Pairs Trading Question
Benchmark selection for market-neutral pairs trading
Working on a cointegration-based pairs trading strategy across S&P 500 constituents (daily frequency). Since the strategy is market-neutral by construction, I've been going back and forth on the most appropriate way to benchmark performance.
Approaches I've considered:
- S&P 500 as a passive alternative
- Risk-free rate only, evaluating purely on Sharpe
- Fama-French factor regression to decompose alpha from market/size/value/momentum exposure
- Some combination
Gatev et al. (2006) and Do & Faff (2012) both use factor models to strip out systematic risk, but a lot of other pairs trading papers just report Sharpe ratios without any benchmark comparison.
For those who've built or evaluated market-neutral strategies, is factor regression sufficient on its own, or should cumulative returns also be compared against the index? Any papers that handle benchmarking particularly well for this type of strategy?
Appreciate any input.
Any pointers to papers that handle this well would be appreciated. Thanks.