Conditional Factor Models Explain Momentum and Long-Term Reversal

Summary

The article examines whether momentum and long-term reversal reflect changing exposures to priced risks rather than standalone return patterns. It tests whether past returns predict future realized betas, then uses an Instrumented Principal Component Analysis (IPCA) framework to estimate latent factors and time-varying stock loadings from equity characteristics. It compares model-implied expected returns, conventional momentum, and residual momentum. In the reported US stock sample, momentum and long-term reversal predict future exposures to market and other common factors. The conditional model accounts for much of momentum’s predictive content: its model-based strategy is stronger than simple momentum in and out of sample, while residual momentum is weak. A market-only conditional model is insufficient, and short-term reversal appears distinct. The authors caution that the model does not explain why the underlying factors are priced, and the evidence comes from a particular historical sample and modeling framework.

Key ideas

• Past momentum and long-term reversal signals are associated with future changes in factor betas.
• IPCA uses stock characteristics to estimate latent factors and time-varying factor loadings.
• Model-implied conditional expected returns capture much of the predictive content of momentum.
• Residual momentum is weak after accounting for the conditional factor model.
• The analysis does not identify the economic reason the latent factors command risk premia.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.