Pairs trading involves identifying two related securities, typically stocks, that have historically exhibited a strong correlation in price movements. Traders then look for deviations from this historical relationship, buying the underperforming security while simultaneously selling the outperforming one. The goal is to profit from the convergence of prices back to their historical relationship. Pairs trading can be implemented using various metrics to measure the relationship between securities, such as cointegration or correlation coefficients.
Reference [1] examines the profitability of pairs trading in the German market. It investigates four methods for selecting pairs: the Distance Method, Cointegration Method, Copula Method, and Ensemble Method. The author pointed out,
This study revisits well-established pairs trading strategies and applies them to the German stock market, for which – despite its large size and sophisticated trading infrastructure – comprehensive studies are lacking. All standard approaches are tested over the period 2000 through 2023 and its multiple market regimes, complemented by the proposal for an ‘ensemble technique’ that combines the signal of individual methods.
…
Opportunities in the market for basic approaches such as the Distance and Cointegration Method have decreased after the 2008/09 financial crisis. Copula and Ensemble Method provide consistently positive portfolio returns on employed capital over the entire period. Their equivalent performance for committed capital is reduced primarily because of limited trading opportunities. With the Copula Method still the best-performing over the entire period, its advantage over, say, the Distance Method is much less pronounced.
In short, when factoring in transaction costs, pairs trading using the Distance and Cointegration methods becomes unprofitable. However, pairs trading based on the Copula and Ensemble methods remains profitable.
We believe that commonly used pairs selection methods are overused and no longer yield profits. To trade pairs profitably, traders should explore less conventional methods.
It’s worth noting that this study examined the German market. It would be interesting to see similar research conducted in other developed markets.
Let us know what you think in the comments below.
References
[1] Sascha Wilkens, Pairs Trading in the German Stock Market: There’s Life in the Old Dog Yet, SSRN
The way I see it is, it's impractical and probably not worth it if you're seeking returns out of your own personal capital
I think it's worth it if, you're 1) running a fund seeking diversified returns via some kind of beta neutral or stat arb strategy, 2) have real edge in building models and applying them, or 3) simply interested in exploring the concept.