We assess whether data-driven statistical methods and, in particular, forecast combination strategies can provide additional information about expected market returns beyond that of theoretically motivated predictors. The results indicate that averaging forecasts from the theoretically motivated predictors and combination strategies enhances prediction accuracy relative to using each forecasting approach individually. Our findings demonstrate that flexible statistical methods could be used to boost economic theory rather than dilute its importance for equity premium predictability. Yet, forecast combination approaches can extract additional information and no theoretical predictor in isolation is likely to be the expected return on the market.
Complete tool for submission, evaluation and publication of proceedings for scientific events.
