國內對新上市股票超額報酬的實證研究雖然很多，但尚未有利用發行公司與投資人間資訊不對稱為基礎進行的實證結果。由於此一不對稱現象相當符合國內現況，因此值得以此為基礎進行實證。另外，國內相關研究到目前為止；皆只著重在探討超額報酬與解釋變數間正負關係，而忽略解釋變數間關係或其交互作用。因此本研究以Grinblatt & Hwang (1989)的訊號模型進行實證以瞭解是否能以發行公司與投資人間資訊不對稱來解釋國內新上市股票超額報酬，並對相關變數彼此間關係作一完整探討。 以Ibbotson & Jaffe (1975)的判別方式為基礎，判別出民國七十七年七月至七十九年一月為上市旺季、民國七十九年四月至八十二年六月為上市淡季。上是旺季的蜜月期為三十一天，經市場模式調整的累積超額報酬約90％、淡季的蜜月期為十三天，累積超額報酬約34％。由訊號模型推出的八個假設命題，在本研究中，六個可被實證結果支持，至於有關業主股份保留比率與公司價值風險成負相關的假設命題則不被實證結果支持。

There are lots of domestic empirical research about the underpricing of new issues, but none of them based on the information asymmetry between issuer and investors. In fact this information asymmetry is the real mode of domestic stock markets, so an empirical research based upon this information asymmetry is worthwhile. Besides, almost all domestic empirical research only focused upon the statistical correlation between excess return and explaining variables but omit the relations and interactions among variables. Consequently, this paper tests the statistical implications of the singalling models of Grinblatt & Hwang (1989) in order to find evidences to support the information asymmetry explanation and to get an exhaustive relation among variables. Based upon the method proposed by Ibbotson & Jaffe (1975), hot issues period and cold issues period are identified. Market model-adjusted CAR in hot issues period is some 90 percent, cold issues one some 34 percent. Among eight implications, six of them are supported by multiregression results, regardless of hot issues period or cold one. The others about negative correlation between issuers fractional holding and firm value variance are not supported by multiregression results.