All those hundreds of 'factors' investors use to predict the stock market are really just the same few economic signals in disguise.
April 2, 2026
Original Paper
Factor Identity
SSRN · 6404361
The Takeaway
The 'factor zoo' in finance has ballooned to include hundreds of variables, but this Bayesian analysis shows they lack distinct species. Almost all pricing information is actually just a handful of identities related to long-term macro growth and market risk.
From the abstract
The empirical dominance of dense factor models for pricing the cross-section of expected returns has come at a cost: the economic narrative of priced risk has disappeared. We introduce factor identities-groups of factors conveying the same pricing information despite modest correlations-and a Bayesian method to recover them, restoring legibility to the dense SDF. Only a handful emerge: a macroeconomic identity revealed through long-horizon growth in output, consumption, and industrial production