Discussion: (0 comments)
There are no comments available.
| American Enterprise Institute
View related content: Higher Education
Student loan debt in the United States now stands at more than $1 trillion, surpassing the nation’s collective credit card debt and second only to the nation’s mortgage debt.1 As student debt has grown, so has the number of graduates across the nation struggling to manage it.2
According to the most recent federal National Postsecondary Student Aid Study (NPSAS), nearly 70 percent of students who graduated with a bachelor’s degree in 2012 borrowed money to finance their education. And their average debt was $29,400, which is more than the average yearly earnings of young adults with bachelor’s degrees.3 According to the same study, the average monthly payment to service that debt was more than $300, a substantial burden for many recent graduates.
Just how burdensome student debt is can be judged only in relation to how much a college graduate earns. A graduate earning $60,000 per year with $30,000 in loans is far better off than a graduate earning $25,000 per year who also owes $30,000. Unfortunately, the wage and debt data needed to calculate the potential burden of student loans are scarce. And when the data are available, they are not reported in a manner that will help students make wise decisions about how to invest their time and money in higher education.
Although several pieces of legislation pending in Congress may improve the flow of student information about earnings and debt levels, some states are taking the lead in making the relevant information available.4 Among the best data we have to calculate the debt burden shouldered by graduates are from Texas, which reports average earnings and debt levels for graduates from every program of study in the state’s public colleges and universities. Moreover, Texas has a strong commitment to transparency and is one of the lead partners with College Measures in putting data into the public sphere in a way that is useful and understandable.5
Data on both earnings and debt can help students make better-informed choices as they choose colleges and majors. Besides providing valuable consumer information, these data are also used in the federal government’s effort to monitor the “gainful employment” of graduates. Currently, the federal government’s gainful employment regulations are limited to career-oriented programs, but students in all sectors of higher education (proprietary for-profit, private not-for-profit, and public) are borrowing more to pay for their education.6
In this paper, I use earnings and debt data from graduates from more than 500 programs in public universities to extend the federal government’s gainful employment efforts to public institutions. Although the Texas data I use are not exactly comparable to the data the federal government has proposed, I believe that the results show that extending gainful employment efforts from just career-oriented programs to all programs is justified.
Federal Gainful Employment Standards
The United States Department of Education (ED) has spent the last several years trying to promulgate regulations pertaining to the gainful employment of career education graduates. This effort relies heavily on debt-to-earnings ratios to determine the extent to which these graduates are making enough money to pay off their debts-that is, whether they are “gainfully employed.” Under proposed federal regulations, programs whose graduates are not earning enough income relative to their debt levels could lose their eligibility to receive federal student aid, which will almost certainly lead to the programs’ closure.
The focus on debt-to-earnings ratios is understandable: as student debt has accumulated and as the number of students having trouble repaying their debt has grown, the question of which schools and programs should be eligible to participate in federal student aid programs has become more important.
ED is authorized by the Higher Education Act to regulate “career-oriented” programs. These are programs authorized by law to participate in the federal student aid system because they lead to “gainful employment in a recognized occupation.” Most of these programs are offered by proprietary colleges and universities, although a number of nondegree community college programs are also covered.7
ED’s efforts to create gainful employment regulations were initially unsuccessful. The first round of rules was put on hold after a federal district court found one of the accountability measures it proposed, the repayment rate, to be arbitrary.8 Facing this legal challenge, in 2013, ED launched a new round of negotiated rulemaking on gainful employment.
On March 14, 2014, after these negotiations ended, ED released new proposed rules.9 Under these, the earnings of graduates from career-oriented programs who have received federal grants or loans through Title IV of the Higher Education Act will be calculated at the program level, and the eligibility of these programs to continue to provide these federal monies will be judged against several criteria. Two of these are based on debt-to-earnings ratios.10 The first ratio compares annual debt service to annual income. The second is the ratio of annual debt service to discretionary income, where discretionary income is calculated as annual income minus 150 percent of the national poverty level (in 2013, this was $11,480).
Programs would be put into different categories depending on the relationship between how much their graduates borrowed and how much they went on to earn. Programs pass if their graduates have an annual debt-to-earnings (aD/E) ratio of less than or equal to 8 percent or a discretionary debt-to-earnings ratio (dD/E) of less than or equal to 20 percent. Programs fail if their annual ratio is more than 12 percent and their discretionary ratio is more than 30 percent. ED also created a warning zone into which programs fall if their aD/E was between 8 and 12 percent or their dD/E was between 20 and 30 percent.
This classification will have consequences once the proposed rules go into effect: a program would be ineligible to provide Title IV aid to its students for three years if it
Given existing law, the Department of Education focused specifically on career-oriented programs, most of which are offered by proprietary colleges. The data show wide variation in the debt that graduates from different programs accrue. The resulting variation in the burden graduates will carry in servicing that debt demonstrates why debt and earnings data should be made available to all students, whether their school is public, proprietary, or not-for-profit. These data also suggest that the gainful employment standards, if applied to more programs, including those public institutions offer, might put a number of programs at risk of adverse consequences.
Calculating Gainful Employment Ratios
Although the gainful employment regulations developed by ED are complex and cannot be replicated exactly using the Texas data, I developed an approach that aligns with the spirit of the federal gainful employment regulations and adheres to the criteria that ED uses in the proposed regulation.
I use program-level data to compute the two ratios reflecting how much students would need to pay to service their loan debt relative to their postgraduation earnings.11 In my ratios, the numerator is the average annual debt service of graduates. In my calculations, I follow ED guidelines and use a 15-year repayment period and an interest rate of 5.42 percent. I compute the two debt-to-earnings ratios using average graduate wages by program for one calendar year (2012) following graduation.12 Wage data are provided by the Texas unemployment insurance (UI) system. Only students who graduated during the 2011 financial year (Fall 2010-Summer 2011) are included.
Because the ED and Texas parameters differ, the ratios I compute are not identical to those ED has proposed (and my data include graduates of public institutions and liberal arts and science programs, while ED’s are only from career-oriented programs).13
Keeping that in mind, I have calculated the two debt-to-earnings ratios and have used the “cut points” that ED has proposed for classifying programs as passing, failing, or falling in the warning zone. Although I do not calculate these ratios over multiple years as ED has proposed, the data from Texas suggest that many programs in public institutions-if subject to the same gainful employment standards as career-oriented ones-would have a hard time meeting ED’s standards.
National data also suggest the debt-to-earnings ratio, if applied more broadly, could put a large number of programs in jeopardy. At ED’s proposed interest rate of 5.42 percent and using the nationwide average loan debt of $29,400, falling below the 8 percent annual debt/earnings cutoff for the proposed zone requires a salary of just over $32,000. But nationwide, according to the Census Bureau in 2012, on average, 18-24-year-olds with bachelors’ degrees earned significantly less than that-just $26,100.14 Similarly, Degrees of Debt, a recent study by the National Center for Education Statistics (NCES), shows that the overall debt-to-earnings ratio nationwide is 13 percent, and almost one-third of all graduates had debt-to-earnings ratios higher than 12 percent in the first year after graduation.15 In short, both the Census and NCES data suggest that many graduates, not just those from career-oriented programs, are likely encountering problems servicing their loans.
Although these reports are based on averages across the nation, the federal gainful employment regulations will apply to specific programs. Since Texas reports earnings and debt levels by program, we can calculate the two debt-to-earnings ratios for 520 individual programs in public universities that have the data needed to calculate these ratios and that meet ED’s minimum program size of 30 students.
Who Is Included in the Ratios?
One contentious issue that arose in the negotiated rule-making panel charged with considering gainful employment was who should be included in the calculations of debt-to-earnings ratios. Given the government’s concern for protecting its large outlays of Title IV student aid, in ED’s proposed regulations the debt-to-earnings metrics are calculated using the median annual loan payment of students in a program who received Title IV funds. Some of these students may have loans, but others may have had just grants (for example, Pell grants) and no loans. In ED’s proposal, these students would be included in calculations of loan amounts, with their loans set to zero. Adding these students reduces the median loan amount for the program, making it easier to pass the thresholds.
Not all students free of federal loans are Title IV recipients, and excluding them may lower program success rates. For example, given that Title IV is means-tested, students without federal loans are likely to be more affluent than Title IV recipients. Since more affluent students on average are likely to be more successful in the job market than Title IV students, excluding them could reduce the overall earnings of graduates, suppressing measures of the labor market success of program graduates.
We need to be aware of the biases of each approach to which students are included. Because the Texas database used in this analysis does not have student-level data, we also need to be aware of our limits in calculating the debt-to-earnings ratios.
In table 1, I estimate the debt-to-earnings ratios for students with and without loans. Texas reports only the average debt of students who have borrowed, so I adjust this number to reflect the experience of all program graduates. (This mirrors ED’s approach of including zeroes in the estimate of program-level debt.) To do this, I multiply the loan amount by the percentage of students who have taken out loans.16
The results using this adjustment are presented in the first line of table 1. As shown, 6 percent of Texas public programs fail outright-both of their debt-to-earnings ratios exceed the threshold-and 22 percent of Texas programs fall in the warning zone. In short, applying the proposed standards to public programs, and using first-year wages, more than one-quarter of Texas bachelor’s programs would be at risk.
Table 1 also presents the results when the loan amount is not discounted by the percent of students who borrowed (approximating the experience of those students with loans). Since the loan amounts are higher, more programs are at risk. These results show that 19 percent of programs would fail and an additional 35 percent would fall in the zone, for a total of 54 percent of all programs being at risk.17
Some of the at-risk programs under either approach are liberal arts programs such as anthropology, English, history, philosophy, and political science. But some of the programs that would be at risk are outside the liberal arts-with biology as the most commonly found at-risk program. Appendix 2 lists the number of programs of study that either failed or are in the zone.
Same Major, Different Debt Burdens
Table 2 presents summary data for the two debt-to–earnings ratios and the underlying data for graduates from four undergraduate programs of study with large numbers of students: biology, English, history, and psychology. These tables document wide variation in the experiences of graduates with the same major from different universities in Texas.
With a median debt level of more than $46,000 and median first-year earnings of just half that, the median biology aD/E and dD/E ratios across state programs are above the proposed cut scores. Indeed, about 90 percent of Texas biology programs would be at risk-and most of those would have been judged as failures.
One reason for this high failure rate-and an important lesson for the development and application of gainful employment measures overall-is that for some fields of study, the time at which earnings are calculated matters greatly. Biology is the clearest case in point: because many biology graduates go on to medical school, short-term earnings are low, and therefore the debt-to-earnings ratios computed using early postgraduation earnings are high. But since Texas does have longer-term data, it is possible to calculate debt-to-earnings ratios using earnings later in the careers of these graduates.18 Not surprisingly, 10 years postgraduation, many of these biology students have medical degrees and high earnings, driving down their debt-to-earnings ratios. I should note that biology is the most notable example of radical change over time--graduates from most other programs with high debt-to-earnings ratios experience far slower growth in earnings over the decade after graduation, and graduates from many of these programs will struggle to manage their debt for a decade or more.
The maximum and minimum discretionary debt-to-earnings ratios reported in table 2 suggest another potential problem with ED’s metrics. Recall that the discretionary measure subtracts $17,220 from the annual income.19 But mathematically, if the observed annual earnings of graduates from any program are close to this amount, then the resultant ratio will be very large (because the denominator will be so small). In biology, the dD/E ranges from around -300 percent to around 450 percent.
The percentage of programs at risk in the other three disciplines is lower, but in each case, it is still more than half. This is clear when we look at the median debt-to-earnings ratios.
Consider English programs, where the median aD/E (8.5 percent) and dD/E (21.3 percent) ratios are above ED’s cut scores. The range in the annual debt-to-earnings ratio is from around 5 percent to 13 percent, while the discretionary debt-to-earnings ratio varies by a factor of 10, showing just how wide the variation is in the debt and earnings of students graduating from different programs across the state.
Psychology is one of the most popular majors across the nation’s colleges. More than three-quarters of psychology programs in Texas would be at risk and, like biology and English, the median program values for both debt-to–earnings ratios are above the suggested cut points. Average loan amounts vary from around $15,000 to more than $38,000, while earnings are more tightly clustered-ranging from around $21,000 to $28,000. In turn, the debt-to-earnings ratios vary widely.
Surprisingly, graduates with history bachelor’s degrees fare best on the debt-to-earnings ratios. The median scores on both ratios are below ED’s cut scores, and the maximum ratios are lower than any of the other programs reported in table 2. But even here, more than 60 percent of the programs are at risk. Moreover, the growth in wages for graduates of history programs is below the average for all bachelor’s graduates in the state, so history graduates will likely have a hard time servicing their debt for many years to come.
Student Debt Is a Problem
According to the New York Federal Reserve Bank, as of the fourth quarter of 2013, more than 11 percent of student loans were at least 90 days behind in payments.20 On top of that, nearly half of outstanding student loans do not currently require any payment, because the student is either still in school or has taken advantage of other ways to defer payment. But, sooner or later, these loans will be due and many graduates will fall behind.
Although studies such as this are based on national samples, gainful employment is focused on program-level results. The data I present illustrate that student debt-and the burden of servicing that debt-varies across programs of study and across institutions. The data also suggest that this variation is found among graduates from public institutions, not just the career-oriented ones on which the federal government currently focuses.
Many efforts to improve the measurement of the success of college graduates are torn between providing better measures of outcomes to help inform student choice and using the power of government to regulate institutions and programs by penalizing programs that fail to meet benchmarks. Gainful employment is not immune to this tension.
Students have traditionally been bombarded with the message that higher education is a good investment. Although this continues to be true overall, the rate of return to students varies widely with choice of majors and of colleges. If higher education pays, it pays a lot more for some graduates than for others.21
In turn, as one of the pending pieces of federal legislation aptly puts it, students need to “know before they owe.” Unfortunately, the federal government has been slow in getting this information into the hands of students so that they can make better decisions. Efforts such as the College Scorecard and the proposed Post-Secondary Institutional Ratings System will focus on institution-level data, but much of the variation in student success is at the program level.
The federal effort to regulate programs based on measures of student labor market success has also run into repeated roadblocks and measurement problems. ED’s first attempt at creating regulations surrounding gainful employment was derailed by a court decision challenging one of its key measures. ED’s latest attempt to regulate career-oriented programs produced an 841-page document that will likely be challenged in the courts. Measurement issues abound in the proposed rules. I have noted several in my analysis here.
But perhaps most clear is that the Higher Education Act’s focus on career-oriented programs is insufficient. As the Texas data show, many students graduating from public universities are leaving with levels of debt that will be hard for them to pay off. In some cases, the overhang of servicing that debt will last for a decade or more, impeding their ability to marry, have children, buy a house or a car, or otherwise launch life as an adult. And because more students attend public than for-profit institutions, far more students in public universities will be affected by adverse debt-to-earnings ratios.
Finally, even though the data I use in this report do not allow me to precisely replicate ED’s approach, they point to a high likelihood that many public programs would be at risk under the gainful employment measures if these regulations were extended to them. This may be just the tip of the iceberg. Remember, these data are from relatively low-cost public universities. The cost of attending private not-for-profit colleges is even higher.
Although most states do not at present collect data on the wages of graduates from these private schools, a few states that College Measures has partnered with do. There is no evidence that the far greater cost of attending these private schools, on average, is associated with higher wages.22 In turn, the debt-to-earnings ratios for programs in these schools may be even higher than for public institutions.
The bottom line: all students and their families deserve the kind of information that the federal government has been releasing for career-oriented programs so that they can make informed decisions about their futures.
Appendix 1: Technical Appendix
The data available for this paper were aggregated by program and by institution. Average earnings are shown for students who worked at least three quarters during the first calendar year after graduation (2012).
Only students who graduated with a bachelor’s degree during the 2011 fiscal year (Fall 2010, Spring 2011, or Summer 2011) are included. The following student records were not eligible for inclusion: students who did not complete an undergraduate program, students who earned a higher degree (beyond the baccalaureate) from a Texas public or independent institution, and students who were employed in a state other than Texas.
The program is defined based on Classification of Instructional Programs (CIP) four-digit codes. For program information, see www.txhighereddata.org/Interactive/CIP/.
1. About 60 percent of this total is owed by undergraduates, and the other 40 percent is owed by graduate and professional students. See Jason Delisle, “The Graduate Student Debt Review,” New America Foundation, March 25, 2014, www.newamerica.net/publications/policy/the_graduate_student_debt_review.
2. See, for example, Floyd Norris, “The Hefty Yoke of Student Loan Debt,” New York Times, February 20, 2014, www.nytimes.com/2014/02/20/business/economy/the-hefty-yoke-of-student-loan-debt.html.
3. The Census Bureau reports that in 2012, 18-24-year-olds with bachelors’ degrees earned an average of $26,100. See US Census Bureau, “2012 Person Income Table of Contents,” Current Population Survey, www.census.gov/hhes/www/cpstables/032013/perinc/pinc04_000.htm.
4. Among the pieces of legislation filed were the Student Loan Borrower Bill of Rights (S.1803), sponsored by Sen. Richard Durbin (D-IL); the Know Before You Owe Private Student Loan Act of 2012 (HR.6273), sponsored by Rep. Jared Polis (D-CO); and the Understanding the True Cost of College Act of 2013 (S.1156), sponsored by Sen. Al Franken (D-MN).
5. I am the president of College Measures, a startup company whose mission is to help state agencies in their efforts to make information about the earnings of graduates from their higher education programs publicly accessible. See www.collegemeasures.org/esm. Also see www.myfuturetx.com
6. Proprietary for-profit institutions include not only the large publicly traded systems such as the University of Phoenix and Strayer but also much smaller trade schools such as the Fortis College of Cosmetology in Mobile, Alabama. Private not-for-profits similarly encompass well-known campuses such as Harvard and Stanford and lesser-known ones like the Academy of Chinese Culture and Health Sciences in Oakland, California.
7. Given the current ban on a student-level data system, ED has no easy way of gathering and reporting earnings data for students other than those in career-oriented programs. Congress could amend the Higher Education Act to gather this information for more students. See the discussion of student unit records in Clare McCann and Amy Laitinen, College Blackout: How the Higher Education Lobby Fought to Keep Students in the Dark, New America Education Policy Program, March 2014, http://education.newamerica.net/sites/newamerica.net/files/policydocs/CollegeBlackoutFINAL.pdf.
8. See US District Court for the District of Columbia, “Memorandum Opinion, Association of Private Colleges and Universities v. Arne Duncan and US Department of Education,” Civil Action 11-1314, June 30, 2012, https://ecf.dcd.uscourts.gov/cgi-bin/show_
9. US Department of Education, “Negotiated Rulemaking 2013-2014: Gainful Employment,” http://www2.ed.gov/policy/highered/reg/hearulemaking/2012/gainfulemployment.html.
10. A third measure is a program-level three-year cohort default rate, which must be less than 30 percent to pass.
11. There is a divergence between the debt I use and what ED uses. Most notably, ED caps the debt level used in its calculation to the maximum of tuition plus fees, excluding debt that students incurred for things such as living expenses. The rationale behind this choice was that some students borrow excessively and colleges should not be responsible for this excess borrowing. However, students will have to pay off the totality of their school loans, not just the part covering their tuition and fees. Moreover, Texas debt data also includes any debt incurred for higher education in Texas. The federal proposal limits its program-level accountability formula to debt accrued at each institution for each program.
12. Average earnings are for students who worked at least three quarters during the first calendar year (2012) after graduation. See the technical appendix for more information on calculations and definitions.
13. Appendix 1 gives more technical details about the construction of the cohort and data limitations.
14. According to NPSAS, the average monthly payment for student debt is $312. Given average annual earnings of $26,100, young college graduates on average will spend 14 percent of their earnings on debt service-above the 12 percent proposed by ED to define failure.
15. Jennie H. Woo and Matthew Soldner, “Degrees of Debt: Student Borrowing and Loan Repayment of Bachelor’s Degree Recipients 1 Year after Graduating-1994, 2001, and 2009,” US Department of Education Stats in Brief (October 2013), http://nces.ed.gov/pubs2014/2014011.pdf.
16. For example, say Texas reports that 75 percent of graduates from a particular biology program have taken loans, and the average loan is $49,429. I multiply this loan amount by 75 percent and use $37,072 (the resulting number) in calculating the debt-to-earnings ratios. This is not identical to the procedure that ED has proposed-since my procedure includes both Title IV grant recipients with no loans and students who are have not received any Title IV monies. Clearly, having individual student-level data and more detailed information about the Title IV status of each student would lead to a more faithful replication of ED’s proposed regulations. However, my estimate is the most accurate picture of an entire program that existing data can provide. I should note that even ED has a hard time generating enough information to calculate rates exactly conforming to the regulation’s terms. See ED’s 2012 Informational Rates and their explanation, released with the proposed regulation.
17. Since the earnings data includes all students (regardless of Title IV status), to the extent that non-Title IV students earn more, the ratios are higher than they might be for just Title IV students.
18. Longer-term wage data are available at www.myfuturetx.com/.
19. This is 150 percent of the poverty line for individuals in 2013.
20. Federal Reserve Bank of New York, “Household Credit,” www.newyorkfed.org/regional/householdcredit.html.
21. See Mark Schneider, Higher Education Pays: But a Lot More for Some Graduates Than for Others, American Institutes for Research, September 3, 2013, www.air.org/resource/higher-education-pays-lot-more-some-graduates-others.
22. See the especially rich College Measures data from Virginia at http://esm.collegemeasures.org/esm/virginia/.
There are no comments available.
1150 17th Street, N.W. Washington, D.C. 20036
© 2015 American Enterprise Institute for Public Policy Research