3

David C. Wheelock

David C. Wheelock is a senioreconomist at the Federal ReserveBank of SL Louis. Kevin White provided research assistance.

us the Banking Industry inDecline? Recent Trends andFuture Prospects from a

Historical Perspective“Banking is essential to a modern economy. Banks are not.”

—Edward Furash (1993)

OMMERCIAL BANKS ENJOYED recordprofits in 1992 and during the first half of 1993.Many observers believe the industry’s futuremay be far less bright, however, if banks con-tinue to lose market share to other intermediariesand providers of transaction services. Some poli-cy makers, including Federal Reserve BoardChairman Alan Greenspan and Comptroller of theCurrency Eugene A. Ludwig, have questionedwhether banks will be able to maintain theirrole as providers of financial services in the fu-ture without significant changes in regulation.’

A shrinking banking industry may reflect areallocation of resources toward more efficientuses, and, hence, should not necessarily beviewed as undesirable. But, because banks areheavily regulated, policy makers need to considerwhether their policies are either hastening orinterfering with changes in the size and structure

of the industry. Such changes may impose sig-nificant social costs, since commercial banks areat the heart of the payments system and continueto be important sources of credit for small firmsand other borrowers. Furthermore, changes inthe size or importance of the banking industrymight also affect the ability of the FederalReserve to implement monetary policy, sincemonetary policy is conducted primarily by alter-ing commercial bank reserve balances.2

The banking industry has been on a rollercoaster ride since 1980. From World War IIthrough the 1970s, banks generally had stableearnings and no more than 10 bank failures oc-curred in any year. The banks that did failwere usually tiny and largely unnoticed. Thenumber of failures rose sharply in the 1980s,reaching 206 in 1989. Although the number offailures has since declined and bank profits have

1Greenspan (1993); Ludwig’s remarks before the MerrillLynch Financial Services Conference are reported inBureau of National Affairs (1993).

2Duca (1993) investigates the impact of a diminishing rolefor banks on monetary aggregates and the implications forFederal Reserve control of the money supply.

4

recently been high, the size and structure of theindustry have continued to change dramatically.

The upheaval of the past 10 years and uncer-tain outlook for the industry’s future make thisan appropriate time to put recent changes in thesize of the banking industry in a longer-termperspective. This article first examines the ap-parently diminishing role of commercial banksas intermediaries and providers of transactionservices. Commercial bank shares of U.S. finan-cial assets, commercial lending and transactionaccounts have fallen in recent years, whichsome observers believe reflects an industry inlong-term decline. Others, however, argue thatbanks will remain central to the payments sys-tem and important lenders for large classes ofborrowers, and note that banks are generatingan increasing amount of their income from off-balance-sheet” activities. This article presentssome evidence on both sides of the debate andaddresses two major policy changes that manyobservers believe are needed for banks to remainprominent providers of financial services in thefuture: 1) removal of limits on branch bankingand 2) relaxation of restrictions on the servicesthat banks may offer.

IS THE ROLE OF COMMERCIALBANKS DECLINING?

Commercial banks specialize in the evaluationof credit risks and monitoring of borrowers, aswell as clearing transactions. Traditionally, theyhave been important sources of loans for firms,households and even governments. In addition,banks are integral to the payments system, issu-ing most of the nation’s transaction deposits andclearing domestic and international payments.As of December 31, 1992, the 11,461 U.S. com-mercial banks and trust companies whosedeposits are insured by the FDIC held $3,506

billion of assets, $2,699 billion of deposits andemployed 1,477,827 people.3

Although commercial banks remain the largestsingle class of financial institutions in terms ofassets, banks have lost market share to otherintermediaries and providers of transactionservices.~Dramatic improvements in communi-cations and computer technology have teducedthe cost of performing information-intensive ac-tivities, and thereby permitted growing roles forfirms and markets that provide specializedfinancial services. They have also subjectedAmerican banks to increased competition fromforeign lenders, as have reductions in govern-ment barriers to the flow of goods and capitalbetween countries.~

Figures 1 and 2 illustrate the declining marketshare of banks as intermediaries. Figure 1 showsthat commercial banks’ share of financial assetsheld by all financial institutions has fallen sincethe early 1970s. Indeed, the trend has beendownward since World War H. Commercialbank loans as a share of the short-term debt ofnonfinancial corporations has similarly declined,as figure 2 illustrates.’

One rapidly growing source of funds for largecorporations is the commercial paper market,which has expanded rapidly since the 1960s.Many corporations, especially established firmswith good credit histories, have discovered thatthey can acquire short-term funds less expen-sively by issuing commercial paper than by bor-rowing from banks. Figure 3 shows that the ratioof commercial paper outstanding issued by non-financial corporations to the commercial and in-dustrial (C&J) loans of banks has risen markedlysince 1980.~

While facing new competition for borrowersfrom the commercial paper market and fromnonbank lenders, banks have also faced greatercompetition for funds. In the 1960s, the expand-

‘Federal Deposit Insurance Corporation (1992).~Atthe end of 1992, U.S. commercial banks had total finan-cial assets of $2,774.6 billion, while thrift institutions had$13452 billion, life insurance companies had $1,622Bbillion, other insurance companies had $624.4 billion,private pension funds had $2,349.4 billion, state and localgovernment employee retirement funds had $972.3 billion,finance companies had $607.1 billion, mutual funds had$1,050.2 billion and money market mutual funds had $543.6billion. The source of these data is the Board of Governorsof the Federal Reserve System’s Flow of Funds.

5See Edwards (1993, pp. 9-11), Barth, Brumbaugh and Litan(1992, pp. 59—64) and Kaufman (1991) for further discussionof the sources of increased competition for commercial

banks. Aguilar (1990) documents the growth during the1980s of bank services provided by nonbank firms.

~Atthe end of 1992, nonbank private financial institutionshad total financial assets of $10,012.0 billion. Commercialbank loans to nonfinancial corporations totaled $516.5billion, while the sum of loans and other short-term paperof nonfinancial corporations was $866.0 billion. The sourceof these data is Board of Governors of the Federal ReserveSystem, Flow of Funds.

7At the end of 1992, C&l loans totaled $597.6 billion andoutstanding commercial paper of nonfinancial corporationswas $107.1 billion (Board of Governors of the FederalReserve System, Flow of Funds). Hahn (1993) provides adescription of the commercial paper market.

Figure 1

5

Commercial Banking Industry Share of U.S. FinancialAssets080

0.68-

0.56-

0.44-

0.32 -

Figure 2Commercial Bank Loan Share of All Short-Term Debt ofNonfinancial Corporations0.90—

0.85-

0.80-

0.75-

0,70-

0.65-

0.60-

0.55-

0,50

0.20— I I I I I I I I I I I I

1972 74 76 78 80 82 84 86 88 90 1992

I I I I I I I I I I I I I I I I I I I1972 74 76 78 80 82 84 86 88 90 1992

6

Figure 3Ratio of Commercial Paper of NonfinancialCorporations to Commercial Bank C & I Loans0.20-

0.18-

0.16-

0.14-

0.12-

0.10-

0,08-

0.06-

004-

0.02-

0.00- I I I I I I I I I I I I I I I I I I1972 74 76 78 80 82 84 86 88 90

ing money market offered corporations andwealthy individuals new alternatives to bankdeposits. Inflation and rising interest rates causedextensive deposit outflows from banks becausethe rates that banks were permitted to pay de-positors were limited by interest rate ceilings.Banks responded by introducing negotiable cer-tificates of deposit, which were not subject torate ceilings. Many banks also adopted the one-bank holding company organizational form toacquire funds from the money market by issu-ing debt instruments through the holding com-pany. Still, commercial banks were unable tomaintain their share of the market for financialassets.

The money market mutual fund was perhapsthe most important financial innovation affect-ing the ability of banks to compete for deposits.Introduced in the 1970s, money market mutualfunds offer small depositors the opportunity tohold highly liquid accounts yielding market in-terest rates. Although uninsured and with mini-mum transaction amounts, money marketmutual funds grew rapidly in the 1970s andearly 1980s as reserve requirements and in~terest rate ceilings left banks competitively dis-advantaged. Commercial banks also lost theirmonopoly on the issuance of insured transaction

accounts when thrifts and credit unions beganoffering share accounts, often on more favora-ble terms than banks were permitted to pro-vide. Deregulation of deposit interest rates inthe 1980s enabled commercial banks to competefor funds, but did not reverse the declines inbank market shares for transaction accounts.Access to the payments system remains almostexclusively the domain of commercial banks;mutual funds, thrifts and other financial institu-tions issuing transaction accounts must ultimatelyrely on banks to make payments. Nonetheless,increased competition for transaction accountcustomers helps explain why the banking indus-try’s share of aggregate financial assets has fallen.

Increased competition for traditional bank ser-vices affecting both the asset and the liabilitysides of bank balance sheets has caused numer-ous observers to be pessimistic about the futureof the banking industry. Edwards (1993) goes sofar as to argue that “if our financial markets andinstitutions were being created for the first timein 1990, banks might not be among the survivinginstitutions.” Others, such as Barth, Brumbaughand Litan (1992), Gorton and Rosen (1992) andKaufman (1991) contend that government poli-cies that restrict the geographic location andservices that banks can provide and impose

1992

7

regulations (such as minimum capital and reserverequirements and community investment man-dates) that are not placed on other intermediarieshamper the ability of banks to compete. Theseresearchers argue that the banking industry isdestined to decline substantially in the futurewithout significant regulatory changes.

Other researchers are more sanguine aboutthe banking industry’s future. Boyd and Gertler(1993), for example, argue that much of the in-creased risk-taking by banks and their subse-quent poor performance during the 1980s canbe attributed to the failure-resolution policyknown as “too-big-to-fail.” To reduce the possi-bility that the failure of a very large bank couldlead to a systemic crisis, with depositor runs onmany banks, regulators adopted a policy thattended to protect all of the depositors and oftenother creditors of large banks that failed. Incontrast, uninsured depositors of small banksthat failed were less frequently protected fromloss, and the assets of such institutions weremore often liquidated.

Although the “too-big-to-fail” policy was im-plemented to limit the repercussions stemmingfrom the failure of very large banks, Boyd andGertler (1993) contend that the policy encouragedlarge banks to assume greater risks than theywould have otherwise. The consequence wasthat banks with more than $10 billion in assetshad the lowest average profit rate during1983—91 of any size class, regardless of location.Boyd and Gertler thus applaud recent increasesin hank capital requirements and restrictions onthe use of the “too-big-to.fail” closure policy im-posed by the Federal Deposit Insurance Corpo-ration Improvement Act of 1991 (FDICIA),because both changes should limit the incentivefor large banks to take excessive risks.~

Boyd and Gertler point out that many banksare now generating substantial earnings fromnon-traditional, or off-balance-sheet, sources.The unique position of banks in the paymentssystem and their access to the federal lender oflast resort enhances the role of banks in provid-ing services to nonbank lenders and financial

8WaIl (1993) describes how the “too-big-to-fail” policy and thelikelihood of systemic risk will likely change under FDICIA.

9Remolona and Wultekuhler (1992) report that finance com-pany lending shares increased the most in niches, such aslease financing, where finance companies have traditional-ly played a major role, which also suggests that the largegains made by finance companies during the 1960s maynot signal a long-run trend.

markets, Ironically, banks have played a signifi-cant role in the development of new financialmarkets and services that compete in traditionalbank niches. For example, bank loan guaranteesto issuers of commercial paper have spurredthe growth of the commercial paper market. Be-cause many such lines of credit are provided toback finance company issues of commercialpaper, the decline of bank loans relative to com-mercial paper issues overstates the decline ofthe commercial banking industry’s role in inter-mediation.”

Other off-balance-sheet activities that provideincome for banks include loan sales, loan servic-ing, mutual fund sales and participation in themarkets for derivative securities, such as op-tions and swaps.’°Many banks originate andthen sell loans to third parties before they comedue. Often home mortgages or other commonloan types are bundled and sold in secondarymarkets, a practice known as securitization. Thefee income that banks generate from issuing let-ters of credit, providing loan commitments,securitization and dealing in foreign exchangeand derivative securities has increased morerapidly in the last decade than interest income,as figure 4 illustrates.

The unique position of banks in the paymentssystem and their access to the Federal Reservediscount window also suggest that banks couldremain important lenders in the future. By mon-itoring transaction accounts, or by requiringthat borrower receipts and payments be pro-cessed by the bank, for example, banks can ac-quire useful information about current andpotential borrowers. Even though specialtylenders and financial service firms, such asmutual funds, provide some intermediation serv-ices at lower cost, banks may continue to havean advantage for loans that are especiallyinformation~intensive}1

SHOULD BANKS BE GIVENGREATER POWERS?

Despite the advantages of access to the pay-ments system and Federal Reserve discount win-

‘°SeeNapoli (1992) for a discussion of derivative securitiesmarkets.

‘1Using data from 1978—92, Pulley and Humphrey (1993) esti-mate that the economies from supplying loan and transac-tion services jointly are small.

8

Figure 4Banking Industry IncomeIncome as a percentage of assets

4,00-

3.50—

3.00-

dow, and the substantial growth of income fromoff-balance-sheet activities, many observers, in-cluding lloyd and Gertler, advocate significantchanges in bank regulation. Most state restric-tions on branching have already been removedand regulators allow banks to sell mutual fundsand offer other securities services, Interstatebranching remains largely prohibited, however,as do commercial bank underwriting andownership of most corporate securities.12 In ad-dition, banks are largely prohibited from offeringa variety of related financial services, includinginsurance and real estate services.” Advocatesfor further reduction of restrictions on branch-ing and the services that banks are permitted tooffer argue that deregulation would enable banksto achieve greater diversification and, hence,reduce their chance of failure. Proponents alsocontend that large, diversified banks would bemore efficient and provide financial services at

‘2Several states have recently enacted or are consideringreciprocal legislation that would permit branch offices ofout-of-state banks based in states with similar laws. Theprivilege, however, applies only to state-chartered banksthat are not members of the Federal Reserve System.

‘“Barth and Brumbaugh (1993) provide an up-to-date summaryof commercial bank powers at both the state and nationallevels.

lower cost to the public than presently available,though critics rebuff many of these claims.

A review of all of the arguments in favor ofand against expanded branching authority andderegulation of bank assets and services is be-yond the scope of this article. The followingsections, however, will describe the evolution ofbranching laws and restrictions on bank serv-ices in the United States, focusing especially onhistorical lessons that could inform the currentdebate on these issues.

BRANCH BANKING

Unlike the United States, most countries have asmall number of commercial banks, nationwidebranching and virtually no bank failures. Thehigh costs of communication and transportationmay have made it difficult to operate vast branchbanking networks in the United States during

Neflnterest Income

~~~~nterest income

2.50—

200-

1.50—

1.00-

0.50—1980 82 84 86 88 90 1992

I I I I I I I I I

9

the 19th century, but today the main impedimentto branching is government regulation.14

Where permitted, banks might operatebranches to capture economies of scale, diver’-sification opportunities or lower overhead costs.States that allow branch banking tend to havefewer banks (but not fewer bank offices) percapita than states that restrict branching.’” Abank locates in a particular market only if thereappears to be sufficient demand for it to operateat a profitable scale. Because the minimum pro-fitable scale for a branch is smaller than for anindependent bank, in the absence of branchingresti-ictions, small markets that are served byone or two independent banks might instead beserved by several branch offices.

Proponents of easing branch banking restric-tions contend that branching affords banksgreater opportunity to diversify and, hence, les-sens their chance of failure. Banks whose assetsare comprised largely of loans in a particularregion could be pulled down by an adverse localeconomic shock, whereas a bank with loans inseveral regions could withstand a downturn inthe economy of any one region it serves. Branch-ing restrictions do not necessarily prevent banksfrom diversifying geographically. Banks some-times purchase loans made by banks in otherlocations. In addition, multiple bank holding com-panies and limited facility branches, such as loanproduction offices, are often permitted wherefull-service branches are not, including acrossstate lines.’8 Nonetheless, branching restrictionslimit flexibility and increase the cost of diversifi-cation, and proponents, such as Kaufman (1993),argue that interstate branching would “permiteven greater geographic and product diversifica-tion than interstate holding company bankingand improve bank safety.” Similarly, Clair andO’Driscoll (1991) argue that branching restrictionsexplain why Texas had so many bank failuresduring the 1980s while the banking industry as awhole had record profits. Such arguments werealso made in the 1920s, when thousands of unit

banks failed in rural farming states, even as thebanking industry as a whole was profitable.17

Banking Before World War H

The geographic dispersion of American bankingmarkets and restrictions on branch banking havea long history. In the 19th century, sparselypopulated states often permitted banks to operatewith little capital to ensure the presence of bank-ing facilities in rural areas. Branch banking net-works could have met the demand for bankoffices, but fears that branching would reducecompetition and pull savings away from ruralareas to urban centers made branching politicallyunfeasible. Calomiris (1992a) argues that agricul-tural landowners had an incentive to favor unitbanking to ensure that local banks would con-tinue to lend to them following an adverse localshock. Branch banking organizations, by contrast,could close offices or restrict loans to areas ex-periencing distress. In agricultural states, the in-terests of farmers thus tended to coincide withthose of community bankers and state regulatorsto favor restrictions on branching. Consequently,reductions in minimum capital requirementswere almost always the policy response to in-creased demand for banking facilities.”

Figure 5 plots the number of U.S. commercialbanks at the end of each year since 1900, andthe number of branch offices since 1920. Thenumber of banks increased dramatically after1900, when the Gold Standard Act halved theminimum capital requirement from $50,000 to$25,000 for chartering national banks in townssmaller than 3,000 persons. Many states respond-ed by reducing minimum capital requirementsfor state-chartered banks to as low as $5,000.The lowering of this barrier and generally strongeconomic growth encouraged the entry of manynew banks. The strict limits on branching im-posed by the federal government and by moststates meant that new banks, rather than newbranches of existing banks, largely met increasesin the demand for banking services.

“Branch banking networks operated in several Southernstates before the Civil War, and the First and Second Banksof the United States operated branches in several cities.Branching was not widespread, however, and in manystates banks did not operate branches even where theywere not prohibited.

‘“Evanoff (1988) found that in 1980 the number of bank of-fices per square mile was higher in both urban and ruralareas of states permitting branching than in unit bankingstates. He found, however, more bank offices in states per-mitting limited branching than in states permitting state-wide branching.

‘“See Hanweck (1992) for a discussion of policy issues per-taining to interstate banking.

“For example, the Comptroller of the Currency, John W. Pole,cited the widespread failures of “one-crop” and “fair-weather” banks, i.e., small, undiversified farming banks, asevidence of the need for branch banking. See histestimony in U.S. House of Representatives (1930).

‘“See also White (1982).

10

Figure 5Number of Commercial Banks and Branches

60,000

50,000

40,000

30,000

20,000

10,000

0

The rapid increase in the number of bankscontinued through 1920. The increase was espe-cially large between 1914 and 1920 in agricul-tural states, where wartime demand for farmproducts, dispersed populations and sti-ict prohi-bitions on branching encouraged a plethora ofsmall unit banks. The number of banks and ag-gregate bank assets also grew rapidly in theeight states that adopted deposit insurance sys-tems. Because insurance premiums were lowand unrelated to the probability of failure,deposit insurance encouraged greater investmentin commercial banks than would have otherwiseoccurred.

In 1921, the number of commercial banksreached 30,456, an all-time peak. By then, theshock that would bring about widespread bankfailures and reduce the number of banks hadalready occurred. The wartime increase in com-modity prices reversed in mid-1920, and the AllCommodities Price Index declined 36.8 percentbetween 1920 and 1921.20 A sharp increase in

‘“Oklahoma initiated deposit insurance in 1908 and was fol-lowed by Kansas, Nebraska, South Dakota and Texas in1909, Mississippi in 1914, and North Dakota and Washing-ton in 1917. See Calomiris (1992b) and Wheelock (1993) forevidence that deposit insurance strongly influenced thegrowth of bank assets and the number of banks per capitabefore 1920,

loan defaults and bank failures in agriculturalstates followed. Between 1921 and 1929, 5,712

banks suspended operations, including 976 banksin the peak failure year of 1926.” As in the1980s, bank failures in the 1920s were regionallyconcentrated. Rural areas, especially in the Mid-west and the South, suffered high failure rateswhile failure rates were low in urban areas, theNortheast and the West Coast.

While the number of commercial banks de-clined, the number of branch bank offices roseduring the 1920s, from 1,281 in 1920 to 3,353in 1929. Before 1920, many state banking lawswere silent on the issue of branch banking,though administrative or judicial interpretation,or simply custom, prevented branching. By 1924,however, many state legislatures had consideredthe issue, with Arizona, California, Delaware,Georgia, Maryland, North Carolina, Rhode lsland,South Carolina, Tennessee and Virginia permit-ting state-wide branching, and nine others per-mitting limited branching. Other states either

2cU,S, Department of Commerce (1960), series E13.

“‘Board of Governors of the Federal Reserve System (1943).

11

prohibited all branching or had no law eitherallowing oi’ prohibiting branching. Wherebranching was permitted, only state-charteredbanks that were not members of the FederalReserve System could establish branches. TheNational Bank Consolidation Act of 1918,

however, permitted national banks to keepbranches acquired through consolidation withstate-chartered banks. Later, the McFadden Actof 1927 permitted national banks to establishbranches within their home-office cities in statesthat granted the same privilege to state-charteredbanks.”

Despite increased branching, differences inbtanch banking laws explain little of the variationin bank failure rates across states during the1920s. Alston, Grove and Wheelock (1993) findthat a state’s failure rate was determined mainlyby the severity of agricultural distress it suffered,and that distress had a greater impact on failurerates in states having deposit insurance systems.Differences in the extent of branch banking andother suggested causes of failure account forcomparatively little of the variation in failurerates across states. The lack of an apparent re-lationship between branching and performancein the 1920s probably reflects the limited extentof branching in the Midwest and the South,where the worst of the agricultural collapsewas felt. Furthermore, state economies in theMidwest and the South were not sufficientlydiverse to provide banks with much protectionfrom a general collapse of commodity prices,even where statewide branching was permitted.Interstate branching, however, might have al-lowed sufficient diversification to limit failures.Canada, which also experienced sharp incomedeclines in commodity-producing regions, hadnationwide branch banking and just one bankfailure in the 1920s.’”

The Great Depression ushered in further wavesof bank failures, and though initially concentratedin farming areas, failures later spread through-

out the country. From 1930 to 1933, 9,096 bankssuspended operations. In 1933 alone, 4,000 banksclosed, including 2,122 that closed during theBank Holiday in March 1933 and never re-opened.’~Between December 1929 and December1933, the number of commercial banks in theUnited States declined 43 percent, from 24,970to 14,207.

Major banking legislation was enacted duringthe Depression, beginning with the Banking Actof 1933 (also known as the Glass-Steagall Act af-ter’ its principal authors). This legislation in-troduced federal deposit insurance, imposedlimits on the interest rates that banks were per-mitted to pay depositors and constrained bankactivities in numerous ways. Though SenatorCarter Glass believed that banks should be res-tricted to making short-term commercial loans,he was an advocate of branch banking. On thisissue he was in the minority, and except for’ aprovision in the Banking Act of 1933 providingnational banks with all branching authoritygranted to state-chartered banks, branching waslargely ignored in banking legislation during the1930s. The large money center’ banks that tend-ed to favor liberal branching laws and authorityto perform a broad array of securities services,and which opposed deposit insurance, werewidely viewed as the villains that caused theGreat Depression. Little wonder that New Dealbanking legislation enacted deposit insuranceand curbs on bank securities activities, withoutexpanding the opportunities for branching.’”

Banking Since World War H

The number of banks changed relatively littlein the 50 years following passage of the BankingAct of 1933. Since 1984, however, when thenumber of commercial banks in the United Statesreached a post-World War II peak of 15,126,the number of banks has fallen 25 percent to11,406 firms, and today there are fewer com-mercial banks in the United States than at anyother time in the 20th century.’6

““See White (1983) for further detail about branch banking inthe United States during the 1920s.

23For comparison of the U.S. and Canadian experiences, seeWhite (1983) or Bordo, Rockhoff and Redish (1993).

“4Federal Reserve Bulletin, December 1937, p. 1208.25Burns (1974) examines the politics involved in New Deal

banking legislation.26The shaded insert on p. 12 presents an accounting of

changes in the number of banks since 1984, highlighting thecontribution of Texas and discussing other changing aspectsof market structure.

12

Recent changes in Bank Market StructureI tic’ ac’i-ornparl\ jug table shows the frintri— liisinltatiori and declining pc’tIi)lI!Uui ~llttiS

t.iiitioxis iliitilC l)~,1)(’,\ c’iitrants,rlostirc’s arid evenltiallv brought a nIlap~.eof liii iotiliiit’rryir’rgers to annual changes in the ntiniher ol cial re il c’slatc’ market thrnugbcitil the ~‘oi.rtliriisui’ed roinriieicial banks in the L iined Suites nest increasing nail defaults arid (tiflItlIelilalironi 15181) to 1 992. I he net change iii I lit hank lLtiltlIc’,. I e\it~ tilotic’ had one third cr1iitjiiiht’r nt banks c’qual’— [Iii’ ntriiib.’i of nns al] 1 .S. bank tzultir’es cltir-irig the I 9805, ifirdbanks Itt~sthose \~hich ceased operations tjvc’i’ half of all I .5. lailure-, iii 1988 when

Ofl\ c’rted to hiaritlics ol other hiuks or run— ITS ‘I e~asbtiriks tailed.’solidatni. I3c’c’atr%tr an’, ol tic’ categoric’,, couldneilc’c t tlit’ l(’sOltitiOn cii a hantI~Iaikn’e, I also throughout the t niln’d States, main oflist the total number of l,rih_imes in each pan 2 banks that ha~e failed or men absorbed Int)c’pile rising failures. the number ol nisured citlici coriimnerc:ial bank~,ha~c lic’en ‘itinall. tic’—

ornmncremah banks ~ricreased through 1984 ic’— tnecmni 198-I and 1992 tIre miczrnhei ol hank,,cause thi’ nunubei of net’,- hank,, c~sc’tic’rJc’d illi less hart than S51) niillion ol a’4sets tintc;se dosed. converted to hranc’lic’s ot ol her 1984 twin’s) tfeclmcrrl by 2.52.3. I rum 9.2 I? tobanks nm corlsolLdatccl After 1981. hon et or. 6.692 . atcoririting for 86 pc’r’cc’rit of the totalhe miririiimr’ nil new hank (‘halters declined dec-line in the miummihei cii riisurc’d rcimrnem’cial

while failures continued to rise. Although banks.’ ‘the shame of’ total I’S. bank asset,,I ailures ha~e Ut-upped off iere.nth-, large held by small banks has also declined. froiriritiritbers of banks-- both failed and solvent— 8.6 per’re.nt in 1981 to 6.3 p t-ce.nl in 1992.continue to lie coin erted to hranelic’s of n-hue tltrir share of U.S. hank eqciitv hasother kinks. (‘onsequentlv. the number’ ol fallen From 12.6 pei’t’ent to 8.1 percentcuinniercial lntnl~shas contiiuetl to declineby several hundr ci banks per \‘etm’ since thin .\c’roinpan\ ing the increase in bm’anchingnitd—198fts. has been a grcn~iig atiiltation of banks with

l’ex as ac’eotlllts for ninr’h of the rec:c’nt fluc’— holding companic’s Although batik hofciuigtuation in the number’ of banks in the t ‘ititc’d c’onrtpelliies ha~e existed since I lie I 920s, theStates Between 1980 and I 9851. 24 pi’rcc’.clt oh titririhen cit banks aililialc’d wit Ii holding coni—al] new I’ S batik charters nc-re issued in panics has increased shar’pl’, in recent ears.I exits, and the stale itc’c’citinted for o’, ci’ half In 1970, 1.1)69 c-oinnic,rt’ial banks 7.9 per’c’c’ntof’ a]l new rhar’ter’s clurirrg 1983 wul 1984 of all bariksi n c-re meriihers of multi—hankthe energ~ boom of the t970s hecaiiir’ the holding c’oliipanies, n lirn’c’as in 199! 1 501real c’state boom ol the 1 980s. and I exas eli- hunks (311.8 pmr’e.riU belonged to multi—batikjot i’d sc’ven’aI ~-ean’sof strong ec’onoriiir; gi on th. holding uortipanic’~ holding c’cirilparn’ banksl.xisting banks grew r’apicfk and many rien are typically larger than oilier batik’,, andbank, opernc’d. Iiec’,rirsr’ the ,,tate str’irtl’,- jointed during 1970—92 thin share of total ronnmc’i-c’ialbranehinti pn-ior’ LI) 15)87 aitncist all nc’,’, hank batik assets liekl I kinks ii liliateci n itlioltic’e,, ‘s-er c’ either’ independent titiit banks cii’ riicrlti—hiaiik holding COiil~iIlli(’5 iric’r’ctast’cl I ‘ourmembers of bank holding collipanrits. 3(1.2 pen’cc’nit ci 73.7 pc’i’c’ent

The sources ot these data are tho Board ot Governors “Because data on the number of banks by asset size areof the Federa! Reserve System 11991a, 1991b 1992. Tn- not publ’sh&~these figures were computed directlyble 72) ano the Boa”d of Govornors of the Fed~ral from the quarterly call reports for insured commorcalReserve System (1993. Table 151 banks and limiteo to the 50 states. There were 14.348

- - - . . banks in ths category ‘n 1984 and 11406 in 1992 The•Failu’e data s’nce 985 reflect oniy bangs closea be- asset cutofi point of $50 milton in 1984 was adjusted:~~::~°:~:~re for inflation by the GDP (gross domestic product) defIa-Federal Depostt Insurance Corporation (1991. Table A) nor Ii 1992 trio cutoff punt was 866.275 citron

and Federal DeposI lnsu’ance Oorporanion (1992~ iBank hoiding company data are an unpublished series

~O’Keefe0990. p ~ ~ Boana of Governors o4 Inc Federal Reserve.:Hcrvnz f19921 ~xam’nesthe causes of lexas bank andtb’ift ~a’lures

rtnc.OA: DCC~D~’ D~~I’t’~CC” ‘IC’

13

Economists usually attribute the stability ofthe banking industry from the mid-1930s to1980 to the system of deposit insurance andregulation imposed on banks, and the generallysteady growth of the U.S. economy.~’Depositinsurance and controls on entry to the industrymade bank charters valuable, and deposit in-terest rate controls stabilized the largest compo-nent of bank costs. But, over time, risinginflation and higher interest rates, coupled withgreater competition among banks and from non-bank intermediaries, increasingly threatened thestability of the banking system.

Rapid improvements in computer and commu-nications technology began to seriously erodetraditional bank niches in the 1960s. Nonbanksources of intermediation, such as the commercialpaper market, grew rapidly in this decade, whiledeposit interest rate ceilings hampered the abilityof banks to compete for funds against the ex-panding money market. The introduction ofmoney market mutual funds and transaction ac-counts at thrifts in the 1970s, and ever-increasing market interest rates, led Congress toenact the Depository Institutions Deregulationand Monetary Control Act of 1980, which dere-gulated deposit interest rates. Though it helpedbanks compete for deposits, the legislation cameat a time when market rates were exceptionally

As competition from nonbank intermediariesgrew, the commercial banking industry itself be-came increasingly competitive. In the 1960s,regulators began issuing new bank chartersmore freely, and many states reduced barriersto branching and interstate holding companies.As their charter values declined, banks had anincentive to increase their assets-per-dollar-of-equity and to make increasingly risky loans.28Increased competition benefited consumers ofbanking services, but left the banking systemmore vulnerable to exogenous economic shocks.

Declining agricultural income and a collapseof agricultural and energy prices in the early1980s resulted in the insolvency of many banksduring the subsequent decade. Many farmerswho had borrowed to buy land in the 1970swere unable to repay their loans when incomesfell, resulting in the failure of many agriculturallenders. Similarly, the economic boom in Texas,Louisiana, Oklahoma and other energy producingstates that accompanied rising energy prices inthe 1970s gave way to falling incomes in the1980s. Banks lent heavily to energy producersand later to real estate developers in these states,and profited from the fortunes of their borrow-

27Friedman and Schwantz (1963) attribute much of the declinein failures and absence of banking panics after 1933 todeposit insurance. More recently, scholars such as Keeley(1990) and Flood (1993) note the importance of regulationsthat encouraged banks to act conservatively and offset theincentive for excessive risk-taking created by deposit in-surance.

28When deposit insurance premiums are unrelated to risk,banks have an incentive to take greater risks than theyotherwise would. A decline in charter value magnifies thisincentive. See Keeley (1990).

high. So, banks experienced sharp increases inthe cost of funds and many suffered substantiallosses.

14

ers. When incomes realized from real estate de-velopment failed to meet expectations, borrow-ers defaulted on their loans and many banksbecame insolvent. A similar phenomenon oc-curred in New England in 1990—92, when a col-lapse of local real estate markets led to thefailure of numerous commercial banks.

While the number of commercial banks re-mained fairly constant until the 1980s, the num-ber of branch bank offices has increaseddramatically over the entire post-World War IIera. The initial surge in branches accompaniedthe great population migration from central citiesto suburbs following the war. And, since 1970,the number of branch offices has more thandoubled, from 21,424 to 53,744 in 1992, whilethe percentage of banks with multiple officeshas increased from 29 to 57 percent. Since 1984,the total number of bank offices (banks plusbranches) has increased, despite a 25 percentdecline in the number of banks, because thenumber of branch offices has risen from 41,740to 53,744.

In the 1920s, branch banking was too limited toprevent widespread bank failures in areas suf-fering sharp income declines. The high numberof failures in agricultural and energy-producingstates since 1980 suggests that, despite the largeincrease in the number of branch banks in recentyears, branching may still be too limited to pro-tect the banking system from regional or sectoralshocks. As in the 1920s, many of the states ex-periencing the greatest economic distress in the1980s also had the most restrictive branchinglaws. Even states that permit statewide branch-ing, such as Arizona and Connecticut, have hadrelatively high failure rates since 1980, suggest-ing again that statewide branching alone maynot result in enough diversification to preventhigh numbers of hank failures.

Figures 6—8 provide a rough indication of theimpact of state branching lan’s during the 1980s.Figure 6 shows states classified according towhether they permitted statewide branching,permitted limited branching or prohibited allbranching, as of December 31, 1979.29 Figure 7shows that states tended to have more banks

per 1,000 inhabitants if they limited branching,or prohibited it altogether, than if they permit-ted statewide branching. No such clear correla-tion between branching restrictions and bankfailure rates is apparent, however. Figure 8plots the average annual bank failure rate dur-ing 1980—92 for each state. Among the 48 con-tinental states, Texas had the highest averagefailure rate, a relatively high number of banksper capita in 1980, and in 1980 prohibited allbranching. Other states with high failure rates,however, such as Arizona, Oregon, Utah, NewHampshire, Connecticut and Massachusetts, per-mitted at least some branching in 1980 and hadlow numbers of banks per capita. Moreover, theaverage failure rates in some unit banking states,such as Montana, North Dakota and Illinois,were low.’°

The lack of an obvious relationship betweenstate branching laws and bank failure rates isnot surprising for two reasons. First, failureswill not be high unless borrowers are unable torepay their loans. In ‘Fexas, for example, real es-tate developers and energy producers sufferedsharp income declines and defaulted on loans.There were fewer loan losses and consequentlylower bank failure rates in other states, includ-ing most unit banking states. Second, statewidebranching will not prevent failures if the timingand extent of economic distress is similar through-out a state. Texas banks were able to achieve ameasure of geographic diversification throughmulti-bank holding companies, and undoubtedlystatewide branching would have enabled evenmore diversification. The collapse of energy andreal estate prices affected the entire state,however, and, thus, many banks would likelyhave failed even if statewide branching hadbeen permitted. New England banks suffered asimilar fate in 1990—92 when the diversificationpermitted by statewide branching and regionalinterstate holding companies again was insuffi-cient to prevent numerous failures followingcollapse of the region’s real estate market.fl

Banks have avenues for geographic diversifica-tion even where bn’anching is prohibited, andmay not take advantage of greater diversification

291he source of this information is the Conference of StateBank Supervisors (1980). South Dakota is listed in thissource as having unlimited statewide branching. After July1, 1969, however, branch banks could be established onlythrough merger.

301he coefficient of correlation between the number of banksper capita a state had on December 31, 1979, and itsaverage annual bank failure rate during 1980—92 is —0.17.

“Randall (1993) investigates the causes of recent NewEngland bank failures.

15

Figure 6State Branching Laws on December 31, 1979

L.

0.000-0.050

0.051-0.1000.101-0.200

0.201-0.300

Unit

Figure 7Banks Per 1,000 Persons on December 31, 1979

N

SEPTEMBER/OCTOBER 1993

16

Figure 8Average Commercia’ Bank Failure Rate, 1980-1992

opportunities if they are available, but most ob-servers accept the argument that the freedomto branch makes diversification easier. Furtherliberalization of branching laws might lead tofurther declines in the number of banks and tolarger average bank size, however, and todaythe debate about branch banking centers largelyon the issues of service, competition and the ef-ficiency of large banks.

Larger average bank size might reduce industrycosts because of economies of scale, i.e., thataverage unit cost declines as total firm outputrises. A widely cited study by Boyd and Graham(1991) suggests, however, that large banks maynot be more efficient than small banks. Theyfind that between 1976 and 1987, banks withassets in the range of $25—$100 million had ahigher average rate of return on assets than

either smaller or larger banks. Between 1988and 1990, those in the $100-million-to-$1-billionclass had the highest average rate of return.Moderate-size banks also had the highest aver-age returns on equity. Not only have largerbanks been less profitable, but Boyd and Grahamalso find that they tend to have lower equi-ty/asset ratios, implying that large banks aregenerally less well protected against insolvency.Because their evidence comes from a periodwhen interstate branch banking was almost en-tirely prohibited, however, the usefulness oftheir evidence for indicating whether or not in-terstate branching networks would be efficientmay be limited.32 Their apparent success in theSouth before the Civil War, and the prevalenceof large branching organizations in other coun-tries, suggest that such organizations can beprofitable.”

‘2Furthermore, large banks may be less profitable becausethey tend to operate in more competitive markets thansmall banks. Like Boyd and Graham (1991), however,most studies, such as those surveyed by Humphrey (1990),find that scale economies are exhausted for banks of rela-tively modest size and that the largest banks in the UnitedStates are less efficient than smaller banks. These studiestend to focus on the entire banking firm, though, and Toevs(1992) finds evidence of substantial scale economies for in-dividual banking functions.

“Moreover, Calomiris and Schweikart (1988) find that inter-state branching organizations in the South withstood finan-cial distress better than unit banks in Northern states.Similarly, Canada, whose banking system is comprised ofa few large banks with nationwide branches, did not ex-perience the extent of financial disruption present in theUnited States during the Great Depression. See Haubrich(1990) and Kryzanowski and Roberts (1993).

0.0000-00025

0.0026-0M050

~ cL0051-0.015O

0.0151-0M450

FEDERAL RESERVE BANK OF St LOUIS

1~7

Increased branching within states explains someof the recent decline in the number of banks inthe United States. Interstate branching wouldlikely cause further consolidation. If branchingleads to greater diversification of bank assets,or if large branching networks are able to deliverbanking services more efficiently than is pre-sently possible, then expanded branching powersmight strengthen the industry and help it copewith increased competition. Moreover, becausethe fixed costs of operating a branch office arelow, consumers would likely benefit from in-creased competition if banks were given greaterauthority to branch.’4 The experience withbranch banking in the United States is perhapsstill too limited to sho~vclearly how much thesize and structure of the American banking sys-tem would change if interstate branching be-comes a reality. The rapid growth of branchbanks in recent years, however, indicates thatbanks have responded to the easing of branch-ing restrictions and that further easing wouldlikely lead to further extensions of branchingnetworks.

FINANCIAL SERVICES

The consequences of permitting commercialbanks to sell insurance, underwrite corporatesecurities or offer a variety of other financialservices are perhaps even less certain than theimplications of interstate branching. The Bank-ing Act of 1933 forced the separation of com-mercial and investment banking, and thoughfinancial innovation and regulatory interpreta-tion has, over time expanded the securities ac-tivities of banks, strict limits on manysecurities-related services remain.” For example,banks are largely prohibited from underwriting,distributing and owning securities issued by pri-vate corporations. Researchers have conjecturedhow authority to offer a variety of financialservices might alter banking markets in theUnited States by simulating mergers of commer-cial banks and other financial service firms, andby studying banks engaged in such activities in

other countries.’8 This section reviews a thirdresearch area which examines the performanceof U.S. commercial banks with securities opera-tions prior to the Banking Act of 1933.

In the United States, commercial banks becameinvolved in the securities business on a largescale during World War I, when many banksinvested in and sold war bonds. The success ofwar bond drives and the growing interest ofthe public in securities ownership led manybanks to offer an increasing variety of invest-ment services. Commercial banks also ex-perienced a decline in commercial loan demandmuch like that of recent years, as more andmore firms found that issuing securities in themoney and capital markets was less expensivethan borrowing from banks.” Furthermore, thepromise of high returns drew many bank depo-sitors to securities markets, and, thus, banks ex-perienced increased competition for funds thatis again very reminiscent of recent experience.in the 1920s, banks responded to increased com-petition by offering a variety of securities serv-ices, such as underwriting, distribution andbrokerage services. Following waves of bankfailures in the early 1930s, Congress imposedstrict limits on the securities activities of banks,however, because commercial bank involvementwith securities was widely thought to have beenan important cause of the banking crisis.

Senator Carter Glass and other proponents ofthe Banking Act of 1933 viewed bank ownershipof securities and bank lending to investors onsecurity collateral as harmful to banks, deposi-tors and the economy. Security underwritingand ownership had increased bank risk, theycharged, and made it more likely that a collapseof security prices, such as the stock marketcrash in October 1929, would cause widespreadbank failures. Glass (1933) argued that by makingloans on security collateral, underwriting newissues and trading in securities, banks had fueledsecurity speculation, drawn funds away from“legitimate” uses and contributed to instabilityof both securities markets and the bankingsystem:

‘4Calem (1993) argues that interstate branching would in-crease competition and improve access to banking servicesfor consumers.

‘5Kaufman and Mote (1990) describe the expansion of banksecurities activities since 1933 and argue that liberal inter-pretation by regulators and the courts have permitted banksto offer a variety of securities services. Benston (1990) alsodetails the securities activities that banks are permitted toperform under federal law.

‘6See Boyd and Graham (1988) and Benston (1990).‘7See Currie (1931).

SEPTEMBER/OCTOBER 1993

18

There seems to be no doubt that a large factorin the overdevelopment of security loans—andin the dangerous use of the resources of bankdepositors for the making of speculative profits,with the risk of hazardous losses—has been theperversion of national and state banking laws.The greatest danger is seen in the growth of‘bank affiliates’ which devote themselves inmany cases to perilous underwriting operations,stock speculation, and maintaining a market forthe hank’s own stock...

Supporters of the Banking Act of 1933 also al-leged widespread abuses of fiduciary responsi-bility by banks that underwrote and distributedsecurities. Such allegations included misrepresen-tations to unsophisticated customers about therisk of securities, the sale of low-grade securitiesto bank trust accounts, the disposal of non-performing loans via securitization and manipu-lation of a bank’s own stock.

Today, the most vigorous opponents of ex-panded powers for banks are securities dealers,insurance agents and others with a financial in-terest in limiting competition, though econo-mists have not reached a consensus on theappropriate activities of banks, particularly theiruse of government-insured deposits.” This sec-tion reviews recent research on commercialbank securities activities before 1933. Thisresearch generally finds little support for thearguments used to justify the separation of com-mercial and investment banking in the 1930s,which is significant because many of these argu-ments are still used to justify their continuedseparation.

Early studies of bank involvement in the secu-rities business generally accepted the view thatsuch activities increased bank risk and weresubject to abuse.” This view remained largelyunchallenged for many years and was the basisof regulatory and court decisions further definingbank securities-related activities. Recent re-examinations of bank securities operations in theperiod before 1933, however, have found littleevidence of increased risk to the banking systemor abuse.

White (1986) investigates whether security op-erations increased the probability of commercialbank failure during the Great Depression. Inthese years, most bank securities operations wereconducted by in-house bond departments orseparate security affiliates. Affiliates wereseparately incorporated entities but typicallyhad the same shareholders as the parent bank.During the Great Depression, the failure rate ofthe 145 national banks operating bond depart-ments in 1929 was 7.6 percent, and the failurerate of the 62 national banks operating securi-ties affiliates was 6.5 percent. By contrast, 26.3percent of all national banks failed between1930 and 1933, and, hence, national banks withsecurities operations had a substantially lowerfailure rate than other national banks. Whitesuggests that economies of scale and greater op-portunities for diversification might explain thedifference in failure rates since banks withsecurities operations were typically larger thanother banks.~°

Proponents of separating commercial and in-vestment banking frequently claimed that varia-bility in the earnings of securities affiliatescaused excessive fluctuations in the earnings oftheir parent banks. Nevertheless, White foundno significant correlation between the rates ofreturn of securities affiliates and those of theirparent banks. His evidence casts doubt on thevalidity of the argument that securities activitiesincrease the risk of bank failure.

Benston (1990) addresses a second justificationfor the separation of commercial and invest-ment banking—that the commingling of invest-ment and commercial banking activities producessignificant conflicts of interest. The original pro-ponents of separating commercial and investmentbanking charged that commercial banks hadtaken advantage of small, unsophisticated inves-tors who trusted banks to give them prudentadvice about securities. Today, similar argumentsare made in favor of limiting bank involvementwith securities and mutual funds. Proponents ofmaintaining the status quo note a possible conflictbetween an investment bank’s desire to promotesecurities it underwrites and a commercial bank’s

“Some economists favor so-called narrow” banking pro-posals by which banks are restricted in their use of insureddeposits to a limited number of very safe assets. See Spong(1993) for a discussion of narrow banking. Other economistsprefer the opposite extreme of universal banking, by whichbanks are permitted to hold a variety of assets and offersecurities, insurance, real estate and other financial serv-ices. See Benston (1990) for arguments in favor of universalbanking.

“The most comprehensive of such studies is Peach (1941).40White also estimates a probit regression model in which

he finds that, after controlling for various financial measuresreflecting management, banks with securities affiliates hada significantly lower chance of failure than other banks, whilebanks with an active bond department had neither a greaternor lesser chance of failure.

FEDERAL RESERVE BANK OF St LOUIS

19

duty to provide disinterested investment advice.Other arguments suggest the temptations a com-mercial bank might have to underwrite securitiesso a borrower can repay unprofitable loans, orto place unsold securities in the bank’s trust ac-counts. Benston, however, contends that whencommercial banks were allowed to deal in secu-rities, conflicts of interest did not lead towidespread abuses. Moreover, he argues thatmany of the conflicts cited by critics of com-mercial banks were or could be eliminatedthrough regulation. For example, when commer-cial banks were permitted to sell securities, theirsecurities affiliates were prohibited from sellingto the trust accounts of the parent bank.

Benston points out that potential conflicts ofinterest exist with nearly all financial transac-tions, and the separation of commercial and in-vestment banking does not guarantee disinter-ested advice from either commercial or invest-ment banks. For example, a commercial bankmight attempt to persuade a customer to bor-row from the bank rather than issue securitiesthrough an underwriter, or steer customersseeking investment advice to a particular securi-ties dealer if the dealer maintains a large depositwith the bank. Similarly, securities firms mightattempt to persuade investors to buy securitiesthey underwrite rather than those underwrittenby another firm, or induce small savers to buysecurities instead of bank certificates of deposit.

Benston contends that conflicts of interest arelikely to be less prevalent for commercial bankswith securities operations than for specializedsecurities firms. For example, a potential hot’-rower might get better advice from a bank thana securities firm about whether to take a loanor issue securities if the bank could offer both.Similarly, there would seem to be less potentialfor biased advice to an investor from a bankthat could provide a variety of investments, frominsured deposit accounts to mutual funds to in-dividual securities.

Kroszner and Rajan (1993) examine one elementof the conflict-of-interest argument. They com-pare the performance of securities underwrittenby commercial banks with those underwrittenby investment banks during the 1920s as ameasure of whether commercial banks were

able to fool the public into buying low-qualitysecurities. Among a matched sample of bondsunderwritten by commercial banks or theirsecurities affiliates and bonds underwritten byinvestment banks, Kroszner and Rajan find thatsecurities underwritten by commercial bankshad the lower default rate. The default rate oflow-grade securities underwritten by commer-cial banks was especially low relative to that ofinvestment banks. Commercial banks also tendedto underwrite higher-quality securities than in-vestment banks. Kroszner and Rajan suggestthat securities markets may have forced com-mercial banks to issue more high-grade, safesecurities because of possible conflicts of in-terest for commercial banks involved in under-writing. If true, their evidence indicates that theinvesting public disciplined commercial bankswith respect to the securities they underwrote.So, even if potential conflicts of interest be-tween commercial and investment banking exist,Kroszner and Rajan conclude that repeal of le-gal constraints on commercial bank underwrit-ing would not likely harm the public.

Proponents of allowing banks to offer securitiesservices point to potential benefits for the econ-omy. Banks offering securities services mightbenefit from economies of scope. Economies ofscope exist if the total cost of providing a varietyof services within a single firm is lower than ifthe services are provided by different firms. Animportant function of banks is to gather andevaluate information about potential borrowersand provide access to the payments system. Com-inercial banks might be able to underwrite secu-rities, sell insurance, offer real estate servicesand provide other financial services at lowercost to consumers than currently available be-cause of economies in the processing and use offinancial information.~’

Calomiris (1993) argues that, historically,branching restrictions and regulations limitingthe scope of banking activities significantly in-creased the cost of capital for U.S. firms. Be-cause of their comparative advantage asevaluators of information about potential bor-rowers, commercial banks have tended to bethe principal lenders to new firms. As firms ma-ture and their creditworthiness becomes widelyknown, they tend to borrow relatively less from

4tMester (1987) provides an introductory discussion of econo-mies of scope in banking. Pulley and Humphrey (1993) findlittle evidence of economies of scope in the joint pro-

duction of loans and deposits by U.S. banks during 1978—92,but do not address the possibility of significant economies inproviding loan and securities services.

20

commercial banks and issue more securitiesthrough investment banks. Banks that provideboth loans and underwrite security issues couldeconomize on the gathering of information aboutfirms and accelerate the process by which a firmmoves from relatively high-cost borrowing in theform of bank loans to lower-cost placement ofdebt and equity. If commercial banks were per-mitted to provide both loans and underwritesecurities, they could provide a firm’s financingneeds over its entire life cycle. Banks could thenspread the cost of acquiring and evaluating in-formation about a new firm over many yearsand reduce the cost of credit for firms in theirearly years when investment needs are highand cash flow is low. Restrictions on branching,on the other hand, “implied a mismatch betweenthe scale and scope of firms and those of theirbankers.’~’Such restrictions impeded banksfrom achieving the size necessary to finance thebort-owing needs of large firms. They alsoprecluded a potentially very efficient means offinancing industrial development—the placementof corporate securities underwritten by com-mercial banks through their branch offices.Moreover, branching restrictions probably in-creased the cost of transactions because firmsoperating in multiple markets were forced torely on different banks in each market for pay-ments services.

The United States remains a long way fromhaving a system of universal banking, and anysignificant change in regulation, such as repealof restrictions on the securities activities of com-mercial banks, requires careful study of numer-ous issues. Recent research suggests that manyof the arguments for separating commercial andinvestment banking, which once seemed com-pelling, are not supported by historical evidence.The historical record is silent on some issues,however, such as the question of what securitiesactivities commercial banks should be permittedto fund with insured deposits. Leaving asidethese issues, removal of legal restrictions onbank securities activities would likely strengthenthe industry by allowing it to compete more ef-fectively with other~intermediaries and banks inother countries. Broadening securities powerswould probably favor larger banks, however,and might also encourage further consolidationof the industry.

CONCLUSION

Although the number of commercial banks hasdeclined sharply since 1984, by other measures,such as the number of bank offices and totalbank assets relative to gross national product(GNP), the banking industry has not been shrink-ing. Some economists contend that the bankingindustry has excess capacity. Even more arguethat without significant changes in regulation,the banking industry is destined to wither. Therelaxation of restrictions on interstate branchingand the securities activities that commercialbanks are permitted to perform are two of themost frequently proposed regulatory changesthat proponents view as necessary to preservethe health of the banking industry.

History provides useful evidence about the ef-ficacy of branch banking and the securities ac-tivities of commercial banks. The United Statesexperienced a high number of bank failures inthe 1980s because its banks were not sufficientlydiversified. As in the 1920s and other decadeswhen failures were high, a lack of geographicdiversification left many banks vulnerable toeconomic downturns. The strongest argumentin favor of interstate branch banking is that itenhances the ability of banks to diversify acrossregions so that they can offset losses in someregions with profits in others. Questions aboutthe technical efficiency of large branching net-works and the effects of interstate branchingon competition and access to banking servicesremain controversial.

Perhaps even more controversial is the debateabout the separation of commercial and invest-ment banking. Securities activities continue tobe widely viewed as too risky for commercialbanks, though, ironically, many commentatorswho hold this view also complain that banks arenot making enough loans. Evidence from theera before legal separation of commercial andinvestment banking indicates that commercialbanks with securities operations were betterdiversified and less likely to fail than otherbanks. Moreover, it appears that commercialbanks underwrote higher-quality securities thaninvestment banks. Although the historicalrecord cannot address all issues of relevance to-day, the evidence generally supports the viewthat banks would benefit from an increased var-iety of financial services.

42Calomiris (1993).

21

REFERENCES

Aguilar, Linda. “Still Toe-To-Toe: Banks and Nonbanks at theEnd of the ‘80s?’ Federal Reserve Bank of Chicago Eco-nomic Perspectives (January/February 1990), pp. 12—23.

Alston, Lee J., Wayne A. Grove and David C. Wheelock.“Why Do Banks Fail? Evidence from the 1920s,” Explora-tions in Economic History (forthcoming, 1993).

Barth, James H., and Dan R. Brumbaugh, Jr. “The ChangingWorld of Banking: Setting the Regulatory Agenda,” JeromeLevy Economics Institute of Bard College, Public PolicyBrief No. 8 (1993).

_______ and Robert E. Litan. The Future of American Bank-ing (M. E. Sharpe, Inc., 1992).

Benston, George J. The Separation of Commercial and Invest-ment Banking: The Glass-Steagall Act Revisited and Recon-sidered (Oxford University Press, 1990).

Board of Governors of the Federal Reserve System. AnnualReport, 1992 (1993).

_______- Annual Statistical Digest, 1991 (1992)._______- Annual Statistical Digest, 1980—89 (1991a)._______- Annual Statistical Digest, 1990 (1991b)._______- Banking and Monetary Statistics, 1914—1941 (1943).Bordo, Michael D., Hugh Rockhoff and Angela Redish. “A

Comparison of the United States and Canadian BankingSystems in the Twentieth Century: Stability vs. Efficiency?”prepared for the annual meeting of the Economic HistoryAssociation (1993).

Boyd, John H., and MarkGertler. “U.S. Commercial Banking:Trends, Cycles, and Policy?’ NBER working paper No. 4404(July 1993).

________ and Stanley L. Graham. “Investigating the BankingConsolidation Trend,” Federal Reserve Bank of MinneapolisQuarterly Review (spring 1991), pp. 3—15.

_______ and _______- “The Profitability and Risk Effects ofAllowing Bank Holding Companies to Merge with OtherFinancial Firms: A Simulation Study?’ in The FinancialServices Industry in the Year 2000: Risk and Efficiency,proceedings of a conference on Bank Structure and Com-petition, Federal Reserve Bank of Chicago (1988), pp.476— 514.

Bureau of National Affairs. “Comptroller Promises Push forPowers, Says Banks Need More Lines of Business?’ Bank-ing Report (September 20, 1993), pp. 417—18.

Burns, Helen M. The American Banking Community and NewDeal Banking Reforms (Greenwood Press, 1974).

Calem, Paul S. “The Proconsumer Argument for InterstateBranching?’ Federal Reserve Bank of Philadelphia BusinessReview (May/June 1993), pp. 15—29.

Calomiris, Charles W. “Corporate-Finance Benefits FromUniversal Banking: Germany and the United States,1870—1914?’ NBER Working Paper No. 4408 (July 1993).

“Regulation, Industrial Structure, and Instability inU.S. Banking: An Historical Perspective?’ in Michael Klaus-ner and Lawrence J. White, eds., Structural Change inBanking (New York University Salmon Center, 1992a).

_______ “Do ‘Vulnerable’ Economies Need Deposit Insur-ance? Lessons from U.S. Agricultural in the 19205?’ inPhilip L. Brock, ad., If Texas Wore Chile: A Primer on Bank-ing Reform (Institute for Contemporary Studies, 1992b), pp.237—314.

Calomiris, Charles W., and Larry Schweikart. “Was the SouthBackward? North-South Differences in Antebellum Bank-ing, working paper, University of Pennsylvania (1988).

Clair, Robert T., and Gerald P. O’Driscoll Jr. “Learning fromOne Another: The U.S. and European Banking Experience,”Research Paper Na 9108, Federal Reserve Bank of Dallas,(May 1991).

Conference of State Bank Supervisors. A Profile of State-Chartered Banking (1980).

Currie, Lauchlin. “The Decline of the Commercial Loan?’Quarterly Journal of Economics (August 1931), pp. 698—709.

Duca, John V. “Regulation, Bank Competitiveness, and Epi-sodes of Missing Money?’ Federal Reserve Bank of DallasEconomic Review (second quarter, 1993), pp. 1—23.

Edwards, Franklin R. “Financial Markets in Transition Or theDecline of Commercial Banking?’ prepared for a symposiumon Changing Capital Markets: Implications for MonetaryPolicy, Federal Reserve Bank of Kansas City (1993).

Evanoff, Douglas D. “Branch Banking and Service Accessibil-ity,” Journal of Money, Credit and Banking (May 1988), pp.191—202.

Federal Deposit Insurance Corporation. Statistics on Banking(1992).

_______ Annual Report (1991).Federal Reserve Board. Bulletin (December 1937).

Flood, Mark. “The Great Deposit Insurance Debate?’ thisReview (July/August 1992), pp. 51—77.

Friedman, Milton, and Anna J. Schwartz. A Monetary Historyof the United States, 1867—1960 (Princeton University Press,1963).

Furash, Edward E. “Banking’s Critical Crossroads?’ TheBankers Magazine (March/April 1993), pp. 20—6.

Glass, Carter. “Carter Glass Urges the Need of Banking Re-form?’ Review of Reviews and World’s Work (January 1933),pp. 20—22.

Greenspan, Alan. “FDICIA and the Future of Banking Lawand Regulation?’ remarks prepared for the 29th AnnualConference on Bank Structure and Regulation, FederalReserve Bank of Chicago (May 6, 1993).

Gorton, Gary, and Richard Rosen. “Corporate Control, Portfo-lio Choice, and the Decline of Banking?’ Federal ReserveBoard Finance and Economic Discussion Series, Divisionof Monetary Affairs (December 1992).

Hahn, Thomas K. “Commercial Paper?’ Federal Reserve Bankof Richmond Economic Quarterly (spring 1993), pp. 45—67.

Hanweck, Gerald A. “Interstate Banking, Bank Expansionand Valuation?’ in R. Alton Gilbert, ed., The Changing Mar-ket in Financial Services (Kluwer Academic Publishers,1992), pp. 41—92.

Haubrich, Joseph G. “Nonmonetary Effects of FinancialCrises: Lessons from the Great Depression in Canada?’Journal of Monetary Economics (March 1990), pp. 223—52.

Horvitz, Paul M. “The Causes of Texas Bank and ThriftFailures?’ in Philip L. Brock, ed., If Texas Were Chile: APrimer on Banking Reform (Institute for ContemporaryStudies, 1992), pp. 131—60.

Humphrey, David B. “Why Do Estimates of Bank Scale Econ-omies Differ?” Federal Reserve Bank of Richmond Eco-nomic Review (September/October 1990), pp. 38—50.

Kaufman, George G. “The Current State of Banking Reform?’prepared for Conferences on Financial Prosperity in the 21stCentury, Jerome Levy Economics Institute of Bard College(1993).

22

_______ “The Diminishing Role of Commercial Banking inthe U.S. Economy,” Issues in Financial Regulation, workingpaper series, Federal Reserve Bank of Chicago (May 1991).

_______ and Larry R. Mote. “Glass-Steagall: Repeal byRegulatory and Judicial Reinterpretation,”The Banking Law Journal (SeptemberlOctober 1990),pp. 38 8—421 -

Keeley, Michael C. “Deposit Insurance, Risk, and MarketPower in Banking?’ American Economic Review (December1990), pp. 1183—1200.

Kroszner, Randall S., and Raghuram G. Rajan. “Is the Glass-Steagall Act Justified? A Study of the U.S. Experience withUniversal Banking before 1933?’ working paper, GraduateSchool of Business, University of Chicago (April 1993).

Kryzanowski, Lawrence, and Gordon S. Roberts. “CanadianBanking Solvency, 1922—1940?’ Journal of Money Creditand Banking (August 1993). pp. 361—76.

Mester, Lorreta J. “Efficient Production of Financial Services:Scale and Scope Economies,” Federal Reserve Bank ofPhiladelphia Business Review (January/February 1987),pp. 15—25.

Napoli, Janet A. “Derivative Markets and Competitiveness?’Economic Perspectives, Federal Reserve Bank of Chicago(JulylAugust 1992), pp. 13—24.

O’Keefe, John. ‘The Texas Banking Crisis: Causes and Con-sequences 1980—89?’ FDIC Banking Review (winter 1990),pp. 1—34.

Peach, W. Nelson. The Security Affiliates of National Banks(The Johns Hopkins Press, 1941).

Pulley, Lawrence B., and David B. Humphrey. “The Role ofFixed Costs and Cost Complementarities in DeterminingScope Economies and the Cost of Narrow Banking Pro-posals?’ Journal of Business (July 1993), pp. 437—82.

Randall, Richard E. “Lessons from New England BankFailures?’ New England Economic Review (May/June 1993),pp. 13—38.

Remolona, Eli M., and Kurt C. Wulfekuhler. “Finance Compa-nies, Bank Competition, and Niche Markets,” FederalReserve Bank of New York Quarterly Review (summer1992), pp. 25—47.

Spong, Kenneth. “Narrow Banks: An Alternative Approach toBanking Reform?’ working paper, Division of Bank Supervi-sion and Structure, Federal Reserve Bank of Kansas City(January 1993).

Toevs, Alden L. “Under What Circumstances Do Bank MergersImprove Efficiency?” in Credit Markets in Transition, pro-ceedings of the 28th Annual Conference on Bank Structureand Competition, Federal Reserve Bank of Chicago (1992),pp. 602—28.

U.S. Department of Commerce. Historical Statistics of theUnited States, Colonial Times to 1957 (Washington D.C.,1960).

U.S. House of Representatives, Committee on Banking andCurrency. Branch, Group and Chain Banking (Washington,D.C., Volume 1, parts 1—3, 1930).

Wall, Larry D. “Too-Big-to-Fail After FDICIA,” Federal ReserveBank of Atlanta Economic Review (January/February 1993),pp. 1—14.

Wheelock, David C. “Government Policy and Banking MarketStructure in the 1920s?’ The Journal of Economic History(December 1993), pp. 857—79.

White, Eugene N. “Before the Glass-Steagall Act: An Analysisof the Investment Banking Activities of National Banks?’ Ex-plorations in Economic History (January 1986), pp. 33—55.

_______- The Regulation and Reform of the American BankingSystem, 1900—29 (Princeton University Press, 1983).

- “The Political Economy of Banking Regulation,1864—1933,” Journal of Economic History (March 1982), pp.33—40.