One of the most contentious tax policy issues during the 2012 presidential election season involved the federal income tax rate reductions defined in the Jobs and Growth Tax Relief Reconciliation Act of 2003. Signed into law by President George H. W. Bush, the act lowered the marginal income tax rate structure across all income levels but was set to expire on January 1, 2011. The candidates of both parties debated whether the act should be extended, modified, or left to expire and revert to the previous, higher marginal tax rate structure. Concerns over the US economy’s faltering recovery after a deep recession pushed Congress to pass, and President Obama to sign, the Tax Relief, Unemployment Insurance Reauthorization and Job Creation Act, which extended many of the Bush tax cuts until then end of 2012. Then on January 2, 2013, President Obama signed a last-minute “fiscal cliff” tax bill produced by Congress that, among other things, extended the lower federal income tax rates for all but the percent of US households with the highest incomes (as defined by the latest IRS data on the adjusted gross incomes of all taxpaying US households). The debate over federal income tax burden fairness continues today.
Many aspects of the debate have involved contradicting claims that were, ostensibly, empirically testable. Examples include whether the lower marginal income tax rates increased or decreased total income tax revenues via a supply-side effect, or whether the lower income tax rates encouraged more or less economic growth in the long run. However, the most volatile aspects of the debate have concerned whether upper-income American households were shouldering their fair share of the total federal tax burden. Indeed, the media have given widespread attention to the increasing gap between the highest- and lowest-income groups in American society, which suggests that federal tax policy fairness is an issue of particular interest to voters.
Whether a given income tax scheme is “fair” is a decidedly normative question. However, analysis with positive claims about competing federal income tax schemes can be developed to enlighten this normative debate. While previous attempts have been made to compare the individual ability of American taxpayers to pay their taxes with the actual tax burdens they face, they often fall short of their goal because the issue of assessing whether upper-income households are shouldering an appropriate level of the federal income tax burden is complicated. Americans face a complex maze of federal taxes, including income taxes, payroll taxes, corporate income taxes, estate taxes, excise taxes, and more. Further, it is difficult to measure each American’s ability to pay these taxes. Should tax policy makers be concerned with the tax burden that each individual will bear over a lifetime? Should the stock of an individual’s personal wealth be added to the flow of personal income when assessing the fairness of his or her tax burden? These are challenging questions. Yet, once a consensus on the proper set of overall income and tax-burden distribution data is reached, there remains the difficult task of properly comparing income and tax-burden distributions in a clear and testable manner.
In this light, economists have developed the concept of tax progressivity. A given tax rate scheme is effectively progressive if a person’s average tax rate on income increases with the level of his or her income. This means that over time, a tax structure has become more progressive if the ratio of income used to pay federal taxes rises among higher-income earners or falls among lower-income earners. The degree to which a proposed tax scheme shifts a greater share of this tax burden onto the higher-income groups should be a quantifiable proposition that can meaningfully inform the normative debate on federal tax policy fairness, provided a proper dataset is chosen and an appropriate index is developed.
The goal of this paper is not to answer the normative question of whether America’s federal tax scheme is fair. Nor is it to identify the proper dataset to employ in this endeavor. Rather, we set out to establish the best way to effectively measure tax progressivity, which is an important step toward enlightening the ongoing debate about federal tax policy fairness. The following is a proposition for determining the best methodology and the resulting index for measuring the degree of an income tax scheme’s progressivity, as well as an improved method for estimating such indexes from empirical data comparing the income and tax-burden distributions across an entire tax base.
First, we propose a set of qualitative principles for critically evaluating any taxprogressivity methodology that can produce measures of progressivity, referred to hereafter as indexes. We apply these principles to two well-established indexes (Kakwani 1977; Suits 1977) and to a more recent index (Stroup 2005). We also examine the informal method of analyzing progressivity developed by Piketty and Saez (2007). We show that the Stroup index arises from a methodology that is superior to these others, making it a more accurate and more reliable index of overall tax progressivity.
Next, we propose a set of principles for evaluating any process to estimate tax progressivity using income and tax-burden distribution datasets. We apply these principles to well-established estimation procedures as well as to a revised method of estimation. We use publicly available IRS data to illustrate the properties of the different estimation methods and find that this revised process is superior for estimating any tax-progressivity index.
Finally, we use annual IRS data from the last quarter century to estimate and observe the behavior of the three income tax progressivity measures over time. We show how the Stroup index indicates that federal income tax progressivity has increased over time, whereas the other three methodologies imply that income tax progressivity has declined. We note that this disparity may not be driven by the choice of data used in their analysis, but may be the result of inherent design flaws in the three methodologies other than Stroup’s. Further, we conclude that the Stroup index could accurately reflect the overall federal tax burden by using a more comprehensive income and tax-burden distribution dataset, such as that used by Piketty and Saez. We conclude that in this case, the Stroup index would provide a cardinal measure of overall tax progressivity that is unbiased, comprehensive, and reliable.
Download the PDF