Inexperienced banks and interstate mergers

Inexperienced banks and interstate mergers

Journal of Economics and Business 54 (2002) 313–330 Inexperienced banks and interstate mergers Jeffrey R. Hart a,∗ , Vince P. Apilado b a b Edwin L...

87KB Sizes 1 Downloads 42 Views

Journal of Economics and Business 54 (2002) 313–330

Inexperienced banks and interstate mergers Jeffrey R. Hart a,∗ , Vince P. Apilado b a

b

Edwin L. Cox School of Business, Southern Methodist University, Dallas, TX 75275, USA College of Business Administration, University of Texas at Arlington, Arlington, TX 76019, USA Received 17 October 2000; received in revised form 28 March 2001; accepted 29 June 2001

Abstract This paper implements an event study, operational performance study, and a financial market accuracy study with respect to interstate bank mergers in general and interstate mergers vis-à-vis the Riegle–Neal Interstate Banking Act. The focus is on banks inexperienced in the merger process. Through the use of a GARCH-M event study model our results show that interstate mergers have been positively received by financial markets with respect to stock price reactions. Also, these banks appear to be less profitable after their merger post-IBBEA. Lastly, the financial markets seem to have difficulty accurately predicting subsequent performance successes/failures of such mergers. © 2002 Elsevier Science Inc. All rights reserved. JEL classification: G21; G14; C40 Keywords: Interstate bank mergers; Event studies; GARCH-M

1. Introduction The majority of past bank merger research derives from a period that did not allow full and unfettered access to interstate banking. Each state had its own laws (with varying degrees of restrictions) regarding interstate bank mergers and acquisitions. However, after the passage of the 1994 Riegle–Neal Interstate Banking and Branching Efficiency Act (hereafter referred to as IBBEA), individual state laws regarding interstate banking are no longer valid.1 Therefore since September 1995 there have been no legal restrictions on out of state bank acquisitions, and therefore interstate mergers, previously deemed illegal, have become legal. ∗

Corresponding author. Tel.: +1-214-768-3150; fax: +1-214-768-4099. E-mail address: [email protected] (J.R. Hart).

0148-6195/02/$ – see front matter © 2002 Elsevier Science Inc. All rights reserved. PII: S 0 1 4 8 - 6 1 9 5 ( 0 2 ) 0 0 0 6 3 - 2

314

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

Prior to IBBEA there may have been potentially profitable mergers that could not occur. Due to this deregulation of interstate banking, the potential acquisition pool dramatically increased, yielding the possibility of additional and more suitable fits between acquiring banks and target banks. Yet this deregulation may also have engendered unsuitable mergers that should never have occurred, but were now possible. This paper looks at banks involved in interstate mergers that are inexperienced in the market for corporate control. The desire is to explore interstate merger deregulation (and thus increased competition) in the context of the relatively unskilled acquiring bank, which finds itself in surroundings where mergers have become noticeably the standard. This is accomplished by first examining stock market reactions to the bank merger announcements (an event study), by analyzing operating performance changes after the bank merger (an operational performance (OP) study), and by investigating the precision of stock market reactions with respect to the subsequent operating performance changes (a financial market accuracy study which combines the OP study with the event study). The first goal of this paper is to examine stock price reactions to mergers using a unique bank merger event study methodology. Previous bank merger event studies use the traditional methodology of estimating some variation of the standard market model prior to a merger followed by testing to see if the merger announcement created abnormal stock returns compared to the returns formed by the already estimated market model.2 The standard market model assumes linearity, homoscedasticity, and independence in stock returns. Yet Akgiray (1989) and Carroll and Wei (1988) contest the linearity assumption. Akgiray (1989) and Engle and Mustafa (1992) show that daily index stock returns and individual stock returns produce autocorrelated second moments; specifically these papers show how the daily stock returns exhibit autocorrelated conditional heteroscedasticity (ARCH) and generalized ARCH (GARCH) effects. In addition, French and Roll (1986), Jennings and Starks (1985), Patel and Wolfson (1984), and Roll (1984) all report autocorrelation in stock returns.3 Therefore past event study research has shown that daily stock returns may not follow the standard assumptions of traditional regression models. However, most past bank merger research incorporates the above assumptions. Violation of the assumptions may lead to inefficient estimators in the market model. Improving estimators’ efficiency is particularly significant in certain circumstances; for example, Pindyck and Rubinfeld (1998) state, “efficiency is desirable because the greater the efficiency associated with an estimation process, the stronger the statistical statements one can make about the estimated parameters.” They also state that when the objective of a model is to maximize the precision of predictions, efficiency is particularly desirable. An event study develops a regression model by estimating certain parameters that subsequently are used in predictions; thus event studies definitely involve circumstances where efficiency in parameter estimates is of importance. Therefore this paper incorporates a more statistically efficient event study methodology that accounts for the violated regression assumptions of the traditional methodology. The next goal of this paper is, through the use of an OP study, to determine whether the easing of interstate banking restrictions has actually improved profitability due to the fact that there are potentially more acquirer-target mixes or whether the increase in the potential merger pool generated the opportunity for more inappropriate mergers. Lastly, this paper combines the two empirical approaches implemented above (event study and OP study) to determine how well

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

315

financial markets assessed potential improvements in profitability of interstate bank mergers during 1994–1997. Thus we measure how the financial markets’ initial stock price reaction to a merger announcement ends up corresponding to the banks’ actual performance change. The goal of testing financial market accuracy is to determine whether financial markets’ stock price reactions to bank merger announcements accurately predict subsequent changes in the newly merged bank entities. Also we test whether there are any differences in how financial markets predict future performance changes pre- and post-IBBEA. Our results show that during the sampled years, banks not very active in the market for corporate control have been positively received by the financial markets with respect to the banks’ stock price reactions on the merger announcement dates. Support is also given for the premise that an alternative event study methodology is more statistically efficient than the traditional methodology. In addition, banks, which lack merger experience, appear to not improve profitability after the merger in the post-IBBEA environment compared to the pre-IBBEA environment. Lastly, the financial markets seem to have difficulty accurately predicting subsequent performance successes/failures of bank mergers when the banks are inexperienced in the market for corporate control. In fact, pre-IBBEA correlations are almost significant whereas post-IBBEA correlations are negative. The rest of this paper is organized as follows: Section 2 reviews relatively recent event study, OP study, and financial market accuracy study papers; Section 3 establishes data requirements; Section 4 details the three employed methodologies; Section 5 presents the results; and Section 6 summarizes the findings.

2. Literature review 2.1. Event studies The fundamental idea of bank merger event studies is to determine if there are any abnormalities in stock returns to the bidding (acquiring) banks, and/or to the target (acquired) banks, and/or to the combined entities around the merger announcement date (the event period). These papers first estimate the market model, then use the estimated market model’s parameters in determining the size and direction of the price changes. Prior event studies that look specifically at acquiring banks have found a variety of results that seem to depend on sample size, years studied, type of merger, event window length, etc. Although there is a difference in results among studies, the majority point toward the acquiring bank producing abnormally negative returns or toward no significant change in returns around the merger announcement date. For example, studies that have shown negative returns include the following: Baradwaj, Dubofsky, and Fraser (1992) analyzing 108 mergers from 1981 to 1987; Cornett and Tehranian (1992) investigating 30 mergers where the purchase price was at least $100 million from 1982 to 1987; Holdren, Bowers, and Mason (1994) assessing bank holding company acquirers from 1977 to 1987 pre-DIDMCA [Public Law 96–221, March 31, 1980, 94 Stat. 132]; Houston and Ryngaert (1994) examining 153 bank mergers from 1985 to 1991; Madura and Wiant (1994) examining 152 mergers from 1983 to 1987; Palia (1994) evaluating 48 mergers from 1984 to 1987; Siems (1996) testing 19 megamergers in

316

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

1995 (those with values exceeding $500 million); and Subrahmanyam, Rangan and Rosenstein (1997) analyzing 225 mergers from 1982 to 1990. The event studies of acquiring banks showing no significant effects to their stock price around the announcement of a merger include the following: Allen and Cebenoyan (1991) studying 138 mergers from 1979 to 1986; Holdren et al. (1994) analyzing post-DIDMCA and their entire sample taken as a whole (44 bank holding company acquirers) from 1977 to 1987; Zhang (1995) investigating 107 bank mergers from 1981 to 1990; and Zhang (1997) examining first time FDIC acquirers, first time unassisted acquirers, and repeat unassisted acquirers. It should be noted, that under certain circumstances, some event studies do find statistically abnormal positive returns to the bidding bank during their respective event period. For example, Zhang (1997) finds that repeat acquirers of FDIC assisted acquisitions have significant positive abnormal returns. Past event studies that look at target banks basically show the same results regardless of sample size, years studied, or event period examined. They reveal that the target receives an abnormal statistically significant positive return during its respective event period. Some of the prior research showing a positive return to the target banks are the following: Cornett and Tehranian (1992) examining mergers from 1982 to 1987; Houston and Ryngaert (1994) inspecting mergers from 1985 to 1991; Siems (1996) investigating mergers in 1995; and Zhang (1995) studying mergers from 1981 to 1990. Previous event study papers also examine net wealth changes in stock prices; thus they incorporate both stock returns of the acquiring bank and the target bank during the event period. Net wealth effect is usually calculated as some type of market value-weighted sum of the acquirer and the target. There is a variety in results of net wealth event study papers. Baradwaj, Fraiser, & Furtado (1990) analyze 53 mergers and find positive net wealth stock effects for hostile takeovers and no significant net wealth effects for nonhostile bank takeovers that occurred from 1980 to 1987.4 Houston and Ryngaert (1994) examine combined wealth effects from 1985 to 1991 and assess that on average bidding banks show abnormal negative returns, target banks show abnormal positive returns, and combined wealth effects are not significant. Looking at bank mergers cross-sectionally, they find positive net wealth effects for mergers where there is a high amount of market overlap (in-market mergers) and for mergers where both the target and acquiring banks are operating above the industry average with respect to return on assets. Also cross-sectionally, they identify a net positive association for bank mergers that use either a conditional stock financing arrangement or a preferred stock financing arrangement, yet find a negative association for bank mergers that use a traditional common stock fixed exchange rate method.5 Zhang (1995) looks at the 1981–1990 period and finds that the combined wealth effects of bank mergers create statistically significant net wealth. The wealth creation is due to efficiency gains in relatively small mergers and geographic diversification in relatively large mergers. In addition, he finds that wealth created by the mergers does not depend upon the method of financing (via cash or by security) used by the acquirer to take over the target bank.6 Pilloff (1996) examines stock price changes of 48 bank mergers that occurred from 1982 to 1991. Pilloff notes that consolidated stock price abnormal returns are greater when the pre-merger target’s total expenses are larger than the pre-merger acquirer’s total expenses. Further, he finds that the returns are greater the higher the pre-merger target’s noninterest expenses. In summary, past event studies generally find that the target receives positive abnormal returns during the event window. Prior analysis of acquiring banks has somewhat mixed results, but

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

317

basically show negative to no abnormal returns. And when the net wealth effect is analyzed, past papers by and large reflect a positive to no significant impact. Yet none of the past event studies utilize the type of methodology employed in this paper. 2.2. Operational performance studies OP approach basically looks at accounting data before and after the bank mergers to determine if there have been any significant changes in the merged banks’ performance. OP studies that look at changes in profitability will examine changes to the merged banks’ profitability ratios (i.e., return on assets (ROA) and/or return on equity (ROE)). There are numerous OP studies that look at profitability in the last few decades with a variety of methodologies and results. Frieder and Apilado (1983) look at 106 mergers of four bank holding companies from 1973 to 1977 and find that the mergers did improve profitability of the bank holding companies. Rose (1987) analyzes national bank mergers from 1970 to 1980 and shows that acquired banks are more profitable and less capitalized than their acquirers prior to their merger. Also, he indicates that profitability does not increase for the acquiring banks post-merger. Linder and Crane (1992) examine 47 New England merger transactions from 1982 to 1987 and find no improvements in operating ROA or growth in operating income. Cornett and Tehranian (1992) examine the 1982–1987 period and find improvements in operating cash flow returns due to increases in employee productivity, asset growth, loans, and deposits.7 They also find that ROE improves, but not ROA. Rose (1992) concentrates on 279 interstate mergers from 1980 to 1989.8 He finds no relationship between interstate acquisitions and changes in operating expense/revenue, changes in noninterest operating expenses/total expenses, changes in ROE, and growth in assets. Spindt and Tarhan (1993) separate bank mergers in 1986 into 79 small mergers where the resultant bank is less than $150 million in assets and 75 medium-to-large mergers where the resultant bank is more than $150 million in assets. They find significant improvements in ROE, margin, and employee cost for both sizes of mergers compared to a control sample.9 Akhavein, Berger, and Humphrey (1996) analyze 69 megamergers (i.e., those involving banks with assets exceeding $1 billion) from 1981 to 1989. They find that on average, mergers do improve profit efficiencies mainly due to a product shift by merged banks from securities to loans, which are more profitable. These profit efficiencies were more prevalent when the mergers consisted of one or both banks being inefficient prior to the merger. In summary, past empirical OP studies looking at profitability show mixed results. Yet, note that most of those cited studies look at 1980s data where the legal atmosphere was quite different than in the post-IBBEA period of the 1990s. 2.3. Financial market accuracy studies of bank mergers Financial market accuracy is measured by how the markets’ initial stock price reaction to a merger announcement ends up corresponding to the banks’ actual performance change. The literature of financial market accuracy combines the two traditional empirical approaches, event studies and OP studies, to determine how accurate the financial markets forecast

318

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

potential changes to the new combined bank. Previous bank merger literature that combines these two traditional empirical approaches is not as extensive as the literature that utilizes the two traditional empirical approaches separately. Thus so far there are only two noted papers (both examined earlier) on this subject, each with different results. Cornett and Tehranian (1992) examine 15 interstate and 15 intrastate bank mergers from 1982 to 1987. They find a correlation between stock market reactions and subsequent changes in cash flow and accounting performance measures. Pilloff (1996) examines 48 mergers from 1982 to 1991 to determine if there is an association between consolidated abnormal returns and subsequent performance changes. He finds that abnormal returns and subsequent performance changes are insignificant, which suggests that market expectations at the time of the merger announcement have not accurately predicted eventual outcomes.

3. Data We first look at bank mergers that involve publicly traded acquirers taking over publicly traded targets. The merger announcement date must have occurred between January 1, 1994 and June 1, 1997, and the merger must have eventually been consummated. A provision of IBBEA relevant to this study states that after September 29, 1995 bank holding companies can acquire banks in other states and the laws of states that allowed interstate banking only on a regional or reciprocal basis are invalidated. Therefore in this paper pre-IBBEA is defined as a merger announcement date prior to September 29, 1995 and post-IBBEA is a merger announcement date after September 29, 1995, since after this date IBBEA allowed unconstrained access to interstate banking. We want to analyze the effects of IBBEA, and thus we want only to examine a snapshot of time around IBBEA. The snapshot of time in this paper is 20 months prior to IBBEA and 20 months after IBBEA (January 1, 1994–June 1, 1997). The 20-month period is felt to be appropriately long enough to get a reasonable sample and short enough to not let other various changing regulatory and/or structural variables of the banking industry affect the results. The initial sample size is 333 mergers. The focus of our study is interstate mergers (state to state); therefore we take out all intrastate (within the state) mergers, leaving 158 mergers. We want to examine only the banks that we consider to be inexperienced in the market for corporate control. In this paper inexperienced is defined as when both the acquirer and the target must not be active in an interstate or an intrastate merger with another publicly traded bank two quarters through five quarters prior to the merger completion date. Also both the acquirer and the target must not be involved with any other interstate or intrastate merger one quarter to four quarters after the merger completion date for the newly combined entity. This leaves us with 28 mergers. We lose six more mergers because the targets do not trade on the NYSE, AMEX, or NASDAQ.10 The reason that two quarters prior to the merger is the starting point for analysis instead of one quarter prior to the merger is due to the fact that if the merger is completed within 45 days after the quarter ending, the acquired bank does not have to file their previous quarter’s financial statements. Also, 1 year (four quarters) after the merger completion date is used for the calculations in this study due to the fact that 50% of gains of a merger should be realized

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

319

within the first year (Rhoades, 1994). Therefore four quarters is chosen so as to be able to detect effects of mergers while maximizing sample size. The final sample size is 22, with 12 mergers announced pre-IBBEA and 10 merger announcements occurring post-IBBEA.11 4. Methodology 4.1. Event study Past research by Akgiray (1989), Bollerslev (1987), Elyasiani and Mansur (1998), and Laux and Ng (1993) has found that GARCH(1,1) most appropriately represents economic time series. Therefore a GARCH(1,1)-M model (as opposed to the above mentioned traditional market model) is utilized to account for the risk–return relationship, nonlinearity, and the heteroscedasticity problems found in daily stock price movements. To correct for the serial correlation found in daily stock price data, a residual lagged one-period is added to the GARCH(1,1)-M equation as another explanatory variable. Therefore the new model is specified as the following equation (hereafter referred to as the GARCH-M model): 2 2 Rt = β0 + β1 RMt + β2 (γ0 + γ1 + εt−1 + λ1 σt−1 ) + β3 εt−1 + µt

(1)

where Rt is the return on the bank on day t; β 0 is the constant term; β 1 is the slope parameter; RMt is the return on the S&P 500 on the announcement day; β 2 is the GARCH coefficient; γ 0 2 2 is the GARCH constant; εt−1 is last period’s residual squared; σt−1 is last period’s variance; εt−1 is last period’s residual; and µt is a white noise random error term.12 One year prior to the merger announcement date of daily stock returns is used to build the estimated parameter coefficients in Eq. (1). The next step in determining abnormal returns is tabulating the following equation: ARt = Rt − [GARCH-M model]

(2)

where ARt is the abnormal return of the bank on day t (where t = 0, the day the merger is announced); Rt is the actual return of the bank on day t = 0; and GARCH-M model is the return predicted by the previously estimated GARCH-M model. Abnormal returns are calculated for each acquirer, each target, and a combined market-capitalization average of the acquirer and target. 4.2. Operational performance study Next this paper investigates the effects of interstate banking via an OP study. It will expand upon prior OP bank studies by examining inexperienced banks and by investigating the influence of IBBEA on the performance of banks involved in interstate mergers. We first determine the change in profitability from pre-merger to post-merger for all mergers in the sample. Then we determine if there is any significant change in the merged banks’ performance whether they occurred pre-IBBEA versus post-IBBEA. The first step in studying the change in profitability is to calculate the statistical change of profitability measurements in all the mergers. To determine the change in profitability we

320

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

calculate the change in ROA. The following methodology is used to determine the statistical difference between the merged banks’ pre-merger performance versus their post-merger performance: ROABANKit = α1i + α2i ROAMKTit + α3i DUMit + εit

(3)

where ROABANKit is the combined entity weighted average (by asset size of the acquirer and target) ROA of acquiring bank i and target bank i at time t; α1i is the constant term; α2i is the coefficient of the weighted average return on assets of banks of similar asset size; ROAMKTit is the weighted average return on assets of banks of similar asset size as that of acquiring bank i and target bank i at time t, α3i is the dummy variable coefficient for the combined entity weighted average of acquiring bank i and target bank i at time t; DUMit is a dummy variable of the combined weighted average of acquiring bank i and target bank i at time t; pre-merger t’s are five quarters (t = −5) prior to the merger completion date to two quarters (t = −2) prior to merger completion date; post-merger t’s one quarter (t = +1) after the merger completion date to four quarters (t = +4) after the merger completion date; and εit is a random error term.13 The dummy variable DUMit is 1 if −5 ≤ t ≤ −2 (pre-merger) and 0 if +1 ≤ t ≤ +4s (post-merger). Thus the coefficient measures change in ROA due to the merger. The dummy variable approach is befitting since it not only determines the statistical change in profitability pre- and post-IBBEA, but it also measures the direction of the change. If the coefficient is negative then the pre-merger market-weighted combined ROA is less than the post-merger ROA, which suggests that the merger of the two banks improved profitability. Yet if the coefficient is positive then the pre-merger ROA is greater than the post-merger ROA, which suggests that the merger of the two banks actually decreased profitability. Next we average the results from above and test for the statistical significance to see if there are any differences between mergers that occurred pre-IBBEA versus mergers that occurred post-IBBEA. 4.3. Financial market accuracy study Financial market accuracy is determined by combining the OP study results with the event study results. Each merger’s measure of change in ROA pre- to post-merger (its dummy variable t-statistic) is matched up with its corresponding abnormal return ARi (combined market-cap weighted-average abnormal return of acquirer and target). To test the financial accuracy of the financial markets regarding these mergers, a similar methodology to that of Cornett and Tehranian (1992), and Pilloff (1996) is employed: N ¯ (ARi − AR)(t i − t¯t ) rARi ,ti =  t=1 (4)  N N 2 2 ¯ ¯ (AR − AR ) (t − t ) i i i t=1 t=1 i where rARi ,ti is the correlation coefficient of the combined weighted-average abnormal return of merger i with its corresponding profitability measurement variable t-statistic; ARi is the combined weighted-average abnormal return for merger i; and t is the merger’s corresponding dummy variable t-statistic for merger i.

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

321

5. Results Before we get into the results of the event study, we want to see if the use of the GARCH-M methodology is indeed a more statistically efficient method of determining abnormal returns in event studies. To do this we construct a total of 88 models. Thus we build 44 (22 acquirers and 22 targets) GARCH-M models as well as 44 (22 acquirers and 22 target) traditional models.14 Next we test each of the 88 equations for heteroscedasticity and autocorrelation violations. To test for heteroscedasticity we use the ARCH Lagrange multiplier (LM) test and to test for serial correlation we use the Ljung–Box Q-statistic.15 Table 1 shows the heteroscedasticity tests for the 44 traditional models, of which 19 (7 acquirers and 12 target equations) exhibit ARCH/GARCH effects. Table 2 displays that only one equation (a target) of the GARCH-M models exhibits ARCH/GARCH effects. With respect Table 1 Traditional event study heteroscedasticity and autocorrelation results Acquirer

p-value

Target

LM

Ljung–Box

Susquehanna Advantage Mellon First of America CFX TCF Synovus First Midwest ALBANK First Chicago Banc One National City

0.0006 0.4500 0.7498 0.3882 0.0496 0.6805 0.9595 0.0009 0.0154 0.3002 0.7127 0.4299

0.0000 0.0100 0.0550 0.7150 0.0190 0.0030 0.6220 0.0100 0.9190 0.9110 0.1330 0.0060

IBBEA Standard Federal Wells Fargo Magna Vermont Financial FNB Banc One CCB Pacific Century TCF Peoples Bancorp

0.0118 0.1856 0.3019 0.3689 0.3985 0.1916 0.0089 0.0645 0.0274 0.5735

0.1530 0.3650 0.0760 0.0500 0.0000 0.5150 0.5500 0.2450 0.4440 0.0000

p-value LM

Ljung–Box

Atlanfed Amity Glendale F&C Orange Great Lakes NBSC CF Bancorp Marble Financial NBD Premier Integra

0.5585 0.0000 0.0000 0.2963 0.0010 0.0163 0.3582 0.0223 0.0923 0.9391 0.0831 0.0095

0.0000 0.0000 0.0000 0.0000 0.0000 0.3310 0.0000 0.0000 0.0000 0.2110 0.0000 0.3990

Bell Bancorp First Interstate Homeland Eastern Bancorp West Coast Bancorp Liberty American Federal CU Bancorp Standard Financial Gateway

0.1049 0.9325 0.0000 0.9380 0.0000 0.2640 0.0001 0.0498 0.0160 0.0000

0.0100 0.6390 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.7410 0.0000

2 To test for heteroscedasticity we use the following ARCH Lagrange multiplier (LM) test: εt2 = θ0 + θ1 εt−1 +. . .+ 2 2 2 θp εt−p where εt is the residual squared from the regression; θ0 is a constant term; θ1 εt−1 is last period’s residual squared; p is the number of lagged squared residuals. The null hypothesis is that each of the lagged squared residuals 2 is equal to zero. The test is χ distributed with p d.f. To test for serial correlation we use the following Ljung–Box Q-statistic: QLB = T (T + 2) pj=1 (rj2 /T − j ) where rj is the jth autocorrelation and T is the number of observations. Statistical significance for both tests is a p-value of 0.05 or less.

322

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

Table 2 GARCH-M event study heteroscedasticity and autocorrelation results Acquirer

p-value

Target

LM

Ljung–Box

Susquehanna Advantage Mellon First of America CFX TCF Synovus First Midwest ALBANK First Chicago Banc One National City

0.9861 0.4811 0.4262 0.8863 0.2073 0.8140 0.6711 0.7831 0.3332 0.9917 0.2753 0.5145

0.1040 0.4150 0.3440 0.8990 0.3530 0.6830 0.7450 0.6500 0.0000 0.7960 0.8240 0.8310

IBBEA Standard Federal Wells Fargo Magna Vermont Financial FNB Banc One CCB Pacific Century TCF Peoples Bancorp

0.4181 0.9064 0.6367 0.9580 0.7826 0.8129 0.6719 0.4986 0.3653 0.1405

0.0880 0.7340 0.7130 0.5900 0.2410 0.8200 0.6670 0.9380 0.1750 0.7950

p-value LM

Ljung–Box

Atlanfed Amity Glendale F&C Orange Great Lakes NBSC CF Bancorp Marble Financial NBD Premier Integra

0.9292 0.8355 0.5363 0.0163 0.6282 0.7502 0.9134 0.8491 0.4880 0.9123 0.8939 0.7570

0.2050 0.0800 0.0280 0.0350 0.1550 0.5790 0.0490 0.1360 0.4340 0.7490 0.2230 0.6840

Bell Bancorp First Interstate Homeland Eastern Bancorp West Coast Bancorp Liberty American Federal CU Bancorp Standard Financial Gateway

0.7159 0.9546 0.8510 0.7858 0.8335 0.9083 0.4413 0.3498 0.8611 0.7230

0.3540 0.8660 0.4400 0.2560 0.0060 0.7250 0.3880 0.2690 0.0730 0.0130

2 2 To test for heteroscedasticity we use the ARCH Lagrange multiplier (LM) test: εt2 = θ0 + θ1 εt−1 + . . . + θp εt−p 2 2 where εt is the residual squared from the regression; θ0 is a constant term; θ1 εt−1 is last period’s residual squared; p is the number of lagged squared residuals. The null hypothesis is that each of the lagged squared residuals is equal to 2 zero. The test is χ distributed with p d.f. To test for serial correlation we use the following Ljung–Box Q-statistic: QLB = T (T + 2) pj=1 (rj2 /T − j ) where rj is the jth autocorrelation and T is the number of observations. Statistical significance for both tests is a p-value of 0.05 or less.

to autocorrelation problems, Table 1 shows that of the 44 traditional models, 26 (9 acquirers and 17 target equations) exhibit serial correlation problems. Whereas with the 44 GARCH-M equations in Table 2, only six (1 acquirer and 5 targets) display serial correlation effects. Thus, this assessment reflects that the GARCH-M’s event study results are more efficient than that of the traditional event study results. Table 3 presents abnormal returns for each acquirer, each target, as well as the mean abnormal returns (MARs) for the entire sample, pre-IBBEA, and post-IBBEA. Yet before the results of Table 3 are discussed it should be noted that all the same computations from Table 3 were also calculated using the traditional methodology (Table 4). When the GARCH-M (Table 3) methodology abnormal return results are compared to the traditional methodology abnormal return results (Table 4), there are no statistical differences between the numbers.16 This is expected since the gain in efficiency by using the GARCH-M model does not alter the statistical unbiasedness

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

323

Table 3 GARCH–M event study abnormal returns (%) and mean abnormal returns (MAR%) Acquirer

Abnormal return (%)

Target

Abnormal return (%)

Market-cap weighted combined abnormal return (%)

Susquehanna Advantage Mellon First of America CFX TCF Synovus First Midwest ALBANK First Chicago Banc One National City

1.0083 −1.8059 0.8941 −0.6800 1.2887 0.3256 −1.1670 0.5291 −0.8662 0.1987 −0.2623 −3.2966

Atlanfed Amity Glendale F&C Orange Great Lakes NBSC CF Bancorp Marble Financial NBD Premier Integra

12.5315 2.6469 8.0744 0.0660 2.7241 2.7143 2.1470 4.4731 2.6811 4.4992 0.2196 10.3403

2.0893 −0.7677 0.9334 −0.6606 1.4230 0.8721 −0.8477 0.8892 −0.4207 2.3258 −0.2621 0.7321

IBBEA Standard Federal Wells Fargo Magna Vermont Financial FNB Banc One CCB Pacific Century TCF Peoples Bancorp

0.5370 −0.9100 −0.0807 −2.7116 0.5107 0.2015 −2.0769 0.9900 −3.1526 1.3836

Bell Bancorp First Interstate Homeland Eastern Bancorp West Coast Bancorp Liberty American Federal CU Bancorp Standard Financial Gateway

−1.6677 0.0678 5.7091 2.0682 8.0823 20.7175 26.9539 8.8109 11.6739 −3.1762

0.1083 −0.4095 1.2260 −1.1471 1.3218 0.6894 4.6847 1.6248 −0.1618 0.7433

Entire sample: MAR(%)

−0.4156

Entire sample: MAR(%)

6.0162

Entire sample: t-statistic

−1.3654

Entire sample: t-statistic

3.9142

Pre-IBBEA: MAR(%)

−0.3195

Pre-IBBEA: MAR(%)

4.4265

Pre-IBBEA: t-statistic

−0.8322

Pre-IBBEA: t-statistic

3.9245

Post-IBBEA: MAR(%)

−0.5309

Post-IBBEA: MAR(%)

7.9240

Post-IBBEA: t-statistic

−1.0478

Post-IBBEA: t-statistic

2.5681

0.6812 ∗

2.4095∗ 0.5255



1.6473 0.8680

∗∗

1.7226

Abnormal returns and mean abnormal returns (MARs) are calculated as of the merger announcement date. The capital-weighted combined entities are determined by the weighted average market capitalization of the acquirer and target as of the merger announcement date. Pre-IBBEA is defined as a merger announcement made prior to September 29, 1995 and post-IBBEA is defined as a merger announcement made after September 29, 1995. ∗ t-statistics at the 1% significance level. ∗∗ t-statistics at the 5% significance level.

of the results. Yet (as shown above), what we do gain by using a more efficient model is the strength of the statistical statements we can make regarding the results shown in Table 3. Table 3’s results for the entire sampled period are consistent with prior research: the acquirers show a negative nonsignificant MAR of −0.4156% (t-statistic of −1.3654) whereas the targets

324

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

Table 4 Traditional event study abnormal returns (%) and mean abnormal returns (MAR%) Acquirer

Abnormal return (%)

Target

Abnormal return (%)

Market-cap weighted combine abnormal return (%)

Susquehanna Advantage Mellon First of America CFX TCF Synovus First Midwest ALBANK First Chicago Banc One National City

−1.2184 −1.9137 0.9342 −0.6891 1.1063 0.3534 −1.2119 0.5541 −0.9792 0.1747 −0.0435 −3.3943

Atlanfed Amity Glendale F&C Orange Great Lakes NBSC CF Bancorp Marble Financial NBD Premier Integra

7.7163 2.9698 7.9247 0.0685 2.1795 2.7366 3.6472 5.4607 2.7683 4.5141 −0.0809 10.1756

−0.3802 −0.7750 0.9725 −0.6694 1.2067 0.8987 −0.7438 1.0021 −0.5086 2.3211 −0.0435 0.6146

IBBEA Standard Federal Wells Fargo Magna Vermont Financial FNB Banc One CCB Pacific Century TCF Peoples Bancorp

0.6385 −0.6816 −0.4644 −2.6670 −0.1595 0.1895 −3.5144 0.9304 −3.1204 1.2545

Bell Bancorp First Interstate Homeland Eastern Bancorp West Coast Bancorp Liberty American Federal CU Bancorp Standard Financial Gateway

−1.7795 0.0812 4.9993 1.9821 6.2248 20.4357 27.6889 8.3596 11.2942 −1.6014

0.1683 −0.2911 0.7686 −1.1453 0.5244 0.6709 3.7533 1.5334 −0.2127 0.8535

Entire sample: MAR (%)

−0.6328

Entire sample: MAR(%)

5.8075

Entire sample: t-statistic

−2.0030

Entire sample: t-statistic

3.8871

Pre-IBBEA: MAR(%)

−0.5273

Pre-IBBEA: MAR(%)

4.1734

Pre-IBBEA: t-statistic

−1.3973

Pre-IBBEA: t-statistic

4.5903

Post-IBBEA: MAR(%)

−0.7594

Post-IBBEA: MAR(%)

7.7685

Post-IBBEA: t-statistic

−1.3860

Post-IBBEA: t-statistic

14.1775

0.4781 ∗

1.9811 0.3246



1.1390 0.6623



1.5926

Abnormal returns and mean abnormal returns (MARs) are calculated as of the merger announcement date. The capital-weighted combined entities are determined by the weighted average market capitalization of the acquirer and target as of the merger announcement date. Pre-IBBEA is defined as a merger announcement made prior to September 29, 1995 and post-IBBEA is defined as a merger announcement made after September 29, 1995. ∗ t-statistics at the 1% significance level.

show statistically positive MAR of 6.0162% (t-statistic of 3.9142). It appears that the market has approved of the mergers due to the fact that the combined market-cap weighted MAR is a significant 0.6812% (t-statistic of 2.4095). Thus for the entire sampled period, net wealth was created by these mergers.

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

325

Table 5 OP study results

Entire sample Pre-IBBEA Post-IBBEA

Mean of dummy variable coefficients

t-statistic

−0.7354 −1.1508 −0.2370

−2.1666∗∗ −2.2077∗∗ −0.6282

The mean of dummy variable coefficients is an arithmetic average of the α3i coefficient from the equation Eq. (3). Pre-IBBEA is defined as a merger announcement made prior to September 29, 1995 and post-IBBEA is defined as a merger announcement made after September 29, 1995. ∗∗ t-statistics at the 5% significant level.

Now looking at pre-IBBEA versus post-IBBEA, we see that the pre-IBBEA MAR for the acquirers is a nonsignificant −0.3195% and the post-IBBEA MAR for the acquirers is a nonsignificant −0.5309%. Yet again, the pre-IBBEA MAR for the targets is a significant positive 4.4265% (t-statistic of 3.9245) and the post-IBBEA MAR is a significant positive 7.924% (t-statistic of 2.5681).17 Table 5 presents the results of the OP study. Here we have the averages for the dummy variable coefficients and t-statistics for the profitability ratio ROA. The first interesting aspect of the entire time period examined is that the ROA dummy variable averages for the entire sampled period are significantly negative (t-statistic of −2.1666). What this means is that there was a statistically significant improvement in profitability post-merger. When separating the mergers pre-IBBEA versus post-IBBEA we see that pre-IBBEA mergers tended to improve profitability, yet post-IBBEA mergers showed no post-merger profitability improvements (a nonsignificant t-statistic). Table 6 displays the correlation coefficients (and their t-statistics) for the entire sample, pre-IBBEA, and post-IBBEA. On average, the financial markets have not been particularly statistically accurate (on the merger date) of predicting subsequent changes in profitability (based on the ROA ratio measure) for the combined bank entities. The correlations are indeed positive, but not significant. When post-IBBEA is compared to pre-IBBEA, it appears that the financial markets have declined in their ability to assess potential benefits for banks not

Table 6 Financial market accuracy study results

Entire sample Pre-IBBEA Post-IBBEA

Correlations of market-weighted abnormal returns with dummy variables

t-statistic

0.2023 0.5489 −0.0993

0.9238 2.0766 −0.2822

The correlation of market-weighted abnormal returns with dummy variables is the correlation coefficient of the combined weighted-average abnormal return (of the acquirer and target) from the event study merger with its corresponding profitability measurement variable t-statistic from the OP study.

326

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

active in the market for corporate control. We have a pre-IBBEA correlation coefficient of 0.5489 (t-statistic of 2.0766) which is almost significant whereas the post-IBBEA correlation is a nonsignificant −0.0993.18

6. Conclusions This paper examines banks involved in mergers during the mid-1990s. The mid-1990s saw the inception of the Riegle–Neal Interstate Banking and Branching Efficiency Act, which deregulated the interstate bank merger environment. Within this new era of interstate mergers this study specifically looked at banks that were not very active in the market for corporate control. The results of this study first show the advantage of the use of the GARCH-M methodology versus the traditional methodology when employing an event study. This is due to the fact that the traditional methodology is associated with heteroscedastic and autocorrelated result-related problems and therefore restrictive in its statistical efficiency. The results of the event study show that acquirers tend to receive slightly negative insignificant abnormal returns on the merger announcement date whereas targets tend to show positive significant abnormal returns. For the entire time period examined the newly combined bank entities show significantly positive abnormal returns, thus net wealth was created. Also the market does not seem to statistically distinguish between pre- versus post-IBBEA mergers with regards to abnormal merger announcement returns. Next this paper analyzes the effects of IBBEA vis-à-vis bank merger profitability. Prior to this legislation there may have been potentially profitable mergers that could not occur. Due to this deregulation of interstate banking, the potential acquisition pool dramatically increased, spawning two possibilities: an increase in the number of potentially profitable fits between acquiring banks and target banks or an increase in their number of less proper fits. Results of this study suggest that interstate mergers for inexperienced banks from 1994 to 1997 have shown statistically significant improvements in profitability from pre- to postmerger. When pre-IBBEA mergers are compared to post-IBBEA mergers, pre-IBBEA mergers do appear to demonstrate improvements in profitability relative to post-IBBEA mergers. These results show that since the sampled data did not include acquirers that were actively involved in the market for corporate control (and therefore experienced in acquisitions), the increase in potential targets due to the deregulation may have prompted some unpromising mergers to occur. Thus, this suggests that the IBBEA may have brought on less profitable merger combinations for banks that are less experienced in the market for corporate control. Lastly this paper’s results indicate that markets do not on average accurately predict how well the newly combined entity will subsequently perform as measured by the changes in ROA; in fact, markets have become worse in the post-IBBEA world at their predictions. A potential explanation is that financial markets’ knowledge of these banks’ expertise in regards to merger activities is suspect. Further, given the basically unlimited size of the post-IBBEA interstate potential merger pool, the proficiency to accurately predict a merger’s success or failure may have actually decreased, as exhibited by this study’s results.

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

327

Notes 1. Public Law 103-328, September 29, 1994, 108 Stat. 2338. 2. The following is the standard market model: Rit = αi + βi RMt + εit where Rit is the actual returns to the banks (acquirer or target or combined entity) of stocks i at time t; αi is the intercept coefficient of the market model; βi is the slope coefficient of the market model; RMt is the actual returns to a market portfolio of bank stocks at time t; and εit is the random error term (Roll, 1977). 3. Besides the problems of serial correlation and heteroscedasticity, another possible problem with event studies using daily stock returns is the fact that it has been shown that daily stock returns are nonnormal (Fama, 1965; Mandelbrot, 1963). Although daily stock market returns have shown nonnormalities, Brown and Warner (1985) demonstrate that the nonnormality of daily stock returns is not as prevalent for cross-sectional mean excess returns, especially for large samples. Since most event study literature examines the cross-sectional sample mean of stock excess returns, the nonnormality question should not alter event study results. 4. Note that the banking industry does not encounter many hostile takeovers. Prowse (1997) looks at the various ways that banks have a change in corporate control. Prowse examines a sample of bank holding companies from 1987 to 1992 and finds that about 23% have a change of corporate control by means of a friendly merger, a hostile takeover, or management turnover. What is of interest here is that of this sample, only 1.7% of corporate control changes occurred by hostile takeovers. Whereas in Morck, Shleifer, and Vishny (1989), examination of 454 manufacturing firms about 9% of corporate control changes were hostile takeovers. Hence when it comes to the market for corporate control, banks play in a different arena than nonfinancial firms due to various regulatory issues. 5. Of their entire sample of 153 mergers from 1985 to 1991, Houston and Ryngaert (1994) find that approximately 80% of the banks use stock financing arrangements (as opposed to cash arrangements). Of this group about 26% use some type of conditional stock arrangement and 4% use preferred stock financing. 6. It is thought that the method used to finance investment projects by a firm may send signals to investors about the managers’ expected success of the proposed investment project, thus affecting the market value of the firm. See Myers and Majluf (1984) for more details on asymmetric information or signaling theories. 7. Cornett and Tehranian (1992) define operating cash flows as earnings before depreciation, goodwill, interest on long-term debt, and taxes. 8. In the 1980s interstate mergers became more prevalent as a result of many states relaxing out-of-state bank holding company entry. For example, in 1980 Maine was the only state that allowed out-of-state bank holding company acquisition whereas by 1990 45 states allowed some type of out-of-state entry (Rose, 1992). 9. Spindt and Tarhan (1993) define margin as net income before extraordinary items as a percentage of total revenue, and employee cost as total salaries and benefits as a percentage of total revenue.

328

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

10. It should be noted that some publicly traded banks are not listed nor trade on the National Association of Securities Dealers Automated Quotation System (NASDAQ). Thus they only trade over-the-counter (i.e., pink sheets). Although the NASDAQ stock market is technically an OTC market, in this paper OTC means that the stock is neither a listed security nor traded on the NASDAQ. 11. Stock data taken from CRSP and accounting data taken from COMPUSTAT. 12. Maximum Likelihood (Marquardt algorithm) is utilized to estimate the GARCH-M models. 13. It should be noted that, when available, one quarter to four quarters prior to the merger completion date are used instead of the two to five quarters prior to the merger completion date. 14. To determine the traditional market model coefficients we used the same data as with the GARCH-M models, except we stop 15 days prior to the merger announcement date. This standard practice of stopping days prior to announcement date is to account for any leakages of information about the merger that may distort the parameter estimates of the traditional model. The GARCH-M model does not have to do this because any leakages of information prior to the merger would be incorporated in the ARCH and GARCH terms. 15. For more details on the ARCH LM test see Engle (1982). For more details on the Ljung–Box Q-statistic see Box and Pierce (1970). 16. It should be noted that the traditional model (Table 4) did not statistically show a market-cap weight combined abnormal return at the 5% significance level whereas the GARCH-M model (Table 3) did. The traditional model did show a significant abnormal return at the 10% level. Yet, as stated in the text, there was not a statistically significant difference between the GARCH-M market-cap weighted combined abnormal return and the traditional market-cap weighted combined abnormal return. 17. To make sure that no one merger was distorting the targets’ abnormal return results, we took out the largest abnormal return pre-IBBEA and the largest abnormal return post-IBBEA, and performed the same tests. None of the results changed. 18. The pre-IBBEA correlation coefficient is significant at the 10% level. References Akgiray, V. (1989, January). Conditional heteroscedasticity in time series of stock returns: Evidence and forecasts. Journal of Business, 62(1), 55–80. Akhavein, J. D., Berger, A. N., & Humphrey, D. B. (1996). The effects of megamergers on efficiency and price evidence from a bank profit function. Review of Industrial Economics, 12, 95–135. Allen, L., & Cebenoyan, A. S. (1991, April). Bank acquisitions and ownership structure: Theory and evidence. Journal of Banking and Finance, 15(2), 425–448. Baradwaj, B. G., Dubofsky, D. A., & Fraser, D. R. (1992, December). Bidder returns in interstate and intrastate bank acquisitions. Journal of Financial Services Research, 5(3), 261–273. Baradwaj, B. G., Fraser, D. R., & Furtado, E. P. H. (1990, December). Hostile bank takeover offers: Analysis and implications. Journal of Banking and Finance, 14(6), 1229–1242. Bollerslev, T. (1987, August). A conditional heteroscedastic time series model for speculative prices and rates of return. Review of Economics and Statistics, 69(3), 542–547.

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

329

Box, G., & Pierce, D. A. (1970, December). Distribution of autocorrelations in autoregressive moving average time series models. Journal of the American Statistical Association, 65(332), 1509–1526. Brown, S. J., & Warner, J. B. (1985, March). Using daily stock returns: The case of event studies. Journal of Financial Economics, 14(1), 3–31. Carroll, C., & Wei, K. C. J. (1988, October). Risk, return, and equilibrium: An extension. Journal of Business, 61(4), 485–499. Cornett, M. M., & Tehranian, H. (1992, April). Changes in corporate performance associated with bank acquisitions. Journal of Financial Economics, 31(2), 211–234. Elyasiani, E., & Mansur, I. (1998, May). Sensitivity of the bank stock returns distributions to changes in the level and volatility of interest rates: A GARCH-M model. Journal of Banking and Finance, 22(5), 535–563. Engle, R. F. (1982, July). Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation. Econometrica, 50(4), 987–1007. Engle, R. F., & Mustafa, C. (1992, April/May). Implied ARCH models from options prices. Journal of Econometrics, 52(1/2), 289–311. Fama, E. F. (1965, January). The behavior of stock market prices. Journal of Business, 38(1), 34–105. French, K. R., & Roll, R. (1986, September). Stock return variances: The arrival of information and the reaction of traders. Journal of Financial Economics, 17(1), 5–26. Frieder, L. A., & Apilado, V. P. (1983, Spring). Bank holding company expansion: A refocus on its financial rationale. Journal of Financial Research, 6(1), 67–81. Holdren, D. P., Bowers, H. M., & Mason, W. J., Jr. (1994, May). Bank holding company acquisition activity: Evidence from pre- and post-deregulation period. The Financial Review, 29(2), 275–292. Houston, J. F., & Ryngaert, M. D. (1994, December). The overall gains from large bank mergers. Journal of Banking and Finance, 18(6), 1155–1176. Jennings, R., & Starks, L. (1985, Spring). Information content and the speed of stock price adjustments. Journal of Accounting Research, 23(1), 336–350. Laux, P. A., & Ng, L. K. (1993, October). The sources of GARCH: Empirical evidence from an intraday returns model incorporating systematic and unique risks. Journal of International Money and Finance, 12(5), 543–560. Linder, J. C., & Crane, D. B. (1992, February). Bank mergers: Integration and profitability. Journal of Financial Services Research, 7(1), 35–55. Madura, J., & Wiant, K. J. (1994, December). Long-term valuation effects of bank acquisitions. Journal of Banking and Finance, 18(6), 1135–1154. Mandelbrot, B. (1963, October). The variation of certain speculative prices. Journal of Business, 36(4), 394–419. Morck, R., Shleifer, A., & Vishny, R. W. (1989, September). Alternative mechanisms for corporate control. American Economic Review, 79(4), 842–852. Myers, S. C., & Majluf, N. S. (1984, June). Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics, 13(2), 187–221. Palia, D. (1994, December). Recent evidence on bank mergers. Financial Markets, Institutions & Instruments, 3(5), 37–59. Patel, J., & Wolfson, M. (1984, June). The intraday speed of adjustment of stock prices to earnings and dividend announcements. Journal of Financial Economics, 13(2), 223–252. Pilloff, S. J. (1996). Performance changes and shareholder wealth creation associated with mergers of publicly traded banking institutions. Journal of Money, Credit and Banking, 28, 294–310. Pindyck, R. S., & Rubinfeld, D. L. (1998). Econometric models and economic forecasts. New York: Irwin/McGraw-Hill. Prowse, S. (1997, Winter). Corporate control in commercial banks. Journal of Financial Research, 20(4), 509–527. Rhoades, S. S. (1994). A summary of merger performance studies in banking, 1980–93, and an assessment of the ‘operating performance’ and ‘event study’ methodologies. Staff studies 167. Washington, DC: Board of Governors of the Federal Reserve System. Roll, R. (1977, March). A critique of the asset pricing theory’s tests, part 1: On past and potential testability of the theory. Journal of Financial Economics, 4(2), 129–177. Roll, R. (1984, September). A simple measure of the effective bid-ask spread in an efficient market. Journal of Finance, 39(4), 1127–1139.

330

J.R. Hart, V.P. Apilado / Journal of Economics and Business 54 (2002) 313–330

Rose, P. A. (1987, November). The impact of mergers in banking: Evidence from a nationwide sample of federally chartered banks. Journal of Economics and Business, 39(4), 289–312. Rose, P. A. (1992, Fall). Interstate banking: Performance, market share, and market concentration issues. Antitrust Bulletin, 37(3), 601–630. Siems, T. F. (1996). Bank mergers and shareholder wealth: Evidence from 1995s megamerger deals (pp. 1–12). Federal Reserve Bank of Dallas, Financial Industry Studies. Spindt, P., & Tarhan, V. (1993, April). Performance economies associated with small and medium sized bank mergers. Journal of Banking and Finance, 17(2–3), 460–462. Subrahmanyam, V., Rangan, N., & Rosenstein, S. (1997, Autumn). The role of outside directors in bank acquisitions. Financial Management, 26(3), 23–36. Zhang, H. (1995, October). Wealth effects of U.S. bank takeovers. Applied Financial Economics, 5(5), 329–336. Zhang, H. (1997, October). Repeated acquirers in FDIC assisted acquisitions. Journal of Banking and Finance, 21(10), 1419–1430.