David Roodman

Independent methodologist (formerly Center for Global Development, now Open Philanthropy) who publishes his own statistical re-checks of contested empirical disputes rather than taking a side by assertion. On the Mariel Boatlift wage-effect dispute, his own formal test of Borjas's adjusted wage series found no statistically distinguishable 1980 trend break (p = 0.82 and 0.37 against a constant-decline-rate null) and judged the Clemens & Hunt census-composition argument plausible — but his review predates Clemens's later race-interaction model and does not adjudicate that specific modeling choice. He discloses his own conflict: a former CGD colleague of Clemens, now at an organization that funds CGD and Clemens's migration work specifically. Matters to this vault as the closest thing yet found to a named, credentialed third-party referee in the Mariel dispute, with the limits of that role stated plainly rather than borrowed.

References

• claim-roodman-independent-assessment-finds-census-artifact-plausible-predates-race-interaction-model
• Captures: 2026-08-19-which-race-adjustment-model-is-the-econometrically-correct