本研究旨在探討1997至2008年台灣上市櫃IPO公司後報酬波動性與期初報酬率的影響因素,本研究具有下列特色:第一,台灣的IPO承銷方式對折價與市場後波動性之影響確實有差異,採詢價圈購方式者,其折價與市場後波動性最高,此與Sherman(2005)之主張相反。第二,本研究延伸Gleason,Johnston,and Madura (2008)的模型,並且加入席次控制與投票權偏離程度(SEATCON)、投票權與現金權偏離程度(CONCASH)新的變數來衡量公司治理對IPO後風險的影響,實證發現,席次控制與投票權偏離程度(SEATCON)與市場後波動性為顯著正相關、投票權與現金權偏離程度(CONCASH)為顯著負相關。第三,本研究考慮到折價與IPO後風險時兩模型間存在著內生性的問題,因此首次利用聯立方程模型來完整探討兩者間之關聯性,並且分析各外生變數對兩內生變數的淨效果,實證發現,單一迴歸方程式,顯著的解釋變數與聯立方程時大致相同,其影響方向亦一致,唯有IPO市場後個股報酬標準差與期初報酬率兩變數,其在OLS與3SLS時之係數差異甚大,本研究方法提供了更完整、更符合實際的模型來探討台灣IPO市場相關現象之關聯性。
The purpose of this study is to investigate possible factors influencing IPO under-pricing (initial abnormal returns) and IPO aftermarket volatility. The relationship between IPO abnormal returns and IPO aftermarket volatility is investigated as well. A total of 459 and 513 IPO firms of Taiwan exchange and OTC market, respectively, are collected from 1997 to 2008. The major findings of the study are as follows. First, for IPO initial return equation, there are significant differences in IPO initial returns and aftermarket volatility under different IPO underwriting mechanisms. IPO firms adopting the book-building mechanism are found to have the highest initial returns and aftermarket volatility. Results are different from that of Sherman (2005). Second, for an extension of the Gleason, Johnson, and Maduru (2008) aftermarket volatility equation specification, results show that the seat controlling right of the board and the deviation of voting right are positively related to aftermarket volatility, while the deviation between voting rights and cash flow rights is negatively related to aftermarket volatility. Lastly, in order to incorporate the relationship between the IPO under-pricing and aftermarket volatility into consideration when there do exist endogeneity, this study is the first attempt to model the above two equations into a simultaneous specification framework. Three Stage Least Square (3SLS) estimation results show that most coefficient estimates are the same as those of OLS either in magnitudes, signs, and significant levels. However, the 3SLS coefficient estimates of IPO initial return in the aftermarket volatility equation and of aftermarket volatility in the IPO initial return equation are found to be much higher than those of OLS. Results suggest that the consideration of a simultaneous equation framework of the study is important to future IPO empirical research.