Reading Between the Lines of the Government's Federal Student Loan Data Reports


On March 17th the U.S. Department of Education (DOE) announced the release of its most recent Quarterly Student Aid Report, its compendium of data on the performance of the government's federal student loan portfolio.

As DOE points out there's a lot to be happy with. More folks than ever are enrolling in affordable, Income-Driven Repayment (IDR) plans and economic hardship deferments have fallen dramatically. The percentage of borrowers who are more than 31 days late on their payments is in decline and in the last quarter of 2015 ED collected some $2.2 billion in defaulted loan debt, largely through a process called loan "rehabilitation" (getting borrowers back into routinely making payments again).

Good news indeed but as famed ESPN announcer Lee Corso would say, "not so fast my friend."

Jason Delisle over at the New America Foundation quickly noted that the data also shows 40 percent of borrowers in repayment are currently NOT making payments on their federal student loans. He also pointed out that the percent of dollars in default is around 12.7 percent yet the percent of individuals who are actually defaulting has climbed to around 20 percent.

The fact that neither set of observations is necessarily wrong reinforces my old college statistics professor's favored claim that if you torture data long enough it'll basically say whatever you want it to say.

There is still talk that Congress may actually reauthorize the Higher Education Act this year and that means greater than usual attention to questioning what works and doesn't work when it comes to handling what is increasingly being referred to as the student debt "crisis." Despite the statistician's lament, there are things we should pay more attention to than not when thinking about the data that informs higher education policy. I've outlined a few below.

Borrowers versus dollars - When it comes to portfolio performance the federal government wants to maximize the return of dollars it lends out. At the same time, DOE and its contractors prioritize recovery based on how delinquent individual borrowers are, not individual loan accounts that are in arrears.

What this buys pundits and policymakers is a whole lot of grief. Policies get designed to help borrowers and families but loan performance metrics look at how well individual loans perform. The coupling between individuals and loans is tidier in other credit markets because people usually only have a single mortgage or auto loan. In lots of cases though, student borrowers can literally have more than a dozen federal student loans.

Where portfolio performance is concerned, there's a big difference between writing off $50,000 in bad debt versus $2,000 but where policy is concerned, the amount is irrelevant. It's the act of defaulting that creates a potential lifetime on financial hardship, not how much wasn't repaid.

The difference between affordable and cheap debt - Growing enrollment in IDR plans implies more borrowers' monthly payments are becoming more affordable, regardless of loan balance. On the other hand, since it is not the default repayment option, large-scale movement into IDR also means borrowers are financially unwilling or unable to make standard term payments.

Decisions involve tradeoffs and in the case of IDR plans, the tradeoff to affordable monthly payments is having to pay more overall for the debt. Think of it like a car payment. A $25,000 loan over 4 years has a monthly payment of around $570 a month and with interest you'll end up paying around $27,300 in total. However, spread it out over say 7 years and you can get the monthly payment down to $347 a month. Problem is now you end up paying almost double the amount in interest so when the whole thing's paid off it'll cost around $29,200.

The second option is certainly more affordable but the first option would've cost $2,000 less. The parallel to higher education is actually magnified since the difference between the standard and IDR repayment terms is much longer (10 versus 20 to 25 years). Affordability can be challenging when a borrower first graduates and, sure, over time IDR balances are eventually written off. Still, the greatest irony of IDR plans is that - for some - they actually saddle people with more debt not less.

Benchmarking performance to other credit sectors - Like student loans, this past autumn the volume of outstanding auto loan debt surpassed $1 trillion for the first time. Also like student loans, the average individual borrows just less than $30,000.

Which is where the similarities end. Despite there being a trillion dollar loan market, despite the absence of income-driven repayment plans, despite the typical auto loan term being less than six years, and despite the fact that the asset being purchased loses almost half its value the second it drives off the lot, the percent of individuals who default on an auto loan is less than one percent.

That's right.

Of course, we're still talking about two different goods and pundits will claim people simply need cars to buy food or get to work (though nobody can explain why they need "new" cars). Nevertheless, the reality is that even in a market where not everyone has ready access to attractive credit terms, a lender's ability to repossess an asset in the case of default is a strong motivator for timely payment.

Is it possible that the default problem is being driven in part by making student loan repayment too easy? The truth is, between income-driven repayment options, almost a year to not make a payment before defaulting and hundreds of required outreach attempts by servicers, there's no reason anyone should default on a federal student loan.

Give policymakers better decision-making data - A considerable amount of time and dollars have been put into creating the five existing income-driven repayment options on the belief that defaults are driven largely by unaffordable monthly payment structures. Now that literally millions of borrowers are enrolled in IDR plans what seems to be missing from loan performance data reports like these are the number of borrowers enrolled in these plans who are either delinquent or in default.

This isn't the first time we've seen a large-scale investment in education that fails to include metrics on how well the program performs. The Pell Grant program costs the federal government almost $40 billion a year and we still have no idea to what extent students who actually receive these federal grants actually go on and complete a college education.

What DOE also needs to provide across the entire loan performance portfolio are better dispersion metrics. If I have 9 borrowers who've each borrowed $2,000 and one borrower who's taken on $100,000, calculating that the "average" borrower has about $12,000 in debt is basically meaningless and not at all reflective of any of the pool's borrowers. We should be focused providing policymakers with more reflective measures, such as the characteristics of median borrowers, that can better inform the decision-making needed to efficiently manage hundreds of billions of dollars in federal debt.

Note: the original version of this piece was posted on LinkedIn.

Header image source: By Kane5187 (Own work) [Public domain], via Wikimedia Commons