Don't Mix What Should Be Separated: Why Combining Value and Momentum Signals Destroys Alpha
This paper compares two approaches to combining value and momentum factors in U.S. equities: integrated ranking versus separate sleeves. Using long/short dollar-neutral portfolios over 2000-2026, the study finds that maintaining independent factor portfolios delivers superior risk-adjusted returns (Sharpe 0.168 vs 0.157), substantially lower drawdowns (-17.5% vs -26.6%), and outperforms by 52 basis points annualized when volatility-matched, driven by preserving the negative correlation between the two factors.
Read on SSRNFactor Timing: Dynamic Portfolio Allocation via Machine Learning
This paper explores the application of multinomial logistic regression for dynamic factor timing across Size, Value, Momentum, Quality, and Low Volatility factors. Using macroeconomic, sentiment, momentum, and valuation indicators with backward selection, the model dynamically adjusts portfolio weights based on predicted relative factor performance. Out-of-sample results show the Hierarchical Portfolio achieves a Sharpe ratio of 0.74 and 13.51% annual return, outperforming both equal-weighted (0.67) and Sharpe-optimized (0.68) static benchmarks.
Read Paper (PDF)