Regression Results

Firstly, I use board effect variables as explanatory variables and bidder characteristics and deal characteristics as control variables. The dependent variable is five-day CAR around each acquisition announcement. In Table V (1), we can see that no board effect variable has a significant effect on the acquisition CAR but some of bidder characteristics variables show significant effects on the acquisition CAR. Firm size has a positive relationship with CAR, it shows that the larger the firm size is, the higher the CAR is. Asquith, Bruner and Mullins (1983) use 214 acquisition events as the research sample and their result shows that firm size and the acquisition CAR have a positive relation. Hence my result is consistent with theirs. Tobin’s Q can be seen as a proxy for the firm operation performance and in my regression and it also

has a significant positive effect on bidder returns. Lang, Stulz, and Walking (1989) and Servaes (1991) document a positive relation between Tobin’s Q and tender offer acquisitions and between Tobin’s Q and public firm acquisitions, respectively. Hence

this result is within our expectation. Leverage and stock price runup both have significant negative effects on the acquisition CAR. Runup is highly significant and this is coherent with the result of Masulis, Wang, and Xei (2007). I expect the leverage has a positive effect on the announcement CAR because if firms with higher leverage the creditor will monitor the firms more severe. However, the result is not as

22

Table V

Regression Analysis of Bidder Returns¹

The sample consists of 118 completed mergers and acquisitions between 2000 and 2008 made by firms which are listed in TWSE and GTSM. The dependent variable is the bidder’s 5-day cumulative abnormal return in percentage points. Variable definitions are in the Appendix I. In parentheses are t-statistics. ***, **, and * stand for statistical significance based on two-sided tests at the 1%, 5%, and 10% level, respectively.

(1) (2) (3)

23

my forecast and leverage negatively affects CAR. Maybe because better firms in Taiwan tend to have lower leverage than worse firms and better firms is apt to make value-adding M&A, there is a negative relationship between leverage and CAR.

Next, I use Ownership effect variables as explanatory variables and bidder characteristics and deal characteristics as control variables. The dependent variable is also five-day CAR. In Table V (2) we can see that some of ownership effect variables have significant influences on the dependent variable. As blockholders holding rate is higher, the acquisition CAR is larger and this result is inherent with Radd and Wu (1994) who use 105 acquisition events from 1981 to 1986 as their sample and find the same result. Shleifer and Vishny (1986) also find that a firm with higher ownership concentration can have higher acquisition CAR. However, foreign investors holding

rate has a negative effect on the CAR, which is consistent with

Pound’s (1988) strategic alignment hypothesis which consider that when there is

cooperation between the institutional investor and the management of the firm, due to the relationship benefit, the institutional investor will not monitor the firm well and tend to indulge the management of the firm. Hence this firm’s management is apt to

1 I use the Breusch-Pagan test to test whether conditional heteroskedasticity is present.

The test statistics for the Breusch-Pagan test in three regression models are 13.92, 18.11, and 20.24 respectively and the one-tailed critical value for a chi-square distribution with 11, 13, and 17 degree of freedom and alpha equal to 5% is 19.68, 22.36, and 27.59. Hence I do not reject the null hypothesis that the regression models have conditional heteroskedasticity and conclude that there are no conditional heteroskedasticity problems in three regression models.

24

make value-reducing acquisition decisions. Meanwhile, the significant bidder

characteristics variables in first regression are significant as well in this regression and they are firm size, Tobin’s Q, leverage, and stock price runup.

In the last regression (Table V (3)), I use all variables in this article to explain the

acquisition CAR. As before, every board effect has no significant effect on the bidder’s return and those significant variables in previous two regressions are still

significant in this regression. It means that these variables have genuine influence on

acquisition CAR. The bottom line is that blockholders holding rate, firm size, and Tobin’s Q have significant positive effects on the CAR and on the contrary, foreign

investors, leverage, and stock price runup have a negative effect on the bidder’s return.

On the other hand, board effect variables and deal characteristics variables have no significant influences on the CAR.

We can see that board size have no significant effect on the CAR and the coefficients in the regression model (1) and (3) are different in their direction. It means that no matter how big or small the board size is, it has no influence on the CAR. Board pledge rate is predicted to have a negative influence on the CAR.

Although they are not significant in two regression models, their coefficients are both negative and show that they negatively affect the announcement CAR. Surprisingly, Independent board is a variable which has a negatively insignificant effect on the

25

dependent variable. Maybe because these independent directors do not monitor the companies well or they do not have enough professional knowledge in dealing with merger and acquisition. CEO also has a direction different from my expectation. In Taiwan it is common for a big company whose CEO is also the chairman of the board of director. Due to this phenomenon, CEO variable has positive influence on the announcement CAR.

In my prediction, excess control should have negative effect on the CAR;

however, in my research it is positively but insignificant related with the CAR. Excess control is voting right minus cash flow right and higher excess control means higher voting right or lower cash flow right. A company with higher excess control tends to have higher cost of equity but it does not mean that this company will make bad acquisition decisions. This result needs more following research to justify. Board holding rate is predicted to have a positive effect on the CAR and the result is identical to my expectation; however the effect is not significant in both regression models. Managers holding rate is expected to have the same effect as board holding

rate on the CAR but the result is not like our forecast. It has an insignificantly negative influence on the dependent variable. Generally, higher the manager’s holding

rate, lower the agency cost and managers with higher holding rate tend to make better M&A decisions for themselves and other stockholders. Maybe it is because managers

26

with higher holding rate are more desirous of expanding the scale of company to acquire more benefit, they lack ripe deliberation and hence make value-reducing M&A decisions. Family holding rate shows a negative but insignificant effect on the CAR. Family holding rate has an insignificantly negative relation with the CAR.

Although higher family holding rate can reduce the conflict of interest between stockholders and managers, it may give the management too much power to make good M&A decisions. The family have high holding rate so other stockholders and creditors can not possess enough power to monitor or affect their decision-making.

Hence, higher family holding rate may result in lower announcement CAR.