臺大管理論叢第31卷第1期

83 NTU Management Review Vol. 31 No. 1 Apr. 2021 In addition to the precautionary effect of excess cash, one may argue the effect of excess cash holdings on the ESOs-induced M&As is merely attributed to the contracting effect from the vega compensation. That is because the high vega compensation not only encourages managers to conduct more risky-taking activities but leads firms to maintain greater liquidity, which is required by their creditors (Liu and Mauer, 2011). Meanwhile, prior studies have documented the endogenous issue for executive stock options. Accordingly, we provide a robust analysis in the latter subsection to verify the endogeneity problem between excess cash, vega compensation and M&A decisions. 3.3.2 Announcement Effect Next, we examine whether investors would react more favorably to ESOs-induced M&A decisions in cash-rich firms than in non-cash-rich firms. Based on the model from Harford and Uysal (2014) and Fuller, Netter, and Stegemoller (2002), we employ the following equation: CARi,t = β0+β1Vegai,t-1+β2Deltai,t-1+β3Cashrichi,t-1 +β4Vegai,t-1×Cashrichi,t-1+β5 Deltai,t-1×Cashrichi,t-1 +β6Cashi,t-1/ TAi,t-1+β7Salesi,t-1+β8Market leveragei,t-1 +β9Stock returni,t-1+β10Market_to_Booki,t-1+ β11EBITDAi,t-1/ TAi,t-1 +β12 Industry M&A liquidityi,t-1+ β13Herfindahl indexi,t-1 +β14InsideOwni,t-1+β15 Relative Sizei,t +β16 Private Target i,t +β17 Within–industry acquisitioni,t +β18 All Cashi,t +β19 Competedi,t+β20 Hostilei,t +εi,t , (5) where CAR is the acquirer's cumulative abnormal returns over a five-day event window (two days before and two days after the announcement date). We calculate abnormal returns as the deviation from the acquirers’ predicted returns to the actual returns, and we estimate the predicted returns of acquirers from the market model regression. The benchmark returns are the value-weighted index returns, including dividends, for the combined New York Stock Exchange, American Stock Exchange, and Nasdaq. The estimation window is 200 days, from 205 days before to 6 days before the announcement date. Vega and Delta remain under the same definitions as in 3.3.1. We expect Vega to be positively related to announcement returns while Delta would have ambiguous results

RkJQdWJsaXNoZXIy ODg3MDU=