Vol. 35, No. 1

CONOMIC REVIEW

1999 Quarter 1

The Effects of Vertical Integration 2on Competing Input Suppliersby R. Preston McAfee

Banking Consolidation and 9Correspondent Bankingby William P. Osterberg and James B. Thomson

FEDERAL RESERVE BANKOF CLEVELAND

Vol. 35, No. 1

1999 Quarter 1

The Effects of Vertical Integration on Competing Input Suppliersby R. Preston McAfee

Banking Consolidation and Correspondent Bankingby William P. Osterberg and James B. Thomson

22

99

http://clevelandfed.org/research/review/Economic Review 1999 Q1

The Effects of Vertical Integration on Competing Input Suppliersby R. Preston McAfee

When a downstream firm buys an input supplier, it can reduce its costs ofusing that input. Other input suppliers typically respond by pricing moreaggressively, given the demand reduction, which tends to lower input sup-ply costs to other firms. Thus, a vertical merger may lower rivals’ costsinstead of raising them.

Banking Consolidation and Correspondent Bankingby William P. Osterberg and James B. Thomson

Banking consolidation, spurred on by interstate branching deregulation, ischanging the competitive structure of banking markets. Policymakers andregulators have focused on the implications of the ongoing consolidationfor customers of banks in retail and wholesale markets. Little attention,however, has been paid to the impact of interstate consolidation on corre-spondent banking markets—those markets where banks buy and sellinputs used to produce banking services. By studying the era of intrastatebranching deregulation, the authors provide some insights on the implica-tions of interstate branching for correspondent banking.

Economic Review is publishedquarterly by the Research Depart-ment of the Federal Reserve Bankof Cleveland. To receive copies or to be placed on the mailing list,e-mail your request tomaryanne.kostal@clev.frb.org or fax it to 216-579-3050.

Economic Review is alsoavailable electronically throughthe Cleveland Fed’s site on theWorld Wide Web:http://www.clev.frb.org/research.

Editors: Michele LachmanLisa McKennaDeborah Zorska

Design: Michael GalkaTypography: Liz Hanna

Opinions stated in Economic Review are those of the authorsand not necessarily those of theFederal Reserve Bank of Cleve-land or of the Board of Governorsof the Federal Reserve System.

Material may be reprinted if thesource is credited. Please sendcopies of reprinted material to the editors.

ISSN 0013-0281

1

E C O N O M I C R E V I E W

1999 Quarter 1Vol. 35, No. 1

2

9

2

The Effects of VerticalIntegration on CompetingInput Suppliersby R. Preston McAfee R. Preston McAfee is a professor

of economics at the University of Texas at Austin. He thanksJoseph G. Haubrich and James B. Thomson for their beneficialcomments.

Introduction

Vertical integration is a booming phenomenonin many U.S. industries. Massive consolidationof the defense industry has left only three orfour developers producing many of the compo-nents used in military platforms.1 Banking alsois consolidating at a rapid pace, with integra-tion of related financial services (insurance,credit cards) and input services (check clearing,payments, electronic funds transfer) into parentcompanies. Telecommunications firms’ mergerscombine cable, wireless, local wireline, andlong-distance services. Simultaneously, firms inother industries are concentrating on their corecompetencies and selling off related lines ofbusiness. Automobile manufacturers, for exam-ple, are becoming more reliant on independentor semi-independent parts suppliers.

What are the effects of vertical integration?The large body of literature on this subjectmight reasonably be described as disjointed. Ithas focused mostly on providing a rationale foropposing vertical mergers on antitrust grounds.When a firm buys an upstream input supplierthat also supplies its downstream competitor,the vertically integrated firm can raise the priceof the input to its competitor, thereby obtaining

an advantage in the downstream market. This isthe standard “raising-rivals’-cost” argument pio-neered by Salop and Scheffman (1987).2 Inextreme cases, the vertically integrated firmmight refuse to sell to competitors and, if theinput supplier’s product was necessary for pro-duction, might be able to foreclose its competi-tors from the downstream market.

This paper examines an opposing effect ofthe raising-rivals’-cost theory. In particular, theanalysis focuses on other input suppliers’ reac-tion to vertical integration. Its main insight isthat vertical integration, by providing easieraccess to one input, reduces demand for theother inputs and tends to lower their prices.This, in turn, encourages the vertically in-tegrated firm to sell its input at a lower price aswell, which may reduce the costs of all inputs.Thus, accounting for the reaction of substituteinput suppliers may reverse the traditional con-clusions of the raising-rivals’-cost theory.

■ 11 An item like an aircraft or a submarine is a platform, which holdsa variety of weapons systems, detection systems like radar or sonar, andother systems like landing gear, engines, and so on.

■ 22 Ordover, Saloner, and Salop (1990) is the best-known treatment.Salinger (1988) and Hart and Tirole (1990) also provide related, and moregeneral, analyses.

3

I. Structure ofthe Model

The model’s general form is set out in figure 1.Two upstream suppliers, X and Y, sell to twodownstream firms, which in turn sell to thefinal consumers. I will focus on the effects ofthe vertical integration that occurs when firm 1purchases firm X. Suppose that the products ofthe upstream firms are imperfect substitutesand that, having purchased firm X, firm 1 willcontinue to use some of the inputs supplied byY. Then vertical integration will affect firmY’spricing decision.3

In this paper, I set aside the incentive toraise rivals’ costs by assuming that firms 1 and 2do not compete in the output market. I focusinstead on vertical integration’s effects on thealternative input supplier and find that verticalintegration tends to lower the prices of bothinputs to firm 2. The intuition is straightfor-ward: The purchase of firm X by firm 1 lowersthe price of input x to firm 1, reducing thedemand for y by firm 1. In response to thisreduction in demand, firm Y lowers the price ofy to both firms. The response of the integratedfirm is to lower the price of x to firm 2.

The situation is less clear when firms 1 and 2compete imperfectly in the downstream market.There are two direct effects of the merger: First,the price of x to the combined entity falls, tend-ing to reduce the price of y. Second, the price ofx to firm 2 rises, tending to increase the price of y. Either effect can dominate, and the price ofy may rise or fall, depending on the extent ofsubstitution between the outputs of firms 1 and2 and the substitutability of the two inputs.

Let xi, yi denote the demand for the twoinputs by firm 1. I assume constant returns toscale. The timing is that the input sellers simul-taneously set input prices px, py, respectively.Then firms 1 and 2 choose their input quanti-ties and output prices.

II. Raising Rivals’ Costs

The standard raising-rivals’-costs theory is bestexplained by eliminating firm Y. In this case, X is a monopoly supplier of the input. If x isnecessary for production, the merged firm hasthe ability to foreclose firm 2 from production.Even if x is valuable but not strictly necessaryfor production, the merged firm can raise thecost of x to firm 2, thereby increasing firm 2’soverall costs.

Even in this simple scenario, the price thatfirm 2 is charged for x can fall. Suppose 1 and 2barely compete in the final output market.Moreover, suppose firm 1 has significantly moreinelastic demand for x, so that the monopolyprice for firm 1 exceeds the monopoly price forfirm 2. Prior to vertical integration, the price of xwill lie between the two monopoly prices. Afterthe merger, the price of x will fall to approxi-mately the monopoly price for firm 2. Moreover,insofar as firms 1 and 2 do compete, the mon-opoly price for firm 2 will fall, since firm 1’slower marginal cost will make it a more aggres-sive competitor after the merger.

The standard analysis focuses on the casewhere X and Y are Cournot competitors withconstant marginal costs. In this case, if themerged firm uses its own inputs and withholdsoutput from firm 2, firm 2 is facing a monopolyand will generally experience higher inputprices. This is true even when firm Y is actuallyseveral firms that are in Cournot competition,although the more firms there are in the inputsupply market, the smaller is the effect.

The results concerning Cournot input supplygeneralize to increasing marginal costs. Withincreasing marginal costs, firm 1 may wisheither to sell to or buy from firm Y, even after

F I G U R E 1

Competition Layout

■ 33 It might seem that services such as check clearing are homoge-neous. However, distinct banks have an advantage in being able to cleartheir own checks quickly, and large banks may have a greater netting out ofchecks. In addition, distinct suppliers of check clearing may have distinctregional advantages.

Firm X Firm Y

Firm 1 Firm 2

Final Consumers

Upstream:

Downstream:

π 11 π 12 dpx2

+0

dpx1 =

0

ψ 221 ψ 1

22 +ψ 222 dpy ψ 1

21 0

This gives

dpx2

=1 π 12ψ 1

21 dpx1.

dpy ∆ –π 11ψ 121

Stability implies that

(6) ∆ = π 11(ψ 122 + ψ 2

22) – π 12ψ 221 > 0.

The terms π12 and ψ21 are similar in that theyrepresent the effect of a competitor’s price in-crease on the marginal profitability of a priceincrease for the firm. If the input pricing gameis one of strategic complements, then theseterms are positive. Alternatively (and equiva-lently), if an increase in the price of one inputmakes the demand for the other input less elas-tic, these cross-partials will be positive.

It can be shown that if the downstream pro-duction functions have constant elasticity ofsubstitution with constant returns to scale, anddemand is constant up to a choke price (whichis tantamount to assuming that the downstreamquantity is exogenous), then the cross-partialsare positive. This special case will be exploredin the following numerical simulation.

When these input profit cross-partials arepositive, then

dpx2

> 0,dpy > 0.

dpx1 dpx

1

Thus, the merger, which lowers the price ofx to firm 1 by eliminating firm X ’s marginaliza-tion, lowers both of the input prices to firm 2 aswell as the price of y to firm 1.

This result is intuitive. The reduction in theprice of x to firm 1 makes that firm’s demandfor y more elastic, since it now has a less ex-pensive substitute. This causes Y to lower theprice of y.4 The lower price of y induces a re-action from the combined firm—it lowers theprice of x.

4

the merger with firm X. However, consider thesymmetric case, in which X looks like Y and 1looks like 2. After the merger, firm 1 does notneed to buy from firm Y, and thus can increasethe costs to firm 2 by refusing to sell.

This literature has played an important roleby showing that vertical mergers could poten-tially foreclose competition downstream. How-ever, the literature has focused primarily onmergers’ bad effects on rivals, without examin-ing their potential for good effects.

III. Lowering Rivals’ Costs

I start the analysis using the demand for inputsas primitives. To facilitate the analysis, I distin-guish between the prices firm X charges tofirms 1 and 2. With independent downstreamdemands, the effect of the merger of X with 1 isto change the input price of x to firm 1. Firm Xearns profits on its sale to firm 2 of

(1) π = (px2 – cx)x2(px

2, py ).

I divide firm Y’s profits into the componentsearned on firm 1 and firm 2:

(2) ψ 1 = (py – cy )y1(px1, py ),

ψ 2 = (py – cy )y2(px2, py ).

Using numerical subscripts to denote partialderivatives, profit maximization yields

(3) π1 = ψ12 + ψ2

2 = 0.

The direct effect of the merger of X and 1 onthe input supply prices is to lower px

1 from itsmonopoly level to marginal cost cx. Because of the assumed independence of demands forthe outputs of 1 and 2, the merged entity willchoose the price of px

2 to maximize π , and firmY will maximize the sum of ψ1 and ψ 2.

I am assuming that firm Y cannot price dis-criminate. If both input suppliers are able to doso, nothing changes in the prices charged tofirm 2. If firm Y can price discriminate but firmX cannot, then the merger permits firm X toprice discriminate, since the only relevant priceis that charged to 2. As a consequence, px

2 willincrease if firm 2’s demand for x is less elasticthan firm 1’s. This will have effects on the priceof y, usually of the same direction.

Differentiating the first-order conditions, one obtains

3 4 21 21 21 .(4)

1 2 1 2(5)

■ 44 This is where the assumption that firm Y cannot price discrimi-nate is critical. If Y could price discriminate, the reduction in the price of xto firm 1 would reduce the price of y to firm 1 but not to firm 2.

(7)

With VerticalIntegration

When firms 1 and X merge, firm 1 can pur-chase x at price c. In this case, the combinedentity will price x to maximize6

(13) π = (px – c)x2

= (px – c) .

As before, this gives the first-order condition

(14) 0 = (1 – α)py1 – α – αpx

1 – α + cpx1 – α .

Firm Y faces a more complicated problembecause it will generally sell to both firms 1 and2, and these two firms generally face distinct in-put prices for x. Firm Y chooses py to maximize

(15) ψ = (py – c)(y1 + y2) = (py – c)

+

While closed forms for the first-order condi-tions exist (and are sufficient to characterizeequilibrium in the range posited), it is not pos-sible to solve the first-order conditions for theequilibrium prices explicitly because the pricesenter these equations in complex ways. Conse-quently, I have used Mathematica 3.0 to findthe roots of the first-order conditions and toplot the outcome as a function of α.

To simplify the calculations, note that c canbe set to unity without loss of generality (pricesmeasured in cost units). Moreover, for scalingpurposes, it is useful to plot the markup reduc-tions associated with vertical integration rather

5

IV. NumericalExample

A numerical example illustrates and quantifiesthe effects described in the theory. Supposethat the two downstream firms can sell one uniteach at a price high enough so that each firmwill always buy inputs sufficient to produceone unit. The downstream firms have a con-stant elasticity-of-substitution production tech-nology with constant returns to scale and para-meter αε [½,1]5:

(8) q = (xα + yα)1α.

Let the marginal production costs of theupstream firms be c.

Without Vertical Integration

If there is no vertical integration, the down-stream firms minimize pxx + pyy s.t. q = 1.This gives

(9) x = ,

y = .

Firm X chooses px to maximize

(10) π = (px – c)x.

Routine calculations yield

(11) 0 = (1 – α)py1 – α – αpx

1 – α + cpx1 – α .

For α > ½ and px > c, this equation charac-terizes a maximum. A symmetric solution to thefirst-order conditions yields

(12) px = py = 2α – 1.

As α →1, the goods become perfect substi-tutes, and prices fall to marginal costs.

■ 55 For α $ 1, only one input is chosen. For α , ½, demand isinelastic and the input pricing equations solve with infinite prices.

■ 66 Because the vertically integrated firm is assumed not to competewith firm 2, downstream profits can be ignored. However, if there is a lowlevel of competition, then downstream profits must be included here, dra-matically complicating the analysis.

py1 – αpx

1 – αα α

py1 – α

α

1 2α1

py1 – αpx

1 – αα α

px1 – α

α

1 2α1

α α 2α –1

c

py1 – αpx

1 – αα α

px1 – α

α

1 2α1

py1 – αpx

1 – αα α

py1 – α

α

1 2α1

α α 2α –1

py1 – αc1 – α

α α

c1 – αα

1 2α1

+

+

+

.

+

+

6

than the actual prices. Thus, the prices relativeto the nonintegrated prices are plotted; in par-ticular, for z = x, y,

(16) Pz = = .

When Pz = 1, there is no cost reduction,while Pz = 0 would be competitive or marginalcost pricing. When the outcome is plotted (seefigure 2), several observations emerge. First, thereductions in input prices are significant, on the

order of 5 percent or 10 percent. Second, firm Yreduces its prices more than firm X does. Thisshould be a reasonably general property, sincefirm X is responding to firm Y’s price reduction.In symmetric models like the one examined, theprice of the unintegrated input should fall morethan the price of the integrated input (to the restof the world).

In asymmetric models, there is an additionaleffect. As an independent firm, X priced toserve both firms 1 and 2; therefore, firm X’sprice is an average of the two monopoly pricesassociated with 1 and 2. After the merger, firm Xwill price only for firm 2; this could increase ordecrease the postmerger price. The effect identi-fied in this article, however, should continue tohold, using as a benchmark the monopoly pricefor firm 2 rather than the monopoly price forboth downstream firms.

Third, the markups are not monotonic in α.This is interesting because the prices are mono-tonic, with prices diverging as α→½, and pricesgoing to costs as α→1. Simulations suggest thatthe markup on y is below the markup withoutvertical integration (as a proportion of the van-ishing markup without vertical integration),even in the limit.

While prices are more indicative of theasymmetric effects on the individual input sup-pliers, firm 2 cares primarily about its marginalcost. In figure 3, the marginal cost is plotted rel-ative to the marginal cost in the absence of ver-tical integration.

Firm 2’s marginal cost is lowered as much as8½ percent. This amount is, of course, less thanthe reduction enjoyed by firm 1, but substantialnevertheless.

V. BankingApplications

The Federal Reserve System recently studiedthe implications of its potential exit from theprovision of retail payments services such ascheck clearing and electronic payments. In thecase of many community banks, especiallythose in rural markets, the remaining providersof these services would be vertically integratedcompetitors. Moreover, as Osterberg andThomson (in this issue) show, branching dereg-ulation is leading to consolidation of bothupstream and downstream banking markets.The extant industrial organization literature sug-gests that increased vertical integration in bank-ing, coupled with consolidation in the inter-bank market, may have deleterious effects ondownstream banking competition.

F I G U R E 2

Input Markups underVertical Integrationa

a. As a proportion of the markup without vertical integration, plotted against α.SOURCE: Author’s calculations.

F I G U R E 3

Firm 2’s Marginal Cost under Vertical Integrationa

a. As a proportion of its marginal cost without vertical integration, plottedagainst α.SOURCE: Author’s calculations.

pz – c

– cc2α – 1

(2α – 1)(pz – 1)

2(1 – α)

7

However, the literature itself is a poor guideto the likelihood of anticompetitive effects. First,much of the research arose from a desire to un-derstand how vertical mergers might matter forantitrust enforcement. In particular, the ClaytonAct, which gave courts the ability to block a va-riety of vertical practices, preceded clear under-standing of any circumstances in which verticalintegration might be harmful to competition.Consequently, a literature developed to showthat vertical integration might have a negativeeffect, instead of assessing the likelihood thatvertical integration is harmful to competition.

Second, much of the literature focuses onthe Cournot model, primarily for tractabilityreasons. While Cournot competition might be areasonable characterization of downstreamcompetition for customers, where firms’ capaci-ties are relatively inflexible (at least comparedto prices), Cournot competition seems a poormodel of the provision of many upstream ser-vices like check clearing, where capacity con-straints are unlikely to bind. Results from mod-els employing Cournot competition upstreammay not be applicable to integration in thebanking industry.

Third, the key ingredient of the raising-rivals’-costs story is that the primary motivation for ver-tical integration is a wish to damage competi-tors. One plausible future for the bankingindustry features a handful of very large inter-state banks, along with a great number of rela-tively small local banks. The large institutionswill offer banking, mortgage, insurance, finance,and other services, will mainly operate electron-ically, and will be vertically integrated into mostor all financial services areas. In contrast, thelocal banks will be predominantly rural and willoffer personalized service, creating a marketniche by exploiting the superior informationand goodwill that local interaction provides.These different styles of banks will probably notcompete with each other in the minds of mostcustomers. The large banks will competestrongly with other large banks, at least untiltheir numbers are whittled down to three orfour in any given region. The rural banks willonly face the threat that their best (largest) cus-tomers may be induced to use large, inexpen-sive banks; for most customers, a given ruralbank will compete with other rural banks.

In this scenario (which does not result fromany study on my part), rural banks will notcompete significantly with large banks. As a re-sult, they are not likely to be the target of anti-competitive vertical integration, nor are theylikely to be harmed by a reduction in the num-ber of large, vertically integrated banks. The

largest banks may try to harm one anotherthrough their pricing of banking service inputs,but these institutions are in the best position tofend for themselves.

The lowering-rivals’-costs story is inapplica-ble to the present analysis as long as the FederalReserve continues to provide check clearingand other services at some reasonable approxi-mation of cost. As a consequence, rural bankscould benefit from vertical integration of largebanks only if the Federal Reserve were ineffi-cient, so that a mechanism for price reductionsexisted. Prices cannot be dropped below mini-mum cost.

VI. Conclusion

The standard analysis of vertical integration’seffect on competitors emphasizes the verticallyintegrated firm’s incentive to foreclose down-stream rivals or raise their costs. While thiseffect is natural in some applications, there isan offsetting effect on suppliers of substituteinputs. If firms 1 and 2 compete weaklyenough in the output market, the effect onother input suppliers may dominate the fore-closure effect, causing vertical integration tobenefit downstream rivals while harmingupstream competitors. The harm to upstreamcompetitors, however, lies in reducing theirmarkups over marginal cost, that is, damagingtheir monopoly power.

A significant aspect of the effect of verticalintegration is that both the unmerged inputsupplier and the vertically integrated firm lowertheir input prices. The mechanism is that themerger eliminates the markup on the input bythe purchased firm. The other input supplierreduces its prices in response to this lower-priced substitute from the merged firm. Thevertically integrated firm lowers its input pricesto the rest of the world in response to the low-ered price of the other input supplier.

Simulations with constant-elasticity-of-substitution production functions indicatepotential reductions of as much as 8½ percentin the downstream competitor’s marginal costs.

8

References

Hart, Oliver, and Jean Tirole. “Vertical Inte-gration and Market Foreclosure,” BrookingsPapers on Economic Activity, Special Issue(1990), pp. 205–76.

Ordover, Janusz A., Garth Saloner, andSteven C. Salop. “Equilibrium Vertical Fore-closure,” American Economic Review,vol. 80, no. 1 (March 1990), pp. 127–42.

Salinger, Michael. “Vertical Mergers andMarket Foreclosure,” Quarterly Journal ofEconomics, vol. 103, no. 2 (May 1988), pp. 345–56.

Salop, Steven C., and David T. Scheffman.“Cost-Raising Strategies,” Journal of Indus-trial Economics, vol. 36, no. 1 (September1987), pp. 19–34.

9

Banking Consolidation andCorrespondent Bankingby William P. Osterberg and James B. Thomson William P. Osterberg is an

economist at the FederalReserve Bank of Cleveland, andJames B. Thomson is a vicepresident and economist at theBank. The authors thank SandySterk and Guhan Venkatu fortheir research support.

Introduction

Throughout most of the United States’ financialhistory, correspondent banking has been anunderpinning of our banking system. Banks usecorrespondent banking relationships to deliverservices to customers in markets where thebank has no physical presence. For example,international correspondent banking relation-ships are used by large banks seeking to pro-vide services to multinational corporations andin the finance of foreign trade. Due in part tothe historical limitations on geographic ex-pansion by banks, correspondent banking hasbeen an important channel for delivering ser-vices to domestic customers who may be oper-ating in markets beyond the bank’s geographicreach. Correspondent banking markets oftenallow banks to purchase intermediate goodsand services at a lower cost than producingthem in-house—hence these markets may havebeen critical to the success of community banks.

The ongoing consolidation of the U.S. bank-ing system and the increasing geographicscope of large banking institutions could haveimportant implications for the competitivestructure and, in turn, the efficiency of corre-spondent banking markets. Whether these

changes will lead to more or less competitionin correspondent banking is unclear. On onehand, consolidation will inevitably lead to areduction in the number of banks offering cor-respondent banking services, thereby increas-ing the market power of the remaining players.On the other hand, given that correspondentbanking markets’ services are regionally orlocally based, interstate consolidation mayincrease the number of providers in a localmarket—even though the total number of sup-pliers has been reduced nationally. Finally, theshrinking number and increased average sizeof banks may lead to a reduction in thedemand for correspondent banking services.

Banking industry consolidation could haveimportant implications for the Federal ReserveBanks in their traditional role as providers ofcorrespondent banking services. As bankingbecomes less fragmented and more nationallyintegrated, there is less need for the public cor-respondent banking network operated by theFederal Reserve Banks. However, if bankingconsolidation appears to have materially dimin-ished the competitiveness of private correspon-dent banking markets, then the continued roleof Federal Reserve Banks as public competitorsmay be warranted.

In this article we reexamine some of theseissues, focusing on the impact of regulatorychanges to permit intrastate branching andinterstate banking. Our concern is primarilywith the impact of such changes on concentra-tion in correspondent balances and domesticdeposit markets. We utilize the call report datacompiled by the Federal Financial InstitutionsExaminations Council (FFIEC) and the Sum-mary of Deposits data prepared by the FederalDeposit Insurance Corporation (FDIC). Anychanges in concentration could have importantimplications for the efficiency and competitive-ness of banking markets.

The paper is organized as follows: Section Iprovides an overview of correspondent bank-ing. The correspondent banking literature isreviewed in section II. Section III furnishes adescription of the data and the empirical strat-egy. The results are discussed in section IV.Finally, conclusions and recommendations arepresented in section V.

I. An Overview of CorrespondentBanking

All firms face the fundamental decision ofwhether to make or buy a particular input usedin production. For example, an automobilemanufacturer must decide whether to make itsown engines and transmissions or to buy themfrom an outside supplier. Computer manufac-turers must decide whether to make or buy theprocessors used in their machines. Likewise, abank must decide whether to sort and present

10

for collection checks drawn on other banks thathave been deposited in customer accounts, orto contract with a third party to perform thisfunction. For firms, the make-or-buy decisiondepends on a number of factors, including thenature of the input’s production function, thefirm’s demand for the good relative to the mar-ket, and the competitive structures of the mar-ket for the input good and the market for thefinal good.

Banking literature refers to correspondentbanking as the purchase (by banks) of inputfrom other banks, central banks, and bank clear-inghouses. For instance, when a bank in Cleve-land sends checks to its local Federal ReserveBank for collection, it has purchased correspon-dent banking services from that Reserve Bank.Another example is a recent agreement betweenJ. P. Morgan and Chase Manhattan Bank, inwhich Chase provides European currency clear-ing services for Morgan.1 The main services pro-vided by correspondent banks are discussed insection II and in the appendix.

Correspondent banking relationships arerelevant to the cost structure of the U.S. deposi-tory institutions sector. As table 1 shows, thelegacy of our unit-banking system—a conse-quence of intrastate and interstate branchingrestrictions—is a highly fragmented bankingsystem with a large number of small, locally andregionally based institutions. Economies of scalein the production of inputs associated with theprovision of many types of bank services, espe-cially payments services, exceed the range ofoutput for most community banks.2 Further-more, community banks lack the geographicscope needed to capitalize on network external-ities. Hence, in the absence of correspondentbanking markets, community banks wouldlikely be less efficient providers of financial ser-vices and would have more difficulty survivingin increasingly competitive banking markets.

In a typical correspondent banking relation-ship, the supplier of services is another bank.The provider of services in the input market canbe viewed as a vertically integrated firm thatmay compete with the community bank in theoutput market. For the integrated firm, thebenefits of supplying correspondent bankingservices are clear. First, there are economies ofscope between production of input for its ownproducts and production of correspondentbanking products. Second, providing correspon-dent banking services may allow the integrated

■ 1 See Steven Marjanovic, “Morgan Taps Rival Chase for EuropeClearances,” The American Banker, August 11, 1998, p. 1.

■ 2 See Bauer and Ferrier (1996).

T A B L E 1

FDIC-Insured Commercial Banksa

Bank size Number of banks Total assetsb

Less than $25 million 1,370 22,496

$25 to $50 million 2,026 75,077

$50 to $100 million 2,251 161,502

$100 to $300 million 2,259 371,720

$300 to $500 million 400 152,924

$500 to $1 billion 304 209,387

$1 to $3 billion 206 332,204

$3 to $10 billion 104 596,790

$10 billion or more 64 3,260,657

Total institutions 8,984 5,182,759

a. As of June 30, 1988.b. In millions of dollarsSOURCE: Federal Deposit Insurance Corporation, Statistics on Banking(http://www.fdic.gov/databank/sob/).

11

bank to more fully exploit economies of scale ornetwork effects. This, in turn, lowers the cost ofproducing (or increases the demand for) its owndownstream products.

The competitive structure of the correspon-dent banking market may also be relevant tothe structure of the markets for bank products.Industrial organization theory tells us that if the integrated firm has substantial monopolypower in the upstream (input) market, then itmay use that power to damage its rival in thedownstream market.3 Thus, if the supplier ofcorrespondent banking services competes withits customer (the community bank) in the out-put market, then it may price its services abovethe average cost of production—thereby dam-aging the community bank’s ability to compete.The integrated bank’s ability to do this dependson the competitive structure of the upstreamand downstream markets.

II. Literature Review

A correspondent banking relationship involvesa correspondent bank, which provides the ser-vices, and a respondent bank receiving them.The respondent usually pays for the services bymaintaining correspondent balances at thelarger bank. The mix of services that might beprovided is broad, but appears to emphasizecheck processing, especially for smaller banks,and loan participation. Other services includeproviding reports on economic conditions,making securities recommendations, and safe-keeping securities. Correspondents also havemade markets for federal funds, in effect reduc-ing the minimum size of such transactions.4

Correspondent services are not paid fordirectly with fees, but rather implicitly throughmaintaining deposit balances.5 Some criticshave claimed that greater efficiency wouldresult from direct payment and have impliedthat smaller banks might not always haveknown the true cost of the services.6 Banks,however, historically have opposed the intro-duction of direct fees.

The early literature on correspondent bank-ing appears to have grown in response to twodevelopments: One was the decline in FederalReserve membership, which led to the estab-lishment of regional Federal Reserve check pro-cessing centers in the early 1970s and the pas-sage of the Depository Institutions Deregulationand Monetary Control Act of 1980. The otherwas the prospect of increased merger activity inbanking, rationalized by a purported positiveimpact on banking efficiency. Early research

■ 3 See McAfee (this issue) for a discussion of the damage-your-rivalargument and a counterexample.

■ 4 Knight (1970a) reported that about 90 percent of the banks sur-veyed indicated that their correspondents offered to help with internationalbanking services, collections, bank wire, and advice on consumer credit,credit information, and electronic data processing.

■ 5 Implicitly, as was pointed out by Flannery (1983), the correspon-dent should provide services costing [1 – ρ]r, where ρ is the reserverequirement and r is the market interest rate.

■ 6 Although fees are not charged directly for correspondent ser-vices, detailed account analyses were performed to estimate the revenueand expenses from correspondent accounts (see Knight [1970a]).

■ 7 See Knight (1970a, 1970b).

explored whether correspondent bankingallowed smaller banks to gain access, or gainaccess at a lower cost, to some of the servicesthat might be provided through mergers oracquisitions. This would imply, for example,that correspondent relationships provided alter-natives to mergers permitted by interstate bank-ing, and that the correspondent system mightbe affected by such developments. Anotherclosely related issue was economies of scale inthe production of important banking services.Certain services, such as processing of interna-tional financial transactions or specialized loanprograms, were more likely than others to beprovided at a lower cost by larger banks. Anissue related to the impact of mergers on effi-ciency was whether holding-company affilia-tion might allow smaller banks to gain accessto economies of scale in certain services pro-vided by the lead bank, which is presumablylarger. The argument had been made thatallowing small banks to join holding compa-nies would allow them to reduce the numberof correspondent accounts—and the totalamount of interbank balances held—but main-tain the same level of correspondent services.However, evidence as early as 1970 showedthat, contrary to this argument, the average sizeof interbank accounts appeared the same forsmaller banks in or out of holding companies.This suggested that holding-company affiliationdid not provide small banks with meaningfulopportunities to economize on holdings ofinterbank balances.7 Another alleged advan-tage of holding-company affiliation was theincreased ease of getting loan participations.While some evidence in favor of this wasfound, the related claim that banks facing fund-ing or capital constraints would have an easiertime placing loan participations with theirrespondents (that is, the banks purchasing cor-respondent services) could not be supported.

12

We are not aware of any studies that directlytackle the question of whether efficiency isenhanced by the presence of a correspondentbanking network. A few studies take an indirectlook at efficiency by analyzing overall econo-mies of scale for banks, taking account of theprovision of correspondent banking services—something that previous work on economies ofscale using functional cost analysis (FCA) datafailed to control for. Dunham (1981) found thatafter controlling for the provision of correspon-dent banking services, noncorrespondent bank-ing services exhibited economies of scale inproduction—something that previous studieswere unable to find. This could be the result oftwo factors: First, if small banks trade relativelymore correspondent balances for services (withsuch payments not reflected in the FCA data),they would appear more efficient. Second,larger banks producing more “due to” accounts,which are more service-intensive, would appearto have higher costs.

Flannery (1983) also re-estimated bank costfunctions after adjusting for the understatmentthe cost to the bank purchasing correspondentservices. He found that branch bank scaleeconomies were overestimated, but those ofunit banks were not affected. Gilbert (1983)found economies of scale in the provision ofcorrespondent services, which he claims weredue to an inverse relation between the amountof demand balances due to banks and theirshort-run variability.

Prior to the move toward interstate banking,it was already apparent that correspondentbanking might allow smaller banks to over-come the obstacles to geographic diversificationposed by interstate banking restrictions. Onepossibility is that correspondent banking wouldhelp effect a transfer of funds from surplus todeficit areas. Loan participations are a majorvehicle for this. Funds are transferred from thelarger correspondent bank to a smaller re-spondent bank to meet a loan request thateither exceeds funds available through localdeposits or exceeds legal lending limits. Pay-ment for such correspondent services is madein terms of the respondent’s balances at thecorrespondent bank. Knight (1970b) presentsevidence supporting the hypothesis that corre-spondent banking provides a channel forfunds to flow from surplus to deficit areas.This author documents a large net flow offunds to correspondents, even from respon-dent banks that originated loan participations.The appendix provides further detail on thisbranch of the literature.

The possibility that correspondent bankingmight affect the competitiveness of downstreammarkets was recognized early on. Consistentwith occasional concerns that correspondentsmight steal business from their respondents,Knight (1970a) indicates that money marketcorrespondents would participate only if theoriginating bank would reciprocate in participa-tions. Anecdotal evidence suggests that somecorrespondents had stated the view that bor-rowers consistently unable to obtain loans fromthe respondent should switch to the correspon-dent for lending.

Early work presaged a concern over theimpact of regulatory policies on the correspon-dent banking system. Obviously, any regulatorychange that influenced correspondent services,such as check-clearing arrangements, mighthave a direct effect on the use of private corre-spondents for such services. Knight (1972)focuses on the impact of the Federal ReserveSystem’s development of regional check pro-cessing centers (RCPCs) and the change in Reg-ulation J requiring all banks to pay for cash let-ters received from Federal Reserve Banks onthe day of receipt in immediately availablefunds. The creation of the RCPCs was seem-ingly intended to improve the efficiency of thecheck-clearing mechanism by permitting allparticipating banks to route items drawn onother participating banks to the clearing centeron the day of the deposit. The Regulation Jchange may have had the effect of transferringcollected funds from outlying (rural) banks(which had been granted a delay in payingafter receipt of a cash letter) to city banks. ForFederal Reserve member banks this regulatorychange effectively reduced the burden ofreserve requirements. However, during thistime in most states, nonmember banks couldcount correspondent balances toward reserverequirements set by their state banking regula-tory agency. Therefore, a differential impactmight have been felt by outlying nonmemberbanks since they would not have had theadvantage of reduced reserve requirements.

Other Federal Reserve System policies havehad key regulatory influences on the develop-ment of the correspondent banking system.Kane (1982) discusses Title I of the DepositoryInstitutions Deregulation and Monetary ControlAct of 1980 (DIDMC), which mandated that theFed make its correspondent services availableto all depository institutions and that they beexplicitly priced. Historically, the Fed hadoffered correspondent services to members freeof charge, in part to offset costs associated withthen-higher reserve requirements faced by

13

member banks. Rising interest rates increasedthe opportunity cost of holding reserves. Hence,rising interest rates during the 1970s would haveincreased the cost of Fed membership, and theability of banks to leave the System may haveincreased pressure on the Fed to offer addi-tional services. These included access to the dis-count window and the hope of receiving prefer-ential regulatory treatment.8 Other aspects ofFederal Reserve membership and relevant regu-lation are discussed in the appendix.

III. Framework forAnalysis and Data

Data limitations constrain our choice of an ana-lytical framework likely to have empiricalapplicability. Many of the studies cited aboveutilize data generated by one-time surveys ofspecific geographical regions. “Due to” and“due from” balances, corresponding to the lia-

bility and assets entries for the deposits of therespondent with the correspondent, respec-tively, are provided on the FFIEC’s call reportforms. However, these reports do not allow usto match up the two banks. In addition, thedue-to numbers are not available for smallbanks. Aside from information about organiza-tional structure, location, and mergers andacquisition history, few of the variables ana-lyzed in the studies cited previously are avail-able in regular reports. Our approach herefocuses on the correspondent balance numbersand the data on overall domestic deposits. Sup-plemental data on deposit markets is con-structed using the FDIC’s Summary of Depositsdata, which are available on an annual basis.

T A B L E 2

State Branching Status

■ 8 In Kane’s analysis, the fact that private correspondents offer awider array of services implies that the balance requirements set by themexceed the Fed’s reserve requirements on the same balances.

Year Year Effectiveswitched switched to date for

Current to limited statewide interstateState statusa branching branching banking

Alabama Statewide 1987 5/31/97Alaska Statewide 1/1/94Arizona Statewide 8/31/96Arkansas Limited 5/31/97California Statewide 10/2/95Colorado Limited 1991 6/0/97Connecticut Statewide 6/27/95Delaware Statewide 9/29/95District of

Columbia Statewide 6/13/96Florida Statewide 5/31/97Georgia Limited 6/1/97Hawaii Statewide 6/1/97Idaho Statewide 7/1/95Illinois Statewide 1988 1994 6/1/97Indiana Statewide 1991 3/15/96Iowa Limited 6/1/97Kansas Statewide 1988 1990 6/1/97Kentucky Limited 6/1/97Louisiana Statewide 1990 6/1/97Maine Statewide 1/1/97Maryland Statewide 9/29/95Massachusetts Statewide 8/2/96Michigan Statewide 1987 11/29/95Minnesota Limited 6/1/97Mississippi Statewide 1990 5/1/97Missouri Statewide 1987 1991 6/1/97Montana Limited 1990 3/21/97

Year Year Effectiveswitched switched to date for

Current to limited statewide interstateState status a branching branching banking

Nebraska Limited 5/31/97Nevada Statewide 9/28/95New

Hampshire Statewide 6/1/97New Jersey Statewide 4/17/96New Mexico Statewide 1991 6/1/96New York Statewide 2/6/96North

Carolina Statewide 6/22/95North

Dakota Limited 1991 5/31/97Ohio Statewide 1990 5/21/97Oklahoma Statewide 1993 5/31/97Oregon Statewide 2/27/95Pennsylvania Statewide 1990 7/6/95Rhode Island Statewide 6/20/95South Carolina Statewide 7/1/96South Dakota Statewide 7/1/96Tennessee Statewide 1990 6/1/97Texas Statewide 1990 8/28/95Utah Statewide 6/1/95Vermont Statewide 5/30/96Virginia Statewide 7/1/95Washington Statewide 6/6/96West Virginia Statewide 1988 5/31/97Wisconsin Statewide 1990 6/1/97Wyoming Limited 1991 5/31/97

a. As of June 30, 1996.SOURCE: Federal Deposit Insurance Corporation, Annual Reports (various).

14

The sample period studied here coversJune 30, 1984, to June 30, 1996, and includesmany shifts from unit banking to limited branch-ing, and from limited branching to statewidebranching. Table 2 details the history of suchregulatory changes during our sample period.

While numerous analyses have focused onthe issue of whether banking efficiency hasbeen enhanced by the recent wave of mergersand acquisitions (M&As), the role of correspon-dent banking has been unclear. Ideally, ouranalysis of changing concentration in corre-spondent banking markets would take this fac-tor into account. It is not possible, however, todirectly examine the effect of M&As on effi-ciency using the existing data.

One possibility is that M&As would renderunnecessary the pre-existing correspondentrelationships involving the formerly indepen-dent banks. On the other hand, they couldreduce the competitiveness and efficiency of theremaining system. Similar concerns arise withthe changes in branching status that we identify.Branching economies might be affected. Thepresumption is that the largest bank is the corre-spondent. After the absorption of unit banks asbranches of the new bank, the correspondentdeposits no longer appear on the reports at thebank level upon which we focus.

Finally, data on market prices charged forprivate correspondent banking services do notexist, especially at the individual market level.Therefore, we rely on measures of market con-centration, such as the Herfindahl index, toinvestigate the impact of branching deregula-tion on the structure of the interbank market. Itis important to note, however, that market con-centration measures are not always good prox-ies for the degree of competition in a market.Hence, increased concentration may not neces-sarily indicate a less competitive market—especially if the event driving market consol-idation increases the degree of potentialcompetition (contestability) in that market.

IV. Results

Tables 3 and 4 illustrate the ongoing consolida-tion of the domestic commercial banking indus-try and its implications for the correspondentbanking market nationwide. While the numberof banks has fallen steadily from 1984 to 1996,the market share of the top 50 correspondentbanks has risen from 28.18 percent to 36.57percent of all deposits due to banks over thesame time period. Similar results are foundwhen looking at the market share held by the

T A B L E 3

Banks and Branches

Year Banks Branches Offices

1966 13,529 16,842 30,3711967 13,506 17,884 31,3901968 13,479 18,966 32,4451969 13,464 20,149 33,6131970 13,502 21,597 35,0991971 13,602 23,080 36,6821972 13,721 24,566 38,2871973 13,964 26,403 40,3671974 14,218 28,384 42,6021975 14,372 29,929 44,3011976 14,397 31,068 45,4651977 14,397 32,836 47,2331978 14,378 34,524 48,9021979 14,351 36,521 50,8721980 14,421 38,458 52,8791981 14,401 40,500 54,9011982 14,435 39,485 53,9201983 14,454 40,548 55,0021984 14,483 41,485 55,9681985 14,402 42,970 57,3721986 14,193 44,054 58,2471987 13,705 45,017 58,7221988 13,119 46,036 59,1551989 12,697 47,650 60,3471990 12,329 50,017 62,3461991 11,909 51,591 63,5001992 11,449 51,544 62,9931993 10,944 52,467 63,4111994 10,431 54,656 65,0871995 9,921 56,028 65,9491996 9,511 57,258 66,7691997 9,125 59,773 68,898

SOURCE: Federal Deposit Insurance Corporation, Statistics on Banking(http://www.fdic.gov/databank/sob/).

T A B L E 4

National Correspondent Banking Deposit Shares

Market share (percent)

Yeara Top 50 banks Top 10 banks Top 5 banks

1984 28.18 11.92 7.511985 27.79 14.78 10.511986 26.89 11.87 8.071987 26.47 11.53 7.781988 25.53 10.94 7.051989 27.82 13.15 8.661990 29.79 12.99 8.641991 28.61 12.89 8.951992 29.43 12.38 7.761993 30.44 13.39 7.991994 32.64 13.33 8.401995 33.52 14.94 8.521996 36.57 16.07 9.99

a. As of June 30.SOURCE: Authors’ calculations.

15

top 10 and top five correspondent banks nation-wide. Despite the increased concentration ofcorrespondent banking deposits shown in table4, the national market for correspondent bank-ing remains relatively unconcentrated, with nofirm controlling more than 3.6 percent ofdeposits due to banks.

Unfortunately, most markets for correspon-dent banking services are likely to be local or

regional in scope and, therefore, national con-centration measures may prove misleading. Forexample, if correspondent banking markets areeffectively segmented by branching restrictions(intrastate or interstate), then low levels of mar-ket concentration at the national level may beconsistent with highly concentrated markets atthe state or local level. Moreover, to the extentthat banking consolidation increases the con-testability of correspondent banking markets,increases in market concentration at thenational level may lead to more competitivecorrespondent banking markets at the state andlocal levels.

Data on correspondent banking markets are collected at the bank level, limiting ourability to accurately gauge the competitivenessof these markets. Moreover, the degree ofaggregation in the data precludes us fromlooking at measures of concentration belowthe state level for both deposits due to banks(correspondent deposits) and deposits duefrom banks (respondent deposits).

To examine trends in market concentrationat the state level, we construct Herfindahlindexes for deposits due to banks (correspon-dent deposits) and for deposits due from banks(respondent deposits). Tables 5 and 6 reportthe average Herfindahl for all states, states withstatewide branching, states with limited branch-ing, and unit banking states. Table 5 shows thatcorrespondent banking is more concentrated,on average, in statewide branching states thanit is in states with more restrictive branchinglaws. However, while mean concentration forthe total of all states has risen over the sampleperiod, holdings of correspondent depositshave become less concentrated in statewidebranching states. The increasing mean concen-tration of correspondent deposits for all statesover time likely reflects the ongoing consolida-tion of the banking system, especially in thosestates switching to less restrictive branchinglaws during the sample period. Moreover, thedecline in the mean concentration of corre-spondent deposits for the statewide branchingstates over the sample period may simplyreflect the inclusion of new states in the sam-ple. If correspondent deposit markets in statesthat switched to statewide branching during thesample period are less concentrated than theaverage statewide branching state in the sam-ple, then mean concentration for the sampleshould decrease over time.

Table 6 shows that respondent deposits (de-posits due from banks) are also more concen-trated in statewide branching states. To the ex-tent that banking markets are more consolidated

T A B L E 5

Deposits Due to Banks:State-Level Herfindahl

All Statewide Limited UnitYeara states branching branching banking

1984 1,302.32 1,814.80 730.67 987.02

1985 1,328.69 2,060.31 577.37 632.83

1986 1,315.64 1,989.67 632.66 746.03

1987 1,322.13 1,863.52 657.61 847.76

1988 1,467.51 1,861.73 804.60 1,646.37

1989 1,379.42 1,826.39 612.90 1,635.88

1990 1,416.91 1,549.98 803.99 2,271.65

1991 1,476.31 1,580.03 1,139.21

1992 1,507.27 1,622.51 1,132.74

1993 1,379.61 1,396.87 1,316.83

1994 1,617.10 1,679.40 1,361.66

1995 1,602.90 1,665.53 1,346.11

1996 1,510.14 1,611.97 1,092.68

a. As of June 30.SOURCE: Authors’ calculations.

T A B L E 6

Deposits Due from Banks:State-Level Herfindahl

All Statewide Limited UnitYeara states branching branching banking

1984 2,121.30 3,084.66 1,210.53 1,160.00

1985 2,014.97 3,140.98 968.51 850.74

1986 2,240.79 3,561.01 1,021.35 858.86

1987 2,011.15 2,994.41 921.99 863.66

1988 1,854.20 2,714.22 895.89 488.00

1989 1,944.99 2,927.31 772.17 666.13

1990 2,176.47 2,794.63 699.32 787.21

1991 2,316.39 2,778.43 814.77

1992 2,591.18 2,951.34 1420.68

1993 2,708.97 3,186.43 972.77

1994 2,700.65 3,141.37 893.69

1995 3,022.00 3,486.61 1,117.09

1996 2,984.88 3,307.98 1,660.20a. As of June 30.SOURCE: Authors’ calculations.

16

in statewide branching states, one wouldexpect to see fewer banks placing depositswith correspondent banks and, therefore,greater concentration. Unlike correspondentbalances, however, respondent deposits do notclearly increase or decrease through time eitherfor the full sample or for statewide branchingstates. Therefore, the implication of interstatebranching on the concentration of respondentdeposits is unclear.

The negative relationship between concen-tration in interbank deposit markets and thestringency of geographic limitations (that is,branching restrictions) may simply be a conse-quence of more concentrated deposit markets.That is, relaxation of branching restrictions mayincrease the concentration of banking deposits—a likely outcome of a more consolidatedbanking system—which, in turn, leads to moreconcentrated interbank-deposit markets. Tables7 and 8 seem to bear this point out: Table 7exhibits average Herfindahls at the state levelfor domestic deposits. Table 8 presents averagemarket-level Herfindahls grouped by statebranching status.9 In both tables, statewidebranching states tend to exhibit more depositmarket concentration than other states, and thelevel of concentration has tended to increaseover the sample period.

Simple correlation analysis confirms a highdegree of correlation between interbank-depositconcentration and concentration measures ofdomestic deposit markets. The Spearman (Pear-son) correlation coefficient between the state-level Herfindahl indexes for domestic depositsand correspondent deposits is 0.7033 (0.7153).The Spearman (Pearson) correlation coefficientbetween the state-level Herfindahl indexes fordomestic deposits and respondent deposits is0.7131 (0.6756). Moreover, the structure of thecorrespondent deposit market and the respon-dent deposit market are highly correlated with aSpearman (Pearson) correlation coefficient of0.45665 (0.52298).

One problem with looking at time trends in Herfindahl indexes across states with differ-ent branching laws is that a number of stateschanged their laws during the sample period.This is evident from the disappearance of unitbanking in the sample in 1991. All changes inbranching status favored a less restrictive formof branching regulation. Therefore, a measureof caution is warranted when comparing trendsin concentration across subcategories in tables5 through 8.

Fortunately, we can directly test the impactof relaxing intrastate branching restrictions ondeposit market concentration. For all states thatchanged branching status during the sampleperiod (there were 22 such switches), we con-structed state-level Herfindahls for domesticdeposits, correspondent deposits, and respon-dent deposits two years prior to and two years

T A B L E 7

Domestic Deposits:State-Level Herfindahl

All Statewide Limited UnitYeara states branching branching banking

1984 803.57 1,413.75 245.13 155.41

1985 839.97 1,476.55 264.17 146.19

1986 829.74 1,434.67 290.64 152.30

1987 822.19 1,341.83 275.91 144.49

1988 874.87 1,363.18 305.17 191.21

1989 924.22 1,431.53 324.04 243.95

1990 948.53 1,216.49 310.75 284.19

1991 975.79 1,175.25 327.55

1992 1,037.58 1,254.06 334.01

1993 1,035.08 1,208.44 404.70

1994 1,079.96 1,233.59 450.09

1995 1,165.78 1,318.10 541.27

1996 1,301.08 1,449.48 692.67a. As of June 30.SOURCE: Authors’ calculations.

T A B L E 8

Domestic Deposits: Weighted Average Herfindahls for All Markets

All Statewide Limited UnitDate states branching branching banking

1984 2,269.58 2,502.29 2,309.44 1,379.66

1985 2,281.17 2,512.56 2,368.29 1,320.86

1986 1,886.94 1,983.10 2,288.14 1,337.33

1987 1,888.29 1,970.49 2,192.11 1,262.59

1988 2,451.90 2,597.36 2,231.45 1,750.54

1989 2,566.49 2,665.32 2,298.59 2,318.25

1990 2,674.95 2,786.77 1,850.64 2,046.39

1991 2,717.42 2,800.73 1,709.98

1992 2,936.77 3,057.20 1,723.09

1993 2,946.38 3,061.80 1,726.98

1994 3,048.91 3,024.30 2,298.79

1995 3,120.80 3,100.03 2,306.78

1996 3,280.86 3,182.87 2,922.62a. As of June 30.SOURCE: Authors’ calculations.

■ 9 In table 9, we assume that for bank offices located in MSAs therelevant market is the MSA. For bank offices located in non-MSA counties,the market is defined to be the county where the office is located.

17

after the event. We then tested to see if thechange in the Herfindahls due to the change in branching status was significant. The resultsof the analysis are reported in table 9. For do-mestic deposits and correspondent deposits(deposits due to banks), a switch to more lib-eral branching significantly increases the Her-findahl; however, the change in the respondent

deposit (deposits due from banks) Herfindahlis not significant.

In all, univariate analysis of the data sug-gests that a relaxation of branching restrictionsis associated with increased concentration inthe market for domestic deposits and in theinterbank deposit market. Neither of theseresults is surprising, as the removal of an artifi-cial constraint to geographic consolidation ofthe banking system would be expected toincrease concentration in banking markets.However, increased concentration does notnecessarily translate into less competitive mar-kets, as the removal of branching restrictionsincreases the potential for entry.

To separate the effects of branching statusfrom deposit market concentration, we conducta simple regression analysis—with the Herfin-dahl for correspondent deposits as the depen-dent variable. The results can be found in table 10. Model 1 regresses the correspondentdeposit Herfindahl on the state-level Herfind-ahl for domestic deposits and a time dummy.Model 2 adds a statewide branching dummyvariable (DSBRANCH = 1 for statewide branch-ing, zero otherwise). A significant coefficienton DSBRANCH suggests that after controllingfor the structure of the domestic deposit mar-ket, branching restrictions affect the structureof the correspondent banking market. Model 3controls for the market concentration for re-spondent deposits. To the extent that there arescale and scope economies associated with theprovision of correspondent banking services,increased concentration in respondent depositsmay reduce the number of correspondentbanks that can profitably operate in a market.Finally, model 4 extends the previous regres-sion by including information on local marketstructure—HLOCDEP, the average local de-posit market Herfindahl in each state. Inclusionof HLOCDEP allows us to control for the effectof local deposit market structure on correspon-dent banking.

The coefficient on the domestic depositHerfindahl (HDOMDEP) is positive and signifi-cant from zero in all four models. In addition,the coefficient on the proxy for local depositmarket structure (HLOCDEP) in model 4 is alsopositive and significant. This confirms the uni-variate results that find increased deposit mar-ket concentration associated with increasedconcentration in correspondent deposit mar-kets. The negative and significant coefficienton DSBRANCH in models 2, 3, and 4 suggeststhat once domestic deposit market concentra-tion is controlled for, relaxing branching re-strictions leads to less concentrated and more

T A B L E 9

Event Analysis: Changein Branching Status

2 years 2 years Percentprior to after Change in change in

Variable switch switch Herfindahl Herfindahl

Domestic 264.96 403.33 138.37a 81.78a

deposits 4.50 3.78

Deposits due 789.67 791.48 1.82 1.42from banks 0.02 0.17

Deposits due 828.22 1436.55 608.32a 144.26a

to banks 3.17 3.27

a. Significant at the 1 percent levelSOURCE: Authors’ calculations.

T A B L E 10

Regression Results

Dependent Variable: HDEPIDOM

Model 1 Model 2 Model 3 Model 4

Intercept 533.373 607.623 578.900 338.7347.175 7.987 7.581 3.402

Timedum –13.546a –6.202b –6.987b –12.697b

–1.546a –0.698b –0.790b –1.424b

HDOMDEP 1.022 1.114 1.028 0.94926.267 24.510 18.796 16.730

DSBRANCH –328.334 –369.127 –381.807–3.839 –4.277 –4.457

HINTBDEP 0.061 0.0702.815 3.160

HAVGDMKT 0.1083.854

Adj-R2 0.512 0.522 0.527 0.527

F-Value 348.217 241.887 185.302 145.451

Prob >F 0 0 0 0

a. Significant at the 10 percent level.b. Not significant.NOTE: Unless otherwise noted, all coeffcients are significant at the 1 percentlevel.SOURCE: Authors’ calculations.

18

competitive correspondent banking markets. Fi-nally, the positive and significant coefficient onthe respondent deposit Herfindahl (HINTBDEP)index is consistent with the hypothesis that a re-duction in the number of respondent banksreduces the number of correspondents that cancoexist in a market—and hence increases con-centration of correspondent deposits.

V. Conclusion

Interstate branching promises to change thecompetitive landscape in banking. As illustratedby the mega-mergers of 1998—which includedthe merger of Bank of America and NationsBank—geographic consolidation of the banking sys-tem is well under way. This consolidation willcertainly increase the concentration of depositmarkets at the national and regional level.Moreover, the preceding analyses suggest thatinterstate consolidation may even increasedeposit market concentration at the state andlocal level.

The evidence presented here indicates thatintrastate branching deregulation and the sub-sequent geographic consolidation of the bank-ing industry has led to increased concentrationin the correspondent and respondent depositmarkets. However, this increased concentrationin the interbank market appears to be a conse-quence of increased concentration in thedomestic deposit market associated with moreliberal branching rights. Controlling for thelevel of concentration in the domestic depositmarket, the effect of statewide branching is toreduce concentration in the correspondentdeposit market. This result is consistent withthe hypothesis that any positive effects ofincreased contestability of interbank marketsresulting from the removal of branching restric-tions mitigate (and may dominate) any negativeeffects on competition from increased concen-tration in interbank markets.

Overall, the evidence presented here sug-gests that interstate branching will result in moreconcentrated interbank markets, as the geo-graphic consolidation of the banking industry atthe national level will certainly reduce the num-ber of correspondent and respondent banksnationwide. This increased concentration in cor-respondent banking markets will not necessarilyreduce the competitiveness of these markets atthe state and local level because branchingderegulation also increases their contestability.

There are several caveats to these results:First, data limitations preclude our controlling

for important nonbank competitors in the corre-spondent banking market such as banker’sbanks, private clearinghouses, data processingfirms, and the Federal Reserve Banks. Second,to the extent that banking organizations use themultibank holding company to circumventbranching restrictions, the measured impact ofbranching deregulation on market concentrationand contestability will be overstated.

From a public-policy standpoint, if interstatebranching leads to the establishment of trulynational correspondent banks, then there maybe less justification for the Federal Reserve Sys-tem to provide correspondent banking services.However, more work needs to be done in thisarea before we can begin to seriously recon-sider the role that Federal Reserve Banks playin this market.

Appendix

Loan Participations,Federal ReserveMembership, andOther Determinantsof CorrespondentBalances

Early literature indicates that loan participa-tions are the second-most important serviceoffered by correspondents. However, informa-tion about this activity is not available on theperiodic reports submitted by banks. Knight(1970b) reports that 75 percent of banks expe-riencing an increased need for loan assistancecited the size of the loan as the main factor.The percentage of banks requiring loan assis-tance increased with bank size, and require-ments for assistance are positively related to theloan-to-deposit ratio, the latter possibly indicat-ing the importance of liquidity constraints.Knight (1970a) indicates that although mostloan participations originate with the smallerbanks, banks that do not experience excessloan demands can still buy loans or participa-tions from their correspondents. Many respon-dents maintained credit lines or borroweddirectly from their correspondents.

The correspondent services offered by theFederal Reserve System to its member bankshave differed somewhat from those offered byprivate correspondents, and this has been a sub-ject of contention and regulatory reform. Accessto the discount window is probably the mostobvious correspondent service that at some timemight have been available only to members.

19

Federal Reserve membership has been iden-tified as an important factor by several authors.The two reasons most often cited for nonmem-ber banks being more likely to use correspon-dent services were the possibility that stateswith reserve requirements would allow non-members to count balances due as reserves,and the granting of immediate credit by corre-spondents for cash letters received fromrespondents. Lawrence and Lougee (1970)found that balances due from banks, but notthe number of correspondent ties or their geo-graphical distribution, is related to Fed mem-bership. Member banks have higher balancesdue if balances at the Fed are included, but notif they are excluded. Knight (1970a) confirmsthis finding and also reports that the benefits ofFed membership appear to increase with banksize since, unlike overall banks, nonmemberbanks over a certain (small) size have corre-spondent balances that increase with bank size.Knight (1970a) reported that about 90 percentof banks surveyed preferred to send checksdrawn on nonlocal banks through correspon-dents rather than the Fed, apparently becauseimmediate credit was offered by the former.Although larger banks appear to have a greaterpreference for using the Fed, this could be mis-leading if correspondents send checks receivedfrom smaller banks on to the Fed for clearing.

Summers and Segala (1979) focus on thedeterminants of usage of Fed correspondentservices. They find that bank size, holding-com-pany affiliation, and metropolitan locationincrease the probability of a bank using Fedcheck-clearing services. However, only banksize affected usage of Fed wire services. Size isinterpreted as indicating administrative capacityto manage the services.

Several other determinants of correspondentbalances have been identified in the literature.For example, Lawrence and Lougee (1970)report that for banks in the Denver area, banksize, ratio of demand to total deposits, and dis-tance of the bank from Denver are positivelyrelated to the amount of domestic “due from”balances. The number of correspondent ties isrelated to the first two characteristics. Meinsterand Mohindru (1975) conclude that correspon-dent balances are influenced by liquidity consid-erations as well as the need to pay for corre-spondent services. The volatility of deposits isanother factor: Kane (1982) found that suppliersof correspondent services imposed higherreserve requirements on more volatile deposits,a practice which is presumably consistent withvolatility being related to the benefits derivedfrom the services. Gilbert (1983) identifies a rela-

tionship between deposit volatility and the scaleof correspondent services, with transaction costbeing lowered by a larger number of respon-dents and thus a higher level of variability.

Geographic variables also play a role. Dis-tance from the correspondent obviously is re-lated to the cost of providing certain services, as well as familiarity with the correspondent.Size of city could, at times, be related to sophis-tication and consequently to the need for cer-tain services.

References

Auerbach, Robert. “Changes in the Provisionof Correspondent-Banking Services and the Role of Federal Reserve Banks under the DIDMC Act: A Comment on Kane,”Carnegie–Rochester Conference Series on Public Policy, vol. 16 (Spring 1982), pp. 127–36.

Baker, George, Robert Gibbons, and Kevin J. Murphy. “Implicit Contracts andthe Theory of the Firm,” Harvard BusinessSchool, Working Paper No. 98-009, Decem-ber 1997.

Bauer, Paul W., and Gary D. Ferrier. “ScaleEconomies, Cost Efficiencies, and Techno-logical Change in Federal Reserve PaymentsProcessing,” Journal of Money, Credit, andBanking, vol. 28, no. 4 (November 1996),pp. 1004–39.

Dunham, Constance. “Commercial BankCosts and Correspondent Banking,” FederalReserve Bank of Boston, New England Eco-nomic Review (September/October 1981),pp. 22–36.

Finney, Katherine. Interbank Deposits: ThePurpose and Effects of Domestic Balances,1934–54. New York: Arno Press, 1958.

Flannery, Mark J. “Correspondent Servicesand Cost Economies in Commercial Bank-ing,” Journal of Banking and Finance, vol. 7,no. 1 (March 1983), pp. 83–99.

Gilbert, R. Alton. “Economies of Scale in Cor-respondent Banking,” Journal of Money,Credit, and Banking, vol. 15, no. 4 (Novem-ber 1983), pp. 483–88.

20

Kane, Edward J. “Changes in the Provision ofCorrespondent-Banking Services and theRole of Federal Reserve Banks under theDIDMC Act: Rejoinder,” Carnegie–RochesterConference Series on Public Policy, vol. 16(Spring 1982), pp. 93–126.

Knight, Robert E. “Correspondent BankingPart I: Balances and Services,” FederalReserve Bank of Kansas City, MonthlyReview (November 1970a), pp. 3–14.

———- . “Correspondent Banking Part II: LoanParticipations and Fund Flows,” FederalReserve Bank of Kansas City, MonthlyReview (December 1970b), pp. 12–24.

———- . “Correspondent Banking Part III:Account Analysis,” Federal Reserve Bank ofKansas City, Monthly Review (December1971), pp. 3–17.

———- . “The Impact of Changing CheckClearing Arrangements on the Correspon-dent Banking System,” Federal Reserve Bankof Kansas City, Monthly Review (December1972), pp. 14–24.

Lawrence, Robert J., and Duane Lougee.“Determinants of Correspondent BankingRelationships,” Journal of Money, Credit,and Banking, vol. 2, no. 3 (August 1970),pp. 358–69.

McAfee, R. Preston. “The Effects of VerticalIntegration on Competing Input Suppliers,”Federal Reserve Bank of Cleveland, Eco-nomic Review, vol. 35, no. 1 (1999), pp. 2–8.

Meinster, David R., and Rajesh K.Mohindru. “The Determinants of the De-mand for Correspondent Balances by Smalland Medium Sized Banks,” Journal of BankResearch, vol. 6, no. 1 (Spring 1975), pp. 25–34.

Merrett, D. T. “Global Reach by AustralianBanks: Correspondent Banking Networks,1830–1960,” Business History, vol. 37, no. 3( July 1995), pp. 70–88.

Milgrom, Paul, and John Roberts. Econom-ics, Organization and Management. Engle-wood Cliffs, N.J.: Prentice–Hall, 1992.

Moulton, Harold G. Principles of Money andBanking. Chicago: University of ChicagoPress, 1916.

Redlich, Fritz. The Molding of AmericanBanking: Men and Ideas. History of Ameri-can Business Leaders Series, vol. 2. NewYork: Hafner Publishing Company, 1951.

Summers, Bruce J., and John P. Segala.“Determinants of Correspondent BankingRelationships with the Federal Reserve Sys-tem,” Federal Reserve Bank of Richmond,Working Paper No. 79-1, March 1979.

Thomson, James B. “Interbank Exposure inthe Fourth Federal Reserve District,” FederalReserve Bank of Cleveland, Economic Com-mentary, August 1, 1987.

Todd, Walker F., and James B. Thomson.“An Insider’s View of the Political Economyof the Too Big to Let Fail Doctrine,” PublicBudgeting and Financial Management: AnInternational Journal, vol. 3, no. 3 (1991),pp. 547–617.

Williamson, Oliver E. The Economic Institu-tions of Capitalism: Firms, Markets, Rela-tional Contracting. New York: The FreePress, 1985.