CHICAGO - Good morning everyone. Thank you for inviting me to speak at the 56th annual meeting of the National Association for Business Economics. I’m happy to be joining you today.
I was invited to share “all my wisdom” on the subject of the macroeconomic effects of student loan debt, and the invitation made me smile because I remembered a story I once heard John Kenneth Galbraith tell. He, too, was invited to give a speech sharing “all of his wisdom” and he insisted that his wife come along because this was such a big deal. When the speech was over, Galbraith turned to his wife proudly and asked how he had done, and she said, “well, I don’t know how much wiser they are, but they’re definitely older.”
Well, I hope to leave you at least a tiny bit wiser and not too much older today.
My subject is student loans and their macroeconomic consequences. To get the proper perspective on the subject, I want to invoke another economist who I know is a hero to many economists in the room, and that is Adam Smith. He argued that private economic prosperity and public investment in social goods go hand in hand; they reinforce one another and indeed they are impossible without each other.
Of all the many social goods out there, Smith especially championed the importance of education for both the individual and for society as a whole. In The Wealth of Nations , he urged society to make education universally available and accessible, especially to the working poor whose development may be stifled by the all-consuming necessity of labor and the ceaseless hardships attendant to poverty.
Smith insisted that education is not just an imperative for disadvantaged individuals but a general social good. In his view, an educated population will develop habits of mutual civic respect, thoughtful political deliberation, cultural uplift, and the reduction of religious fanaticism and political extremism—all positive social forces.
Smith stated in 1776, that public payment for the expense of institutions for education “is…no doubt, beneficial to the whole society.”
We of course provide free public schooling for grades K through 12 in America, but leave college and university education much more up to market forces. Indeed, we have seen a steady public disengagement from the funding of state universities and colleges over the last decade, even as college and graduate degrees become more, not less, essential to upward mobility and financial security . Our approach to higher education contrasts sharply with that of some other countries, and there are obviously both benefits and costs to the way we organize funding for college tuition. But when all is said and done, it is widely understood, both here and abroad, that we have the best system of higher education in the world. It is a key force in both our society and our economy and it is something we can all be proud of.
But we also now have a higher level of college student loan debt than any other country in the world . This sobering reality has consequences not just for the individual student carrying large student debt, but for the economy as a whole. How do we as a nation maximize the returns on our investment in higher education in such a way as to catalyze economic growth and prosperity? In these remarks I will identify aspects of the increase in student loan debt that are relevant to macroeconomic growth.
In short: We cannot understand the macroeconomic impact of student loans without incorporating the full economic and societal benefits of a more educated workforce. That is, what would the macroeconomic effect be if borrowers had no debt, but also lacked the higher education that comes with it? Student loan debt is not inherently bad, until it becomes too costly to propel people to a place they could not reach without it; until the investment cannot pay off at both the individual and societal level: until it leads to an education system that is accessible to the wealthy. When public taxpayer costs associated with the growing student loan portfolio are quantified, there should also be recognition of the benefits taxpayers receive from the investment in our nation’s human capital.
As we explore the various aspects of this question we should keep in mind Adam Smith’s point that social and personal spending on higher education is not a form of extravagance or profligacy, but an important investment in both the individual potential of our citizens and the general health of our society. Indeed, it is the crucial nature of this investment that requires us to make sure that student loan debt does not have a destabilizing macroeconomic effect that might cause the country to turn away from a proper and healthy investment in higher education for our people or disincentivize individuals’ investment in their own potential.
Outstanding student debt is growing. Americans owed just over half a trillion dollars in student loans at the end of 2007. This increased to over $1 trillion by the end of 2013, suggesting a growth rate of approximately 11 percent per year. Student loan debt is now larger than household credit card debt and auto loan debt, though it remains substantially smaller than levels of household mortgage debt. The recovery from the recession has seemingly brought little respite: in the first two quarters of 2014 alone, outstanding student loans have grown by another four percent to $1.12 trillion . Nevertheless, total federal originations have fallen since their 2012 peak and originations per borrower have fallen since 2010 . If these trends continue as the recovery strengthens further, they may noticeably slow the growth of outstanding student loan debt.
The steady growth in federal student loan debt looks different from trends for other types of household credit, which grew prior to the financial crisis before collapsing sharply. This may be explained in part by the fact that, because federal student loans are not underwritten, availability did not tighten after the financial crisis, unlike private lending in various markets . It may be the case that the demand for student loan credit rose, as the recession eroded household wealth and caused job loss, reducing households’ ability to pay for higher education while increasing people’s desire to postpone employment and enroll in school . These drivers for student loans were reinforced by the long-term trend of rising tuitions that had been underway even before the crisis. Affordability is a key data point that is driving the growth of student loan debt.
Some of the growth in debt volume is fueled by higher tuition costs, and some of the growth is fueled by rising student enrollments and a rising share of students borrowing. Survey data from the Department of Education show that the average 4-year degree recipient in 2012 who borrowed money for college owed $31,000 in federal student loans, up from an inflation-adjusted $25,000 in 2008 . Debt for graduate school has increased even more quickly: the median graduate degree recipient in 2012 who borrowed money owed about $58,000, up from $44,000 in 2008.  So, recent cohorts appear to be taking out more debt to pay for education. Rather than looking only at recent graduates, data from the New York Fed include borrowers at all stages of repayment, including those who left college many years ago and those who have not yet finished. These data show that real outstanding debt per borrower increased from $21,000 in 2007 to $25,000 in 2012, with the number of borrowers increasing from 28.9 million to 38.8 million. 
These averages mask considerable heterogeneity, however. Both debt amounts and incomes vary greatly across borrowers, and many borrowers are struggling to make loan payments. Data from the New York Fed show that the fraction of borrowers who are struggling to make payments has trended up between 2003 and 2012. More recently, approximately nine percent of outstanding student loan volume in Q2 2012 was seriously delinquent, and that has increased to 11 percent in Q2 2014.  Undergraduates experience more difficulty repaying than graduate students, with a default rate seven times higher.  These are the big data and big trends setting the stage for analysis.
Student Loans Are a Special Kind of Investment
Now we can’t forget that student loan borrowing is different from other forms of household debt for another reason. And that is this: student loans fund an investment in human capital and human development.
Unlike many other forms of credit, like credit cards, student loans fund an investment, rather than consumption. 
So when we evaluate macroeconomic impact, we have to consider that rising debt has paid for an increase in the numbers of people able to receive higher education and professional training, and has therefore raised incomes and increased growth, as well as provided other benefits. When we consider a counterfactual situation with less debt, we need to consider whether the particular hypothetical policy reducing debt would also reduce access to higher education, together with its well-understood social and economic benefits.
For example, the strong relationship between higher education and individual earnings is well known. A four-year degree yields approximately $570,000 more in lifetime earnings than a high school diploma alone, while a two-year degree yields $170,000 more.  In addition, the likelihood of positive externalities from higher education, such as increased productivity, increased innovation, and increased civic engagement—the benefits spoken of by Adam Smith—means that these economic returns are an underestimate of the true return that the American economy and society receive from investments in higher education. Higher education is also crucial to social mobility, as everyone from Thomas Jefferson and Abraham Lincoln to Lyndon Johnson and Barack Obama have maintained. Without a college degree, children born in the bottom income quintile have a 45 percent chance of remaining there as adults. With a degree, they have less than a 20 percent chance of staying in the bottom quintile of the income distribution and a roughly equal chance of ending up in any of the higher income quintiles. 
The Payment Burden of Student Loans
Having established the general social utility of investment in education does not establish the investment value of individual student loans, and does not mean that borrowing is always the right decision for everyone. The investment value is influenced by the borrower’s experience and expectation about future financial burdens, but on this point we lack a lot of significant and concrete data.
Indeed, one of the challenges with analyzing student loan burdens is that we have only an imperfect dataset about them. The Federal Reserve Board’s Survey of Consumer Finances includes the appropriate variables of income, debt, and other characteristics, but it is at the household, rather than the individual, level. In addition, the Survey of Consumer Finances is relatively small, which makes it difficult to draw strong conclusions about the most vulnerable households. It is also a triennial survey, which comes out with considerable lag, limiting its use for capturing emerging trends at any given time. 
That said, the data in the Survey is being used for analysis. Recent analysis by researchers at the Brookings Institution uses the Survey of Consumer Finances and finds that the median payment-to-income ratio for young households with student loans hovered around four percent between 1992 and 2013.  But again, this masks considerable heterogeneity and data showing larger fractions of borrowers struggling to make payments.
A more fundamental challenge in analyzing the effect of student loan burden on economic growth is the need to determine how student loan borrowers assess their own financial burden. In other words, do student loan borrowers consider their monthly debt burden when they incur the loan or when they make decisions about purchasing homes or starting businesses or choosing careers? Or are student loan borrower decisions more focused on total debt? These questions tie to an anecdotal trend we have noticed. Academic studies cite monthly burden amounts, but borrowers tend to speak of their debt in the aggregate: “I owe $75,000” or “I owe $40,000.” This is a different way of looking at debt than understanding a burden as a monthly obligation—“I pay $450 per month” or “I pay $300 per month.” We simply do not yet know whether monthly debt burden, rather than total balance owed, is the relevant factor that individual borrowers have in mind when they are thinking about their debt. This is important because decision-making regarding whether to buy a home, launch a business, or start a family may be more a function of the ability to juggle student loan payments with other bills, and the type of job your education will lead you to, especially in a still-recovering labor market, and the psychological burden of owing tens of thousands of dollars in total student debt, and for example, just the amount of your monthly student loan payments. Because individuals tend to report the amount they owe, rather than the amount they pay per month, when asked about student loan burden, it may be that “total debt” is a cognitively salient feature for borrower decision-making. Accordingly, a framework for assessing macroeconomic impact should consider the need for data that can be a relevant proxy for understanding how borrowers themselves are assessing their debt levels.
The Macroeconomic Effects of Delinquencies and Defaults
While the vast majority of student loan borrowers are making their loan repayments and managing their debt, and many are seeing their investment in college and/or graduate school pay off well, student loan delinquency and default have afflicted a small but growing percentage of borrowers. Delinquency and defaults in sufficient numbers can undermine economic growth by crowding out other kinds of investment.
First, let’s consider the effect of a current student loan on the decision to buy a house, for example. The channel through which student loan payments may dampen growth is intuitively obvious: money spent on debt service cannot be spent on housing .
Indeed, the average monthly pace of single-family housing starts and permits so far this year remains only slightly above 600,000, far below pre-crisis levels. Sales of new single-family homes were up just two percent during the first eight months of 2014, slowing sharply from a double-digit pace of growth in 2013, and the average sales pace for existing single-family homes through August was five percent lower than average during 2014. Household formation has been less than half its historical rate since mid-2006. To be sure, many factors have contributed to the weakness in the housing market such as changing demographics, changing housing preferences, weak job growth, flat wage growth, and tight credit conditions . But perhaps the financial burden of student loans, too, is contributing to low homebuying. What if debt service going to student loans could instead be saved for a down payment or used to get a larger mortgage?
For example: the nominal student debt per borrower in 2007 was $18,957. This nominal student debt per borrower increased to $24,803 in 2012. A higher debt balance means higher payments are needed to service that debt. Indeed, the average borrower in 2012 would have saved about $800 per year in student loan payments if he or she had borrowed at 2007 levels .
In terms of the size of a typical mortgage or house, this extra $800 per year could have purchased an additional $13,300 in housing, assuming a 30-year 4.5 percent mortgage.
Alternatively, we can think about this $800 relative to a down payment. A typical 10 percent down payment on the median home is $16,300, so the $800 represents five percent of the needed amount. So, the effect of a current student loan on the decision to buy a house is—or is not—significant depending on the borrower’s other financial resources or how short the borrower is from having enough for a complete down payment.
However we evaluate this effect with loans that are current, the effect seems magnified with loans that are delinquent. What I call the “damaged credit” channel may be the most dramatic channel through which to see macroeconomic effects, including in the housing market. So consider now a graduate who is delinquent or in default. Delinquencies for young adults are particularly relevant, since this is the population of potential first-time homeowners. This delinquency constitutes a negative credit event, which would limit access to credit, making it harder for the potential homeowner to get a mortgage. Negative credit events exacerbate access to not just mortgage credit, but also to other forms of credit. And when delinquencies become defaults, the full weight of the government’s collection efforts drop on the borrower, accentuating the negative credit event of a delinquency. While federal loans come with borrower protections—such as income-driven repayment options—that other forms of consumer credit do not, student loan debt generally is not dischargeable in bankruptcy and therefore, the relief of unmanageable repayment obligations that otherwise eventually come with bankruptcy appears impossible.
Concerns for the Future
In light of this analysis, what can be done to address the student loan burden, and the commensurate, even if difficult to measure, impact on our economy?
First, without a doubt, student loan debt has become a serious burden for far too many borrowers. Although none of us now expects a student loan meltdown because there is a great deal of integrity and stability in the student loan market, consumer advocates continue to report large numbers of complaints and testimonials from distressed borrowers who have received misinformation from student loan servicers or who have unnecessarily entered default or delinquent status . We must do everything in our power right now to guarantee accuracy and fairness in the loan servicing industry, and transparency and disclosure for borrowers in the loan process.
Despite the amount of data we have on the student loan debt situation, there are many things we do not know. Even so, macroeconomic analysis would be more nuanced if we had more granular data at the individual level to better capture the heterogeneity of the student loan borrower population.
First, we need to better understand how borrowers view their loan repayment burden. Is it a function of their total amount owed and the strength of entry-level jobs? Or is it a function of monthly debt payments? What other financial resources do they use? How do they evaluate their different options for higher education? To what extent do they understand all their options for repayment?
Second, we need to understand more precisely why borrowers reach delinquent or default status. We need to consider ways to reduce unnecessary delinquencies and defaults so as not to impair borrowers’ future access to credit and compromise their ability to avoid garnishment and loss of federal benefits.
Making policy decisions to support the vitality and vigor of higher education in the face of troubling trends in delinquent or defaulted student loan balances is a responsibility we cannot shirk. We possess, as a nation, a wealth of highly talented and motivated individuals who chose the rigors of advanced studies to improve their lives. Among these individuals are many who will themselves be catalysts for economic growth and job creation.
We have a solemn covenant with all generations of Americans to treat them fairly and openly through the educational process. And we have a system of government with a rule of law that provides the foundation for business growth. With the right policies in place to support the tens of millions of Americans who are investing in their own potential, we will reach a point in the not so distant future where the knowledge and understanding acquired by so many will become the motor for innovation. Even with these hopes for the collective future, we need to remain attuned to the individual voices detailing the economic pressures and financial burdens associated with being a student loan borrower.
I thank you for the opportunity to speak to you today, and I invite your help as we work to fulfill our nation’s historic commitment to universal educational opportunity while maximizing macroeconomic growth and prosperity.
 The Deputy Secretary would like to thank the following Treasury Department leadership and staff members for contributing to this work: Seth Carpenter, Tiffany Chou, Karen Dynan, Jennifer Hunt, Lepi Jha, Melissa Koide, Jake Liebschultz, Adam Looney, Monique Rollins, Matthew Rutherford, Guari Subramani, and Sam Valverde.
 There is also the possibility that some of the rise in student debt may reflect households’ substituting one form of credit for another.
 Tabulated from the National Postsecondary Student Aid Survey (NPSAS) 2008 and 2012.
 FRBNY (2013). Student Loan Debt by Age Group. http://www.newyorkfed.org/studentloandebt/. The nominal increase is from $18,957 to $24,803. FRBNY’s published numbers are not directly comparable to ED statistics since their credit bureau data cannot distinguish between different “types” of borrowers (i.e., those who borrowed for their own undergraduate education, graduate borrowers, and parents who borrowed for their children) or lifecycle effects due to borrowers at being in different stages of repayment.
 FRBNY (2014).
 This distinction between student loans funding consumption and investment is not as sharp in practice. First, higher education could have some direct consumption value, as some people enjoy college in and of itself. Second, institutional amenities (e.g., rock-climbing walls, upgraded gyms, fancy dormitories) are one form of consumption, and if students are borrowing more to attend schools with generous amenities, then that is a form of consumption-smoothing. It is not necessarily suboptimal or inefficient for students to choose to fund such consumption with student loans, though policymakers may need to evaluate whether this is a desirable use of federal loans.
 Isaacs, Sawhill, and Haskins (2011). Professor Amir Sufi has suggested that repayment of debt should depend on economic returns. This idea drives from the notion that the value of the investment depends on the job market and economic prospects. Amir Sufi, Who is the Economy Working For? The Impact of Rising Inequality on the American Economy, Testimony Before Sen. Comm. on Banking, Housing, and Urban Affairs, Subcomm. on Economic Policy, Sept. 17, 2014.
 There are also potential consumption effects if a student loan impairs borrowers’ access to credit, making it more expensive or impossible to debt-finance consumption via credit cards or auto loans. Quantifying such effects is very difficult, however.
 Note that delayed household formation and homeownership could also be the result of choices that are not the result of economic constraints. In particular, changes in norms or preferences around cohabitation and home ownership could lower household formation and the demand for homes.
 At these levels of debt, annual student loan payments would have increased from $2,618 to $3,425, assuming the interest rate was the one charged on federal student loans over the past few years, 6.8%.