It’s become common knowledge that the combination of enormous student loan debt loads and a tepid economic recovery has left recent graduates with a crushing burden, so when a Brookings Institution pushed back against that narrative is was bound to create waves. It concludes that the growth in student loan debt isn’t a serious problem, which would be great news if the report held up well to scrutiny.
“We find that, on average, increases in lifetime incomes among households with student loan debt more than offset increases in borrowing,” write Beth Akers and Matthew M. Chingos, fellows at the Brown Center on Education Policy at the Brookings Institution. “We also find that the average burden of monthly payments for student loans has not increased over time, likely due to the fact that borrowers are repaying debts over longer periods of time.”
Jim Chanos has a new short target in his sights. Earlier this week, the hedge fund manager disclosed that he is betting against "legacy" data centers that face growing competition from the trio of technology giants, which have previously been their biggest customers. The fund manager, who is best known for his winning bet against Read More
Study’s population doesn’t address recent graduates
If this were the only information you had, the student debt situation would look promising. It tells you that, yes, people are paying more for their educations but that they’re also getting better returns on that investment. The first problem is that, as Akers and Chingos point out, borrowers are taking a lot longer to pay their loans back – 13.4 years in 2010 compared to 7.5 years in 1992. Keeping relative payments the same and doubling the term shouldn’t be dismissed so quickly.
But that’s just a nitpick compared to the bigger methodological problems in the study. Akers and Chingos decide to look at the level of indebtedness of households aged 20 – 40 who are currently paying off their student loans.
First, the age range averages out current debt levels across a long period of tuition levels. You can see this clearly in the trends of published tuition rates and trends in debt over time published in the report. Notice how one starts rising around 1982 and the other starts rising between 1995 and 1998. The inclusion of people who went to college as much as 22 years ago means that the study’s results demonstrate a lot of lag, as you would expect (and the lag increases along with average loan term). It also means that a 35 year old who still owes a few grand is being averaged in with a recent graduate with six-figure debt. Using current indebtedness instead of initial debt can only underestimate the debt burden facing current graduates.
Second, households aged 20 – 40 excludes recent graduates who are now living at home with parents older than 40, while restricting the study to people who are paying off their student loans excludes the unemployed and the working poor. Another way of looking at this study is that if you take on a lot of college debt and land a good job, you’ll be fine, but that was never really the question. It’s the people who don’t land a decent job and need to declare bankruptcy (but can’t) who are struggling.
Brookings claims that tuition only explains half of rising Student loan debt
Another startling conclusion from the Brookings report is that the rise in tuition doesn’t explain the rise in indebtedness. Akers and Chingos run a regression that tests the impact that race (who is going to college/grand school), educational attainment (how many people are going to college/grad school), and rising tuition on average student loan debt. They argue that rising tuition accounts for just 51% of the rise, educational attainment for about 28%, and race was slightly negatively correlated.
There’s a danger of overfitting here, so it would be nice if the report had more information about the regression’s confidence level, especially since the model only explains 67% of the rise in student loan debt. Akers and Chingos say that the remaining one-third of debt level rise could be due to changing borrower behavior, but the model as presented doesn’t explain the data that we have so it’s hard to take it too seriously.