State progressive taxes have no effect on personal income growth
Chernick, Hunter College, Professor of Economics, 10
(Howard, 2010, SAGE, “Redistribution at the State and Local Level: Consequences for Economic Growth”, DOI: 10.1177/1091142110375694, Accessed 7-9-23, ED)
The first row of table 3 indicates that the progressivity of a state’s tax structure does not have a statistically significant effect on the rate of growth in personal income. The first 2 rows of table 4 show that neither high- nor low-income tax burdens separately affect economic growth. A variety of robustness tests are presented below.
The statistical significance of the neighbor effects is robust to the use of instrumental variables (columns 3, 5, and 7). The magnitude of the Instru- mental Variable estimate is close to the Ordinary Least Squares results for per capita income growth. However, the IV estimate of the neighbor coeffi- cients in the ten-year aggregate model (column 7) is almost twice as large as the OLS estimate. This is consistent with the expected downward bias in the aggregate growth equation.
The main structural feature of more progressive state–local tax systems is greater reliance on income taxation. Because the income tax is the tax most likely to depart from the benefits received principle, any negative influence of tax incidence on growth should be most strongly revealed by focusing on the income tax. Therefore, in columns 1 and 2 of the top panel of table 5, tax incidence is measured in terms of the first and fifth quintile burdens of the income tax alone. The results are consistent with the previous results. Income tax burdens do not a have a significant effect on growth, whereas the neighbor burden on the top quintile has a significant positive effect on own-state growth.
The various tax burden coefficients (BURD1, BURD5, INCTAX1, and INCTAX5) show the effect on economic growth, controlling for burdens elsewhere in the income distribution. Because the model controls for redistributional spending, an increase in the quintile-specific tax burden implies an increase in government revenues, which are allocated to mainly non-redistributional functions (Helms 1985). The lack of a tax effect on growth could reflect the offsetting effects of additional spending on public goods. Therefore, to control for total spending, columns 3 and 4 of table 5 include total taxes as a share of personal income. Only the pro- gressivity and average tax burden variables are shown in the table. In this specification, the tax progressivity variable measures the effect of a shift in the distribution of taxes, controlling for the overall level of taxation. None of the basic results are affected by including the aggregate tax burden.
Table 6 presents a number of additional robustness tests for the tax results. To investigate whether fiscal redistribution on the tax side operates with even longer lags than accounted for in our basic specifications, con- temporaneous fiscal variables are replaced by their five-year lagged values. The results, presented in columns 1 to 4 of table 6, show no significant dif- ference in effect. I also estimated a model that included both the initial and the lagged values of the fiscal variables. The results, not presented here but available from the author on request, are unchanged.
Up to this point, tax burdens have been measured for state-specific income quintiles. Tax rates at a given income level have been used more commonly in the tax and growth literature (McGuire and Wasylenko 1985). Average income in the state-specific top quintile is substantially higher in rich states than in poor states.21 Hence, I am not measuring the difference in potential tax burdens faced by an individual with a given income. To address this issue, I adjusted each state’s tax burden to reflect the national income distribution.22 The adjusted rate serves as a pseudo- instrument for the tax progressivity that would be faced by a family with given earnings capacity in different states. The results for five-year growth are shown in table 6, column 5. The absence of an own-tax effect is unchanged.
Theoretically, tax burdens should be measured net of the federal deduct- ibility offset. However, if taxpayers perceive taxes in terms of gross bur- dens, then the latter burdens may be more relevant for economic performance. The results for five-year per capita growth are shown in table 6 (column 6). The lack of effect of tax progressivity is again unchanged. Although not shown here, a similar result holds for the other growth measures.