President Trump’s Office of Management and Budget has proposed a 14.1 percent cut in funding for the Internal Revenue Service. If made permanent, we conservatively estimate that this cut will cost the Treasury $60 billion, just from direct revenue loss due to reduced enforcement (and netting out the reduced spending on IRS enforcement). This new cut follows a more than 20 percent real reduction in resources for IRS enforcement since 2010. Undoing that reduction could save the federal government an additional $30 to $40 billion. All together, the President’s failing to restore these previous cuts and then cutting even more allows handouts to tax evaders costing the Treasury about $100 billion over ten years.

For several reasons, these figures probably substantially understate the effects on the federal budget.

First, consistent with government scoring, these estimates count only collections that enforcement efforts bring in from their targets. This is like accounting for the effect of a police department on bank robberies by measuring the number of thieves caught in the act or afterwards. In fact, the vast majority of Americans pay their taxes without getting a nasty note from the IRS. Some are afraid; others understand their responsibility to do their part. As the IRS’s enforcement budget shrinks, some will feel emboldened to cheat more. Why not? And as the IRS’s professionalism erodes and call wait times grow—they’ve already more than doubled—others will be infuriated into cheating.

While these indirect effects are hard to measure and uncertain, they are probably large. In fact, the Obama Administration previously said that these kinds of cuts could mean three times the size of the direct revenue effects from changes in enforcement (see here). If true, this would bring net revenue loss from $100 billion over a decade into the range of $400 billion, including both direct ($100 billion) and indirect ($300 billion) effects.

Second, there is talk in Washington of fundamental tax reform. The frameworks that Republicans have put forward all lose trillions of dollars, mostly to the benefit of those at the top of the income distribution. However, the estimates of these plans generally assume that the IRS has the resources to implement major changes to the tax system. A fundamental rewrite of the corporate tax code will require substantial new guidance from the IRS. In its absence, the system could bleed even more revenue. You simply cannot be serious about tax reform and fail to give the IRS the resources necessary to transition to a new system.

Slashing the budget of the IRS is classic faux populism. Taxes funds programs that are core for the middle class, like Social Security and Medicare. And most Americans face little audit risk, even from a well-funded IRS. In 2010, before recent budget cuts, only about 1 in 100 households were audited. The rate rose above 2 percent only for households making more than $500,000 per year.

Even so, cutting the IRS budget is a long-standing priority for the right. Between 2010 and 2015, the Congress rejected five consecutive budgets from the Obama Administration seeking increased IRS funding, and instead imposed large cuts at the IRS. The IRS’s staff has shrunken by 13,000 people.

During the 2011 negotiation over the grand bargain, both sides understood that increasing the budget for the IRS would be a key element of a budget deal. When the Congressional Budget Office (CBO) formally recognized that the Obama Administration’s modest proposal to increase enforcement beyond 2010 levels would result in net deficit reduction of $29 billion, a key Republican staffer reacted with delight. He knew that spending more on enforcement was a way to raise revenues without raising tax rates, and he thought this approach would be acceptable to conservatives who opposed higher rates.

Alas, only the first part was true.

Methodology

We assumed different multipliers for different components of the “direct revenue” effects of changes in IRS enforcement.

The first multiplier—8-to-1—applied to the cuts that Trump is floating. While the reductions of 2010 to 2015 hit enforcement especially hard, we assume that the 14.1 percent cut will apply across-the-board to enforcement and other functions. This would reduce enforcement activities at the IRS by $800 million. We further assume that cut continues—adjusted for inflation—going forward. At a ratio of 8-to-1, this reduces government revenue by $70 billion for a net cost of about $60 billion.

The 8-to-1 ratio is conservative in that it is less than the average bang-per-buck of IRS enforcement efforts. For 2015, the IRS estimated that the cost of enforcement activities, combining enforcement and the relevant part of operations support, was $5.6 billion, while the revenue generated was $54.2 billion, for a return on investment of $9.60 for every $1 spent. (See here, p. 112.)

To arrive at a possible bang-per-buck of the activities that would be eliminated due to the Trump cuts, we averaged the 9.6 bang-per-buck with the estimated bang-per-buck of new enforcement activities—above current levels (if funded and when fully operational) of around 6. (See here, p. 1050.)

It’s possible that the Administration could focus cuts on enforcement activities with lower bang-per-buck than this. That is what happened in recent years under the Obama Administration. But it happened because the Obama Administration was willing to go where the money is, the very wealthiest. Between 2010 and 2015, the likelihood of individuals being audited declined from 1.11 percent to 0.84 percent, but the audit rate for those earning more than $10 million increased from 29.7 percent to 34.7 percent. (Compare pages 23 and 27 here with pages 22 and 26 here.) (Though new reports suggest that audit rates fell further in 2016, even at the top of the income spectrum.)

Nothing about the Trump Administration’s economic and regulatory agenda suggests that it will focus enforcement at the top of the income scale. In fact, there is talk about focusing enforcement on the working poor and the earned income tax credit, for which there is very low bang-per-buck.

We assumed a lower rate of return—averaging a multiple of about 4-to-1—for restoring the funding cuts since 2010. This reflects the ratio used by the Obama Administration in estimating the effects of its proposed increases in IRS enforcement funding, and this takes into account the time it takes to bring new enforcement programs in line, as well as the lower bang-per-buck of the additional activities. (As noted above, these activities—when fully in effect—would have a bang-per-buck of about 6.) This is roughly in line with estimates from the CBO.

And, again, our estimates do not count the effects of enforcement on voluntary compliance or the risks surrounding corporate tax reform.