One sentence summary: State tax revenues in the U.S. would benefit from higher wage–income, sales or property taxes, while any increase in dividend–income tax would result in a reduction of revenues.
The corresponding paper by Hakan Yilmazkuday titled "Individual Tax Rates and Regional Tax Revenues: A Cross-State Analysis" has been published at Regional Studies.
The working paper version is available here, where you can find state-specific elasticities of tax revenues with respect to alternative tax rates.
This paper analyzes the effects of state-level personal tax rates on state tax revenue and individual welfare. The policy analysis based on a general equilibrium model suggests that tax revenues would benefit from higher wage-income, sales or property taxes, while any increase in dividend-income tax would result in a reduction of revenues. It is also shown that individuals would suffer from an increase in state-level wage-income tax, dividend-tax or sales tax, while they would benefit from an increase in property taxes. The heterogeneity across states is determined by a TaxIndex, a weighted average of initial taxes at the state level.
The Great Recession of 2007-2009 had a devastating effect on state finances in the U.S. when states took in $87 billion less in tax revenue from October 2008 through September 2009 than they collected in the previous 12 months; this corresponds to a decline of 11 percent, the steepest on record, resulted from the impact on tax collections of reduced wages and lowered economic activity. The requirement that states have balanced budgets has increased the pressure on states to deal with the unprecedented revenue shortfalls in a variety of ways; to recoup lost revenue, states have taken actions such as increasing tax rates. For example, according to the U.S. Census of Governments, the share of tax revenue in overall state revenue has increased from 38% to 42% between 2007 and 2012, on average across states. However, when it comes to the policy details, which type of tax should be modified in each state to improve the state budget? We answer this question by introducing an economic model that is matched by the U.S. data.
The empirical results show that state tax revenue increases with wage-income tax for any state with an average (across states) elasticity of 0.88. When we make a comparison across states, Louisiana, Hawaii and Alabama would benefit most in terms of their tax revenue out of a wage-income tax increase, while states of Iowa, South Dakota, New Hampshire and Texas would benefit less. When we search for a systematic explanation for this result, the economic model introduced in the paper implies that TaxIndex, which is a weighted average of initial taxes at the state level, is the main determinant of the difference across regions. As is evident in Figure 1, the states that have benefited more in relative terms are the ones with a lower TaxIndex.
Regarding the short-run effects of an increase in sales taxes, the average (across states) elasticity of tax revenue with respect to sales taxes is about 0.20; the tax revenue of any state would increase after an increase in its sales tax rate. The elasticity of tax revenue with respect to sales taxes goes down with the TaxIndex; therefore, states with low TaxIndex such as Louisiana, Hawaii and Alabama would benefit most from an increase in sales taxes, while New Jersey, Wisconsin, Vermont and Nebraska would not have any significant changes in their tax revenues. An increase in state-level property taxes would also increase tax revenue of any state; once again, the TaxIndex seems to be the main determinant of tax revenue differences across states in Figure 1. Dividend-income tax is the only tax type that results in lower state tax revenue.
This paper has been mentioned, among others, in Carolina Journal, Bladen Journal, NC Spin The Robesonian, The Mountaineer, New Bern Sun Journal, and The Daily Dispatch (Henderson, NC).