Households have reduced some forms of debt in recent years, but student loans more than doubled between 2005 and 2012.1 In fact, student loan debt is now more prominent than any non-mortgage household debt, including auto loans and credit card debt.2 Unfortunately, growing student debt has also been accompanied by rising default rates. For instance, default rates currently stand at 13.4 percent and 14.7 percent for students entering repayment in 2009 and 2010, respectively. These rates are much higher than a decade ago.3
As with any investment financed with external funds, a high level of student debt is not problematic provided the investment in education has high returns. I briefly review two key components pertinent to obtaining a high return on college investment: success in school (i.e., earning a degree) and then success in the labor market (i.e., finding jobs that pay well enough to justify the investment).
The figure shows education attainment as of 2009, by student debt levels, for the cohort of students who finished high school in 2003-04. Notice that 43 percent of this cohort did not borrow at all for school. Some of these individuals may not have pursued further education, as suggested by the high fraction of nonborrowers without a postsecondary degree. On the other end of the spectrum, a significant fraction of nonborrowers attained a bachelor's degree. The remaining debt is distributed roughly equally among intermediate categories, such as a certificate and an associate degree, and those who are still enrolled in school. The figure also shows that low debt (less than $10,000) and intermediate debt ($10,000-$30,000) tend to be concentrated among those with associate degrees and individuals still enrolled, perhaps suggesting that a fraction of them are part-time students. Degree attainment increases as the amount of borrowing increases: More borrowers with high debt ($30,001-$50,000) attained either an associate or bachelor's degree than those with low debt.
The figure also shows that more than half of those with the highest debt level (more than $50,000) attained a bachelor's degree and more than 20 percent of them are still enrolled. Only a small fraction of individuals with high debt left school with only a certificate or no degree at all. Therefore, default can arise from different sources: Individuals with low debt may have faced difficulties repaying their loans in recent years because of slow labor market conditions for individuals with less than a bachelor's degree. Those who earned bachelor's degrees might have faced better labor markets, but still not good enough to service their higher debt. The most serious problems are for the small but significant fraction of those with high debt and no advanced degree. The fact that they borrow suggests they lack a significant source of wealth, and even if the recovery of the labor market accelerates, they may have problems repaying their high loan balances.
Evaluating delinquency or default on student loans can be difficult if based on the first years of labor market experience, since the early years tend to be more volatile than later ones, when individuals have settled in jobs that fit their needs and are appropriate for their skills. Indeed, student loans typically offer the options of deferment and forbearance to temporarily avoid repaying without falling into delinquency or default. These options are meant to help in the transition from school to the labor market but can further complicate analysis and projections of default rates.
Lochner and Monge-Naranjo (2014) use data on the cohort who earned bachelor's degrees in 1993. They find interesting patterns between repayment of student loans and labor market experience. The first table shows that nonpayment rates are roughly the same 5 and 10 years after school, but the type of nonpayment is very different. Cases of deferment and forbearance are more frequent five years after school, while default rates are much higher 10 years after school.
The second table shows the transition probabilities across the different states of repayment between 5 and 10 years after school. Clearly, having defaulted after 5 years increases the probability of being in default after 10 years. Moreover, the same pattern applies if the borrower is in deferment or forbearance after 5 years. If such a pattern holds for more recent cohorts, the true default rates for recent years are probably even higher than the measured ones because some default might be temporarily masked as deferment or forbearance. The default projections for recent cohorts might need to be downgraded even further, since the results of Lochner and Monge-Naranjo (2014) are for those with bachelor's degrees entering a much better labor market.