Debt Burden and Postgraduation Behavior

A concern often raised by those studying student loans is the impact of borrowing for undergraduate education on students’ postbaccalaureate plans.

Excessive levels of borrowing or, more specifically, high levels of debt burden (the relationship between student loan repayment amounts and income) are often cited by policymakers and in the media as impediments to students’ decisions to attend graduate school or take jobs in low-paying yet socially valuable fields.

As with the research on undergraduate enrollment behavior, the ideal method of approaching this question would be to conduct a randomized trial, where students graduating from college would be randomly saddled with varying levels of student loans.

One could then examine the choices students made in coping with their student loans and repaying them, and ascertain whether higher levels of debt burden (controlling for other factors) led students to decide against graduate school enrollment or entry into low-paying fields.

But once again, these kinds of true experimental studies are impossible to carry out. So we are left with the observational studies that try to exploit the statistical power of research databases, or surveys of students about their attitudes toward loan repayment and debt burden, to try to ascertain how big a problem it is.

Nellie Mae, a lender and subsidiary of Sallie Mae, has conducted four surveys of students on their attitudes toward their student loans and debt burden. The last two were conducted after the increase in loan limits enacted in the 1992 reauthorization of the Higher Education Act (Baum and O’Malley, 2003; Baum and Saunders, 1998).

In these most recent two surveys, the authors reported that overall, students appeared to be managing their debt burden adequately, and students reported that loans were an important source of financing for college.

They did find, however, that even though most students were managing debt burden reasonably well, there were pockets of students who faced high levels of debt burden and who reported that this was affecting some of their education decisions.

In addition, they found that negative attitudes toward debt were increasing over time and that students from low-income families—even controlling for postcollege earnings—were more likely to report problems with repaying their loans.

In a 2001 study (Heller, 2001), I examined the relationship between levels of borrowing, debt burden, and the postbaccalaureate career and graduate school decisions of students. Using data from the 1992 NPSAS (which also followed graduating seniors longitudinally), I found that levels of undergraduate borrowing had little impact on students’ decisions to attend graduate school or not.

While higher levels of borrowing were found to be statistically significant in predicting whether students would enroll in graduate school, this effect was much smaller than the effect of characteristics such as academic performance, choice of major, and overall degree expectations.

However, in that study I provided the caveat that the cohort of students being studied graduated from college before the loan limit increases enacted in 1992, so that one needed to be cautious in applying the findings to later cohorts who incurred larger levels of borrowing.

A similar study conducted by Price (2004) compared the debt burden levels of students from different racial groups. While my 2001 study found no differences in average debt burdens across racial groups, Price did find higher levels for African American and Hispanic students.

The difference, however, is due to methodological approaches. My study examined the debt burden only of fulltime workers, while Price included all graduates. The difference in the findings between the two studies can be attributed to the fact that African American and Hispanic graduates were less likely to be working full-time.

Monks (2001) examined a cohort of students who graduated from a set of expensive, selective, private institutions in 1998. The majority of these students attended college after the 1992 loan limit increase. The author looked at the relationship between levels of student loan borrowing and graduate school and career decisions made by students. He concluded that:

While a majority of student borrowers do not feel overly burdened by their level of indebtedness, a sizeable minority of borrowers does feel adversely affected by their student loans. Despite these feelings of anxiety, student loans do not appear to have a significant impact on the post-graduation plans of students, in terms of planning to attend graduate or professional school in the fall following graduation, the type of degree program sought, and the career plans of seniors. (p. 550)

This study highlights the potential disconnect between reported feelings of concern or anxiety over levels of borrowing, and the actual behaviors these students exhibit. An important caveat of this study is that it used students from elite institutions, those who likely have the brightest graduate school and employment outcomes by virtue of their human capital.

Harrast (2004) examined cohorts of students who graduated in the post-1992 era from the University of Memphis, a large public university. The study merged institutional records with state unemployment data to measure levels of debt burden (ratio of loan payments to income), and examined students with different undergraduate majors.

The author found that while median debt levels did not appear to be problematic (using the 8 percent benchmark) for the students graduating from this university, in all but three majors (finance, management information systems, and sales) students whose total borrowing was in the upper quartile did face debt burden exceeding the benchmark.

In a recent National Bureau of Economic Research working paper, Rothstein and Rouse (2007) examined the impact of a policy change that served to eliminate loans for students eligible for aid at one institution.

The institution studied was described as a highly selective university.10 In 1998 the institution first removed all loans from financial aid packages (and replaced them dollarfor-dollar with grants) for students from families with incomes below $40,000, and then three years later, the same policy was put in place for all students receiving aid (approximately half the institution’s students were eligible for aid).

The authors examined the occupations and salaries of students from cohorts before and after the policy changes, in order to determine whether the elimination of student loan debt was related to students’ postcollege career choices.

In order to control for any macroeconomic changes in the economy that could have affected differences across cohorts of students, they also examined students who did not receive financial aid. The authors did find that the policy changes had an impact on the occupational choices of students:

Aid recipients shifted out of industries with high average salaries and into lower-salary industries, while there was little change in the industry composition of jobs taken by students not on aid. While there was no relative decline in the share of aid recipients taking jobs in the consulting, investment banking, and finance sectors—the most prominent high-salary employers of Anon U graduates—there was a notable increase in the share taking jobs in the nonprofit, government, and education sectors [after the policy changes]. (p. 23)

The authors acknowledge some limitations about the generalizability of their results to other college students, given the highly selective nature of the institution and the high level of academic ability of the students admitted. But they provide data to bolster a claim that the debt effects on employment outcomes they found would actually be larger among the general population of college graduates.

Field (2005) examined the prospective forgiveness of loans for law students entering public interest law or other low-paying professions.

New York University introduced a program, known as the Innovative Financial Aid Study (IFAS), that offered grants to entering students worth two-thirds of tuition costs for students entering public interest law jobs. If the student did not pursue such a career, the grant converted to a loan.

Students interested in the program applied, and participants were chosen by random lottery to participate. Nonparticipants were still eligible for NYU’s Loan Repayment Assistance Program (LRAP), which forgave loans for students working in public interest, low-income positions in the legal profession.

The key difference between the two programs was that IFAS provided tuition grants up front, while LRAP forgave loans for debt already incurred. The study indicated that both programs were designed to have the same net present value from the student’s perspective.

The study examined whether the IFAS program (as compared to LRAP) influenced the matriculation and job placement rates of students. The author concluded that:

Under a career-contingent financial aid program that offers tuition waivers rather than an equivalent amount of loan repayment assistance, matriculation rates are nearly twice as high. Furthermore, rates of first job placement in public interest law are roughly one third higher when students are offered tuition subsidies rather than loan repayment assistance. …The fact that career-contingent tuition subsidies are associated with higher rates of public interest law than are financially equivalent backward-looking loan repayment schemes provides strong evidence of the influence of debt aversion on job choice in a high stakes setting. (p. 22)

The author does not speculate as to the applicability of the results to undergraduate students. It is probably fair to argue, however, that the results should be applied with caution to undergraduates because of the uncertainty with respect to the linkage between major and the first job out of college.

While the education-tooccupation linkage is fairly straightforward for law students even given the diversity of jobs available to graduates of law schools many undergraduate fields are much less tightly coupled.

Nevertheless, the results of the Field study (2005) may have some applicability for other postsecondary settings. The studies of the impact of debt burden and borrowing levels on postcollege plans are an excellent example of the importance of analyzing not just students at the median, but also those at the margins.

Most of the analyses of student debt burden levels, even those conducted as student loan borrowing has exploded since 1992, find few problems with average debt levels using what has been the generally accepted debt burden benchmark of 8 percent of income (and perhaps even less concern would exist using the more income-based benchmarks suggested by Baum and Schwartz, 2006).

However, once you start to examine students with debt burdens above the median driven either by higher levels of borrowing and/or lower average earnings levels warning signs begin to appear.

The American Council on Education reported that 62 percent of all students graduating with a bachelor’s degree in the 2003-2004 year borrowed in the federal loan programs (American Council on Education, 2005), an increase from 37 percent in 1993.

According to the National Center for Education Statistics, 1.4 million bachelor’s degrees were received in the United States in 2003-04 (National Center for Education Statistics, 2006).

Extrapolating from the finding of Harrast (2004) that the top quarter of borrowers may be exceeding the 8 percent debt burden benchmark, this could represent more than 200,000 graduates annually who may run into problems repaying their loans.

Another area neglected by researchers is the question of what happens to those students who begin a postsecondary education, take out student loans, but never complete a degree.

This may put students in the worst of all worlds, in that they may incur significant levels of debt without attaining the labor market (and other) benefits of a college degree. One of the few studies to examine these students concluded that, “borrowers who drop out face greater economic hardship due to their debt burden” (Gladieux and Perna, 2005, p. 7).

The authors found that nearly a quarter of students who borrowed and subsequently dropped out of college defaulted at least once on their loans, a default rate three to four times higher than default among all student loan borrowers.

The evidence on whether debt burden is influencing the decisions borrowers are making after they leave college—the issue generally raised by policymakers and the media is mixed. A key limitation is that most of this research was conducted before the large increase in borrowing in recent years.

While some studies were conducted on cohorts of students who were in college after the 1992 loan limit increases, the largest increase in borrowing has occurred since 2000, and no studies are available that looked at very recent graduates.

In the five years from 1995, when the College Board first began tracking private loans, to 2000, borrowing in the federal and private loan programs increased at an 8 percent annual rate (author’s calculations from Figure 1).

From 2000 to 2005, the most recently available data, total borrowing increased at a 15 percent annual rate, or almost double that of the earlier period. And as mentioned earlier, the fastest growing form of borrowing is in the private loan programs, which offer the leastgenerous repayment terms to borrowers.

It is fair to conclude that these increased borrowing levels may lead to problems down the road. Of course, if incomes are sufficient to repay student loans without impinging on students’ labor market or education decisions, then there would be little cause for concern.

Yet there is some evidence that for many students, particularly those in lower-paying jobs and with higher levels of debt, student loans may be causing problems.

As student loan borrowing continues to increase, and presumably rises at rates faster than the income of students when they enter the labor force, the issue of debt burden will become more and more problematic.

This issue is, however, distinct from the question of whether the availability of student loans increases access to college and choice among institutions.

Location: Indonesia