本研究主要目的為探討影響機構投資人持股比例的因素以及對未來績效的預測能力。由於機構投資人持股賣空訊息無法揭露,所以資料呈現左截分配形式,因此採用OLS將產生之估計偏誤。據此,本文以Tobit迴歸以校正傳統迴歸所產生的估計偏誤問題。此外,在訊息交易者的假設下,以及股票蜜月期與穩定性操作的存在,因此本文以股票上市後第一個季財務報表揭露日作為公司訊息揭露的時點,進一步檢驗機構投資人持股比例與績效預測的關係。經由實證結果發現:機構投資人持股集中在電子產業,而公司規模以及董監事持股將左右其持股比率高低。再者,無論是否控制動能因素,機構投資人高持股比率會有高報酬與只有在季報日後第一個月仍有獲利。就特徵因素而言,規模因素與淨值市價比因素均呈現顯著性。至於動能效果仍然在預測未來報酬中仍有顯著性的影響。
In this study, we examines determinants and return predictability of institutional ownership in IPOs. Owing to short selling, the ratio of the aggregate shares held by the institution is not less than 0%. The distribution of the observations would be truncated on the left. So we employ the Tobit model for left censored data assuming a normally distributed error term. Beside that, we assume there are two types of investors. The first type is unsophisticated investors, presumably individuals, who are prone to optimistic or pessimistic sentiment. The other type of investors, presumably institutional investors, has an unbiased estimate of the firm's future prospects.We find that IPOs widely held by institutions are electronic stocks. The significant coefficients confirm that there is much preference by institutional investors for bigger market value and higher director ownership stocks. To find the better predictors of returns, we employ the Fama & Macbeth (1973) procedure. Our findings show the effects of firm characteristic including the firm's scale and the ratio of book equity to market valve are significant. Moreover, the effect of institutional ownership is only significant during the second month after the report date. Overall, these findings reinforce the importance of price momentum in the price behavior of IPOS.