Administration

Student-Loan Default Rates Are Easily Gamed. Here's a Better Measure.

March 26, 2015

The University of the Rockies has a good federal student-loan default rate. Only 7 percent of borrowers from the for-profit graduate school default on their loans within three years of leaving the university — a better record than its peer institutions. But a closer look at the data reveals a strange pattern. From October 2008 to September 2011, nearly 1,300 borrowers from the University of the Rockies started repaying their loans. Yet the university awarded only 316 degrees during the 2008-2011 academic years. In other words, for every degree earned at the University of Rockies, four borrowers have their debts come due. So while many of the students aren’t defaulting, they aren’t graduating either.

This disconnect between repayment, completion, and default exposes a significant flaw in the key metric the federal government uses to police its student-loan programs and decide which colleges to bar from eligibility. And it highlights the weakness of current attempts at accountability around student loans.

The current measure of student-loan performance ignores what research shows is the most important indicator in determining whether students are able to avoid defaulting on their loans: whether they finished college. Borrowers who drop out with debt are also more likely to be unemployed and earn less than their peers who graduated. In fact, of student borrowers who entered college in the 2003-4 academic year and defaulted by 2009, 63 percent had dropped out.


Proof: An occasional series in which higher-education insiders work out new arguments using data. If you’re interested in contributing, email jeff.young@chronicle.com.


Here’s how the system works now: The cohort default rate — which since the 1990s has been the federal government’s main metric to evaluate institutions in the federal student-loan programs — tracks how many students default on their loans within three years of beginning repayment. Colleges that have a default rate of 30 percent or more for three consecutive years or 40 percent in one year risk losing access to federal student loans.

But because the measure tracks results for such a short time, it is possible for colleges to game the metric by artificially lowering the number of students who default within three years. How? A college can encourage borrowers to ask for a forbearance — an option in which the federal government allows borrowers to stop making payments without their loans becoming delinquent or heading toward default. Since it takes almost a year to default, the college needs borrowers to enter forbearance only for a couple of years, ensuring they cannot show up in the default rate, even if they never make a single payment.

Unfortunately, there’s no easy way to spot colleges that may have inaccurate loan-default rates. There’s no data on institutional usage of forbearance or a similar option called deferment. Nor do we know anything about the long-term default-rate picture.

Fortunately, there is a way to identify colleges that may be a greater student-loan risk than they otherwise appear. The trick is to compare the number of borrowers in repayment to the number of graduates.

Finding the number of borrowers entering repayment is easy. We know that borrowers can end up in repayment in only one of three ways: They graduated, they dropped out, or they started taking less than a half-time course load. These individuals all show up in the denominator of existing student-loan default rates. Knowing the number of students who entered repayment thus provides an accurate picture of students who are separating from their college in a given year.

While existing data cannot tell us how many borrowers are entering repayment because they graduated, we can approximate it by looking at the number of credentials a college awarded each year. If the number of borrowers entering repayment is greater than the number of credentials awarded, then we know that there must be some indebted dropouts. And since not every student borrows, even a ratio that approaches 1:1 should indicate an institution that is producing a lot of students with debt but no degree.

Over all, higher education produces almost exactly as many borrowers in repayment as graduates. Over a three-year period, American colleges and universities handed out 12.4 million credentials and had 12.1 million borrowers enter repayment. But the results vary a great deal by the type of institution. For example, the public four-year sector produced about 84 borrowers in repayment for every 100 graduates, slightly below the rate of 87 for every 100 graduates in the private nonprofit sector.

The results also add depth to comparisons of loan performance at community colleges versus for-profit institutions. On paper, community colleges, and four- and two-year for-profit institutions all have nearly identical default rates of around 20 percent. But community colleges produced just 75 borrowers in repayment for every 100 graduates compared with 136 per 100 at two-year for-profit colleges and an alarming 233 borrowers in repayment for every 100 graduates at four-year for-profit institutions.

Comparing graduates and borrowers shows that some colleges’ default rates may not be as good as they appear. For example, Wilson College, a private, nonprofit four-year institution in Pennsylvania, has an official default rate of 6.8 percent over three years, just below average for colleges of its type. But it actually sends 200 students into repayment for every 100 credentials it issues, well above the average of 87 for its sector.

And some colleges with already poor default rates look even worse. ITT Technical Institute, for example, already has a poor cohort default rate of 27 percent, but it sent more than 123,000 borrowers into repayment over three years while awarding only 51,500 credentials.

To be sure, this comparison is better suited for capturing results at degree-granting institutions than those that award certificates. That’s because students who complete only a certificate are more likely to default than other types of graduates. The short length of these programs means they often have high graduation rates, even if they are not producing much economic return. That is why a place like Tint School of Makeup and Cosmetology in Dallas — a school whose 38 percent default rate suggests it should never be allowed near a student loan — had a low ratio of 51 borrowers entering repayment for every 100 credentials issued. Clearly looking at completion alone is insufficient for colleges that mostly grant certificates.

The current method of judging default rates has noble aims, but its limited focus has also been its undoing. By confronting only the worst possible outcome — student loan default — the cohort default rate has made it too easy to ignore other important issues regarding student debt.

As the White House increases its efforts to talk to consumers about college quality, it’s time for a more reliable method of ensuring student-loan accountability.

Correction (4/6/2015, 2:30 p.m.): The author of this article originally miscalculated the number of borrowers from Chapman University who were entering repayment. As a result, the article erroneously singled out Chapman as an institution whose low loan-default rate understated its ratio of credentials awarded versus borrowers who were in repayment. Chapman's corrected ratio of 80 borrowers for every 100 graduates places it just below the average for similar institutions. The Chapman paragraph has been replaced with an example from Wilson College.

Ben Miller is research director for higher education at New America.