Regression Models and Variable Definition

4.1 Firm Performance and CEO compensation after an acquisition

Equation (1) is the regression model for examining the impact of firm performance on CEOs‟ compensation.

The left-hand-side variable is the logarithm of CEO‟s total pays in company i for year t, and the right-hand-side are the firm characteristics variables and deal status dummies applied as the independent variables.

This study estimates the model based on panel data set that includes all ExecuComp firms over the entire sample period. The estimation takes industry and year fixed effects into account to examine both individual-specific effects and time effects. This study includes 48 Fama-French industry dummies (Fama and French (1997)) to control for the difference in the demand for managerial talent.

This study expects that larger firm with higher growth opportunities will demand higher quality managers and thus will offer higher pay. This study proxies firm size and growth opportunities by logarithm of sales and sales growth rate respectively. Previous literatures suggest that the level of executive pay should be an increasing function of firm performance. Thus

12

this study employs ROA and the annual stock market return to evaluate the sensitivity between firm performance and CEO pay. Since the positive and negative returns may result in different consequence when it comes to CEO incentives evaluation, this study divides the returns into two categories:

Positive Return and Negative Return. Positive Return (Negative Return) is the fiscal year stock return from year t-1 to t if the stock return is positive (negative), and it is set to zero otherwise. This contributes to clarifying the pay-performance sensitivities. If the coefficient of negative return is positive, the CEO compensation is accordingly declined with firm performance.

Moreover, to assess the pay differential that acquiring CEOs realize after the acquisition, this study introduces the variable Acq to denote the year after the deal completion year. This study expects that under different situations, the composition and policy for CEO incentives may differ accordingly. Thus this study employs SucAcq and WithAcq to denote the year after the completion year of successful deals and the withdrawn deals respectively. That is, is equal to 1 if the company i conducted a successful acquisition deal that completed at year t-1, otherwise the value is set to 0. The same rule applies to .

To capture the possible differential sensitivity of pay to performance for sample firms after the acquisition in different situations, this study creates the interaction term of the return variables and the acquisition indicator variables. The interaction term, , for instance, represents the group of firms that completed a successful acquisition with positive stock return. Four interaction terms are thus appear in the model since there are two deal status indicator variables and two return variables.

4.2 Bid Characteristics and CEO Post-Acquisition Incentives

Bid characteristics play an important role in acquisition evaluations. Previous literatures include bid characteristics to assess the post-acquisition firm performance. This study further examines the post-acquisition payment sensitivity.

Equation (2) to (4) shows the modification of equation (1) by adding different bid characteristics into the original regression model.

13

The first introduced characteristics variable in equation (2) is Cash. Cash is a dummy variable whose value is equal to one if the acquiring firm conducts 100%

cash offer to acquire the target firm. If the method of payment includes stock payment or a combination of cash and stock, the Cash variable takes the value of 0. The past research suggests that stock financed acquisitions generated significantly lower returns than cash financed ones (Loughran and Vijh, 1997).

In acquisition cases, the advisor firms play an important role. An acquiring firm may lack of acquisition skills and thus use an advisor to assist on the transaction.

For more important deals, acquiring firms are more likely to employ professional advisor firms to assist them. Thus this study introduces the second indicator variable, Advisor, to examine whether the use of advisor firms affect CEO compensation. Advisor is set equal to one if the acqiror uses the advisor service, and equal to zero, otherwise. Equation (3) presents the regression model that includes Advisor dummy.

The third deal characteristics variable is RelativeAcqSize. RelativeAcqSize is the final purchase price of the acquisition as a percentage of the market capitalization of the acquiring firm at the time of acquisition announcement. For acquiring firms, RelativeAcqSize counters for the importance of the deal. That is, the higher the

14

RelativeAcqSize ratio, the more important the acquisition. If an acquiring firm decides to acquire a company that is much bigger than itself, this may arouse the disgorging issues of firms‟ available fund. If the Relative Size is extremely large, the CEO compensation may be inevitably influenced, both in the composition structure or the way of payments. Equation (4) consists of RelativeAcqSize for estimation.

4.3 Anti-takeover Provisions and CEOs Compensation After The Acquisition

This study further examines the relation between anti-takeover provisions and CEOs compensation. The boards have the power to decide the structure of CEO compensation, and even to provide downside protection of CEO when it comes to unpleasant situations. Therefore this study uses the corporate governance proxy to investigate the possible changes in CEO compensation after an acquisition. In previous literatures, many researchers used different proxies for measuring the level of corporate governance. Gompers, Ishii, and Metrick(2003) constructed a broad index, G-index, of antitakeover provisions using five governance rules for a total of twenty four possible provisions. The five governance rules are delay, protection, voting, state, and other. The index employs a point scale from one to twenty four. For every firm, the index adds one point for every added provision that restricts shareholder rights, which means the managerial power is enhanced.

The index with highest value has the weakest shareholder rights, and the index with lowest value has the strongest shareholder rights. The G-Index is pervasively used in measuring the power between shareholder and CEOs, thus this study employs G-index as corporate governance proxy for the following estimation.

Equation (5) shows the regression model considering corporate governance proxy.

To further investigate the effect of firm performance, this study also includes the PositiveReturn and NegativeReturn variables to examine the interaction impact on CEO payoff. Equation (6) includes the both GIndex and the interaction terms.

15