Mergers

Financial Institutions Center The Value Effects of Bank Mergers and Acquisitions by Steven J. Pilloff Anthony M. Santomero 97-07 THE WHARTON FINANCIAL INSTITUTIONS CENTER The Wharton Financial Institutions Center provides a multi-disciplinary research approach to the problems and opportunities facing the financial services industry in its search for competitive excellence. The Center's research focuses on the issues related to managing risk at the firm level as well as ways to improve productivity and performance. The Center fosters the development of a community of faculty, visiting scholars and Ph.D. candidates whose research interests complement and support the mission of the Center. The Center works closely with industry executives and practitioners to ensure that its research is informed by the operating realities and competitive demands facing industry participants as they pursue competitive excellence. Copies of the working papers summarized here are available from the Center. If you would like to learn more about the Center or become a member of our research community, please let us know of your interest. Anthony M. Santomero Director The Working Paper Series is made possible by a generous grant from the Alfred P. Sloan Foundation The Value of Bank Mergers and Acquisitions October 29, 1996 1 Abstract: The banking industry has experienced an unprecedented level of consolidation on a belief that gains can accrue through expense reduction, increased market power, reduced earnings volatility, and scale and scope economies. A review of the literature suggests that the value gains that are alleged have not been verified. The paper then seeks to address alternative explanations and reconcile the data with continued merger activity. Steven J. Pilloff is at the Board of Governors of the Federal Reserve System. Anthony M. Santomero is at the Wharton School, University of Pennsylvania. The views expressed do not necessarily reflect those of the Board of Governors or its staff. 1 1. Introduction Over the past decade, the banking industry has experienced an unprecedented level o f consolidation as mergers and acquisitions among large financial institutions have taken place a t record levels. In the last three years alone more t han 1500 mergers have occurred in the US market. To a large extent, this consolidation is based on a belief that gains can accrue through expens e reduction, increased market power, reduced earnings volatility, and scale and scope economies . Whether or not bank mergers actually achieve the expected per formance gains is the critical question. If consolidation does, in fact, lead to value gains, then shareholder wealth can be increased. On the other hand, if consolidating entities does not lead to the p romised positive effects, then mergers may lead to a less profitable and valuable banking industry. A reading of the literature suggests that the value ga ins that are alleged to accrue to the large and growing wave of merger and acquisition activity have not been verified. This has left the research community in a quandary. Has the indu stry followed a path of massive restructuring on a misguided belief of value gains' Is management in this sector just incompetent' Or, are they merely lying to shareholders about the effect of their activity on shareholder value' The present paper seeks to address these alternative explanations and reconcile the data with the empirical reality of continued merger and acquisition activity. Its not easy. The banking industry is following a path of widespread consolidation even while academics seek answers to why it is going on. The literature in the area is rich in terms of data and variation of techniques employed. But, the results are disappointing in that they leave the read er with no greater understanding of why banks are merging, or which banks are merging, or when this transformation of the industry will stop. In this paper we begin with a review of this literature on the effects of merger and acquisitio n activity. Then we attempt to put it in context, and we finally try to learn from the experience. W e argue that the literature, in fact, has done a good job in researching the question at hand. It is just Throughout this paper, the terms interchangeably. 1 1 merger and acquisition are used that we do not understand the answer. Gains by many measures are either small or non existent. The researchers in this field try to explain away the overwhelming message of the literature. In general, these explanations are rationalizations for the non existence of positive value outcomes from th e previously reviewed literature. They are not alterative, testable the ories, in and of themselves. In fact, these rationalizations can be seen as attempts to defend the approaches taken in the literature from professional economists or bankers who do not find the results credible. But, recently, a new thread of the literature has developed which we find somewhat more promising. The recent interest i n understanding individual cases, looking into the process of change for a particular merger and th e realized outcome from the event, seems potentially rewarding and rev ealing. Where these case studies will lead, and what we will learn that can be generalized is still an open question. Our review of the subject proceeds as follows. In Section 2, we begin with a synopsis of th e benefits that are supposed to accrue to a merger transaction. Section 3 outlines the major empirical methodologies used to investigate a nd measure these alleged gains. This is followed by a discussion of the findings reported in the major studies in the area, although we have, no doubt, left out a few. As noted above, the results obtained are overwhelmingly unsupportive of the value effects. Bot h accounting and event studies offer no evidence of value gains. The average merger has either n o effect on total firm value, or a slightly negative one. The reasons offered for these (non)results are enumerated and critiqued in Section 5, which we appropriately refer to as rationalizations. Then, in Section 6, we present a discussion of the newly emerg ing case approach. We attempt to explain why this research procedure is being employed and present both sides of the issue of whether such field studies will yield results. We end with Section 7, where our final thoughts are offered. 2. Traditional Views of the Value of Mergers and Acquisitions Merger and acquisition activity results in overall benefits to shareholders when the consolidated post-merger firm is more valuable than the simple sum of the two separate pre-merger firms. The primary cause of this gain in value is supposed to be the performance improvement following th e merger. The research for post-merger performance gains has focused on improvements in any one of the following areas, namely efficiency improvements, increased market power, or heightene d diversification. Several types of efficiency gains may flow from merger and acquisition activity. Of these , increased cost efficiency is most commonly mentioned. Many mergers have been motivated by a belief that a significant quantity of redundant operating costs could be eliminated through th e consolidation of activities. For example, Wells Fargo estimated an nual cost savings of $1 billion from its 1996 acquisition of First Interstate. 1 Consolidation enables costs to be lowered if scale or scope economies can be achieved. Larger institutions may be more efficient if redundant facilities and personnel are eliminated within th e post-merger organization. Moreover, costs may be lowered if one bank can offer several products at a lower cost than separate banks each providing individual products. Cost efficiency may also be improved through merger activity if the management of the acquiring institution is more skilled a t holding down expenses for any level of activity than that of the target. Bank merger and acquisition activity may also encourage improved revenu e efficiency in a manner analogous to cost efficiency. Some recent deals, such as the proposed acquisition of Boatmen' s Bancshares by NationsBank, have been motivated by potential gains in this area. 2 Accordin g to this view, scale economies may enable larger banks to offer more products and services, and scop e economies may allow providers of multiple products and services to increase the market share o f targeted customer activity. Additionally, acquiri ng management may raise revenues by implementing superior pricing strategies, offering more lucrative product mixes, or incorporating sophisticated sales Barton Crockett, "First Bank Claims Wells Overstates Deal Savings," American Banker (November 20, 1995). 1 Kenneth Cline, "NationsBank Sees Boatmen's Revenue Potential," American Banker (September 26, 1996). 2 and marketing programs. Banks may also generate gr eater revenue by cross-selling various products of each merger partner to customers of the other partner. T he result is supposed to be higher revenue without the commensurate costs, i.e., improved profit efficiency. The latter term in general refers to the ability of profits to improve from any of the sources noted above, cost economies, scop e economies or marketing efficiency. In a sense, it represents the total efficiency gains from the merger without specific reference to the separately titled efficiency improvement areas. Merger-related gains may also stem from increased market power. Deals among banks wit h substantial geographic overlap reduce the number of firms in markets in which both organization s compete. A related effect of in-market mergers is th at the market share of the surviving organization in these markets is raised. These changes in market structure make the affected markets mor e vulnerable to reduced competition. The increased market power of the surviving organization may enable it to earn higher profits by raising loan rates and lowering deposit rates. It should be noted that antitrust policies of the Federal Reserve and Department of Justice ar e designed to prohibit mergers with substa ntially anti-competitive effects. However, to the extent that a local market can be exploited by a merger which results in substantial market power, the potential gain could be substantial. Finally, mergers may enhance value by raising the level of bank diversifica tion. Consolidation may increase diversification by either broadening the geographic reach of an institution or increasing the breadth of the products and services offered. Moreover, the simple addition of newly acquired assets and deposits facilitates diversification by increasing the number of bank customers. Greater diversification provides value by stabilizing returns. Lower volatility may rais e shareholder wealth in several ways. First, the expected value of bankruptcy costs may be reduced. Second, if firms face a convex tax schedule, then expected taxes paid may fall, raising expected net 1 income. Third, earnings from lines of business wh ere customers value bank stability may b e increased. Finally, levels of certain risky, yet profitable, activities such as lending may be increased without additional capital being necessary. 2 Any one of these reasons for gains from mergers is sufficient, and different ones presumably are relevant in different circumstances. Not all mergers are expected to result in cost efficiencies, no r does each one result in higher revenue and/or diversificati on gains. However, for any specific merger to create value, at least one of these gains appears to be necessary to achieve it. A casual review of the press suggests most mergers assert cost advantages, while revenue and diversification gains are less often mentioned. When firms of dissimilar franchise merge, on the other hand, revenue efficiency or diversification are often the indicated reasons. Participants in in-market mergers trumpet cos t efficiencies, while others allege market power outcomes. In the U.S., the geographic expansion of the franchise of major super regionals, by contrast, often speaks of the transference of best practice in production and the stability of a large geographic distribution channel. Whether any of these gains are obtained is another matter. Bankers, and their investment bankers can allege all sorts of benefits. The key issue for the researchers in the area is whether or not these gains are observable. To address this issue, the literature has examined mergers in several different, but hopefully complementary ways. It is to this area that we now turn. 3. Traditional Approaches to Testing the Effects of Consolidation Most academic studies follow one of two approaches to estimate and evaluate the significance of merger-related gains. 1 3 The first compares the pre-merg er and post-merger performance o f See Santomero (1995) for a discussion of the benefits of diversification on shareholder value. This area of risk management has grown substantially as this review will illustrate. This is a key rationalization in favor of a move toward universal banking. See Saunders and Walters (1994). 2 Rhoades (1994) provides an excellent discussion not only of the two approaches, but of many of the studies conducted from 1980 to 1993 that adopt them. 3 institutions using accounting data to determine whether consolidation leads to changes in reported costs, revenue or profit figures. The strengths of this approach are that accounting performance can be directly measured and the data used are both easily obtained and well understood. The approach is also fairly straightforward. Data from both pre-merger and post-merger data are used in th e analysis and evaluated for evidence of a change in the performance around the merger activity . Proponents of this methodology argue that accounting data measure actual performance conditions, not investor expectations and are therefore likely to be somewhat more reliable than equity returns. However, studies of accounting data have several drawbacks. Although accounting data ar e designed to measure actual performance, they may be inaccurate in an economic sense. Data ar e based on historical figures and often neglect current market values. In addition, measured changes between the pre-merger and post-merger period may not be solely due to the merger. Other events may have occurred during the period that is being investigated which may more accurately account for the observable performance changes. Failing to account for such extraneous events may lead to improper conclusions regarding merger-related changes. The second approach to analyzing merger be nefits evaluates the stock market reaction to merger announcements. Proponents of this approach argue that in as much as it relies on market data rather than accounting figures, it more accurately conveys the implied value of merging two independent entities. In essence, they argue that a ccounting data are unreliable and the market s reaction is likely to be a better indicator of the real economic effects of the announced deal. Most studies examine the abnormal returns of acquirers and targets separately, but several papers analyze the total change in shareholder wealth. In such cases, the value-weighted sum of acquirer and target abnormal returns is the appropriate measure of overall gains stemming from merger and acquisition activity. Thi s measure quantifies the value creation that the market be lieves the merger will provide. Studies of the abnormal returns experienced by individual merger parties can not differentiate between the effect on value of consolidation and the effect of the wealth transfer embodied in the purchase price. As with any approach, market price studies themselves are not perfect. They have their share of detractors who focus most often on the tim ing of the analysis. While abnormal return studies, which use market prices, have the benefit of not relying on potentially misleading accounting data, most studies only measure returns during a short period around the merger announcement. The analysis is, therefore, based solely on market expectations of unrealized events. Studying abnormal returns during the post merger period, however, has its own set of problems. The interpretation of the few studies of abnormal returns that extend into the post-merge r period is unclear. Observed returns may be attributable to expected bank performance or to the actu al outcome which may be unrelated to the merger transaction under investigation. In all studies, it is unclear what event period is most preferred. How many days before th e announcement are necessary to capture the effect of information leakage without including too much unrelated noise' How many days after the announcement are appropriate to enable the market t o fully trade on information regarding the proposed transaction' If the event period is important , results may be influenced substantially by the window selected for analysis. Recently, a third approach has emerged in the literature which incorporates and extends the two basic methods of analysis. Several pape rs not only analyze the relationship between mergers activity and both changes in accounting figures and stock market returns, b ut go a step further. These studies measure the correlation between changes in accounting data and abnormal returns. In so doing, the ability of the market to accurately forecast subsequent performance changes is examined. Th e extended analysis addresses the question of whether the market is able to differentiate among mergers that ultimately achieve improved performance and those that fail to achieve gains. 4. A Review of the Published Literature Before examining the applications of the three approaches enumerated above, a practical issue arises, namely, what should be the level of the analysis. Two primary approaches to defining mergers exist in the literature. Mergers are defined at the bank level and at the holding company level. A bank-level merger occurs when previously distinct banks are consolidated into one institution . Consolidations of individual banks unde r the same holding company are often included in samples of mergers defined this way. Analyzing mergers at the bank level is appealing for several reasons. Not only have there been a great number of mergers at this lev el, but because the FDIC Report of Income and Report of Condition (call report) measure perfor mance at the bank level, data are easily obtained for these types of mergers. However, studying bank-level mergers centers the study on the impact of changing organizational structure. It does not clearly assess the gains brought about by ne w ownership which economists general view as the centerpiece of the analysis of mergers an d acquisitions. Therefore, most studies focus on mergers of holding companies. A merger at th e holding company level is defined by a change in ownership of a subsidiary bank or a group o f subsidiary banks. This type of merger is viewed in the same manner regardless of whether the newly acquired banks are consolidated into a single institution or continue to operate as separate entities under new ownership. By construction, analysis of this type of merger is particularly useful i n examining the effect of changes in ownership. A large portion of the emp irical work examining the benefits of mergers focuses on changes in cost efficiency using available accounting data. Berger and Humphrey (1992), for example, examine mergers occurring in the 198 0s that involved banking organizations with at least $1 billion in assets. The results of their paper are ba sed on data aggregated to the holding company level, using frontier methodology and the relative industry rankings of banks participating in mergers. Frontie r methodology involves econometrically estimating an efficient cost frontier for a cross-section o f banks. For a given institution, the deviatio n between its actual costs and the minimum cost point on the frontier corresponding to an institution similar to the bank in que stion measures X-efficiency. The authors find that, on average, mergers led to no significant gains in X-efficiency. 1 Berger and See Berger, Hunter, and Timme (1993) for an excellent discussion on the topic of financial institution efficiency. 1 Humphrey also conclude that the amoun t of market overlap and the difference between acquirer and target X-efficiency did not affect post-merger efficiency gains. In addition to testing X-efficiency, they also analyze return on assets and total co sts to assets and reach a similar conclusion: no average gains and no relation between gains and the relative performance of acquirers and targets . Non-interest costs yield significant results, but the findings are opposite of expectations that th e operations of an inefficient target purchased by an efficient acquirer should be improved. Akhavein, Berger, and Humphrey (1997) analyz e changes in profitability experienced in the same set of large mergers as examined by Berger and Humphrey. They find that banking organization s significantly improved their profit efficiency ranking after mergers. Ho wever, rankings based on more traditional ROA and ROE measure s that exclude loan loss provisions and taxes from net income did not change significantly following consolidation. DeYoung (1993) also utilizes frontier methodology to examine cost e fficiency and reaches similar conclusions as Berger and Humphrey. Cost benefits from mergers did not exist for 34 8 bank-level mergers taking place in 1986 and 1987. In addition t o the lack of average efficiency gains, improvements were unrelated to the difference between acquirer and target efficiency. However , DeYoung does find that when both the acquirer and target were poor performers, mergers resulted in improved cost efficiency. In addition to frontier methodology, the literature contains several papers that solel y employ standard corporate finance measures to analyze the effect of mergers on performance. For example, Srinivasan and Wall (1992) examine all commercial bank and bank holding compan y mergers occurring between 1982 and 1986. They find that mergers did not reduce non-interes t expenses. Srinivasan (1992) reaches a similar conclusion. Both of these studies focus solely on non-interest expenses resulting in an incomplete picture of the cost savings associated with mergers. In order to gain a complete view of bank costs, the total of interest and non-interest expenses must be examined. Various funding and investment strategies have different impacts on the two cost components. For example, an increase in purchased fund s raises interest costs, but lowers non-interest costs. Therefore, to avoid attributing efficiency gains to changes in funding methods or investment choices, total costs must be evaluated. 1 Toward this end, Rhoades (1993) c onducts a thorough examination of in-market mergers taking place between 1981 and 1986. He re gresses the change in several performance measures on control variables and a dummy variable differentiating banks that engaged in an in-market merger from those that did not. Rhoades also conducts several logit analyses where the dependent variables measure whether the efficiency quartile of a bank increased, decreased, or remained unchanged. In both sets of tests, cost reductions and efficiency gains were not significantly related to horizontal mergers. The 1993 study is the most recent of a number of studies on the subject by this author. In a n earlier study, Rhoades (1987) examines the impact of mergers on the ratios of net income befor e extraordinary items to assets and non-interest expenses to assets. He runs probit analyses in which a dummy variable distinguishing non-acquired banks from banks acquired by multibank holdin g companies is the dependent variable. Performance measures and several control variables serve as the independent variables. Rhoades finds that neither income no r non-interest expenses were affected by merger activity. In Rhoades (1990), a similar study to Rhoades (1993) is conducted with 1 3 acquisitions involving billion dollar banks. Consistent with his other work, Rhoades finds n o performance effect due to mergers. The work of Linder and Crane (1992) is also noteworthy. They analyze the operatin g performance of 47 bank-level intrastate mergers that took place in New England between 1982 and 1987. Of the 47 mergers in the sample, 25 were consolidations of bank subsidiaries owned by the same holding company. The authors aggregate acquirer and target data one year before the merger and compare it to performance one and two years after consolidation. The performance of merged banks is adjusted by the performance of all non-merging banks in the same state as the mergin g entities. The results indica te that mergers did not result in improved operating income, as measured by net interest income plus net non-interest income to assets. The problems with ignoring total expenses and only analyzing non interest costs are discussed in Berger and Humphrey (1992). 1 Several studies find evidence of merger gains, but the results of these studies must be scrutinized carefully. Spindt and Tarhan (1993) find gains in their sample of 192 commercial bank merger s completed in 1986. Non-parametric tests compa ring the performance changes of merged banks with a group of matched pairs indicate that mergers led to operati ng improvements. The results, however, may be due primarily to economies of scale. The existing evidence in t he literature suggests that scale economies do exist for institutions holding less than $100 million in assets. 1 Spindt an d Tarhan's results are based on a sample that is dominated by mergers involving banks of this size. Because the results may be driven by economies of scale at small institutions, it is unclear whether their findings are relevant to large mergers--the transactions most severely transforming the banking industry. Chamberlain (1992) demonstrates the importance that sample selection can have in influencing the results of a merger study. Her sample consists of 180 bank subsidiaries that were acquired b y bank holding companies between 1981 and 1987. The unit of analysis is the individual target bank that experienced a change in ownership, but was not consolidated into another bank. For eac h merger, matched pair analysis is conducted. Pre-merger and post-merger perf ormance of the acquired bank are compared to those of a non-acquired bank from the same area and of similar size an d leverage. While Chamberlain finds evidence of overall gains when Texas mergers are omitted from the sample, the full sample yields no evidence of gains. Turning to studies of stock market reactions to merger announce ments, researchers also generally fail to find total gains from consolidation. Most abnormal return studies typically analyze target and acquirer returns separately. However, in order to measur e the overall anticipated gains resulting from a merger, the value-weighted average of bidde r and target abnormal returns must be analyzed. Most research on abnormal returns does not do this. Hannan and Wolken (1989) conduct a study of th e value-weighted abnormal returns experienced in 43 deals announced between 1982 and 1987. The authors find that, on average, total shareholder For good surveys of the literature on economies of scale, see Clark (1988) or Humphrey (1990). 1 value was not significantly affected by the announcem ent of the deal. The authors do, however, find that one determinant, target capitalization, cross-sectionally influenced expected synergistic gains . Target capital was negatively related to the change in total value. Houston and Ryngaert (1994) examine abnormal returns from four days before the target wa s initially declared a takeover candidate (by any bank) to the announ cement day. In their sample of 153 mergers announced between 1985 and 1991, acquirers suffered a loss in value and targets enjoyed a gain. However, there was no significant aggregate effect on the overall value of the tw o organizations. The amount of value that was created was highest when acquirers were stron g pre-merger performers and when substantial overlap e xisted. This relationship of value creation with the degree of overlap is consistent with the market expecting mergers best suited for improve d efficiency and/or increased market power to experience the greatest level of post-merger benefits. Madura and Wiant (1994) study abnormal returns of acquirers over a lengthy period following the merger announcement. They find that average cumulative abnormal returns of acquirers in a sample of 152 deals taking place between 1983 and 1987 were negative during the 36-month period following the merger announcement. Moreover, abnormal returns were negative in nearly ever y month. Acquirer losses around the time of the announcement may reflect a loss of wealth from an overly generous acquisition price. Negative abnormal returns in months after the announcement , however, are not likely to be due to the price. They seem more attributable to either the merge r achieving fewer benefits than projected, or the market revising downward its expectations for th e merger. The only serious study of the European market on this issue is the recent work by Cybo-Ottone and Murgia (1996). In it they analyze 26 mergers of European financial services firms (not jus t banks) taking place between 1988 and 1995 in 13 European banking markets. Their results ar e qualitatively similar to much of the analysi s conducted on American banking organizations. Average abnormal returns of targets were significantly negative and those of acquirers were essentially zero. This pattern suggests that there was a transfer of wealth from acquirers to targets. Also comparable to mergers of American banks, the change in overall value of European financial firms at the time of the announcement was small and not significant. This pattern continued for at least a year. In the year following the merger, (exc luding the first 10 days after the announcement), the combined value of the acquirer and target did not change significantly. The findings of Zhang (1995) on U.S. data contradict those of most abnormal return studies . Among a sample of 107 mergers taki ng place between 1980 and 1990, the author finds that mergers led to a significant increase in over all value. Although both merger partners experienced an increase in share price around the merger announcement, target shareholders benefitted much more on a percentage basis than the acquiring shareholders. Cross-sectional results suggest that increases i n value were smallest when improved efficiency and increased market power were expected to have their greatest potential impact. Changes in value decreased as targets got larger relative to acquirers and as the amount of geographic overlap bet ween acquirers and targets increased. The latter finding is consistent with diversification creating value. Recently, several papers i ncorporate both approaches in the literature. The first of these studies is conducted by Cornett and Tehranian (1992) and examines 30 large holding company merger s occurring between 1982 and 1987. The authors find that profitability, as measured by cash flo w returns on the market value of assets, improved significan tly after the merger. This finding, however, must be viewed closely for several reasons. First , the market value of assets may be an inappropriate measure for standardizing income. It is defined primarily from the liability side of the balance sheet as the market value of common stock pl us the book value of long-term debt and preferred stock less cash. Given the nature of banks as financial intermediaries, it is unclear why deposits are not included in this liability-based definition. The appropriateness of subtracting cash holdings is also debatable. Cornett and Tehranian find that net income to assets, a more traditional measure of bank profitability, does not change by a significant amount. In addition, the findings of Cornett and Tehranian may also be partially driven by adjustin g performance by an improper benchmark. The authors use, as their peer group, a sample of bank s located throughout the country that were traded on either the NYSE or AMEX and that did no t merge during the sample period. This comparison set of banking organizations may not be relevant to the sample institutions which had significantly different regional characteristics. This problem is accentuated by a set of sample observations which has a number of questionable deals. 1 As a result, Cornett and Tehranian's findings of post-merger improvements relative to a benchmark may be due to the unique data used for the study. Cornett and Tehranian also examine value-weighted abnormal returns around the time of th e merger announcement. They find that the market responded to announced deals by raising th e combined value of the merger partners. The authors also find that changes in several performance measures, including cash flow returns on the market value of assets, were positively correlated with value-weighted abnormal returns. These relationships suggest that the market may have been able to accurately forecast the eventual benefits of individual mergers. Net income to total assets is not one of the variables that was correlated to value-weighted abnormal returns, however. Pilloff (1996), like Cornett and Tehranian, combines both approaches found in the literature to analyze a sample of 48 mergers of publicly traded banking organizations that merged between 1982 and 1991. His study improves upon Cornett and Tehranian by addressing many of the problems in that paper. First, results are based on traditional measures of performance that are appropriate for a study of banking organizations. Second, the performance of merging banks is compared to a more accurate benchmark that controls for geographic location. Third, and perhaps most importantly, the merger sample is larger with substantially fewer observations that are poorly suited for analysis. Pilloff obtains results that are consistent with the bulk of the merger literature. In general , mergers were not associated with any significant change in performance, suggesting that managers were unable to generate benefits from deals on average. Moreover, the mean overall change i n shareholder value was also quite small. Although there was no average change in either operating performance or shar eholder value, there was a great deal of variation among banks. Some mergers proceeded successfully and others resulted 1 See Pilloff (1996) for a detailed discussion of this problem. in failure. Likewise, the dispersion of changes in market values indicates that investors expecte d some mergers to increase and others to decrease firm value. A particularly important result of this paper is that merger-related changes in performance we re found to be unrelated to changes in market value at the time of merger announcement. Investors recognized that although the mergers would not create benefits on average, some would result in better performance and some would lead t o worse performance. However, the market was unable to distinguish between the two types of deals at the time the mergers were initially announced. In summary, most studies fail to find a positive relationship between merger activity and gains in either performance or stockholder wealth. This conclusion of no economic benefits holds across a wide variety of methodologies, samples, and levels of analysis, (individual bank or bank holdin g company). Moreover, there appears to be no relationship be tween changes in value at announcement and subsequent outcomes. Although Cornett and Tehranian find the existence a relationship, Pilloff provides stronger evidence for nonexistence. 5. Explanations and Rationalizations The general findings of the merger literature raises the question of why bank consolidation has been and continues to be so prevalent. When gains are not observable on average. Moreover, equity returns indicate that they have been difficult, if not impossible, to accurately forecast. Severa l answers to this question have been offered . The first of these is the most straightforward. Th e argument goes that the nature of the data may obscure the true economic impact of mergers. There are a number of arguments offered in this regard. It is often contended that the lack o f market data biases studies of accounting data. It is also argued that perhaps the post-merger tim e period is insufficiently long to capture the gains. Consolidation includes well-known transition costs, which may disguise operating gains achieved shortly after merger completion. Moreover, man y 1 performance gains may take time to either be achieved or be reflec ted in financial reports. Extending Discussions with a number of bank consultants and analysts indicated that as much as one half to three quarters of all merger-related cost savings should be achieved within a year. 1 the post-merger period doe s not alleviate the issue, because this solution includes its own problems. Beyond a certain point, the analysis of the merged firm relative to some peer control group relates less to the merger itself than to the idiosyncratic circumstances of each market or firm-specifi c strategy. There is a limit to the ex tent that the merger can be held accountable for the firm s relative performance. Studies of abnormal returns are al so not immune from these data problems. In these studies, the time period under consideration is typically short, but the exact interval is uncertain. With firm s engaged in discussions prior to announcement and investors speculating on potential acquisitio n targets, the researcher is frequently unsure about the exact period to analyze. In truth, the correct timing is likely to be deal-dependent and probably requires careful analysis in each and every case. However, to do so somewhat vitiates the benefits of a broad based cross-section analysis. Th e researcher hopes to gain some insight into the average benefits by looking at a large number of such events. If the time period of analysis varies across the transactions investigated, the standardization is lost. So, researchers have tended to examine fixe d pre- and post-merger time periods for purposes of gaining insight on market value effects. There is another issue that frequently comes up when critiquing the empirical studies, som e selection bias. Researchers often examine mergers in a way that excludes relevant data points in an effort to obtain a clean data set. Typically, only deals that involve banks that engaged in no othe r major merger activity during t he period surrounding the deal of interest are studied. Deals by banks that either had multiple mergers in the same year or eng aged in several deals over a given time period are often excluded from analysis. Requirements like these omit the very firms that are most relevant to analysis of consolidation. For example, banks such as Nations Bank and Bank One, just to name two, have been active acquirers, and are reputed to be especially efficient in the integration process. However, these very firms are often excluded from the samples investigated due to the selectio n criteria listed above. The consequence may be a sample that b iases results. These firms may be active participants that achieve substantial gains. T heir elimination may lead to the no-gain results reported above. At the very least, their omission from many of the samples may exclude interesting data from the analysis. Is this list of data problems sufficiently severe to lead one to discount the entire literature' We think not. While there are always data problems in every field, it is hard for us to imagine that the results reported above by so m any different researchers can be dismissed because of data impurities. While these problems should clearly be of concern and the effect of sample bias and timing need s ongoing scrutiny, it seems unlikely that those technical features alone can reverse the overwhelming quantity of evidence against gains in performance and value. Other researchers seem to concur with the conclusion that data problems are not masking rea l gains, and have recently offered stories to explain the inconsistency of the empirical evidence with the ongoing wave of bank mergers. The explanations center on two allegations concernin g managerial behavior. The first of these explanations is managerial hubris. In short, managers have an unrealistic view of their ski ll and talent, leading them to believe that they are capable of obtaining gains from the acquisition of another institution. However, in truth, they are no more capable than 1 others. Therefore, ex post results do not lead to superior performance. The argument , then, comes down to the contention that ex ante expecta tions of performance gains systematically exceed ex post performance. Variability in realized gains are not attributable to differences in managerial skill, but to unpredictable noise in the integration process. This rationalization of the empirical results i s appealing, at least to the research community that studies, but does not operate within, the banking industry. The argument states that the research is correct and the managers are systematicall y incorrect. However, the managerial hubris argument is a bit incredulous. It is based upon a view tha t managers are systematically blind to the reality of the situation, and that they do not observe th e 1 See Roll (1986) for a further discussion actual outcomes of their past actions or the actions of their peers. Further, it contends tha t shareholders and boards are oblivious to the reality of the situation and allow management to engage in activity that systematically has no positive shareholder value. This story might work for a one-time event, as in an industry where mergers and acquisitions are rare. In such circumstances, the truth of the situation is not apparent until after the action and there is no way to develop reasonabl e expectations of the ex post earnings effects. These co nditions do not seem consistent, however, with the U.S. banking market. With the number of mergers totaling 420 in 1995, 564 in 1994 and 477 in 1993 , it is hard to believe that shareholders are u naware of the consequences of managerial action in this area or the likely outcome of the next acquisition. In addition, it is not at all clear why managerial hubr is in the mergers and acquisitions area should be any greater than in other areas of bank activity. Perhaps managers have over-inflated egos and unrealistic views of their own talents. However, elsewhere in finance we presume that realit y impinges upon delusion and markets have a way of sorting these things out. Why is it the case that this does not occur here' The literature is silent on this point. A second rationalization offered for the absence of observed gains in the literature centers around agency problems. It is well known that there is a general lack of alignment between the interests of shareholders and managers. This point has received considerable attention in the recent corporate 2 finance literature. To many, the recent wave of mergers in the banking industry is one more piece 1 of evidence of this phenomenon. According to this view, merg ers are in the best interest of managers but not necessarily shareholde rs. The former engage in the activity to increase their own power and 1 Mergers and Acquisitions Roundup , American Banker (1996) The obvious strategy point here is Fama (1980), but some more recent references include Morck, Schleifer, and Vishney (1990). 2 remuneration, which are both assumed to be related to institutional scale. However, this behavior comes at the expense of the shareholders of the acquiring institutions who, in general, overpay for such acquisitions and suffer dilution if not decline, in firm value itself. Managers in the acquire d institution seem oblivious to the issue of agency problems. They seem able to exploit the interests of the acquiring manager by obtaining systematic gains to shareholders of acquired firms even while they are displaced in the process. Some may appropriately object to this characterization of acquired firm managers. They seem more than capable of obtaining golden parachutes and lucrative buyout agreements. However, the evidence is that, on average, they negotiate a price which increases shareholder value, even while the acquiring management is following another agenda of self-interest. This contrast between the managerial behavior of acquired and acquiring firms is problematic . It seems that if manager behavior is driven by self-interest, rath er than purely increasing market value, then the behavior of managers on both sides of the negotiation should be explainable using the same paradigm. This is indeed possible. Perha ps the gains from mergers are a reality. However, the side payments to the two groups of managers completely exhaust them, resulting in a neutral effect o n value and reported performance measures. This would be consiste nt with the data and the allegations of an agency problem. Continuing mana gement, according to this view, obtains the gains associated with running a larger organization, such as greater power, pr estige and remuneration, while departing management receives the present value of their gain in terms of a buyout compensation package. The mystery is why the costs of such a transaction are born by the acquiring organization, rather than split by the two groups of shareholders. Perhaps th ere is a winners curse , where acquiring firms bid up the price of other firms who are willing to sell out. However, this theory has not been investigated empirically, and remains a puzzle. In light of our inability to explain the results reported above using conventional economi c arguments some recent researchers have attempted to investigate the effect of merger activity using a different approach. They propose investigating merger events with close scrutiny of manageria l process and the extent to whic h plans are both put in place and accomplished. This approach, while still in a nascent state, argues that the performance gains that can be forthcoming from merger activity are often unique to a specific merger and difficult to estimate or even see using the standard crosssection tools heretofore employed. Rather, to understand and obtain estimates of the value gain s requires case by case examination of merger activity. This has led to a growing number of case specific studies which have tried to analyze specific mergers and document the outcomes and their associated efficiency gains or, in some cases, losses. At the core of this examination of merger results is a view that it is diff icult to go from the specific to the general. It argues that averages obscure the fact that many good mergers occur, which add efficiency gains, and that can be explained on a case by case basis. And, there are still others where gains, while potentially sizable, never accrue to the surviving firm s shareholder. In the latter case, the management may never follow through with plans to downsize operations or eliminate exces s capacity, or does not fully exploit the merger's potential. Yet, from a technical point of view th e efficiency gains are both feasi ble and estimable. Researchers pursuing this line of investigation hope to obtain results on managerial best prac tice by examining events on a case by case basis. They look for estimates of pure efficiency gains and a better understanding of why the standard cross-section analysis does not systematically find evidence of performance gains. 6. The Managerial Process Or Follow Through Approach Economists are understandably suspect of an examination of empirical regularities that begin s with an assertion that every case is unique. Nonetheless, the reality is that existing explanations of the data seem hard-pressed to explain the ongoing and substantial merger activity in the bankin g sector. It does seem that the research in this area is somehow missing an important factor in thi s ongoing trend. Whether the managerial process studies can unearth new insights or result in mere rationalizations for the fact that traditional empirical approaches are incapable of revealin g performance gains is still open to question. But certain facts suggest that this approach t o understanding the effect of mergers warrants at least some consideration. It is a well-documented fact that the variability of efficiency across banks of a specific size and product mix is far greater than any variation in efficiency that has been associated with either scale 1 or scope. Accordingly, there seems to exist an opport unity for revenue enhancement as well as cost efficiency gains from the merger of firms with different efficiency levels. In fact, the data reported by Berger and Humphrey (1992) suggest that on average acquiring firms are more efficient tha n acquired institutions. Yet, the data show no evidence of across-the-board ex post efficiency gains. The goal of this new field based research is to investigate why these efficiency gains do no t materialize in the average merger data by closely examining mergers on a case by case basis . According to this approach, one needs to examine each merger to estimate the gains or losse s associated with different parts of the m erger process in order to understand the gain or loss from the entire merger transaction. These studies investigate the fixed costs associated with the transition to a merged entity, the time and the effort needed to downsize, restructure and close redundan t production facilities, and management s attention to aggressively slashing expenses. They seek t o understand where merged institutions quickly align the ir practices to least cost procedures and where they do not. They separately estimate trans ition costs, and gains from long term operating efficiency accruing to the combined organi zation. In fact, one of the goals of these studies is to separate these two features of any merger. They also study the process and timing issues surrounding th e convergence to a single operating environment, as it is well known that the gains from the mergers materialize more quickly when the process of transition is completed early. 2 At the very least, these studies may help provide an explanation for the observed cross section behavior reported in Section 4 above. In addition, by concentrating on a wider array of data , institutional detail and idiosync ratic evidence surrounding the process, they may offer new insight to the research community. See Berger, Hanweck and Humphrey (1987), Berger and Humphrey (1991), Berger, Humphrey and Timme (1993) or Berger and Humphrey (1994) for a discussion of the overwhelming evidence. 1 2 Frei, Harker and Hunter (1995). Doubters may argue that this p rocedure is too idiosyncratic. By moving to the specific from the general, broad patterns are lost. However, as has been reported above, there are no broad patterns! This type of investigation may permit some to emerge. It may also be suggested that this approach is nothing more than documenting the reasons for the failure of the average merger. However, it is important to understand that the focus of the hubri s and the agency stories is that gains in fact do not exist. In the present context, the argument is quite different. Gains may exist, but they requir e specific know how and process talent to be achiev ed. The focus is on why some firms succeed while others fail, and it attempts to document the cause s of these differences. As such, it is quite a different approach than the ones enumerated above. If success ful, the literature may emerge from this exercise with a better understanding of both the merger process an d the difficulties it encounters. In any case, it seems worthy of investigation. Thus far there are three studies of which we are aware, of this case by case approach. The first reported by Calomiris and Karcuki (1996) e xamines nine bank merger cases. The second is a broadbased study at Wharton, including w ork by Frei, Harker and Hunter (1995) and Singh and Zollo. In both cases, the studies are on-goin g and results are scarce. The last is an on-going effort at the Fed, in which a similar project will produce results of nine in-depth analy ses and, hopefully, lessons learned from the exercise. While each of these studies is far f rom definitive and some have not even been formally released, the expectation is that they will begin to provide insights into the drivers of efficiency gains an d obstacles to achieving them. Already, participants in s uch studies have suggested that process design and merger planning play a large rol e in the short and intermediate term benefits from mergers, (Frei Harker and Hunter (1995), Frei and Harker (1996)). Each case study provides an insight into th e reasons for the empirical outcome, not merely a rationalization of them. Whether they will trul y provide fundamental economic insights is still an open question. However, as noted by Akhavein, Berger and Humphrey (1997), the data on X-efficiency suggests that many banks have cost levels that are 20% to 25% above those using best practice. Understanding firms practices in merge r consolidation may illustrate where gains are, which firms have exploited th em, and by extension which managers have superior abilities in this area. At this point, however, it is only a hope. 7. Summary and Conclusions The literature on the value of bank mergers and acquisitions presents a clear paradox. Empirical evidence indicates clearly that on average there is no statistically significant gain in value o r performance from merger activity. On average, acquired firm shareholders gain at the expense of the acquiring firm. This is documented over the course of many studies covering different time periods and different locations. It is true whether one looks a t accounting data or the market value of equity. Even more disturbing is that the market is unable to accurately forecast the ultimate success o f individual mergers, as indicated by the absence of any correlation between changes in accounting based performacne measures and stock market returns around the merger announcement. Yet , mergers continue. Indeed the merger wave that has swept the U.S. shows no sign of abating an d there is increasing evidence of a similar move in Europe. How can these facts be reconciled' The answers that appear in the literature are reasonabl e enough and follow a predictable p ath. Maybe the empirical work is wrong. But, this does not seem likely because there have been many studies by many authors with incredibly robust results. Maybe managers are suffering from self-delusion' But, it is hard to believe that a massiv e restructuring of the world financial structure is occurring because of a misguided view of one s own managerial talent. Maybe managers are lying, telling the shareholders that they are creating value but merel y expanding their own power base and compensation' But, here too, it se ems incredulous to argue that major institutions have engaged in massive acquisition plans with the blessing of shareholders tha t gain no value from the exercise. The truth may be that we do not understand what exactly is going on in the process of industry consolidation. The efforts that have recently begun to examine the managerial process and follow through that is associated with a specific merger deal may help shed some light on the matter. The key question is: can we explain the process and the recent expanded activity level from investigating merger deals on a case by case basis' Can this process lead us to an appreciation of the potentia l efficiency gains, and why they do not on average appear in the data that most economists examine' While we are not convinced, it appears a productive path to pursue, at least for a while. Bibliography Akhavein, J., A. Berger and D. Humphrey, The Effects of Megamergers on Efficiency and Prices: Evidence from a Bank Profit Function, Review of Industrial Organization 12, forthcoming, 1997. Amel, D., State Laws Affecting Commercial Bank Branching, Multibank Holding Company Expansion, and Interstate Banking, Working Paper, Board of Governors of the Federal System, 1991. American Banker, Mergers and Acquisitions Roundup, January 29, 1996. Beatty, R., A. Santomero, and M. Smirlock, Bank Merger Premiums: Analysis and Evidence, Monographs in Finance and Economics , Salomon Brothers Center, New York University, 1988. Berger A., D. Hancock, and D. Humphrey, Bank Efficiency Derived from the Profit Function, Journal of Banking and Finance 17: 317-47, April 1993. Berger, A., G. Hanweck and D. Humphrey, Competitive Viability in Banking: Scale and Scope and Product Mix Economies, Journal of Monetary Economics 20, 1987. Berger, A. and D. Humphrey, Bank Scale Economies, Mergers Concentration and Efficiency: The U.S. Experience, Working Paper 94-25, Wharton Financial Institutions Center, University of Pennsylvania, 1994. Berger, A. and D. Humphrey, The Dominance of Inefficiencies over Scale and Product Mix Economies in Banking, Journal of Monetary Economics 28: 117-48, August, 1991. Berger, A. and D. Humphrey, Megamergers in Banking and the Use of Cost Efficiency as an Antitrust Defense, The Antitrust Bulletin 37, 541-600, 1992. Berger, A., W. C. Hunter and S. G. Timme, The Efficiency of Financial Institutions: A Review and Preview of Research Past, Present and Future, Journal of Banking and Finance , vol. 17, p. 221-249, 1993. Calomiris, C. and J. Karenski, The Bank Merger Wave of the 1990s: Nine Case Studies, University of Illinois, 1996. Chamberlain, Sandra L., The Effect of Bank Ownership Changes on Subsidiary-Level Earnings, Unpublished Paper, The Wharton School, University of Pennsylvania, 1992. Clark, Jefferey A., Economies of Scale and Scope at Depository Institution: A Review of the Literature, Economic Review, Federal Reserve Bank of Kansas City, 16-33, 1988. Cline, Kenneth, "NationsBank Sees Boatmen's Revenue Potential" American Banker, September 26,1996. Cornett, Marcia Millon and Hassan Tehranian, Changes in Corporate Performance Associated with Bank Acquisitions, Journal of Financial Economics , 31, p.211-234, 1992. Crockett, Barton "First Bank Claims Wells Overstated Deal Savings" American Banker , November, 20,1995. Cybo-Ottone, Alberto and Maurizio Murgia, Mergers and Acquisitions in the European Banking Market, Working Paper, University of Pavia, Italy, 1996. DeYoung, Robert, Determinants of Cost Efficiences in Bank Mergers, Working Paper 93-1, Office of the Comptroller of the Currency, 1993. Fama, Eugene, Agency Problems and the Theory of the Firm, The Journal of Political Economy , 13, 1980. Frei, Frances and Patrick Harker, Measuring the Efficiency of Service Delivery Processes: With Application to Retail Banking, Working Paper 96-31, Wharton Financial Institutions Center, University of Pennsylvania, 1996. Frei, Frances, Patrick Harker and Larry Hunter, Performance in Consumer Financial Services Organizations: Framework and Results from the Pilot Study, Working Paper 95-03, Wharton Financial Institutions Center, University of Pennsylvania, 1995. Hannan, Timothy H. and John D. Wolken, Returns to Bidders and targets in the Acquisition Process: Evidence from the Banking Industry, Journal of Financial Services Research , 3, p. 5-16, 1989. Hawawini, Gabriel and Itzhak Swary, Mergers and Acquisitions in the U.S. Banking Industry: Evidence from the Capital Markets , Elsevier Science Publishers B.V., Amsterdam, the Netherlands, 1990. Healy, Paul M., Krishna G. Palepu and Richard S. Ruback, Does Corporate Performance Improve After Takeovers' Journal of Financial Economics , 31, p. 135-175, 1992. Houston, Joel F. and Michael D. Ryngaert, The Overall Gaining from Large Bank Mergers, Journal of Banking and Finance , 18, p.1155-1176, 1994. Humphrey, David B., Why Do Estimates of Bank Scale Economies Differ' Economic Review 76, p. 38-50, 1990. Linder, Jane C. and Dwight B. Crane, Bank Mergers: Integration and Profitability, Journal of Financial Services Research , vol. 7, p.35-55, 1992. Madura, J. and K. J. Wiant, Long-Term Valuation Effects of Bank Acquisitions, Journal of Banking and Finance, vol. 18, 1135-1154, 1994. Morck, R., A.Schleifer and R.Vis hney, Do Managerial Objectives Drive Bad Acquisitions' Journal of Finance, 45, 1980 Pilloff, Steven J., Performance Changes and Shareholder Wealth Creation Associated with Mergers of Publicly Traded Banking Institutions, Journal of Money, Credit and Banking, vol. 28, p. 294-310, 1996. Rhoades, Stephen A, The Efficiency Effects of Bank Mergers: Rationale for Case Study Approach and Preliminary Findings, Proceedings of a Con ference on Bank Structure and Competition, Federal Reserve Bank of Chicago, 1993b. Rhoades, Stephen A., The Operating Performanc e of Acquired Firms in Banking, in Wills, Caswell, Culbertson eds.: Issues after a Century of Federal Competition Policy , Lexington Books, Lexington, MA, 1987. Rhoades, Stephen A., Billion Dollar Bank Acquisitions: A Note on the Performance Effects , Mimeom Board of Governors of the Federal Reserve System, 1990. Rhoades, Stephen A., Efficiency Effects of Horizonta l (in-market) Mergers, Journal of Banking and Finance, vol 17, p.411-422, 1993a. Rhoades, Stephen A., A Summary of Merger Performance Studies in Banking, 1980-1993, and an Assessment of the Operating Performance and Event Study Methodologies, Staff Study 167 , Board of Governors of the Federal Reserve System, 1994. Roll, Richard, The Hubris Hypothesis of Corporate Takeovers, Journal of Business, 59, 1986. Santomero, Anthony M., Financial Risk Management: The Whys And Hows" Financial Markets, Instruments and Institutions 4,1995. Saunders, Anthony, and Ingo Walters, Universal Banking For the US , Oxford University Press,1994 Spindt, Paul A. and Vefa Tarham, The Impact of Mergers on Bank Operating Performance , Working Paper, Tulane University, 1993. Srinivasan, Aruna, Are There Cost Savings from Bank Mergers' Economic Review, Federal Reserve Bank of Atlanta, 17-28, 1992. Srinivasan, Aruna and Larry D. Wall, Cost Savings Associated with Bank Mergers, Working Paper 92-2, Federal Reserve Bank of Atlanta, 1992. Zhang, Hao, Wealth Effects of U.S. B ank Takeovers, Applied Financial Economics , vol. 5, p. 329336, 1995.