Abstract:
|
This paper uses the 2019 American Housing Survey to evaluate the representativeness of two different sources of property tax records. Leveraging the U.S. Census Bureau's linkage infrastructure, we compute the fraction of AHS housing units that could be matched to a unique property parcel (coverage rate), as well as the extent to which survey and property tax data contain the same information (agreement rate). We analyze heterogeneity in coverage and agreement across states, selected housing characteristics, data sources, and 11 AHS items of interest. We find little variation in coverage or agreement across data sources. Our results suggest that partial or adaptive replacement of AHS data with property tax data, targeted toward certain survey items or single-family detached homes, could reduce respondent burden without compromising data quality. Our work can guide other comparisons between survey and administrative housing data, and orient those who wish to conduct population research with property tax records.
|