I can’t recommend enough Kelly Field’s excellent piece on the front of this week’s Chronicle of Higher Education about the true story of long-term default rates (and not just because Erin and Robin’s report on this topic gets a mention). What Field turns up isn’t pretty:

According to unpublished data obtained by The Chronicle, one in every five government loans that entered repayment in 1995 has gone into default. The default rate is higher for loans made to students from two-year colleges, and higher still, reaching 40 percent, for those who attended for-profit institutions.

We’ve known from other Department documents that default rates were always higher than expected,  but to see it laid out in such clear terms is worth the read by itself. But there are three other things about the story (and the accompanying piece from Goldie Blumenstyk) that are worth highlighting.

Lots of Complexity

Over the past decade or so the Department of Education successfully established a complex and opaque system of default rates. While Field obtained some unpublished data showing the 15-year default rate for loans, the figures are impossible to compare to any other shorter-term numbers we have. In fact, there’s a chart showing how not even basic things like the year and institution type are guaranteed to match. That’s ridiculous given that we aren’t even talking about trying to see institutional level data. I’d even settle for a single default rate figure that would be comparable to the cohort figures.

Defaulted Student Loans Don’t Make Money

This is an argument that pops up from time to time—the government makes money off the collection costs of a defaulted loan, so it doesn’t matter if a borrower can’t stay in repayment. This argument stems from Office of Management and Budget assumptions that the government takes in between $111 and $122 for every $100 of a defaulted loan, but Field correctly notes “those numbers do not take into account collection costs or inflation.” In other words, a dollar of a defaulted loan collected 10 years from now is treated exactly the same as a dollar lent with no consideration of how expensive it is to obtain it or the various fees that must be paid to collection agencies or guaranty agencies.

(For what it’s worth, federal student loans don’t make money either.)

Recognizing defaulted loans have a cost is an important shift and one that would hopefully put a greater emphasis on both not wasting dollars at bad programs and also providing evidence-based assistance to keep people in repayment.

Our Existing System Understates Default

The chart leading off Blumenstyk’s piece does a nice job illustrating how it’s possible for a combination of deferment and forbearance can keep a borrower from being caught in the default rate measure without ever making a payment.  (I once took a stab at a far more MS Paint heavy version in this post here.) And the Chronicle’s chart assumes that the borrowers never make a payment—in many cases even a couple months of meeting payments will be sufficient to avoid being captured in the default window. No wonder the actual default rate can be 15 or more percentage points higher. Such large discrepancies between our short-term accountability rates and actual long-term figures just underscores the weakness of the existing cohort default rate system.

Adding it Up

Field and Blumenstyk’s pieces conclusively show that the default rate problem is worse than we think, costs more than we assume, and has an insufficient regulatory structure to do anything about the first two problems. That’s a big deal, and a bad outcome for both borrowers and taxpayers.* Rather than continuing to hide these effects with non-compatible rates and misleading figures, it’s time for a frank discussion of what the default problem actually looks like and how to address it. Some of that might involve limiting or shutting down colleges and some of it might mean more work with students. But first we need a more honest accounting for these loans than we’ve gotten so far.

*This isn’t a bank-based versus government-based lending problem either. Defaulted loans are reimbursed by the government and either it or guaranty agencies handle collection activities, so the costs all go to the same place, with the cut for guarantors more than making up for the 3 percent of a defaulted loan that lenders aren’t paid for.  As for the default aversion activities provided by all these groups, there’s no evidence testing their effectiveness.