreementby the U.S. monetary authorities totaled $11.3 billionat the

frb_091998

Documents

Transcript of frb_091998