Property Tax Regressivity: Tax Policy, Assessments, and Housing Markets

• In Event: Property Taxation and Local Fiscal Health

Abstract

The property tax is the primary funding mechanism for public education and the main revenue source most local governments control. Yet how the property tax distributes its burden across income levels is poorly understood. A regressive tax takes a larger share of income from lower-income households than from higher-income ones, and existing evidence suggests property taxes are often regressive in this sense. But commonly reported statewide estimates are misleading: they conflate the distribution of the burden of the tax within a community with the sorting of households with differing incomes to communities that tax at higher or lower rates. To understand the property tax itself, these two things need to be separated.

Policymakers choosing between the property tax and alternatives like sales taxes need accurate local estimates of who bears the burden. State school-finance systems, which use property wealth per pupil to gauge how much a district can raise on its own and set aid accordingly, face a subtler problem. Two districts with the same property wealth per pupil are not equally able to raise revenue if the property tax falls much more heavily on lower-income households in one place than the other. Residents of a high-regressivity district bear higher costs of financing school spending, and equal property wealth does not imply equal fiscal capacity. Measuring local regressivity, and understanding what drives differences across districts, would give state policymakers a more accurate and equitable basis for targeting aid.

This paper measures property-tax regressivity at the local jurisdiction level and asks why it differs across places, decomposing it into contributions from tax policy choices, assessment practices, and underlying housing markets.

Data

The analysis uses restricted-access American Community Survey (ACS) data linked to administrative property-tax records. A key feature of the restricted-access files is that they include fine geographic identifiers, among them school district codes unavailable in public-use data, making it possible to assign households to the specific taxing jurisdictions they face. The linked parcel records contribute administrative taxes owed and assessed values, the two quantities needed to measure tax burdens and trace how assessments translate into bills.

Research Design and Methods

I estimate regressivity within tax rate areas (sets of properties subject to the same combination of local taxing authorities) using three measures: percentile-ratio measures, the Kakwani index, and a log-linear measure of how after-tax income co-varies with pre-tax income. Cross-jurisdiction variation is then decomposed using a Shapley counterfactual framework into contributions from tax liability rules (statutory rates, exemptions, and credits), assessment practices (estimated from arm's-length sales within each jurisdiction), and a residual that largely reflects the sorting of higher-income households into higher-value homes.

Results

Preliminary evidence using publicly available ACS data shows substantial within-state variation in local regressivity; statewide measures often fall outside the interquartile range of local ones, confirming that conventional estimates are a poor guide to how the property tax distributes its burden within any given community. Full results are pending Census Bureau disclosure review.

Author

• Ross Teichert Milton, University of Wisconsin-Madison
  Presenting Author