Evaluating the Accuracy of American Community Survey Data on Housing Values and Property Taxes

How accurate are the self-reported property tax burdens and property values often used by policy organizations to develop national property tax comparisons? We compare American Community Survey (ACS) data on home values and property tax burdens for four individual years in 12 Minnesota counties and over a three-year period in 44 counties with amounts reported by local assessors to the Minnesota Department of Revenue (DOR). We find that actual property tax burdens fell within the ACS’ 90 percent confidence margins of error two-thirds of the time in individual years and only half of the time in the three-year period studied. Evidence strongly suggests respondents do not include the effects of homestead property tax relief programs that do not directly lower property tax bills when reporting property tax burdens to the ACS.

Concerning property valuation, we find that actual home values fell within the ACS’ 90 percent confidence margins 30% of the time in individual years and in only 20% of instances during the three-year period we reviewed. Homeowners overestimated the value of their residences nearly 90% of the time in the one-year data and over 80% of the time in three-year data.

These burden and valuation discrepancies have implications for actual and self-reported effective property tax rates. ETRs generated using DOR data were greater in 21 of the 24 cases—5.5% on average—than those derived using ACS data. Over-reporting to ACS of tax burdens in the numerator were more than offset by larger overstatements of home values in the denominator.

The exclusion of special assessments from the DOR data might be a contributing factor in explaining at least some of the discrepancy between ACS-reported and actual tax burdens. Similarly, the time period constituting our analysis—the bursting of the national housing bubble and a period of exceptional real estate market volatility—is likely a major factor in explaining the difference between ACS-reported and assessed home values. Further study in an era of more stable real estate markets is needed to determine if ACS reported home values convey market conditions or systemic homeowner perception of inflated levels of property wealth.