The RP-PCA factors and stock return predictability: An aligned approach

Our study first investigates robust evidence for the predictive power of risk premium principal component analysis (RP-PCA) in forecasting equity returns and macroeconomic activity. We use the partial least squares (PLS) method to extract the optimal information from five RP-PCA factors, and the aligned RP-PCA index appears to outperform the original RP-PCA factors in various in-sample and out-of-sample diagnostic tests with little evidence of instability. Furthermore, the aligned RP-PCA index can generate adequately more profits than most of the other RP-PCA factors in an active market-timing trading strategy in excess of the historical mean forecast strategy. A vector autoregression-based stock return decomposition shows that the economic source of the forecasting power for the aligned RP-PCA index predominantly comes from the future cash flow channel.