邻里变化、住房市场动态和财产税中的种族不平等（英）

Neighborhood Change, Housing Market Dynamics, and Racial Inequality in Property TaxesKELSEY O’HOLLARENThis document was submitted as a dissertation in September 2025 in partial fulfillment of the requirements of the Frederick S. Pardee Ph.D. in Policy Analysis at the RAND School of Public Policy. The faculty committee that supervised and approved the dissertation consisted of Jonathan Welburn (chair), Jason Ward, and Skylar Olsen. This dissertation was generously supported by RAND’s Rothenberg and James Q. Wilson Dissertation Awards along with money from RAND’s Tech and Narrative Lab.Approved for public release; distribution is unlimited.For more information on this publication, visit www.rand.org/t/RGSDA4498-1.About the RAND School of Public PolicyThe RAND School of Public Policy has specialized in graduate-level policy education since its founding in 1970. The RAND School is home to the Frederick S. Pardee Ph.D. in Policy Analysis, which is the original public policy Ph.D. program in the United States and the only Ph.D. program based at an independent public policy research organization. To learn more about the RAND School of Public Policy, visit www.rand.edu.Published in 2025 by the RAND Corporation, Santa Monica, Calif. is a registered trademark. ii Abstract Persistent racial wealth inequality in the United States is closely tied to differences in homeownership rates and property values. Racial inequality in property taxes contribute to this inequality yet remain an essential source of revenue for local governments. This dissertation investigates the mechanisms and extent of racial inequality in property tax burdens. It combines national quantitative analyses with interviews from county assessors in Washington State. The study uses administrative datasets, including Zillow’s ZTrax, mortgage application data sourced from the Home Mortgage Disclosure Act, and the American Community Survey. To analyze properties from a wider range of regions, the research imputes homeowner race using RAND’s Bayesian Improved First name-Surname Geocoding algorithm. The study also compares models that use imputed race with those that use self-reported race to clarify the strengths and limitations of each method. By integrating these methods with census-tract-level socioeconomic variables, the research estimates racial gaps in property assessment ratios. The analysis examines how sales volume, gentrification, and other neighborhood characteristics affect assessment equity for Black, Latino, Asian Pacific Islander, and American Indian/Alaska Native homeowners relative to White homeowners. The key findings show significant racial disparities in property tax assessments, with minority homeowners experiencing higher relative tax burdens in areas with lower housing market activity. In census tracts with high sales volume, these gaps are smaller or, in some cases, statistically insignificant for certain minority groups. This suggests that increased transaction data a