Shareholders wealth and mergers and acquisitions ( M & As )

We re-examine the abnormal returns (ARs) around merger announcements using a large sample of 8,945 announcements. We estimate the ARs using the Carhart (1997) fourfactor model under the standard ordinary least square (OLS) method and the Glosten et al.’s (1993) asymmetric GARCH specification (hereafter, GJR-GARCH). Under the OLS method, acquirers do not generate significant cumulative ARs (CARs) in line with prior work. Our new results, however, show that under the GJR-GARCH estimation, acquirers generate positive and significant cumulative CARs. We attribute the gains to the use of the GJR-GARCH estimation method, as the GJR-GARCH method is more effective in capturing conditional volatility and asymmetry in the excess returns. Justice Kyei-Mensah (Ghana), Chen Su (UK), Nathan Lael Joseph (UK) BUSINESS PERSPECTIVES LLC “СPС “Business Perspectives” Hryhorii Skovoroda lane, 10, Sumy, 40022, Ukraine www.businessperspectives.org Shareholders wealth and mergers and acquisitions (M&AS) Received on: 19th of May, 2017 Accepted on: 11th of September, 2017


INTRODUCTION
There are a number of review studies in the areas that take on different perspectives. For example, Haleblian et al. (2009) consolidate the materials from management, economics, finance, accounting and sociology. Tuch and O'Sullivan (2007) review short-and long-run performance of firms engaged in M&As deals. Napier (1989) reviews the materials from a human resource perspective. 2 Aslinger and Copeland (1996) find increases in acquires' value of 14.3% above the S&P 500 index (see also Savor and Lu, 2009;Martynova et al., 2007). However, Sharma and Ho (2002) report that M&As have an insignificant effect on the adjusted operating performance of firms. Bild et al. (2002) also conclude that up to 30% of the pre-acquisition value of U.K. firms is destroyed following the completion of M&As (see also Dickerson et al., 1997).
There is no doubt that the area of mergers and acquisitions (M&As) has been heavily researched 1 . However, the empirical findings are not always consistent. To date, accounting and finance researchers provide definitive answers on the economic gain arising from M&As deals. For example, the finance literature indicates that acquirers' abnormal returns (ARs) around merger announcements are either zero or negative and significant (Campa & Hernando, 2006;Dutta & Jog, 2009;Stunda, 2014). These results hold fairly consistent, except when targets are unlisted (Faccio et al., 2006;Fuller et al., 2002). Only targets tend to consistently generate positive ARs (Goergen & Renneboog, 2004). The empirical results are mixed in accounting research 2 . The economic question is: Why do acquirers undertake M&As deals that do not generate gains to their shareholders?
This paper focuses on the estimation issues around the determination of the ARs. How the ARs are estimated is important as it affects inferences about the gains to shareholders in M&As deals. Using the bid price observed in capital markets is the most appropriate measure of the gains to shareholders (Grinblatt & Titman, 2002). This is because managers and executives have less control over capital markets, thereby causing market valuations to be more representative of true value 3 . Uncertainty about both the acquirer and target prices can dictate the form of payments, which, in turn, can affect the ARs. Myers and Majluf (1984) indicate that a share exchange occurs if acquirers believe that their shares are overvalued. Thus, adverse selection on the part of acquirers could lead to an exchange of acquirer's own stocks with targets, so that target shareholders share the risk of overpayment (Eckbo & Thorburn, 2000).
The next section briefly discusses the theories underpinning mergers and acquisitions deals and relates them to existing evidence. Sections 2 discuss the methodologies in prior work. Section 3 presents our data and research methodology. Section 4 presents empirical results and we conclude in the final section. 3 The argument relies on rational behavior and market efficiency. However, stock prices around M&As announcements can be mis-priced ( These exceptions are studies that examine acquirer returns when targets are private and/or not listed. These studies report positive ARs for acquires (see Faccio et al., 2006;Fuller et al., 2002). It is useful to acknowledge that private and/or unlisted targets do not have market valuations or reference points. Thus, they may be undervalued. Similarly, asymmetric information is more pronounced in capital markets when targets are not listed.  5 . While these studies do not necessarily emphasize the economies of scale and cost effectiveness motive for mergers, the general result is that the acquirers' ARs are not positive. In contrast, most empirical studies document positive ARs for targets (Fuller et al., 2002), suggesting that all of the gains go to target shareholders.

REVIEW OF PRIOR WORK
Economic impact of mergers and acquisitions -The economic impact of M&As is significant as they affect several interest groups, i.e., employees and creditors. Studies that investigate the economic and social effects of mergers suggest that M&As have prolong negative repercussions due to lay-offs following mergers (Blonigen & Pierce, 2016). M&As can also lead to excessive market concentration and contribute to price increases and reduction in consumer welfare (Carletti et al., 2015). Other studies suggest that restructuring following M&As help safeguard the workforce of targets (see Inoue et al., 2010). Other studies suggest that M&As enhance operational efficiency of firms (Carline et al., 2009).
Synergy motive -Synergy theory suggests that M&As take place due to the economic benefits of unification following mergers. Dutordoir et al. (2014) report that disclosing synergy forecasts prior to mergers leads to an increase in returns. M&As also take place to exploit financial (Leland, 2007) and operational (Lewis & Webb, 2007)

Sample selection and descriptive statistics
We identify US M&As using the Thomson Financial Securities Data Company's (SDC) Database over the period January 1, 1991 to December 31, 2013. Similar to Moeller et al. (2005), we require that: (i) each merger announcement leads to successful completion and that there are less than 1,000 days between the announcement and completion; (ii) the deal value is one million dollars or more and the deal value relative to the market capitalization of acquirer is more than 1%; (iii) the acquirer is publicly quoted nonfinancial U.S. firm listed on the NYSE, AMEX, or NASDAQ; (iv) the acquirer also has financial and accounting data on the Center for Research in Security Prices (CRSP) and Compustat databases; (v) the target is a U.S. public or private nonfinancial firm; and (vi) the acquirer controls less than 50% of shares of the target at the announcement day, but ends up with 100% on completion.
Following Chang (1998), we include only firms with M&As announcements in the event window. We exclude acquirers with stock prices below two dollars at the announcement date. Our final sample comprises 8,945 successful M&As made by 2,970 acquirers. Following Martin (1996)

Methodology
To capture the ARs, we first estimate the Carhart (1997) four-factor model over the estimation-window ( ) -240, -6 tt relative to the merger announcement date , t thus:     Kolari and Pynnӧnen (2010) is used to test statistical significance. The Wilcoxon signedranks statistic tests for differences in the ARs over the estimation methods. a, b, and c denote the statistical significance at the 1%, 5%, and 10% levels, respectively.

ARs for acquirers and targets
Panel A of Table 2  Panel B of Table 2 shows the corresponding results for targets. Over the two-day window t to 1 , t + the CAR of 2.33% is positive and significant under the OLS method, corroborating prior results (Goergen & Renneboog, 2004). The GJR-GARCH still outperforms the OLS method. Here, the CARs are significant over the four-day window -1 t to 2. t + The Wilcoxon signed-ranks test rejects the null hypothesis that the magnitude of the AR are similar for both estimation methods (p-value ≥ 0.05). Furthermore, Table 3 shows that using bootstrapping, the simulated CARs are similar to those estimated under the OLS and GJR-GARCH methods.

ARs and payment methods
Following Myers and Majluf (1984), high value acquirers tend to make cash payment or a large pro- Note: This table presents the average ARs measures, i.e., ARs, CARs, and SCARs, for acquirers (in Panel A) and targets (in Panel B) around merger announcements using the Carhart (1997) four-factor model under the OLS method and the GJR-GARCH method. The corresponding simulated returns (boot.) around merger announcements are based on the nonparametric bootstrapping simulations using 1,000 runs with replacements for each estimation method.
portion of cash payment to close the deal, to signal the higher value of their stocks. Adverse selection on the part of acquirers can cause them to exchange stocks, as this allows targets to share the risk of over-payment using cash (Eckbo & Thorburn, 2000). This argument suggests that acquirers will make stock payment to shareholders of targets when there is high uncertainty about market value of targets. In contrast, acquirers will make cash payment when there is high uncertainty regarding their own market value. This means that payment methods will affect the magnitude of the CARs. So we analyze the ARs according to the method of payments. Table 4 shows the estimated ARs for acquirers according to the payment methods. Under the OLS method, the CARs are positive and significant (p-value ≤ 0.10) when cash payments are made, Corresponding results for targets are shown in Table 5. The CARs are positive and significant across the payment methods, except for stock payment. As before, the persistence in the CARs is much stronger under the GJR-GARCH method. The Wilcoxon signed-ranks test also confirms that the CARs different under the estimation methods (p-value ≤ 0.05). Note: This table presents the average ARs measures similar to Table 3 but for targets. The Wilcoxon signed-ranks statistic tests for differences in the ARs over the estimation methods. a, b, and c denote the statistical significance at the 1%, 5%, and 10% levels, respectively.