Be very skeptical about how much revenue Elizabeth Warren's wealth tax could generate

Sen. Elizabeth Warren, D-Mass., has made her proposed 2% wealth tax on those worth more than $50 million a central part of her presidential campaign. Emmanuel Saez and Gabriel Zucman, two economists at the University of California, Berkeley, who helped developed the proposal, estimated that it would rake in $187 billion a year.

In April, we published a piece in the Washington Post suggesting that this estimate was likely overly optimistic. This week, Saez and Zucman published a rejoinder.

Their response is disingenuous. They focus most of their fire on what they label as our revenue estimate: that the proposed wealth tax would raise $25 billion annually, rather than the $187 billion they estimate. In reality, we are explicit that $25 billion is a rough back-of-the-envelope number and state that "we would be surprised if the $25-billion-a-year figure we suggest was not a significant underestimate of the revenue potential of a 2% wealth tax." The purpose of our piece was not to provide an alternative revenue estimate for the wealth tax but to call into question the naively high estimate provided by Saez and Zucman.

Readers may find it difficult to evaluate the technical aspects of this debate. But they should keep in mind that Saez and Zucman are writing in a commissioned letter to a presidential candidate for a campaign proposal. Meanwhile, experts with experience in tax policy scorekeeping and academics on the progressive end of the political spectrum — such as ourselves and MIT economist Jonathan Gruber — are much more pessimistic on the revenue-raising potential of wealth taxation. In this context, Saez and Zucman's indignant insistence that their $187 billion number is not a "best case scenario" but instead a "middle-ground" estimate seems unsound.

As we explain at some length in our piece, naive estimation of the kind offered by Saez and Zucman tends to be way optimistic relative to scorekeeping by government experts. This point is well illustrated by the difference between academic and government estimates of taxing carried interest as ordinary income or of the value-added tax. Nothing in Saez and Zucman's response suggests they are immune from this problem.

They attempt a broad allowance for tax avoidance, assuming the rich would successfully shelter at most 15% of their wealth from taxation. They base this guess on four academic studies that consider the international experience of wealth taxation, which find that a 1% wealth tax reduces reported wealth by 0.5 and 35%, which they simply average to 15%. But this strikes us as too low.

First, Saez and Zucman's interpretation of the international experience differs from ours. They rely on estimates suggesting that a 1% wealth tax in Denmark and Sweden results in evasion of less than 1% (which makes their 15% estimate look huge). But in both countries, wealth taxation proved so easy to avoid and so difficult to administer that these taxes were repealed. In fact, of the 12 nations in the Organization for Economic Cooperation and Development that had wealth taxes in 1990, only three still have them today.

It is worth noting that estimating the tax base for those worth more than $50 million in itself is a difficult task — let alone estimating the revenue that taxing these households can raise. Different approaches to measuring top wealth can paint very different pictures. And the numbers reported in the Forbes 400, which Saez and Zucman rely on repeatedly in their rejoinder, are thought of by many as dubious.

This isn't to say that our method should be viewed as definitive, but it does suggest the Saez and Zucman estimation is likely too high. As an illustration of the crudity of their analysis: They neglect to contend with behavioral responses that would inevitably follow a 2% tax on a small group of wealth holders. For example, there would be a significant incentive to accelerate charitable giving, which would decrease the wealth tax base. It seems important to account for the fact that the wealthy (and their tax planners) will inevitably be motivated to limit tax liability.

Saez and Zucman are at pains to suggest that their proposal is for a ramped-up Internal Revenue Service that is much more serious about collections than the current estate tax. We share their view that more could be done to collect estate tax revenue (and tax revenue more generally).

And we certainly do not start from "the premise that the rich cannot be taxed," as Saez and Zucman allege. Instead, we share their objective that it is imperative to raise more tax revenue from those at the very top, and we propose a variety of progressive reforms to this end.

However, government budget scorekeepers properly score proposals in the form in which they would likely be enacted, not on the basis of the aspirations of their academic authors. So, we stand by our position — which will possibly be tested someday — that official scorekeepers would be very unlikely to validate the Saez-Zucman estimate of Warren's proposed wealth tax, and that the gap would likely be substantial.