Timing Equity Factors with Conditional Expected Returns and Covariances

Summary

The report presents a framework for dynamically adjusting multi-factor portfolio weights by forecasting factor returns and their covariance. It replaces unconditional expectations with conditional expectations based on external market variables, allowing allocations to respond to changing conditions. It compares different factor sets, historical estimation windows, and conditioning variables, then introduces Akaike information criterion (AIC) screening to select multiple conditioning variables.

The reported backtests suggest that volatility measures, some price-change measures, selected index valuation measures, and turnover can help time factors in some settings. Gains depend on the portfolio and factor set: timing improved TOP100 results in some models without a size factor, while only a few volatility measures improved annualized returns when size was included. Annual results varied, and the AIC-selected model had slightly lower full-period annualized return than its baseline despite a more even yearly distribution. These historical comparisons are specific to the report’s sample and do not establish robust future performance.

Key ideas

• Factor timing adjusts portfolio weights using forecasts of factor returns and their covariance.
• Conditional expectations incorporate external variables to reflect changing market environments.
• The report compares conditioning variables, factor sets, and historical estimation windows.
• AIC screening is used to construct a model with multiple conditioning variables.
• Reported improvements vary by portfolio composition, market year, and evaluation period.

Tags

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