Our preliminary analysis of 2017 and 2018 tax return data indicates that, on average, effective tax rates were lower for ZIP codes where most of the population identified as white than for ZIP codes with more racial diversity. Also, the analysis indicates that after passage of the Tax Cuts and Jobs Act, larger percentage point reductions are seen in ZIP codes with predominantly white populations.
Equity for Underserved
On January 20 President Biden signed an executive order announcing that the federal government should “pursue a comprehensive approach to advancing equity for all, including people of color and others who have been historically underserved.” Noting that many federal data sets lack the demographic detail “to measure and advance equity,” the order established an Interagency Working Group on Equitable Data; membership consists of several high-ranking officials, including the Treasury assistant secretary for tax policy. Thomas S. Neubig recently recommended that the U.S. tax system “collect information on racial identifiers . . . to better assess the effects of current and potential tax policies on racial inequality.” (Prior analysis: Tax Notes Federal, Mar. 8, 2021, p. 1555.)
It will be a while, if ever, before tax returns include demographic information, such as race identification, that would allow direct statistical analysis of the relationship between taxes and racial equality. In the meantime, we can use indirect methods like the one used here. As a first step toward measuring the relationship between tax and race, we combine the tax information of individual taxpayers for more than 27,000 ZIP codes with their associated census demographic data.
Computations using data from the IRS Statistics of Income division tell us that in 2017, the national average effective tax rate was 14.5 percent, the effective tax rate calculation being income taxes paid divided by adjusted gross income. The TCJA was enacted on December 22, 2017. In 2018 the national average effective tax rate declined to 13.2 percent. In 2018 the Census Bureau estimated that the U.S. population identified as 72.7 percent white, 12.7 percent Black or African American, 5.4 percent Asian, 0.8 percent American Indian and Alaska native, 0.2 percent native Hawaiian and other Pacific islander, 4.9 percent some other race, and 3.2 percent two or more races.
Horizontal Racial Equity
In a 1996 article, Beverly I. Moran and William C. Witford carefully reviewed major provisions of the U.S. individual income tax law and concluded that “even if income is held constant, the Internal Revenue Code systematically disfavors the financial interests of blacks” (Moran and Witford, “A Black Critique of the Internal Revenue Code,” Wis. L. Rev. 751 (1996)). As a simple check on this conclusion, we created a data subset of 400 ZIP codes, where the average AGI was approximately equal to the national median AGI of $56,000 in 2018. We then matched demographic data to those ZIP codes, ranked them by the shares of their population that identified as a racial minority, and divided that sample into deciles of 40 ZIP codes each. We then calculated the (unweighted) average effective tax rates for each decile.
Figure 1 illustrates the relationship between ZIP code racial makeup and effective tax rates for 2017 and 2018. It shows that generally, as the racial minority proportion of a ZIP-code population grows, so does the effective tax rate. In 2017 the average effective tax rate of the lowest decile (0.2 percent racial minority) is 10.1 percent, and the average effective tax rate of the highest decile (47.1 percent racial minority) is 10.9 percent. In 2018 the average effective tax rate of the lowest decile (0.2 percent racial minority) is 8.5 percent, and the average effective tax rate of the highest decile (47.1 percent racial minority) is 9.8 percent.
Of course, the change in tax burdens between 2017 and 2018 can be attributed to more than passage of the TCJA. One major additional factor weighing heavily on any change in effective tax rates between years is changing levels of income. Because we have progressive tax rates, a ZIP code with a significant increase in income between 2017 and 2018 can be expected to have a larger increase in its effective tax rate (all other things being equal) between 2017 and 2018. To remove this bias, which would skew any interpretation of changes between those years, Figure 2 repeats the calculations of Figure 1, but this time the 400-item sample includes ZIP codes with close to median income and requires those ZIP codes to have growth of AGI close to the national average (about 5.7 percent).
Figure 2 shows that in 2017, the average effective tax rate of the lowest decile (0.6 percent racial minority) is 10.3 percent, and the average effective tax rate of the highest decile (47 percent racial minority) is 11.1 percent. In 2018 the average effective tax rate of the lowest decile (0.6 percent racial minority) is 8.6 percent, and the average effective tax rate of the highest decile (47 percent racial minority) is 10 percent. Of the 400 observations used in the second figure, approximately 360 are different from those in the first figure. The similarity of results from these different data subsets provides some assurance that the results don’t depend on some fluke in the samples.
Both figures show a clear pattern: ZIP codes with larger racial minority populations have higher effective tax rates. This conclusion is derived only for ZIP codes with median income. Future work will examine other income levels.
The TCJA and Race
To examine the effects of the TCJA, we gathered from our data set of 27,000 the 400 ZIP codes with AGI growth rates closest to the national average of 5.7 percent (ranging from 5.6 percent to 5.8 percent). We then divided that data into 10 deciles, sorting them by size of the tax reduction from 2017 to 2018. Figure 3 shows that in general, for ZIP codes with large racial minority populations, the percentage point reduction in tax rates was smaller than the reduction in predominantly white ZIP codes. For the lowest decile (37.1 percent racial minority), the effective rate reduction was 0.6 percentage points. For the highest decile (11.4 percent racial minority), the effective rate reduction was 2.3 percentage points. These results suggest that the TCJA provided more benefits as a percentage of AGI to white over racial minority taxpayers.
It is important to emphasize that racial skewing of tax benefits can be entirely unintentional. Neubig points out that “although U.S. federal laws don’t have explicit differences in tax rules by race (disparate treatment), the outcomes of those facially ‘race neutral’ rules can have differential effects across racial groups (disparate impact).” On that point Moran and Whitford write: “We make no accusations of discriminatory intent. We suggest that the Code reflects systematic black political underrepresentation in the halls of power.”
Conclusion
A great deal more analysis can be conducted with merged tax and demographic ZIP code data. Because of the heightened interest in measuring potential unintentional racial biases in government programs in the future, we hope to provide insight through the use of different comparisons, larger samples, and more sophisticated statistical techniques.
In addition to making a preliminary attempt to address two questions using specific data — about racial equity in the tax law and about the racial equity of the TCJA using merged tax and census ZIP code data — this article is a preview of analyses that will likely become more common in light of Biden’s executive order on advancing racial equity.