The Logic of Factor Portfolios: Boolean Rules, Interactions, and Trading Costs
In briefLogic-based factor portfolios: how rule interactions shape returns and trading costs
SummaryThe paper expresses investment rules that hold stocks by signal rank as Boolean functions of threshold statements and uses a Walsh decomposition to split returns into single-signal and interaction exposures. It argues that interactions are less persistent than their components, so rules loading on higher-order interactions trade more. In U.S. stocks from 1973 to 2024, predicted persistence and turnover closely match realized values. Higher-order interaction portfolios face higher trading costs, yet out of sample they raise the Sharpe ratio of single-signal tangency portfolios.
WhyThe authors unify factor portfolio returns and turnover via Boolean rules and Walsh decomposition, reporting close agreement between predicted and realized persistence and turnover on U.S. stocks.