Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit...

download Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

of 46

Transcript of Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit...

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    1/46

    Review of Industrial Organization12:9 5 - 1 3 9 , 1 9 9 7 . 95 1997Kluwer Academ ic Publishers. Printed in the Netherlands.

    The Effects of Megamergers on Eflficiency andPrices: Evidence from a Bank Profit FunctionJALALD.AKHAVEIN*Department of Economics,N ewYork University, N e w York,N Y10012a n dWharton FinancialInstitutions Center, University of Pennsylvania. Philadelphia, PA 19104, U.S.A.ALLEN N. BERGER*Board of Governors of the Federal Reserve System, Washington, DC 20551 a n dWharton FinancialInstitutions Center, University of Pennsylvania, Philadelphia, PA 19104, U.S.A.DAVID B. HUMPHREYF. W.Smith Eminent Scholar in Banking, Departm ent of Finance, Florida State University,Tal lahassee, FL 32306, U.S.A.,

    Abs t ract This paper examines the efficiency andprice effects ofmergers byapplyinga frontierprofit fiinction to data on bank 'megamergers'. We find that merged banks experienceastatisticallysignificant 16percentage p oint average increase in profit efficiency rank relative to other large banks.Most of the improvement isfrom increasing revenues, includingashift inoutputs from securitiestoloans, a higher-valued product. Improvements were greatest for the banks with the lowest efficienciesprior to merging, who therefore had the greatest capacity for improvement By comparison, the effectson profits from merger-related changesinprices were found to be very small.Keywords: Bank, merger, efficiency, profit, price, antitrust.JEL Classification Codes:L l1,M l ,L89,G 2 1 ,G 28

    I. IntroductioiiThe recent waves of large mergers and acquisitions in both manufacturing and ser-vice industries in the United States raise important questions concerning the publicpolicy tradeoff between possible gains in operating efficiency versus possible socialefficiency losses from a greater exercise of market power. If any improvements inoperating eflBciency from these mergers are large relative to any adverse effects ofprice changes created by increases in market power, then such mergers may be inthe public interest. For an informed antitrust policy, it is also important to know if

    ' Theviews expressed do not necessarily reflect thoseof the BoardofGovernorsoritsstaflF.The authors thank Anders Christensen for very useful discussant's commenits, Bob DeYoung, TimHannan, Steve Pilloff, Steve Rhoades, and the participants in the Nordic Banking Research Seminarfor helpful suggestions, and Joe Scalise and Seth Bonime for outstanding rcsearch assistance.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    2/46

    96 JALALD.AKHAVEINETAL.there are identifiable ex ante conditions that are good pred ictors of either efficiencyimprovem ents or increases in the use of market power in setting prices. W hetheror not these mergers are socially beneficial on av erage, there may be identifiablecircumstances that may help guide the policy decisions about individual mergers.Current antitrust policy relies heavily on the use of the ex ante Herfindahl indexof concentration for predicting market power problems and considers operatingefficiency only under limited circum stances.'

    The answers to these policy questions largely depend upon the source ofincreased operating profits (if any)firomconsolidation. Mergers and acquisitionscould raise profits in any of three major w ays. First, they could imp rovecost effi-c iency , reducing costs per imit of output for a given set of output quantities andinput p rices. Indeed, con sultants and managers have often justified large mergerson the basis of ex pected cost efficiency gains.Seco nd, mergers may increase profits through imp rovements\n profit efficien cythat involve superior combinations of inputs and outputs. Profit efficiencyisa moreinclusive concept than cost efficiency, because it takes into account the cost andrevenue effects of the choice of the output vector, w hich is taken as given in theme asure men t of cost efficiency. Th us, a merger could imp rove profit efficiency

    w ithout improving cost efficiency if the reconfiguration of outp uts associated w iththe merger increases revenues more than it increases costs, or if it reduces costsmore than it reduces revenues. We argue below that analysis of profit efficiency ismore ap prop riate for the evaluation of mergers than cost efficiency be cause o utputstypicallyd ochange substantially subsequent to a m erger.Third, mergers may improve profits through the exereise of additional mar ke tpo werin setting prices. An increase in market concentration or market share mayallow the consolidated firm to charge higher rates for the goods or services itproduces, raising profits by extracting more surplus from consumers, without anyimprovement in efficiency.These policy issues are of particular imp ortance in the banking industry beca userecent regulatory changes have made possible many mergers among very largebanks.The 198 0s w itnessed the beginning ofatrend toward 'm egamergers' in the

    U . S . banking industry, mergers and acquisitions in which both banking organi-zations have more than $1 billion in assets. This trend - w hich was p recipitatedby the removal of many intrastate and interstate geographic restrictions on bankbranching and holding comp any affiliation - has continued into the 1 990s. At theoutset of the 19 80s, only 2.1 % of bank assets were controlled by out-of-state ban k-ing organizations. HalfW ay through the 1 990s, 27 .9% of assets w ere controlled byout-of-state bank holding companies, primarily through regional compacts amongnearby states.^ The R iegle-N eal Interstate Banking and Branching Efficiency A ctof 1994 is likely to accelerate these trends, since it allows bank holding compa-' Se eU .S. Department of Justice and Federal Trade Cotn mission (1992).^ SeeBerger .Kashyap, and Scal ise (1995).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    3/46

    T H E E F F E C T S O F M E G A M E R G E R S O N E F F IC I E N C Y A N D P R I C E S - 9 7nies to acquire banks in any other state as of September 29, 1995, and will allowinterstate branch ing in almost every state by June1,1997.There are other reasons why banking provides such an interesting academic andpolicy experiment for mergers. First, competition in banking has been restrictedfor a long time by geograp hic and other restrictions, so past inefSciencies might beexpected to persist. The market for corporate control in banking has also been quitelimited, since nonbanks are p rohibited from taking overbanks,and the geographicbarriers to competition have also reduced the potential for takeovers by moreefficient banks. T hese restrictions on comp etition both in the p roduct ma rkets andin the market for corporate control may have p rotected inefficient manag ers. Bothtypes of restrictions are now b eing lifted.

    Seco nd, the ba nking industry has relatively c lean, detailed data available fromregulatory reports that give information on relatively homogenous products in dif-ferent local markets w ith various market structures and econom ic conditions. Thismakes for an almost ideal controlled envirotiment in which t otest various industrialorganization theories. As a result, banking is one of the most heavily researchedindustries in industrial organization, yielding a relatively strong background liter-ature upon which to build.Unfortunately, the academic literature ha s made little progress in determiningthe sources of profitability gains^if any, associated w ith bank me rgers. Of the threemain sources of potential profitability gains from mergers, the literature has focusedprimarily on cost efficiency improvements. As discussed below, the empiricalevidence suggests that mergers have had very little effect on cost efficiency onaverage. Moreover, there has also been little progress in divining any ex anteconditions that accurately predict the chang es in cost efficiency that do occur forpossible use in antitrust policy.

    Despite the advantages of the profit efficiency concept over cost efficiency, weare not aw are of any previous studies in banking or any o ther industry of the profitefficiency effects of mergers. Although many studies have examined changes insome profitability rat ios pursuant to mergers, such studies cannot determine theextent to which any increase in profitability is due to an improvement in profitefficiency (w hich is a change in quantities for given prices) versus an increase inmarket p ow er (a change in price for a given efficiency level).Similarly, there are very few academic studies of which we are aware thatdetermine the changesi n prices associated w ith bank mergers. Price changes wouldreveal the effects of increases in market power plus any price effects that mayresult from changes in operating efficiency. The lack of analysis of the marketpower effects of bank mergers is perhaps surprising given that a major thrust ofcurrent antitrust enforcement is to prevent mergers w hich are exp ected to result inprices less favorable to consumers (higher loanrates,lower dep osit rates) or requiredivestitures that may accom plish this goal.

    The purpose of this paper is to add some of t h e missing infonnation about theprofit efficiency and market power effects of mergers. We analyze data on bank

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    4/46

    98 JALALD. AKHAVEINE T AL .megamergers of the 1980s, using the same data set as employed in an earliercost efficiency analysis (Berger and Hum phrey 199 2). In this way, all three of thepo tential source s of increased ope rating profits from mergers cost efficiency, profitefficiency, and market power in setting prices can be evaluated and comparedusingthe same dataset. In addition,we test several hypotheses regarding the ex anteconditions that may help p redict w hich mergers are likely to increase efficiency orpromote the exercise of market power.

    By w ay of an ticipation, the findings suggest that there are statistically significantincreases in profit efficiency associated with U.S. bank megamergers on average,although there do not appear to be significant cost efficiency improvements onaverage. The improvement in average profit efficiency in part refiects a productmix shift from securities to loans, increasing the value of output. The data areconsistent with the hyp othesis that megamergers tend to diversify the portfolio andreduce risk, wh ich allows the consolidated bank to issue more loans for about thesame am ount of equ ity capital, raising p rofits on average. The p rofit efficiencyimprovem ents can be fairly w ell predicted- they tend to occur w hen either or bothof the merging firms are inefficient relative to the industry prior to the merger.The changes in market power associated with megamergers as refected in

    changes in prices subsequent to the mergers - are foimd to be very small onaverage and not statistically significant, although they are predictable to somedegree. These results are consistent with the hypothesis that antitrust policy hasbeen fairly successful in preventing m ergers that would have brought about largeincreases in market pow er. H owever, it is not known w hether this policy may ha vealso prevented some mergers that might have increased efficiency substantially.Section n summarizes prior emp irical studies of merger efficiency and marketpow er, showing how our ap proach differs from p ast efforts. SectionEHpresents thefrontier profit function model used to measure profit efficiency and describes thedata set. Section IV gives the estimated profit efficiency effects of mergers and aregression a nalysis of some ex ante factors that may pre dict these efficiency effects.Section V gives a similar analysis ofthe changes in market power as reflected inthe price changes associated with the mergers. Section VI concludes.

    II. The Merger Literature Versus Our ApproachMergers and Cos t Ef f ic iency . Mergers can potentially improve cost efficiency byincreasing scale efficiency, scope (product mix) efficiency, or X-efficiency (man-agerial efficiency). The findings in the banking literatme'^^gest that scale andscope efficiency changes are unlikely to change unit costs by m ore than a few per-cent forlargebanks w hichwe studyhere.A ny meaningful cost scale economies thatare found typically ap ply only to relatively small banks. The p otential is greater forcost X-efficiency gains by moving closer to the 'best-p ractice' cost frontier whe recost is minimized for a given output bundle. The X-efficiency empirical findingssuggest that on average, banks have costs that are about 20% to 25% above those

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    5/46

    T H E E F F E C T S O F M E G A M E R G E R S O N E F F IC I E N C Y A N D P R I C E S 99of the observed best-practice banks. This result suggests that cost efl&ciency couldbe considerably imp roved by a m erger in which a relatively efficient bank acquiresa relatively inefficient bank and spreads its superior management talent over moreresources.^The em pirical bank merger literature confirmsthis potent ia l for cost efficiencyimp rovem ent from mergers. * How ever, this literature also sugge sts that the poten tialfor cost efficiency improvement generally was n o t realized. Most merger studiescompared simple cost ratios, such as the operating cost to total assets ratio, andtypically found no substantial change in cost perfonnance associated with bankmergers (e.g., Rhoades, 1986, 1990; Srinivasin, 1992; Srinivasin and Wall, 1992;Linder and Crane, 1992;Pilloff, 1996). There are methodological problems withusing simple cost ratios to meastire cost efficiency, including the fact that suchratios do not control for differences in input prices and outp ut mix.^ N evertheless,the results of these ratio studies are consistent with the small number of studiesthat calculated the efficiency effects of mergers by meastiring the distance iromthe best-practice cost frontier and found little or no improvement on average incost efficiency (Berger and H umphrey, 1 992 ; R hoades, 1 99 3; Peristiani, 1995 ;D e Young, 19 96). For example, Beiger and Hump hrey (1992 ) found about a 5percentage point average improvement in cost X-efficiency rank relative to peergroup, but the improvement was not statistically significant.^

    These acade mic findings seem to cotiflict with co nsultant studies which forecastconsiderable cost savings from large bank mergers - as much as 30 % of theoperating ex penses of the acquired bank. H owever, as disctissed in detail in Bergerand Humphrey (1992), the acadetnic and consultant results do not necessarilydisagree substantively. Rather, the academics and consultants tend to state theirfindings differently or use different denominators that may make their resultsapp ear inconsistent when they are actually fairly consistent with each o ther. 'A ll of the cost efficiency analyses share the problem that outp uts are taken asgiven and the revenue eflFects of mergers are not considered. As noted above, thetotal output of t h e consolidated firm typically changes after a merger and there is

    ' See the survey by Berger, Hunter,a n d Timme (1993) for summaries of the cost scale, scope , andX-efficiency l iteratures.* Savage (1991) and Shaffer (1993) showed by simulation methods that the potential for scaleefficiency gains from mergers between large banks is negligible, but that large X-efficiency gainsare possible. Similarly, using actual merger data, Berger and Humphrey (1992) found that acquiringbanks were substantially more cost X-efficient than the banks they acquired on average. This resultconfirms the potential for cost X-efficiency gains if the managers of the acquiring bank are able torun the consolidated bank after the merger as efficiently as they ran the acquiring bank before themerger.

    ' See Berger and Humphrey (1992) for more discussion of these problems.' See Rhoades (1994) for a survey of the cost and perfonnance merger studies from 1980 to 1993.' For example, since the average acquired bank represents about 30% of the consolidate d bank,and since operating costs currently are about4 5 of total expenses, a savings of 3 0 of the acquiredbank's operating cos ts a s claimedb y consultants translatesintoonly about4 o ft h e total consolidatedexpenses [(30% 45%)0.30], close to the results of academic studies.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    6/46

    100 J A L A L D . A K H A V E I N E T A L .no w ay to determine from cost analysis alone whether the cost changes are greaterthan or less than the revenue changes. Th us, a determination that cost efficiencyimproved or worseneddoesnot by itself necessarily imply that the firm h as becom emore or less efficient overall, or become m ore or less profitable. As w ill be show n,profit efficiency solves this problem.Me r ge r s and Re v e nue and Pr o f i tEf f ic iency . Mergers might also improve revenueor profit efficiency by improving revenue or profit scale, scope, or X-efficiency,but the literature here is much more limited and therefore less definitive than forcost efficiency. Revenue X-inefficiency is the failure to produce the highest valueof output for a given set of input quantities and output prices. A firm may berevenue X-inefficient because it produces too few outputs for the given inputs, or isinside its production- possibilities frontier (analogous to the cost X-inefficiency ofa firm that uses too many inputs to produce the givenoutputs).Alternatively, a firmmay b e revenue X-inefficient if it responds poorly to reiadve prices and produ cestoo little of a high-priced output and too much of a low-priced output, even if itis on the p roduction-possibilities frontier (analogous to the cost inefficiency of atechnically efficient firm that employs too much of a relatively high priced input).Thus,revenue X-inefficiencies are fully analo gous to cost X-inefficiencies, as bothinvolve a net loss of real value, but just differ as to whether the loss is in termsof a lower value of output produced or a higher value of inputs consumed.* If theassumption of exogenously determined prices is dropped and allowance is madefor market power in price setting, revenue scale and scope economies can alsoocc ur.' Thus, revenue efficiencies app ear to offer the sam e type of opp ortunity forimprovement from m ergers as cost efficiency, but there has been n o investigationof w hether this potential has been realized in actual mergers.

    Profit efficiencies incorporate both cost and revenue efficiencies and their inter-actio ns, but have received little academ ic attention. Profit efficiency stu dies of U . S .banks foimd that estitnated inefficiencies were usually quite large, about one-thirdto two-thirds of potential profits may be lost due to inefficiency. In addition, it wasfoimd that most inefficiencies were due to deficient output revenues rather than' Revenue X-inefficiency is not usually directly measured, but can be inferred from analysis ofan output distance function, whicb is an altemative way to measure output inefficiencies. An outputdistance function applied to banking data suggested that revenue or output inefficiencies were on thesame order of magnitude or pertiaps somewhat greater than the typical co st ine fficiencies findings inother research (English,Grosskopf,Hayes, and Yaisawamg 1993).' Revenues can more than double if output doubles (scale e con omie s), or revenue may increaseby producing two products jointly rather than separately (scope econ omies) if large firms o r joint-production firms can charge higher prices for their services. This may occur if customers preferservice s that can only be provided by a larger finn, or if customers enjoy the additional conven ienceof 'one-stop shopping,' havingagreater variety of service s deliveredb y t h e same firm. These customerpreferences m ay be reflected in higher revenues fort h efir ms that provide the extra ser vices, providedthat these firms have the market power to extract some of this consumer surplus. The one study ofthis topic in batiking found revenue scale econ omies to be 4 or less of revenues , andrevenuescopeecon omies to be small and statistically insignificant (Berger, Humphrey, and Pulley, 1995).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    7/46

    THE EFFECTS OF MEGAMERGERS ON EFFICIENCY AND PRICES 1 0 1excessive inputcosts.The estimated ineflSciencies w ere prim arily techn ical, so thatbanks were generallywell inside their production-possibilities frontiers. Allocativeinefficiencies, or errors in responding to m arket prices for inputs and o utputs, w ereusually reladvely small. '

    There have been no profit efficiency studies of mergers in any industry to ourknow ledge. W e argue that analysis of profit efficiency is more app ropriate to theevaluation o f mergers than cost efficiency. Profit efficiency takes into acco unt boththe cost and revenue effects ofth e changes in output scale and scope that typicallyocciu- subsequent to a merger. Cost efficiency analysis, which takes outputs asgiven, cannot evaluate whether any revenue changes from shifts in output offsetthe cost changes except in the special case in which outputs r e m ain c ons tan t(i.e.,the output vector ofth e consolidated firm equals sum of tbe output vectors ofth eacquire r and acqu ired firms prior to the merger). In addition, profit efficiency is themor e general con cep t that includes cost efficiency, so evaluation of profit efficiencychanges a ssociated with mergers incorporates w hatever changes in cost efficiencyoccur plu s the revenue and co st effects of changes in output. For po licy analysis, itis approp riate to considerboththe change in the value of real resources consum ed,which is represented by the change in costs, and the change in the real value ofoutput produ ced, w hich is represented by the change in revenues for given prices,and this is accomplished through evaluating profit efficiency.Although there are no profit efficiency studies of mergers, some studies havecompared simple pre- and post-merger profitability ratios, such as the retum onassets (ROA) or retum on equity (ROE) based on accounting values. There is noconsensus as to whether mergers increase profitability some of these studiesfound improved profitability ratios associated with bank mergers (e.g., Comettand Tehranian 1992, Spindt and Tarhan 1992), although most others found no

    ' These find ings primarily reflect the results of Berger, Hancock, and Humphrey (1993 ) andDeYoung and Nolle (1996). Akhavein, Swamy, and Taubman (1997) also obtained qualitativelysimilar results when their analysis was restricted to those observations in which the predicted netputswere o fth e correct sign (i .e. , positive outputs and inputs). When this restriction was dropped, tfieirmeasured profit ine fficiencies became very small. Berger, Cummins, and Weiss (1995) found profitinefficiencies of similar magnitudes in the insurance industry. Humphrey and Pulley (1997) foundsmaller profit inefficiencies for banks,but they were examining interquartile differences in efficiency,rather than average inefficiencies. Berger, Cummins, and W eiss (I99S) and Humphrey and Pulley(1997) used both the standard profit function (which takes output prices as give n) and a nonstandardprofit function (which takes output quantities as given).'' Other advantages of profit efficiency over cost efficiency are discussed in Berger, Hancock, andHumphrey (1993).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    8/46

    102 J A L A L D . A K H A V E I N E T A L .improvement in these ratios (e.g., Berger and Humphrey, 1992; Linder and Crane,1992;Pilloff, 1996).'2.>3

    These profitability ratio studies have similar methodological problems to thecost ratios discussed abovethey do not control for inp ut prices, and they simplydivide by a crude indicator of bank scale (assets or equity). However, ih e moreimportant problem is that without controls for output prices, there is no way todetermine the source of any profitability change. The ROA and ROE ratios mightincrease because of an improvem ent in profit efficiency associated with m ergersin whichquantitiesof outputs andinputsw ere altered for a given vector of input andoutputprices.Altematively, an increase in market power associated with mergers in w hich thepr ices of bank products are made less favorable to consumersmight be respon sible for a finding of higher ROA or R OE after mergers. These tw osourc es of profitability chang es cannot be disentangled w ithout a profit efficiencyanalysis. Similarly, merger event studies, which use market equity values ratherthan accounting data, cannot differentiate between efficiency and market powereffects of mergers, since markets value profitability increases from all sotircesequally.'^Bank Me r ge r s and Mar ke t Power. Under certain conditions, bank mergers alsohave the potential to raise profits through an increased exercise of market powerin settingprices.Mergers betw een banks that have significant local ma rket overlapex ante may increase local market concentration and market share and allow theconsolidated banks to raise profits by setting prices less favorable to consumers(higher loanrates,lower depositrates).Mergers between b anks in different regionsgenerally d o not affect local market structure significantly anda r e less likely to raisemarket power. If anything, such mergers may bring new aggressive competitionto bear on previously imperfectly competitive markets and reduce the effects of

    See Berger and Humphrey (1992) and Rhoades (1994) for extensive discussion s of these ratioana lyses. More rec ently, Schrantz (1993) also found higher profitabil ity ra tios for banks in stales withrelatively l iberal takeover policies that make mergers and acquisitions relatively easy. However, itcannot b e determined from such an analysis whether the profitability is derivedit>mactual mergersand acquisitions or simply the greater perceived threat of them.

    '' In a related analysis, Fixler and Zieschang (1993) measured relative efficiency by the ratio of avalue-weighted output index to a value-weighted input index. They found that acquiring banks weremuch more efficient than other banks prior to merger and maintained this a dvantage after merger.G i venthat otherstudies typicallyindacquiring banks to b e more efficientthant h e banksthey acquire,this suggests an improvement in total efficiency fi'om mergers. Since they include an output index aswell as an input index, the improvement could be fivm either revenue or cost sourc es. Event studies usually find no improvement in the total market value oft he consolidating banksassociated with merger announcements. The market usually bids down the equity value of acquiringbanks and bids up ^e value of the acquired banks, so the change in the combined equity value isusually n o t significantly different from zero (Hannana n d Wolken, 1 9 8 9 ; Houstona n d Ryngaert, 1994;

    Pilloff, 1996). See Rhoades (1994) for a more complete summary of event study findings for bankmergers. In aiddition, the post-merger performance improvement has bee n found to be insignificantlyrelated to theeqcatymarket's response to merger announcements(Pilloff, 1996).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    9/46

    T H E E F F E C T SO FM E G A M E R G E R SO N E F F I C I E N C YA N D P R I C E S 103market power.'^ Note that increases in local market concentration and marketshare need not affect prices substantially if the local market is highly contestable,if there are significant nonbank alternative sources of similar services, or if thereis a substantial coincident improvement in bank efficiency from the merger that ispartially passed on in consumer prices.

    D espite the antitrust policy focus o n price effects of mergers, few academ ic stud-ies exist which compare prices before and after mergers. An exception is Hannanand Prager (1995), which finds that mergers that violate the Justice Departmentguidelines for banks (local market Herfindahl over 1800, increase of over 200)sometimes substantially lower the dep osit rates paid by banks in the affected mar-kets, consistent with market power effects of mergers. They did not control forthe efficiency effects of mergers, so that their results may incorporate some priceeffects of any change in efficiency as w ell. That is, if mei^ers increase op eratingefficiency and part of the change in efficiency is passed on in p rices, the measuredeffect of m ergers on prices ma y imderstate the market pow er effects. A ltematively,th e measured market pow er effects may be overstated if mergers reduce efficiency.' *

    Som e further insights into this problem may be gained by ex amining the larg-er literature regarding the effects of market concentration and market share onprices and p rofits. It should be bome in mind that there may be many differencesbetw een the dynamic effects of mergers on perfonnanc e and the static equilibriumrelationships between market structure and perfonnance.

    There are two opposing sets of theories regarding the relationships betweenmarket structure (concentration and market share) and both prices and profits.Accordingt o traditional market pow er hypotheses (includingt h e structure-conduct-performance hypothesis), high concentration and/or market share are associatedw ith prices that are less favorable to consume rs, which in tum create higher profitsfor producers. In co ntrast, accordingt o the efficient-structure hypothesis (Demsetz,1 9 7 3 ; Peltzman, 1977),concentration and market share are positively related tofirmefficiency, with more efficient firms growing larger and gaining dominant marketshares. Under the efficient-structure hypothesis, high concentration and marketshare may be associated w ith prices that arem o r e favorable to consumers if someof the efficiency savings are passed on to consumers (possiblya s part of the processof gaining dominant marketshares).Tbe g reater averag e efficiency of firms in moreconcentrated markets and w ith higher market shares w ill also yield higher profits

    Some banking products do trade in national markets, such as large corporate loans and largecertificates of deposit. However, any increase in U.S. national concentration from individual bankmergers is unlikely to create significant market power at present because the national market iscurrently so unconcentrated. This may or may not remain the c ase in the future. Price effects o f mergers have also been studied outside of banking. Kim and Singal (1993) foundthat airlines raised prices substantially after meigers. They acknowledged, however, that because theydid not control for efficiency cha nges,theirprice changes incorporated confounding effects of marketpower and efficiency changes which could not be separately identified.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    10/46

    104 J A L A L D . A K H A V E I N E T A L .for these firms. The emp irical literature on the determination of prices and profitsprovides some support for both sets of theories.*^This brief sutnmary of theory has two main implications for empirical studiesof the effects of market pow er. First, the analysis should focus primarily on p rices,rather than profits, since the market-pow er theories have opposite predictions fiomthe eflScient-structure theory regardhig relationship between market structure andprices but sometimes yield the same p rediction for the market structure-profit a sso-ciation. Second, any analysisof e i therpric es or profits shou ld control for efficiency.Otherw ise, the observed relationship betw een m arket structure and prices or profitsmay confound t h e efFects of marke t pow er and efficiency, w ithout allow ing sep arateidentification of either effect.In sum , the literature suggests that bank mergers have the potential to increaseprofitability through increases in cost efficiency, profit efficiency, or market powerin setting prices. Studies of cost ratios and cost efficiency generally found thatthe potential for cost efficiency improvement was not realized for most mergers.In contrast, there have been no aca dem ic studies of the profit efficiency effectsof mergers and very little research on the market power effects of bank mergers.Studies of the effects of mergers on profitability ratios or equity values may con-foimd c hanges in profit efficiency w ith changes in the exercise of market pow er insetting price s. S tudies of the effects of equilibrium m arket structure on prices andprofits provide some support for both market power and efficient structure effectsof concentration and market share. In the remainder of this paper, we investigateboth the profit efficiency and m arket pow er effects of merge rs and try to identify e x

    Studies of the effects ofmarketpowero n bankpn'cas have generally foundthat banks thatoperatein more concentrated local markets pay lower rates on deposits (e.g. , Berger and Hannan 1989) andcharge higher rates on loans (e.g., Hannan 1991), consistent with the maricet power hypotheses.However, these studies generally failed to con trol for efficiency in their analyses, creating a possiblebias in the measured effect of market power, since efficiency may be correlated with the regressors(concentration, market share) and efficiency may be an important determinant of the dependentvariable (price). Berger and Hannan (1996) addressed this problem by including direct measures ofefficiency in the analysis and stil l found strong evidence of marketpower in setting loan and d epositprices.Otherstudies of the associationbetween profitability an d market structurei nbanking and elsewhereoften found that market share (but not concentra tion) was positively related to profitability when bothmarket share and concentration were included in the profitabil ity regression. However, there isdisagreement over whether market share represents the exercise of market power on differentiatedproducts (e.g., Rhoades 1985, Shepherd 1986) or firm efficiency which was left out of the model(e.g., Smirlock, G ill igan, and Marshall 1984, Smiriock 1985). Recent ana lyses (Berger 1995a, Bergerand Hannan 1996) tried to resolve this problem by adding direct measures of efficiency to theanalysis. They gene rally found that concentration and market share had little effect on profitabil ityafter con troll ing for efficiency, de spite the market power effects on prices.Thus, substantial market power from high levels of concen tration or market share appear to havesubstantial effects on prices, but not on profitabil ity. One possible explanation of this discrepancymay be a 'quiet-l ife' effect in which firms take part of any benefit of market power in the form ofless rigoro us adherence to efficiency maximization. In this event, part of the gains from pricing maybe reflected in lower efficiencyrather han in higher profits. Berger and Hannan (1995, 1996) foundevidence consistent with quiet-l ife effects in banking.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    11/46

    THE EFFECTS OF MEGAMERG ERS ON EFFICIENCY AN D PRICES 105ante conditions that predict when either profit efficiency or market power is likelyto be increased.m . The Measurement of Profit EfficiencyDetermining Profit Efficiency.For the piupose of evaluating whether andb yhowmuch ban k megam ergers affect profit efficiency, w e estimate the profit efficiencyof all largeU.S. banks (assets over$1billion) overthep eriod 198 0-199 0, whetheror not they were involved in megamergers. For each megamerger, we calculate theimprovement in efficiency associated with the merger as the efficiency rank of theconsolidated bank after the merger less the weighted average rank of the acquiringand acquired banks before the merger. In all cases, the efficiency rank is calculatedrelative to the p eer group of all large banks that had data available over exactly thesame time period as the consolidated or m erging bank.Inthis way, we controlforany industry-wide changesinprofits or efficiency that may occur andkeepthe dataconsistent and comp arable over time.

    Thespecification of the profit fimction and estimation of profit efficiency closelyfollow the procedures of Berger, H ancock, and H umphrey (1 99 3). W e estimateamodified Fuss normalized quadratic variable profit function aswell as quantityequations which embody cross-equation restrictions that help identify th emodel(similar to the more commonly specified input cost share equations). Thus,theprofit model is given by:

    7 r ( P ^ T ^ ) P FPn i^i Pn j^i j^ i

    (1 )r=l ^r=l =l t=l r=l P

    n-1 *gi = a i + X l

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    12/46

    106 J A L A L D . A K H A V E I N E T A L .prices is imp osed by normalizing the variable profits and prices by the price of thelastnetput. '^Allocative inefficiency is defined as the loss of profits from making non-profit-maximizing choices of netputs in the production plan. Allocative inefficiency ismodeled as if the bank we re responding to shadow relative prices rather than ac tualrelative pricesm aximizing p rofits as if T Jpi /pn w ere the relative price of netputt to netput n rather thanPi/pn- Allocative inefficiency is measiwed as the loss ofprofits from Tbeing different from a vector of 1' s , or7 r ( p , z,l,^) -IT(P,Z, T ,^ ) =

    i , vTechnical inefficiency is defined as the loss of profits from failing to meet theproduction plan. Technical inefficiency is modeled as each of the netputsibeing ^jbelow the efficient frontier, i.e., the outputs being too low or the inputs being toohigh. T echnical inefficiency is mea sured as the loss of profits from ^ being differentfrom a vector of O ' s , or 7 r ( p , z , T 0) - T { P Z T 4) = Et=i,...,n^P-A fuUy efficient firm with no allocative or technical inefficiencies would earnthe maximum or optimal level of profits for its given variable netput prices pand fixed netput quantities z, or n = trip^z, 1,0). The total profit efficiencyratio for each bank is measured as the ratio of actual profits to optimal profits,7r/7r = 7 r ( p , z, r, O / ^ ' C P I ^ J 1> 0 ) . Both the numerator and the denominator in thisformula are measured as predicted values that exclude the random errorterms.T heefficiency ratio varies over the range (- o o , 1]- the best a firm can do is earn all ofoptimal profits (7r/7r = 1 ) , but the worst a firm can do is unbounded since the firmcan alw ays make arbitrarily larger losses by using more inp uts without prod ucingmore outputs. For the puip oses of this study, we focus on the total efficiency ratio,w hich incorporates both allocative and technical efficiencies.Our specification of the profit model in (1) and (2) includes four variable netputs(n = 4). Total loans and total securities (securities measure d as all assets otherthan loans) are the outputs, and total deposit funds (including purchased funds)and labor are the inp uts. Equity capital is the sole fixed netput. T he specificationis parsimonious because of the difficulty of estimating a nonlinear system withcross-equation restrictions. The choices of outputs and inputs are consistent with

    '^ A concern with this specification is that it takes prices as given, an assumption that may beviolated if the firm exercises market power in setting prices. However, our results be low suggest thatno ser ious bias has been created by this assumption. First, in separate regressions of the price effectsof mergers, we find that estimated market power effects of mergers are extremely small relative toprofits. These effects a re also very small relativet o the estimated effects o f mergerso n profit efficiencythrough changes in netput quantities. Second, the allocative inefficiencies, which depends on priceeffects, are found to b e negligible in the results below, suggesting that prices are fairly unimportantin determining profits. Finally, we try specifying an alternative, nonstandard profit function belowthat removes output prices from the specification in favor of output quantities. The main results o fthe model are materially unchanged, strongly suggesting that any problems with the specification ofprices in the profit function are not important. By construction, allocative and technical inefficiencies add up to total inefficiency and donot interact, since the level of allocative inefficiency is unaffected by f and the level of technicalinefficiency is unaffected by T .

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    13/46

    T H E E F F E C T S O F M E G A M E R G E R S O N E F F IC I E N C Y A N D P R IC E S 107the intermediation or asset approach of Sealey and L indley (1 97 7), under whichintermediated assets are the ouQ)uts and sources of funding are the inputs of afinancial institution.^ The speciiScation of equity as a fixed input addresses thepotential problem that the size of a bank, especi^ly its loan portfolio, is stronglytied by both regulators and maikets to its quantity of equity capital available toabsorb loanlosses.Equity is very difficult and costly to chang e substantially ex cep tover the long run, and so we treat this important input as fixed. If equity were notspecified as fixed, the largest banks may b e me asured as the mo st profit efficientsimply bec ause their higher capital levels allow them to have the mostloans.^ Ourspecification as a wh ole may be thought of as measuring efficiency by how w ell thefirm is able to eam a retum on equity by using de posit funds and labor to produceloans and securities.

    The model in Equations (1) and (2 )i s estimated by nonlinear iterative Seem inglyUnrelated Regression techniques (NTTSUR).^ We use annual data from 1980 to199 0 for all U . S . banking organizations - both merging and nonmerging - tha t hadassets of at least $1 billion in at least one year over that interval. However, theorganization need not be present in all years to be included in the estimation dataset Besides eliminating ttie small banking organizations, the only deletion is thatdata from the merger yea r itself are left out for the con solidated banks involved inmegam ergers. This is because such data are likely to contain very significant on e-time transition costs. Efficiency is calculated for each of the at least three entitiesinvolved in a merger (1) the acquiring bank during the available years before themerger, (2) the acquired bank or banks du ring the available years before the merger,and (3) the consolidated bank during the available years after the merger. All ofthe efficiency levels and ranks of the m erging ban ks are determined relative to p eergroup s of large banks that have data available over exactly the same time intervals.As described below, this generally involves tracking separate peer groups of largebanks for each merger.

    The allocative inefficiencies for each bank (the losses from a poor productionplan) are estimated from the T < , the conventional profit function parameters, andthe prices for that bank. To keep the model manageable, the n , i - 1, 2 , 3 aretreated as parame ters that are constant across banks . Unfortunately, this limits thevariability of allocative inefficiency across firms, but most prior research suggests^ Deposits havebothinput and output attributes, and have been mod eleda s suchin cost functions byspecifying both deposit quantities and prices (e.g., Berger and Hun^hrey, 1991). However, depositscannot be modeled as having both traits in the variable profit function, which does not allow forquantities o f variable outputs.^' Equity capital is prefened to the value of fixed asse ts (premises and equipment) as a fixed inputFixed assets are very small in banking, only about 2 0 as large as equity, and can b e increased muchmore quickly and e asily than equity. Profit ftmctio n convexity in prices is imposed by constraining the matrix of ^y to be positivesemidefinite, which assures nonnegative allocative inefficiency. The model is first estimated uncon-strained and the positive semidefinite matrix that is 'clos est' to the estimated * matrix (in the senseof minimizing the Euclidean norni of th e difference) i s selected . The other model paiam etos are thenre-estim ated given this revise d * matrix. See A khavein, Swamy, and Taubman (1997).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    14/46

    108 J A L A L D . A K H A V E I N E T A L .that this may not be important because allocative inefficiencies are usually foundto be small relative to technical inefficiencies (e.g., Aly, G rabow ski, Pasurka, andRangan, 1990; Berger and Humphrey, 1 99 1; Berger, Hancock, and H umphrey,

    23To estimate the technical inefficiencies (the losses from failing to meet theproduction plan), we follow the 'distribution-free' approach of Berger (1993),which is based on Sickles and Schmidt (1984). Each of the 4 equations in (1)and (2) contains a composed error term (et ^i), a random error term minusthe technical inefficiency in netput i for the individual firm. The distribution-freeapproach separates the technical inefficiency from the random error by assumingthat inefficiency is constant over the time interval of measurement, whereas therandom error tends to average out over time. Thus, the ^ j , i = 1 , . . . , 4 for eachbank involved in a merger is estimated by the difference between the maximiunaverage residual from the equa tion containing C j - ^ifor the sample of banks withcomplete data over the corresponding time interval and the average residual forthe bank in question. If the efficiencies are not perfecdy constant over the timeinterval, the measured technical inefficiencies may be interpreted as the deviationsof the average practice of the bank from the best average practice frontier. Whe n

    computing the level of efficiency (not the rank), we also truncate the averageresiduals of each bank by assigning the most extreme 5% at the top and bottomof the distribution to the 95th and 5th percentage points, respectively, to furtherreduce the effects of random error.The Me game r ge r Da ta Se t . W e collected data on all mergers of U.S . bankingorganizations diuing 1981-1989 in which both partners had at least $1 billion inassets.All but a few of these mergers were between h olding comp anies rather thanbetween individual banks. For our analysis, we treat the 'high' holding companythe holding comp any w hich is not owned by any other holding company - as thedecision making unit that tries to max imize profits. That i s , w e sum together all thecommercial banks that are jointly owned through the holding company structureand treat them as a single profit-maximizing unit (although for convenience wesometimes still refer to the consolidated entity as a bank) . This is consistent withthe efficiency claims made by bank consultants, which imply that the merger-related efficiency improvements are made in a coordinated w ay through the holdingcomp any structure. Bank regulators and the Justice Dep artment similarly focus onmarket structure (concentration, market share) at the holding comp any level.

    The ex pected time pattem of costs and revenues associated with bank mergers isthat some ex tra nonrecurring or transitional costs (e.g., legal exp enses, consultantfees, severance pay, etc.) will occur in the short term, but that other recurringexp enses w ill fall and longer-term revenues ma y rise. Thus, it i s exp ected that profitsmay be temporarily lower during this transition but possibly higher afterwards. Exceptions that sometimes fitid allocative inefficiencies to be relatively large are Fenie r andLovell (1990) and A khavein, Swamy, and Taubman (1997).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    15/46

    T H E E F F E C T S O F M E G A M E R G E R S O N E F F IC I E N C Y A N D P R I C E S 109Ideally, t o judg e the benefit of a merger, one w ould detennine the p resent value ofall fiiture profit imp rovements after a m erger. Since this is not possible, we simplydrop the datafi-omhe yea r of the merger to reduce the effects of transition co sts.Fortunately, Berger and H ump hrey (199 2) foimd that merger cost results remainedmaterially unchanged whether the single year of data, 2 years of data, or 3 yearsof data subsequent to the mergers were dropped, suggesting that our treatment oftransition costs likely does not create serious biases.The netp ut quantities and prices w ere constructedfiromCall Rep ort informationover 1980-1990. The ex post efficiency and perfonnance indicators for the con-solidated bank after the merger were based on all the years following the mergerimtil either another megamerger involving that bank occurred or the year 1990 w asreached . The pre-merger efficiency of the acquiring bank and the acquired bank(s)w ere based on all the years going backward in time prior to the merger until eitheranother mega merger involving that bank or the year 1 9 8 0 w as reached. In the usualcase in w hich exactly two banking organizations merged, this involves comp utingthe efficiency measures for 3 banks, each over a different time interval - the app ro-priate yea rs after the m etger for the consolidated entity and the app ropriate yearsprior to the merger for both the acquiring and the acquired bank separately.^ If abank acquired more than one other bank in the same or consecutive years, thesewere combined into a single merger observation with additional pre-merger inter-vals over w hich efficiency i sm easured. Ina U cases,the data on merging banks w erecomp ared to the set of large banks w ith complete data over the same time interval.If data w ere unavailable for the merging bank s for any of the intervals, the mergerw as not used in the efficiency an alysis. In many case s, the data w ere imavailablebecause one of the entities w as a thrift or a foreign-owned institution w ithout com-parable data available. In all , 69 of the 114 megamergers over 1981-1989 wereretained, although they appear in only 57 observations since some observationscontain mergers among 2, 3, or 4 entities. See Berger and Humphrey (1992) foradditional details about the data set.

    IV. The ProfitEfficiency EfFects of MegamergersT h e Level o f Prof i t Ef f ic iency . W e begin discussing the results with some informa-tion on theleve lsof profit efficiency of merging and nonmerging b anks. The morerigorous comparisons of the efficiency r anks will be discussed below. AppendixTable AI describes the data that went into the standard profit efficiency model inEquations (1) and (2) above, and Table AH gives the estimated parameters of themodel.

    For example, suppose bank A (in existence since before 1980) acquired bank B in 1988, andbank B had acquired bank C in 1984. F or the 1988 A -^ merger, the ex post data on the consolidatedbank would be on A's perfonnance over 1989-90, the ex ante data on the acquiring bank would beon A's perfonnance over 1980-87, and the ex ante data on the acquired bank would be on B over1985-87.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    16/46

    110 JALALD.AKHAVEINETAL.The level of profit efficiency the ratio of predicted profits to maximum oroptimal profits on the frontier {tt/n) - w as measured for all large banks over19 80- 19 90, w hether or not they were involved in m egamergers. For each firm'spricesp and fixed netputs z, we take the ratio of the predicted values of profitsusing the estimated values ofTand^, TT{P,Z,T,^) , divided by the predicted value

    with vectors of1's and O's replacing Tand ^, respectively, orn{ p , z,l ,O).Merging banks improved their profit efficiency substantially after mergers. Theasset-weighted average of the acquiring and acquired banks prior to megamergerswis44% , and rose to 7 1 % for the consolidated banks after merging, a statisticallysignificant increase of 27 percentage points. That is, the asset-weighted averageof banks that participated in mergers earned 44% of optimal profits before themerge rs, and the consolidated banks earned 7 1 % of maximum profits after themergers. H owever, thisdoesnot necessarily indicate a merger-related imp rovementin efficiency because profit efficiencies may vary with the number of observationsavailable and the economic environment of the banks, which can change fairlyrapidly w ith variations in open-m arket interest rates (w hich fell substantially overthe 1980s).W hat matters instead is how the measured improvement for a merging bank

    compares with the measured improvement for its peer group of large banks withdata over the same pre- and post-merger periods as the merging bank. Puttingtogether the peer groups for all 57 megamergers, the weighted average pre-mergerprofit efficiency level for all banks w as 24 % before the mergers, and rose to 34 %after the mergers, for a statistically significant increase of 10 percentage points.That is, the asset-weighted average ofalllarge banks that had consistent data overthe same time periods as the acquiring and acquired banks prior to the mergerseam ed 2 4% of optimal profits, while those in existence over the same dme periodsas the consolidated ban ks after the mergers eamed 3 4% of maximum profits. Thisresult suggests that profit efficiencies do vary with the economic environment ofthe banks, but not enough to explain the efficiency improvement of the mergingbanks. Subtracting the 10 percentage point improvement of all banks fi-om the27 percentage point increase for merging banks leaves a 17 percentage pointadditional increase in efficiency associated with megamergers. Thus, banks thatchose to m erge w ere more profit efficient on average than other banks ex ante, andapp eared to add to this advantage by improving their efficiency by mo re than otherbanks ex post.

    Profit efficiency for merging banks can be decom pose dintotechnical and alloca-tive compon ents. The averagelevelof technical efficiencywas46 %beforeamergerand 73% afterwards, rising 27 percentage points and mirroring the situation foroverall profit efficiency. Allocative efficiency was already high for these banksprior to merging, an average of9 8 . 3 % ,and was little changed after merging, fallingby 0 .1 % to 98 .2 %. This confirms our speculation based on prior research thatallocative inefficiency would be small relative to technical inefficiency, so that

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    17/46

    T H E E F F E C T S O F M E G A M E R G E R S O N E F F IC I E N C Y A N D P R I C E S 1 1 1cross-sectional variations in allocative inefficiency (which are mostly suppressedby our assumption of constant r's) would likely not be important.Profit efficiency can also be decomposed into output and input components.Output inefficiency in the profit function includes the output technical inefficiency(failure to produce as much output as planned) and allocative inefficiency frommisresponding to output prices (including the cost and revenue effects of devi-ating from the profit-maximizing production plan). Input inefficiency is definedsimilarly.^^ For merging banks, output efficiency climbed 13 percentage points(from 69% pre-merger to 82% post-merger), while input efficiency rose 14 per-centage p oints (from 7 5 t o 8 9 ) . Thus,both input and output efficiency im provedsubsequent to m ergers. Note that the rise in input efficiency does not nece ssarilyimply any changei n cost X-efficiency. Thisi s because the change in input efficiencyincorporates part of t h e change in outputs subsequent to the merger. For examp le,if planned outputs are smaller and require fewer inputs, inputs may be closer totheir optimal levels and input efficiency may be improved, but cost X-efficiencyis imchanged because it takes outputs as given. Only in the special case in whichoutputs remain constant does the change in input inefficiency necessarily reflecta change in cost X-efficiency, and as shown below, the outputs do change after amerger.

    In the remainder of the analysis, w e focus simply on the total efficiency ran ks ofthe merging bank s, rather than dealing w ith the cumbersome array of comp onentsof efficiency or w ith the level of efficiency. Th e use of total efficiency, the ratio ofpredicted profits to optimal profits {it/ir ), corresponds well to the social benefitconcept of the real value of output produced less the real value of resourcesconsumed. The rank of total efficiency is preferred to the level because the rankis neutral with respect to changes in tiie distribution of measured efficiency overt ime, which do seem to occur. In addition, our 'distribution-free' methodologyin which the random error is averaged out over time - introduces some biases intothe measurement of the levels of relative efficiency because different numbers ofobservations are available for difFerent mergers. Fortunately, the expected value ofthe efficiency rank does not depend on the number o f observations (although thevariation in the measured rank around the true rank obviously does depend on thenumb er of observations). As w ill be shown, our main results are also robust to theuse of rank or level.Changes in Prof i t Ef f ic iency Rank. W e comp ute the rank of a merging bank(acquiring, acquired, or con solidated) relative to its peer group oflargebanks w ithcontemporaneous data as the proportion of peer group with efficiency below thatof the merging bank. Thus, a merging bank with total efficiency (7r/7r) better than8 0 of its peer group is assigned a rank of 0.80.

    ^' Forth e purposes of discussion here, we somewhat arbitrarily count the very small amount ofinput-output price allocative ine fficie ncy -the interaction of misresponding to inputa n d output prices- a s output inefficiency (see Berger, Hancock, and Humphrey, 1993).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    18/46

    1 1 2 JALALD.AKHAVEINETA L.TABLE I. Changes in profit efficiency and perfonna nce raik frommegamergers (1981-1989)

    Profit efficiency:i:litPre-merger rankPost-merger rankImprovement in rank

    57Mergers

    0.740.900 . 1 6

    Pre-merger rank:Lowestl/3rd

    0.560.830 . 2 7

    Adjusted retum on assets: WTTA = ROA*Pre-merger rankPost-merger rankImprovement in rank

    0.520.570.05

    0.500.610.11*

    Adjusted retum on equity:NI*/EQ = ROE*Pre-merger rankPost-merger rankImprovement in rank

    0.620.650.03

    0.510.670 . 1 6

    Highestl/3rd

    0.920.950 . 0 30.470.500.030.670.59

    - 0 . 0 8* (**) Improvement is statistically significantly different fTom zeroat the 10% (5%) level, two sided.Note: The profitabil ity ratios, ROA* and ROE*, are adjusted toremove the 'no isy' fluctuations asso ciated with loan loss provisionsand taxes.

    Boththepre-and post-mergerranks,along with the resulting chang e in rank, arereported in Column 1 ofTableI. The pre-m erger profit efficiency rank of mergingbanks, which is an asset-weighted average of the acquiring and acquired bank'sefficiency rank, averaged 0.74. Consistent with the eflBciency levels discussedabove, merging banks are more efficient on average than 74% of all large banksprior to merger. After the mergers, the average profit efl&ciency rank increasedto 0.90.^^ Thus, the average bank megamerger is associated with a statisticallysignificant 16perce ntage point increase in rank. This is consistent with the m erger-related 17 percentage p oint improvement in average profit efficiency level relativeto the peer group change reported above.

    W hile profit efficiency is our preferred m easure to gauge the profit effects ofbank megam ergers, it is helpful to comp are the results w ith standard profitabilityratios,retum on assets (ROA ) and retum on equity(ROE),which should incorporatesome p rofit efficiency effects as well as any m arket pow er effects of m ergers. W eremove from the standard measures the confounding effects of variations in taxespaid and loan loss provisions, w hich often fluctuate substantially over time in w ays^ Altho ugh this profit efficien cy rankof 0.90 is seemingly high, it is not necessarily indicative ofa high efficiencylevel.On average, consolidated banks had a profit efficiency level of71% - i .e. , theylost an estimated 29% o f their potential profits to inefficiency.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    19/46

    T H E E FFE CT S O F M E G AM E RG E RS O N E FFICIE NCY AND PRICE S 1 1 3

    that do not reflect op erating efficiency. W e refer to these measures w ith the noisycomp onents of net income removed as adjusted retums on assets (ROA ^) and equityAs shown in Table I, the ROA* rank of merging banks increased an average 5percen tage points (fi-om a rank of 0.52 to 0.57) and the rank of ROE* improved by3 pereentage points (fi-om 0.62 to 0.65). Neither of these changes are statistically

    significant.^* Thus, the measured improvements in profit efficiency are not wellreflected in the ranks of the more com monly used profitability ratios ROA and R OE(adjusted or imadjusted). A s noted earlier, the profitability ratios may b e inaccurateindicators of performance because they do not take account of differences in theprices faced by the banking firms, and these ratios divide eamings by cmde mea-sures of bank size, total assets or equity cap ital, rather than by the potential profits7r benchmark, which is the highest profits that can be eamed for the equity andprice s faced by the firm. W hen adjusted n et income w as divided by potentialprof-its (NP/n) , merging firms imp roved by a statistically significant 12 p ercentagepoints in average rank fi'om 0.76 to 0.88 (not show n), much closer to the profit effi-ciency results. Perhaps more important from an analytical viewpoint, the changesin profitability ratios confound two major eflfects that should be separated whenevaluating mergers - changes in efficiency and changes in market pow er in pricesetting. Thus, it is perhaps not too surprising and perhap s partially ex plainable w hythe changes in the stanckrd profitability ratios differ fiim the measured efficiencyrank improvements.

    Columns 2 and3of Table I show the same p re- and post-merger efficiency andprofitability ranks and chang es in rank for merging banks w ithin low est and highestthirds,respectively, in terms of pre-merger profit efficiency itr/tr ) rank (weightedaverage of acquirer and acquired). Banks in the low est third in terms of pre-m ergerprofit efficiency rank moved from a rank of 0.56 up to a rank of 0 .8 3 ,yielding asignificant 27 percentage point improvement. Banks with the highest pre-mergerrank moved from a rank of 0.92 up to 0.95, giving only a 3 pereentage pointimprovement in rank, although it is statistically significant. Thus, the banks thatimproved the most w ere generally those w ith the lowest pre-merger profit efficiencyrank, and vice versa for ban ks w ith the highest pre-merger efficiency. T his is clearlyrelated to the opportunity to improve banks in the highest third were already atthe 92nd percentage point in the distribution for profit efficiency, and so could nothave the mean efficiency improvement of 16 percentage points shown in the firstcolumn of thetable.A similar result is show n fortheprofitabilityratios.Banks w ithlow pre-merger profit efficiency experienced statistically significant improvementsin theirROA*and RO E' of11and 16 perce ntage points, respectively. Thus , when

    Thus, ROA* = NI*/TA and ROE* = NT/EQ , where NI* is net income plus taxes and provisions,TA = total a ssets and EQ = equity capital. The average ROA* (ROE*)level for merging banks was1.4% (22%) pre-merger and impro ved to 1.6% (23%) post-merger.^ When the ranks of unadjustedROA and ROE values were used (not shown ), the merger-relatedimprovements were 0.006 and 0.034, respectively, neither o f which wa s statistically significant.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    20/46

    1 1 4 JALAL D. AKHAVEINETA L.profit efficiency movements are pronounced, they do appear in the profitabilityratios, suggesting that these ratios do incorporate efficiency movements, but arenot as sensitive as the profit efficiency rank when the efficiency movements arerelatively small.How Prof i t Ef f ic iency Improves . The question now arises as to how or w hy w efind a profit efficiency improvement from megamergers, whereas no cost efficiencyimprovement was found from these same mergers using the same data set inBerger and Hump hrey (19 92 ). As discussed above, the only special case in w hichit is guaranteed that the cost and profit efficiency effects of mergers will be thesame is if the consolidated firm produces exactly the same ou^ut vector as theacquirer and acquired firms produced com bined prior to the merger.

    To explore this issue, we examined the output behavior of firms engaged inmegamergers versus the same peer groups used in the efficiency calculations.A lthough merging and nom nerging firms both grew in size substantially after themergers owing to trends in banking in the 1980s, firms engaged in megamergershad a decided shift in product mix relative to the other firms. Prior to mergers,merging firms had an average of 56.4% of total assets in the loan output, belowthe 59.5% of all large banks. Subsequent to the mergers, merging firms raisedtheir average loan/total asset ratio almost 7 percentage points to 6 3 . 3 % , passingthe peer group average for all large banks, which increased only 3.4 percentagepoints to 62.9%. One hypothesis is that consolidated firms may have increasedtheir focus on loans because their larger size, greater geographic spread, and/orbroader industrial coverage allowed better diversification of risk, allowing for aprop ortionately larger loan portfolio w ith about the same am ount of total risk. Thatis , under the Divers i f icat ion Hypothes is , the market rewarded the merging firmsfor an improved diversification of loan risks by allowing them to hold a higherloan/asset ratio, all else equal. Consolidated banks may also have chosen to bemore aggressive in obtaining market share in loan markets because of their moreprom inent stature. In any event, the movement into lending represents an increasein the value of output produced and an improvement in profit efficiency, all elseequal, since loans have higher retums on average than securities. This increasein profit efficiency will not be captured in cost efficiency, which takes outputs asgiven.^'

    To examine theDivers i f ication H ypothes is further, we looked at the behaviorof the equity/asset ratio before and after the merger. If the hypothesis is correct,then the higher loan ratio should not require an increase in the equity/asset ratio tofinance. That is, w e need to rule out the p ossibility that the higher loan/asset ratiowas made possible by a decrease in leverage risk, rather than an improvement inrisk diversification. The data show that prior to merger, the combined acquiring^ A n increase in profit efficiency iTom diversifying loan risks would be consistent with the findin gelsewhere that U.S. banks that reduced their risks in the 1980s tended to have higher eamings,primarily through reduced rates paid on uninsured debt (Berger, 199Sb).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    21/46

    T H E E F F E C T S O F M E G A M E R G E R S O N E F F IC I E N C Y A N D P R I C E S 1 1 5and acquired banks had a mean equity/asset ratio of 53 basis points below themean of the peer group before the merger. This difference widened by 6 basispoints after the m er ge r- consolidated banks had an average equity/asset ratio of 59basis points below the mean ofth e peer group of laige banking organizations overcomp arable years. These data are consistent with theDivers i f icat ion Hypothes is -merging firms w ere able to take on both a substantially higher loan/asset ratio anda slightly increased leverage risk after the m erger, possibly ow ing to tbe reductionof risks from diversification. It is possible that the merging banks w anted to low ertheir equity/asset ratio by more than 6 basis points relative to the peer group mean,but w ere prevented from doing so by regulatory cap ital standards, since most largebanking organizations w ere very close to or in violation ofth e regulatory standardsduring the 198 0s. Interestingly, the increase in the loan/asset ratio may have been asecond choice for m any of these firms to increase exp ected eam ings after receivingthe benefits of risk diversification through mergers they may have raised lendingbecause the regulatory capital requirements blocked a reduction in funding costsby sw itching from equity to debt and raising leverage.^

    Benston, Hunter, and Wall (1995) tested another implication ofthe Divers i-f icat ion Hypothes is using data on the prices bid to acquire banks in the earlyto mid-1980s. Under this hypothesis, it would be expected that acquiring banksw ould bid more for other banks that tended to reduce the variance of their combinedeamings stream. Under an altemative hypothesis that the purpose of mergers is toincrease the value of the deposit insurance put option, banks would bid more forother banks that would increase total risk or put tbe bank in a position in whichit might be considered to be 'too big to fail'. The data supported theD ivers i fica-t ion Hypothes is - acquiring banks bid more for other batiks w ith low er eaming svariation and higher equity/asset ratios, all else held equa l.

    H ughes, Lang, Mester, and Moon (19 96 ) also presented infonnation consistentwith the Divers i f icat ion Hypothes is . They found that as banking organizationsincrease in size (through merger or otherw ise), their risk-expected retum tradeoffimprove s, presumab ly because of better diversification of portfolio risks. T he largerfirms tend to respond to these incentives by increasing both risk and expected retum ,consistent with the increased profit efficiency observed here for large bankingorganizations that merge, and consistent withilieDivers i f icat ion Hypothes is .Similar to our analysis above of ex ante efficiencies of merging ban ks, we alsotried constructing the thirds of the data by the pre-merger loan/asset ratio. Theresults (not shown) were that merging banks with weighted average loan/assetratios in tbe lowest third of the data had lower average profit efficiency thanmerging banks in tbe highest third (0.73 versus 0.82 ), and also bad bigber average

    ^^ In nonfinancial industries,irms ypically increase their leverage afteramerger, bringing the riskof bankruptcy back u p close to the desirable level after diversifying the risks (see Kima n d McConnel l ,1977; Asquith and Kim, 1983). Similarly, priorto implementation of formal capital requirements inthe early 1980s, acquired banks were often found to increase their leverage following mergers (seeBenston , Hunter, and Wall, 1995 for a summary of these studies).

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    22/46

    1 1 6 JALALD.AKHA VEINETAL .improvem ents in profit efiBciency (.18 versus 0.1 1). These findings again suggestthat the loan/asset ratio refiects profit efficiency and that efficiency imp rovem entssubsequent to mergers are related to the potential or capacity to improve.A nother potential exp lanation of the increase in measured profit efficiency fi-ommegamergers may be the specification of the profit function. The standard profitfunction takes prices and fixed netputs as given and assumes that firms will beable to choose freely the size of their variable ouQ)uts (loans and securities). Wegenerally find larger firms to be more profit efficient, a result that is typicallyfound elsew here. A p otential problem is diat it may be difficult or time-consumingfor firms to change the size of their asset portfolio because of the size of theirmarkets and/or because of regulatory restrictions on their ex pansion. In this event,the efficiency of smaller firms may be understated because they were unable toachieve efficient scale quickly and the profit efficiency gains from megamergersmight be overstated because the merging firms tended to have fewer barriers toovercome in achieving more efficient size. As noted above, the specification ofequity as a fixed netput in the profit function partially addresses this potentialproblem by only requiring banks to achieve the most profit for their given cap italpositions.

    We go a step further here and specify a 'nonstandard' profit function, whichtreats a ll of the outputs as fixed, so that smaller firms that cannot expand arenot disadvantaged. That is, we replace the output prices in the standard profitfunction with output quantities, so that profits are a function of output quantities,fixed netput quantities, and input prices.'' This nonstandard specification shouldeffectively eliminate most of any scale bias or merger bias.^^ The coefficients ofthe nonstandard profit function are shown in Ap pendix TableA I I I .Fortunately, ourmain results are materially unchanged fi-om the standard specification. Under thenonstandard sp ecification, m erging banks risefi-oma weighted average efficiencyrank of 0.77 pre-merger to 0.90 post-merger (not shown in tables), quite similarto the rise in rankfi-om0.74 to 0.90 found using the standard specification. T hus,it seems unlikely that any scale bias in our econometric specification is primarilyresponsible for our findings ofasubstantial imp rovemen t in profit efficiency fi-ommegamergers in banking.Sources o f Prof i t Ef f ic iency Improvem ent . A s noted above, it is important for thepurp oses of antitrust policy to know if there are any identifiable ex ante conditions

    See Berger, Cununins, and Weiss (1995). Berger, Humphrey, and Pulley (1997), Pulley andHumphrey (1997) for previous spec ifications o f the nonstandard form o f revenueandprofit functions.Note that the nonstandard form is usually altematively motivated by a desire to ailow for marketpower in output pricing, i .e. , an assumption that ou^u ts are relatively fixed and prices are chosen bythe f irm.'^ Note that the nonstandard specification specifies the same e xogeno us variablesasa standard cos tfunction ou^ut and fixed netput quantities and input prices. The &ct that cost studies in bankingusing these variables typically finds very few scale eco nom ies or disecon omies strongly suggests thatcontroll ing for these variables rem oves any significant scale or merger bias.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    23/46

    THE EFFECTS OF MEGAMERGERS ONEFFICIENCY A ND PRICESTABL E II. Ex ante variables used to predict efficiency improvements

    117

    VariableW2(EFF1-EFF2): wt difference in pre-metger profitefficiency rank (using acquired bank's weigjit).W1 (EFF1): wt pre-merger profit efficiency rankof acquiring bank.W2(EFF2): w t pre-merger profit efficiency rankof acquired bank.W2: total asset size o f acquired bank relative to thesize of the acquirer plus acquired banks.OVERLAP: percent of merging banks' deposits in samelocal marketSTINCOME: average annual growth rate in statepersonal income.HERF: wt average Herfindahl index formerging banks.SHARE : wt average deposit market share ofmerging banks.RETAIL: pre-metger deposit/asset ratio ofmerging banks.SCALE: wt average of pre-merger asset rankings formeiging banks.

    Sample mean (range)0 . 0 5( - 0 . 1 6 , 0 . 1 7 )0 . 5 8(0.15,0.96)0 . 1 6(0.00,0.57)0.28**(0.01,0.73)0 . 3 4(0 , 0.99)0 . 0 1 9( - 0 . 0 0 6 , 0 . 0 4 4 )0 . 1 8 5 0(0.0800, 0.3628)0 . 2 1(0.06, 0.45)0 . 6 9(0.36, o;93)0 . 7 6(0.40,0.99)

    ( ) Mean is statistically significantly different from zero at the 10% (5%) level,two-sided.

    that are good p redictors of efficiency improvements or ch anges in the ex ercise ofmarket power in setting prices. There is a substantial dispersion in the findings- some mergers ap pear to result in large increases in efficiency, w hereas othersappeartoresult in efficiency lossesand being ab le to predict which is which maybe helpful in the merger app roval/denial process.We put forth two major hypotheses about the prediction of merger efficiencygains and exa mine the effects of several possible ex ante conditions in a regressionanalysis in which the dependent variable is the profit efficiency rank improvementafter the m erger. The independent variables representing these ex ante conditions,shown in Table n, will also be used below to help predict changes in bank prof-itability and price s.

    Conventional wisdom in banking asserts that well-run banks seek out andacquire poorly-run banks. A preference for poor performers as acquisition tar-gets would not be surprising. Poorly performing banks typically have a relativelylow market to book value of equity, making them comparatively cheap to acquireon a per dollar of assets or deposits basis. This value of the acquired part of theconsolidated bank can potentially be increased by ap plying the manage rial policies

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    24/46

    1 1 8 JALAL D. AKHAVEIN ET AL.and procedures of the more efficient acquiring bank to it. Because of regulatoryrestrictions on combinations of banking and commerce, other commercial banksand bank holding companies are virtually the only type of firm that can purchasea commercial bank. Therefore, the market for corporate control can usually onlywork to improve managerial efl&ciency through the process of relatively efficientbanking organizations acquiring relatively inefficient banks and raising their effi-ciency. As noted in the introduction, the po tential market for corporate co ntrol inbanking will be greatly expandedin the future w hen nationw ide bankingisallowed,and banking organizations from virtually anywhere in the nation will be able tocomp ete for control of almost any banking organ ization.

    We call this effect of relatively efficient banks taking over and reforming thepra ctices of relatively inefficient ban ks theRelat ive E f f ic iency Hypothes is .We mea-sure it using the variable W 2 (EFF 1- EFF2 ), the difference in efficiency betw een theacquiring bank (EFFl) and the acquired bank (EFF2), weighted by the proportionof their combined pre-merger total assets accounted for by the acquired bank (W 2= TA2 /(TA1 + TA2 )). The greater the difference in efficiency betw een the acquir-ing bank and the acquired bank - i.e., the greater is EFF1- EFF2 - the greater is thescope for improving perfonnance. The W2 weight on this term is needed becausethe overall efficiency improvement of the consolidated firm should be directlyproportional to the relative size ofthe acquired bank, the part ofthe consolidatedbarik that is postulated to improve. T hus, under this hypothe sis, if the con solidatedfirm's efficiency is raised all the way to the level ofthe acquirer, the improvementin the consolidated firm (the dependent variable) w ill equal W 2(EFF1EFF2). Theregressio n coefficient of this variable in pre dicting merger efficiency imp rovem entsis expected to be positive under the Relat ive Ef f ic iency Hypothes is , and may beinterpreted as the propo rtion of this potential ex ante imp rovement that is achievedex post.

    A s shown in TableII , the average W 2(EFF1-EFF2)is only0.05. This relativelysmall average suggests that if theRelat ive E f f ic iency Hypothes isis true, it is likelyto raise the average efficiency rank ofth e merging banks by5percentage points orless.Moreover, the range ofthis variable - from - 0. 1 6 (where the acquired bankwas less efficient than the acquired bank) to +0.17 (where the acquiring bank ismo re efficient)suggests that adoption ofthe managerial policies and proceduresof the acquiring bank may have a wide array of consequences, including someoutcomes in which mergers reduce efficiency.A n altemative theory is that profit efficiency is more likely improved w hen theacquired and the acquiring banks are both poor performers prior to their merger.H ere the merger event itself may have the effect of w aking up manageme nt andbe used as an ex cuse to implement substantial restructuring (including job cutsand reassignments) and efficiency improvements to increase the profitability ofboth parts of the combined institution. In the absence of a merger, a significantrestructuring m ay not be undertaken because of the difficult and disruptive natureofth e change . The remaining employees may have serious morale problem s unless

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    25/46

    T H E E FFE CT S O F M E G AM E RG E RS ON E FFICIE NCY AND PRICE S 1 1 9

    there is a merger or other external event on which to blame the restructuring. Wecall thistheLow E f f ic iency Hypothes isand it predicts that the ex post improvementin eflBciency after the merger will be higher if either or both of merging firmshave low efficiency ranks prior to the merger, leaving room for both pa rts of theconsolidated bank to improve. To test this hypothesis, we specify as exogenousvariables the weighted ranks of the efficiency of both the acquirer and acquiredbanks, W 1(EFF1 ) and W 2 (EFF2 ), respectively. The coefficients of both of thesevariables are predicted to be negative under this hypothesis because low efficiencyindicates more capacity for the waking up to improve perfonnance. The W land W 2 w eights are needed because the contribution of each ofthe merging firmsto the improvem ent of the consolidated firm should be proportional to that firm'sshare of the conso lidated b ank. It may be difficult to distinguish the Low Ef f ic i e nc yHypothes isfrom theRelat ive E f f ic iency Hypothes isbecause both depend on almostthe sam e ex ante efficiency variables and neither hypo thesis nests the other one.In addition to theRelat ive E f f ic iency Hypothes isand theLow Ef f ic iency Hypoth-esis ,w e control for/investigate a number of other possible ex ante exp lanations ofmerger efficiency gains. We include W 2 , the measure of the relative size of theacquired bank to control for potential differences betw een mergers of equa ls andacquisitions of smaller banks. Mergers of equals, in w hich W 2 is relatively close

    to .5 , may increase the expected gainsfi-ommergers because there may be greatercost savings from the elimination of parallel managem ent structures. For exam ple,in the 1986 Wells Fargo-Crocker merger of near equals, large cost savings wereachieved by eliminating most of the C rocker management structure. Alternatively,acquisition of a relatively small bank also has potential advantages, such as aneasier integration of comp uter and accoimting systems and fewer intemal strugglesfor control. Note that these arguments are in addition to and separate from theRelat ive Ef f ic iencyandLow E f f ic iency Hypotheses ,w hich also specify W 2 .It is often argued by bank consultants that the greater the overlap in the localdeposit marke ts of merging banks, the greater should be the cost savings from themerger.^^ This cost-based argument reflects an expectation that there is likely tobe greater scope for consolidating back-office payment processing operations aswell as eliminating duplicate branch offices, tellers, and mid-level management

    when bank offices are geographically proximate. For this reason, we include thevariable OVERLAP the proportion of the deposits of the merging banks thatare in the same Metropolitan Statistical A reas (MSA s) or non- MS A counties - inthe regressions. H ow ever, prior academic studies of cost ratios and cost efficiencyusually did not find significant effects of OVER LA P (Berger and H ump hrey 1 992 ,Srinivasin 1 9 9 2 ,Srinivasin andWall1 9 9 2 ,R hoades1993 ,Pilloff 1996) .The degreeof overlap in the megamergers of the 1980s averaged 34%, although 70% of themergers have at least some dep osits in the same m arket. For example, Toers (1992) suggested that cost savings as much as35 % of the operating cost ofthe acquired bank may be expected for in-maiket mergers, but savings of o nly 15%may be expectedwith an out-of-market merger.

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    26/46

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    27/46

    THEEFFECTSOFMEGAMERGERSONEFFICIENCYA NDPRICESTABLE i n . E xante sources o fmerger-related changes inefficiency an dperfonnance(1981-1989; t-ratiosinparentheses)

    121

    Intercept

    W2(BFFWEFF2)

    Wl(EFFI)

    W2(EFF2)

    W2OVERLAP

    STINCOME

    HERF

    SHARERETAIL

    SCALE

    AdjustedR^Num. of Obs.

    A(V^)0.82(6.74)0.30(2.03)-

    ---0.15(2.46)-0.05*( -1.72)-0 .22(-0.28)-0 .39( -1.56)0.26(1.41)

    - 0 . 2 0( -2.30)- 0 . 7 1( -9.33)0.7157

    ACTT/^ )0.71(5.24)

    --- 0 . 5 5( -4.27)- 0 . 7 2( -3.51)0.23(1.28)-0 .03( -0.99)0.13(0.19)-0 .05( -0.22)0.04(0.25)-0.15*( -1.93)

    -0 .09(-0.58)0.7857

    A(7r/7r'')0.87(5.88)0.48(2.27)

    - 0 . 6 3(-4.90)-0 .17( -0.53)-0 .25(-0.94)-0 .04( -1.36)-0.11(-0.17)-0 .17( -0.73)0.09(0.58)- 0 . 1 6( -2.19)

    -0 .18( -1.21)(UO57

    AROA'0.64(0.82)1.36(1.22)

    - 0 . 5 5( -0.82)1.29(0.79)-1 .00(-0.70)0.04(0.25)- 8 . 1 2(-2.32)1.63(1.35)

    -0 .74(-0.90)-0 .50(-1.29)0.09(0.11)0.0957

    AROE'0.85(1.05)1.41(1.22)-0 .37( -0.53)1.35(a79)-1 .00( -0.67)-0 .07( -0.43)- 1 1 . 5 8( -3.18)1.20(0.95)

    -0 .70( -0.82)-0 .18( -0.44)-0 .43( -0.52)

    0.1757 ( ) Statistically significantly differentromzero at the 10%(5%) level, two-sided.

    difference in the pre-m erger efficiency p ositions ofthe merging banksW 2(EFF1EFF2), consistent with theRelat ive E f f ic iency Hypothes is . The point estimate of0.30 is statistically significantly different from zero, but it is not very preciselyestimated. It suggests that the typical acquiring bank brings the acquired part ofthe consolidated bank an estimated30%ofthe w ay toward its own pre-m erger effi-ciency rank.W etemp orarily defer discussion ofthe other variables in the regression.In the second coliunn ofTableHI, we examine the Low Ef f ic iency Hypothes is ,which emphasizesthepotential for improvement w hen either or both ofthe acquirerand acquired ba nks are inefficient. The variables W1 (EFF1)and W2 (EFF2) replaceW 2 (EFF 1- EFF 2 ) in (3), so that the efficiency ranks of acquiring and acquiredbanks are weighted by their respective importance to the consolidated firm. Thestatistically significant negative coefficients o f- 0. 5 5 and - 0. 7 2 on W 1(EFF1) andW 2 (EFF2 ), respectively, are consistent with the predictions oft he L ow E f f ic i e nc yHypothes is - ibe merger efficiency gains appear to be greater when there is more

  • 8/13/2019 Berger - 1997 - Effect of Megamergers on Efficiency and Prices - Evidence From a Bank Profit Function

    28/46

    12 2 J A LA LD .AKHAVEIN ETA L.room for the improvement of both the acquirer and acquired. The adjusted R^ ofthis regression is also improved relative to the prior equation (rising from 0.71to 0.78), further suggesting that the ex ante efiBciencies of both the acquirer andacquired are important, rather than just their difference.The third column of Table III show s a comp lete