How Robust are Empirical Factor Models to the Choice of Breakpoints?

We comprehensively investigate the robustness of well-known factor models to altered factor-formation breakpoints. Deviating from the standard 30th and 70th percentile selection, we use an extensive set of anomaly test portfolios to uncover two main findings: First, there is a trade-off between specification versus diversification. More centered breakpoints tend to result in less (idiosyncratic) risk. More extreme sorts create stronger exposures to the underlying anomalies and, thus, higher average returns. Second, the models are robust to different degrees. The Hou, Xue, and Zhang (2015) model is much more sensitive to changes in breakpoints than the Fama-French models.