NYU Tax Policy Colloquium, week 12: James Alm's and Jay Soled's "Whither the Tax Gap?"

Yesterday at the colloquium, James Alm and Jay Soled presented their paper, “Whither the Tax Gap?”

The paper argues that, contrary to widespread anxiety about the tax gap, there are actually several reasons for thinking it may shrink.  In particular:

1) The use of cash, which is much harder to trace than banking, card, and phone app payment transactions, is in decline.  Even if the use of cash doesn’t disappear, the convenience advantages of using other payment methods may reduce the net expected benefit from using it to facilitate evasion.

2) Computerization generally makes it easier for tax authorities to access data and process it readily.

3) The relative shift of employment from small to large businesses means that there will be more reporting and less opportunity for under-the-radar screen evasion.

To each of these observations, there are possibly counter-arguments.  For example, as to (1) and (2), perhaps it’s hard to predict which way technology will go, as between the “cops” and the “robbers.”  Bitcoin is an example of a non-cash technology that might end up having the same advantages as cash for would-be evaders.  As to (3), perhaps in some ways different types of problems may pertain to large businesses.

For example, suppose we define the tax gap as the difference between what was paid in taxes and what should have been paid.  Say that a large business does ten things on its tax return, each purporting to save $1 million of tax, and classified as tax avoidance rather than tax evasion, except that the legal permissibility of each is open to question.  Suppose, for example, that each of these items has an independent 40 percent chance of being upheld if the IRS understands and fully reviews.  (Suppose further for simplicity, albeit perhaps naïvely, that this 40 percent probability is truly a frequentist one, rather than just expressing someone’s subjective probability.)

Then, on average, the “true” tax gap as to this company is an expected $6 million.  To be sure, under standard measuring techniques, none of this would be included in the tax gap (nor, perhaps, could it be).  But it would fit within the broad conceptual definition that I gave above.  And it might be viewed as indicating that the nature of enforcement problems had merely changed, rather than necessarily easing.

Still, there's something to be said both for a thought-provoking against-the-grain prediction, and (if the prediction proves correct) for lessening outright fraud as an enforcement problem.