An improved pairs trading strategy based on switching regime volatility

Marco Bee Giulio Gatti
A pairs trading strategy on energy, agricultural and index futures is developed. The strategy uses different parameters according to a volatility regime detected using a threshold evaluated in two ways, namely by means of a mixture of two Gaussian densities and a Markov switching model. The performance is assessed using different time frames and filters. When associated to cointegration, this investment algorithm gives a larger Sharpe ratio with respect to classical methods; on the other hand, the correlation filter does not work well with the regime switching algorithm.

Анализ стат. арбитража с использование hidden markov model для определения порогов входа. Тестирование осуществлялось на 7 парах: CL-BRN; HO-RB;ZM-ZL;ZS-ZC; YM-NQ; YM-ES; ES-NQ, на 10 и 20-ти минутках.

К-ты считали по OLS;границы в зависимости от режимов волатильности (по HMM). Шарпы 2-3 без использования режимов и 3-5 при использовании режимов.