Explore our maps here.
Since the Great Recession, outstanding student loan debt in the United States has increased by 122% in 2019 dollars, reaching the staggering sum of \$1.66 trillion in June of this year. Student loan debt has grown faster than other debt types, including auto, credit card and mortgage debt. For many, education is the only pathway towards good employment with benefits, leading to economic and social opportunities later in life. But as college becomes more unaffordable with each passing year, student loans are bridging the ever-expanding chasm between college savings and obtaining an education. The crisis has reached the national political arena, with policymakers recently calling for debt cancellation up to \$50,000 for federal borrowers.
Our research demonstrates that the student debt crisis has deepened existing inequalities. We found that all young borrowers are saddled with dramatically rising debt since 2009, but low-income groups, BIPOC, and those in their thirties fare worse than others. While richer students have higher absolute debt, low-income students experience massive and growing relative debt burdens. And students in majority-Black and Hispanic zip codes, who are more likely to attend for-profit private institutions, have seen larger debt increases than those in majority-white zip codes. Debt levels have jumped in states like Wisconsin, Michigan, Pennsylvania, and Ohio. Gaining insight into broad trends in debt accumulation, as well as details about the particular demographic or labor market characteristics that shape changes in individuals’ debt burden over time, allows us to more effectively tailor our policy recommendations. For example, our research finds that forgiving $50,000 in student loans would make 80% of young adult borrowers student debt free.
In December of last year, we undertook a long-term investigation into the roots of America’s higher education crisis. Using a novel measure of access to higher education, the School Concentration Index, we mapped the distribution of higher education institutions across the country and analyzed its relationship to factors like race and income. In our June 2020 report, we uncovered several worrying trends. Worsening enrollment concentration, higher prices, fewer campuses, more education deserts, and the adoption of for-profit models by non-profit institutions have significantly skewed access to higher education across the country. In the latest iteration of our series, we present new findings on the most important symptom of these developments: student debt.
Multiple factors have contributed to the explosion in student loan debt. Higher tuition and living expenses have made being a student more costly, and labor market credentialization has enticed growing numbers of students to become borrowers.1 But problems with repayment and capitalized interest are also to blame. Behind the scenes, broader aspects of the American political economy impact borrowers and the debt they carry.
In the past decade, the growth in total student debt balances outpaced the growth in the real price of college, as well as the growth in tuition and fees. Using annual random samples of credit archives on one million 18–35 year-olds with student loan activity, we closely examined this dramatic rise, documenting a drastic increase in balances per person. Among young adults with at least one student loan in 2019, the median total loan balance owed was \$18,727, a 17.93% increase from 2009 in 2019 dollars; the average total balance for this group increased by 23.09% to reach \$34,839. Percent increases varied across regions as well, with the Great Lakes, Southeast, and New England experiencing the largest percent increases in median student loan debt since 2009.
The number line below visualizes the ranking of states according to median income and 18-35 year-old median student debt, including the percent change in median debt since 2009 (all dollar amounts adjusted to USD 2019). Certain states like South Carolina exhibited a concerning combination of extremely low median incomes and high median debt balances, which are increasing rapidly. Michigan, Wisconsin, Pennsylvania, and Ohio saw median student debt grow by over 26% between 2009 and 2019. These four states also had median debt among majority-Black zip codes rise by over 68% between 2009 and 2019, whereas the nation as a whole saw majority-Black zip codes’ median debt increase by 54%.
This increase is also present when examining total balance by deferral status. Growing overall load debts have led to higher balances per person, as referenced in the total balance distribution above and by deferral and non-deferral status below. In 2013, median and average total balances really started to climb, likely due to peaking enrollment expansion. Research on outstanding federal direct loans shows that the majority of debtors owe under \$20,000, but a mere 7% of borrowers with balances above \$100,000 hold 36% of overall student debt. Our samples are also skewed in this way, as reflected in the graphs below. Across both federal and non-federal loans, only 15% of borrowers in 2009 owed over \$50,000 in student loans; in 2019, that number had jumped to 20%. In 2019, only 7% of young adult borrowers had over \$100,000 in student loans. By forgiving \$50,000 in student loan debt, 80% of borrowers become student debt-free and 13% have at least half of their debt burden erased.
The student debt crisis is characterized by an increasing number of debtors and a greater prevalence of falling behind on payments. With such a large growth in the number of borrowers, the percentage increases in the median and average loan balance among borrowers should be understood as particularly severe. Increasing balances for people with student loans, and an increased number of people with student loans, are crucial vectors for understanding the advent of the crisis.
|Postsecondary Enrollment at Institutions (excludes online-only and graduate-only institutions)|
|2007-2008||22 M||17 M||4.9 M|
|2010-2011||25 M||18.9 M||6.1 M|
|2014-2015||23.4 M||17.7 M||5.6 M|
|2018-2019||22.2 M||17.4 M||4.7 M|
Our data highlights the significance of several underlying trends. Post-recession enrollment growth means that more students in aggregate are experiencing deeper indebtedness; during the post-recession enrollment boom, a growing percentage of full-time first-time students at two-year and four-year schools received student loans (45.5% in 2007-08, peaking at 51.2% in 2011-12, and settling to 46.1% in 2017-18). The percentage increase in debtors was more pronounced at certain institutional types—especially concerning is the trend at for-profits, which have poor outcomes, low completion rates, and the highest average prices and default rates.
|Percent of Enrolled Students Receiving Student Loan Aid by School Type|
|Institution Type||2000-01||Peak post-Recession percentage||2016-17|
|Private not-for-profits||57.7%||64.3% (2010-11)||62.3%|
|Private for-profits||63.5%||82.0% (2010-11)||73.5%|
However, the numbers above exclude below-two-year institutions; when we include these in the calculation, percentage bloat across institution types became even more pronounced (59.5% in 2007-08, 66.3% in 2010-11, and 59.7% in 2017-18).2 The increase in aggregate students, the increase in the proportion that borrow, and the occurrence of these trends within the confines of schools with poor labor market outcomes is a clear indication of how the Recession escalated the student debt crisis.
The number, share, and composition of overdue student loans have vacillated over the time period alongside enrollment. Student loan accounts that are past due accounted for 9.9% of loans in 2009, peaked at 11.6% in 2015, and landed at 7% in 2019. But while the proportion of overdue accounts has decreased, the total number of overdue accounts has increased, only possible due to the growth in the aggregate amount of student loans. That means young adults today are juggling more loans than young adults a decade ago. The total number of overdue accounts in 2019 is well above 2009, at 251,469 accounts behind payments—this is down from its staggering peak at 408,385 in 2015.
When we zoom in on these overdue accounts, the situation becomes even starker, with the share in severe delinquency at 76%. The 2015 expansion in income-driven repayment (IDR) likely explains the improvement we have seen in the share of severe delinquency since 2016 and the attendant reduction in total overdue account balances. Future research in this series will more closely examine the student loan payment patterns and inferred IDR enrollment.
The gravity of the debt crisis becomes more pronounced in light of demographic disparities. Looking at the relationship between zip code tabulation area (ZCTA) median income and median per-person student loan debt for 18 – 35 year-olds, it’s quite clear that richer zip codes generally experience higher debt burdens.3 But it’s also clear that all income groups are seeing their debt burdens rise at the same pace, with increasing prevalence of extremely high balances among the poorest ZCTAs—more so than in the past. Notably, the debt burden of the lowest income groups in 2018 is higher than that of the high income groups in 2009. All income groups saw increases in median total student debt of at least \$4000 (USD 2019), with the lowest income and highest income zip codes experiencing increases greater than \$5000 (USD 2019).
The above graphs also reveal the disparity in total balances between majority-minority and majority-white zip codes, which has become more pronounced over the last decade. Research shows that student debt balances are statistically higher for borrowers from Black and Hispanic ZCTAs, and that higher proportions of students from these ZCTAs take out loans. Black students are more likely to attend for-profits and two-year colleges, which have higher default rates than four-year schools. Black students also have a default rate much higher than Hispanic and white students, regardless of school type. While in our data, median total balances increased by 17.93% over the 11 year period, this increase was not uniform across zip codes according to their racial predominance: median total balances in majority-Black zip codes experienced a 54% increase, majority-Hispanic a 42.81% increase, majority-white a 38.12% increase and majority-Asian a 20.63% increase (there was not enough data to determine percent increases in Pacific Islanders, Native Alaskan or First Nation Peoples dominated areas).
Absolute debt burden is one problem, and relative debt burden is another. Though lower income ZCTAs usually have smaller balances, the debt burden in relation to income in these areas is alarmingly high. And the disparities have widened between 2009 and 2018. The lowest income ZCTAs saw relative debt burdens increase from 56% to 94% over the 11 year time period, whereas the richest ZCTAs saw relative debt burden increase from 32% to 42%. Students from lower income families are enrolling in postsecondary education at increasing rates and borrow at the same rate as middle income households. This trend corresponds to the higher rate of default among lower income borrowers, even though they carry smaller balances than their wealthier peers. Lower income borrowers are also increasingly likely to attend for-profits and two-year colleges, where the default rate is higher than that of four year schools.
Disparities by age group are an enduring trait of the student loan crisis, with 31–35 year-olds holding the highest median student loan debt. Although student loans make up a smaller share of this age group's debt, 31–35 year-olds also have a higher prevalence of delinquency on student loans, and they consistently account for a third of the student loan borrowers in our random 18–35 year-old samples. This indicates that the oldest age group in our analysis is incredibly debt-ridden—they have the highest student loan balances, and that student loan debt, on average, accounts for only 20% of their overall debt. It is likely that today’s 31–35 year-olds are benefiting more from IDR plans, which the Obama administration expanded in 2010 and 2015, and are thus making lower payments while interest continues to accrue. Regardless of risk or repayment status, balances are trending upwards for this age group at a rate much higher than borrowers on the whole: 30% and 18%, respectively.
Student debt and school concentration
Analyzing the relationship between student debt and the School Concentration Index4 revealed a meaningful overlap. In the binned scatter plot above, ZCTAs were clustered by median income and student loan balance on loans opened in the last twenty-four months, then plotted according to their demeaned SCI and median student loan balance. Our previous reports already outlined the troubling negative relationship between income and SCI; the above graph illustrates the equally worrisome positive relationship between debt and SCI—more concentrated higher education markets have higher balances on recently-opened loans. The causes behind this overlap, and the policy recommendations it necessitates, require further research. But the findings indicate that students in highly concentrated markets may be taking on more debt, either by borrowing according to high priced local markets or by attending schools non-locally. Making the connection between enrollment concentration, prices, and debt levels for geographies is an important part of assessing the debt crisis.
|Median Net Price and Total Student Loan Balance According to Dominant School Type in a ZCTA 2017–18|
|More Publics||More For-Profits|
|Majority White Zips|
|Majority Black Zips|
|Majority Hispanic Zips|
|Majority Asian Zips|
ZCTAs with a higher proportion of for-profits than public schools in our analysis, throughout 2009–2019, (for example, Madison, Columbus, Little Rock, Grand Rapids, Salt Lake City, Boise) also experienced higher student loan balances and higher net prices. When classifying zip codes according to racial majority, we see the same relationship across school types. Majority-Black ZCTAs with more for-profits than public institutions see the highest debt balances and net prices. This consistent relationship points to the dual nature of the US higher education crisis: unequal access to an already inadequate public good.
The student debt crisis is a symptom of various concurrent trends. We discussed the impact of growing enrollment, rising prices, larger balances, the number and composition of borrowers and overdue loans, disparities in balance and repayment on the basis of race and income, and finally market concentration in education. Following the Recession, the enrollment boom and growing percentage of students who borrow exacerbated the already burgeoning crisis. Rising prices, costs of living, and default rates, in tandem with poor labor market outcomes, worked with the enrollment boom to increase balances for borrowers across the income distribution.
Low income borrowers, while holding smaller balances than their wealthier counterparts, are constrained by significantly higher debt burdens, and as a result, they fall further behind on payments. BIPOC and GenXers have higher balances than their peers and face higher default rates to boot. The make-up and distribution of higher education institutions across the country also seems to have a strong relationship with the debt burdens and prices, with students in highly-concentrated education markets and ZCTAs dominated by for-profit institutions seeing higher debt burdens. The findings support arguments for targeted and broad debt cancellation or relief. But just as the cause for crisis is diverse, so should be the solutions, whether they include increased funding, expansion, regulation, or debt cancellation.
Julie Margetta Morgan and Marshall Steinbaum, The Student Debt Crisis, Labor Market Credentialization,
and Racial Inequality: How the Current Student Debt Debate Gets the Economics Wrong, Roosevelt Institute, October 2018. ↩
The relationship between high income and high debt burden has been a chief argument against designing a truly universal student debt jubilee. For more on these arguments and the racial wealth gap, see here. ↩
The school concentration index (SCI) is a measurement that assesses the concentration of higher education institutions in a given geographical region. SCI varies from 0 (perfectly unconcentrated) to 10,000 (complete monopoly) to null (education desert). We’ve computed the SCI for every zip in the US and US territories for 2008–2017, including satellite campuses offering degree programs. (The analysis we discuss in this post was completed using forty-five minute driving distances around higher ed institutions.) For more on the index and its relevance for examining the crisis of American higher ed, see our report from June 2020. ↩