The 33% Solution

Debtor audits begin today. In the 2005 bankruptcy amendments, Congress decreed that one of every 250 cases shall be randomly audited under procedures established by the Executive Office for U.S. Trustees. In addition, Congress declared there shall be an audit . . . and here is where it gets really fun . . . of every schedule "of income and expenses that reflect greater than average variances from the statistical norm of the district in which the schedules were filed." Greater than average variance from the statistical norm. How erudite. How brilliant. Who says Americans lag the world in math education?

Well, I'm not sure this piece of legislative drafting doesn't provide more evidence that we don't know our pi from our hypotenuses. In statistics, variance is a term generally used to describe a sample or population. Thus, a sample of schedules of income and expenses would have variance, but a single schedule in the sample would not. The single schedule would deviate from the population (or sample) variance.

OK, Congress is not a bunch of pointy-headed, protractor wielding nerds like yours truly. We all know what Congress meant--audit all schedules of income or expenses that are "further away" from the average than the average schedule of income or expenses. To figure out how far away the average schedule of income and expenses is from the average, we have computational issues because some schedules will be above or below the average. That is why we calculate a variance by squaring the amount each observation is away from the average. Squaring gets rid of the negative numbers, but I suppose we can get past all of these problems by saying Congress was really directing audits of all schedules more than the average amount "further away" from the average after Pointdexter finishes his pedantic calculations.

If we get past all those calculations, then we run into another inevitable fact of mathematical life. Assuming a normal distribution (i.e., a "bell curve"), approximately 33% of the observations will be further away from the average than the average observations. In other words, Congress appears to have directed random audits of one in every 250 chapter 7s and 13s and mandated audits for 33% of all bankruptcy filings. Maybe we're only worried about schedules of income that are too low or schedules of expenses that are too high (although the statute does not so limit it), which would at least cut the number in half, to 16.5%. That's still a lot of audits.

My guess is that we won't actually see that many cases audited. Rather, a rule of thumb will develop where schedules that look bad get audited. This problem, however, is just the beginning of the new audit rules. I'll try to write some more postings about on this topic.

Comments

[Mark Agee]

With the IRS auditing less than 1% of tax returns, I continue to believe that this program will quickly die a quiet death. When Justice discovers that virtually all of the debtors are, in fact, broke they will find a better use for the money. Maybe like fighting crime instead of enforcing civil rights for credit card companies.