Copula approach for pairs-trading

Abstract

Pairs trading is a market neutral trading strategy that was first introduced and implemented by Morgan Stanley in the 1980s. This simple strategy involves trading securities with similar trends in pairs, placing the undervalued security in long position and the overvalued security in short position. Investors will make profits when the gap closes, no matter what the market trend is. However, the common approaches to this issue rely on correlation coefficient, which is based on the assumption of normality of financial data. This oversimplified assumption may be biased as financial data in real world are not guaranteed to be normally distributed. In fact, most of them are skewed with heavier tails. Copula is a type of distribution function which describes the dependency between variables by relating the joint distribution with their respective marginal distributions, without being subject to any assumptions. Therefore copula is an obvious candidate for improving description of dependency between the pairs, and is likely to achieve higher profits when implemented properly to pairs trading strategy. This project is an empirical study on the effect of copula approach in pairs trading, with comparisons to traditional distance approach and random trading. [5th Award]