本研究延伸徐語辰(2019)並加上其他相關文獻所提出的配對交易方法，對台灣股票市場進行實證研究。本文使用的模型為Engle and Granger (1987)提出的二階段共整合檢定，搭配上徐語辰(2019)所用的動態移動窗格共整合檢定法，來檢定台灣股市中的股票，從中找出具共整合性質之股票配對，並利用Lin et al.(2006)提出的最小獲利界限交易策略，找出價格異常的時間點進行交易；另外本研究也延伸徐語辰(2019)的設定，將殘差的性質分為兩種假設去討論，一為獨立同分配，另一為服從OU process。實證結果顯示兩個策略的結果多數都很不錯，不只是報酬率優異且穩定度方面也還可以；另外比較和徐語辰(2019)的結果，我們發現假設殘差服從OU process的交易結果比假設為獨立同分配來的好。;This paper extends the work of Hsu Yuchen (2019) and incorporates pair trading methods proposed in other related literature to conduct an empirical study in the Taiwanese stock market. The model used in this paper is the two-step cointegration test proposed by Engle and Granger(1987), combined with the dynamic moving window cointegration test method used by Hsu Yuchen(2019), to identify pairs of stocks in the Taiwanese stock market with cointegrated properties. Furthermore, the paper adopts the minimum profit bound trading strategy proposed by Lin et al.(2006) to identify unusual price movements for trading. Additionally, this research extends Hsu Yuchen′s(2019) settings by considering two different assumptions for the nature of the residuals: one assuming i.i.d. and the other assuming residuals following an OU process. The empirical results show that the outcomes of both strategies are favorable, not only demonstrating not only excellent returns but also reasonable stability.Unlike Hsu Yuchen′s(2019) findings, we discovered that the trading results under the assumption that the residuals follow an OU process are better than those under the assumption of i.i.d. residuals.
