在資訊不對稱的狀況下，大股東及內部人藉由詢價圈購的特性，於轉換價格訂定期間操控股價，使得外部股東財富被移轉給大股東或特定人。有鑑於此，2004年主管機關修正轉換價格訂定規定，以杜絕特定人圖利情事。本研究以修正後之年度，2005至2013年上市(櫃)公司所發行之可轉債為研究對象，可轉債轉換價訂價基準日為事件日，分別觀察各事件期之平均異常報酬及累積異常報酬。 實證結果顯示：國內可轉換公司債，其於轉換價格訂定基準日前幾日即開始產生顯著正的平均異常報酬，並持續至基準日後2個交易日，顯示大股東及特定人在訂價期間無壓低股價之情事，反而拉高訂價。因此本研究認為在2004年修正轉換價格基準日設定後，使得過去大股東或內部人訂定較低轉換價格，於發行後立即得利狀況不復存在。即證明主管機關修改轉換價格基準日後，有效壓抑大股東或內部人於訂價前1、3、5個交易日壓低股價的意圖。而上述結論在控制上市(櫃)別、產業別、研究期間等因素下，仍持續存在。 In the presence of information asymmetry, blockholders and insiders may take advantage of the characteristics of bookbuilding to manipulate their stock price during conversion price setting, in an attempt to transfer the wealth of external shareholders to them or any specific people. To avoid this behavior, the authority concerned has amended regulations governing conversion price setting in 2004. Focusing on convertible bonds issued by publicly listed firms in Taiwan during 2005~2013, a period where the amended regulations have gone into effect, this study uses the record date for setting conversion price of convertible bonds as the event day to observe the average abnormal return (AAR) and cumulative abnormal return (CAR) for these firms in different event periods. Empirical results highlight a tendency among all the sample firms that a significantly positive AAR emerges several days before the record date for conversion price setting and persists until two trading days after the record date, suggesting that there is no underpricing of stock prices by blockholders or specific people and stock prices are raised during the conversion price setting period. It can be inferred that since the 2004 amendment of the regulations governing the record date for conversion price setting, the behavior of gaining an instant profit by setting a lower conversion price has been eliminated. In other words, this amendment has effectively inhibited blockholders’ and insiders’ intention to underprice their stocks one, three or five days preceding to the conversion price setting day. It should be noted that this finding persists even if the effects of market type, industry type or research period are controlled for.