Neighborhood Demand Estimation and Ex-Ante Policy Evaluation

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

Government housing policies, such as the Housing Choice Voucher program or the Empowerment Zone program, can change the relative prices of neighborhoods, thereby altering households' demand for neighborhoods. Despite the importance of understanding these neighborhood demand patterns, the existing literature has not estimated neighborhood demand systems intended to capture realistic substitution patterns and evaluate the housing policies. This project seeks to understand how government housing policies would affect demand for neighborhoods, equilibrium neighborhood demographic compositions and price-levels, and ultimately households' well-being. The investigators will develop empirically tractable models of local demand for neighborhoods which place minimal ex ante restrictions on the substitutability of neighborhoods. Using detailed panel data on location choices and credit information, the investigators will estimate the models and further perform program evaluations of existing and yet-to-be enacted government housing programs. This project develops and estimates empirically tractable models of local demand for neighborhoods that allow for unrestricted patterns of substitution across neighborhoods and elasticities of demand for particular neighborhoods. The demand model achieves flexibility by positing a large number of observable types of households, with minimal restrictions on how each type's flow indirect utility differs across neighborhoods. Using panel data on the location choices from 1999 to present for 5% of the U.S. population from the NYFRB/Equifax Consumer Credit Panel, the investigators estimate the model's structural parameters by maximum likelihood, exploiting the conditional-choice-probability inversion techniques of Hotz and Miller (1993) and Arcidiacono and Miller (2011) to construct likelihood functions without having to repeatedly solve the full dynamic model. The estimated model facilitates a variety of policy evaluation methodologies, including regression analyses that treat the estimated neighborhood demand elasticities as explanatory variables; partial equilibrium counterfactual policy experiments that study the impact of various subsidies on households' location choices; and equilibrium counterfactual policy experiments that study the impact of various housing policies on neighborhood rent levels and neighborhood demographic compositions. Specifically, the investigators use the estimated model to study the following policy questions: (1) In which locations is the incidence of place-based subsidies -- Empowerment Zones, Enterprise Zones, etc. -- likely to fall on prior zone residents?; (2) Would a large expansion to Section 8-style housing vouchers lead to more mixed-income neighborhoods in equilibrium?; (3) What targeting rules for housing vouchers are most likely to induce households to choose neighborhoods that contribute positively to their children's academic achievement?