最大跌幅個股市場效率收斂性之研究

Abstract

市場效率性一直是學者們研究的議題之一，股價會即時反應所有的資訊只存在於理論下完美的效率市場中。在現實世界裡，市場仍舊需要時間來反應資訊，本文即是致力於研究市場反應資訊的過程，期望以實證結果顯示出市場效率的收斂性。 經由過去的研究顯示，買賣單不均衡與股價報酬率間有顯著的關係存在。因此本文以每日之最大跌幅個股為研究標的，測試此顯著關係是否隨著市場反應時間增加而遞減來證明市場趨於效率。首先，我們以多元迴歸模型和GARCH(1,1)模型來檢驗買賣單不均衡與股價報酬率之間的關係，我們發現兩個模型皆檢驗出兩者之間有顯著關係且此關係隨著時間遞減。 接著，我們以GARCH(1,1)模型來研究買賣單不均衡與股價報酬波動性之間的關係，同樣的，我們可以觀察到兩者之間隨著時間而明顯下降的顯著關係，隱含市場隨著時間增加而逐漸趨於效率。 另外，我們使用簡單迴歸模型來研究買賣單不均衡對報酬的預測力是否明顯存在於小型股中，然而，沒有顯著的實證結果證明我們的資料中存在小型股效果。 最後，本文嘗試建立一套以買賣單不均衡為指標的交易策略，當看到賣方市場的買賣單不均衡時放空該股，而當買賣單不均衡轉換為買方市場時即回補。經由實證結果測試，當我們使用交易價格計算報酬率時，此交易策略可以成功的賺取正向報酬。

The concepts of market efficiency have drawn much attention in finance field. In the real world, however, it takes time for the market to react and reveal the information. The central goal of this thesis is to present evidence about the important issue: convergence to market efficiency. Firstly, we selected 70 samples of daily top losers from July 2006 to December 2006 and applied the OLS method and GARCH model to examine whether order imbalance have significant influence on stock return. We found that the significance of order imbalance coefficients decreased with increasing time interval (5, 10 and 15-min), indicating evidences on convergence to market efficiency. Similarly, we used the GARCH model to test the relation between order imbalance and volatility. Again, the significant coefficients have a declining pattern which also supported the convergence to market efficiency. Then, we ran the regression on small firm effect. The expected negative relation between order imbalance coefficients and market capitalization was not achieved; as a consequence, the empirical results cannot make any conclusion on small firm effect. Finally, we developed an imbalance-based trading strategy and made profit from these daily top losers. Our strategy was to short sell when seeing the first seller-initiated order imbalance and immediately buy back the underlying when the order imbalance transfer to buyer-initiated. We applied many methods in testing our strategy, such as using trading price or bid-ask price to evaluate the performance of the strategy, and selecting order imbalance with 0% or 90% truncation. All of them performed better than the original buy and hold rate of return.