A Cooperative Dynamic Approach to Pairs Trading

J P Ramos-Requena,M N López-García,J E Trinidad-Segovia,M A Sánchez-Granero,Paulo Jorge Silveira Ferreira

doi:10.1155/2021/7152846

Abstract

Based on recent works on stocks comovement, Pairs Trading’s strategy is enhanced by reducing the stock universe to the stocks with the lower volatility on a given date. From this universe of low volatility stocks, pairs are selected by looking for pairs whose series present a high degree of antipersistence. Finally, a “reversion to the mean” strategy is applied to these pairs. It is shown that, with this approach to Pairs Trading, positive results can be obtained for stock from the Nasdaq stock exchange, mainly during bull markets and low volatility periods.

Highlights

Statistical arbitrage techniques became a popular research topic after the pioneer paper of Gatev et al [1] who proved that this methodology was able to obtain positive results in US market during a period of 30 years (1962–1997)
Pairs trading strategies are structured as follows: during the formation period, two securities are identified, whose prices have moved together historically. ese securities are considered in equilibrium, which is identified by a proposed model
We propose the following selection method: in the first stage, we select the 50 stocks from our universe with the lower volatility at the selection date; in the second stage, we look for pairs from stocks of this 50 low volatility stocks with a high degree of comovement, based on the Hurst exponent of the pair

Summary

Introduction

Statistical arbitrage techniques became a popular research topic after the pioneer paper of Gatev et al [1] who proved that this methodology was able to obtain positive results in US market during a period of 30 years (1962–1997). The authors derived the Hamilton-Jacobi-Bellmann (HJB) equation to find closed-form solutions for the value and policy functions Following this line, Liu and Timmermann [21] derived optimal portfolio holdings for convergence trades under diverse arbitrage opportunities. During noncrisis periods, the degree of comovement among stocks with low volatility is significantly higher that among random stocks Based on these results, we propose the following selection method: in the first stage, we select the 50 stocks from our universe (stocks from the Nasdaq stock exchange) with the lower volatility at the selection date; in the second stage, we look for pairs from stocks of this 50 low volatility stocks with a high degree of comovement, based on the Hurst exponent of the pair (we look for pairs with low Hurst exponent). It is clear that the comovement of the 50 stocks with the lower volatility is quite high and well above the general market comovement, so this selection of stocks seems a good candidate to apply pairs selection methods

Hurst Exponent

Pairs Trading Strategy

Empirical Results

Conclusions