Managerial rights, use of investment banks, and the wealth effects for acquiring firms’ shareholders

Journal of Banking & Finance 34 (2010) 44–54

Contents lists available at ScienceDirect

Journal of Banking & Finance journal homepage: www.elsevier.com/locate/jbf

Managerial rights, use of investment banks, and the wealth effects for acquiring firms’ shareholders Weishen Wang a, Ann Marie Whyte b,* a b

Marshall University, One John Marshall Drive, Huntington, WV 25755, United States University of Central Florida, 4000 Central Florida Blvd., Orlando, FL 32816-1400, United States

a r t i c l e

i n f o

Article history: Received 7 January 2008 Accepted 5 July 2009 Available online 9 July 2009 JEL classification: G24 G30 G34

a b s t r a c t We examine the relation between managerial rights in acquiring firms and the decision to use an investment bank in merger and acquisition deals, and explore whether this relation impacts the wealth effects for acquiring firms’ shareholders. We find that acquiring firms whose managers have relatively strong rights are more likely to use investment banks to facilitate deals and are more likely to use reputable banks. The wealth effects to acquiring firms are inversely related to the use of investment banks when managerial rights are relatively strong. However, the wealth loss is mitigated when acquiring firms use reputable investment banks. Ó 2009 Elsevier B.V. All rights reserved.

Keywords: Investment banking Mergers and acquisitions Shareholder rights Wealth effects Managerial rights

1. Introduction Investment banks perform many important functions within the economy including underwriting securities, providing venture capital, conducting capital market research, and facilitating mergers and acquisitions (M&As). Theoretical models show that the general functions of financial institutions include reducing transaction costs (Benston and Smith, 1976), alleviating asymmetric information in imperfect markets (Leland and Pyle, 1977), and simultaneously producing information (Campbell and Kracaw, 1980). Consistent with these theoretical models, empirical studies document the positive contributions of investment banks in several functional areas such as underwriting and venture capitalism. Whether investment banks also add value as financial advisors in M&As remains unclear in the finance literature. We address this issue by analyzing how managerial rights in acquiring firms impact the decision to use an investment bank in M&As, and examining the value consequence of the interaction between managerial rights and the use of investment banks for acquiring firms’ shareholders. We draw on key studies from * Corresponding author. Tel.: +1 407 823 3945; fax: +1 407 823 6676. E-mail addresses: [email protected] (W. Wang), [email protected]. edu (A.M. Whyte). 0378-4266/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.jbankfin.2009.07.002

two strands of the existing literature in developing our hypotheses. The first strand examines the role of investment banks in M&As and is best exemplified by Servaes and Zenner (1996). They find that investment banks are more likely to be used in complex deals where their knowledge and expertise are needed to facilitate deal completion. The second strand focuses on whether antitakeover provisions, a measure of managerial rights, impact the market for corporate control. For example, Gompers et al. (2003) find that firms with more antitakeover provisions (ATPs) are more likely to undertake M&As. Masulis et al. (2007) find that acquirers with more ATPs experience lower announcement period abnormal returns compared to firms with less ATPs during the merger announcement period. Combining the two sets of studies we argue that managerial rights may influence both the propensity to use an investment bank in M&As and the resulting wealth impact for acquirers. Acquirers with strong managerial rights may be more likely to use an investment bank, since investment banks are skilled in deal completion and the fee contract between the investment bank and the acquirer stresses deal completion (Rau, 2000). On the other hand, investment banks need to maintain their reputation capital and may focus on adding value. We investigate whether managerial rights in the acquiring firm dominate the quality of the investment banks’ service.

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54

We find that strong managerial rights in acquiring firms are positively associated with the use of investment banks in M&As. This relation holds after controlling for deal and firm features such as transaction size, method of payment, type of transaction, and firm performance. Acquiring firms with relatively strong managerial rights are also more likely to use reputable banks. Further, the wealth effects to acquiring firms are inversely related to the use of investment banks when managerial rights are relatively strong. However, this effect is mitigated when acquiring firms use reputable investment banks. Our study makes two important contributions to the literature. First, we document the contingent nature of investment banks’ service: investment banks may help managerial empire building at the expense of shareholders when they are hired by acquiring firms with strong managerial rights. Stated differently, the participation of an investment bank, per se, does not have a negative impact on shareholders – the negative effect only arises when investment banks interact with managers with strong rights. Second, we show that the investment bank’s reputation capital does have an impact on the acquiring firm. It seems that investment banks use their reputation capital to counter the negative side of strong managerial rights to some extent. To the best of our knowledge, these results have not been documented in the literature. The remainder of the paper is organized as follows: Section 2 develops the hypotheses, Section 3 discusses the methodology and data, Section 4 presents empirical results, and Section 5 provides concluding remarks. 2. Hypotheses In this section, we develop hypotheses regarding the possible relation between managerial rights in the acquiring firm and the decision to use an investment bank, and the wealth impact of using investment banks. 2.1. Managerial rights and use of investment banks The separation between ownership and management allows managers to maximize their interests at the expense of shareholders (Jensen and Meckling, 1976). Managerial empire building is one manifestation of the agency problem, a well documented merger motive in the literature (Trautwein, 1990). Gompers et al. (2003) find that firms with more severe agency problems (relatively strong managerial rights) are more likely to pursue M&As. Masulis et al. (2007) find that deals that firms with strong managerial rights enter are associated with poor market reactions. Deals with poor market reactions may confront strong resistance from shareholders and may be more difficult to complete. Since the investment bank has skills to facilitate deal completion (Rau, 2000), managers with strong rights may be more likely to use investment banks. Managers may also be motivated to use investment banks because they receive personal benefits for granting business to investment banks. For instance, they may receive an allocation of initial public offerings (IPOs) underwritten by these investment banks.1 Since the use of investment banks typically costs millions of dollars in advisory fees, managers with strong rights may have more freedom to consume excessive perquisites, and can easily ex-

1 Piper Jaffray was fined $2.4 million and censured by the National Association of Securities Dealers (NASD) for allegations related to the allocation of IPOs from 1999 to 2001. In agreeing to the penalty, Piper Jaffray neither admitted nor denied the charges. The NASD alleged that Piper Jaffray developed a tiered system for awarding shares to the executives of corporate clients. A ‘‘0” ranking, according to NASD, meant ‘‘no stock for you”. (Forbes, Wall Street Fine Tracker, 07.15.04.)

45

pense these fees without the concern of being disciplined (Shleifer and Vishny, 1989). Managers not only receive direct cash bonuses for successfully completing the deal (Grinstein and Hribar, 2004), their total compensation may also be adjusted upwards as firm size increases (Bliss and Rosen, 2001). The pursuit of increased compensation may affect the market reaction and also the use of investment banks. Alternatively, when managerial rights are weak (i.e. shareholders have strong rights relative to managers), agency problems are mitigated since managers are more easily disciplined in such an environment. Therefore, we expect that when shareholder rights are strong, managers are less likely to enter value-destroying deals. They may choose not to enter a deal (Gompers et al., 2003) or enter less complex deals, since complex deals are not in the interests of shareholders (Servaes and Zenner, 1996). Indeed, Masulis et al. (2007) find that firms with weak managerial rights lose less during merger announcements. If the deal is relatively simple or fundamentally sound, it may be more easily executed making the use of investment banks less necessary. These arguments suggest that firms with weak managerial rights may be less likely to use investment banks in M&As. Thus, we formulate hypothesis 1 as follows: Acquiring firms with strong managerial rights are more likely to use investment banks to facilitate M&A deals. Admittedly, it is also possible that managers may still use investment banks to facilitate the deal even when managerial rights are weak; that is, when agency problems are less severe. Investment banks provide many valuable services including optimizing accounting, tax, and legal treatment for the deals (Stouraitis, 2003) and managers may still require their expertise to complete the deal. Thus, the impact of managerial rights on the use of investment banks is an empirical question.

2.2. Managerial rights and the use of reputable investment banks The participation of a reputable investment bank may have a certifying effect on the quality of the deal. This may be especially meaningful for deals motivated by managerial self-interest. A reputable bank enables the manager to make a strong case to obtain shareholder approval. In fact, Rau (2000) reports that first-tier banks complete a higher percentage of deals than third-tier banks. Thus, based on the bank’s expertise in completing the deals and their certification effect, we expect that strong managers are more likely to use reputable investment banks. This gives rise to hypothesis 2: Acquiring firms with strong managerial rights are more likely to use reputable investment banks because of the certification effect and their superior deal completion skills However, reputable banks may be unwilling to participate in deals that are not value maximizing. A bank may self-select to avoid the deals motivated by managerial self-interest because of reputation concerns. This mutual selection between the banks and the deals has been documented in the IPO market (Fernando et al., 2005). Indeed, reputation is one reason that IPO firms switch underwriters in their follow-up seasoned equity offerings (SEOs) (Krigman et al., 2001). Probably also because of reputation concerns, prestigious underwriters are associated with good quality IPOs (Beatty and Ritter, 1986; Carter et al., 1998). Investment banks may constrain managerial agency behaviors and advise them to increase value for shareholders. If self-selection and monitoring are more prevalent among top-tier investment banks, firms with strong managerial rights may use less reputable investment banks. However, it is likely that managers with strong managerial rights are likely to do more deals in the future, which may offset

46

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54

the reservations of top-tier investment banks about working on value-destroying deals. Consequently, firms with strong managerial rights are more likely to associate with reputable banks.2 2.3. Wealth effects and the interaction of managerial rights and use of investment banks The literature is mixed regarding the wealth impact of investment banks. Servaes and Zenner (1996) find that the use of investment banks is not associated with wealth enhancement for acquiring firms. Kale et al. (2003) document that the relative reputation between the bidder’s and target’s advisors helps increase the absolute wealth gain as well as the share of the total takeover wealth gain accruing to the bidder. Bowers and Miller (1990) find that in acquisitions where either the bidding or target firm uses first-tier investment banks, the total incremental wealth is greater than when neither firm employs a prestigious banker. Rau (2000) shows that top-tier investment banks are associated with less value increase or reduced value for acquiring firms than lower-tier investment banks. However, these studies do not explore the impact of managerial rights in the acquiring firms. We link managerial rights and the use of investment banks to the wealth effects for acquirers. We contend that the use of investment banks adds value for shareholders. Investment banks have a vested interest in preserving their reputation and may self-select to avoid deals motivated by managerial self-interests. Investment banks may play a strong monitoring role and serve as a check against bad managerial decisions. In support of this view, Kale et al. (2003) find that when using investment banks, acquiring firms are more likely to withdraw from those deals that are not in the interests of shareholders. This gives rise to hypothesis 3: The use of investment banks is value enhancing. Both acquirers with strong and weak managerial rights benefit from the use of investment banks in M&As. Alternatively, the nature of managerial rights in the acquiring firm may interact with the use of an investment bank and negatively impact the wealth effects. Investment banks may not serve the best interests of shareholders. Investment banks are compensated for the successful closure of a deal, not for the deal’s performance. McLaughlin (1990) reports that 70–80% of the advising fee is contingent on the completion of the deal and the fee is increasing in the deal value. The fee is attractive to both reputable and non-reputable banks and may be too lucrative to forego. Rau (2000) shows that the pursuit of fee income may be a priority even for top-tier banks. Advising a deal does not require an investment bank to engage its own capital as in IPOs. Thus, investment banks may enter the deal regardless of its wealth impact on shareholders, and this problem may be exacerbated when acquiring firms have strong managerial rights. The banks’ performance may also be affected by the standards set by the hiring managers. If the managers do not make the interests of shareholders a priority, it is difficult to expect the banks to do so. Thus, the interaction between managers with relatively strong rights and investment banks may simply reinforce the importance of completing the deal, with no explicit requirements to maximize shareholder wealth. Conversely, when managerial rights are weak, managers may emphasize the interests of shareholders when interacting with investment banks. The interacting process may provide investment banks with explicit or implicit wealth requirements beyond deal completion. The managers may reject the bank’s proposals during

2

We thank the anonymous referee for pointing this out.

the advising process if the proposals are not in the best interests of shareholders and force them to develop other alternatives which serve the interests of shareholders. These arguments give rise to hypothesis 4: The use of investment banks by acquirers with strong (weak) managerial rights has value reducing (enhancing) effects for acquiring firms’ shareholders in M&As. The reputation of the investment bank may also impact the wealth effects. We conjecture that the use of more reputable investment banks will have a positive impact on the wealth effects. As noted previously, existing studies have shown that the bank’s reputation may impact the distribution of gains. For example, Kale et al. (2003) find that the relative reputation between the bidder’s and target’s advisors helps increase the absolute wealth gain as well as the share of the total takeover wealth gain accruing to the bidder. Bowers and Miller (1990) find that in acquisitions where either the bidding or target firm uses first-tier investment banks, the total incremental wealth is greater than when neither firm employs a prestigious banker. This gives rise to hypothesis 5: The use of more reputable investment banks should be positively related to the wealth effects However, Rau (2000) finds that top-tier investment banks are actually associated with lower announcement period returns than lower-tier investment banks in merger deals. Further, given our hypothesis that firms with strong managerial rights may be more likely to use reputable investment banks, the wealth effects are not clear cut. The ultimate outcome depends on which effect dominates. 3. Methodology and data We begin by identifying all firms in the Governance Index (GI) dataset created by Gompers et al. (2003), since we use the GI as our proxy for managerial rights. The index has values for the years 1990, 1993, 1995, 1998, 2000, and 2002. For each year in which the GI data are available (GI year), we retrieve financial data from Compustat for all firms in the GI year and the subsequent years prior to next GI year. For the years in which GI data are not available, we use the GI data for the closest year prior to the current year. For example, for firms with GI data in the year 1990, we download the financial data for the year 1990, 1991 and 1992. The GI for firms in the years 1991 and 1992 has the same value as in 1990. This parallels the procedure used in Gompers et al. (2003) and Masulis et al. (2007), and is reasonable because the GI data are relatively stable over short periods. This procedure results in 19,755 observations. For each firm with financial data, we obtain the M&A data from Securities Data Corporation (SDC) during the sample period ranging from 1991 to 2004. These deals take the form of mergers, acquisitions of main interest, acquisitions of partial interest, acquisitions of remaining interest, and acquisitions of assets, resulting in 18,949 observations. We merge the deals with the acquiring firm’s financial information from the prior year. This produces 10,900 observations. For these deals, we use standard event study methodology to estimate the cumulative abnormal returns (CARs) for acquiring firms using a sample of 10,767 observations. 3.1. Measuring managerial rights and bank reputation 3.1.1. Measuring managerial rights We use the GI as a proxy for managerial rights. The GI is the total number of antitakeover provisions adopted by a given firm and ranges from 1 to 24. Among these provisions, the most popular are

47

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54

those that stagger the terms of directors, provide severance packages for managers, and limit shareholders’ ability to act. Gompers et al. (2003) created the GI to proxy for the level of shareholder rights. They find that firms with higher GIs, that is, firms in which managers have strong rights, have lower value, lower sales growth, higher capital expenditures, and make more acquisitions. Complementing Gompers et al. (2003), Core et al. (2006) find that firms with strong managerial rights exhibit significant future operating underperformance. DeAngelo and Rice (1983) and Mahoney and Mahoney (1993) document that managerial entrenchment motivates the adoption of antitakeover amendments. More recently, Masulis et al. (2007) show that acquiring firms with strong managerial rights lose more in response to their M&A announcements. They argue that this may be driven by empire building of managers who are subject to weak monitoring from the corporate control market. Fahlenbrach (2003) examines the relationship between shareholder rights (measured by the GI) and the compensation contracts of executives, and finds that non-founder Chief Executive Officers (CEOs) receive higher total compensation, a higher annual increase in compensation, and have smaller fractional ownership if the managers have more power relative to shareholders. In the framework of agency theory, Jiraporn et al. (2006) link the GI with the firm’s probability to diversify and a large diversification discount. Harris and Glegg (2009) show that that the premium paid for shares is inversely related to the strength of shareholder rights while Autore et al. (2009) find that analyst recommendations are influenced by shareholder rights as measured by the GI. Chae et al. (2009) also document that corporate governance (measured in terms of the GI) affects payout policy. These studies consistently provide evidence that firms with higher GIs have greater agency costs. Thus, we use the GI to proxy for managerial rights relative to shareholders.3 For each year, we divide the sample equally into five groups based on the firms’ GI values. We use the GI quintile rank (GIRank) to replace the raw value of the GI in the analyses. The quintile rank captures investor’s perceptions of the firm’s agency problem relative to other firms in the same year. The rationale for this is that investors may only perceive the difference in managerial rights when firms have a relatively large difference in the number of antitakeover amendments. Previous studies have also used rank as the independent variable including Desai and Jain (1997), Mikhail et al. (2004), and Dong et al. (2006). 3.1.2. Measuring investment bank reputation The literature has examined the issue of investment bank reputation primarily in the context of the underwriting market. The banks’ reputation in the M&A market is seldom explored. In the IPO market, besides the position on the tombstone advertisement (Carter and Manaster, 1990), market share is the most widely used measure of bank reputation (Megginson and Weiss, 1991). In the merger advising market, Rau (2000) divides banks into different tiers based on their market share of advising services. In this study, we follow a similar logic. Specifically, we utilize the league tables

3 In order to increase our confidence in this measure, we also conduct robustness tests using other measures of the agency problem. We combine the GI data file with the director data file compiled by the Investor Responsibility Research Center (IRRC) and conduct a test of the relation between the GI and other variables related to agency costs; CEO duality (Core et al., 1999), management holdings (Yermack, 1996), and board size (Yermack, 1996). Consistent with the agency argument, both CEO duality and board size are positively associated with the GI. Management shareholding is negatively associated with the GI consistent with the notion that as shareholdings increase, the interests of the board converges with shareholder interests thereby reducing the agency problem. In all instances, the coefficient estimates are significant at the 1% level. Thus, we are confident that the GI is a good measure of managerial rights and we use GI in the remainder of the paper. In the interest of brevity, the results are not tabulated.

provided by SDC and use the market share of advising services to measure reputation. When a bank participates in a deal as a financial advisor, the transaction value (total value of consideration paid by the acquirer) is credited to that bank. For each year, the total transaction values of the bank are ranked. The bank with a rank of 1 has the highest total transaction value for that year. We select the top 25 banks each year and use their ranks to measure the reputation. As a robustness check, we also use market share in percentage and the number of deals completed as measures of bank reputation. We match each bank’s reputation measure in the current year with the deals the bank advises in the following year.4 3.2. Managerial rights and the decision to use an investment bank We first examine whether managerial rights in acquiring firms are associated with use of investment banks. For each deal, SDC records whether a financial advisor is used by target and acquiring firms, how many advisors each side uses, and identifies the advisor by name. For consistency with the existing literature, (Bowers and Millers, 1990; Servaes and Zenner, 1996; Kale et al., 2003; Rau, 2000) we do not differentiate specific functions performed by investment banks in this study. We do not consider how many banks the acquiring firm uses and when the banks enter and exit the deal. As long as SDC discloses that a financial advisor is used by the acquiring firm, we code the dummy variable ‘‘IB” with a value of 1 and 0 otherwise. To test the relation between managerial rights and the likelihood of using an investment bank, we set up the following Probit choice model:

ProbðIBÞ ¼ a þ b1  GIRank þ

X

bi  Controlsi þ e1

ð1Þ

where IB is a binary variable equal to 1 if the acquiring firm uses an investment bank and 0 otherwise, and the GIRank is the acquirer’s GI quintile rank for the year in which the deal is announced or 1 or 2 years preceding the announcement year. We use the following control variables (Controls) in the model. Ln(TA) is the natural log of total assets of the acquiring firm, ROA is the acquirer’s return on assets; Relative size is the transaction value divided by the acquiring firm’s total assets; Stockpay is the percentage of the deal value that is paid with the acquiring firm’s stock; Industry is equal to 1 if the acquiring and target firms share the first three digits of their primary Standard Industry Classification (SIC) code, and 0 otherwise; Hostile is a dummy variable equal to 1 if the target firm opposes the deal and 0 otherwise; Merger is a dummy variable equal to 1 if the deal is a merger and 0 otherwise; Major is a dummy variable equal to 1 if the deal is an acquisition of major interest in the target firm and 0 otherwise; Partial is a dummy variable equal to 1 if the deal is a partial acquisition and 0 otherwise; Remaining is a dummy variable equal to1 if the deal is an acquisition of remaining interest in the target firm and 0 otherwise. The underlying benchmark group is asset acquisitions. PercentIB is the percentage of deals in which the acquiring firm uses an investment bank in the 20 deals prior to the current deal. PriorUse is a dummy variable which has a value of 1 if the deal immediately before the subject deal undertaken by the same acquiring firm uses a bank, and 0 otherwise. The focus of Eq. (1) is to test the significance of the coefficient estimate, b1. Since the GIRank is a proxy for managerial rights, a

4 It should be noted that the league tables provided by SDC uses the bank’s current name. For instance, the deals advised by Salomon Brothers in 1992 are added to Citigroup for its year 1992 ranking, even though Salomon was not acquired by Citigroup until 1998. We exclude the deals advised by Salomon Brothers in 1993 since we do not know the ranking of Salomon Brothers in 1992. Our sample includes only deals advised by the banks ranked among the top 25 based on the previous year’s market share under exactly the same bank name.

48

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54

positive and significant b1 means that acquiring firms with strong managerial rights are more likely to use an investment bank. A negative and significant b1 indicates that acquiring firms with strong managerial right are less likely to use an investment bank. In order to examine whether strong managerial rights in the acquiring firm results in a preference for reputable banks, we also run regressions using three measures of bank reputation as independent variables: the ranking of the investment bank, the bank’s market share, and the number of deals completed by the bank in the previous year. We expect that firms with strong managerial rights are more likely to use reputable investment banks.

We use the abnormal returns around the announcements to capture the wealth effects. Abnormal returns are computed using the standard event study approach. All stock price data are obtained from the Center for Research in Security Prices (CRSP). Following Datta et al. (2001), we compute 2-day (1, 0) cumulative abnormal returns (CARs) for acquirers in response to deal announcements. We report CARs for all acquirers and for subgroups based on whether an investment bank is used in the deal and the GI quintile. We use the following regression model to investigate the wealth effects:

The second IV we identify is a dummy variable indicating whether the same acquiring firm used an investment bank in a deal immediately prior to the subject deal (PriorUse).6 The use of an investment bank in the prior deal may increase the probability that the acquiring firm will use an investment bank in subsequent deals. Thus, we expect that PriorUse will be positively related to the use of banks in the current deal. Given the exogenous nature of the two IVs, we do not expect them to be correlated with the market reaction in the second stage of the analysis. We provide specific tests regarding whether the IVs satisfy the two conditions required for IVs in the discussion of results below. To examine whether the bank’s reputation has a significant impact, we also regress the CARs on the measures of bank reputation for the sub-sample of deals which use at least one buy-side top 25 bank (if multiple banks are used in the deal, only the most reputable bank is considered). If reputable banks constrain managers’ agency behavior alleviating the wealth loss caused by self-interested management, we expect that the wealth effects should be higher when more reputable investment banks are used to facilitate the deal. Since the deals using reputable or non-reputable banks are a subset of all the deals which use investment banks, there is a sample selection issue. We apply traditional Heckman (1979) two-stage method and include the Inverse Mills Ratio (IMR) to correct for the sample selection as in Kale et al. (2003).

CAR ¼ a þ b1  GIRank þ b2  IB þ b3  GIRank  IB X þ bi  Controlsi þ e2

4. Empirical results

3.3. Wealth effects of the use of investment banks

ð2Þ

where CAR is cumulative abnormal return over the (1, 0) window and the other variables are as defined in Eq. (1). Our primary interest is to test whether the coefficient on the interaction between the GI and use of investment banks (GIRank*IB) is significant. 3.3.1. Controlling for possible endogeneity The use of investment banks, IB in Eq. (2), is a choice variable. If some factors that affect the firms’ choice to use an investment bank also affect the market’s reaction, IB may be correlated with the error term in Eq. (2). For example, if the use of investment banks is partially driven by managerial self-interests, such agency problems may affect the market reaction as well. Further, although we use the GI to measure the agency problem, we recognize that it may not capture all aspects of the agency issue such as the positive link between firm size and CEO total compensation. Thus, Eq. (2) may be affected by the omitted variable issue which would exacerbate the endogeneity problem and ordinary least squares (OLS) estimates may be biased. We apply the instrumental variable (IV) approach to address this possible endogeneity issue. A valid instrumental variable (IV) must satisfy two conditions. First, the IV must be correlated with the endogenous variable, in this case, the use of investment banks. Second, the IV itself should be uncorrelated with the market reaction in the second stage of the analysis (Eq. (2)). Using these criteria, we identify the first IV as the frequency with which investment banks are used in the deals prior to the subject deal. The fact that many prior deals use investment banks may provide justification for the CEO to use a bank as well for the current deal. In order to find the frequency that investment banks are used in prior deals, we first sort the observations in our sample based on their announcement dates. Then we obtain the percentage of deals using investment banks among 20 deals prior to a subject deal (PercentIB).5 We expect that PercentIB will be positively correlated with the use of investment bank for a subject deal.

5

The 20 deals are not necessarily undertaken by the same acquiring firm.

4.1. Descriptive statistics Descriptive statistics are reported in Table 1 for the sample as a whole and for subgroups based on the GI quintiles. Quintile 1 is the group in which the acquiring firms have the lowest average GI (GIQ1, weak managerial rights) and quintile 5 is the group in which acquiring firms have the highest mean value of GI (GI-Q5, strong managerial rights). The minimum and maximum values of the GI for each quintile are given in brackets. Panel A in Table 1 shows the results for acquiring firm features. The average GI for acquiring firms is 9.26, with a standard deviation of 2.77. Acquirers with strong managerial rights (quintile 5) have a higher level of total assets compared to those with weaker managerial rights (quintile 1). The average return on assets (ROA) for acquiring firms is 4.36%. Panel B presents deal features. The average transaction size is approximately 13% of the acquiring firm’s total assets (relative size). Acquiring firms use investment banks in 19% of deals. Acquiring firms in the lowest GI quintile use an investment bank in 16.0% of their deals compared to 19.0% among acquirers in the highest GI quintile. This is suggestive evidence that firms with strong managerial rights may be more likely to use investment banks. On average, approximately 13.5% of the deal value is financed with the acquiring firms’ stock. Acquiring and target firms share the first three digits of their SIC code in 39.5% of deals and only 0.5% of all deals are opposed by the target firm. Twenty five percent of the sample is outright mergers, 2.0% are acquisitions of major interest, 9.5% are acquisitions of partial interest, 2.2% are acquisitions of remaining interest, and 61.4% are asset acquisitions. In the prior 20 deals, approximately 17% involve an investment bank whereas a given acquiring firm uses an investment bank in 18% of its prior deals.

6 Using this as an IV leads to excluding those observations in which the acquiring firm is only involved in one deal. This reduces the sample size for the second stage regression.

49

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54 Table 1 Descriptive statistics. Variables

N

Mean

Standard deviation

GI-Q1 [1, 6]

GI-Q2 [7, 8]

GI-Q3 [8, 10]

GI-Q4 [10, 11]

GI-Q5 [12, 18]

Panel A: firm features GI 10,767 Ln(TA) 10,735 ROA 10,735

9.26 8.064 0.0436

2.77 1.6968 0.1321

5.21 7.88 0.04

7.51 7.83 0.03

9.09 8.03 0.04

10.63 8.24 0.05

12.95 8.29 0.04

Panel B: deal features Relative size 5378 IB 10,767 Stockpay 10,767 Industry 10,767 Hostile 10,767 Merger 10,767 Major 10,767 Partial 10,767 Remaining 10,767 Asset 10,767 PercentIB 10,767 PriorUse 9188

0.13 0.19 0.135 0.395 0.005 0.247 0.020 0.095 0.022 0.614 0.167 0.180

0.39 0.39 0.327 0.488 0.076 0.431 0.140 0.293 0.147 0.486 0.093 0.384

0.15 0.16 0.14 0.42 0.002 0.250 0.022 0.110 0.020 0.595 0.167 0.156

0.16 0.18 0.13 0.43 0.005 0.237 0.021 0.0978 0.020 0.623 0.165 0.173

0.124 0.203 0.15 0.38 0.010 0.273 0.019 0.087 0.026 0.594 0.168 0.187

0.11 0.206 0.13 0.38 0.004 0.229 0.021 0.095 0.023 0.631 0.167 0.192

0.12 0.19 0.13 0.36 0.008 0.249 0.018 0.086 0.020 0.626 0.169 0.187

This table shows descriptive statistics for the data. Panel A shows firm features and Panel B shows deal features. GI is the Governance Index for the acquiring firm in the year of the deal or 1 or 2 years prior to the deal; ln(TA) is the natural log of total assets of the acquiring firm; ROA is the return on assets of the acquiring firm; Relative size is the transaction value divided by the acquiring firm’s total assets; IB is a dummy variable equal to 1 if the acquiring firm uses an investment bank, 0 otherwise; Stockpay is the percentage of the deal value that is paid for with the acquiring firm’s stock; Industry is a dummy variable equal to 1 if the acquiring and target firms share the same 3-digit SIC code, 0 otherwise; Hostile is a dummy variable equal to 1 if the target firm opposes the deal, 0 otherwise; Merger is a dummy variable equal to 1 if the deal is a merger, 0 otherwise; Major is a dummy variable equal to 1 for acquisitions of major interest, 0 otherwise; Partial is a dummy variable equal to 1 for partial acquisitions, 0 otherwise; Remaining is a dummy variable equal to 1 for acquisitions of remaining interest, 0 otherwise; Asset is a dummy variable equal to 1 for asset acquisitions, 0 otherwise; Percent IB is the percentage of deals using investment banks among the 20 deals prior to a subject deal; PriorUse is a dummy variable equal to 1 if the acquiring firm uses an investment bank in the deal immediately preceding the subject deal and 0 otherwise. GI-Q1 through GI-Q5 are quintiles based on the GI. GI-Q1 is the quintile with the lowest GI (weak managerial rights) and GI-Q5 is the quintile with the highest GI (strong managerial rights). The values in brackets are the minimum and maximum GI value for each quintile.

4.2. Univariate tests of hypotheses Table 2 presents the CARs in response to the deal announcements. It presents preliminary results of the CARs over the 2-day window (1, 0) based on whether an investment bank is used in the deal. Interesting results emerge from the analyses. The deals that do not use investment banks (No IB) have an average CAR of 0.09%, significant at the 5% level. The deals that do use investment banks (IB) have an average CAR of 0.42%, significant at the 1% level. The difference is statistically significant at the 1% level. The negative association between participation of investment banks and the market reaction may be due to the fact that investment banks are more often used in complex deals (Servaes and Zenner, 1996). However, because of the lack of controls we cannot filter out the marginal contribution of investment banks in the deals based on these preliminary results. The average values of CAR (1, 0) across the five GI groups also suggest a negative association between the market reaction and

managerial rights in acquiring firms. When an investment bank is present, the group with the lowest GI (weak managerial rights) has an insignificant 2-day CAR of 0.28%. In contrast, the group with the highest GI (strong managerial rights) has a 2-day CAR of 0.49%, significant at the 5% level. When investment banks are not present, the negative association between managerial rights and the market reaction is seemingly not apparent – both the high and low quintile acquirers experience insignificant CARs. The finding of a negative association between managerial rights and the market reaction for the subset of firms that use investment banks is consistent with findings in Masulis et al. (2007). To test the existence of interaction effects, we conduct two-factor analysis of variance (ANOVA). The results are reported in the last column of Table 2. The first F-test with a value of 28.75 is testing the main effect of the use of investment banks. The F-test is statistically significant, indicating that the involvement of investment banks has a significant impact on the CARs. The significance of the F-test on the main effect of managerial rights (GIRank) confirms

Table 2 Market reaction to deal announcements. Variables

N (%)

Mean (standard error)

GI-Q1 [1, 6]

GI-Q2 [7, 8]

GI-Q3 [8, 10]

GI-Q4 [10, 11]

GI-Q5 [12, 18]

F-tests (ANOVA)

CAR (1, 0)

10,767 8706 (80.9%) 2061 (19.1%)

0.0004 (0.0010) 0.0010 (0.0010) 0.0028 (0.0031)

0.0002 (0.0008) 0.0003 (0.0008) 0.0001 (0.0025)

0.0009 (0.0009) 0.0023 (0.0009) 0.0023 (0.0023)

0.0021 (0.0007) 0.00001 (0.0007) 0.0100*** (0.0022)

0.0004 (0.0006) 0.0007 (0.0006) 0.0049** (0.0021)

IB

No IB

0.0001 (0.0003) 0.0009** (0.0004) 0.0042*** (0.0011) 0.0051*** (0.0009)

IB Difference

GIRank GIRank*IB

28.75*** (<0.001) 4.08*** (0.0026) 2.63** (0.0273)

This table presents 2-day cumulative abnormal returns (CAR (1, 0)) based on whether an investment bank is used in the deal. The results are also reported by Governance Index (GI) quintile. CAR (1, 0) is obtained using standard event study methodology. ‘‘No IB” refers to the deals in which the acquiring firm does not use an investment bank. ‘‘IB” refers to the deals in which the acquiring firms uses at least one investment bank. GI-Q1 through GI-Q5 are quintiles based on the GI. GI-Q1 is the quintile with the lowest GI (weak managerial rights) and GI-Q5 is the quintile with the highest GI (strong managerial rights). The values in brackets are the minimum and maximum GI value for each quintile. GIRank is the GI quintile rank for the acquiring firm in the year of the deal or one or 2 years prior to the deal. The F-tests are based on two-factor analysis of variance (ANOVA). The standard errors are in parentheses. * Indicates significance at 10% level. ** Indicates significance at 5% level. *** Indicates significance at 1% level.

50

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54

Table 3 Probit regression analysis of the choice to use investment banks by the acquiring firms. Variable

Intercepts GIRank Ln(TA) ROA

Use a bank

Bank reputation

Model 1

Model 2

Model 3 (IV approach)

Model 4 (ranking)

Model 5 (market share)

Model 6 (number of deals)

1.3023*** (0.0826) 0.0254*** (0.0098) 0.0561*** (0.0082) 0.1505 (0.1007)

1.8886 (1.8212) 0.0562*** (0.0181) 0.1392 (0.1090) 0.1244 (0.5299) 3.3078*** (0.1554) 0.0951* (0.0530) 0.0834** (0.0379) 0.2356 (0.1966) 0.5238*** (0.0473) 0.1521 (0.1269) 0.6069*** (0.0831) 0.4188*** (0.1038)

2.6300*** (0.1565) 0.0597*** (0.0153) 0.1638*** (0.0147) 0.0540 (0.1383) 3.2861*** (0.1832) 0.1248** (0.0620) 0.0273 (0.0438) 0.0471 (0.2261) 0.5682*** (0.0550) 0.1812 (0.1473) 0.5339*** (0.0937) 0.4402*** (0.1211) 0.4919** (0.2307) 0.4563*** (0.0500)

Yes 10,735 0.001

Yes 5997 0.219

Yes 4502 0.227

2.1539*** (0.1145) 0.0403*** (0.0097) 0.0442*** (0.0082) 0.4438** (0.1389) 0.0723*** (0.0278) 0.0669* (0.0355) 0.0909*** (0.0267) 0.0645 (0.0870) 0.0316 (0.0384) 0.1476 (0.1008) 0.2762 (0.0828) 0.0481 (0.0687) 0.0085 (0.1335) 0.1850*** (0.0306) 0.9108** (0.4220) Yes 1040 0.062

7.1333*** (1.3326) 0.2634** (0.1103) 0.2285** (0.0952) 3.6710** (1.8364) 0.6980** (0.3517) 0.5241 (0.4007) 0.4023 (0.3036) 0.6993 (0.9414) 0.0648 (0.4459) 1.1551 (1.1120) 1.9769** (0.9065) 0.7641 (0.8071) 0.0521 (1.5272) 1.9823*** (0.5544) 1.9823*** (0.5544) Yes 1057 0.04

3.4567*** (0.0354) 0.0303*** (0.0028) 0.0128*** (0.0025) 0.0473 (0.0527) 0.0019 (0.0082) 0.0017 (0.0102) 0.0007 (0.0078) 0.0434* (0.0227) 0.0378*** (0.0115) 0.2315*** (0.0291) 0.0936*** (0.0228) 0.0361* (0.0211) 0.1111*** (0.0376) 0.0054 (0.0090) 0.0349** (0.0141) Yes 1057 0.375

Relative size Stockpay Industry Hostile Merger Major Partial Remaining PercentIB PriorUse IMR Control years N Pseudo R-square

This table shows the results of testing whether managerial rights in the acquiring firm are associated with the use of an investment bank by estimating the following model:

ProbðIBÞ ¼ a þ b1  GIRank þ

X

bi  Controlsi þ e1

where IB is a binary variable equal to 1 if the acquiring firm uses an investment banks, 0 otherwise; GIRank is the Governance Index quintile rank for the acquiring firm in the year of the deal or 1 or 2 years prior to the deal. Controls include ln(TA) which is the natural log of total assets of the acquiring firm; ROA is the return on assets of the acquiring firm; Relative size is the transaction value divided by the acquiring firm’s total assets; Stockpay is the percentage of the deal value that is paid for with the acquiring firm’s stock; Industry is a dummy variable equal to 1 if the acquiring and target firms share the same 3-digit SIC code, 0 otherwise; Hostile is a dummy variable equal to 1 if the target firm opposes the deal, 0 otherwise; Merger is a dummy variable equal to 1 if the deal is a merger, 0 otherwise; Major is a dummy variable equal to 1 for acquisitions of major interest, 0 otherwise; Partial is a dummy variable equal to 1 for partial acquisitions, 0 otherwise; Remaining is a dummy variable equal to 1 for acquisitions of remaining interest, 0 otherwise; PercentIB is the percentage of deals using investment banks among 20 deals prior to a subject deal; PriorUse is a dummy variable equal to 1 if the deal immediately before the subject deal undertaken by the same acquiring firm uses an investment bank, 0 otherwise. The underlying benchmark group is asset acquisitions. IMR is the inverse mills ratio which corrects for the sample selection issue. Models 1, 2 and 3 are Probit regressions. Model 3 also includes the two instrumental variables (IVs). Models 4 and 6 are Poisson models with bank ranking and the number of deals advised in prior year as dependent variables, respectively. Model 5 is Tobit model with the market share (percentage) in the prior year as the dependent variable. The calculation of the inverse mills ratio (IMR) in Models 4–6 is based on using Model 3 as the first stage regression to correct for the sample selection issue. The decrease in sample size in Model 3 is due to the variable PriorUse, which is a dummy variable equal to 1 if in a prior deal, the same acquiring firm also uses an investment bank, 0 otherwise. Models 4–6 are based on this reduced sample in which at least one top 25 bank is used. * Indicates significance at 10% level. ** Indicates significance at 5% level. *** Indicates significance at 1% level.

that managerial rights are significant in explaining the market reaction. The F-test on the interaction between the investment banks and managerial rights (GIRank*IB) is also significant at the 5% level. These preliminary univariate tests indicate that managers with strong rights may interact with investment banks to affect shareholder wealth around the deal announcements.

4.3. Multivariate tests 4.3.1. Managerial rights and the choice to use an investment bank Table 3 presents the results of Probit models identifying the factors influencing the decision to use an investment bank. Model 1 regresses the likelihood of using the bank on managerial rights and includes controls for firm features and year effects only. Model 2 includes all the variables in Model 1 and also controls for deal features. Model 3 includes the IVs.

The results for Model 1 show that the GIRank is positively associated with the decision to use an investment bank. The coefficient estimate on GIRank is 0.0254, significant at the 1% level. The results for Model 2 also confirm this finding that strong managerial rights in the acquiring firm increase the likelihood that the firm will use an investment bank consistent with hypothesis 1.7 Model 3 includes all variables in Model 2 plus the two IVs. The results in Model 3 confirm that managerial rights are positively related to the use of

7 We run Probit models using CEO duality, management holdings, and board size as alternative measures of managerial rights. We find that management holdings are negatively associated with the use of investment banks, a result that is consistent with the agency explanation for the use of investment banks. Board size is also positively associated with the use of investment banks. However, CEO duality is not significantly related to the use of investment banks. Overall, these results validate our contention that the use of investment bank is directly related to the severity of the agency problem. In the interest of brevity, the results are not tabulated.

51

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54 Table 4 The wealth effect of the interaction between investment banks and managerial rights in the acquiring firms. Model 1 (OLS) Intercepts GIRank

Model 2 (OLS)

***

***

0.0177 (0.0038) 0.0007* (0.0004)

0.0177 (0.0038) 0.0007* (0.0004) 0.0005 (0.0015)

0.0016*** (0.0004) 0.0051 (0.0048) 0.0092*** (0.0032) 0.0067*** (0.0018) 0.0020* (0.0012) 0.0028 (0.0052) 0.0057*** (0.0016) 0.0027 (0.0044) 0.0009 (0.0019) 0.0040 (0.0040)

Yes 5378 0.027

IB

ROA Relative size Stockpay Industry Hostile Merger Major Partial Remaining

***

Model 4 (IV approach) ***

0.0016*** (0.0004) 0.0050 (0.0047) 0.0090*** (0.0033) 0.0067 (0.0018) 0.0020* (0.0012) 0.0029 (0.0052) 0.0056*** (0.0015) 0.0027 (0.0044) 0.0009 (0.0020) 0.0041 (0.0039)

0.0168 (0.0036) 0.0003 (0.0005) 0.0027 (0.0023) 0.0016* (0.0009) 0.0016*** (0.0004) 0.0052 (0.0047) 0.0092*** (0.0016) 0.0068*** (0.0018) 0.0020* (0.0012) 0.0032 (0.0056) 0.0054*** (0.0016) 0.0029 (0.0042) 0.0009 (0.0021) 0.0042 (0.0039)

0.0105 (0.0040) 0.0011 (0.0007) 0.0045 (0.0101) 0.0053** (0.0025) 0.0012** (0.0004) 0.0021 (0.0043) 0.0091* (0.0051) 0.0067*** (0.0018) 0.0026** (0.0012) 0.0050 (0.0052) 0.0042* (0.0022) 0.0052 (0.0037) 0.0001 (0.0023) 0.0042 (0.0033)

Yes 5378 0.027

Yes 5378 0.028

Yes 4462 0.032

GIRank*IB Ln(TA)

Model 3 (OLS)

PercentIB PriorUse Control years N Adjusted R-square

Model 5 (IV approach) 0.0109** (0.0043) 0.0011 (0.0007) 0.0006 (0.0113) 0.0052** (0.0025) 0.0011** (0.0004) 0.0023 (0.0042) 0.0080 (0.0054) 0.0069*** (0.0019) 0.0027** (0.0012) 0.0052 (0.0052) 0.0034 (0.0025) 0.0053 (0.0037) 0.0005 (0.0023) 0.0045 (0.0033) 0.0032 (0.0066) 0.0028 (0.0019) Yes 4462 0.033

This table shows the relationship between cumulative abnormal returns (CAR 1, 0) and the use of investment banks:

CAR ¼ a þ b1  GIRank þ b2  IB þ b3  GIRank  IB þ

X

bi  Controlsi þ e2

GIRank is the Governance Index quintile rank for the acquiring firm in the year of the deal or 1 or 2 years prior to the deal; IB is a dummy variable equal to 1 if the acquiring firm uses an investment bank, 0 otherwise; ln(TA) is the natural log of total assets of the acquiring firm; ROA is the return on assets of the acquiring firm; Relative size is the transaction value divided by the acquiring firm’s total assets; Stockpay is the percentage of deal value that is paid for with the acquiring firm’s stock; Industry is a dummy variable equal to 1 if the acquiring and target firms share the same 3-digit SIC code, 0 otherwise; Hostile is a dummy variable equal to 1 if the target firm opposes the deal, 0 otherwise; Merger is a dummy variable equal to 1 if the deal is a merger, 0 otherwise; Major is a dummy variable equal to 1 for acquisitions of major interest, 0 otherwise; Partial is a dummy variable equal to 1 for partial acquisitions, 0 otherwise; Remaining is a dummy variable equal to 1 for acquisitions of remaining interest, 0 otherwise. The underlying benchmark group is asset acquisitions. PercentIB is the percentage of deals using investment banks among the 20 deals prior to a subject deal, and PriorUse is a dummy variable equal to 1 if the acquiring firm uses an investment bank in the deal immediately preceding the subject deal, 0 otherwise. Models 1–3 are ordinary least squares (OLS) regressions and differ in the number of control variables. Models 4 and 5 are the second stage regressions using Model 3 in Table 3 as the first stage Probit regression. Model 5 includes the instrumental variables (IVs) PercentIB and PriorUse. The insignificance of their coefficients indicates that these two IVs are not correlated with the market reaction. The heteroskedasticity robust standard errors are in parentheses. * Indicates significance at 10% level. ** Indicates significance at 5% level. *** Indicates significance at 1% level.

investment banks (0.0597, significant at the 1% level) even after controlling for the two IVs. The coefficient estimate on the first IV, (PercentIB) is 0.4919, significant at the 5% level. The coefficient estimate on the second IV (PriorUse) is 0.4563, significant at the 1% level. The significance of both coefficients indicates that they are correlated with the use of investment banks, satisfying the first condition for being valid IVs. However, if the IVs are weakly correlated with the endogenous variable, the use of investment banks, the second stage estimates from the market reaction equation will be biased (Staiger and Stock, 1997). To test whether we have a weak IV problem, we regress the use of investment bank dummy variable on all variables in Model 3 and find the incremental F-statistic for these two IVs is 45.35 with p-value less than 0.0000. Our F-statistic is much higher

than the 11.59 benchmark for two IVs suggested by Stock et al. (2002) and Larcker and Rusticus (2008). It also exceeds the benchmark suggested by Staiger and Stock (1997). This alleviates the concern that the IVs may be weak. We will use Model 3 from Table 3 as the first stage regression for Models 4 and 5 in Table 4 when estimating the wealth impact of the use of investment banks using the IV approach. Models 4–6 in Table 3 test the association between managerial rights and investment bank reputation using different measures of bank reputation as dependent variables. In Model 4, reputation is measured with rank (based on the bank’s market share in the previous year) and a value of 1 indicates the most reputable bank. Model 5 uses the market share (in percentage) where a higher market share indicates a stronger reputation

52

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54

and Model 6 uses the number of deals completed in the previous year by the bank. A higher number of deals reflects stronger reputation. Due to the discrete and continuous nature of the dependent variables, Models 4 and 6 are Poisson models and Model 5 is a Tobit model.8 The three models present similar results; strong managerial rights are positively associated with more reputable investment banks. For instance, in Model 4, the coefficient estimate on GIRank is 0.0403, significant at the 1% level, indicating that acquiring firms whose managers have more rights relative to shareholders are more likely to use a bank with a high ranking in the advising market.9 Models 5 and 6 present similar results: relatively strong managerial rights are positively associated with market share and deals completed in the previous year. This is consistent with hypothesis 2 that because of the certification effect associated with the use of a reputable investment bank, firms with strong managerial rights are more likely to use reputable investment banks. The IMR in Models 4–6 corrects for the possible sample selection bias in these models, since we only use the sample of deals in which the acquiring firm uses at least one top 25 bank. Table 3 also presents other variables that significantly affect the firm’s choice to use an investment bank. The relative size measure is positively associated with the use of investment bank, consistent with Servaes and Zenner’s (1996) deal complexity argument. Investment banks are also more likely to be used in merger deals and acquisitions of remaining interest. We also see some evidence that large acquiring firms are more likely use reputable banks. 4.3.2. Use of investment banks and the wealth impact for acquiring firms Table 4 reports the results of estimating the cross-sectional regression model. CAR (1, 0) is the dependent variable. Models 1–3 use OLS estimates. Model 1 includes GIRank but does not include whether an investment bank is used in the deal so that we can isolate the impact of managerial rights. Model 2 includes both GIRank and IB and controls for deal and firm features. In both Models 1 and 2, GIRank carries coefficients of 0.0007, both significant at the 10% level. These results are consistent with Masulis et al. (2007) who find that acquiring firms with strong managerial rights lose more at the time of deal announcement. Interestingly, the impact of the investment bank itself (IB in Model 2) is not significant, suggesting that the use of an investment bank per se, does not have a significant impact on the abnormal returns. Model 3 includes the interaction term (GIRank*IB). As shown, the coefficient estimate on the interaction term is negative and statistically significant at the 10% level, confirming that it is the interaction between strong rights and use of investment bank that drives the results. Models 4 and 5 report the results using the instrumental variable approach with Model 5 including both instrumental variables. The models are estimated using 2-stage least squares (2sls) using Model 3 in Table 3 as the first stage regression. In Models 4 and 5, the coefficients on the interaction term (GIRank*IB) are negative and significant at the 5% level.10 The significance provides support for the interaction between managerial rights and use of investment banks: when investment banks are used by acquiring firms with rel-

8 Both Tobit and Poisson regressions produce similar results as straight OLS does in the testing. 9 Note that a high ranking corresponds to a low number. In other words, the highest ranking (indicating the strongest reputation) is a rank of 1. 10 We re-estimate the models using the raw GI values instead of GIRank and the results are qualitatively the same. In the interest of brevity, the results are not tabulated.

atively strong managerial rights, shareholders lose.11 Model 5 shows that the IVs are not significantly partially-correlated with the CAR (1, 0). This indicates that the second condition for a valid IV is satisfied; that is, the IVs are not correlated with the dependent variable CAR (1, 0) in the market reaction equation. The interaction items have significant coefficients in both the OLS regression (Model 3) and the 2sls regressions (Models 4 and 5). This is consistent with our assertion in hypothesis 4 that strong managerial rights interact with the use of investment banks to have a negative impact on the acquiring firm’s shareholders.12 It appears that the banks may mainly serve the purposes of managers instead of those of shareholders. The insignificance of the coefficients on IB in Models 4 and 5 indicates that the use of investment banks, per se, could be value neutral. Thus far, our results suggest that acquiring firms with strong managerial rights are more likely to use investment banks in M&A deals. Further, they are also more likely to use reputable investment banks to facilitate the deal either to certify the legitimacy of the deal or to facilitate successful completion of the deal. Does the reputation of the banks make a difference in terms of the wealth impact for acquiring firms’ shareholders? The results in Table 5 address this issue. In Table 5, the 2-day CARs to acquiring firms are regressed on the bank reputation measures and the interaction terms between bank reputation and the acquiring firm’s GIRank. Models 1 (a) and 1 (b) use investment bank ranking as the reputation measure whereas Models 2 (a) and 2 (b) use market share as the reputation measure.13 The (a) versions of the model do not control for sample selection whereas the (b) versions include the IMR to control for sample selection.14 In Model 1 (a), the coefficient estimate on the interaction term, GIRank*IBRank, is 0.0006, significant at the 10% level, showing that investment banks with top ranking can alleviate the value loss when interacting with firms with strong managerial rights. The results remain relatively unchanged when we include the IMR in Model 1 (b). Model 2 (a) directly confirms this point when market share is the measure of reputation. The coefficient estimate on the interaction term, GIRank*MarketShare, is 0.0004, significant at the 5% level, showing that a bank with high market share can help increase shareholder wealth when working with acquiring firms with strong managerial rights. The coefficient of the interaction item in Model 2 (b) is still positive, although it is not significant at traditional levels. These results provide some evidence that reputable banks can alleviate the value loss when working with acquiring firms with strong managerial rights. 11 We also estimate the model using long-term abnormal returns as the dependent variable to determine whether our central result persists in the long-term. The results show that although the GIRank continues to be significantly negatively related to the abnormal returns, the interaction term (GIRank*IB) is not significant. In the interest of brevity, the results are not tabulated but the complete results are available from the authors. We have more confidence in the short-term tests rather than the long-term tests, given the problems associated with the measurement of long-term abnormal returns. During the long-term period acquiring firms often enter multiple deals. This makes it difficult to accurately measure the long-term impact of one specific deal. Datta, Iskandar-Datta, and Raman (2001), for example, include only the first acquisition by a firm during the study period. Although this action achieves independence of observations, it measures the long-term performance for an acquiring firm after one specific deal with a bias. Subsequent deals also affect firm performance during long-time period. In contrast, announcement day returns avoid this problem to a great extent. 12 The significance of the interaction item in all models also makes the Hausman test unnecessary. The test compares whether 2sls is better than OLS. 13 We also use the number of deals to proxy for bank reputation but the results are not significant. In the interest of brevity, we do not report the results. 14 Since the testing sample includes only deals in which at least one top 25 bank is used, there is a sample selection issue. In the market reaction equation, the key variables of interest are the proxies for bank reputation. We consider these reputation measures to be exogenous and unrelated to the acquiring firm or deal features, observed or unobserved. The performance of the acquirers advised by the bank is unrelated to its market share (Rau, 2000).

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54 Table 5 Investment bank reputation and abnormal returns. Model 1 (a) Intercept GIRank IBRank GIRank*IBRank

**

0.0306 (0.0129) 0.0016 (0.0020) 0.0009 (0.0008) 0.0006* (0.0003)

Model 1 (b) 0.0039 (0.0169) 0.0027 (0.0021) 0.0012 (0.0009) 0.0006* (0.0003)

MarketShare *

GIRank MarketShare Ln(TA) ROA Relative size Stockpay Industry Hostile Merger Major Partial Remaining

0.0029*** (0.0010) 0.0282* (0.0164) 0.0095* (0.0057) 0.0141*** (0.0041) 0.0001 (0.0030) 0.0084 (0.0074) 0.009** (0.0036) 0.0138* (0.0073) 0.0034 (0.0081) 0.0141 (0.0105)

IMR Control years N Adjusted R-square

Yes 1228 0.0788

0.0016* (0.0009) 0.0178 (0.0263) 0.0043 (0.0068) 0.0145*** (0.0042) 0.0000 (0.0031) 0.0045 (0.0068) 0.0049 (0.0047) 0.0141* (0.0084) 0.0097 (0.0098) 0.0069 (0.0062) 0.0091 (0.0072) Yes 1040 0.073

Model 2 (a) ***

Model 2 (b)

0.0430 (0.0118) 0.0057*** (0.0021)

0.0191 (0.0172) 0.0034* (0.0021)

0.0006 (0.0006) 0.0004** (0.0002) 0.0030*** (0.0010) 0.0278* (0.0163) 0.0087* (0.0052) 0.0139*** (0.0041) 0.0003 (0.0029) 0.0080 (0.0074) 0.0094*** (0.0036) 0.0157** (0.0073) 0.0057 (0.0078) 0.0149 (0.0104)

0.0008 (0.0006) 0.0003 (0.0002) 0.0018* (0.0009) 0.0178 (0.0265) 0.0046 (0.0068) 0.0143*** (0.0042) 0.0007 (0.0030) 0.0044 (0.0068) 0.0049 (0.0047) 0.0166* (0.0085) 0.0111 (0.0095) 0.0076 (0.0062) 0.0088 (0.0071) Yes 1057 0.051

Yes 1264 0.059

This table shows the relationship between the 2-day cumulative abnormal returns (CAR (1, 0)) and investment bank reputation. CAR (1, 0) is the dependent variable; GIRank is the Governance Index rank of the acquiring firm in the year of the deal or 1 or 2 years prior to the deal; IBRank is the rank of the investment bank based on its previous market share; MarketShare is the market share of the investment bank in the previous year; ln(TA) is the natural log of total assets; ROA is the return on assets of the acquiring firm; Relative size is the transaction value divided by the acquiring firm’s total assets; Stockpay is the percentage of the deal value that is paid for with the acquiring firm’s stock; Industry is a dummy variable equal to 1 if the acquiring and target firms share the same 3-digit SIC code, 0 otherwise; Hostile is dummy variable equal to 1 if the target firm opposes the deal, 0 otherwise; Merger is a dummy variable equal to 1 if the deal is a merger, 0 otherwise; Major is a dummy variable equal to 1 for acquisitions of major interest, 0 otherwise; Partial is a dummy variable equal to 1 for partial acquisitions, 0 otherwise; Remaining is a dummy variable equal to 1 for acquisitions of remaining interest, 0 otherwise. The underlying benchmark group is asset acquisitions. IMR is the inverse mills ratio, estimated from the Probit Model 3 in Table 3. All models in the table are ordinary least squares (OLS) regressions. The heteroscedasticity robust standard errors are in parentheses. * Indicates significance at 10% level. ** Indicates significance at 5% level. *** Indicates significance at 1% level.

5. Conclusions We investigate whether managerial rights in acquiring firms impact the decision to use an investment bank in M&A deals. We then examine the value consequence of the interaction between managerial rights and the use of investment banks for acquiring firms’ shareholders. We find that strong managerial rights in acquiring firms are positively associated with the use of investment banks in mergers and acquisitions. This relation holds after controlling for deal and firm features such as transaction size, method of payment, type of transaction, and firm performance.

53

We also find that acquiring firms with strong management are more likely to use reputable banks. Confirming the existence of an interaction effect between strong management and the use of investment banks, we find that when acquiring firms have strong management, the use of investment bank is associated with a significant wealth loss for the shareholders. However, the use of more reputable investment banks could be alleviating the value-loss. The findings suggest that investment banks may help managerial empire building at the expense of shareholders only when they facilitate M&As among acquirers with strong rights. Acknowledgements We thank Melissa B. Frye, Honghui Chen, Michael Fuerst (the discussant at the 2008 Financial Management Association meetings), and the anonymous referee for helpful comments. The usual disclaimer applies. References Autore, D.M., Kovacs, T., Sharma, V., 2009. Do analyst recommendations reflect shareholder rights? Journal of Banking and Finance 33, 193–202. Beatty, R.P., Ritter, J.R., 1986. Investment banking, reputation, and the underpricing of initial public offerings. Journal of Financial Economics 15, 213–232. Benston, G.J., Smith, C.W., 1976. A transactions cost approach to the theory of financial intermediation. Journal of Finance 31, 215–242. Bliss, R.T., Rosen, R.J., 2001. CEO compensation and bank mergers. Journal of Financial Economics 61, 107–139. Bowers, H.M., Miller, R.E., 1990. Choice of investment banker and shareholders’ wealth of firms involved in acquisitions. Financial Management 19, 34–45. Campbell, T.S., Kracaw, W.A., 1980. Information production, market signalling, and the theory of financial intermediation. Journal of Finance 35, 893–913. Carter, R.B., Dark, F.H., Singh, A.K., 1998. Underwriter reputation, initial returns, and the long-run performance of IPO stocks. Journal of Finance 53, 285–311. Carter, R.B., Manaster, S., 1990. Initial public offerings and underwriter reputation. Journal of Finance 45, 1045–1067. Chae, J., Kim, S., Lee, E.J., 2009. How corporate governance affects payout policy under agency problems and external financing constraints. Journal of Banking and Finance 33, 2093–2101. Core, J.E., Holthausen, R.W., Larcker, D.F., 1999. Corporate governance, chief executive officer compensation, and firm performance. Journal of Financial Economics 51, 371–406. Core, J.E., Guay, W.R., Rusticus, T.O., 2006. Does weak governance cause weak stock returns? An examination of firm operating performance and investors’ Expectations. Journal of Finance 61, 655–687. Datta, S., Iskandar-Datta, M., Raman, K., 2001. Executive compensation and corporate merger decisions. Journal of Finance 56, 2299–2337. DeAngelo, H., Rice, E.M., 1983. Antitakeover charter amendments and stockholder wealth. Journal of Financial Economics 11, 329–362. Desai, H., Jain, P.C., 1997. Long-run common stock returns following stock splits and reverse splits. Journal of Business 70, 409–433. Dong, M., Hirshleifer, D., Richardson, S., Teoh, S.H., 2006. Does investors’ misvaluation drive the takeover market. Journal of Finance LXI, 725–762. Fahlenbrach, R., 2003. Shareholder rights and CEO compensation. Working paper, University of Pennsylvania, Institute for Law & Economic Research Paper 03-05. Fernando, C.S., Gatchev, V.A., Spindt, P.A., 2005. Wanna dance? How firms and underwriters choose each other. Journal of Finance LX, 2437–2469. Gompers, P., Ishii, J., Metrick, A., 2003. Corporate governance and equity prices. Quarterly Journal of Economics 118, 107–156. Grinstein, Y., Hribar, P., 2004. CEO compensation and incentives: Evidence from M&A bonuses. Journal of Financial Economics 73, 119–144. Heckman, J., 1979. Sample selection bias as specification error. Econometrica: Journal of the Econometric Society 47, 153–161. Harris, O., Glegg, C., 2009. Governance quality and privately negotiated stock repurchases: Evidence of agency conflict. Journal of Banking and Finance 33, 317–325. Jensen, M.C., Meckling, W.H., 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics 3, 305–361. Jiraporn, P., Kim, Y.S., Davidson, W.N., Singh, M., 2006. Corporate governance, shareholder rights and firm diversification: An empirical analysis. Journal of Banking and Finance 30, 947–963. Kale, J.R., Kini, O., Ryan, H.E., 2003. Financial advisors and shareholder wealth gains in corporate takeovers. Journal of Financial and Quantitative Analysis 38, 475– 502. Krigman, L., Shaw, W.H., Womack, K.L., 2001. Why do firms switch underwriters? Journal of Financial Economics 60, 245–284. Larcker, D.F., Rusticus, T.O., 2008. On the use of instrumental variables in accounting research. SSRN Working paper. Leland, H.E., Pyle, D.H., 1977. Informational asymmetries, financial structure, and financial intermediation. Journal of Finance 32, 371–387.

54

W. Wang, A.M. Whyte / Journal of Banking & Finance 34 (2010) 44–54

Mahoney, J.M., Mahoney, J.T., 1993. An empirical investigation of the effect of corporate charter antitakeover amendments on stockholder wealth. Strategic Management Journal 14, 17–32. Masulis, R.W., Wang, C., Xie, F., 2007. Corporate governance and acquirer returns. Journal of Finance LXII, 1851–1889. Megginson, W., Weiss, K., 1991. Venture capitalist certification in initial public offerings. Journal of Finance 46, 879–904. Mikhail, M.B., Walther, B.R., Wills, R.H., 2004. Do security analysts exhibit persistent differences in stock picking ability? Journal of Financial Economics 74, 67–91. McLaughlin, R.M., 1990. Investment-banking contracts in tender offers: An empirical analysis. Journal of Financial Economics 28, 209–233. Rau, P., 2000. Investment bank market share, contingent fee payments, and the performance of acquiring firms. Journal of Financial Economics 56, 293–325. Servaes, H., Zenner, M., 1996. The role of investment banks in acquisitions. Review of Financial Studies 9, 787–816.

Shleifer, A., Vishny, R.W., 1989. Management entrenchment. Journal of Financial Economics 25, 123–139. Staiger, D., Stock, J.H., 1997. Instrumental variables regression with weak instruments. Econometrica: Journal of the Econometric Society 65, 557– 586. Stock, J.H., Wright, J.H., Yogo, M., 2002. A survey of weak instruments and weak identification in generalized method of moments. Journal of Business and Economic Statistics 20, 518–529. Stouraitis, A., 2003. Acquisition premiums when investment banks invest their own money in the deals they advise and when they do not: Evidence from acquisitions of assets in the UK. Journal of Banking and Finance 27, 1917–1935. Trautwein, F., 1990. Merger motives and merger prescriptions. Strategic Management Journal 11, 283–295. Yermack, D., 1996. Higher market valuation of companies with a small board of directors. Journal of Financial Economics 40, 185–211.