From Prediction to Causation: Heterogeneous Policy Effects of Rent Stabilization and Rental Assistance on Severe Housing Cost Burden in New York City
Year: 2026
Type: preprint
Study design: Observational Study
Abstract: This paper develops an AI-driven causal inference framework to evaluate the heterogeneous effects of eight major housing policy instruments on severe housing cost burden in New York City. We integrate ensemble machine learning for prediction with triple quasi-experimental designs---difference-in-differences, regression discontinuity, and double/debiased machine learning with causal forests---to establish causal effects while characterizing effect heterogeneity across neighborhood contexts. Our approach advances beyond pure prediction by identifying not only which policies work but for whom and under what conditions. We find that the 2019 Housing Stability and Tenant Protection Act rent stabilization expansion reduced severe burden by 6.2 percentage points, while tenant protection ordinances and housing court access showed the largest complementary effects. Contrary to policy intentions, Emergency Rental Assistance eligibility is associated with increased burden during the pandemic period. Critically, we uncover significant effect heterogeneity: stabilization policies are most effective in high-opportunity, high-eviction neighborhoods, whereas protection and access reforms show stronger impacts in vulnerable areas. Our ensemble ML approach achieves superior out-of-sample forecasting accuracy (R² = 0.87) while providing interpretable causal estimates through doubly robust estimation. These results demonstrate how AI-enhanced causal inference can guide precision housing policy design in urban settings facing affordability crises.
Source: Research Square
Language: English
Topic: Housing Market and Economics
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7204197102