Horrendous tragedies gradually recede, as they have to if you want to go on living, and things begin to return to normal. Yesterday's Tax Policy Colloquium, my fifth since a string of about 250 sessions (counting AM and PM) over 10 years with David Bradford came to an end, was easier than it's been, although certainly a heavy cloud still hangs over it. It has begun to feel permissible and feasible to laugh again.
Yesterday's paper by Kyle Logue of Michigan Law School (visiting at Virginia), available here, discusses the recent, apparently very small, trend of companies buying insurance against losing on a given big-money tax issue. The market is small because insurers can't easily make the law of large numbers work for them, as with car insurance, so instead it amounts to betting that their tax-expert underwriters are setting the odds right. In other words, the insurers are like bookies who don't have matching bets on both sides (the IRS hasn't gone to them to bet on its side of the over-under). But the issues are interesting, extending, for example, to why we might not (if we would not) want to let criminals insure against the sanctions they would face if caught. The question is not as obvious as it sounds (e.g., insurance doesn't lower expected value), and you can learn a lot, concerning how to think about more realistic issues, from identifying the conditions under which such insurance would be fine, even if you agree that those conditions do not actually hold.
Other places where Logue has presented this paper apparently focused on the fairly predictable question of, gee, isn't this just a trick by corporate tax shelter promoters to increase their sales and reduce compliance? The answer seems to be no, although one could imagine conditions where that would be the case. We try to take a more unpredictable slant at our colloquium, also one more focused on fundamentals, an approach that owes a lot to David Bradford but that is now engrained enough here to continue without him. So we focused on why the tax insurance market remains small, how tax insurance relates to the broader problem of legal ambiguity, how risk aversion and deterrence are related, whether Logue is right to emphasize the distinction between merits risk (will I win if audited?) and detection risk (will the IRS spot the issue?), and how the IRS should price advance rulings. On merits risk versus detection risk, we seem to have persuaded Logue he was wrong to focus on it, but then to have persuaded ourselves (and perhaps him) that it might be a useful distinction after all, though for different reasons than the paper suggested. Not to be mysterious about all this here; I just don't want to make this post too long and detailed.
I say "we" because actually this time I didn't talk much. Practitioner Ed Kleinbard and economist Jason Furman did most of the heavy lifting. As these sessions have gone on over the years, I have been aiming for the same transition that Stephan Marbury is supposed to be trying with the Knicks: less scoring, more playmaking. I hope I am being more successful than he is.