A growing number of papers measuring U.S. wealth inequality and its continuing growth were published over the past year. One of those key papers, by economists Emmanuel Saez of the University of California-Berkeley and Gabriel Zucman of the London School of Economics, finds that the share of wealth held by the top 0.1 percent of families in the United States grew from about 7 percent in the late 1970s to 22 percent in 2012. Yet it’s important to note that Saez and Zucman’s results and similar estimates look at the distribution of wealth before accounting for the impact of taxation. A new paper looks at the post-tax distribution of wealth and finds that the federal income tax system is doing significantly less to reduce wealth inequality than in the past. And there are signs that the federal tax system in recent years might actually be increasing wealth inequality.

The paper by economists Adam Looney at the Brookings Institution and Kevin B. Moore at the U.S. Federal Reserve looks at trends in wealth inequality from 1989 to 2013 using data from the Fed’s Survey of Consumer Finances. This survey is particularly useful for measuring for wealth inequality because the survey oversamples families at the very top of the income and wealth ladders.

What Looney and Moore are looking at specifically is the mechanical impact of total income taxes on wealth inequality. Two trends emerged during the period of 1989 to 2013 that affected the relationship between taxes and wealth inequality. The first is the increasing ownership of tax-deferred assets among U.S. families. These assets include 401(k) retirement plans and the capital gains on a financial asset that hasn’t been sold yet (also known as unrealized capital gains). Families up and down the wealth ladder have increased their holding of these kinds of assets, but those at the top are by far more likely to hold a lot more of them. Almost every family in the top 1 percent has unrealized capital gains compared to 25 percent of the bottom 90 percent of families.

These unrealized capital gains are obviously unequally distributed—the difference between pre-tax wealth inequality and post-tax wealth inequality. These tax-deferred capital gains will be realized and taxed eventually, yet at the same time that more assets have been deferred from taxation the two authors find that federal income tax rates have been on the decline, both on ordinary income (essentially wages and salaries) and on capital gains.

Looney and Moore show how these rate cuts, particularly for realized capital gains, resulted in significant declines in the effective tax rates for these kinds of assets. While rates declined for every taxpayer, they declined the most for families in the top 1 percent. In 1989, the rate was quite similar for all wealth levels, but starting in 1998 the rates started to diverge, with the rate for the top 1 percent declining the most. The authors point to the large decline in the capital gains tax rate—from 28 percent to 15 percent during this time period—as the main driver of this decline in top tax rates.

The result of an effective tax cut that favors those at the top is, unsurprisingly, a decreasing reduction in wealth inequality. One way Looney and Moore show this is by comparing the share of pre-tax wealth owned by different sections of the population to the share of post-tax wealth they own. For the top 1 percent, the after-tax share is lower than the pre-tax share from 1989 to 2007, but the difference is declining over this period of time. But by 2010, the post-tax share of the top 1 percent is actually larger than the pre-tax share.

In short, the federal income tax code seems to be increasing wealth inequality.

One possible concern is how the timing of the Great Recession in 2007-2009 affects the two authors’ finding that the tax code shifted from doing less to reduce wealth inequality to outright increasing inequality right after 2007. Maybe the change is due to the huge collapse in housing prices, which resulted in a big hit to unrealized capital gains and the actual value of housing wealth for households in the bottom 90 percent of the wealth spectrum. Looney and Moore don’t include deferred housing capital gains in their analysis due to complications in how those gains have been taxed over the years. So it’s possible that this omission makes tax liabilities for the bottom 90 percent of households high as a share of total wealth, reducing their post-tax wealth, and thus increasing wealth inequality.

Looney and Moore’s analysis is, as they note, the first attempt to analyze trends in post-tax wealth inequality. So their paper is just the beginning of the investigation into this area. But if their results hold up they would have strong implications for how we think about the tax code and wealth inequality.