Evaluating how factor premia vary over time and across asset classes is challenging due to limited time series data, especially outside of U.S. equities. We examine four prominent factors across six asset classes over a century. We find little evidence for arbitrage activity influencing returns, though some novel evidence of overfitting biases. We identify meaningful time variation in factor risk-adjusted returns that appears unrelated to macroeconomic risks, supporting other theories of dynamic return premia. Attempting to capture this variation, we evaluate various factor timing strategies, but find relatively modest predictability that likely fails to overcome implementation frictions.

Keywords: Factor premia, factor timing, multi-asset class, overfitting bias, arbitrage activity

Continue reading here.