Select Page

# A Primer on the Past, Present, and Future of Higher Education

To begin with, the study takes a deeper dive into the data on college tuition and the leading theories seeking to explain its stubborn rise. One striking feature of American higher education is the considerable variation in outcomes across different types of institutions. For example, while average sticker price tuition stands at nearly $20,000 overall, it ranges from less than$10,000 for nonselective, teaching-focused public colleges to nearly $45,000 for selective private schools. Thus, any comprehensive explanation for college tuition inflation must account for not just the overall average but also the full distribution of changes over time. No single theory has yielded an authoritative explanation for tuition inflation, but this paper discusses the evidence regarding some of the most prominent explanations—namely, regarding the roles of financial aid, state funding, productivity differentials, and even the higher education decision-making model itself. Turning from costs to benefits, the paper presents evidence that the labor market benefits of college attainment remain strong, though recent stagnation in the college wage premium along with continued increases in student debt show that the race between debt and opportunity may become tighter if trends continue. Moreover, the economic benefits of a college degree vary widely by field of study, quality of institution, and the academic preparation of incoming students, implying that there is no single economic rate of return to college. Particularly with regard to academic preparation, efforts to increase enrollment may even be counterproductive to the extent that they disproportionately attract nontraditional students who are less likely to graduate. The distinction between the economic returns to college enrollment and the returns to attainment (i.e. successful graduation) also goes a long way toward explaining student loan default. Rather than being driven by college graduates taking on large sums of debt, student loan default is concentrated among college dropouts with relatively small debt balances but sufficiently bad labor market prospects to inhibit their ability to pay. This paper documents several negative consequences of student debt besides default, ranging from constrained career choices to delayed homeownership and reduced entrepreneurial activity. Finally, this paper discusses several guideposts for reform as well as directions for future research. Research has not yet revealed a smoking gun that can explain rising higher education costs. However, evidence suggests that expanded financial aid and a growing college wage premium have increased the demand for college and thus have increased prices, particularly in the face of slower productivity growth in higher education compared to the rest of the economy. When it comes to dealing with these costs, research provides a strong justification for pursuing structural reforms to the student loan program as well as for allowing greater innovation and flexibility rather than pursuing one-size-fits-all policies aimed at guaranteeing free (i.e., taxpayer-financed) college or instituting blanket student loan forgiveness. ### Introduction Increasing alarm about the cost of college and the toll of student debt has led a growing chorus of voices to question the viability of the current model for the US higher education system. Some have gone so far as to raise doubts about whether college is still a sound investment. While many Americans continue to benefit from a college education, each year millions of borrowers become delinquent on their student loans, and many more struggle under the financial burden of debt payments to the point that they are unable to purchase a house, start a family, or launch a business venture. 1Federal Reserve Bank of New York, “Press Briefing on Household Debt, with Focus on Student Debt,” April 3, 2017, slide 24. Although attitudes and hiring practices have begun to shift away from treating a bachelor’s degree as an absolute necessity, a four-year degree is still seen by many as a reliable path to economic opportunity, which makes the challenge of rising costs all the more salient. Three statistics begin to summarize the current state of higher education in America:$1.5 trillion, $20,000, and 70%. The first—$1.5 trillion—reflects the current balance of outstanding student loans, which have surpassed credit cards and auto loans to become the second largest source of household debt, behind mortgages.2Federal Reserve Bank of New York, New York Fed Consumer Credit Panel 2003–2018. Moreover, the growth of this statistic during the past 15 years, at over 300%, has far eclipsed that of any other household liability. 3 Federal Reserve Bank of New York, New York Fed Consumer Credit Panel 2003–2018, inflation-adjusted using the personal consumption expenditure index Part of the spike in student debt reflects increased enrollment, but rapidly rising tuition at four-year colleges plays a front-and-center role. Tuition at such institutions has more than doubled since the late 1980s, and now stands at approximately $20,000 per year on average.4 Integrated Postsecondary Education Data System data from the National Center for Education Statistics The last statistic—70%—represents the average wage premium recipients of a bachelor’s degree receive over people with only a high school diploma. 5Table 9.1 in Robert G. Valletta, “Recent Flattening in the Higher Education Wage Premium: Polarization, Skill Downgrading, or Both?,” in Education, Skills, and Technical Change: Implications for Future US GDP Growth, ed. Charles R. Hulten and Valerie A. Ramey (Chicago and London: University of Chicago Press, 2019). Although college degrees have always conferred a wage advantage, the average premium stood at only 40% in 1980. Thus, while college costs and debt receive disproportionate attention, a more accurate characterization of the current state of US higher education ought to take into account rising costs, debt, and labor market rewards—at least for the average graduate. Headline statistics tell only part of the story, however, and several key challenges still confront researchers, policy makers, and the public. Perhaps the most critical issue involves determining what forces are most responsible for driving up tuition. At the moment, the relevant research is nascent and researchers have not yet reached a consensus. The extent to which financial aid may be contributing to higher tuition stands out as one of the most important policy quandaries. On one hand, if colleges set tuition independently of students’ ability to pay, then financial aid (in the form of grants, loans, work-study opportunities, or some mix of programs) may be an effective method to expand access to the economic opportunity that college attainment provides. On the other hand, if colleges increase tuition in response to students’ ability to pay, then financial aid may serve more as a subsidy to the institutions than as a bona fide source of assistance to students and their families. Apart from financial aid, others have flagged lagging state support for higher education as another potential culprit behind rising tuition, which, if true, may become more salient over time as states struggle to fund competing priorities such as Medicaid, K-12 education, and infrastructure. Then there’s the issue of the returns to college enrollment and attainment. Although a simple comparison of college graduates those with only a high school diploma reveals a large wage premium for the former, part of this differential is attributable to higher-ability youth self-selecting into college attendance when they could plausibly have received high pay in any number of occupations that do not require a college degree. Whether the observed financial return to college is an indication that colleges provide marketable skills or whether it is just a credential that signals the graduates’ innate ability matters greatly for deciding optimal policy. That is to say, if college degrees serve only as a signal, then efforts to expand enrollment are counterproductive, whereas if college attendance is a form of investment in valuable human capital, positive spillovers may result from expansion. Even under optimistic scenarios for the expected return to college, there is still significant uncertainty about the realized rate of return for different students, owing to the presence of risk both during and after college. At the front end, nearly 40% of students who enroll in four-year institutions do not go on to graduate—they drop out, either of their own volition or because they fail academically. Although college attendance provides some economic benefits even without graduation, there is a sizable “sheepskin effect” whereby wages jump upon receipt of the degree.6See James J. Heckman, Lance J. Lochner, and Petra E. Todd, “Earnings Functions and Rates of Return,” Journal of Human Capital 2, no. 1 (2008): 1–31. Students who succeed in obtaining a bachelor’s degree still experience considerable variation in labor market outcomes based on their choices of institution, major, and occupation—and on the state of the economy at the time of their graduation, over which they obviously have no control. Given these considerable sources of risk, students’ reliance on debt to finance college can be problematic, because it saddles graduates and dropouts alike with a stream of payment obligations regardless of their economic fortunes. Uncovering the consequences of student debt for the economy is an ongoing task for researchers, as is developing potential ideas for reforms. This paper discusses some of the most relevant trends in higher education costs, financing, and returns, as well as research that attempts to explain such trends. It then assesses the likely impact of various reform proposals. Such proposals range from modest changes to the existing policy landscape to more sweeping overhauls aimed at fundamentally altering the structure of the higher education market and the role of government. ### The Rise in College Tuition: Trends and Perspectives Statistics for the average costs and benefits of college gloss over the wide range of experiences of students with different family backgrounds, academic preparation, and choices of institution and major. While$20,000 is the average sticker price tuition across all US four-year colleges, tuition varies tremendously by institution. For example, figure 1 shows that, as of 2018, most students at public colleges attend institutions with published tuition and fees that amount to less than $15,000, whereas a large fraction of students at private colleges are charged more than$50,000.

Figure 1. Distribution of Undergraduate Students by Published Tuition and Fees, 2018.

### 4.2 Credit as a Gateway to Opportunity

With income as a limiting factor to college attendance, the ability to borrow through both the federal student loan program and supplemental private loans acts as a gateway to opportunity for academically prepared students for whom the college investment is most likely to yield a high rate of return. However, the sensitivity of enrollment to loan limits should not be overstated. Indeed, multiple studies find that raising loan limits causes only a modest increase in enrollment because students are reluctant to borrow large sums of money in the face of uncertainty during and after college. For example, one paper concludes that allowing students to borrow up to the full cost of college would only raise college completion rates by 2.4%.71See Matthew T. Johnson, “Borrowing Constraints, College Enrollment, and Delayed Entry,” Journal of Labor Economics 31, no. 4 (2013): 669–725. Another study finds that increases in credit supply caused by banking deregulation from the 1970s to the early 1990s raised college enrollment by 2.6 percentage points.72 See Stephen Teng Sun and Constantine Yannelis, “Credit Constraints and Demand for Higher Education: Evidence from Financial Deregulation,” Review of Economics and Statistics 98, no. 1 (2016): 12–24.

### 4.3 Trends in Student Debt and Default

The data reveal that students have indeed availed themselves of the opportunity to borrow to finance college. As shown in figure 10, inflation-adjusted student debt balances since 2003 have more than doubled in nearly every state, and in parts of the southeast they have nearly quadrupled. During this same period, the student loan delinquency rate has also skyrocketed, often in the same states where student debt has risen by the most.

In the face of these stark statistics, it is tempting to sound the alarm and point to a growing student debt crisis, but some perspective is still in order. While average student debt per borrower now exceeds $20,000, nearly 20% of students graduate with less than$5,000 of debt, and only 9% graduate with more than $75,000 in student loans, as shown in figure 11.73Federal Reserve Bank of New York, 2018 Student Loan Update. Data comes from the Federal Reserve Bank of New York Consumer Credit Panel/Equifax Moreover, perhaps surprisingly, the highest frequency of default occurs among borrowers with the lowest balances, rather than among those who have accumulated the largest amount of debt. One possible explanation for such an unexpected pattern is that the colleges that charge the highest tuition also tend to yield the greatest labor market returns, thereby enabling graduates to service their debt using their higher earnings. To test this hypothesis, one recent study examines administrative data on federal student borrowing matched to earnings data from tax records. The paper concludes that most of the increase in defaults is associated with for-profit schools, two-year colleges, and nonselective institutions. The authors argue that much of the recent rise in student loan defaults can therefore be attributed to changes in the characteristics of borrowers and which schools they attend.74 See Adam Looney and Constantine Yannelis, “A Crisis in Student Loans? How Changes in the Characteristics of Borrowers and in the Institutions They Attended Contributed to Rising Loan Defaults” (Brookings Papers on Economic Activity, 2015). Figure 12 gives a visual representation of the five-year default rate over time and across institution types. Compared to the tuition patterns documented earlier, there appears to be little relationship between tuition and student loan default. In many cases, the most expensive institutions have the lowest default rates, whether because they offer greater value-added or, perhaps more importantly, because they tend to attract the students who are already best situated for success later in life owing to their academic ability and family resources. Either way, the lesson that emerges is that high default rates are endemic to certain classes of institutions and borrowers rather than being a generalized feature of rising student debt levels. Recent evidence indicates that other factors also play a role in driving student loan defaults. For example, one study links the decline in house prices between 2007 and 2010 to the increase in student default rates observed during the same period.75See Holger M. Mueller and Constantine Yannelis, “The Rise in Student Loan Defaults,” Journal of Financial Economics 131 (2019): 1–19. While in principle both phenomena could be driven by the same adverse economic changes, the authors’ analysis attempts to extract causality and finds that the drop in house prices was responsible for between 24% and 32% of the increase in student loan defaults by causing negative employment spillovers. The authors attribute a further 30% of the rise in student defaults to changes in borrower composition. Although much attention has been paid to financial conditions, recent research also demonstrates that strategic motives seem to play a role in student loan defaults, indicating that financial distress is not necessary for default to occur.76 See Constantine Yannelis, “Strategic Default on Student Loans” (working paper, 2016). ### 4.4 The Consequence of Student Debt Increasingly onerous student dent burdens have negative side effects besides simply crowding out consumption or raising the risk of default. Over the past decade, several academic studies have found that student debt alters borrowers’ career choices, impacts their marriage prospects, and can seriously impinge on plans for homeownership or starting a business.77Jesse Rothstein and Cecilia Elena Rouse find that student debt causes students to take higher-salary jobs and reduces the probability that they will take low-paid “public interest” jobs. Rothstein and Rouse, “Constrained after College: Student Loans and Early-Career Occupational Choices,” Journal of Public Economics 95 (2011): 149–63. Dora Gicheva shows that each$10,000 increase in student debt is associated with a 3–4 percentage point drop in the probability of first marriage for men and a 1 percentage point drop for women. Gicheva, “Student Loans or Marriage? A Look at the Highly Educated,” Economics of Education Review 53 (2016): 207–16. Dora Gicheva and Jeffrey Thompson demonstrate that student debt impairs access to financial markets after graduation. For a given amount of education, more student debt is associated with a greater likelihood of declaring bankruptcy. Gicheva and Thompson, “The Effects of Student Loans on Long-Term Household Financial Stability,” in Student Loans and the Dynamics of Debt, ed. Brad Hershbein and Kevin M. Hollenbeck (Kalamazoo, MI: W.E. Upjohn Institute for Employment Research, 2015), 287–316. Holger Sieg and Yu Wang find that student debt has negative effects on marriage prospects, career prospects, and investments in educational quality for female lawyers. Sieg and Wang, “The Impact of Student Debt on Education, Career, Marriage Choices of Female Lawyers,” European Economic Review 109 (2018): 124–47. Vyacheslav Fos, Andres Liberman, and Constantine Yannelis find that each $4,000 increase in debt reduces the probability of graduate school enrollment by 1.3–1.5 percentage points (relative to the 12% mean). Fos, Liberman, and Yannelis, “Debt and Human Capital: Evidence from Student Loans” (working paper, 2017). Alvaro Mezza, Daniel Ringo, Shane Sherlund, and Kamila Sommer show that a$1,000 increase in student loan debt lowers the homeownership rate by 1.8 percentage points for public four-year college-goers during their mid-20s. Mezza et al., “Student Loans and Homeownership,” Journal of Labor Economics (forthcoming). Daniel Cooper and J. Christina Wang reveal that student debt lowers the likelihood of homeownership and that there is also a negative correlation between student debt and wealth accumulation. Cooper and Wang, “Student Loan Debt and Economic Outcomes” (working paper, 2014). Zachary Bleemer, Meta Brown, Donghoon Lee, Katherine Strair, and Wilbert van der Klaauw demonstrate that American youth have accommodated recent increases in tuition by amassing debt rather than forgoing school. Moreover, the increase in student debt can explain 11%–35% of the 8 percentage point decline in homeownership among 28- to 30-year-olds from 2007 to 2015. Bleemer et al., “Echoes of Rising Tuition in Students’ Borrowing, Educational Attainment, and Homeownership in Post-recession America” (NY Fed Staff Report No. 820, Federal Reserve Bank of New York, 2017). Sarena Goodman, Adam Isen, and Constantine Yannelis conclude that an additional $10,000 of student loan and grant dollars implies an estimated 2.4 percentage point increase in homeownership. Goodman, Isen, and Yannelis, “A Day Late and a Dollar Short: Liquidity and Household Formation among Student Borrowers” (working paper, 2018). Karthik Krishnan and Pinshuo Wang discover that graduates from universities that establish no-loans financial aid policies are more likely to start entrepreneurial ventures and that these ventures are more likely to get venture capital backing and to get more venture capital dollars. Krishnan and Wang, “The Impact of Student Debt on High Value Entrepreneurship and Venture Success: Evidence from No-Loans Financial Aid Policies” (working paper, 2018). The overriding picture that emerges is one in which student debt appears to be just as much a force limiting economic opportunity in early adulthood as it is a force that facilitates the acquisition of a college degree that is supposed to open doors. ### 5. The Uncertain Future of Higher Education Although a college degree remains a sound investment for many students toward better future labor market opportunities, decades of rising tuition and student debt have taken their toll on an increasing number of borrowers and created damaging economic side effects. If a trend is unsustainable, the only certainty is that it will not be sustained. Something will change. No sector is immune to disruption, including the higher education sector. A comprehensive analysis of all possible higher education reforms falls outside the scope of this report, but the latest research provides valuable guideposts for policy and suggests directions for additional future study. ### 5.1 Guideposts for Student Loan Reform Student loans facilitate college enrollment, but they result in diminishing returns and may also inflate tuition. The available evidence identifies access to credit as an important factor determining college enrollment, particularly in recent decades as rising tuition has made it more difficult to pay for college out of pocket. Thus, any efforts to pare back student loans are likely to reduce the fraction of college graduates in society. For example, one study finds that completely eliminating student loans would lower college attainment by nearly 9.5 percentage points.78See Brant Abbott et al., “Education Policy and Intergenerational Transfers in Equilibrium,” Journal of Political Economy (forthcoming). Similarly, efforts to impede supplemental private borrowing—even if done in the name of consumer protection—may have the adverse, unintended side effect of reducing college enrollment, particularly for youth from low-income families.79See Lochner and Monge-Naranjo, “Nature of Credit Constraints.” It does not follow, however, that expanding student loans is a sure path toward increased college enrollment. Here, the research is mixed, but multiple studies have found that even allowing students to borrow significantly more than the law currently allows is unlikely to measurably alter college attainment. Instead, it appears that student loans are a close substitute for parental transfers and student employment. More over, policy changes for government-provided student loans impact the private loan market. For example, one recent study finds that increasing government borrowing limits would lead to more default in the private market for student loans.80 See Felicia Ionescu and Nicole Simpson, “Default Risk and Private Student Loans: Implications for Higher Education Policies,” Journal of Economic Dynamics and Control 64 (2016): 119–47. Lastly, several studies have found that colleges “capture” a significant fraction of student aid by raising tuition in response to students’ increased ability to pay, though additional research is needed to reliably and precisely quantify this behavior. Downside risk is a driver of default and a factor limiting college attainment. The evidence also indicates that students may be reluctant to borrow because of fears about their future inability to repay. After all, debt as a financial instrument is simple but blunt. In general, it requires borrowers to repay according to a fixed schedule until the loan terminates. In some cases the payment is fixed, in other cases it fluctuates (if interest rates are variable). In either case, the nature of the contract is such that borrowers are left to manage any economic risks they face in life—such as a deterioration in job prospects—without any adjustment to their payment obligations. In the case of student loans, this problem is compounded by the inherent riskiness of college (given high dropout rates) and the fact that student debt is, except in extreme situations, not dischargeable in bankruptcy. Even though the average college premium is large, there is considerable variation that is not completely predictable at the time of enrollment. Student loan repayment should feature a safety net mechanism. Student debt has always allowed for contingencies of one form or another to prevent undue borrower hardship. In the past, borrowers could more easily discharge their student debt in bankruptcy if they experienced financial distress. However, high default rates in the late 1980s prompted lawmakers to mostly eliminate this option.81For more institutional discussion, see Felicia Ionescu, “Risky Human Capital and Alternative Bankruptcy Regimes for Student Loans,” Journal of Human Capital 5, no. 2 (2011): 153–206. Since then, several different income-based repayment plans have been introduced to protect borrowers who experience the worst labor market outcomes. The earliest iteration was Income-Contingent Repayment, which capped monthly payments at 20% of discretionary income for 25 years and forgave any unpaid balances at the end of that period. Starting in 2009, students were able to enroll in an Income-Based Repayment Plan, which capped payments at 15% of income, and this cap was subsequently lowered to 10%. This same level was enshrined in the subsequent Pay as You Earn Repayment Plan and Revised Pay as You Earn Repayment Plan, with each plan featuring different eligibility criteria and conditions. 82Additional information can be found at the US Department of Education’s Federal Student Aid website: see “Income-Driven Plans,” https:// studentaid.ed.gov/sa/repay-loans/understand/plans/income-driven (accessed November 6, 2019). Importantly, any loan balances that are forgiven are treated as taxable income, which can leave borrowers with hefty tax bills. Currently, only 29% of borrowers are enrolled in some form of income-driven repayment scheme. 83 See figure 13A in College Board, Trends in Student Aid 2018, Trends in Higher Education Series, 2018, 20. In the case of borrowers with high earnings, enrollment may provide little benefit. However, recent research indicates that uptake is low in part because of a lack of awareness and because the current framing of these policies discourages borrowers from participating. 84See Katharine G. Abraham et al., “Framing Effects, Earnings Expectations, and the Design of Student Loan Repayment Schemes” (NBER Working Paper No. 24484, National Bureau of Economic Research, Cambridge, MA, 2018). Regardless of the cause of this low uptake, research indicates that increasing the availability of income-driven loan repayment would provide some benefits both before and after college. On the front end, the availability of plans with income-contingent repayment obligations combined with forgiveness of any unpaid balances at the end of the loan duration would likely increase college enrollment and completion rates.85See Ionescu, “Risky Human Capital.” Another study finds that the expansion of income-based repayment during the Great Recession reduced student loan defaults and made borrowers less sensitive to house price fluctuations.86See Mueller and Yannelis, “Rise in Student Loan Defaults.” Besides enhancing college attainment, insurance against downside risk may provide additional labor market benefits. For example, recent work finds that mitigating earnings uncertainty could lead to higher occupational mobility.87 See German Cubas and Pedro Silos, “Social Insurance and Occupational Mobility” (working paper, 2019). In particular, income-based repayment would allow workers to enter occupations that have either lower or more uncertain starting pay but that may feature steeper earnings gradients over time. In addition to income-based repayment, which provides insurance against low post-graduation earnings, some scholars have suggested adding automatic loan forgiveness in the event that a student does not complete college. To the extent that dropping out is a totally unforeseen event, such an arrangement could be an effective way to provide insurance against dropout risk. According to one study, loan forgiveness only in the event of involuntary dropout (i.e., failing out of school) could be provided at little cost, even taking into account that it could provide an incentive to some students to purposely shirk and do poorly such that they drop out, just to avoid future student loan payments.88See Satyajit Chatterjee and Felicia Ionescu, “Insuring Student Loans against the Financial Risk of Failing to Complete College,” Quantitative Economics 3 (2012): 393–420. However, the “insurance premium” of covering such loan forgiveness would become far more expensive for successful students if the program were expanded to include voluntary instances of dropping out (i.e., when students in good standing leave). Policy makers should distinguish insurance from redistribution. Expanding income-based repayment entails several potential drawbacks as well as advantages, particularly if policy makers view it as a way to expand subsidies or pursue other social policy in an opaque manner. Deliberate underpricing of loans encourages excessive debt and an inefficient allocation of resources, all without the transparency of explicit grant programs. For this reason, income-based repayment should not be seen as a vehicle for loan forgiveness, but rather as a modified financial arrangement whereby borrowers should fully repay their loan on average. In other words, ideally the cost of the insurance would be priced properly into loans. However, the nature of the pricing would likely be the subject of considerable political debate, just as it is for health insurance. On one end of the spectrum, all students could be placed into the same risk pool and charged a common interest rate that covers the pool-wide risk of ex post underpayment from borrowers who end up with low earnings. In this scenario, students likely to receive the largest post-graduation labor market returns to college (for example, engineering majors coming from selective institutions) would enter college knowing that their student loan borrowing would contain an implicit subsidy for other students who are likely to receive lower post-graduation earnings. On the other end of the spectrum, students could undergo individual risk assessments based on their academic preparedness (e.g., college admissions test scores or high school GPA), the institution they attend, and their choice of major. This model would be similar to the income share agreements that were first proposed by Milton Friedman decades ago.89 See Beckie Supiano, “A Closer Look at Income-Based Repayment, the Centerpiece of Donald Trump’s Unexpected Higher-Ed Speech,” Chronicle of Higher Education, October 15, 2016, https://www.chronicle.com/article/A-Closer-Look-at-Income-Based/238085. Each arrangement has advantages and disadvantages, but moral hazard emerges as a common concern. By acting as an implicit marginal tax—that is, the more people earn, the more they pay—income-contingent loans could discourage post-college human capital accumulation or labor effort. Total loan forgiveness would be expensive, untargeted, and distortionary. With student loan defaults reaching troubling levels, proposals for unconditional loan forgiveness have begun to gather steam. While it is difficult to precisely estimate the effects of such sweeping plans, researchers have recently provided some insights by looking at how borrowers responded to an unexpected discharging of their debt following court decisions finding that National Collegiate Student Loan Trust—the largest private holder of student debt in the US—could not properly prove chain of title. In other words, the researchers have been able to compare outcomes for borrowers with similar levels of debt and income who differed only in whether they were lucky enough to have had their debt forgiven as a result of a court ruling. The researchers find that borrowers experiencing debt relief subsequently sought to repair other dimensions of their balance sheets and creditworthiness by deleveraging. As a result, they were less likely to default on other accounts. In addition, such borrowers exhibited increased geographic and job mobility, along with a significant rise in income—equal to two month’s salary over a three year period. 90 See Marco Di Maggio, Ankit Kalda, and Vincent Yao, “Second Chance: Life without Student Debt” (NBER Working Paper No. 25810, National Bureau of Economic Research, Cambridge, MA, 2019). In short, debt relief provided significant benefits. However, such “accidental” loan forgiveness does not provide a proper accounting of what an overt nationwide policy of blanket forgiveness would entail. Most importantly, it ignores costs. Whereas shareholders of National Collegiate have presumably borne much of the cost of the lost revenue from discharged debt, taxpayers would be on the hook for any broad-based loan forgiveness. After all, current forms of debt forgiveness on a much smaller scale already impose significant costs on the government.91For an evaluation of the costs of income consolidation, see Felicia Ionescu, “The Federal Student Loan Program: Quantitative Implications for College Enrollment and Default Rates,” Review of Economic Dynamics 12, no. 1 (2009): 205–31. Besides being expensive, loan forgiveness would also be poorly targeted if the primary goal is to alleviate hardship and reduce default. The evidence discussed throughout this report paints a picture of a growing student loan repayment crisis for certain borrowers rather than a broad-based student debt crisis afflicting anybody with a student loan. Rather than targeting for assistance borrowers who are either already delinquent or likely to be on the verge of default, loan forgiveness would provide a large subsidy to a group of borrowers that would include people with moderately high earnings and little trouble repaying their obligations. Data from the Survey of Consumer Finances show that households in the top 25% by income held 47% of outstanding student debt, compared to only 11% among the bottom 25% of households.92 See figure 19A in College Board, Trends in Student Aid 2015. Furthermore, data from the 2012 National Postsecondary Student Aid Study show that families with incomes over$106,000 were nearly as likely to have large cumulative debt balances as families with more modest incomes. Furthermore, nearly 8% of families with incomes above $106,000 took out private loans, compared to only 7% of families with incomes between$30,000 and $65,000.93 See figures 2014_14B and 2013_9C in College Board, Trends in Student Aid 2018 (Excel file). In addition, data from the Federal Reserve Bank of New York reveal that the student loan default rate for borrowers with above-average income is in some cases only moderately higher than the rate for lower-income borrowers.94See Rajashri Chakrabarti et al., “Who Is More Likely to Default on Student Loans?,” Liberty Street Economics (Federal Reserve Bank of New York), November 20, 2017, https://libertystreeteconomics.newyorkfed.org/2017/11/who-is-more-likely-to-default-on-student-loans.html. Besides being untargeted, unconditional loan forgiveness would likely create unintended negative consequences for taxpayers, future borrowers, and the broader economy. Taxpayers would be affected because students may anticipate that the loan forgiveness is not a one-time event but rather a policy that could be implemented from time to time in the future. In light of evidence that strategic motives already account for some student loan default behavior, borrowers would have an incentive to purposely defer repayment and even become delinquent in the hope that their debt will be wiped out in the future. After all, recent research indicates that the reintroduction of bankruptcy protection—a far less sweeping proposal than unconditional loan forgiveness—would increase student loan default by 18%, and that eliminating wage garnishment would increase default by 50%.95See Yannelis, “Strategic Default on Student Loans.” Future borrowers would be harmed if, in the process of implementing loan forgiveness, the government imposed haircuts on private lenders (i.e., by not compensating them completely for their losses). In that case lenders would inevitably price the risk of future debt wipeouts into new loans, thus increasing the cost of credit for future borrowers. The broader economy would be affected by any attempt to phase out the benefits of loan forgiveness for those with higher incomes: this would create an implicit tax on earnings on top of the explicit taxes that people already pay. For example, phasing out loan forgiveness by$1 for every $3 in income above$100,000 would increase the marginal tax rate for households with incomes over $100,000 by more than 33 percentage points. Policy makers shouldn’t ignore Parent PLUS Loans. While much attention has been paid to the borrowing of students, parents also play a pivotal role in financing college, both through direct transfers to their children and through their own borrowing. In fact, since the first few years after 2000, Parent PLUS borrowing has been one of the fastest-growing sources of debt to finance higher education. Back in 2002, the average annual borrowing amount was$11,750 for such loans, and it has since increased by nearly \$5,000 in inflation-adjusted terms.96See figure 9B in College Board, Trends in Student Aid 2018, 17 Parent PLUS Loans present several policy challenges. First, as with any source of financial aid, they may be contributing to tuition inflation. Second, because parents are not the recipients of their child’s college degree and the higher earnings it confers, they may lack the means to repay the debt. This risk is compounded by the fact that Parent PLUS Loans are ineligible for most of the income-based repayment programs available to students. However, forgiveness of Parent PLUS Loans would create the same unintended consequences as would forgiveness of student loans. Thus, future research is needed to examine different policy options, such as altering Parent PLUS eligibility, altering loan limits, or adjusting other forms of financial assistance to families such that parents do not engage in as much borrowing of their own.

Information is important. There are many additional reforms that have been suggested. At the borrowing stage, the government has in the past instituted changes to the expected family contribution formula that determines eligibility for grants and subsidized loans. For example, as of 1992, home equity on the principal residence is no longer included in the asset calculation. Research indicates, however, that this particular change has had little impact on enrollment.97See Ionescu, “Federal Student Loan Program.” Other scholars have highlighted the complexity of the FAFSA form and pointed out that a similar distribution of federal aid could be achieved by replacing the FAFSA with information from IRS Form 1040, which taxpayers are already required to fill out.98See Susan Dynarski and Mark Wiederspan, “Student Aid Simplification: Looking Back and Looking Ahead,” National Tax Journal 65, no. 1 (2012): 211–34. By providing detailed income and asset information to colleges—thus providing colleges with significant knowledge about students’ ability to pay—the current financial aid system gives colleges significant pricing power that is enjoyed by almost no other sector in the economy. One recent study evaluates the impact of restricting the provision of FAFSA information to colleges and concludes that, while a fraction of students attending elite colleges would be priced out of that market, the majority of students would benefit from lower prices. While colleges like to claim that the existing information available to them facilitates the provision of need-based aid, this study finds that 30% of low-income students are actually harmed by the status quo.99See Ian Fillmore, “Price Discrimination and Public Policy in the U.S. College Market” (working paper, 2018).

### 5.2 Other Lessons for Higher Education Reform

One-size-fits-all pricing has limits. Internally through institutional aid, colleges already differentiate prices from student to student based on academic ability (merit) and family income (need). However, some institutions face external constraints on their ability to provide more generous targeted assistance, such as state-imposed caps on sticker price tuition, even though the sticker price is a highly misleading measure for the price paid by most students. Absent such constraints, colleges could raise their sticker price, charge only a subset of students the higher price (e.g., high-income, low-ability students), and offer more generous aid to low-income, high-ability students. Research indicates that such behavior indeed occurred following tuition deregulation in Texas in 2003 and that “deregulation did not reduce poor students’ outcomes.”100Rodney J. Andrews and Kevin Stange, “Price Regulation, Price Discrimination, and Equality of Opportunity in Higher Education: Evidence from Texas” American Economic Journal: Economic Policy, 11(4): 31 – 65 (citation on page 32). In addition, in-state tuition caps create an incentive for public colleges to substitute for the admission of state residents the admission of out-of-state students, who can be charged higher prices.

Along a different margin, colleges could also charge differential tuition based on a student’s choice of major. Although such a practice would diverge from the liberal arts ethos that has guided much of American higher education, it would reflect the fact that vast cost differences emerge between majors. In the current system, for example, English majors effectively subsidize engineering majors, even though the latter are likely to receive far higher salaries after graduation. While such a shift could dissuade low-income students from enrolling in high-cost majors, financial aid could also be made major-specific. The extensive cross-subsidization across majors may partly explain the research finding that students place greater weight on their tastes than on monetary returns when making decisions about their college major.

The reality of free college would fall short of the promise. For decades, complete subsidization of K–12 education by the government has guaranteed access to elementary and secondary education at no out-of-pocket expense. With college potentially taking the place of high school as the gateway to economic opportunity, some have called for similarly making college free at the point of service through much-more-generous public subsidies. Before addressing such a dramatic move, it is useful to consult existing research on the impact and cost effectiveness of current higher education subsidies as well as on the likely effects of marginal expansions in funding. Several research papers conclude that increased subsidies are unlikely to be cost effective or to achieve the goal of directing assistance toward low-income families. For example, one study finds that a higher subsidy rate would lead to substantial increases in college investment among youth from wealthy families but to only modest effects among financially constrained youth.101 See Lochner and Monge-Naranjo, “Nature of Credit Constraints.” As a result, a universal subsidy would actually amplify earnings inequality. Other work finds that increasing college subsidies is much less effective than raising expenditures for early education, largely because college subsidies come too late in a child’s life to provide significant benefits for children from low-income families. 102See Diego Restuccia and Carlos Urrutia, “Intergenerational Persistence of Earnings: The Role of Early and College Education,” American Economic Review 94, no. 5 (2004): 1354–78.