The Washington PostDemocracy Dies in Darkness

Opinion IRS reform will generate a lot more revenue than the CBO thinks

Signage outside the Internal Revenue Service (IRS) headquarters in Washington, D.C., on March 19. (Samuel Corum/Bloomberg)

The largest offset in the Build Back Better Act is not a tax increase. Instead, it is an $80 billion investment to restore a depleted IRS. The Treasury Department expects this transformative investment to generate $480 billion — $400 billion net — in additional tax collections over the course of the next decade.

That is a large sum. But it is important to put it into context given the scope of the tax evasion problem faced by the federal government. Over the next 10 years, the IRS is on track to collect $7 trillion less than is owed. This enormous tax gap is around 3 percent of gross domestic product on an annualized basis. President Biden’s proposal to revitalize the IRS is projected to net an amount that is merely 5.7 percent of that tax gap.

That modest gain is a more-than-reasonable expectation given the starting point. Today, the IRS has about the same number of auditors as it did during World War II, and the IRS can answer fewer than 30 percent of the phone calls it receives from taxpayers with questions. IRS technology is woefully outdated, and it fails to allow for even simple data analytics to identify evaders. Providing the IRS the resources it needs will go a long way towards shrinking the tax gap.

Yet the Congressional Budget Office is soon to release its estimate of these efforts, which it is likely to score at netting less than half the Treasury number — some experts expect the final CBO estimate to be around $150 billion.

Opinion | IRS commissioner: We desperately need resources so the agency can serve you and our country

In general, I believe policy should be set on the basis of official scorekeeping by nonpartisan scorekeepers. But, in this case, it would be irresponsible to not recognize that the CBO estimate for tax-compliance efforts is conservative to the point of implausibility. Assuming the administration is successfully able to develop and implement sound plans for the IRS — which will require substantial management and strategic planning to go with the new resources — I am confident that the proposed investments can generate much more revenue than the CBO assumes.

Although the CBO has not disclosed its methodology in any detail, I base my conviction on three considerations.

First, as best as can be discerned, the CBO assumes significant diminishing returns — that is, when additional resources are provided to the IRS, first the most promising leads are pursued, then other leads, then less promising leads, thereby reducing the efficacy of large expenditures. Indeed, in the past, when the CBO has estimated doubling the scale of new investment in the IRS, it has concluded that far less than double the additional tax revenue will be collected.

But what this approach fails to recognize is that the president’s budget proposal reinvigorates an IRS that has been gutted during the past decade, with audit rates for millionaires falling by more than 60 percent and halved for large corporations. That audit revenue has declined exactly 1:1 in proportion to reduced audit efforts suggests that audit revenue will increase proportionally when cuts are reversed. Further, the administration proposal calls strongly for focusing new enforcement resources on the high end, where underpayments are greatest. Further, the CBO should recognize systemic changes are more impactful than incremental ones.

Larry Summers discusses this piece in more detail on James Hohmann's podcast, "Please, Go On." Listen now.

Second, the CBO’s approach also takes no account of the benefits of either better taxpayer service or improved technology. Even a 2 percent improvement in voluntary compliance would be enough to shrink the tax gap by well over $100 billion over a decade. Said another way, you gain significant revenue from small service improvements. And there is scope for meaningful, immediate change: Today, dozens of taxpayer assistance centers — primarily in rural areas — are unstaffed. Restoring taxpayer assistance to 2010 levels would mean nearly twice as many taxpayers would have access to the IRS to help them accurately file their taxes. With respect to the value of technology, one important data point is the success of the return review program, which automates the review of returns to prevent invalid refunds. Estimates from the GAO suggest this program saves the IRS $4.4 billion annually and costs only $90 million (a return of nearly 50:1).

Frankly, no one can be certain what the right estimate is for the revenue potential of technological and service overhauls at the IRS, but the CBO’s assumption of zero is a poor approximation.

Finally, and perhaps most important, official estimates leave out the large impact of increased enforcement on taxpayer behavior, due to the CBO’s belief that greater enforcement scrutiny will increase voluntary compliance “only modestly.” Even Treasury is extremely conservative in accounting for deterrence, suggesting that the indirect effect of additional spending on the IRS will be about half of the direct effect. This flies in the face of academic research: Indeed, a recent study from Tulane’s James Alm notes that the magnitude of these spillover effects are in the range of 4 to 12.

It is true there is heterogeneity in the size of deterrent effects estimated in the literature depending on the type of enforcement activity considered: Recent work by the University of Michigan’s Joel Slemrod and co-authors finds an indirect effect that is at least as large as the direct one in the corporate context. But the consensus of the literature is that deterrent effects are important. Excluding them from official scores is clearly incorrect.

Cumulatively, these three adjustments would be sufficient to raise the CBO estimate above Treasury’s $400 billion measure. My judgment is that even Treasury is far too conservative. This estimate is less than half of what I concluded, in work with Natasha Sarin (who is now at the Treasury Department), can be generated by an infusion of resources into the IRS of about the same size as proposed by the Build Back Better Act. And while my estimates are much higher than those of Treasury, they are just less than half what other experts, including former IRS commissioners Fred Goldberg and Charles Rossotti, conclude can be generated by a robust attack on the tax gap.

Effective implementation of IRS reform efforts will be among the most significant challenges facing the Biden administration once Build Back Better is passed. It will be imperative to create vigilant oversight and clear metrics that will help document progress and enable course correction quickly.

But I have no doubt that this is possible. There is a reason that a bipartisan group of former Treasury secretaries and IRS commissioners (including the current commissioner, appointed by the prior administration) have coalesced around overhauling the IRS. Nothing is more important for tax reform efforts.

Larry Summers discusses this piece in more detail on James Hohmann’s podcast, “Please, Go On.” Listen now.