Are Student Loans Becoming a Macroeconomic Issue?

What’s the general economic consensus on the impact of student loans on the household finances of those who hold them? Here’s “Student Loans: Do College Students Borrow Too Much—Or Not Enough?” (Christopher Avery and Sarah Turner, 2012), which argues, “[t]here is little evidence to suggest that the average burden of loan repayment relative to income has increased in recent years.” Using data from 2004-2009, the authors find that “the mean ratio of monthly payments to income is 10.5 percent” for those in repayment six years after initial enrollment.

They boost that number with a 2006 study by Baum and Schwarz to conclude that two trends cancel each other out: there’s rising debt but steady student debt-to-income ratios. How can this happen? It “can be attributed to a combination of rising earnings, declining interest rates, and increased use of extended repayment options.” This is how, though average total undergraduate debt jumped 66 percent to a value of $18,900 from 1997 to 2002, “average monthly payments increased by only 13 percent over these five years. The mean ratio of payments to income actually declined from 11 percent to 9 percent because borrower.”

Let’s put this a different way. If you asked economists looking at the data if student loans could be having a macroeconomic effect, especially through a financial burden on those that have them, they’d say that the actual percent of monthly income paying student loans hasn’t changed all that much since the 1990s. They may be making larger lifetime payments, since they’ll carry the debts longer, but that’s a choice they are making, which could reflect positive or negative developments. Certaintly there’s no short-term strain. So there aren’t any economic consequences worth mentioning when it comes to student loans.

I always thought this approach had problems. First, they were only looking at the pre-crisis era, so we couldn’t see the impact of student loans once we hit a serious problem. And they were just rough averages of short-term income aggregates, rather than looking at specific individuals with or without student-debt and seeing what kinds of spending, particularly on longer-term durable goods, they do. But since I had no data myself, I never pushed on this very hard. Part of the problem is that student loans have happened relatively quickly, so quantitatively it’s hard for data agencies to adjust their techniques to “see” this data easily, and not just lump them in with “other debts.”

That is starting to change. The Federal Reserve Bank of New York is doing some high-end analysis of student loans, and their economists Meta Brown and Sydnee Caldwell have a great post from last week,“Young Student Loan Borrowers Retreat from Housing and Auto Markets.” They find that over the past decade, people with student loans were more likely to have a mortgage at age 30 and a car loan at age 25. In the crisis this edge has collapsed:

There’s a similar dynamic for car loans.

The researchers argue that two obvious explanations stands out for this collapse. The first is that the actual future expected earnings have fallen for this group, so they are going to spend less. The second is that credit constraints are especially binding, as those with student loans have a worse credit score than those without.

Derek Thompson at The Altantic Business responds critically, arguing that: (1) cars and mortgages are falling out of favor with young people, so this is likely a secular trend; (2) young people are essentially doing a “debt swap,” switching cars and mortgages for education to take advantage of an education premium, and the cars and mortgages will come later; and (3) though this is, at best, a short-term drag on the economy and reflecting short-term problems, it’ll super-charge our economy come later.

What should we make of this?

(1) It’s possible that there is a secular trend to it, with young people not wanting mortgages or cars. But why wouldn’t the spread survive? “People with student loans” is a broad category of people, and it is difficult to assume that it’s just people moving to become renters in urban cores driving the entire thing. The collapse of the spread between the two coinciding with the crisis makes it hard to believe it’s just a coincidence.

(2) As discussed at the beginning, the overall idea in the student loan data literature is that student loans shouldn’t have a negative impact on consumption, especially at the national level. The extra cost of servicing the debt is more than balanced out by the extra income earned, even if the length of the debt needs to adjust to meet that. Indeed, there’s often a “best investment ever” or “leaving money on the table” aspect to the discussion of higher education and student loans. So if this data holds, it’s a major change from the normal way economists understand this.

And the issue of student debt is where the problem with the “education premium” is going to hit a wall. The college premium is driven just as much by high school wages falling as it is by college-educated wages increasing, which has slowed in the past decade. So if you have to take on large debt to secure a stagnating college-level income, it suddenly isn’t clear that it is such a great deal, even if there’s a strictly defined “premium” over the alternative.

(3) It isn’t clear that the upswing in people, particularly women, taking on additional education is involved with this collapse in borrowing, as the ages of 25 and 30 cut off many people in school. I think it would reflect the collapse in the housing market, but the auto loan market is there as well. It is true that the economy as a whole is deleveraging, but that is largely reflective of housing and foreclosures.

How much this reverts if we get back to full employment and whether there’s a “swap” that could lead to a better long-term economy are good questions, but the fact that we even have to put the question these way shows a change in what economists believed about student loans. No matter what, this shows that education isn’t enough of an insurance against the business cycle.

And I actually see it the other way – right now Ben Bernanke is working overtime to try and get interest rates to the lowest they’ve ever been, and he still can’t induce borrowing by college-educated young people. Congress also lowered interest rates on new student loans, though too many student loans are out there at high rates given the disinflationary times. If the lower lending isn’t the result of institutional issues with credit scores, that means college-educated young people are particularly battered in this economy. And there could be a low-level drag on the economy for the foreseeable future.

If the New York Fed is taking requests, the biggest question I have is how student loans are impacting household formations. Young people are living with their parents for longer at a point where getting an additional million homebuyers would supercharge the economy. Are they living at home because they are unemployed, or because they are un(der)employed and have student loans? If it is the second, then there’s definitely a serious lag on the economy.

But the real issue revealed by this study is that this stuff is important. It is showing up in national data; the people arguing that student loans simply disappear under higher earnings now have a macroeconomic issue to deal with.