Statistical artifact? Risk premium?
Are the trading rules picking up patterns in the data not accounted for by standard statistical models?
- Bootstrapping tests show that they are
Are the excess returns a risk premium?
- Regressions on portfolio benchmarks suggest not