The full paper is available here as an Adobe Acrobat PDF. 

Introduction

The study of the incidence of the U.S. tax system has a long and impressive history that has proceeded along many different complementary paths. Arnold Harberger (1962) laid the groundwork for distributing the burden of taxes. Beginning with the pioneering work of Joseph Pechman and Benjamin Okner (1974), scholars have explored the impact of a wide array of taxes on the distribution of income. Theoretical work has subsequently incorporated distributive objectives into optimal tax design (see Auerbach 1985). Other work has explored the direct and indirect impacts of various tax reform proposals and also identified the effects that different tax reforms have on the distribution of income across income classes (See Carasso and Steuerle 2003; Burman and Saleem 2004; Gale and Orzag 2004; and Devarajan, Fullerton, and Musgrave 1980).

While the existing literature presents interesting snapshots of the impact of the U.S. tax system on low-income individuals, these snapshots are difficult to assemble into a complete view, something that was last done exhaustively in Pechman (1985). While ample data exist documenting changes in federal tax policy over time, information on state and local taxes, especially sales and property taxes, is less abundant. However, studies of state taxes have been done recently by both the Institute on Taxation and Economic Policy (2003) and the Center on Budget and Policy Priorities (2004).

Income distribution issues are, in addition to being of academic interest, often a key factor in public policy debates. Proponents and opponents of a particular bill often emphasize the distributional characteristics of their proposal. For example, in the 2000 presidential election, GeorgeW. Bush presented evidence suggesting that his tax changes would significantly improve the welfare of low-income individuals. Others presented evidence to the contrary.

Inferences concerning distributional issues, of course, depend on the impact of specific bills on the entire income distribution. In this paper, we abstract from this, and focus our attention on low-income individuals. Specifically, we attempt to gather data on every tax payment made by typical families to government at all levels in the U.S. for the longest period that data will allow. We collect these observations into an aggregate payment--including sales, excise, income, property taxes, and even at times the lottery--and explore the movements of this aggregate variable over time. We document movements both in marginal and in average tax rates and payments over time.

To foreshadow our conclusions, we find that total direct taxes paid by low-income families have declined significantly, dropping especially sharply since the late 1990s. This overall result reflects a number of different factors. Federal income taxes have declined sharply for families because of the refundable child tax credit (instituted in 2001 for taxpayers with less than three children), expansions of the Earned Income Credit (EIC), and lower marginal tax rates. At the same time, state sales taxes paid have increased significantly, but not enough to offset the decline in federal taxes paid. Payroll taxes and property taxes stayed relatively constant throughout the last several decades.

Together, these results lead to our second finding. Non-income taxes as a percentage of total taxes for low-income families has increased sharply over time. One possible explanation for this pattern may be that issues of fairness appear to have put significant downward political pressure on tax liabilities for low-income families for taxes whose incidence is readily observable, but not for taxes, such as sales taxes, with less obvious distributional effects. Another explanation is that states, which levy most of the non-income taxes, tend to be less redistributive than the federal government. Finally, while overall taxes paid have declined notably, the phase-out range of the EIC applies relatively high marginal tax rates to low-income families. Accordingly, while average tax rates have declined, marginal tax rates have, in some income ranges, done the opposite. The next section outlines the methods we use to identify the movement of tax payments over time. Section III presents our results. Section IV concludes. 

Kevin A. Hassett is a resident scholar and director of economic policy studies at AEI. Anne Moore is associated with the Department of Economics at the Univeristy of California-Berkeley.