NASDAQ避險型個股之市場效率收斂性

Abstract

長久以來，成交量在股票市場中一直扮演著重要的角色，而其中買賣單不均衡更是投資人用以提高報酬的主要指標之一。過去的實證顯示，單日買賣單不均衡確實與單日股價報酬率有著顯著的關係，然而，這樣的關係卻會隨著交易時間的增加而逐漸的消失。換言之，市場的不效率性的確會隨著交易時間的拉長，逐漸的回復到效率市場的境界，此時，投機型投資人便無法藉由觀察買賣單的不均衡建立起具有超額報酬的投資策略。 本論文利用價平、量縮、低價等條件篩選出避險型個股，並以持有期間股價最高的當天為採樣樣本，探討在日內資料中市場達成效率境界所需的時間。實證顯示，在五分鐘以及十分鐘的交易區間內，買賣單不均衡與股價報酬仍有50%以上的樣本資料呈現顯著的正相關，而在十五分鐘的交易區間，顯著正相關的樣本資料僅有30.95%，此即表示當交易區間延長，市場的效率也會隨之增加。 其次，本文亦利用OLS模型，分別測試五分鐘、十分鐘以及十五分鐘各交易區間買賣單不均衡與公司規模間是否存在小型股效果。實證結果顯示，買賣單不均衡與公司規模間的負向關係並不顯著。 最後，本文針對買賣單不對稱的情形建立一交易策略，當各時間區間的總和買賣單不均衡為正向則買進，為負向則賣出。實證結果顯示，有篩選的交易策略之報酬明顯高於未經篩選策略之報酬。此外，交易區間愈短報酬亦越高。

For a long time, the trading volume has been playing an important role in the stock market. Furthermore, investors usually utilize order imbalance as an indicator to earn abnormal returns. Among previous studies, Chordia and Subrahmanyzm (2004) found the positive relationship between order imbalances and stock returns. However, this relation will become insignificant as the time interval increases. In the efficient market, speculative investors find it difficult to profit only by observing the order imbalances. In this paper, we select the hedging stocks as our sample stocks using the criteria of stationary price, declining volume and price range. Furthermore, we examine how long within the day the effect of order imbalances on prices lasts. The empirical results of this paper show that more than 50% of the samples have significant positive correlation between order imbalances and stock return in five to ten-minutes time interval, but only 30.95% in fifteen-minute time interval. This result provides strong evidence that the market is getting more efficient as the time interval becomes longer We also employ OLS model to examine the small firm effect in five to fifteen- minutes time interval respectively. According to the empirical results of this paper, we are not allowed to make a conclusion that there exists a small firm effect in our sample data. Finally, with the help of the previous results, we develop the order imbalances based trading strategy. That is to say, in each time interval, if order imbalances are positive, we long this stock, and sell it when the order imbalances turn negative. The result presents that the return from truncated trading strategy is better than non-truncated one. Similarly, the return in shorter time interval also performs better than it does in longer time interval.