Gainful Employment's Quantum Mechanics Problem

In physics, rules that apply to larger bodies fall apart when moving down to nanoscopic levels. The gainful employment framework suffers from a similar quantum mechanics issue. If left unfixed, it could make it easier for programs with extremely low earnings returns to pass than programs with equivalent percentage results but better incomes. The result could be a significant number of programs deemed to be providing gainful employment in a recognized occupation while the federal government and taxpayers forgive every last dollar lent plus interest.

The low-income problem arises due to the interaction between the equal treatment of the two debt-to-earnings rates and the presence of a program cohort default rate instead of a repayment rate. The result is that a program with earnings so low that it has no discretionary income has an easier path to passing the metrics by keeping its annual debt-to-earnings rate within acceptable bounds and sending all borrowers into income-based repayment.

The proposed rule requires a program to only pass one of the two debt-to-earnings tests. But allowing programs to pass solely based on the annual debt-to-earnings measure makes it possible for a program with sub-poverty wages to still avoid failing the metrics. That's true even if passing means graduates have to contribute dollars needed for necessities toward loan payments. Take for example the medical secretary certificate offered by Everest College in Portland, which passes the annual debt-to-earnings rate at 7.75 percent. But graduates from this program earned just $14,595—less than the annual income for a worker who earns minimum wage. These graduates are effectively impoverished with little if any discretionary income and a failing discretionary debt-to-earnings rate.

The notion that it’s acceptable for poverty level workers to be making payments on their student loans actually runs counter to both the way Congress thinks about loan debt as well as other researchers. For example, Sandy Baum and Saul Schwartz in “How Much Debt is Too Much: Defining Benchmarks for Manageable Student Debt,” conclude:

Borrowers with pre-tax incomes less than half the median ($37,543 for full-time workers 25 and older in 2004 in the U.S.) should not be expected to make loan payments.

Baum and Schwartz recognize that manageable student loan repayment is not a linear function. Rather, there’s a level of income where the entire system falls apart due to the need to cover essential expenses. Simply put, there is some level of income so low that any positive debt-to-earnings rate should be considered unacceptable. A discretionary debt-to-earnings rate understands this concept, but the annual one does not.

The cohort default rate is the logical response to this concern, since low-income borrowers may have a harder time paying off debt and be more likely to default. Here's where income-based repayment complicates the picture.

As set by Congress, students can benefit from income-based payment plans if they have a partial financial hardship, which is defined as having loan payments exceed 15 or 10 percent of discretionary income, depending on when a borrower entered repayment. Higher income borrowers who qualify see their payments reduced. But if your income gets too low then payments drop to $0. That’s because the partial financial hardship calculation first deducts 150 percent of the poverty level based on household size from a borrower’s adjusted gross income and then uses 10 or 15 percent of the remaining amount to calculate loan payments. Those deductions establish an effective income floor of about $17,700 for a single person and $24,000 for a two-person household before they would be expected to make a payment.

The combination of income-based repayment and low-income borrowers produces a clear path for gutting the effectiveness of the cohort default rate and potentially the entire gainful employment system. To see why, return to the Everest medical secretary program. It passes the annual debt-to-earnings test despite very low incomes for graduates. But at least half its graduates would have a $0 payment on income-based repayment because they have no discretionary income to put toward loan payments. In that case, all the institution has to do is ensure it keeps the annual debt-to-earnings rate manageable. If it can do that, then borrowers could be easily flipped into income-based repayment, where they will not owe a single dollar, making default impossible. And should they stay low income for their entire loan term, the entire balance plus interest would be forgiven by the federal government. In this scenario, that program is not providing gainful employment. It is indebting impoverished graduates and successfully navigating the borrower safety net to avoid accountability.

By contrast, consider Everest’s master’s program in business administration. It too barely passes the annual debt-to-earnings rate with a figure of 7.99 percent. But it has annual earnings of $42,081. Despite substantially higher earnings than the medical secretary program, it will have a tougher time using income-based repayment to avoid problems on the cohort default rate. That’s because its borrowers likely earn enough money that even those with a partial financial hardship will still have to pay something on their loans, making repayment less passive.

Three ways to solve the low-income problem

Option 1: Require passage of the discretionary debt-to-earnings rate

The simplest way to address low-income programs would be to make the discretionary debt-to-earnings rate the primary requirement for passage. Doing so would cause programs with an acceptably low annual debt-to-earnings rate but poor earnings to fail. It sends a strong message that postsecondary programs should be expected to generate an earnings premium for graduates that afford them the ability to engage in discretionary consumption and it is from that discretionary consumption that they can make payments on their student loans.

Requiring passage of the discretionary debt-to-earnings rate would admittedly result in a stronger standard for programs. That’s because 3,048 of the 4,540 programs with earnings information and some annual debt payments in the latest data release fail to meet the discretionary debt-to-earnings standard of 20 percent, a failure rate of approximately 67 percent. That's partly because more than half of programs that fail the discretionary debt-to-earnings test (1,643 or 53 percent) have earnings so low that they do not earn even have a single dollar in discretionary income.

Option 2: Replace the program cohort default rate with an amortization test

The repayment rate is a fraught metric in this process. But that’s because past versions of the measure captured individual results and set standards for the percentage of individuals that should be successfully repaying their loans. While that may be a useful way to think of a repayment rate for consumer disclosure purposes, the Department should approach the repayment rate through the eyes of its function as a lender.

What lenders care about is amortization—are loans being retired as per the agreed upon terms and conditions. So a lender-focused repayment rate would consider the repayment rate as a measure of the amortization of the entire pool of loans it made to borrowers at a program, hereafter referred to as an amortization rate. For a given dollar amount of loans and a known interest rate, it is very easy to calculate how much of that balance should be remaining at any stated point in time. The Department could then compare the actual amount remaining at that point in time to the expected balance on an amortization table. Programs where the balance is at or below the expected level would pass on the grounds that borrowers are paying down their loans at least as fast as expected. Those with balances above expected levels would fail on the grounds that borrowers are not paying down their loans fast enough to retire their debt in a meaningful timeframe.

This option is noticeably different from the repayment rate suggested by the Department in the second negotiation session. That measure looked at whether the total outstanding amount of loans decreased by at least $1 year over year. But such a measure is not necessarily fair to schools, since borrowers may make a great deal of pre-payments in one year, resulting in a lower balance that might be harder to pay down further.

At the same time, this amortization rate benefits from not making value calls about the acceptable level of individual performance that necessitates passage or failure. Instead, it views a program’s pooled loans from a lender’s perspective and asks whether the entire set is being repaid at an acceptable rate or not.

This option address the cohort default rate problem and IBR problem because borrowers with very low incomes that result in $0 payments would hurt the amortization rate by growing the total outstanding loan balance for the program. And for all borrowers it would require an orderly retirement of debt, which requires much more than the $1 a year reduction required by the old repayment rate test.

An amortization rate would still need standards for acceptable repayment time. There are two logical standards to choose. The first and laxest would be to test whether loans are being retired on a 20-year amortization schedule. This would judge whether borrowers are repaying their loans fast enough to avoid needing forgiveness under IBR. Since it's a pooled rate, it does not preclude some borrowers from using such plans if they need it but only causes problems if large numbers of people are making very small income-based payments.

The other logical thresholds would be to use 10, 15, or 20 years based upon the credential type, as the Department is already proposing to do for the debt-to-earnings rates. In this case, a certificate program would have to show its amount outstanding at the end of three fiscal years is equal to or less than the amount that should be outstanding at that point in time if the loans were going to be completely retired by the end of 10 years, while a bachelor’s degree program would have to do the same thing but for a 15-year repayment schedule.

Option 3: Adopt a minimum earnings level for programs with any expected debt payments

The big reason why the gainful employment framework is ineffective with lower-earnings programs is because there’s no acknowledgement that at some point incomes get so low that any debt payments should be considered unacceptable. This option would fix that problem by stating that any program with typical earnings below a certain level and debt payments greater than $0 would automatically fail the debt-to-earnings tests. It is in effect a requirement of generating some level of a non-infinite discretionary debt-to-earnings rate.

Programs would still have to pass either the annual or discretionary debt-to-earnings tests. But they would also have to show that the earnings level observed exceed at least a minimum threshold. The recommended one to use would be 150 percent of the poverty level for a single person. Graduates with earnings at or above that level are more likely to at least make some payments on their student loans through income-based repayment plans—removing that potential loophole. It also lines up with what’s considered minimum income for discretionary purposes on the second debt-to-earnings test without requiring them to actually pass that measure.

This option would only apply to programs that had some amount of debt in the numerator of a debt-to-earnings rate. The rationale is that as noted by Baum and Schwartz, very low income people should not be diverting any of their earnings to loan payments. Programs with low incomes and debt payments of $0 are not ideal, but do not suffer from this concern and thus should still pass.

If enacted, this requirement would affect an estimated 1,101 programs that currently have annual debt-to-earnings rates of 8 percent or less, annual debt payments greater than $0 and no discretionary income.