Out-of-sample Predictability of the South African Equity Risk Premium Distribution: A Quantile Regression Approach

This paper explores the out-of-sample predictability of the South African equity risk premium (ERP) distribution through a quantile regression framework. Empirical results show that beyond central quantiles, several predictor variables exhibit statistically and economically significant predictive ability, reinforcing evidence against the location shift hypothesis which proposes that predictor variables affect only the location of the ERP conditional distribution. Furthermore, combining out-of sample forecasts from various parts of the ERP distribution, a robust out-of-sample approximation of the mean ERP is attained under a 5-quantile post- least absolute shrinkage and selection operator specification with a time-invariant weighting scheme.