The impact of the changing trading time on volatility, liquidity and efficiency of Taiwan Stock Market

[[abstract]]證券市場交易時間的長短攸關數以百萬計投資大眾的福祉，台灣證券市場從2001/1/2起，每日交易時間由3小時延長為4.5小時，在此之前，已歷經四次的交易時間變革。因此研究主要的目的是探討交易時間延長後是否可提升市場績效：降低波動性、提升流動性及效率性。 本研究運用最近四次的交易時間改變，以2000萬筆以上的日內交易資料與日交易資料，驗證交易時間改變對波動性(volatility)、流動性(liquidity)及效率性(efficiency)等市場績效指標之影響。以本文以報酬率變異數與GARCH模型所得的條件變異數衡量波動性；以週轉率、交易量及Hasbrouck-Foster-Viswanathan模型等方法衡量流動性；以一般自我迴歸模型與Hasbrouck and Schwartz的市場效率係數衡量效率性進行實證。整體而言，由日內資料的實證發現：交易時間延長導致波動性增加，流動性不受影響及效率性減少；日資料實證的結果：交易時間延長使得波動性增加，流動性、效率性不受影響。 延長交易時間對台灣證券市場日內績效成現一致性的負面影響；對日的市場績效影響則是正、負兩面並存。就交易成本及交易公平性，以及流動性及效率性的定義來說，日內的流動性及效率性的重要性及攸關性，可能相對高於日的流動性及效率性。此外，探討交易時間影響市場績效的成因，結果發現主要成因不是公共資訊。交易時間與波動性呈現正相關的實證結果，隱含台灣證券市場之選擇權定價型中的波動性與時間兩個變數，宜採用的衡量基礎是交易時間而非日曆時間。[[abstract]]This study examines the effect of changing trading time on volatility, liquidity and effieciency. Since January 2, 2001, Taiwan Stock Exchange (TSE) has changed its daily trading time form 3 hours to 4.5 hours. Before that, TSE had changed its trading time five times. We take advantage of the last four changes of trading time , and using more than 20 millions observations of the transaction data from the Taiwan Stock Exchange, to verify the impact of the changing trading time on the market performance that contains volatility, liquidity and efficiency. From the intra-day data, the empirical results show that increasing the trading time decreases the efficiency and increase volatility, but not impact on the liquidity; From the inter-day data, the empirical results show that the longer the trading time the higher the volatility, not significant impact on the efficiency and liquidity. These results imply that increasing trading time will lead to poor intra-day market performance, and can not consistently improve the inter-day market performance。From the viewpoint of the transaction cost and fair trade, or from the definition of liquidity and efficiency, the intra-day liquidity and efficiency may be more important and relevant to the inter-day liquidity and efficiency. The result also cast some doubt on the current trend to increase the trading time. Our results indicate that, public information not affect volatility