Pairs Trading with ETFs: Backtested from 2007-2020

Alexander Ludvig Tommerup & Rasmus Bruun Jørgensen

Student thesis: Master thesis

Abstract

Pairs trading is a type of statistical arbitrage strategy created to exploit relative mispricings in the price development between two securities. The trading strategy has primarily been conducted on single stocks which have showcased declining profitability in recent years. This paper examines the application of a pairs trading strategy when substituting the traded securities from single stocks to exchange-traded funds (ETFs). Inferences of this modification is provided through an empirical backtest of the period from 2007 to 2020 by applying the two most prominent methods within the field of pairs trading, namely the distance method and the cointegration method. The findings of this paper reveal that pairs trading with the use of ETFs provides a more efficient alternative to the utilisation of single stocks when executed with the cointegration method. At the same time, it is demonstrated that a pairs trading strategy based on ETFs differs substantially depending on the parameterisations and method applied. The cointegration method generates robust and reliable trading attributes and statistically significant alpha for five out of six trigger settings tested. On the other hand, the distance method does not provide any statistically significant alpha, a consequence of not providing returns robust to transaction costs, but only generating net profit in highly volatile periods. This has shown to be the case as the pairs composition of the distance method primarily rely on pairs of ETFs tracking the same index, which is associated with realised losses on the traded positions after transaction costs. Pairs trading is a type of statistical arbitrage strategy created to exploit relative mispricings in the price development between two securities. The trading strategy has primarily been conducted on single stocks which have showcased declining profitability in recent years. This paper examines the application of a pairs trading strategy when substituting the traded securities from single stocks to exchange-traded funds (ETFs). Inferences of this modification is provided through an empirical backtest of the period from 2007 to 2020 by applying the two most prominent methods within the field of pairs trading, namely the distance method and the cointegration method. The findings of this paper reveal that pairs trading with the use of ETFs provides a more efficient alternative to the utilisation of single stocks when executed with the cointegration method. At the same time, it is demonstrated that a pairs trading strategy based on ETFs differs substantially depending on the parameterisations and method applied. The cointegration method generates robust and reliable trading attributes and statistically significant alpha for five out of six trigger settings tested. On the other hand, the distance method does not provide any statistically significant alpha, a consequence of not providing returns robust to transaction costs, but only generating net profit in highly volatile periods. This has shown to be the case as the pairs composition of the distance method primarily rely on pairs of ETFs tracking the same index, which is associated with realised losses on the traded positions after transaction costs.

EducationsMSc in Finance and Accounting, (Graduate Programme) Final ThesisMSc in Applied Economics and Finance, (Graduate Programme) Final Thesis
LanguageEnglish
Publication date2020
Number of pages125