Sections

Commentary

Taxing capital income

Bar graph illustration

Legislated changes affecting capital income have dramatically reduced the federal income tax base and revenues over the past 25 years. A significant share of capital income is never subject to tax. The massive “leakage” between the generation of economic income and the reporting of income on tax forms calls for careful analysis of (a) which forms of income do not show up on tax forms, (b) where in the income distribution that divergence is occurring, (c) how base erosion is changing over time, and (d) the revenue and distributional effects of broadening the capital income tax base.

In a recent project, we examine these issues. First, we construct tax filing unit data from household data in the 1995-2019 Surveys of Consumer Finances (SCF). The SCF is uniquely qualified for this exercise because it has the wealth data needed to link capital incomes (missing and reported) to specific types of assets and specific tax units. Tax data alone are insufficient to address these issues, since the income reported to the IRS is already affected by laws, avoidance strategies, and evasion practices. Second, we compare data from the National Income and Product Accounts (NIPAs), the Internal Revenue Service (IRS) and the SCF aggregate incomes to better understand which types of capital income do not appear on tax forms. Third, we use the NBER TAXSIM model to calculate income tax liabilities under different scenarios.

The project output includes two research papers and policy briefs as well as a publicly available micro data set that creates tax units from SCF households and reconciles economic and tax concepts of capital income – with documentation, code, and data that allow other researchers to use improved methods and data.

We obtain several key results. Figure 1 shows that—after making conceptual corrections to align the NIPA and IRS income concepts discussed above—the fraction of NIPA income that shows up as SOI income varies by type of income and over time. The ratio of IRS to NIPA measures of income from closely held businesses and financial assets is low and has declined over time. The blue line in Figure 1 shows that the ratio was 44 percent in 1994 and declined to 32 percent by 2018. That is, the United States has shifted from taxing less than half of economic measures of business and financial income to taxing less than a third of such incomes over that period. The dashed line shows that removing the adjustments for qualified dividends and qualified business income leaves the ratio at 44 percent in 1994 and 40 percent in 2018, a smaller decline. In contrast, the yellow line shows that the ratio of SOI to NIPA measures of all “other” income has been high and relatively constant, at 86 percent in 1994 and 84 percent in 2018, with most of this small decline occurring only in the last few years of the sample period.

Figure 1

We compare the revenue and distributional effects of two scenarios: one with SCF business income as reported (unadjusted), the other with SCF business income reduced by 50 percent (adjusted). The latter scenario is motivated by the fact that NIPA and SCF closely held business incomes are generally twice as large as those reported on tax returns in SOI data. Thus, the scenario is a rough approximation of what business owners actually report on their tax forms as opposed to what they report on the Survey of Consumer Finance. Figure 2 shows that revenues in the unadjusted scenario (the solid orange line) are well above published SOI values (the blue line). This occurs because the scenario produces higher total income, most of which is accounted for by higher business incomes. The revenue gap is relatively stable over time, which is consistent with a systematic reporting difference. Income tax revenues in the adjusted scenario (the dotted line) are $1,517 billion in 2018, almost identical to the IRS figure, reported at $1,510 billion.

Figure 2

The tax revenue effects of taxing all business income would be substantial. For example, in 2018, tax revenues in the unadjusted scenario are $249 billion, or 16.4 percent, higher than in the unadjusted scenario. The substantial increase in tax revenues reflects both the doubling of business incomes and the above average marginal tax rates that additional business income faces.

The distributional effects of taxing all business income would be substantial as well.  We focus on distributional effects by wealth class because income is endogenous to this exercise. The first two columns of Table 1 provide some perspective on wealth distribution, including, for example, the fact that SCF households with net worth of $10 million or more account for just 1percent of the population, but own 39 percent of the wealth.

The distribution of taxes is very different. In the unadjusted scenario, households with wealth of $10 million or more account for 30.3 percent of taxes. In the adjusted scenario, those same households account for only 27.6 percent of taxes. The last two columns show that if all business income were included in the tax base average tax liability would jump by roughly $80,000, from $287,830 to $367,145 (28 percent) for families with $10 million or more in wealth. Taxes would be changed much less for households with less than $1 million in net worth, on the order of five percent of their average business taxes. These results indicate the potential revenue and distributional effects of bringing all business income into the tax base.

About one in every six tax dollars that are owed to the federal government are not paid.

More generally, policy makers have a variety of choices if they wish to move in the general direction of broadening the income tax base to include more business and other types of capital income.

The first is cracking down on tax evasion. About one in every six tax dollars that are owed to the federal government are not paid. Evasion is particularly high in areas of the economy that do not feature the third-party withholding that is typical of wage earnings. Thus, evasion rates are high among business taxes including, sole proprietors, farm owners, rent recipients, partnerships and so on. Over the past 10 years, the IRS’s own budget has declined precipitously in real terms and the number of auditors of high-income households has plummeted. Funding the IRS on a sustainable basis and improving information reporting – such as the Biden’s Administration’s proposal for banks to report total deposits and withdrawals for accounts with a certain amount of money flowing in or out – would help address these issues.

A second step would be to tax capital gains more fully. Currently, gains are only taxed when they are realized rather than when they accrue, and the biggest loophole in the income tax system is the so-called “Angel of Death” loophole, namely, that capital gains that have accrued on assets held till death are never subject to income tax. Their basis is stepped-up to the current asset value upon wealth transfer. As a result, many business owners have incentive to defer selling the business because gains on the value of the business will go tax-free upon the owners’ death.  The value of the business would be subject to estate taxes but as discussed below, that will not apply to the vast majority of businesses. The obvious choices for capital gains reform would be to (a) remove step up of basis at death, (b) tax the gain at death, or tax all gains every year on accrual. Taxing marked-to-market assets (such as stocks on the NYSE) on accrual would be straightforward. Taxing non-traded assets provides challenges, but these can be addressed via a system that includes interest charges for delayed realization (Auerbach 1991).  President Biden’s proposal to tax the accrued capital gains of people with more than $100 million in wealth is a significant (though complicated) proposal that could put a real dent in capital gains loopholes and raise significant revenue.

A third step would be to tighten estate tax and other wealth transfer rules. The estate tax, which is in principle an important backstop to the income tax, was all but eviscerated in the 2017 tax act, with the exemption level raised to $22.8 million for a married couple. Reducing the exemption and closing the myriad of loopholes associated with the tax could go a long way toward raising revenues, not just at the time of death but during the lifetime of the decedent as well. Wealth transfer taxes are of particular importance in the next few decades because people over age 65 currently hold a greater share of household net worth than at any time in the last 30 years at least. Over the next 2-3 decades much of that wealth is likely to be inherited by the most affluent members of younger generations.  Having the estate tax provide such generous exemptions and loopholes during the coming period of elevated wealth transfers risks losing public revenue and exacerbating inequality.

All these ideas have the potential to raise significant amounts of public revenue, with the burdens concentrated heavily among the wealthiest households.

To conclude, our work is motivated by changes in the distribution of income and wealth, changes in tax policy over the past 25 years, and the presence of long-term federal fiscal shortfalls. All these trends point to the potential importance and fairness of raising taxes on wealthy households. Capital income in general and business income in particular is highly concentrated among such households but the average tax burden on such income has fallen significantly over time. As this project shows, there is significant revenue and distributional impact from broadening the tax base to include more business income and policy makers could pursue a variety of policies to raise burdens on the wealthy and thus raise revenue in a progressive manner.


This brief was funded by the Peter G. Peterson Foundation. The Brookings Institution is financed through the support of a diverse array of foundations, corporations, governments, individuals, as well as an endowment. A list of donors can be found in our annual reports published online here. The findings, interpretations, and conclusions in this report are solely those of its author(s) and are not influenced by any donation.

Authors