Pairs trading via Regime- switching in Electronic Components Industry Taiwan

Abstract

With the technological innovation and booming application of computing equipment the complex statistical models have become popular and low-entrance level tools towards arbitrage therefore decreasing the profitability of pair trading Previous studies always implement pair trading under the assumption of single regime This kind of traditional trading approach will cause huge loss whenever the market is under volatility anomaly This paper combines the Markov regime-switching model with the least squares method to match pairs testifying whether these pairs could create abnormal returns The samples are chosen from the listing companies in Taiwan's electronic components industry from 2008 to 2013 and adjust pairs in each quarter We not only utilize the model prediction interval criterion as trading principle but also incorporate the minimum squared distance method with high-volatility regime deciding by ex ante probability expecting to enhance returns from pairs trading Empirical results show that during the 6-year transaction periods the cumulative return for our pairs is up to 270% after considering transaction costs The outcome is significantly higher than the cumulative return of Taiwan Electronic Components Index and traditional pairs trading strategy

Date of Award	2014 Jul 22
Original language	English
Supervisor	Meng-Feng Yen (Supervisor)