Chevalier AER 1995

7/28/2019 Chevalier AER 1995

1/22

American Economic Association

Capital Structure and Product-Market Competition: Empirical Evidence from the SupermarketIndustryAuthor(s): Judith A. ChevalierSource: The American Economic Review, Vol. 85, No. 3 (Jun., 1995), pp. 415-435Published by: American Economic AssociationStable URL: http://www.jstor.org/stable/2118181 .

Accessed: 07/06/2013 06:40

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

.

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The

American Economic Review.

http://www.jstor.org

This content downloaded from 14.139.157.22 on Fri, 7 Jun 2013 06:40:10 AMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/action/showPublisher?publisherCode=aeahttp://www.jstor.org/stable/2118181?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/2118181?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=aea


2/22

CapitalStructureand Product-MarketCompetition:EmpiricalEvidence fromthe SupermarketIndustryBy JUDITH A. CHEVALIER*

Thispaper establishesan empirical ink between irm capital structureandproduct-marketompetitionusing data from local supermarket ompetition.First,an event-study nalysisof supermarketeveragedbuyouts(LBO's) sug-gests that an LBO announcementncreasesthe marketvalue of the LBOchain's ocal rivals.Second, showthat supermarkethainsweremore ikely oenterandexpandn a localmarket f a large hareof theincumbentirms n thelocal marketundertookLBO's. Thestudysuggests hatleverage ncreases n thelate 1980's ed to softerproduct-marketompetitionn thisindustry. JEL D43,G14, G32, G34, L13, L81)

During he late 1980'scorporatedebtrosedramatically,due in large part to an un-precedented wave of leveraged buyouts(LBO's). This large-scaleexperiment withfirm capital structurerefocused both popu-lar and academicattentionon the issue ofhow a firm'sfinancingchoices might affectits performanceand behavior. Numerousrecent theoreticalworkshaveexaminedonecomponent of this issue-the question ofhow a firm'scapitalstructureaffectscompe-tition in the market or the firm'sproducts.However,verylittle work has been done todetermineempiricallywhether a real link-age exists between capital markets andproductmarkets.'

In this paper, I test between two classesof theoreticalmodelsbyexamininghe shareprice responseof supermarket hainsto theannouncementof a rival chain's leveragedbuyout and by examining the entry, exit,and expansion behavior of supermarketchains. The first class of models predictsthat increases in firm leverage tend to"soften" product-market ompetition. Thesecond class of models predicts that in-creases in firmleverage tend to "toughen"product-marketompetition.2A finding hateither of these two hypotheses s truewouldbe important n that it would suggest thatfinancingdecisions can have real product-marketeffects.I examine the effect of debt on product-market competition by studying super-marketchains in local markets.The super-marketindustry s a naturallaboratory ortestingthese theoreticalmodelsfor tworea-sons. First, many large supermarket hainsundertook everagedbuyouttransactions nthe late 1980's.This allowsexaminationofhow product-marketcompetition "shakesout" after competitorsundertake sudden,

* Graduate School of Business, University ofChicago, 1101 East 58th Street, Chicago, IL 60637, andNational Bureau of Economic Research. This paper isa revised version of Chapters 1 and 2 of my M.I.T.dissertation. My advisors, Paul Joskow and DavidScharfstein, provided invaluable guidance and advice.Christopher Avery, Tasneem Chipty, Glenn Ellison,Franklin Fisher, Ruth Judson, Steven Kaplan, BrigitteMadrian, Christopher Mayer, Wallace Mullin, JamesPoterba, Nancy Rose, Julio Rotemberg, Fiona ScottMorton, Jeremy Stein, and two anonymous refereesprovided helpful comments. Charles Hadlock providedinvaluable data guidance. Financial assistance from theNational Science Foundation at the beginning of thisproject is gratefully acknowledged. All errors and omis-sions remain my own.1The only empirical work of which I am aware is A.Michael Spence (1985), Jose C. Guedes and Tim C.Opler (1992), and Gordon M. Phillips (1992).

2The terms "tough" and "soft" price competitionare used in the sense of John Sutton (1991). The"toughness"of price competitiondiffersin two mar-kets if, holdingthe concentration n the two marketsconstant,price-costmargins n the two marketsdiffer.415

http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


3/22

416 THE AMERICANECONOMIC REVIEW JUNE 1995dramatic ncreases in debt. Second, super-marketcompetition akesplace at the locallevel. This allows a comparison of super-marketcompetitionacrossmarkets.

The findings n this paper are consistentwith the group of theoretical models ofcapital structureand product-market om-petition that suggest that product-marketcompetitionbecomes "softer" when lever-age increases. These models include workby Drew Fudenbergand Jean Tirole (1986),Patrick Bolton and David S. Scharfstein(1990), and Phillips (1991). The resultsareinconsistent with other models, includingJames A. Brander and Tracy R. Lewis(1986), Vojislav Maksimovic (1988), andJulio J. Rotembergand Scharfstein 1990),which predict that leverage changes man-agerialand shareholder ncentives n a waythat makes product-market competition"tougher."The organization f the rest of the paperis as follows.Section I describesLBO activ-ity in the supermarketndustry.Section IIpresents an event study examining he an-nouncement effects of supermarket ever-aged buyoutson rival supermarket hains.Section III describes the empirical predic-tions of the hypothesesof capital structureand product-market ompetitionfor entry,exit, and expansion. Section IV describesthe data for the study of entry, exit, andexpansion.Section V presents the results.Section VI examinesan alternative xplana-tion for the empiricalresults in Section V,and my conlcusionsare presented in Sec-tion VII.

I. LBO Activity in the SupermarketIndustry

A supermarkets definedby the publica-tion Progressive Grocer (1989) as a retailfood store that has annual sales of morethan $2 million and has greaterthan 9,000square feet of selling space. Supermarketsaccount for 70 percent of retail food storesales but only 10 percent of retail foodestablishments. According to ProgressiveGrocer (1989), there were approximately30,754supermarketsn the United Statesin1988,55 percentof whichbelongedto chainsof 11 or more stores.

At the national level, the supermarketindustryappearsto be relativelyunconcen-trated. The four largest supermarket hainsaccountedfor only 16 percentof U.S. gro-cery store sales in 1982(PhillipK. Kaufmanand CharlesR. Handy, 1989). However,nosupermarket hain in the United States istrulynational.For example, he largestchainin the United States, AmericanStores,op-eratedin only 18 states in 1990.Thus, whilethe industry s relativelyunconcentrated na national evel, local marketscan be highlyconcentrated. The average metropolitanstatistical area in the United States had afour-firm supermarketconcentrationratioof 58 percentin 1982.SupermarketLBO's occurred primarilybetween 1985and 1988.3The largesttrans-actions were the $5.3 billion SafewayLBO,the $4.1 billion Kroger everagedrecapital-ization, the $1.8 billion SupermarketsGen-eral LBO, and the $1.2billionStop & ShopLBO. These four companies alone ownednearly 4,000 U.S. supermarkets t the timeof their LBO's. During this period, it wasalso common for smaller regional chainsand divisionsof largerchains to undertakeLBO's.Altogether,19 of the 50 biggestsu-permarket hainsin the United StateshaveundertakenLBO's.They accountedfor ap-proximately 72 billion of the $297 billioninsupermarket ales in 1991.LBO activityhas not been concentratedin any single geographic egion.In the sam-ple used in this study,LBOfirmsaccountedfor 16 percent of the stores in Midwesternmarkets,17 percent of the stores in South-ern markets, 21 percent of the stores inNortheasternmarkets, and 42 percent ofthe stores in Westernmarkets.Part of thelarge LBOconcentrationn the West is dueto the enormous mportanceof Safewayon

3There is also one instance of a leveraged recapital-ization in this industry, which was undertaken byKroger. A leveraged recap is a transaction in which afirm borrows in order to pay a large dividend to share-holders of at least 50 percent of the former equityvalue of the firm. Because this recap resulted in debtlevels for Kroger similar to typical LBO debt levels, itis included in this analysis as an LBO.



4/22

VOL. 85 NO. 3 CHEVALIER: CAPITALSTRUCTUREAND MARKETCOMPETITION 417the West Coast. Safeway's 1985 marketshare in cities in the sample in the Westtotalednearly25 percent.The vast majorityof the leveragedbuy-outs were not the result of unconstraineddecisionsby managementand shareholders.Instead, most of them were undertaken nresponse to unwanted akeoverattempts.Infact, all four of the biggest deals (andmanyof the smaller ones) were undertakentothwartthe unwantedtakeoverattemptsofthe Haft family, which controls the Dartdrugstorechain.This descriptionof LBO activity in thesupermarketndustryeaves unanswered hequestion of what caused the LBO's in thisindustry.4It has been suggested by PeterMagowan (1989), the CEO of Safewaystores, which undertook an LBO in 1986,that the main effect of the Safeway LBOwas to force Safeway to sell or spinoff divisions which were not profitable.Magowan uggeststhat Safewayexcisedthedivisions in which it was not as strong acompetitoras its rivals see Magowan,1989).This suggestionby Magowan s practicallyrestatementof the basic "empire-building"rationale for LBO's suggestedby MichaelC. Jensen (1989). If a good firm has a fewbad divisions,then value can be gained bybuying the firm and turning over the baddivisions o higher-valued sers.These post-LBO asset sales were commonamongLBOchains. If this model of LBO's is correct,then, in the studyof entry,exit, and expan-sion, by examining the local markets inwhich Safeway chose to remain an active

competitor,one examines those local mar-kets where competitionwas not an impor-tant cause of the LBO.5The local-marketnature of supermarket ompetitionhelps to"cleanout" the endogeneityof the LBO inthe study of entry, exit, and expansion.Theissue of how the endogeneity of LBO'saffectsthe results will be taken up in Sec-tion VI.II. An Event Study of Supermarket LBO's

In this section,I examine he stock-returnresponse of supermarket hains to the an-nouncementthat a rival chain is undertak-ing a leveragedbuyout.6This approachoflooking at the event responsesof rivalfirmswas pioneered by B. Espen Eckbo (1983)and Robert S. Stillman 1983)in the mergerliterature. f leveragedbuyoutsareexpectedto makeproduct-marketompetitionsofter,rivalsupermarket hainsshouldexhibitpos-itive abnormal returns around the timeof LBO announcements; if LBO's areexpected to lead to tougher product-market competition, rival supermarketchainsshould exhibitnegative abnormal e-turns around the time of LBO announce-ments.

A. Methodology and DataI focusmyanalysison a singleindustry norder to separate the hypothesis that theLBO leads to a change in product-market

4Perhaps, given the importance of the Haft family ininitiating takeover attempts in this industry, the betterquestion to ask is: what factors contributed to the Haftfamily's choices of takeover targets? An examination ofnewspaper and magazine accounts of these takeoversreveals no statements by the Haft family indicating whyit chose particular targets. One might hope to analyzethe family's plan by examining the reforms that thefamily instituted at the supermarket chains that wereactually taken over. However, the only supermarketchain that the Haft family actually took over wasShoppers Food Warehouse. It is difficult to try to inferwhat the Haft family would have done with Safeway orKroger by examining what changes they effected at achain with only 30 stores.

5The explanation suggested here for LBO's-thatthey serve to force firms to excise bad divisions, and toleave them only with divisions in which they are rela-tively efficient producers-is supported by the evi-dence. LBO firms are not statistically significantly morelikely to be closing stores in unclosed divisions thannon-LBO firms. At least in this regard, their behaviorin the divisions not sold off seems unaffected by theLBO.61 do not focus on the stock-return response of theleveraging firm itself. A leveraged buyout transactionoccurs when the managers of the firm (or others) offerto pay a premium over the prevailing market price ofthe firm. The stock price rises to reflect this premium.The managers of the firm would not undertake aleveraged buyout if they did not believe that they couldimprove firm value.



5/22

418 THE AMERICAN ECONOMIC REVIEW JUNE 1995competition rom the alternativehypothesisthat abnormalevent returnsare due to in-creased speculationthat more LBO's willoccur in the industry. I use informationabout local-market ompetitionto separatefirms in the industry hat are directlycom-peting with the leveraging irms from firmsthat are not directly competing with theleveragingfirms.7If an LBO greatly im-provesthe financialoutlookof the firmun-dertaking he LBO and the LBO announce-ment increases speculation that othersupermarket chains will also undertakean LBO, then one wouldexpectall firms nthe industry o experiencea positivestock-return response to the LBO announce-ment. However,if the LBO is expected tosoften product-marketcompetition, thenone wouldexpect supermarket hainsoper-atingin the samelocal marketsas the lever-agingchainto exhibita positivestock-returnresponseto the LBO announcement.Super-marketsthat do not compete directlywiththe leveragingchain should have no returnresponse to the LBO announcement.Fi-nally, if LBO's are expected to makeproduct-market ompetition tougher, thendirectcompetitors houldexperiencea neg-ative share price response, while noncom-petitors should experience no significantsharepriceresponse.I examine the leveragedbuyoutsof Safe-way, SupermarketsGeneral, and Stop &Shop and the leveraged recapitalizationofKroger, the largest leveraged transactionsundertaken n the industry.The event win-dow begins 30 days prior to the first an-nouncement suggestive that an LBO orleveraged recapitalizationmight occur andends on the day of the firm's final an-nouncement hatit was undertaking n LBOor leveraged recapitalization.8Announce-

ments and announcementdates were ob-tained from the Wall Street Journal Index.Since all four of the transactionsstudiedhere were undertaken n response to take-over attempts, he eventwindowbeginspriorto the firstpublic announcement uggestingthat a takeover might occur. Because theevent window extends until the announce-ment of the LBO, it can be interpretedasreflecting the market'sexpectationof thechangein the value of the rival firm due tothe LBO, as long as no confounding nfor-mationwas released within the event win-dow.The first announcement leading up toeach leveragetransactionunder study andits date are listed in Table 1. Table 1 alsolists the date of the announcement hat theleveragetransactionwould definitelyoccur.The daily stock returns of 13 supermarketrivals are studied. The rivals are listed inTable 2. These rivalsrepresentall firms hatderivedat least 80 percentof their revenuesfromsupermarketales and tradedcontinu-ously from January1, 1985, throughOcto-ber 10, 1988,the date of the announcementof the Kroger everagedrecapitalization.The eventstudywas conductedusingdailydata on the stock-marketreturns of thesupermarket hains.The equationto be es-timated,a variantof the basicmarketmodelhas the form following orm:9

=it,a +,3iRmt + E,SijDjt + eit

where Rit= firmi's returnat date t, Rmt =the return on the value-weightedNYSE/AMEX index at date t, j indexes the fourevents, ai, f8i,and 3ijare parameters o beestimated,eit is an errorterm,and D- is adummyvariablewhich equals 1 duringtheeventwindowfor event j and 0 otherwise.The stockreturndata used for estimationstart on the first tradingday of 1985 andextend throughKroger'sannouncementof7MichaelD. Whinstonand Scott C. Collins(1992)separate competingairlines from noncompetingair-lines in their event study of the entry of People Ex-press.8This wide event windowwas chosen because con-victed insider-tradervan Boeskywas investigatedby

the SecuritiesExchangeCommissionor insider-trad-ing in at leastone of these transactions. 9For a discussion of the market model, see EugeneF. Fama (1976).



6/22

VOL. 85 NO. 3 CHEVALIER:CAPITALSTRUCTUREAND MARKET COMPETITION 419TABLE 1-EVENTS INCLUDED IN THE EVENT ANALYSIS

Date finaltransactionEvent First announcement Date announcedSafeway LBO Dart Group announces 6/13/86 7/29/86that it has acquired a 6-percentstake in SafewaySupermarkets General LBO Dart Group proposes to buy 3/10/87 4/23/87Supermarkets GeneralStop & Shop LBO Dart Group announces that 1/15/88 3/1/88it seeks a major stakein Stop & ShopKroger leveraged recap The Haft family (who control 9/13/88 10/10/88Dart Group) reveals thatit has a major stakein KrogerNote: The event window is from 30 days prior to the first announcement, through theannouncement of the final transaction.

TABLE 2-RIVAL SUPERMARKET CHAINS INCLUDEDIN THE EVENT STUDY

AlbertsonsAmerican StoresBrunosDelchampsFood LionFoodaramaGiant Food StoresGreat Atlantic and Pacific Tea CompanyHannaford BrothersMarsh SupermarketsRuddickWeis MarketsWinn-Dixie

its leveraged recapitalizationin October1988. These data are obtained from theCenterfor Research in SecurityPrices.The event response parametersare esti-mated using seemingly unrelated regres-sions (SUR). This methodologys employedbecause the error terms from the market-model equation for a supermarketchainshould be contemporaneously correlatedwith the errorterms for other supermarketchains.For each debt event, I calculate theaverageevent coefficientof firms that com-pete directlywith the leveraging irm,and Icalculate the average event coefficient offirms that do not compete directlywith theleveraging irm.The firmsdirectlycompet-ing with the leveraging irm are constrained

to have a single event response, and thefirmsnot directlycompetingwith the lever-aging firm are constrained o have a singleevent response for each event. These con-straints cannot be rejectedat conventionalsignificance evels.For each event, volumesof the Supermar-ket News's annual Distribution Study ofGrocery Store Sales were used to determinewhich chains competedwith the chain un-dertaking the LBO. This book lists thenames of stores operating in each of theMetropolitanStatistical Areas (MSA's) inthe United States. The Supermarket Newsguide lists store names, not parent firms.Informationfrom annual 10K filings withthe Securities and Exchange Commissionand the 1988 Retail TenantsDirectory wereused to link store names to parent firms.Twosupermarket hainswere considered obe in directcompetitionwith one another fthere was any MSA in which both ownedstores listedin the Supermarket ewsguide.

B. ResultsTable 3 shows the results of an SURestimation of returnresponses to the fourevents. The returnresponsesfor the com-peting firms are shown in column 1; thereturnresponses or noncompetingirmsareshown in column 2. The table shows thatthe returnresponsesof the competing irms



7/22

420 THEAMERICANECONOMICREVIEW JUNE 1995TABLE 3-EVENT COEFFICIENTS

Event coefficient for othersupermarket chains(1) (2)Event Competing Not competing

Safeway LBO 0.003168** 0.001614(0.000966) (0.001044)Supermarkets 0.001782 - 0.000175General LBO (0.001141) (0.000817)Stop & Shop LBO 0.001573 -0.000082(0.001381) (0.000793)Kroger leveraged 0.001857* 0.000991recapitalization (0.000930) (0.001229)Notes: Column 1 reports event coefficients for firmscompeting in some of the same MSA's as the "event"firm; column 2 reports event coefficients for firms com-peting in none of the same MSA's as the "event" firm.The coefficients were estimated using seemingly unre-lated regressions. Standard errors are given in paren-theses.*Statistically different from zero at the 5-percentlevel.**Statistically different from zero at the 1-percentlevel.

are positive for all four events, consistentwith the hypothesisthat competition s ex-pectedto become softerfollowing he LBO.The returnresponsesare statistically ignif-icant at the 5-percent evel for the Safewayand Kroger events but statisticallysignifi-cant at only the 12-percentlevel for theSupermarketsGeneralevent and at the 25-percent level for the Stop & Shop event.The joint hypothesis that the event coeffi-cients for all four events equal zero is re-jected at the 1-percentsignificance evel.The returnresponsesfor the noncompet-ing firms are positive for two of the fourevents,butarenot statistically ifferent romzero at standard significance levels. TheSafewayevent is statisticallydifferentfromzero at the 12-percent evel. It is not sur-prisingthat this event is the one in whichthere is some responseof the noncompetingfirmssince the Safeway LBO was the firstlarge LBO in this industry,and thus, thespeculation effect might have been impor-tant for this LBO. However,the joint hy-pothesis that the event coefficients for all

four eventsequal zero cannot be rejectedateven the 55-percentsignificanceevel.The measured return responses for thecompeting firms, while statisticallysignifi-cant as a group, appearto be quite small.The Safeway event, for example, is esti-mated to lead to an abnormal ncrease invalue for the competing firmsof only 0.32percent, andthe othereventcoefficientsareeven smaller.However,this smallevent re-sponsemay be due to the smallamount hateven the "directcompetitors" ompetewiththe firm undertaking he LBO. For exam-ple, at the time of the Kroger leveragedrecapitalization,American Stores ownedapproximately 1,500 stores in 29 states;Krogerownedapproximately ,400stores in25 states. However,I could find records ofthe two firmscompeting n the same MSAin only seven states. Thus, the magnitudesof the coefficientsestimateddo not reflectthe changesin the expectedprofitability fthose operationsof each supermarket ivalthat actuallycompetes with the leveragingchain.10The event-studyresults suggest that thepresent discounted value of the expectedfuture profits of a supermarket hain riseswhen a rivalsupermarket hain announcesthat it is undertakingan LBO or leveragedrecapitalization.These results are consis-tentwiththe hypothesis hatproduct-marketcompetitionfollowingthe LBO is expectedto become softer. In the next section, thishypothesis is tested furtherusing data onthe entryand exit of supermarket hains.

III. Empirical Predictions for Entry andExpansionThe theories of capital structure andproduct-market competition posit thatchanges in capital structure change thetoughness of product-market competition. Ifan LBO changes the toughness of product-market competition, then, following the

10R. Preston McAfee and Michael A. Williams(1988) make a similar point about the estimated magni-tudes of the effects of mergers on rival firms.



8/22


9/22

422 THE AMERICANECONOMIC RE;VIEW JUNE 1995changed between the two years.'4The useof MSA-leveldata may be of concern,sincethereis no reason to assumethat an MSAisthe correctmeasure of the relevantmarketwhen considering upermarket ompetition.However,myexamination f the ProgressiveGrocer data shows that the MSA's corre-spondcloselyto divisionsof largesupermar-ket chains. In general, all of the supermar-kets that one chain has in a division areoverseenby a singledivisionalmanagerandare served by a single division warehouse.Furthermore,one importantway in whichsupermarketscompete is by distributingweeklycircularsn the local newspapershatdescribethe sales in the supermarketshatweek. A single flyer is generally ssued forall of the supermarkets n a division or,approximately,ll of the supermarketsn anMSA.All of the firms in the 85 MSA's areclassified by whether or not they have un-dertakenan LBO. I use this mechanism odivide firms into low-debt and high-debtfirmsbecause actual leverageratios are un-available for privately owned firms. Thepower of the test is weakenedby the factthat manyof the "low-leverage"irmsmayhave reasonablyhighlevelsof debt althoughthey did not undertakean LBO.'5The informationon LBO'swas obtainedin two ways. First, quarterlyeditions ofMergersand Acquisitions contain all owner-ship transactions (including LBO's) ofgreater than $1 million. Second, all refer-ences to transactions nvolvingthe super-marketparentcompanies n the sampleweresearched using Predicasts Funk and ScottIndex, United States, which indexes Super-market News, Supermarket Business, andProgressiveGrocer, the major industry trade

publications.From these sources, a defini-tive list of LBO'swas assembled.A lever-aged or LBO firm is definedas a firmthatunderwentan LBO (or leveragedrecapital-ization) anytime between 1981 and 1990.LBO firms ypically xit several ocal mar-kets followingthe LBO, usually by sellingthe local division to anotherchain or spin-ning it off to the division'smanagers soonafter the LBO. In total, 633 of the 13,512supermarkets n the study were sold in apost-LBO asset sale. Of these 633 super-markets, 187 were sold to the division'smanagement n a second LBO of the divi-sion.My approach s to treat the assets as ifthey were always owned by the eventualpurchaser.'6I take this conservativeap-proach because, otherwise,one would seeincreased entry into LBO markets simplybecause of these asset transfers, not be-cause of a change in post-LBO product-marketcompetition.For example, Safewaysold its SouthernCaliforniadivision o Vons shortlyaftertheSafewayLBOin 1986.Here,I add the storesin my sample that were part of Safeway's

SouthernCaliforniadivision o Vons'sstoretotal for 1985.Thus, for a SouthernCalifor-niacity,the change n storesfor Vons equalsthe net total of Safeway and Vons storesopened or closed in that citybetween 1985and 1991. The change in the number ofSafewaystores in any SouthernCaliforniacity equals zero. In constructing ndepen-dent variablessuch as the LBO share of amarket, he same convention s used.Mergersamongnon-LBO irmswere han-dled in a similarway. The stores of twofirmswhich merged were treated as if theywere alwaysowned by the same firm.Theone exceptionto this rule is that informa-tion about the acquisitionsof very smallindependentchains was not generallyavail-able. Purchasesof small ndependentchainsby chainsin the sampleare thus countedasentryor expansion.

14Unfortunately, because the MSA's were redefinedfor most of New England, the Bridgeport, Connecticut,MSA is the only New England MSA appearing in thesample. This removes from consideration most of oneLBO chain which was very successful (Stop & Shop)and most of another which was very unsuccessful(Supermarkets General).I have confirmed that the debt ratios of non-LBOfirms with publicly traded debt or equity are in fact,much lower than the debt levels of LBO firms.

16Thistreatment s undertaken n constructing oththe dependentand the independent ariables hatwillbe describedater.



10/22

VOL. 85 NO. 3 CHEVALIER: CAPITALSTRUCTUREAND MARKET COMPETITION 423Information about asset sales was ob-tained by checking the Wall Street JournalIndex, Mergers and Acquisitions, Supermar-ket News, Supermarket Business, and Pro-gressive Grocer. Demographic data areobtained from Donnelly MarketingInfor-mation Services, a market research firmwhich providedthe demographicdata forthe Progressive Grocer volume.

V. Methodology nd ResultsA. Full-MarketRegressions

I first test the hypothesis that LBO'schangedproduct-market ompetitionby de-terminingwhether LBO's lead to more orfewer stores"fitting" n the local market.Ifa marketcan supportmore stores followingan LBO than it could before (adjusting orother changesin the market), his supportsthe hypothesisthat LBO's "soften" prod-uct-marketcompetition.If fewer stores fit,this supports the hypothesis that LBO's"toughen"product-market ompetition.The strategyemployedhere is to measurethe percentage change in the number ofstoresin each MSA between 1985 and 1991and to check whether this measure is re-lated to the share of stores in each marketin 1985 owned by chains that eventuallyundertook LBO's. This allows measurementof how market structurechanged over theperiodin which the LBO'stook place.The specifications n this section controlfor several actorsthatmightbe expectedtocontributeto the growth of the number ofsupermarkets n a local market over thisperiod:the growth n the numberof house-holds in the MSA and household growthadjustedfor MSA area, the growthin me-dian incomeand its square,and the changein the share of households that have anincome of less than $10,000were included.These variablesare described n Table 4.The use of these five market characteris-tics implicitlyassumes hat each marketwasin an equilibrium tate in 1985:changesinthe market tructurebetween1985 and 1991should be due to changes in the marketcharacteristicsbetween 1985 and 1991. Toadjustfor the possibility hat an MSA was

in an "over-stored" or "under-stored" equi-librium n 1985, I also includetwo charac-teristicsof the marketin 1985:a measure-ment of a city's deviation romthe expectednumber of stores per household in 1985,and a measure of market concentration n1985. These variables are also described nTable 4. The variableof interest,the shareof LBO firmsin the MSA, is the share ofstores in the market in 1985 owned by asupermarket hain that wouldundertakeanLBOby 1990.Table4 provides ummarytatistics or allof the variablesused in this andsubsequentspecifications. Results for an ordinaryleast-squares OLS) regressionof the per-centage change n the numberof storesin acity between 1985 and 1991 on the LBOshare of the market and the controls formarket conditions described above areshownin column1 of Table 5.Table 5 shows that the LBO share of themarket has a positive coefficient,but thecoefficientis only statistically ignificantatthe 22-percent evel. The magnitudeof thecoefficient mplies that, if a firm n an MSAowning10 percentof the stores undertakesan LBO, the numberof storesin the marketis expectedto growby 1-percentmore thanit would otherwise. This insignificant ffectmay be due to there being,in fact, no effectof LBO's on the toughness of product-market competitionor may be due to thefact that there simply may not have beenenough time for market conditions to re-spondfullyto the LBO's.The second columnof Table 5 makes apreliminary ttempt o ascertainwhetheraneffect might have been observedhad moretime elapsedsince the LBO's.It repeatstheregression specificationof the first columnbut separatesthe LBO share of total storesinto two groups:the share of stores thatundertook LBO's prior to 1988 and theshare of stores that undertookLBO's dur-ing or after 1988.This is done because,if ittakes time for the market to adjust tochanges in competition,one might not ex-pect to see much response to the laterLBO's. Indeed, Table 5 showsthat the co-efficient for the store share of LBO firmswhich took place prior to 1988 is positive



11/22

424 THE AMERICANECONOMIC REVIEW JUNE 1995TABLE 4-SUMMARY STATISTICS

StandardVariable Description Mean deviationA. Variables:Change in households (10,000's) Change in the number of households in 5.50 5.93[percentage change households] the MSA between 1985 and 1991 [12.0] [10.1]Change in households per square mile Change in households per square mile in 49.7 157[percentage change households the MSA between 1985 and 1991. This [0.01] [0.03]per square mile] is included because the change inhouseholds may have a different im-pact if spread over a very large orvery small area.Change in median income ($10,000's) Change in median income in the MSA 1.53 4.73

[percentage change in median income] between 1985 and 1991 [64.2] [16.5]Change in median income squared Change in squared median income in the 9.95 4.66(in $1 x 108) [percentage change MSA between 1985 and 1991 [172] [53.7]in squared income]Change in the share households with Change in the share of households with - 0.0534 - 0.0227income less than $10,000 annual incomes of less than $10,000in the MSA between 1985 and 1991Deviation in mean stores per household MSA's deviation in 1985 from the num- 0.00 30.0[percentage deviation in mean stores ber of stores that it would be pre- [-7.11] [24.8]per household] dicted to have given the number ofhouseholds in the MSA in 1985. Iestimate that MSA's in 1985 have 40stores plus 2.3 x 10-4 stores perhousehold.Share of LBO firms Share of stores in the MSA in 1985 owned 0.220 0.191by firms that would undertake LBO'sby 1991Share of early-LBO firms Share of stores in the MSA in 1985 owned 0.108 0.156by firms that would undertake LBO'sprior to 1988Share of late-LBO firms Share of stores in the MSA in 1985 owned 0.111 0.128by firms that would undertake LBO'sin 1988 or laterHerfindahl index Sum of the squared market shares of the 0.120 0.020five firms with the largest marketshares in the MSA, where a firm'smarket share is defined as its share oftotal stores in a marketB. IncumbentFirm Variables:Store share, non-LBO incumbents Incumbent firm's share of the total stores 0.116 0.098in the MSA in 1985Store share, LBO incumbents Incumbent firm's share of the total stores 0.143 0.108in the MSA in 1985Total stores in chain, non-LBO incumbents Total stores in the sample of 100 MSA's 307 284owned by the incumbent in 1985Total stores in chain, LBO incumbents Total stores in the sample of 100 MSA's 283 227owned by the incumbent in 1985



12/22

VOL. 85 NO. 3 CHEVALIER: CAPITALSTRUCTUREAND MARKETCOMPETITION 425TABLE 4- Continued.

B. IncumbentFirm Variables: Continued)Number of firm-MSA observations in which non-LBO incumbents add stores: 79Number of firm-MSA observations in which non-LBO incumbents neither add nor subtract stores: 20Number of firm-MSA observations in which non-LBO incumbents subtract stores: 85Number of firm-MSA observations in which LBO incumbents add stores: 47Number of firm-MSA observations in which LBO incumbents neither add nor subtract stores: 19Number of firm-MSA observations in which LBO incumbents subtract stores: 47Number of MSA's in which de novo entry occurs: 39

TABLE 5-OLS SPECIFICATIONSCoefficients

Variable (1) (2) (3)Constant - 0.0070 0.0075 - 0.0006(0.1386) (0.1380) (0.1572)Percentage change in households 0.6347** 0.5952** 0.5377*(0.1569) (0.1581) (0.1651)Percentage change in income - 0.6731 - 0.7661 - 0.5778(1.1708) (1.6374) (1.1935)Percentage change in income squared 0.1974 0.2282 0.1918(0.3558) (0.3537) (0.3630)Change in share with income less than - 0.0071 - 0.0076 - 0.0044$10,000 (0.0084) (0.0084) (0.0088)Percentage change in households per 58.7209 60.0863 52.8402square mile (54.0745) (53.6788) (54.2677)Percentage deviation from mean stores -0.0866 -0.0831 -0.1194per household (0.0575) (0.0571) (0.0620)Herfindahl index - 0.1079 -0.1171 -0.2920(0.2818) (0.2798) (0.3036)Share LBO 0.0966(0.0775)Share early LBO 0.1736* 0.1438(0.0931) (0.1369)Share late LBO -0.0175 0.0280

(0.1094) (0.1141)Regional dummies included? no no yesR2 0.30 0.32 0.34N 85 85 85Notes: The dependent variable is the percentage change in the total number of storesin the MSA between 1985 and 1991. Standard errors are in parentheses.*Significantly different from zero at the 5-percent level.**Significantly different from zero at the 1-percent level.



13/22

426 THEAMERICAN ECONOMIC REVIEW JUNE 1995and significantat the 7-percent significancelevel. The coefficientfor the share of laterLBO's is insignificant t standard evels.

It has been suggested that regionaldummyvariablesshould be includedin thisregression to control for unmodeled cityheterogeneity. The regression was reesti-mated using dummy variables for theNortheast, Midwest, and South (with theWest as a base case). None of the dummyvariableswas statistically ignificantat eventhe 30-percent level. The inclusion of thedummies hrank he estimatedcoefficientofthe share of early LBO's slightly, to 0.144from 0.174, and decreased the statisticalsignificance level of the coefficient to30 percent. The coefficientfor late LBO'sbecame positive,but remained nsignificant.These results offer a preliminary ugges-tion that the presenceof leveragedfirms inthe marketdoes lead to a changein marketstructure.However,many hypothesesotherthan a change in the "toughness" ofproduct-market ompetition could be putforth to explain these results.For example,if undertakingan LBO greatly decreasedafirm's otalcosts, then LBO firmsmightfindit profitable o expand.Markets populatedby these firmsmight experiencefaster totalstore growth than other markets. In thefollowing section, I introduce tests whichwill avoid these alternativehypothesesandwhich attemptto measureseparately hosechanges n marketstructurewhichwouldbeexpectedto occur relativelyquicklyfollow-ing the LBO from those which might beexpectedto take moretime to occur.

B. Expansion by IncumbentFirmsIn this section, I examinethe questionofwhy large supermarket hains that are ac-tivelycompetingin a marketmight chooseto add or subtractstores in that market onnet. To do this, I identify he 50 chainswiththe largestnumberof stores in 1985 in thesample.17After adjustingfor mergers and

acquisitionsas described in Section III, 48chains are left for study.It is the expansiondecisions of these firms in each of the 85markets n which they are incumbents hatwill be studied. These firms account for6,068 of the 13,512 supermarkets n theMSA's n the study.There are a total of 297firm-city airs in which the firm s an incum-bent in the city in 1985.Because of the small, integer number ofstores added or subtractedby a chain in alocal market, one should not ignore thediscreteness of the data when analyzingthese decisions; n 36 percentof the obser-vations for incumbentfirms the incumbentadds or subtractsno more than one store.Thus, I adopt an ordered-probitmethodol-ogy, estimating whether each large super-marketchain adds stores in a market,nei-ther adds nor subtractsstores in a market,or subtractsstores in a market.18Becausethe determinantsof these decisions maybevery different for LBO and non-LBO in-cumbent firms, the specificationsare esti-mated separately or the two sets of firms.I measure the relationship between afirm'sdecisionto add or subtract tores andthe share of rival stores in the market in1985 owned by firms that would eventuallyundertakeLBO's.I controlfor demographicchanges in the market and control for thepossibilitythat the market was under- orover-stored n 1985.These controlvariablesare described n Table 4.

170ne set of firms s left out of the sampleof topfirms.Thesefirmsare those involvedn the onlymajor

antitrust hallenge o a supermarketmergerduring heperiod. After the federal antitrustsupervisory odiesdecided not to challenge he purchaseof LuckyStoresby AmericanStores,the CaliforniaAttorneyGeneral'sOffice decided to pursue a challenge of the mergerunder the Californiaantitruststatutes. The case wastiedup in the courts orover a year, duringwhichtimeAmericanStoreswas not allowed o mergethe opera-tions of the two firmsandwas restricted romopeningandclosingnewstores n California.Theparties o thismerger are left out of the specificationshere, thoughresults ncluding hemwere checkedand are extremelysimilar.18The ordered-probitmethodologywas used in thecontext of measuring he determinantsof how manyfirmscompetein a city in TimothyF. BresnahanandPeter C. Reiss(1987,1990).



14/22


15/22

428 THEAMERICANECONOMICREVIEW JUNE 1995percent level. The "marginal effects" showthat, in a city in which all market character-istics are held at their mean, adding theLBO of a firm with a 10-percent marketshare would increase the probability that agiven non-LBO firm will add stores in themarket by approximately 6.5 percent. Thisresult supports the results in Section I whichsuggest that LBO's lead to a decrease in thetoughness of competition in the market;when a firm undertakes an LBO, rival non-LBO firms in the market find expansionattractive.Part B of Table 6 repeats the specifica-tion of Part A, except the expansion deci-sions of LBO firms are used as the depen-dent variables. The coefficients for all of thedemographic variables are statistically in-significant, and many have the opposite signsfrom the previous specifications. In contrastto the specification for non-LBO stores, thecoefficient for total stores is positive, andthe coefficient for the firm's own marketshare is negative.The coefficient for the share of firm i'srivals in market j which are LBO firms ispositive, as in the previous specification, butsignificant only at the 12-percent level. Thisresult provides some evidence for the hy-pothesis that LBO's decrease the "tough-ness" of product-market competition, al-though the results are clearly weaker thanfor non-LBO firms.As mentioned before, one reason for ex-amining expansion by incumbent firms sepa-rately from new entry into local markets isthat one would expect that incumbent firmswould be able to begin to respond relativelyquickly to local market conditions. Thus,one would expect that a firm adding storesin response to an LBO would have added atleast one store by 1991, since the last LBOtook place in early 1990. In Table 7, I divideLBO's into those that took place prior to1988 (early LBO's) and those that took placeduring or after 1988 (late LBO's) and re-peat the specifications of Tables 6. In bothcolumns, the coefficient for the early-LBOshare is only slightly larger than the coeffi-cient for the late-LBO share. The hypothe-sis that the coefficient for early LBO's is

larger than the coefficient for late LBO's isnot rejected at standard significance levels.This is consistent with the view that expan-sion by rival incumbents in response toLBO's should begin quickly.Several tests were undertaken to test therobustness of the results. First, becausethese specifications use firm-level data, theconcern arises that unmodeled firm hetero-geneity may affect the basic results. Theseresults were reestimated including firmdummy variables. Because of the number offirm dummy variables relative to the num-ber of observations, it was necessary to paredown the specification in order to estimatethis relationship. The basic specification wasthus reestimated pooling data from LBOand non-LBO firms and including firmdummy variables. The coefficient for theLBO share of the market remains posi-tive and is statistically significant at the6-percent level.Unfortunately, the results could not bereestimated with city or even state dummyvariables because there are not enough ob-servations per geographic area. However,the results for both LBO and non-LBOfirms are robust to the inclusion of dummyvariables for the Midwest, West, South, andEast Coast. The coefficient for share LBOin the non-LBO firm regressions remainspositive and significant at the 3-percentlevel; the coefficient for share LBO in theLBO firm regressions remains positive andsignificant at the 6-percent level.Cross-firm within-city correlation of theerror term might lead to the estimation ofinflated significance levels. The regressionresults were checked for robustness to thispossibility. Table 7 was reestimated usingone observation per MSA. The dependentvariable took the value of 1 if the supermar-ket chains in the sample in the MSA addedstores on net; it took the value of 0 if theyneither added nor subtracted stores, and ittook the value of - 1 if they subtractedstores on net. The right-hand-side variablestook their mean value for the MSA. Thespecification was done separately for LBOfirms and non-LBO firms, as before. In thespecification for non-LBO firms, there were



16/22

VOL. 85 NO. 3 CHEVALIER: CAPITALSTRUCTUREAND MARKET COMPETITION 429TABLE 7-MAXIMUM-LIKELIHOOD ESTIMATION RESULTS FOR INCUMBENT FIRMS

A. Non-LBO ncumbents B. LBO incumbentsMarginal effects Marginal effects

dPr[y=-1] dPr[y=1] dPr[y=-1] dPr[y=1]Variables Coefficient dx dx Coefficient dx dxChange in households 0.0330a -0.0131 0.0126 -0.0283 0.0110 -0.0057(0.0202) (0.0265)Change in income -2.6630 1.0553 -1.0152 4.2400 -1.6511 0.8565(2.0190) (2.6840)Change in income squared 0.2550 -0.1010 0.0972 -0.3950 0.1538 -0.0798(0.1850) (0.2430)Change in share with income - 0.2125 0.0842 -0.0810 0.2358 - 0.0918 0.0476less than $10,000 (0.1438) (0.1793)Change in households per -0.0008 0.0003 - 0.0003 0.0008 - 0.0003 0.0002square mile (0.0006) (0.0064)Deviation from mean stores 0.0012 -0.0005 0.0004 0.0032 - 0.0013 0.0007per household (0.0038) (0.0046)Total stores -0.0012** 0.0005 - 0.0004 0.0006 - 0.0002 0.0001(0.0004) (0.0006)Market share 2.1043* -0.8339 0.8022 -2.7507a 1.0712 -0.5557(1.0503) (1.5434)Herfindahl index 2.0584 -0.8157 0.7847 - 0.2282 0.0889 - 0.0461(1.8184) (3.6914)Share early LBO 1.7032a -0.6749 0.6493 1.9126 - 0.7448 0.3864(0.9073) (1.1192)Share late LBO 1.4087 -0.5582 0.5370 1.13798 -0.5373 0.2787(0.8865) (1.3221)Exit threshold 0.0996 0.9035(0.6693) (0.9542)Entry threshold 0.4173 1.3866(0.6697) (0.9569)Number of observations 184 113Notes: The dependent variable has the following values: Yij=-1 if firm i withdraws at least one store from market j,Yij= 0 if firm i neither adds nor withdraws stores from market j, and Yij = + 1 if firm i adds stores in market j. Standarderrors are reported in parentheses.aSignificantly different from zero at the 10-percent level.*Significantly different from zero at the 5-percent level.**Significantly different from zero at the 1-percent level.

79 observations, and the coefficient forthe average share of LBO rivals remainedpositive and was statistically significant atthe 5-percent level. In the specification forLBO firms, the coefficient for the averageshare of LBO rivals remained positive butwas statistically significant at only the 25-percent confidence level. The results forthis LBO incumbent specification should betreated with extreme caution, however, asonly 28 observations were available.Finally, the importance of the use of theordered-probit specification was investi-gated. Table 8 reestimates Table 7 using theactual change in the number of stores foreach incumbent firm as the dependent vari-

able. Because the left-hand-side variablevaries so much in scale, heteroscedasticity-robust standard errors are used, followingthe method of Halbert White (1980). Thecoefficient for the share of early-LBO firmsremains positive and statistically significantat the 6-percent level in the regressionfor non-LBO incumbents. The coefficientsfor the share of late-LBO firms is ap-proximately zero. This is not surprising.The ordered-probit methodology examineswhether any response to the LBO has oc-curred, while this specification measures themagnitude of the response. It is not surpris-ing that, by 1991, very few stores have beenbuilt "responding" to the later LBO's. The



17/22

430 THE AMERICANECONOMIC REVIEW JUNE 1995TABLE 8-OLS ESTIMATION RESULTS

FOR INCUMBENT FIRMSCoefficients

(1) (2)Non-LBO LBOVariables incumbents ncumbentsChange in households - 0.0660 - 0.2690(0.1170) (0.2350)Change in income 0.2860 14.0750(6.2420) (14.1650)Change in income squared 0.0900 - 1.3100(0.5750) (1.2500)Change in share with income - 0.4167 0.5441less than $10,000 (0.4084) (0.9351)Change in households per - 0.0033 0.0003square mile (0.0039) (0.0352)Deviation from mean stores 0.0575* 0.0268

per household (0.0250) (0.0435)Total stores - 0.0055* - 0.0005(0.0022) (0.0042)Market share 7.0556 - 28.1176**(5.4637) (10.2838)Herfindahl index 29.6984 30.6566(11.9950) (27.0131)Share, early LBO 7.9665a 0.7648(4.1466) (5.7235)Share, late LBO - 1.0105 - 5.8049(3.9945) (9.0537)Constant - 6.4550* - 3.2238(3.1960) (6.7199)R2 0.12 0.14Number of observations 184 113Notes: The dependent variable is the number of storesthat firm i has added to market j. It is negative if thefirm has subtracted stores from market j. White (1980)robust standard errors are in parentheses.aSignificantly different from zero at the 10-percentlevel.*Significantly different from zero at the 5-percentlevel.**Significantly different from zero at the 1-percentlevel.

results in column 2 of Table 8 offer nosupportfor the hypothesisthat LBO firmsrespondto the LBO'sof their rivals.The results n this subsectionsuggestthatnon-LBOfirmsfind expansionattractive nmarketsdominated by LBO firms. As ex-pected, the results in this subsection arestronger than those which pool the storeadditionsof incumbentfirmsand new en-trants.This is consistentwith the expecta-tion that expansion or contractionby in-

cumbentfirmswould be the first detectableresponseto a change n conditions n a localmarket.C. Entry

In this subsection, I examine de novoentry by a large supermarket hain into alocal market.19 extended the data set to1993 by searching the Nexis data base forannouncementsof new entry. This is donefor the entry specificationsand not for theexpansion specificationsabove for two rea-sons. First, because one expectsentryto lagchanges in market conditions more thanexpansion, t is more important o have asrecent data as possible for new entry. Sec-ond, the supermarket rade press does not,in general, announce that a supermarketchain is opening new stores in a city inwhich it alreadyhas stores, though it doesreport that a large supermarketchain isenteringa new local market.The specification or the entrymodel incolumn1 of Table9 is a simpleprobit.Thedependentvariabletakes the value of 0 ifno entry occurs in the city between 1985and 1993. The dependentvariable akes thevalue of 1 if entry occurs in that period.Entry is defined to occur when a largesupermarket hain of more than 25 storesopens at least one storein an MSA in whichit was not an incumbent n 1985. Also in-cluded as entry is the opening of a hyper-marketby KMartor WalMart.20The variablesincluded in the entry re-gressionsare a subset of those included in

19This s not the first paperto studyde novoentryinto localmarkets n the supermarketndustry.RonaldW. Cotterilland LawrenceE. Haller (1992) studydenovo entry into cities by supermarket hains.They donot considerthe leveragecharacteristicsf incumbentfirms.They considerentryover a different ime periodusinga differentdataset and methodology.20A hypermarkets a full supermarket ombinedwith a generalmerchandise tore.WhileneitherKMartnor WalMartowned 25 supermarkets t the time thatthey opened hypermarkets,hey are obviously argeretailing hains and were thusincluded.



18/22

VOL. 85 NO. 3 CHEVALIER:CAPITALSTRUCTUREAND MVARKETOMPETITION 431TABLE 9-RESULTS FOR NEW ENTRY

(1) (2) (3)Marginal effectsdPr[y =1]Variable Coefficient dx Coefficient

Change in households 0.0841* 0.0126 2.23 x 10-6*(0.0395) (1.13 x 10-6)Change in income 4.6880 0.7038 1.15 x 10-4(2.9970) (7.0x 10-5)Change in income squared - 0.4950a -0.0743 1.22x 10-9a(0.2910) (6.75 x 10- 10)Change in share with incomeless than $10,000 0.2859 0.0429 0.0702(0.1806) (0.0388)Change in households persquare mile - 0.0025 - 0.0004 4.19 X 10-4(0.0032) (3.97x 10-4)Deviation mean stores perhousehold 0.0012 0.0002 6.95 x i04(0.0064) (0.0021)Herfindahl index 0.0887 0.0133 0.0147(2.9218) (1.0290)Share, early LBO's 2.4183a 0.3630 0.7904a(1.1330) (0.3835)Share, late LBO's 0.6756 0.2960(1.2177) (0.4317)Constant -1.518 0.0758(1.0364) (0.3248)R2 0.2010Notes: The first column shows probit results. The dependent variable has the followingvalues: Yj= 0 if no entry occurs in market j; Yj= 1 if entry occurs. The second columnshows the marginal effects implied by the coefficients in column 1. The third columnshows the results of the linear probability specification. Entry occurs in 39 of the 85markets. Standard errors are reported in parentheses.aSignificantly different from zero at the 10-percent level.*Significantly different from zero at the 5-percent level.

the expansion regressions above. Since theseregressions examine entry at the marketlevel rather than the firm level, firm-specificcharacteristics must be excluded.Table 9 shows that the coefficient for theearly-LBO share is of a larger magnitudethan the coefficient for the share of late-LBO firms. However, the hypothesis thatthe coefficient for the share of early LBO'sis different from the coefficient for the shareof late LBO's is rejected at only the 24-percent significance level. As expected, theshare of early-LBO firms in the local mar-ket has a positive coefficient, statisticallysignificant at the 3-percent level. The coef-ficient for the share of late-LBO firms ispositive, but statistically significant at onlythe 58-percent level. This table thus sug-

gests that large supermarket chains find en-try into local markets dominated by firmsthat undertook LBO's prior to 1988 attrac-tive. There is only very limited evidence,however, that entry has responded to LBO'sthat took place during or after 1988.21 Theresults of a linear probability model, shownin column 3 of Table 9 are substantially thesame as the probit results.

21Once again, the results are robust to the inclusionof regional dummy variables. The coefficient for theshare of early LBO's is positive and statistically signif-icant at the 10-percent level; the coefficient for theshare of late LBO's is positive and significant at the42-percent level.



19/22

432 THEAMERICAN ECONOMIC REVIEW JUNE 1995TABLE 10-COMPARISON OF 1985ACCOUNTING VALUES FOR FIRMS THAT WOULD

EVENTUALLY UNDERTAKE AN LBOAND FIRMS THAT WOULD NOTEVENTUALLY UNDERTAKE AN LBO

MeanNon-LBO t statistic ofAccounting ratio LBO firm firm differenceOperating income/sales 0.0363 0.0395 0.48Net income/sales 0.0040 0.0043 0.14Market value/book value of assets 0.8316 0.8194 0.10Capital expenditures/assets 0.1461 0.1300 0.80Retained earnings/net income 0.3703 0.3266 0.76Dividends/net income 0.2375 0.1896 0.39

VI. An AlternativeHypothesisThe results of SectionV suggestthat su-permarketfirms find entry and expansionattractivein markets dominated by firmsthat undertookLBO's.I have suggested hatentry and expansion are attractivebecausecompetitionbecomesless "tough" ollowingan LBO. In this section, I brieflyexplorethealternativehypothesis that LBO's did notchange the toughness of product-marketcompetition,but rather, that firmsthat un-dertook LBO's were weak firms. If weakfirms undertookLBO's,then one mightex-pect to see entry and expansionoccur inthose marketsdominatedby LBOfirms.

A. Asset SalesEvenif LBOfirmswere underperformerson averagepriorto their LBO's,this wouldnot necessarily affect the results of Sec-

tion V if LBO firmssold off all underper-formingdivisions o non-LBO irms n post-LBOasset sales.As discussed n Section IV,the methodologyused in Section V assignsassets sold after LBO's to their eventualowners. The sale of the division is notcountedas a loss of storesfor the LBOfirm,nor is it counted as entryor expansionforthe purchaser.More importantly, he pres-ence in a city of supermarkets hat weresold in a post-LBOasset sale is not countedin calculating he store share of LBO firmsin the city.The asset-assignment rocedure

helps to diminish he effect of these under-performingdivisionson the results.B. Accounting Evidence

If the LBO event selects for firms thatwere underperformers,hen one might ex-pect accountingdata to reflect that LBOfirms were underperformerson averageprior to their LBO's.Table 10 contains ac-counting data for all of the supermarketchains that were publiclytraded in 1985.22Eleven of these firms undertook an LBOafter 1985,and 20 did not.The table shows that, on average, LBOand non-LBO firms do not differ signifi-cantly.In particular,LBOfirmsdo not gen-erate significantlyess operating ncome as ashareof sales or less net income as a shareof sales. Furthermore, he ratio of marketto book value of assets, a proxy for themarket's estimation of a firm's futureprospects, is slightly higher for LBO firms.The ratio of capital expenditures o assetsand the ratio of retained earningsto netincomearesomewhathigherfor LBO firms,although the difference is not statisticallysignificant.2322Thesedata are from Compustat.23FirmswhichundertookLBO's early do not differsignificantly rom firms which undertookthem later,except that firms which undertookLBO's early havehigher ratiosof capital expenditures o assets thando

firms hatundertookLBO's ater.



20/22


21/22

434 THEAMERICANECONOMICREVIEW JUNE 1995competitionbecomesless vigorous ollowingan LBO from the hypothesis that qualitycompetitionbecomeslessvigorous ollowingan LBO.For example, t has been suggestedin the tradeliterature hat LBO firmscom-pete less fiercely n the area of storequality,placing their stores on slower renovationand repair schedules.It has also been sug-gested that LBO firms refuse to becomeembroiled in price wars with rivals. Thisstudycannot identifywhich,if any,of thesemechanisms brings about the change inproduct-market ompetition. An examina-tion of price competitionby leveragedandunleveraged upermarket hains is the sub-ject of future research.Finally,whilethis paper does suggestthatthe nature of competition changes whenfirm leverage changes, the results do notnecessarilymake any contributionto thedebateconcerningwhetheror not leveragedbuyouts were value-maximizing.Clearly,LBO firmswouldrathertheir rivalsdid notexpandandentertheir markets.However, tis impossibleto knowwhetherthe amountthat firmsdeterredthis entry and expansionprior to LBO's was efficientor inefficient.Thus, while these results cannot prescribean optimal capitalstructurebased on prod-uct-market utcomes, he resultsmake clearthat product-market effects of capital-marketdecisionsmustbe considereda com-ponent of the choice of optimal capitalstructure.

REFERENCESBolton, Patrick and Scharfstein,David S. "A

Theory of Predation Based on AgencyProblems in Financial Contracting."American Economic Review, March 1990,80(1), pp. 93-106.Brander, James A. and Lewis, Tracy R."Oligopoly and Financial Structure."American Economic Review, December1986, 76(5), pp. 956-70.Bresnahan,TimothyF. and Reiss, PeterC. "DoEntry Conditions Vary Across Markets?"Brookings Papers on Economic Activity,1987, (3), pp. 833-71."Entry in Monopoly Markets." Re-

view of Economic Studies, October 1990,57(4), pp. 531-53.Cotterill, Ronald W. and Haller, LawrenceE."Barrier and Queue Effects: A Studyof Leading U.S. Supermarket ChainEntry Patterns." Journal of IndustrialEconomics, December 1992, 40(4),pp. 427-40.Eckbo,B. Espen."Horizontal Mergers, Collu-sion, and Stockholder Wealth." Journalof Financial Economics, April 1983,11(1-4), pp. 241-73.Fama,EugeneF. Foundations of finance. NewYork: Basic Books, 1976.Fudenberg,Drewand Tirole, Jean. "A 'Signal-Jamming' Theory of Predation." RandJournal of Economics, Autumn 1986,17(3), pp. 366-76.Guedes, Jose C. and Opler, Tim C. "TheStrategic Value of Leverage: An Ex-ploratory Study." Mimeo, Cox School ofBusiness, Southern Methodist University,1992.Jensen,MichaelC. "The Eclipse of the PublicCorporation." Harvard Business Review,September-October 1989, 67(5), pp.61-74.

Kaufman, Phillip R. and Handy, Charles R.Supermarketprices and price differences:City, firm, and store-level determinants,Economic Research Service PublicationTB-1776. Washington, DC: U.S. Depart-ment of Agriculture, December 1989.Magowan,Peter. "The Case for LBOs: TheSafeway Experience." California Manage-ment Review, Fall 1989, 32(1), pp. 9-18.Maksimovic,Vojislav. "Capital Structure inRepeated Oligopolies." Rand Journal ofEconomics, Autumn 1988, 19(3), pp.389-407.McAfee,R. Preston and Williams, Michael A."Can Event Studies Detect Anticompeti-tive Mergers?" Economics Letters, 1988,28(2), pp. 199-203.Mergersand Aquisitions. Philadelphia, PA:MLR, bimonthly, 1985-1992.Phillips, Gordon M. "Financial Slack, Refi-nancing Decisions, and Firm Interaction."Working paper, Krannert School of Man-agement, Purdue University, 1991.

_ . "Increased Debt and Product Mar-ket Competition: An Empirical Analysis."



22/22

VOL. 85 NO. 3 CHEVALIER:CAPITAL STRUCTUREAND MARKETCOMPETITION 435Working paper, Krannert School of Man-agement, Purdue University, 1992.PredicastsFunk and Scott index, UnitedStates.Cleveland, OH: Predicasts, annual, 1985-1991.ProgressiveGrocer."Annual Report of theSupermarket Industry." April 1989, 68(4).. Market scope. White Plains, NY:MacLean Hunter Media, 1986, 1992.Retail tenants directory.Clearwater, FL: Na-tional Mall Monitor, 1988.Rotemberg, Julio J. and Scharfstein, David S."Shareholder-Value Maximization andProduct Market Competition." Review ofFinancial Studies, 1990, 3(3), pp. 367-91.

Spence, A. Michael. "Capital Structure andthe Corporation's Product Market Envi-ronment," in Benjamin Friedman, ed.,Corporate capital structures in the UnitedStates. Chicago: University of ChicagoPress, 1985, pp. 353-82.Supermarket ews. Distribution study of gro-

cery store sales. New York: Fairchild, an-nual, 1987-1990.Stillman, Robert S. "Antitrust HorizontalMerger Policy." Journal of Financial Eco-nomics, April 1983, 11(1-4), pp. 225-40.Sutton, John. Sunk costs and market struc-ture:Price competition,advertising,and theevolution of concentration. Cambridge,MA: MIT Press, 1991.WaUl treetJournal Index. New York: DowJones, 1985-1991.Whinston, Michael D. and Collins, Scott C."Entry and Competitive Structure inDeregulated Airline Markets: An EventStudy Analysis of People Express." RandJournal of Economics, Winter 1992, 23(4),pp. 445-62.White, Halbert. "A Heteroskedasticity-Con-sistent Covariance Matrix Estimator anda Direct Test for Heteroskedasticity."Econometrica, May 1980, 48(4), pp.817-38.

Date post:	03-Apr-2018
Category:	Documents
Upload:	prasad-hegde
View:	231 times
Download:	0 times

Chevalier AER 1995

Documents