Quantile forecast combinations

Whether it is possible to improve point, quantile and density forecasts via quantile forecast combinations is tested. The models we employ are quantile autoregressive and mean regression models augmented with a plethora of macroeconomic and financial variables. Complete subset combinations of both linear and quantile forecasts enable us to efficiently summarize the information content in the candidate predictors. We also develop a recursive algorithm that selects, in real time, the best complete subset for each predictive regression quantile. We provide two forecasting applications; one related to stock market return forecasting and the second on realised volatility forecasting. We show that our approach delivers statistically and economically significant out-of-sample forecasts relative to both the historical average/autoregressive benchmark and the complete subset regression approach.