在本論文，我們重新衡量美國市場從1990年1月到2003年12月不同類型的公司購併事件宣布後股價長短期的績效表現與差異。實證結果指出，集中式的購併相對於多角化的購併有較高的長短期異常報酬，這個結果也與資本錯誤配置假說相符合，其理論認為一個多角化公司的管理階層可能因為代理問題進而隨便地分配公司資本至各事業體，或是容易發生表現較好的事業群將其自由資金拿來當作表現較差的事業群的資金融通來源，進而產生投資不效率的問題。我們假設多角化所導致的績效下滑是源自投資不效率與代理問題，而我們將樣本根據衡量公司治理程度的指數分為容易產生代理問題與不容易產生代理問題的群體，衡量當一家具有良好公司治理體系的公司去從事多角化購併時，是否可以有效降低發生投資不效率與代理問題發生的可能性。實證結果指出，兩個衡量長期異常報酬的方法都顯現公司治理較不容易產生代理問題的群體相對於公司治理較容易產生代理問題的群體只有在多角化的購併樣本有較高的顯著異常報酬；而在短期下，此二個群體並無顯著績效表現的差異。 In this thesis, we estimate the long-run/ short-run abnormal returns after the different types M&As announcement and compare their differences of stock performance from January 1990 to December 2003 in the US. The results show that short-run/long-run performance in focus M&As is greater than the performance in diversified M&As. This supports capital misallocation hypothesis that diversified firms deliberately make poor allocations because of agency problems, or it results in cross-subsidization, where investments in the firm’s weaker divisions are supported with the cash flows from stronger divisions. We suppose that diversification discount is derived from agency problem and inefficient investment, we divide our sample into Democracy and Dictatorship portfolio based on bidder’s Governance Index. We estimate whether the firm with better corporate governance would efficiently reduces occurrence of agency problem and inefficient investment and the results show that the long-run abnormal returns of democracy portfolio is 9.16% higher than Dictatorship ones in diversified M&As for Buy and Hold Abnormal Returns (BHARs) method and is 0.16% for Calendar-Time Abnormal Returns (CTARs) method. But the differences of abnormal returns between Democracy and Dictatorship portfolio are insignificant no matter any subsamples in the short-run.
