Asset-Pricing Test Design (jf-empirical-design)

When to trigger

• You have a candidate predictor / anomaly / factor and must decide how to test it
• You are unsure whether to run Fama–MacBeth, time-series factor regressions, or a panel
• You report t-stats but have not addressed the standard-error subtleties of cross-sectional asset pricing
• A referee will ask "is this data mining / does it survive multiple testing / does it work out of sample?"

Scope: this skill is for asset-pricing tests. For corporate/empirical causal effects, route to jf-identification.

Choosing the test