Tax Burden and the Mismeasurement of State Tax Policy

Tax Burden, defined as the ratio of total tax revenues over personal income, is prominently used to summarize state tax policy. We analyze the empirical relationship between changes in Tax Burden and changes in state tax policy from 1987 to 2000 – as measured by states’ own forecasts of the revenue impacts of tax legislation – and find that Tax Burden contains substantial measurement error. We explain this result by decomposing Tax Burden changes into three components: (1) changes in state tax policy, (2) income-induced changes in revenue that are unrelated to state tax policy, and (3) other factors that do not measure state tax policy. We empirically demonstrate the statistical significance of the second component, highlighting important consequences for studies that estimate the impact of taxes on economic growth.