Abstract:
This paper critically evaluates the Urban Institute’s Land Use Reforms and Housing Costs (2024) study, which aims to link zoning reforms to housing supply using machine-learning analysis of newspaper articles. The study identifies 180 “major” land-use reforms across eight U.S. metropolitan areas and concludes that “less restrictive” reforms increase housing stock by only 0.8 percent within three to nine years.
Our independent review finds that 60 of articles are invalid due to duplication or geographic miscoding. Of the remainder, 118 are misclassified, minor, or lack sufficient information to determine direction or significance. Only two cases plausibly qualify as major reforms— and even these are uncertain in measurable impact under Urban’s methodology.
Urban’s reliance on unverified media sources, simplistic binary classifications, and limitations in the underlying address data produce substantial methodological errors. In contrast, the AEI Housing Center’s case studies show that simple, by-right zoning reforms can expand supply by 1–2.5 percent per year, underscoring the need for rigorous, context-sensitive approaches to measuring policy impact.
