Learning to Love Progressive Taxes

• In Event: Experiments in Political Economy

Abstract

One of the greatest puzzles in contemporary US politics is the inability of the current political system to counterbalance the massive rise of economic inequality in the recent decades (Bonica et al., 2013). Many observers argue that a primary reason for this is a breakdown of democratic responsiveness: the affluent use their resources to divert policies from the ones preferred by the median voter (e.g. Gilens, 2012; Gilens and Page, 2014). Others blame the ignorance of the mass public: they argue that most citizens are simply unable to articulate preferences that reflect their own material interests (e.g. Bartels, 2005; Sides, 2015).

While there exists research exploring the possibility that the mismatch between redistributive preferences and self-interest is due to voter ignorance, the results of these studies are mixed. For instance, while Bartels (2005) argues that support for the Bush tax cuts was due to voter’s general ignorance about the redistributive consequences of this policy change, Lupia et al. (2007) show that more informed citizens were in fact not less likely to support the cuts. Studies that test the effect of providing information about the redistributive consequences of progressive tax changes (Sides, 2015; Kuziemko et al., 2015) demonstrate that even strong frames and powerful information treatments did not lead to substantial increase demand for progressive taxation.

The experiment that we propose diverges from earlier studies in that our policy of interest is not about the amount of tax revenues or spending but the way in which tax revenues are raised (i.e. progressivity holding the amount of tax revenues fixed). Doing so allows us to estimate the impact of information of preferences about progressivity as measured separately from possibly distinct attitudes about the scope of government in general (see Ballard-Rosa et al., 2018). We focus our attention on the redistributive consequences of the choice of state taxes. While most of the public opinion research on redistributive preferences has centered on the federal tax system or federal spending programs, shifting our focus on state taxes offers important advantages.

First, given that the tax systems in almost all states are regressive (e.g. ITEP, 2015) it is at least plausible that such policies are in effect because voter’s fail to realize the redistributive consequences of state taxes. Second, the progressivity of state taxes are mostly driven by a simple choice on the part of states: the more a state relies on sales taxes as opposed to income taxes, the more regressive overall taxes become (ITEP, 2015). Finally, using the preferred mix of state taxes as an outcome makes it possible to compare how representative the currently observed state taxes are and how they would be if the electorate was more informed.

The hypotheses that we seek to test in this project is that the reason why state tax policy is so disconnected from the material interests of citizens is that most individuals simply do not understand how state tax policy maps to the redistribution of income. This necessitates the measurement of three variables of interest: (1) preferences about the progressivity of state taxes (2) knowledge about the mapping between the use of tax types and progressivity and (3) policy preferences about specific taxes. The key expectation is that the information treatment would align (1) with (3) through increasing (2).

The proposed intervention is extremely simple. We provide respondents in the treatment group with the information that on average, income taxes are progressive, while sales taxes are regressive (see the question wording in the next section). We expect this intervention to have two distinct effects. First, to the extent that most people would prefer taxes that are progressive, we expect individuals in the treatment group to be more favorable to income taxes as compared to sales taxes. Second, we expect that policy preferences in the treatment group will become more strongly related to underlying attitudes about redistribution.

Author

• Gabor Simonovits, New York University