The Revenue and Distributional Effects of IRS Funding

There is substantial focus in Washington policy debates on cutting government spending in the putative pursuit of fiscal sustainability. However, when spending is discussed, it is imperative to think both about the direct costs of the government spending (taxpayer dollars that support different agency missions) and the benefits that society reaps from that spending.

A timely example of the importance of looking at both sides of the government’s ledger is investment in the Internal Revenue Service (IRS), which has been a recent focus. The IRS is responsible for collecting 96% of federal revenue.¹ In the last Administration, the agency received a significant influx of funding from the Inflation Reduction Act, which was geared toward modernization and improved compliance and service efforts. Some of those funds have been rescinded, and so far in this Administration, there have been further efforts to decrease the IRS’s footprint by cuts to its workforce.

The Budget Lab estimates that the expansion of funding ($80 billion) for the IRS in the Inflation Reduction Act (IRA) would have led to a net revenue increase of $637 billion over the full 10-year budget window.² If the IRS shrinks by 50% (a workforce decrease of about 50,000 people),³ we estimate that this significant reduction in IRS staffing and resulting IRS capacity to collect revenues would result in $395 billion ($350 billion net) forgone revenue over the 10-year budget window.⁴ If the lack of IRS resources leads to a substantial increase in noncompliance, net forgone revenue could rise by $2.4 trillion over 10-years.

Additional revenue for the IRS generally generates more revenue than it costs through improved enforcement, particularly focusing on high-income individuals and corporations. The additional funding increases revenue by increasing the efficiency of the service and helping to close the "tax gap” - the difference between taxes legally owed and taxes collected, which currently totals around $700 billion per year.

The revenue increase depends on when those additional funds are spent, what they are spent on, and the return to those investments. Our estimate utilizes the pattern of spending assumed by CBO and the allocation of funds to activities outlined in the IRA.⁵ The return to investment we use comes from several sources. We rely on Holtzblatt and McGuire (2016), CBO, Boning et al (2020), Sarin and Summers (2020) and Debacker et al (2018) for our return to investments estimates (ROI).⁶ The ROI varies from 2:1 for service and IT to 7.1:1 for audits.⁷

Our base estimate also includes an indirect benefit through deterrence. When taxpayers know there's a higher chance of audit, voluntary compliance tends to improve. Economists estimate these indirect effects could multiply the revenue impact by 2-3 times.⁸ We use the midpoint of 2.5. The main constraint on returns isn't the lack of recoverable taxes, but rather the IRS's capacity to conduct thorough audits, even with significantly increased resources. With lower resources, collections will suffer.

Not all categories of taxable income afford the same evasion opportunities. For certain categories of income—like wages and salaries—compliance is essentially 100 percent, because taxes are automatically withheld. For the most opaque categories of income—like proprietorship and rental income—where the IRS has little information available from third party reporting to cross-check the veracity of tax filings, misreporting rates exceeds 50 percent.⁹

Noncompliance is also not evenly distributed across the income distribution. Because the highest earners have the largest tax obligations and tend to accrue income in more opaque ways, their rates of noncompliance are higher. One paper estimates that the top 1% of earners are responsible for nearly 30% of unpaid taxes – according to Budget Lab calculations (Table 1 below) and projection of the latest tax gap numbers that would total about $205 billion annually. The bottom two quintiles are responsible for just over 20 billion annually (less than 5%).¹⁰

Since the provision of a stream of mandatory funding to the IRS in the Inflation Reduction Act, critics of increased enforcement efforts have pushed to rescind these new funds. Most recently, the Trump Administration has pursued a strategy of dramatic “reductions in force” (RIF) at the IRS.

These reductions in force would save the government resources in that the budget of the agency could shrink if it needed to fund fewer employees. However, these reductions would come with dramatic increases in noncompliance which would outstrip any budgetary savings several times over.

We consider two different RIFs:¹¹

1. First, as of February 20, 2025, the IRS has announced around 7,000 layoffs, more than 5,000 of whom reportedly worked in compliance efforts.¹²
2. Second, last week, reports suggested that the Trump Administration was pushing to slash the IRS workforce in half. We assume that these cuts come from across the agency’s functions and are on top of the initial reduction in force, such that if these changes are implemented, the IRS workforce will total around 50,000, down from around 100,000 in January.

Table 2 reports our estimates. We conclude that the aggregate effect of these two RIFs would decrease gross tax collections by about $395 billion over the course of the next decade, which includes a rough estimate of the delays to technology improvements and enhancements that would be the consequence of halving the agency’s workforce. When we account for the fact that salary costs at the IRS will fall as a result of this workforce reduction, we conclude that the cost of cutting the workforce by 50% would be about $350 billion net in lower revenues over the course of the budget window and around $560  billion net in the second decade.¹³

It is generally understood that tax enforcement efforts have direct and indirect effects. Direct effects are direct: more dollars spent on audits means more revenue is collected for the fisc from taxpayers who are audited. Indirect effects have two variants: specific deterrence (a taxpayer who is audited and found to have misstated taxes owed is more careful in the future) and group deterrence (a taxpayer who sees others audited is more careful in filing her own returns).

There are a wide range of estimates of deterrent effects, as Sarin and Mazur (2024) detail. These studies tend to be based on marginal changes, for example Boning et al (2020) use a field experiment to examine the effect of an IRS audit on a tax preparer’s other client’s returns. The authors find that this network deterrence effect raises 1.2 times the revenue of the direct audit. And Boning et al (2023) find a return of $12 for each dollar spent on audits of high-income individuals and a return of $5 for every dollar spent on audits of below median income individuals.

However, the literature has not explored how a change in spending on audits as large as the ones this Administration has proposed would impact revenue collection.¹⁴

It is massively difficult to extrapolate from these estimates to understand what such a sizable RIF would do to taxpayer behavior. Returning the IRS to a 1960s-level workforce in an environment when tax entities have become much more complex and its capacity is already so depleted could potentially meaningfully shift taxpayer behavior, more so than our midpoint deterrence estimate assumes.

To try and guess how taxpayer behavior could adjust, we offer a back-of-the-envelope calculation for illustrative purposes. Here, we rely on the IRS’s tax gap estimates, which document the substantial difference in compliance rates between categories of individual income depending on the nature of third-party reporting that helps the IRS cross-check the veracity of tax filings. Categories of income without any third-party reporting are highly likely to be misreported, as discussed above. Figure 1 is taken from the most recent IRS Tax Gap report and is the basis for this calculation.

If we assume that taxpayer behavior adjusts such that low-compliance taxpayers become somewhat lower compliance, medium-compliance taxpayers become low-compliance taxpayers, etc., we arrive at a much larger impact of the RIF on compliance.¹⁵

Specifically, a 50% cut of the agency workforce would cost $2.4 trillion over the course of the decade, and it would result in around a 25% increase in the tax gap over this period. Note that this adjustment is in some sense also understated because it assumes that the only category of income where taxpayer behavior adjusts is individual misreported income (where the tax gap breaks out income types by compliance rates), whereas other categories of the tax gap, like corporate income and non-filing, are assumed to stay fixed in the new world with a substantially diminished IRS. 

These cost estimates are of course highly uncertain, partly because our calculations are relying on contemporary reporting of what is happening at the IRS. In addition, there is no modern historical precedent for this level of cuts to IRS staffing: It is unclear how the IRS would be able to actually function given the scale of what has been reported. Our estimates are thus meant to reflect reasonable behavioral reactions to a hollowed-out IRS and to what this substantial hit to IRS functionality could potentially mean for revenue.