The Biden administration plans to release new gainful employment regulations. The regulations would terminate federal financial aid for some programs where graduates do not earn more than high school graduates or where the students take on excessive debt, as determined by two debt-to-income tests. My previous list of pros and cons still holds, but having had more time to dive into the details, I’ve found a few more issues.
1. Federal policy shouldn’t vary by state
When determining whether a college program’s graduates earn more than a high school graduate, the administration wants to use the state average high school wage, a value that ranges from $20,859 in Mississippi to $31,294 in North Dakota.
This raises three problems. First, suppose there is a program with median earnings of $30,000. This program would fail in North Dakota but would pass in Mississippi. But why should the location of a college affect whether it is eligible for federal funding?
Second, a statewide average is too broad. The ostensible purpose of varying the wage by state is to account for different labor-market possibilities. Clearly, high school graduates in North Dakota have greater opportunities due to the oil boom than those from Mississippi. However, a state is much too large to accurately account for these opportunities. It takes around 10 hours to drive from El Paso to Houston, so opportunities in one city will not help people who live in the other. A statewide average risks underestimating opportunities in some areas of the state while overestimating opportunities in others.
Third, as a general rule, the federal government takes a uniform approach to states in higher education. The maximum Pell grant doesn’t vary by state. Nor does student loan eligibility. Given that student access to these programs is uniform rather than state-dependent, it seems bizarre to then introduce an accountability system that would limit institutional access to these programs by state.
2. The regulations ignore debt for politically favored sectors
Any higher education accountability program quickly runs into a political problem: community colleges. Outcomes at inexpensive community colleges tend to be terrible (the national graduation rate is around one-third). While associate degree programs at community colleges are exempt from the debt-to-income and earnings-floor tests, certificate programs are not. Certificates make up a larger share of enrollment than most people think, with about an equal number of certificates and associate degrees awarded each year (roughly 1 million). Any accountability system that includes certificates will hit community colleges hard. Since most congressional districts have at least one community college or branch campus, it is therefore politically fraught to introduce an accountability system with teeth.
The Biden administration’s solution to this problem is to measure, but then simply ignore, the debt of community college students. Note the bolded text in this description of how the regulations would measure debt:
… the Department would calculate the annual loan payment for a program by (1) Determining the median loan debt of the students who completed the program during the cohort period, based on the lesser of the loan debt incurred by each student, computed as described in § 668.403(d), or the total amount for tuition and fees and books, equipment, and supplies for each student, less the amount of institutional grant or scholarship funds provided to that student; removing the highest loan debts for a number of students equal to those for whom the Federal agency with earnings data does not provide median earnings data; and calculating the median of the remaining amounts … [emphasis added]
In other words, the regulations completely ignore the actual amount of debt for low-cost (read: community college) programs. Debt used to cover room and board is simply ignored. This is a get-out-of-jail-free card for community colleges, which tend to have low tuition.
3. The discretionary income test is too arbitrary
The Discretionary Income Rate (DIR), one of the debt-to-income tests, is arbitrary. Programs fail this test if graduates’ median debt relative to median earnings would entail a debt-service-to-income percentage of greater than 20% of discretionary income, where discretionary income is defined as income above 150% of the poverty line.
This is an unorthodox choice, so why did the Department choose it? “The acceptable threshold for the discretionary income rate would be set at 20 percent, based on research conducted by economists Sandy Baum and Saul Schwartz.” Yet the Baum and Schwartz report is not sufficient to justify the policy. As the authors write in this same report, is it not possible to define discretionary income non-arbitrarily: “there is weak theoretical rationale for this type of calculation. Any attempt to draw a line between discretionary and nondiscretionary expenditures is fraught with difficulty.” Setting a threshold for discretionary income “clearly requires a subjective judgment.” They argue for 150% of the poverty line, a recommendation that Biden’s proposed regulations implement. But that choice was a subjective judgment with a weak theoretical rationale.
To be clear, Baum and Schwartz did nothing wrong. After conducting an analysis, they made a policy recommendation. The error here lies with the Department of Education, which has presented the Baum and Schwartz recommendation as though it is the consensus opinion or was somehow derived analytically, rather than one idea among many. To be non-arbitrary, the regulations would need to have considered a broad range of options (e.g., 100%, 125%, or 200% of the poverty line), not simply cite one example of external analysts recommending the Department’s preferred value. Moreover, the Department is using a completely different cutoff (225%) for its proposed repayment plan. This lack of consistency clearly suggests that the Biden administration is choosing cutoffs to achieve predetermined outcomes.
4. Dealing with non-borrowers
Another issue with the proposed regulations concerns non-borrowers—namely, should non-borrowers be included when determining the median debt? There isn’t a knockout winner here. The extremes illustrate the danger of either approach. Suppose there is a perfectly good program in which only 1% of students borrow, but these borrowers’ income is so low that the program fails the debt-to-income tests. If non-borrowers are excluded, this program would be eliminated from the federal financial aid programs, including Pell grants. But the converse is even less appealing. Suppose there is a program where 50.1% of students don’t borrow, while 49.9% do, and that all those 49.9% take on excessive debt with ruinous financial consequences. Clearly, extending loans for this program is a mistake. Yet, when non-borrowers are included, the median debt would be $0, and the program would pass the debt-to-income tests, even though the 49.9% of students that did borrow are financially crippled.
The regulations include non-borrowers—while neither option is perfect, excluding non-borrowers is better. To begin with, students who don’t borrow will, naturally, ignore debt-to-income metrics, as these tests are irrelevant for them. Conversely, borrowers will (hopefully) pay attention to debt-to-income metrics, yet the proposed method would severely mislead them about how past borrowers have fared by underestimating debt. In addition, a policy that actively harms is worse than one that fails to assist, which, in this context, means that it is worse to allow and encourage students to take out financially devasting loans than it is to fail to offer worthwhile educational loans.
5. The Department’s data is not useful in evaluating the policy
To help the public evaluate and understand the proposed regulations, the department released a dataset. But this dataset is not particularly useful in evaluating the policy. There are two big issues.
First, earnings appear to be underestimated. Median earnings in the data are much lower than other earnings measures released by the Department. For example, the data reports median earnings of $34,060 among the programs. Yet the Department’s College Scorecard data reports an average of $47,048 for the same cohort of students. It is not clear to me what is driving the difference. One possibility is that non-workers are included when determining the median value in the GE data but not in the College Scorecard data. If this is the case, it is inconsistent with the approach used to measure high school graduate earnings, which is measured “excluding individuals not in the labor force,” meaning that those not in the labor force are counted for college earnings but not for high school earnings. If this is true, the proposed regulations would be biased against colleges by including non-workers when determining their median wage but excluding non-workers among high school graduates.
Second, debt is a key metric in the regulations. The dataset released by the Department includes Parent PLUS loans in the median. Yet the proposed regulations do not include Parent PLUS loas. In other words, the debt data released by the Department is not useful in evaluating the policy.
Given these and the previously highlighted problems, the Biden administration should not move forward with these regulations. If it does, court cases should be filed to determine whether the Department was acting arbitrarily, and the next administration should rescind the regulations.