In an op-ed at the Washington Post Lawrence Summers Natasha Sarin propose that going after tax evaders will bring materially more money into federal coffers:
What is the overall potential? In a study released this weekend, we conservatively priced out a program of increased auditing, IT investment and greater third-party reporting. We estimated that it would be possible to close 15 percent of the tax gap by spending approximately $100 billion on additional enforcement, as would be necessary to return the IRS to its historical scale. Every $1 that is spent would generate more than $11 in greater tax collection.
Congressional scorekeepers have suggested more modest revenue potential from investment in enforcement; however, our study shows these differences can be reconciled. Their approach is based on a program that is more modest in size (only a quarter as large as our proposed restoration of enforcement effort to previous peak levels) and scope (we consider the revenue potential of targeting audit resources on high-income individuals, as well as increasing information reporting). Critically, the Congressional Budget Office does not account for deterrence effects, which Treasury Department reports suggest greatly magnify the revenue gains from increased enforcement.
Why is the federal government leaving so much money on the table? Part of the answer is that there are powerful interests that want to maintain a system that facilitates evasion.
Frankly, I’m skeptical for two reasons. The more important reason is that it isn’t your father’s IRS any more. The IRS has engaged Palantir Technologies to create a system for using modern data mining and artificial intelligence strategies to identify possible cheats and train IRS personnel in its use. Traditionally, the IRS’s procedures have had a two-pronged approach: investigators identify possible cheats, something of a hit or miss proposition, while auditors do the spadework. The new system should identify more tax fraud less expensively and I have found little evidence for a backlog of IRS audits. There’s a backlog in correspondence and FOIA requests, yes, but that doesn’t necessarily translate into a backlog of audits. Said another way giving the IRS the 2009 IRS’s budget adjusted for inflation will probably not produce more results. It’s about 20% lower than it was then.
The second reason is that the world is not linear. Increasing the IRS’s budget by 20% won’t necessarily realize 20% more fines, penalties, and recoveries.