2019-13703. Program Integrity: Gainful Employment

AGENCY:

Office of Postsecondary Education, Department of Education.

ACTION:

Final regulations.

SUMMARY:

The Secretary of the Department of Education (Department) amends the regulations on institutional eligibility under the Higher Education Act of 1965, as amended (HEA), and the Student Assistance General Provisions to rescind the Department's gainful employment (GE) regulations (2014 Rule).

DATES:

Effective date: These regulations are effective July 1, 2020.

Implementation date: For the implementation date of these regulatory changes, see the Implementation Date of These Regulations section of this document.

FOR FURTHER INFORMATION CONTACT:

Scott Filter, U.S. Department of Education, 400 Maryland Avenue SW, Room 290-42, Washington, DC 20202. Telephone [phone]. Email: [e-mail].

If you use a telecommunications device for the deaf (TDD) or a text telephone (TTY), call the Federal Relay Service (FRS), toll free at 1-800-877-8339.

SUPPLEMENTARY INFORMATION:

Executive Summary

Purpose of This Regulatory Action: This regulatory action rescinds the GE regulations and removes and reserves subpart Q of the Student Assistance General Provisions in 34 CFR part 668. This regulatory action also rescinds subpart R of the Student Assistance and General Provisions in 34 CFR part 668.

As discussed in the sections below, the Department has determined that the GE regulations rely on a debt-to-earnings (D/E) rates formula that is fundamentally flawed and inconsistent with the requirements of currently available student loan repayment programs, fails to properly account for factors other than institutional or program quality that directly influence student earnings and other outcomes, fails to provide transparency regarding program-level debt and earnings outcomes for all academic programs, and wrongfully targets some academic programs and institutions while ignoring other programs that may result in lesser outcomes and higher student debt. Although the GE regulation applies to less-than-degree programs at non-profit institutions, this represents a very small percentage of academic programs offered by non-profit institutions.

Table 1-1—Reporting Overview of Gainful Employment Programs

School classification	GE Programs qualifying for calculation (based on NSLDS reporting)	GE programs published	Percent of GE programs published (%)	GE programs not published	Percent of GE programs not published (%)	Failing GE programs	Failure rate (%)
Proprietary	9,838	5,676	57.70	4,162	42.30	727	12.80
Non-profit	18,962	2,956	15.60	16,006	84.40	16	0.50
Foreign	17	5	29.40	12	70.60	0	0.00
Total	28,817	8,637	30.00	20,180	70.00	743	8.60
Data from Federal Student Aid.

As table 1-1 shows only 16 percent (2,956) of the 18,962 GE programs at non-profit institutions meet the 30-student cohort size requirement. Therefore, only a small minority of those programs are subject to the D/E rates calculation and certain reporting requirements. On the other hand, all programs at proprietary institutions—including undergraduate, graduate, and professional programs—are considered to be GE programs, and 58 percent (5,676) of programs meet the minimum student threshold to report outcomes to the public. As a result, the GE regulations have a disparate impact on proprietary institutions and the students these institutions serve. The regulations also fail to provide transparency to students enrolled in poorly performing degree programs at non-profit institutions and fail to provide comparison information for students who are considering enrollment options at both non-profit and proprietary institutions. Specifically, the Department's review of research findings published subsequent to the 2014 Rule, our review of the 2015 Final GE rates (published in 2017),^[1] and our review of a sample of GE disclosure forms published by proprietary and non-profit institutions, has led the Department to conclude the following: (1) As a cornerstone of the GE regulations, the D/E rates measure ^[2] is an inaccurate and unreliable proxy for program quality and incorporates factors into the calculation that inflate student debt relative to actual repayment requirements; (2) the D/E rates thresholds, used to differentiate between “passing,” “zone,” and “failing” programs, lack an empirical basis; and (3) the disclosures required by the GE regulations include some data, such as job placement rates, that are highly unreliable and may not provide the information that students and families need to make informed decisions about higher education options.

In addition, since the Social Security Administration (SSA) has not signed a new Memorandum of Understanding (MOU) with the Department to share earnings data, the Department is currently unable to calculate D/E rates, which serve as the basis of the 2014 Rule's accountability framework.^[3] The GE regulations specify that SSA data must be used to calculate D/E rates, meaning that other government data sources cannot be used to calculate those rates. Because the Department was unaware at the time of negotiated rulemaking and publication of the notice of proposed rulemaking (NPRM) (83 FR 40167) that SSA would not renew the MOU, we did not address this issue, nor did we suggest, or seek comment on, the potential use of earnings data from the Internal Revenue Service (IRS) or the Census Bureau to calculate D/E rates. Therefore, switching to IRS or Census Bureau data for the purpose of calculating D/E rates would require additional negotiated rulemaking. However, since the Department has decided to rescind the GE regulations, the data source for calculating D/E rates is moot.

The 2014 Rule was developed in response to concerns about poor outcomes among GE programs that left students with debt that was outsized, relative to student earnings in the early years of student loan repayment. For example, the Department pointed to cohort default rates (CDRs) that were disproportionately high among students who enrolled at or completed their educational programs at proprietary institutions as an indication that the education provided was of lower quality.^[4]

However, research published in 2014—and discussed throughout this document—but not considered during the Department's development of the 2014 Rule, confirms that CDRs are largely influenced by the demographics and socioeconomic status of borrowers, and not necessarily institutional quality.^[5] This makes CDRs a poor proxy for institutional quality, and therefore insufficiently justifies the GE regulations.

The 2014 paper also shows that CDRs disproportionately single-out institutions that serve larger percentages of African-American students or single mothers, since these demographic groups default at higher rates and sooner after entering repayment than other borrowers.^[6] The authors of this study point to reduced parental wealth transfers to minority students as the reason that defaults are higher among this group. As a result, institutions that serve larger proportions of minority students will likely have higher CDRs than an institution of equal quality that serves mostly white or more socioeconomically advantaged students. Thus, higher CDRs among minority students may be a strong sign of lingering societal inequities among different racial groups, but not conclusive evidence that an institution is failing its students. The Department now recognizes that a number of studies used to support its earlier rulemaking efforts relied on comparisons between costs and debt levels among students who enrolled at community colleges and those who enrolled at proprietary institutions. However, this is an illegitimate comparison since in 2014, 53 percent of proprietary institutions were four-year institutions, and 63 percent of students enrolled at proprietary institutions were enrolled at four-year institutions.^[7] Therefore, with regard to costs and student debt levels, comparisons with four-year institutions are more appropriate.

Comparisons between students who attend community colleges and those who attend proprietary institutions may be appropriate, especially since both are generally open-enrollment institutions. However, research published by the Brown Center in 2016 shows that there are considerable differences between the characteristics of students who enroll at proprietary institutions and those who enroll at two-year public institutions.^[8] Students who enroll at proprietary institutions are far more likely to be financially independent (80 percent vs. 59 percent); part of an underrepresented minority group (52 percent vs. 44 percent); or a single parent (33 percent vs. 18 percent) than students enrolled at community colleges. Students enrolled at proprietary institutions are also slightly less likely to have a parent who completed high school (84 percent vs. 87 percent); and are much less likely to have a parent who completed a bachelor's degree or higher (22 percent vs. 30 percent). These differences in characteristics may explain disparities in student outcomes, including higher borrowing levels and student loan defaults among students who enroll at proprietary institutions.

Research published in 2015 by Sandy Baum and Martha Johnson pointed to student and family demographics, as well as length of time in school, as key determinants of borrowing.^[9] Therefore, research published subsequent to promulgation of the 2014 Rule showed that differences in borrowing levels and student outcomes may well be attributable to student characteristics and may not accurately indicate institutional quality or be influenced by institutional tax status.

The Department has also come to realize that unlike CDRs that measure borrower behavior in the first three years of repayment, lifecycle loan repayment rates more accurately illustrate the challenges that the majority of students are having in repaying their student loan debt and the need to look beyond one sector of higher education to solve this problem. In 2015, the Department began calculating institution-level student loan repayment rates in order to include those rates in its newly introduced College Scorecard and reported that the majority of borrowers at most institutions were paying down their principal and interest.

However, in January 2017, the Department reported that it had discovered a coding error, making the repayment data it had published earlier incorrect.^[10] Though the Department's announcement downplayed the magnitude of this error, both Robert Kelchen, assistant professor of higher education at Seton Hall, and Kim Dancy, a New America policy analyst, independently found that the error was significant.^[11]

Prior to correcting the error, it was determined that three years into repayment, 61 percent of borrowers were paying down their loans—meaning that these borrowers had reduced their principal by at least one dollar. This reinforced the belief that only a minority of borrowers were struggling to repay debt—such as borrowers who attended proprietary institutions.

However, once the error was corrected, it became clear that repayment rates were actually much lower. The corrected data reveals that only 41 percent of borrowers in their third year of repayment were paying down their loan balances by at least one dollar. As noted by Dancy, “the new data reveal that the average institution saw less than half of their former students managing to pay even a dollar toward their principal loan balance three years after leaving school.” ^[12]

The 2017 corrected repayment rate data led the Department to conclude that the transparency and accountability frameworks created by the GE regulations were insufficient to address the student borrowing and under-payment problem of this magnitude, as the GE regulations apply to only a small proportion of higher education programs.^[13] In order to enable all students to make informed enrollment and borrowing decisions, the Department sought an alternative to the GE regulations that would include all title IV-eligible institutions and programs.

The GE regulations failed to equitably hold all institutions accountable student outcomes, such as student loan repayment. However, the Department could not simply expand the GE regulations to include all title IV programs since the term “gainful employment” is found only in section 102 of the HEA. This section extends title IV eligibility to non-degree programs at non-profit and institutions and all programs at proprietary institutions, and at the same time restricts the application of the GE regulations to those same programs and institutions. Therefore, without a statutory change, there was no way to expand the GE regulations to apply to all institutions.

As a result, the Department engaged in negotiated rulemaking to evaluate the accuracy and usefulness of the GE regulations and to explore the possibility of creating a “GE-like” regulation that could be applied to all institutions and programs. The Department sought to develop a new transparency and accountability framework that would apply to all institutions and programs, likely through the Program Participation Agreement (PPA).

Unfortunately, negotiations ended having failed to reach consensus on how to improve the accuracy, validity, and reliability of the GE regulations, and having failed to develop a valid GE-like standard that could serve as the basis for an appropriate and useful accountability and transparency framework for all title IV-participating programs.

In 2018, the Department's office of Federal Student Aid (FSA) determined that the student loan repayment situation was more dire than we originally thought. Analysis of 2018 third quarter data showed that only 24 percent of loans, or $298 billion, are being reduced by at least one dollar of principal plus interest, and that 43 percent of all outstanding loans, or $505 billion, are in distress, meaning they are at risk, either through negatively amortizing Income-Driven Repayment (IDR) plans, 30 plus days delinquent, or in default.^[14] These data reinforce the need for an accountability and transparency framework that applies to all title IV programs and institutions.

Failing to have reached consensus during negotiations, the Department determined that the best way to improve transparency and inform students and parents was through the development of a comprehensive, market-based, accountability framework that provides program-level debt and earnings data for title IV programs. The College Scorecard was selected as the tool for delivering those data, and by expanding the Scorecard to include program-level data, all students could make informed enrollment and borrowing decisions.

Given the Department's general authority to collect and report data related to the performance of title IV programs, the Department is not required to engage in rulemaking to modify the College Scorecard. However, to address concerns that by rescinding the 2014 Rule some students would be more likely to make poor educational investments, the Department describes in this document our preliminary plans for the expansion of the College Scorecard.

As outlined in President Trump's Executive Order on Improving Free Inquiry, Transparency, and Accountability at Colleges and Universities,^[15] the Department plans to expand the College Scorecard to include the following program-level data: (1) Program size; (2) the median Federal student loan debt and the monthly payment associated with that debt based on a standard repayment period; (3) the median Graduate PLUS loan debt and the monthly payment associated with that debt based on a standard repayment period; (4) the median Parent PLUS loan debt and the monthly payment associated with that debt based on a standard repayment period; and (5) student loan default and repayment rates.

In addition to the information above, College Scorecard will continue to include institution-level data, such as admissions selectivity, student demographics, and student socioeconomic status. This information will provide important context to help students compare outcomes among institutions that serve demographically matched populations or that support similar educational missions.

The College Scorecard ensures that accurate and comparable information is disclosed about all programs and institutions. It provides a centralized access point that enables students to compare outcomes easily without visiting multiple institution or program websites and with the certainty that the data they are reviewing were produced by a Federal agency. This eliminates the potential for institutions to manipulate or exaggerate data, which is possible when data are self-reported by institutions.

As a result of these changes, students and parents will have access to comparable information about program outcomes at all types of title IV-participating institutions, thus expanding higher education transparency. Students will be able to make enrollment choices informed by debt and earnings data, thus enabling a market-based accountability system to function. These changes will also help taxpayers understand where their investments have generated the highest and lowest returns.

Summary of the Major Provisions of This Regulatory Action: The Department rescinds 34 CFR part 668, subpart Q—Gainful Employment Programs.^[16] The term “gainful employment” was added to the HEA in 1968 to describe training programs that gained eligibility to participate in title IV, HEA programs. The 2014 Rule defined “gainful employment” based on economic circumstances rather than educational goals, created a new D/E rates measure to distinguish between passing and failing programs, and established other reporting, disclosure, and certification requirements applicable only to GE programs.

By rescinding subpart Q, the Department is eliminating the D/E rates measure, which is an inaccurate and unreliable proxy for quality, including the use of the 8 percent debt-to-earnings threshold and the 20 percent debt-to-discretionary-income threshold as the requirement for continued eligibility of GE programs. By rescinding subpart Q, we also eliminate the requirement for institutions to issue warnings, including hand-delivered notifications, in any year in which a program is at risk of losing title IV eligibility based on the next year's D/E rates.

Rescinding the GE regulations also eliminates the need for institutions to report certain data elements to the Department in order to facilitate the calculation of D/E rates. It also eliminates requirements for GE programs to publish disclosures that include the following: Program length; program enrollment; loan repayment rates; total program costs; job placement rates; percentage of enrolled students who received a title IV or private loan; median loan debt of those who completed and those who withdrew from the program; program-level cohort default rates; annual earnings; whether or not the program meets the educational prerequisites for professional licensure or certification in each State within the institution's metropolitan service area or for any State for which the institution has determined that the program does not meet those requirements; whether the program is programmatically accredited and the name of the accrediting agency; and a link to the College Navigator website. The table in Appendix A compares the information that was made available to students and parents through the 2017 GE disclosure template with the information that will be provided through the expanded College Scorecard or other consumer information tools, such as College Navigator. Disclosure requirements are also being included in other rulemaking efforts, including Borrower Defense regulations and Accreditation and Innovation regulations.

In addition, by rescinding subpart Q, the Department is also eliminating requirements regarding alternate earnings appeals, reviewing and correcting program completer lists, and providing certification by the institution's most senior executive officer that the programs meet the prerequisite education requirements for State licensure or certification.

Finally, the Department rescinds 34 CFR part 668, subpart R—Program Cohort Default Rate, including instructions for calculating those rates and disputing or appealing incorrect rates provided by the Secretary. As the Department only contemplated calculating those rates as part of the disclosures under the GE regulations, we can find no compelling reason to maintain subpart R and did not identify public comments to this aspect of the proposed regulations. We note that the HEA requires the Department to calculate institutional cohort default rates, and regulations regarding the calculation of those rates are in 34 CFR 668.202.

Authority for this Regulatory Action: Section 410 of the General Education Provisions Act provides the Secretary with authority to make, promulgate, issue, rescind, and amend rules and regulations governing the manner of operations of, and governing the applicable programs administered by, the Department. 20 U.S.C. 1221e-3. Furthermore, under section 414 of the Department of Education Organization Act, the Secretary is authorized to prescribe such rules and regulations as the Secretary determines necessary or appropriate to administer and manage the functions of the Secretary or the Department. 20 U.S.C. 3474. These authorities, together with the provisions in the HEA, permit the Secretary to disclose information about title IV, HEA programs to students, prospective students, and their families, the public, taxpayers, and the Government, and institutions. Further, section 431 of the Department of Education Organization Act provides authority to the Secretary, in relevant part, to inform the public about federally supported education programs and collect data and information on applicable programs for the purpose of obtaining objective measurements of the effectiveness of such programs in achieving the intended purposes of such programs. 20 U.S.C. 1231a.

For the reasons described in the NPRM and below, the Department believes that the GE regulations do not align with the authority granted by section 431 of the Department of Education Organization Act since the D/E rates measure that underpins the GE regulations does not provide an objective measure of the effectiveness of such programs.

Costs and Benefits: The Department believes that the benefits of these final regulations outweigh the costs. There will be one primary cost and several outweighing benefits associated with rescinding the GE regulations. The primary cost is that some programs that may have failed the D/E rates measure, and as a result lose title IV eligibility, will continue to participate in title IV, HEA programs. In instances in which the program failed because it truly was a low-quality program, there is a cost associated with continuing to provide title IV support to such a program, especially if doing so burdens students with debt they cannot repay or an educational credential that does not improve their employability. However, there are numerous benefits associated with eliminating the GE regulations, including: (1) Programs producing poor earnings outcomes will not escape notice simply because taxpayer subsidies make the program less costly to students; (2) programs that prepare students for high-demand careers will be less likely to lose title IV eligibility just because those high-demand careers do not pay high wages; (3) students will not inadvertently select a non-GE program with less favorable student outcomes than a comparable GE program simply because non-GE programs are not subject to the GE regulations; (4) institutions will save considerable time and money by eliminating burdensome reporting and disclosure requirements; (5) all students will retain the right to enroll in the program of their choice, rather than allowing government to decide which programs are worth of a student's time and financial investment; and (6) by providing debt and earnings data for all title IV programs through the College Scorecard, all students will be able to identify programs with better outcomes or limit borrowing based on what they are likely to be able to repay. The Department believes that the benefits outweigh the costs since all students will benefit from choice and transparency.

Implementation Date of These Regulations: These regulations are effective on July 1, 2020. Section 482(c) of the HEA requires that regulations affecting programs under title IV of the HEA be published in final form by November 1, prior to the start of the award year (July 1) to which they apply. However, that section also permits the Secretary to designate any regulation as one that an entity subject to the regulations may choose to implement earlier, as well as the conditions for early implementation.

The Secretary is exercising her authority under section 482(c) of the HEA to designate the regulatory changes to subpart Q and subpart R of the Student Assistance General Provisions at title 34, part 668, of the Code of Federal Regulations, included in this document, for early implementation beginning on July 1, 2019, at the discretion of each institution.

Public Comment: In response to our invitation in the NPRM, 13,921 parties submitted comments on the proposed regulations. In this preamble, we respond to those comments, which we have grouped by subject. Generally, we do not address technical or other minor changes.

Analysis of Public Comments: An analysis of the public comments received follows.

Scope and Purpose

Comments: Many commenters indicated they supported rescinding the GE regulations because defining “gainful employment” using a bright-line debt-to-earnings standard is complicated and does not accurately differentiate between high-quality and low-quality programs, or programs that do and do not meet their learning objectives. A number of commenters also supported the Department's decision to rescind the GE regulations because they believe the regulations discriminate against career and technical education (CTE) programs and the students who enroll in them. Some suggested that the GE regulations signal to students that CTE is less valuable than traditional liberal arts education since the Department, as a result of the GE regulations, was holding traditional degree programs to a lower standard. Other commenters expressed concern that the GE regulations discriminate against institutions based on their tax status.

Several commenters stated that the GE regulations threaten to limit access to necessary workforce development programs at community colleges and at proprietary schools, as a result of the increased accountability for CTE programs as compared to liberal arts and humanities programs. Another commenter expressed concern that the D/E rates measure ignores or exempts a significant number of programs with the worst outcomes, simply because those programs are offered by public and non-profit institutions or receive taxpayer subsidies in the form of direct appropriations rather than or in addition to Pell grants and title IV loans.

Multiple commenters supported the rescission of the GE regulations because, in their opinion, the GE regulations would otherwise force the closure of programs and potentially entire institutions that serve minority, low-income, adult, and veteran students.

One commenter highlighted the lack of guidance from Congress on the meaning of “gainful employment,” and asserted that in the absence of that guidance, the Department contrived a complicated regulation that has yielded “a patchwork of complicated and inconsistent rules that have left schools buried in paperwork with no real measure of whether students have benefited.”

Some commenters suggested that any institution could ensure that they will pass the D/E rates measure by lowering tuition. Several commenters submitted a joint comment opposing the rescission of the 2014 Rule. They argued that the rescission is arbitrary and capricious because it ignores both the benefits of the 2014 Rule and the data analysis supporting the 2014 Rule. The commenters noted that Congress had reason to require that for-profit programs be subject to increased supervision. They cited a post on the Federal Reserve Bank of New York's blog that states that attending a four-year private for-profit college is the strongest predictor of default, even more so than dropping out.^[17] They cited evidence that students who attend for-profit institutions are 50 percent more likely to default on a student loan than students who attend community colleges.^[18] The commenters also argued that a rise in enrollment in the for-profit sector corresponded with reports of fraud, low earnings, high debt, and a disproportionate amount of student loan defaults. They claimed that of the 10 percent of institutions with the lowest repayment rates, 70 percent were for-profit institutions. They argued that because poor outcomes are concentrated in for-profit programs, the 2014 Rule is justified.

Commenters also noted that students enrolled in programs that close generally re-enroll in nearby non-profit or public institutions and that shifting aid to better performing institutions will result in positive impacts for students. They also cited evidence ^[19] that, after enrollment in for-profit programs declined in California, local community colleges increased their capacity. They argued that in light of these examples, the 2014 Rule would not reduce college access for students but would rather direct them into programs that are more beneficial in the long term.

One commenter disagreed with the Department for citing the Bureau of Labor Statistics (BLS) Job Openings and Labor Turnover Survey as evidence that certain jobs are “unfilled due to the lack of qualified workers.” ^[20] The commenter also stated that there is no evidence that the job openings in the BLS survey relate in any way to GE programs. Another commenter stated that the Department should withdraw its claim based on this study because the BLS press release did not note any relation to gainful employment.

Discussion: The Department appreciates the support received from many commenters who agreed that the D/E rates measure is a fundamentally flawed and unreliable quality indicator and that the limited applicability of the 2014 Rule to some, but not all, higher education programs makes it an inadequate solution for informing consumer choice and addressing loan default issues. Further, the Department agrees that the formula for deriving D/E rates is complicated and that it may be difficult for students and parents to understand how it was calculated and how to apply it to their own situation to determine what their likely debt and earnings outcomes will be.

The Department shares the concern of commenters who predicted that the GE regulations would result in reduced access to certain CTE focused programs. However, since no programs have lost eligibility as of yet, it is impossible to know for certain what longer-term impacts the GE regulations would have had. That said, some commenters have pointed to programs like Harvard's graduate certificate program in theater,^[21] which was discontinued in part because the university knew that the program would not pass the D/E rates measure, and large closures among art and design or culinary schools as evidence that some schools voluntarily discontinued programs in order to avoid sanctions under the GE regulations.

The Department agrees with commenters that the D/E rates measure does not accurately differentiate between high- and low-quality programs or eliminate programs that produce the worst outcomes, since programs that generate much lower earnings can pass the D/E rates measure simply because taxpayers rather than students pay some of the cost of the education provided, thus reducing the price students pay.

For example, a Colorado public community college's massage therapy program passed the D/E rates measure despite having mean annual earnings of $9,516, whereas a comparable program at a Colorado proprietary institution that resulted in earnings of $15,929 failed the D/E rates measure. The Department understands that high student loan debt can be burdensome to students, especially to those who earn low wages. However, it is difficult to argue that the program yielding earnings of $9,516 is higher quality than one that yields earnings of $15,929. As is the case with four-year public and private institutions, tuition is higher at institutions that receive fewer public subsidies.

To provide another example, consider that in Ohio, a medical assistant program at a community college passed the D/E rates measure even though its graduates had median annual earnings of $14,742. Meanwhile, a medical assistant program at a proprietary institution in Ohio failed the D/E rates measure even though its graduates posted median earnings of $21,737. In Arizona, two proprietary institutions' interior design programs failed the D/E rates measure, despite having significantly higher median annual earnings ($31,844 and $32,046) than a nearby community college program ($19,493).

As stated by Cooper and Delisle with regard to the D/E rates measure, “the danger here is that a program at a public institution may provide a low return on investment from a societal perspective, but pass the GE rule anyway because a large portion of the cost of providing it is not taken into account.” ^[22] Cooper and Delisle state that this creates a distortion effect that may render student choices as rational for themselves, but disadvantageous to society.^[23] In other words, while taxpayer subsidies to public institutions ensure that they pass the D/E rates measure, that may hide from students and taxpayers the amount of funding that is being used to administer ineffective programs and may fool students into enrolling in a program that has passing D/E rates without realizing that the earnings generated by the program do not justify the direct, indirect, or opportunity costs of obtaining that education. Although there are low-performing programs in all sectors, students have received only limited information about them because the GE regulations do not apply to programs in all sectors.

As is the case among all private institutions, the absence of State and local taxpayer subsidies means that students bear a larger portion of the cost of education, which generally means that tuition and fees are higher than at public institutions. Even at public institutions, students who are from outside of the State or the country pay tuition and fees that more closely resemble those of private institutions, thus demonstrating the impact of direct appropriations on subsidizing tuition costs for State residents. Yet title IV programs do not limit financial aid to students who select a public institution or the lowest cost institution available. Instead, title IV programs provide additional sources of aid, including additional funding programs (such as campus-based aid programs), to ensure that low-income students can pick the college of their choice, even if doing so means that the student needs more taxpayer-funded grants and loans.

Congress created the campus-based aid programs, in part, so that low-income students would not be limited to public institutions.^[24] The campus-based aid programs provide the largest allocations to private, non-profit institutions that have been long-term participants in the program. Creating a system of sanctions that penalizes private institutions for charging more than public institutions is contrary to the foundation of the title IV programs, which were designed to promote freedom of institutional choice. Prices will vary among institutions, as will debt levels among students based on the socioeconomic status and demographics of students served.^[25] But those variances do not, themselves, serve as accurate indicators or program quality.

Students make decisions about where to attend college based on many different factors, and they do so understanding that costs vary from one institution to the next. Students also make independent decisions about borrowing, and those decisions are influenced by any number of factors, including family socioeconomic status, cost of attendance, and the degree to which the student is required to support himself or herself and his or her family while enrolled in school. The Department believes that it is important to help inform those decisions so that students understand the impact of their decisions on their longer-term financial status.

The Department recognizes that over-borrowing for a low-value education that does not improve earnings is a serious challenge that could have long-term negative consequences for individual students, and it urges institutions to rein in escalating costs. However, it is unreasonable to sanction institutions simply because they serve students who take advantage of Federal Student Aid programs that Congress has made available to them, or because they operate without generous direct contributions from taxpayers.^[26]

Students have the right to know what the cost of attendance is at any institution they are considering, which is already required by law.

The Department agrees with commenters who expressed concern that the GE regulations established policies that unfairly target career and technical education programs. For example, under the GE regulations, student loan debt is calculated using an amortization term that assumes these borrowers, unlike others, are required to repay their loans in 10 years if they earned an associate's degree or less, 15 years if they earned a baccalaureate or master's degree, and 20 years if they earned a doctoral or professional degree. However, the law provides for students enrolled in both GE and non-GE programs to have as many as 20 or 25 years to repay their loans, and receive loan forgiveness for the balance, if any, that remains at the end of the repayment period. The amortization terms used to calculate D/E rates are in direct conflict with the amortization terms made available by Congress, and the Department in the case of the Revised Pay As You Earn (REPAYE) repayment plan, to all borrowers.

Therefore, for students, especially those sufficiently distressed to provide low repayment, the GE regulations create an inconsistent standard that suggests students who enroll in GE programs should be expected to repay their student loan debts more rapidly than students who enroll in non-GE programs. Therefore, the Department agrees with commenters who expressed concern that the GE regulations send a strong message that those pursing career and technical education are less worthy of taxpayer investment, or that they have greater, or at least faster, repayment obligations than students who enroll in other kinds of programs. This contradicts the purpose of title IV, HEA programs, which were developed to expand opportunity to low-income students. These students are served disproportionately by institutions offering CTE programs.

The Administration does not believe that students who enroll at proprietary institutions are unaware that other options are available, and the assertion that they are unsophisticated is condescending and based on false stereotypes.

According to analysis provided by Federal Student Aid, in 2018, 42.2 percent of students currently enrolled at proprietary institutions had enrolled at a non-profit institution during a prior enrollment,^[27] which suggests that these students are well aware that other, lower cost options exist. Perhaps better access to programs of choice, more flexible scheduling, more convenient locations, or a more personalized college experience compels students to pay more for their education. This is not unlike wealthier students who select an elite private institution over a public institution that offers the same programs at lower cost.

The Department believes it is important to provide earnings information to all students for as many title IV participating programs as possible so that no student or family—regardless of their socioeconomic status—is misled about likely earnings after completion. A program that yields low earnings is no less a problem for low- or middle-income students enrolled in a general studies or an arts and humanities program than it is for a low-or middle-income student enrolled in a CTE-focused program. While the goals of programs may differ, nearly all students who go to college today do so with the expectation of increasing their economic opportunity, and all students, regardless of institution type, are expected to repay their loans.

The Department's review of student loan repayment rates makes it clear that the problem of students borrowing more than they can repay through a standard repayment period is a problem that is not limited to students who attend proprietary institutions or who participate in CTE.

Regardless of institutional type or institutional tax status, colleges that serve large numbers and proportions of low-income students, minority students, and adult learners are likely to have outcomes that are not as strong as those of institutions that serve a more advantaged student population. Therefore, any effort to place sanctions on institutions that does not also take into account the socioeconomic status and demographics of students served unfairly targets those institutions that are expanding access and opportunity to students who are not served by more selective institutions. While the 2014 Rule emphasized that low-income and minority students who go to more elite institutions have better outcomes, it is difficult to know if that is because the institution has done something remarkable or unique, or because the selective admissions process already culls students who are less likely to succeed. Wealthy institutions that enroll small numbers of high-need students also have the ability to have devote significantly more resources to those students than an open-enrollment institution that serves large numbers of high-need students.

There are many reasons why a student might elect to attend a proprietary institution. For example, it is very possible that the insightful student selects a proprietary institution because of the more personalized learning experience and higher graduation rates than might be found at many public, open-enrollment institutions.^[28] Proprietary institutions are more likely to offer accelerated programs, pre-established course sequences, more flexible class schedules and delivery models, and more personalized student services.^[29] The Department is also aware of recent studies that conclude proprietary institutions are more responsive to labor market changes in comparison to community colleges, which may lead students to choose proprietary institutions over their local, public, two-year counterparts.^[30]

The GE regulations also unfairly target proprietary institutions, as explained in the NPRM, because if the D/E rates measure considered the total cost of education relative to graduate earnings, a number of GE programs offered by public institutions would fail the measure.^[31]

The low price of public, two-year colleges may mean that fewer students need to borrow to enroll at those schools, but lower borrowing rates may also be due to the fact that a lower proportion of community college students are Pell eligible, or financially independent students, as compared to students at proprietary institutions.^[32] Despite assertions that community colleges and proprietary institutions serve the same students, as stated above, the data reveal that proprietary institutions serve a much larger population of low-income, older, and minority students.^[33] It is important to consider that despite lower proportions of student borrowers, given the total size of many public institutions, those institutions leave many more borrowers with debt and pose a higher aggregate loan burden and non-repayment risk to students and taxpayers. For example, a public college with 30,000 students and a 17 percent borrowing rate will produce 5,100 borrowers whereas a proprietary institution that serves 500 students and has a 90 percent borrowing rate will produce 450 borrowers. The same is true for small private, non-profit colleges that may have a higher percentage of students who need to borrow to pay tuition, but based on a small total student population, produce fewer total borrowers than public institutions that serve large numbers of students.

Unaffordable student loan debt is an issue across all sectors, including public institutions. The 2015 follow-up to the 1995-96 and 2003-04 Beginning Postsecondary Survey showed that despite the lower percentage of students who borrow at community colleges, among those who do borrow, their debts may be debilitating. For example, among borrowers who enrolled at community colleges in the 2003-04 cohort, twelve years later not only did they have a larger outstanding debt ($21,000) than students who enrolled at proprietary institutions ($14,600), but the level of debt held represented 90 percent of the original loan balance for students who enrolled at community colleges and 82 percent for those who enrolled at proprietary institutions.^[34] Therefore, it is as important for students at non-GE institutions or who are enrolled in non-GE programs to understand their likely earning outcomes so that they can borrow at a level that will not leave them struggling for decades after graduation.

Also, the Department is concerned that some community colleges do not participate in the Federal Student Loan programs because of concerns that high default rates would end the institution's participation in the Pell grant program.^[35] According to data from FSA, 38 community colleges do not participate in the loan programs. While this may be beneficial to students, it may also have a number of unintended consequences, including necessitating students to use more expensive forms of credit—such as credit cards and payday loans—to pay their tuition and fees. Or it may prevent low-income students from having access to higher education at lower cost institutions. An institution that elects to prevent students from taking Federal student loans will automatically pass the D/E rates measure, even if there are no earnings benefits associated with program completion. In some instances, the student may be better off in the long run by borrowing to attend a program he or she is more likely to complete, or that provides a more personalized experience, or that leads to a higher paying job. Despite the Department's interest in reducing student debt levels, it is noteworthy that a recent study showed that increased borrowing among community colleges may have a positive impact on completion and transfer to four-year institutions.^[36]

Student enrollment and borrowing decisions are as complex as the decisions that graduates make about where they want to work, what they want to do for a living, and how many hours a week they want to work. Until the Department has more sophisticated analytical tools that take into account the many variables other than institutional quality that impact both cost and outcomes, it is inappropriate to develop a scheme that imposes high-stakes sanctions without understanding the longer term impact of those sanctions on students and the production of ample workers for occupations that may pay lower wages but are in high demand (such as cosmetology, culinary arts, allied health, social work, and early childhood education).

While some commenters suggested that any institution could ensure that they will pass the D/E rates measure by lowering tuition, such a view oversimplifies college financing realities. In addition to the lack of direct taxpayer subsidies, proprietary institutions may have a higher per-student delivery cost since CTE-focused education can be four or five times more expensive to administer than liberal arts or general studies education.^[37] During times of high enrollment pressure or constrained resources, community colleges tend to reduce the number of vocational programs offered so that they can serve a large number of students in lower-cost general studies and liberal arts programs.^[38] In addition, as noted by Shulock, Lewis, and Tan, comprehensive institutions have the added benefit of cross-subsidizing higher cost CTE programs with low-cost general studies programs that typically enroll larger numbers of students.^[39] Since proprietary institutions are, for the most part, not permitted to offer lower cost general studies programs, the full cost of providing CTE is paid by the student without the benefit of cross-subsidizations from other students enrolled in lower-cost programs.

Therefore, the Department agrees with the commenter who stated that by focusing on GE programs, the Department has ignored worse outcomes generated by other programs. For example, as explained in the NPRM under “Covered Institutions and Programs,” numerous researchers have emphasized the importance of picking the right major in order to optimize earnings.^[40] According to Holzer and Baum's 2017 publication, community college liberal arts and general studies degrees have no market value for the majority of students who earn them, but the students will never transfer to a four-year institution.^[41] Nonetheless, these programs, and more at the baccalaureate level, were not covered by the GE regulations.

According to a 2018 Q3 breakdown of FSA's federally serviced portfolio, 24 percent of the dollars in the portfolio, or $272 billion, are in IDR plans that are current, but negatively amortizing. This substantial percentage of borrowers whose loans are growing rather than shrinking due to their enrollment in an IDR plan are of serious concern.^[42] This is a problem of a magnitude and importance that any action the Department takes must include all borrowers at all title IV participating institutions. Of course, participation in an IDR plan may not be a sign that a student's program was of low quality but could instead be a sign that the student borrowed recklessly or made lifestyle decisions that result in lower earnings.

Since the REPAYE program eliminates the income hardship test and allows any borrower to sign up for a student loan payment that is 10 percent of his or her income, it cannot be said that a borrower in an IDR plan is one who has been harmed by his or her program or institution. In some instances, borrowers may elect to pursue a lower paying job in order to benefit from IDR-derived loan forgiveness. Nonetheless, since so many students are enrolled in IDR programs, the Department believes that any transparency and accountability framework must apply to all title IV programs, which it plans to do through the expanded College Scorecard.

A Department review of the 2015 D/E rates shows that cosmetology and medical assisting programs were disproportionately represented among the programs that failed the D/E rates measure in the first year that D/E rates were calculated under the GE regulations.^[43] Yet both of these occupations are considered by the U.S. Department of Labor to be “bright outlook” occupations,^[44] suggesting that it is possible that GE-related program closures could reduce availability of CTE-focused programs needed to fill high-demand occupations. The Department agrees with the commenter who discussed the complicated patchwork of regulations that the Department has created, without any direction to do so by Congress. The 2015 Senate Task Force on Higher Education Regulation Report reinforces that point, and highlights the GE regulations as an example of the Department's “us[ing] the regulatory process to set its own policy agenda in the absence of any direction from Congress, and in the face of clear opposition to that policy from one house of Congress.” ^[45] By rescinding the GE regulations, we begin to correct that problem.

The Department disagrees that the BLS Job Openings and Labor Turnover Survey does not provide sufficient evidence to support the Department's assertion that many good jobs are currently unfilled, including jobs for which individuals could, in some cases, prepare for by completing a GE program. The Department pointed to the BLS survey to illustrate that the Department cannot predict the long-term impact of removing programs from title IV, including potential workforce shortages that could be caused by eliminating high-quality programs that fail the D/E rates measure for reasons beyond the control of the institution.

The Department disagrees with the commenters who said that the rescission of the GE regulations is arbitrary and capricious. Under the Administrative Procedure Act (APA), an agency “must show that there are good reasons for the new policy.” ^[46] However, “it need not demonstrate to a court's satisfaction that the reasons for the new policy are better than the reasons for the old one; it suffices that the new policy is permissible under the statute, that there are good reasons for it, and that the agency believes it to be better.” ^[47] (emphasis in original) Additionally, the Department provided ample evidence that any transparency and accountability framework must be expanded to include all title IV programs since student loan repayment rates are unacceptably low across all sectors of higher education and because a student may unknowingly select a non-GE program with poor outcomes because no data are available. If we want students to make informed decisions, then we need to provide information about all of the available options. Since the GE regulations cannot be expanded to include all institutions, and since negotiators could not come to consensus on a GE-like accountability and transparency framework that was substantiated by research and applicable to all title IV programs, the Department decided to take another approach.

The Department acknowledges evidence that students enrolled at proprietary institutions may be at higher risk for default and that, on average, students who attended a proprietary institution are more likely to default on their loans than students who enrolled at a community colleges. However, the Department provided ample data in the NPRM and in this document that higher defaults among students who enrolled a proprietary institution could be the result of these institutions serving higher risk students. A much higher proportion of students enrolled at proprietary colleges exhibit many more risk factors—such as being over 25, being a single parent, working full-time while being enrolled, being financially independent, and being Pell eligible—than students enrolled at other institutions, including community colleges.^[48]

The Department agrees that during the Great Recession, proprietary institutions likely grew too rapidly, and some have been accused of committing fraud, but the most rapid growth in the sector was by online institutions, where relatively few programs failed the D/E rates measure. During the Great Recession, many students sought relief by enrolling in college, and the Department does not deny that some institutions took advantage of that. However, there are other mechanisms, such state attorneys general, consumer protection agencies, civil legal proceedings, internal resolution arrangements, and borrower defense to repayment regulations that enable students to take action against institutions that have committed fraud. However, a failing outcome under the D/E rates measure in no way signals, demonstrates, or proves that the institutions committed fraud.

The Department is aware of research demonstrating that as enrollments in California proprietary institutions went down, there was a commensurate increase in enrollments at local community colleges.^[49] California is a State rich with community colleges, so it is not surprising that students were able to find alternatives to proprietary institutions. However, not all States and regions have as many options as those in California. In addition, a student who does not have the opportunity to attend a proprietary institution may be limited to a general studies program at a community college, which may disadvantage the student. Since, on average, graduation rates at proprietary institutions are higher than those at community colleges, a student may not be served if the lower-cost institution reduces the student's chances of completing his or her credential.

The Department agrees that some proprietary institutions serve students poorly and produce unimpressive results. However, there are institutions among all sectors that serve students poorly and produce unimpressive results, and yet the GE regulations do nothing to expose those programs or institutions or protect students from enrolling in them since the GE regulations are limited in their coverage. The point is not to ignore the legitimate challenges among institutions in the proprietary sector but is instead to expand the reach of a new accountability and transparency system to ensure that all students, regardless of institutional sector, can obtain information to inform their enrollment and borrowing decisions.

Changes: None.

Is there a need to define gainful employment?

Comments: One commenter stated that the Department must establish a definition for the term “gainful employment in a recognized occupation,” rather than leaving the term undefined.

Other commenters stated that the Department is violating the law by failing to differentiate between institutions that do and do not prepare students for gainful employment, and that by eliminating the GE regulations, the Department is no longer following the requirements of the HEA in differentiating between GE programs and non-GE programs.

Discussion: The Department does not agree that it needs to define the term “gainful employment” beyond what appears in statute. Since it was added to the HEA in 1968, the term “gainful employment” has been widely understood to be a descriptive term that differentiates between programs that prepare students for named occupations and those that educate students more generally in the liberal arts and humanities, including all degree programs offered by public and private, non-profit institutions.

Congress reaffirmed this interpretation when it added a provision to the 2008 Higher Education Opportunity Act (HEOA) that allowed a small number of proprietary institutions to offer baccalaureate degrees in liberal arts.^[50] Had Congress intended the term “gainful employment” to mean something other than a limitation on HEA section 102 institutions from offering programs that are not CTE-focused, it would not have needed to create a statutory exception to allow some HEA section 102 institutions to offer liberal arts programs.

Therefore, contrary to suggestions by commenters that the Department needs to develop a new definition in order to enforce the law or differentiate between GE and non-GE programs, the Department confirms that it, in fact, is enforcing the law as written and as intended, because it disallows proprietary institutions, other than those exempted by the above-mentioned provision of the HEOA, to offer general studies, liberal arts, humanities, or other programs not intended to prepare students for a named occupation. The Department will continue to enforce the law in this regard—in the same way it enforced it between 1968 and 2011.

In promulgating the 2014 Rule, the Department cited Senate debate in the 1960s as evidence that the GE regulations are consistent with congressional intent. The Senate Report accompanying the National Vocational Student Loan Insurance Act (NVSLI), Public Law 89-287, captured testimony delivered by University of Iowa professor Kenneth B. Hoyt that supported the “concept” of making loans available to students pursuing vocational training. He described findings from a sample of students whose earnings data were collected two years after completing their training, and based on those data, he concluded that “in terms of this sample of students, sufficient numbers were working for sufficient wages so as to make the concept of student loans to be [repaid] following graduation a reasonable approach to take.” ^[51]

The Senate report made no mention of how quickly the student would need to repay his or her loan, and it referred to the “concept” of student loan repayment rather than a particular repayment amortization term or a particular debt-to-earnings threshold. Moreover, the Senate report was focused on legislation other than the HEA and the conversation had a very different focus when Congress was contemplating the inclusion of proprietary institutions in all HEA programs.

What the Department neglected to include in its recounting of the early history of student loans, is that in 1972 when the National Vocational Student Loan Insurance Act (NVSLIA) was passed, Congress decided to incorporate vocational education programs into the HEA, by allowing their participation in the Educational Opportunity Grants as well as the student loan programs. Here the House conference report is clear that the new legislation “not only extends existing programs but creates exciting and long needed (sic) new ones. For the first time, the bill commits the Federal Government to the principle that every qualified high school student graduate, regardless of his family income, is entitled to higher education, whether in community colleges, vocational institutes or the traditional 4-year college or university.” ^[52] Vocational institutions in this context included proprietary colleges that would, for the first time ever, be eligible to participate in title IV grants as well as loans. The inclusion of proprietary schools in the HEA was an important step toward achieving the goals of providing equitable access to postsecondary education, for all students, regardless of whether their interests were in the traditional trades or vocations, or in typical degree programs.

The Department points out that Congress intends for all Federal student loan borrowers to repay their loans, not just those who borrow to attend “vocational training” programs.

However, Congress has elected to address concerns about unmanageable student loan debt by providing numerous extended repayment and income-driven repayment programs that reduce monthly and annual payments and provide loan forgiveness if, after 20 (or in some cases 25) years of income-driven repayment, an outstanding loan balance remains.

While the Department agrees that some of these repayment programs lead to undesirable outcomes for borrowers and taxpayers, in that they allow students to accumulate more debt (through negative amortization) rather than paying down their original student loan balances, the intent of Congress is clear. In fact, in introducing the Income Dependent Educational Assistance (IDEA) Act, which ultimately became the income-based repayment (IBR) program in the College Cost Reduction and Access Act of 2007 (CCRAA), Congressman Tom Petri (R-WI) stated:

Unfortunately, little has been done by way of providing more flexible repayment options for borrowers after graduation. Traditionally it has been expected that the borrower will pay the amortized loan over a standard period, usually 10 years, with the same repayment amount on day one as on the last day. However, this model of repayment fails to take into account that students often face periods of significant unemployment or underemployment during the first years after leaving college . . . I believe the IDEA Act does just that. This legislation would allow any Stafford loan borrower the ability to consolidate into a direct IDEA loan with a repayment schedule that corresponds to the borrower's income once in repayment. This new schedule requires regular payments; however, it ensures that such payments reflect the borrowers' capacity to repay under their current income status. This feature would be particularly useful for those pursuing lower-income, public-service careers. It also would help relieve some of the stress that borrowers face during periods of unemployment or underemployment following graduation.^[53]

Support for income-driven repayment during the 2007 HEA reauthorization was bipartisan, with Congressman George Miller (D-CA) stating that IBR was created because “knowing that they will face a mountain of debt after graduation, some students feel compelled to major in areas that will lead to a high-paying career. The hope is that income-based repayment will encourage students to pursue their real interests, even if careers in the major of their choice don't provide a high income.” ^[54]

Congressional support for IBR in the CCRAA in 2007, and for the Pay As You Earn (PAYE) income-driven repayment program in 2012, makes it clear that Congress does not wish for a student to feel compelled to select the highest paying major or job, to select the lowest cost educational opportunity, or to abandon his or her interests in lower-paying careers, such as public service careers, in order to meet student loan repayment obligations under the standard, 10-year repayment plan. Therefore, the Department's original determination the GE regulations are based upon or align with congressional intent was based on an incomplete review of the legislative record.

It should have been clear to the Department that the GE regulations did not comport with congressional intent when a bipartisan group of 113 Members of the House of Representatives, led by Congressman Alcee Hastings (D-FL), sent a letter in 2011 to President Obama asking him to withdraw the GE regulations.^[55] Further, the Department should have noted that the House of Representatives passed House Amendment 94 to House Resolution 1, the Disaster Relief Appropriations Bill of 2013, with a vote of 289 to 136.^[56] This amendment would have prohibited the Department from implementing the 2011 GE rule. Although the amendment was not included in the final bill, the amendment should have given the Department pause before claiming that the GE regulations were consistent with Congress' intent.

Despite numerous reauthorizations of the HEA between 1964 and 2008, Congress never attempted to define “gainful employment” based on a mathematical formula nor did it attempt to define the term using threshold debt-to-earnings ratios. Congress never attempted to prohibit students who attended GE programs from participating in IDR programs. In addition, the GE regulations were also identified in 2015 by the bipartisan Senate Task Force on Higher Education Regulation as a glaring example of the Department's “increasing appetite” for regulation.^[57]

Despite previous assertions, the Department now recognizes that it had incorrectly described congressional intent and engaged in regulatory overreach, as discussed throughout these final regulations, and for those reasons, and the others described in the NPRM and these final regulations, it is rescinding the GE regulations.

Changes: None.

Protecting Students

Comments: A number of commenters disagreed with the Department's decision to rescind the GE regulations, arguing that minority, low-income, adult, and veteran students are particularly vulnerable and, therefore, need additional protections from unscrupulous institutions and from programs with inferior outcomes, as well as to eliminate waste, fraud, and abuse.

Discussion: The Department shares the concern of commenters who highlighted the need to protect low-income students and taxpayers from programs with poor outcomes, and from waste, fraud, and abuse. However, we do not believe the GE regulations are an effective tool for either of those purposes.

First, the GE regulations do not accurately identify programs with poor outcomes. Many programs that had poor earnings outcomes passed the D/E rates measure due to large public subsidies that reduce the cost of enrollment to students. At the same time, programs that resulted in much higher earnings failed the D/E rates measure since the lack of public subsidies required the students to pay the full cost.^[58] The Department believes that the best way to protect all students is to acknowledge that they select their college and major based on a variety of factors, but provide clear and accurate information about debt and earnings to enable them to compare likely outcomes among the institutions and programs they are considering.

Second, although the Department acknowledges that it plays an important financial stewardship role, and has the responsibility of reducing waste, fraud, and abuse, the GE regulations did not support that goal. Many programs are not subject to the GE regulations, so the regulation would play no role in preventing waste, fraud, and abuse among those programs. The Department does not agree that by charging students for the full cost of their education, rather than accepting direct appropriations and other taxpayer subsidies, is an act of waste, fraud, or abuse. Were that the case, then the Department would need to apply the D/E rates measure to all private institutions, including private, non-profit institutions, since those institutions generally have the highest annual tuition, including for programs that result in modest earnings.

The Department is committed to ensuring that students are provided with accurate outcomes data. All students should be able to view accurate and unbiased outcomes data from a reliable source. The Department seeks to make it much more difficult for institutions to mislead students by making reliable data readily available to all students about the institutions they are considering attending or are attending.

There are many instances of fraud that would never be detected by the GE regulations, either because the programs or institutions are not subject to the GE regulations or student earnings are sufficient to mask misrepresentation that took place. Therefore, complacency based on the mistaken belief that the GE regulations will obviate the need for other efforts to detect and eliminate waste, fraud, and abuse could have serious consequences.

The Department acknowledges that it plays an important financial stewardship role, and has the responsibility of reducing waste, fraud, and abuse. However, the GE regulations did not support that goal.

Moreover, the GE regulations do not necessarily identify instances of fraud or abuse since programs designed to prepare, for example, teachers, community health workers, and allied health professionals may result in low wages simply because the prevailing wages in those fields are low. Therefore, a program could fail the D/E rates measure not because of fraudulent or abusive practices on the part of institutions, but because a number of high-demand occupations pay low wages, especially in the early years of employment, or because in some occupations there is an induction period of several years before a graduate can be fully licensed or be paid at the level of experienced professionals.

There are ample examples of institutions that committed acts of fraud that would never be detected by the D/E rates measure. For example, the Nebraska Attorney General alleges that Bellevue University misrepresented the truth about the accreditation of its nursing program,^[59] City Colleges of Chicago inflated their graduation rates,^[60] Maricopa Community College was found guilty of falsifying student volunteer hours to allow students to receive an education award through the Americorps program,^[61] and a number of law schools admitted to inflating job placement rates ^[62] in order to attract more students. Yet the GE regulations would identify none of these acts of misrepresentation.

The Department will continue to employ its usual fraud prevention mechanisms, such as program reviews, to identify institutions that are not abiding by title IV rules and regulations. In addition, it will continue to rely on States to execute their consumer protection functions and accrediting agencies to evaluate program quality so that the regulatory triad will retain its importance and shared responsibility in the oversight of institutions of higher education. Finally, the Department seeks to make it much more difficult for institutions to mislead students by ensuring that all students are able to view accurate and unbiased outcomes data from a reliable source, and the Department will continue to work with accreditors to try to identify and stop institutions that are reporting false outcomes data.

Changes: None.

Accountability

Comments: Some commenters disagreed with the Department's proposal to rescind the GE regulations, arguing that the GE regulations provide the only standard by which programs might be held accountable for outcomes. Another commenter stated that by eliminating the GE regulations, proprietary institutions would be held to a lower standard than non-GE institutions.

One commenter acknowledged that CDRs currently serve as an accountability standard for all institutions of higher education, but expressed concern that defaults are not an accurate indicator of program quality or an accurate measure of a student's or taxpayer's return on investment.

Another commenter stated that research shows that income increases with the level of degree earned. For example, the research found that students with an associate's degree saw their quarterly incomes increase by more than $2,300 for women and nearly $1,500 for men, while those with a short-term certificate saw an increase of only around $300 per quarter. The commenter also cited a study finding that among certificate holders, workers in female-dominated occupations (healthcare and education) earned less than those in male dominated occupations (technology-based).^[63]

Discussion: The Department strongly disagrees with the commenter who suggested that by eliminating the GE regulations, there will be no more program-level accountability measures. It is the role of accreditors and States, not the Department, to evaluate program quality, and, in some instances, specialized or programmatic accreditors establish quality assurance measures, enrollment caps, and licensure pass rates that determine whether or not specific programs will continue to be accredited. The Department will continue to rely on accreditors and State authorizing agencies to evaluate program quality.

The Department also does not agree with the commenters who argued that by eliminating the GE regulations, proprietary institutions would be held to a lower standard than non-GE institutions. In addition to meeting CDR requirements like all institutions and financial responsibility standards like all non-public institutions, proprietary institutions must also meet requirements that limit title IV revenue to 90 percent of total revenue (the 90-10 Rule). The requirements regarding annual audits and the types of jobs Federal Work Study students can be placed in are also stricter for proprietary institutions. So, they remain subject to additional regulatory requirements.

As pointed out by at least one commenter, CDRs are one of the metrics that Congress has established to determine continuing eligibility for an institution, including proprietary institutions, to participate in title IV programs. We agree that CDRs are misleading indicators of program quality or the current status and risk associated with the outstanding Federal student loan portfolio. As noted earlier in this document, updated repayment rate data revealed, in January 2017, that less than half of all borrowers were paying down a dollar of principal by their third year of repayment, and more recent portfolio analysis has revealed that of the nearly $1.2 trillion in outstanding student loans, only 24 percent, or $298 billion, are in a positive repayment status, meaning that interest and principal are being paid down. The remaining loans are in post-enrollment grace, default, forbearance, deferment, or negative amortization due to income-driven repayment, and 43 percent, or $505 billion, are in distress, as previously mentioned.^[64] Despite these grim statistics, it is noteworthy that the most recent CDR is only 10.8 percent (the 2018 three-year CDR for the 2015 cohort).^[65] Accordingly, although the Department will continue to enforce the law by restricting title IV eligibility to those institutions, including proprietary institutions, that pass the CDR test, it also seeks to expand transparency and market-based accountability through the College Scorecard.

Regarding the comment about credential inflation, the 768 programs that failed the D/E rates measure based on the 2015 D/E rates published in 2017, more than 100 were medical assisting or similar programs and more than 90 were cosmetology/barbering programs. This suggests that these occupations may not pay a wage that is commensurate with the educational requirements for licensure or certification, but institutions do not determine or set those requirements. States and occupational licensing boards or credentialing organizations establish those requirements.

The Department agrees that the financial rewards associated with a postsecondary credential, in general, increase as the credential level increases. However, there are bachelor's, master's, and doctoral degree programs that result in relatively low earnings and that require borrowers to rely on income-driven repayment. In fact, some researchers have pointed out that it is recipients of graduate degrees who are in greatest need of, and who will benefit most from, these programs.^[66] Therefore, the Department continues to believe that the best way to expand transparency and accountability to all students is to expand the College Scorecard to the program-level for all categories (GE and non-GE) of title IV programs.

Changes: None.

Which institutions should be included?

Comments: A number of commenters stated that they fully support the original intent of the GE regulations and that schools must be held accountable to provide equitable value to their students. However, others asserted that given the limited reach of the GE regulations, students may not have had sufficient information to accurately compare the outcomes of a GE program to a non-GE program that was not subject to the regulations. These commenters agreed with the Department that the 2014 Rule should be rescinded.

Other commenters noted that they supported the GE regulations, but indicated that all schools and programs, including proprietary institutions and non-profit institutions, should be held to the same standards and requirements. Those commenters were split on whether the Department should expand the regulations to include all institutions or rescind the regulations.

Several commenters took the position that any new regulations, whether they require a specific outcome threshold, additional disclosures, or overall transparency, should apply equally to all institutions. Of those commenters who favored uniform application of new regulations, some voiced support for a disclosure-only protocol that would provide students with program-level data about all participating institutions regardless of the type of control.

Discussion: The Department agrees that the same standards and reporting requirements should apply to all institutions, regardless of tax status. However, the Department could not simply expand the GE regulations to include all title IV programs since the term “gainful employment” is found only in section 102 of the HEA, which refers to vocational institutions and programs (meaning non-degree programs at non-profit and public institutions and all programs at proprietary institutions). Therefore, there was no way to expand the GE regulations to apply to all institutions. Moreover, although the negotiating committee considered adopting a “GE-like” solution that could be applied to all institutions, the negotiators were unable to reach consensus on an accurate, valid, and reliable outcomes standard that could serve as the basis for an appropriate and useful accountability and transparency framework for all title IV participating programs.

The Department agrees with the commenters that stated that the most effective method to increase accountability and transparency, under current law, for all programs is through a disclosure-only protocol, and it plans to do so using the College Scorecard to make program-level data readily available and in a format that enables easy comparative analysis. Only when students can consider comparable information about all of the institutions and programs they are considering, and that are available to them, can students begin to make data-driven decisions. Part of our goal is to end information asymmetry between institutions and students.

Changes: None.

Location Matters

Comments: One commenter noted that while the Department correctly cites research showing that most students do, in fact, stay close to home for college, the commenter disagrees with the assertion made in the NPRM that eliminating a failing GE program could eliminate the opportunity for a student to gain a credential if a passing program is located farther away. The commenter suggested that this research should not be used as a justification for eliminating the 2014 Rule, but rather to support keeping the GE regulations in effect in order to protect consumers.

Discussion: The Department does not believe that the NPRM mischaracterizes these research findings. The Department continues to believe that since location is important in influencing student enrollment decisions, a less expensive option may be of no benefit for a student who would need to travel too far from home to enroll in it. In addition, the 2015 GE data provides numerous examples of programs that pass the D/E rates measure because they are heavily taxpayer subsidized, even though they result in earnings that are substantially less than the earnings associated with programs provided by proprietary institutions that charge students the full price of educational delivery.

The Department stands by its original point, which is that location matters and that the elimination of a program that fails the D/E rates measure may not result in better long-term outcomes for students if another option doesn't exist in that place. On the other hand, a student who has only one option may decide, when better informed about debt and earnings, that it is best to forfeit that option and find a different workforce preparation pathway. The Department believes that all institutions should provide high-quality educational options to students, but without public subsidies, some of those options could result in higher tuition and fees and increased borrowing.

Regardless of whether information about program outcomes encourages program improvements, encourages institutional selectivity, or encourages students to pursue other kinds of career preparation, the Department believes that, especially when a student has very limited institutional or programmatic options, he or she needs access to data about all available options to better inform enrollment and borrowing decisions.

We are aware that the researcher who wrote the paper about the role of location in student enrollment decisions disagrees with our position on the GE regulations, and does not wish his research to be used to support our conclusions. However, we did not misrepresent his research findings and still believe that they are relevant in explaining that students with limited options in their local geographic area could be better off attending a program that results in debt but also elevates wages, as opposed to attending no postsecondary program at all.

We continue to believe that if the program in which a student is interested in enrolling loses title IV eligibility under the D/E rates measure, and there are no other options to enroll in that program within a reasonable commuting distance, the student may not be well served by the elimination of the program, even if the student would have required more than 10 years to repay their student loan debt.

Changes: None.

Proprietary Institution Outcomes

Comments: Commenters cited a number of studies on outcomes at proprietary institutions, in support of their position that the GE regulations should not be rescinded.

One commenter provided an appendix with research citations believed to be relevant to the GE regulations. The commenter referenced research by Cellini and Turner that found that students who attend proprietary certificate programs experience small, statistically insignificant gains in annual earnings.^[67] Chou, Looney, and Watson found that proprietary schools have relatively poor cohort loan repayment rates, with almost no schools in that sector having a repayment rate above 20 percent.^[68] Looney and Yannelis found that between 2000 and 2011 there was substantial growth in both proprietary college enrollment and student loan default rates.^[69] Armona et al. found that those who enroll in for-profit four-year institutions have the worst outcomes, including more educational debt, worse labor market outcomes, and higher default rates than students attending similarly selective public institutions.^[70]

Research citations in the appendix also included work by Darolia et al. who found that employers were less likely to hire applicants with degrees from proprietary institutions, even compared to those with no degrees.^[71] Chakrabarti and Jiang found that attending a proprietary college yields earnings that are 17 percent lower that earnings of those who attend private, not-for-profit four-year colleges.^[72]

Commenters stated that in the 2014 Rule, the Department showed that in 27 percent of GE programs, the average graduate had an income lower than a full-time worker making the Federal minimum wage. The commenters also noted a study demonstrating that since 2014, 350,000 students graduated from certain GE programs with nearly $7.5 billion in student debt.

Discussion: The Department appreciates the bibliography provided by the commenter and agrees that these papers conclude that students who attend proprietary institutions, in many instances, have outcomes that are inferior to students who attend other institutions. However, the Department believes that our analysis of the outstanding student loan portfolio demonstrates that poor outcomes are not limited to these institutions or the small number, relative to total postsecondary enrollment, of students who attend them. For this reason, the Department believes that it must implement a transparency and accountability system that applies equally to all title IV programs, and that enables all students to make informed enrollment and borrowing decisions.

The Department is aware of the survey results showing that many employers “do not prefer” graduates of proprietary institutions,^[73] or may be less likely to interview a candidate who completed an online degree at one of the well-known, large, online proprietary institutions.^[74] However, the “do not prefer” study shows that employers similarly did not prefer to hire community college graduates over proprietary school graduates. And while employers may have been less likely to interview a candidate who attended one of the large, online, proprietary institutions, there was not an observed bias against graduates of smaller, ground-based proprietary institutions. It is difficult to know if employers were skeptical of large, online proprietary institutions because of negative experiences with prior employees, or because of negative media coverage of, and political opposition to, well-known proprietary schools.

The Department also believes that many of the studies cited have serious limitations that, in some cases, reduce the validity and reliability of their conclusions. For example, a Cellini study found that proprietary institutions are more expensive than community colleges, when tuition as well as opportunity cost is considered.^[75] However, Cellini assumed in this study that it takes students the same amount of time to complete programs at proprietary institutions and community colleges, even though in subsequent publications she cites research showing that students at proprietary institutions tend to complete at higher rates and more quickly than students at community colleges. Since opportunity cost could reasonably be seen as a considerable part of the expense of attending college for adult learners who must leave the workforce or reduce the number of hours worked in order to attend college, the ability to accelerate completion could generate substantial savings compared to a lower cost program that takes longer to complete.

In her more recent work to compare pre- and post-earnings of community college and proprietary certificate programs, Cellini admits that the Great Recession could have introduced bias into her research results, and that the kinds of certificates offered by community colleges and proprietary institutions differ.^[76] In other words, she was comparing what employees earn in fields that may pay very different prevailing wages. She also admits that her methodology for creating demographically matched comparison groups relied on the use of zip codes and birthdates, but every one of the same age in the same zip code is not otherwise socioeconomically and demographically matched. Moreover, she relied on a data set made available exclusively to her, meaning that it is not available for full peer review. Without the advantages of peer review and the ability of other researchers to replicate or challenge her findings, it is difficult to know how credible they are. That said, she concluded in her report that when it came to cosmetology certificates, it appeared that those who completed those certificates at proprietary institutions had higher earnings gains than those who completed those certificates at community colleges, which she attributes to the number of proprietary cosmetology colleges that are affiliated with high-end salons and channel graduates to jobs at those salons.

What her study fails to show, however, are earnings gains realized by students who are unable to enroll in the career and technical education program of their choice at a public institution, and instead enroll in a general studies program. Importantly, her study compared the outcomes of students who enrolled in CTE programs at public and proprietary institutions, but the study did not consider the outcomes of students who are unable to enroll in the career and technical education programs of their choice at a public institution, and instead enroll in a general studies program.

What matters to a student may not be how the earnings compare between a CTE program offered by a public and private institution, but instead how the earnings compare between the CTE program available at the private institution, and the general studies program available at the public institution. We believe that the best way to inform student choice is by providing comparable information about all of the choices a student might have. This is another reason why we are rescinding the GE regulations and proposing to expand the College Scorecard.

The Department agrees with the commenter that the GE regulations could have the unintended consequence of creating workforce shortages in occupations of high societal value. For example, according to the Department of Labor's ONET database, there were 634,000 medical assistants employed in the United States in 2016, with the projected need of almost 95,000 additional workers in this field by 2026. This makes medical assisting a “bright outlook” occupation, meaning that it will experience fast growth in the coming decade.

Unfortunately, medical assisting is also a field that had a median pay of $33,610 per year in 2018.^[77] Yet, medical assistant program costs are rising, possibly because only medical assistants who complete a program accredited by the Accrediting Board of Health Education Schools (ABHES) or the Commission on Accreditation of Allied Health Education Programs (CAAHEP) are eligible to sit for the Certified Medical Assisting exam. Thus, programmatic accreditation may be the driver of escalating program costs given the requirements that accreditors impose on educational institutions.

It is unclear whether the relatively large number of medical assisting programs that failed the D/E rates measure did so because they are low-quality programs, they are overly expensive, high workforce demand in general results in a larger number of these programs (thus the higher failure rate is proportional to the larger number of programs offered) or if the educational requirements for entry to the field are disproportionately high relative to the wages employers pay.

The medical assisting programs that failed the D/E rates measure may be overly expensive or of low quality. However, medical assistant programs prepare students to work in a field necessary to keep our healthcare system working and where employment opportunities are readily available, although they generate low wages. While the Department agrees that a student could benefit from having access to a low-cost medical assisting program, such as by attending a program at a community college, or apprenticeships, National Center for Educational Statistics (NCES) data show that of the 103,589 medical assistants who completed programs in 2013, 84,463 or 82 percent completed programs at proprietary institutions.^[78]

In response to the commenters who raised concerns about the 350,000 students who graduated from career education programs with $7.5 billion in debt, the Department shares the concern that many students take on too much debt. However, by dividing the total debt by the number of students, the average debt for each of the 350,000 students in that group would be $21,429, which is actually lower than the average loan debt for the Class of 2017 ($39,400) ^[79] and the Class of 2016 ($37,172).^[80] Because proprietary institutions confer associate, baccalaureate, graduate, and professional degrees, comparisons of student debt levels must include not just community colleges, but also four-year and graduate institutions.

In response to the comment citing the Department's statistic from the 2014 Rule that 27 percent of GE programs resulted in lower earnings than those of a full-time worker who earns the Federal minimum wage, the Department has further considered this statistic and determined that it was based on an invalid comparison. In calculating annual earnings for minimum-wage workers, the Department assumed that minimum wage workers all work forty hours per week, fifty-two weeks per year.

However, employment statistics for low-skilled workers show that unemployment is higher among this group than others, making the full-time, year-round employment assumption overly generous. This calculation did not include part-time workers or unemployed workers in proportion to actual employment rates, but instead considered only the wage that would be earned by those who work full time. Consider that in 2017, the real median earnings for males was $44,408 and for females was $31,610, and the real median earnings for males working full time, year-round, was $52,146 and for females was $41,977. These data make clear the impact of part-time work on wages, and do not include individuals who are not in the workforce, either by choice or not.

On the other hand, the D/E rates calculation includes, not only full-time workers, but also part-time workers and those who are not in the workforce, perhaps by choice in order to raise children or care for an elderly family member. Among the 10,727,000 married couples with children under the age of 6, there are 3,811,000 in which the husband works but the wife does not but only 339,000 in which the wife works but the husband does not.^[81] This demonstrates the significant impact that age and gender have on workforce participation.^[82]

Additionally, as pointed out by Witteveen and Attwell in their 2017 analysis of Beginning Postsecondary Survey (BPS) data, institutional selectivity and college major, as well as student gender and socioeconomic status, have a significant impact on earnings outcomes.^[83]

If the D/E rates measure, like the projected earnings of minimum wage workers, included only full-time workers, it is likely that the comparison would have yielded very different outcomes.

Changes: None.

D/E Rates Thresholds and Sanctions

Comments: A number of commenters supported the Department's proposal to rescind the GE regulations due to a lack of evidence that an 8 percent debt-to-income ratio sufficiently differentiates between high-quality and low-quality, or between effective and ineffective, programs. These commenters agreed that the lack of an empirical basis for the 8 percent threshold makes it inappropriate to use in determining whether or not a program should be allowed to continue participating in title IV programs. One commenter stated that currently there is not enough data to identify appropriate sanctions for any institution and that this was evident when the 2014 Rule was being negotiated.

Other commenters agreed with the Department that the GE regulations have several shortcomings, including the D/E rates thresholds, but argued that there are aspects of the GE regulatory framework that provide a reasonable and simple methodology for determining whether a program is preparing students for gainful employment. The commenter offered alternative D/E rates and thresholds for consideration, including using a 10% debt-to-income threshold with a 10-year repayment term or a 15% or 20% debt-to-income thresholds. Several commenters recommended that the Department revise the GE regulations rather than eliminate them. Another commenter asserted that sanctions should not apply only to proprietary institutions.

One commenter argued that while there is no justification for eliminating the rule, changes should be made to the measures and thresholds, with the Secretary given discretion to provide relief to programs experiencing the effects of lasting economic trends that might distort the measure or limit its reliability.

Several commenters stated that they thought efforts to reduce an institution's regulatory burden should be made, while also maintaining sanctions for poorly performing programs or, conversely, while maintaining the GE regulations. One commenter acknowledged the challenges associated with the GE regulations, but argued that these challenges are not insurmountable and that low-performing GE programs should be identified through some means and be subject to sanctions.

One commenter stated that while they understood the validity of the D/E rates measure was questionable, without it in place, low-income students would continue to be able to enroll in programs that are at high risk of not providing the students the education they deserve.

At least two commenters stated that the Department only addresses its concerns with the annual D/E rates metric but did not provide any justification for rescinding the discretionary D/E rates measure.

A few commenters were strongly in favor of retaining sanctions, including the loss of title IV program eligibility, for those programs with failing D/E rates. One of these commenters stressed that taxpayers should not pay for educational programs that “don't work well when there are plenty of programs that do work well,” and that it is the government's job to “provide regulations that put the right incentives in place to protect consumers.” Another commenter writing in favor of retaining an accountability framework inclusive of program sanctions recommended that the Department leave the 2014 Rule in place as currently written. The commenter offered that students enrolled at proprietary institutions and in other GE programs have lower employment and earnings gains than students in similar programs in other sectors and are saddled with greater debt for these high-cost programs that they cannot reasonably be expected to repay. Several commenters pointed to studies that control for student demographics, and still find that students in for-profit GE programs have lower employment and earnings outcomes than students in similar programs in other sectors.

Many commenters pointed to a blog post written by Sandy Baum as evidence that the Department mischaracterized research that she and Schwartz published as evidence that the 8 percent D/E rates threshold was an inappropriate or invalid threshold to use in establishing student borrowing limits.

Discussion: The Department appreciates support from the many commenters who agreed that the 8 percent threshold lacks sufficient accuracy and validity to serve as a high-stakes standard that determines whether or not a program may continue to participate in title IV programs. The Department continues to believe that our more careful recent review of the Baum and Schwartz paper confirms that the 8 percent D/E rates threshold is not appropriate to use in determining a program's continuing eligibility in title IV programs. The Department appreciates Dr. Baum's confirmation that the Department accurately reported the findings of her 2006 paper, including the recommendation that the 8 percent debt-to-income standard is a mortgage standard and one that “has no particular merit or justification” for use in establishing student borrowing limits.^[84] The Department understands that Dr. Baum does not wish her paper to be used to support the Department's decision to rescind the GE regulations; however, the Department has never asserted that Dr. Baum supports our decision. Instead, the Department has pointed out that the source it referenced to justify the 8 percent threshold in 2010 and in 2014 is her paper, which states explicitly that 8 percent “has no particular merit or justification” for establishing student borrowing limits. Therefore, the Department has no empirical basis for the 8 percent threshold and will, therefore, no longer use it to determine title IV program eligibility. The Department also recognizes that in its 2011 GE regulation, it used a different set of thresholds that included 12 percent as the passing rate rather than 8 percent. This further demonstrates the absence of a reasoned methodology for distinguishing between passing and failing programs.

In the 2014 Rule, the Department failed to provide a sufficient, objective, and reliable basis for the 20 percent threshold for the debt-to-discretionary income standard. However, in 2015, the Department promulgated regulations to establish a new income-driven student loan repayment program (REPAYE), and it established 10 percent as the debt-to-discretionary income threshold that is used to determine a borrower's monthly payment obligation.^[85] The REPAYE program renders the 20 percent debt-to-discretionary income threshold in the 2014 Rule obsolete since no borrower would ever be required to pay more than 10 percent of their discretionary income. Instead, REPAYE provides a longer repayment period at the 10 percent payment level in order to help borrowers manage their repayment obligations, and after 20 to 25 years (depending upon the level of the credential earned), the remaining debt is forgiven and considered taxable income.^[86]

The Department agrees with the commenter who stated that all institutions should be held to the same standards. This is why we attempted, through negotiated rulemaking, to identify thresholds that could be used to determine the continuing eligibility for all title IV programs. However, despite robust discussion and the Department's willingness to consider the use of other metrics to determine program outcomes, including a proposal from one negotiator to use a one-to-one ratio to report debt-to-earnings, there was no consensus around that proposal. Similarly, negotiators could not identify a threshold that they agreed should be used to determine title IV eligibility for all programs.

The Department appreciates the recommendations from commenters to establish a new threshold for triggering sanctions, but we are not persuaded that any of those recommendations have merit, especially because borrowers have multiple student loan repayment options that use different formulas for determining how much a borrower must pay each month. None of the sanction recommendations properly accounted for differences in repayment rates among the available repayment options.

The Department agrees that students and taxpayers should not continue investing in failing programs. However, the Department does not believe that the D/E rates measure accurately distinguishes between programs that “do or do not work” since the majority of title IV programs are not subject to the GE regulations. Also, it is difficult to argue that a program resulting in higher earnings does not work, simply because the cost of attending that program is paid by students rather than taxpayers, which results in higher student loan debt. The Department also believes that providing direct appropriations and other tax subsidies to low-value programs creates the same financial risks to taxpayers as student loans. Therefore, any future sanctions should also take into account the amount that taxpayers contribute through direct appropriations and preferential tax benefits to programs that do not result in better student or societal outcomes.

Our review of the 2015 D/E rates reveals that a number of programs whose graduates have exceptionally low earnings passed the D/E rates measure simply because taxpayers provide substantial subsidies to students enrolled in those programs in order to reduce the portion of program costs that students pay through tuition. For example, cosmetology programs offered by non-profit institutions in Puerto Rico, such as at Institucion Chaviano de Mayaguez and Leston College, resulted in the lowest earnings among any GE programs in that field.^[87] Yet, these programs passed the D/E rates measure because the taxpayer carried most of the burden of paying the costs of program delivery. Just because the taxpayer covered the majority of the cost of the program, does not change the fact that its graduates earn exceptionally low wages. Even if these students took no loans, the taxpayer's contributions may not have been well spent and will not necessarily generate returns commensurate with investment.

The Department is not surprised that students who attend proprietary institutions accumulate more debt than those who attend public institutions because the same is also true of students who attend private, non-profit institutions versus public, non-profit institutions. In fact, national data indicate that students who attend proprietary institutions, which include four-year institutions and graduate institutions, accumulate less debt on average than those who attend private, non-profit institutions.^[88]

The Department also notes that a number of GE programs offered by public institutions did not meet the minimum cohort size and, therefore, did not report outcomes. For example, as of 2017-2018 award year, 14,476 of 18,184 GE programs, or 79.6%, at public institutions have fewer than 10 graduates.

Unable to demonstrate that the D/E rates measure is an accurate indicator of program quality and unable to identify an alternative threshold that is supported by empirical evidence, the Department is rescinding the GE regulations and plans to report directly to the public the median debt and earnings of program completers. This enables students, parents, and taxpayers to evaluate program value and make informed enrollment and investment decisions.

Perhaps, in time, researchers can develop evidence-based recommendations for thresholds and sanctions that take into account all of the factors that influence program outcomes. More accurate and valid thresholds must also take into account differences in earnings among workers in different fields, the societal benefits afforded by some lower-paying occupations, the educational qualifications demanded by employers (which may exceed the level of education technically required to do a particular job), and the education requirements associated with State or professional licensure or certification.

Since the Department is rescinding the GE regulations, it will no longer use arbitrary thresholds that lack an empirical basis to establish continuing title IV eligibility. However, through the expanded College Scorecard, students and taxpayers will, for the first time, have access to debt and earnings data for the graduates of all categories of title IV programs, which will help students, families, taxpayers, and institutions, determine which investments generate the highest return.

The Department clearly stated in the NPRM that neither it nor negotiators were able to identify a D/E metric that was sufficiently valid and accurate to serve as a high-stakes quality test or to become a new, non-congressionally mandated, eligibility criteria for title IV. Regardless of whether gross income or discretionary income forms the basis of the D/E calculation, the methodology is inaccurate and fails to control for the many other factors other than program quality that influence debt and earnings.

The Department does not agree that it can overlook the limitations of the GE regulations and instead rely on the Secretary to grant relief to institutions facing particular challenges or extenuating circumstances. While identifying a more accurate metric or formula for evaluating program quality may not be insurmountable, the Department does not currently have tools that can differentiate between outcomes that are the result of program quality and outcomes that are the result of institutional selectivity or student demographics.

Changes: None.

Concerns About the Validity and Complexity of the D/E Rates Calculation

Comments: A number of commenters agreed with the Department's decision to rescind the GE regulations due to inaccuracies in the D/E rates formula.

Many commenters agreed with the Department's proposal to rescind the GE regulations because the D/E rates calculation is overly complicated and not easily understood by students or parents, which led those commenters to state it would be unwise to continue using those rates to determine title IV eligibility.

Another commenter said that a study used to illustrate the impact of student demographics on earnings was inappropriate since it did not isolate graduates of GE programs or distinguish them from other individuals.

Discussion: The Department agrees that the D/E rates calculation is too complicated for many students and parents to understand how to translate D/E rates into a meaningful and useful data point.

The Department referenced College Board information in their Trends in Higher Education research series to substantiate our claim that earnings are impacted by a number of factors, including gender, race, geographic location, and socioeconomic status.^[89] The Department agrees that the research does not single out graduates of GE programs, but it need not do that to be relevant to the Department's concerns about the many factors other than institutional quality that can impact D/E rates. The data supports our position that earnings outcomes are influenced by a number of factors, which may include program quality.

Changes: None.

Amortization and Interest Rates

Comments: Among those who agreed with the Department that the GE regulations should be rescinded were commenters who were concerned about the use of amortization terms and interest rates that could have a significant impact on D/E rates outcomes.

A few commenters disagreed with the Department's position expressed in the NPRM that it is not appropriate to use an amortization period in the D/E rates calculation of less than 20 years for any undergraduate program or of less than 25 years for any graduate program, given that the REPAYE program provides 20- to 25-year amortization periods, depending upon the level of the credential earned. The commenters maintained that it is inappropriate to apply the 20- or 25-year amortization period associated with REPAYE to associates or certificate programs since those programs are shorter-term and should be less costly than four-year or graduate programs. However, another commenter agreed with the Department's position on the amortization period that should be used to calculate D/E rates for two-year and certificate programs, offering that though switching to a 20-year amortization period would allow some low-performing programs to pass the D/E rates measure, it is reasonable given that the Department offers a repayment plan of that length.

Another commenter strongly objected to the Department's statement in the NPRM that the problem of unaffordable debt levels has been ameliorated by the creation of IDR plans. The commenter asserted that IDR plans are not a solution to the problem of unaffordable for-profit educational programs and that there is no evidence to suggest IDR plans have improved the landscape of GE programs. One commenter contended that PAYE, REPAYE, and other IDR plans set programs up to fail the D/E rates measure since these repayment plans often lower monthly payments to the point where the minimum payment consists only of interest or, in some cases, allows the loan to negatively amortize.

Discussion: The Department appreciates support from commenters who agree that it would be arbitrary for the Department to use an amortization term for the purpose of calculating D/E rates that differs from the amortization terms made available to borrowers under the law and the Department's REPAYE regulations. The Department agrees that it is desirable for students who completed shorter-term programs to repay their debts more quickly, but it is equally desirable for all borrowers to repay their debts over a standard 10-year repayment plan. However, Congress has created IDR plans to help borrowers manage debt and ensure that student loan payments will always be a fixed percent of discretionary income. For example, in the REPAYE program, introduced by the Department in 2015, the fixed percent of discretionary income is 10 percent.

The Department does not agree that IDR plans lead to a program's failure to meet the required D/E standard, since the D/E formula is a mathematical calculation and not a measure of the amount of debt borrowers are actually paying. However, the Department believes that student participation in IDR plans will negatively impact repayment rates, since it is possible that a student making the required payment is paying so little that the payment will not keep pace with accumulating interest. We share the commenter's concern about the impact of IDR plans on borrowers and outstanding debt, but IDR plans do not have an impact on calculating a program's D/E rate.

Changes: None.

Earnings Data and Tip-Based Occupations

Comments: Numerous commenters raised concerns that earnings data used to calculate D/E rates were not accurate or reliable for a number of reasons, including that SSA data excludes unreported tip income and some self-employment earnings. Several commenters noted that tip-based careers and commission-based employment may adversely impact a program's D/E rates. Others commented that since data collected by the SSA is used to administer the Social Security Act and not evaluate college or university performance, it should not be used to determine continuing title IV eligibility. Another commenter pointed out that SSA data cannot differentiate between wages earned by those working full time versus part time, including when part-time work is the option preferred by the program completer.

On the other hand, one commenter stated that the Department should not make accommodations for the underreporting of tipped income. The commenter argued that those who underreport tipped income are committing an illegal act and the Department should not protect those individuals.

Discussion: The Department agrees with the commenters' critiques of the D/E rates calculation and that institutions may not have the ability to control for the many variables that impact earnings. The Department does not believe that it should sanction institutions for aspects of student debt and earning outcomes that are outside of the institution's control. The Department provided detailed explanations regarding its concerns about the accuracy of the D/E rates formula in the NPRM, including that second- and third-year earnings do not accurately reflect long-term earnings associated with program completion; macro-economic conditions can have a significant impact on D/E rates, even if there are no changes in the program's content or quality; and prevailing wages may differ significantly from one occupation to the next and one part of the country to the next.

The Department also agrees that the exclusion of tip-based income—especially in heavily tip-influenced professions, such as cosmetology—some self-employment income, and household income from the D/E rates measure renders the earnings portion of the D/E calculation subject to significant errors. It also agrees that institutions should encourage graduates to report all income accurately to the IRS; however, institutions do not complete tax returns for students and cannot guarantee accurate reporting.

While the Department agrees that individuals who receive tip income should report that income fully and pay required taxes on that income, it is not the fault of institutions of higher education that many individuals do not. The IRS often assesses the fact that many tipped workers often underreport income, which further demonstrates that the D/E rates calculation is subject to numerous sources of error. The Department provided a means for institutions to survey program graduates to obtain an alternate earnings appeal for the program in instances where IRS data underreported actual earnings.^[90]

However, that mechanism proved more problematic and burdensome to administer than anticipated, and, in American Association of Cosmetology Schools (AACS) v. DeVos, a Federal court ruled that the Department's standard for such appeals was inappropriately high.^[91] The administrative burden and complexity of accounting for underreported income for the purpose of the D/E rates measure is another factor that supports the rescission of the 2014 Rule.^[92]

While not expanding the application of its holding beyond AACS cosmetology programs, in AACS v. DeVos, the D.C. Circuit noted, in dicta, that the problem of underreported income is not reserved solely to cosmetology programs. The court stated: “The problem of underreporting [income] extends across multiple industries and even across individual entities within those industries. While cosmetology schools' graduates engage in, on average, a certain amount of underreporting, other industries likely also experience different levels of underreporting based on factors like the amount of tips their graduates earn, how frequently their graduates are self-employed, and the amount of tax-compliance training their graduates receive. Within these industries, individual schools experience varying levels of underreporting.” ^[93] The consequence of this phenomenon, regardless of the existence of civil and criminal penalties, was an artificial devaluing of programs subject to graduates underreporting their income.^[94]

As stated above, to remedy the underreporting issue impacting a program's D/E rates, the 2014 Rule offered an alternate earnings appeal process. Here, the D.C. Circuit found the process reasonable “on the surface,” but identified the assumption that every program would be capable of mounting an appeal “the fly in the DOE's reasoned decision-making ointment.” ^[95]

The problem, the court found, was AACS's evidence that showed that cosmetology schools were “simply unable to mount appeals.” ^[96] When considering that, according to the reported 2015 GE data, there were over 950 cosmetology programs that could not accurately report graduate income, plus additional GE programs that rely heavily on tips such as massage therapy, hair styling, and barbering, it is difficult to justify a metric that punishes a program harshly, while not fairly, accurately, or without undue burden measuring the value of the program.^[97]

Further, the Department agrees with the commenters that SSA data may be inaccurate, especially for students who are self-employed and for workers in occupations that are highly dependent upon tip income, which may be underreported. SSA data similarly does not provide information about household earnings, which may be adequate to support a family without needing the graduate to work outside of the home. Penalizing programs because the students they serve may decide, for example, to work fewer hours in order to be with children is absurd, especially since daycare challenges and costs may make it economically advantageous to work part-time when family members can provide free or low-cost childcare.

However, SSA has not renewed its MOU with the Department and, therefore, will not currently share earnings data. As a result, the Department is unable to calculate future D/E rates unless it changes the GE regulations to rely on a different data source for earnings information. The 2014 Rule specifically states that earnings data must come from the SSA. Considering the lack of a sufficient alternative data source and that the Department has decided to rescind the GE regulations, it is not necessary to identify a new data source for calculating D/E rates.

Changes: None.

Short-Term vs. Long-Term Earnings

Comments: Multiple commenters noted that the D/E rates measure, as established in the GE regulations, did not account for long-term earnings that accurately reflect the full earnings premium associated with college completion.

Discussion: The Department agrees that D/E rates, based on earnings in the third and fourth year following completion of a program, do not accurately predict how much a graduate will earn over a lifetime.

Changes: None.

Impact of Macroeconomic Changes

Comments: One commenter stated that the earnings data used to calculate D/E rates were not sensitive to macroeconomic changes beyond the institution's control.

Another commenter stated that the impact of economic issues, such as how recessions would be accounted for, are sufficiently addressed in the 2014 Rule by using a cohort that includes multiple years of graduates and considers results over several years. The commenter stated that the Department has not explained why it changed its interpretation of the rule regarding these issues. The commenter also stated that the Department fails to disprove the 2014 Rule's research on adult students and D/E rates in its justification to rescind the GE regulations.

One commenter stated that using the impact of economic recessions to justify the rescission of the GE regulations is inappropriate, because data collected during a recession would be an outlier and would not have a long-term impact on rates or program sanctions. Another commenter said that by the Department's own words, the Great Recession was an exceptional event and exceptional events should not be relied upon as a baseline in policy making.

One commenter stated that the Department misinterpreted research by Abel and Dietz ^[98] in using these data to explain its concerns about the impact of recessions on earnings and employment. The commenter stated that this research is not particularly relevant to the gainful employment conversation and only includes bachelor's degree recipients. The commenter stated that there is a connection between educational qualifications and pay that the Department did not consider. The commenter noted that Abel and Dietz looked at what graduates actually earned. The commenter also took issue with the CNN Money research that the Department cited in the NPRM since the methodology relied upon in that article was not available for review.

Discussion: The Department disagrees that the D/E rates measure under the 2014 Rule sufficiently controls for the impact of recessions. The Great Recession provides a recent example of how prolonged economic challenges coupled with high unemployment and a jobless recovery—with both phenomena lasting longer than the 3-year period afforded to institutions by the 2014 Rule—can have a considerable impact on D/E rates outcomes. It may be true that prolonged recessions of this magnitude are outlier events, but nonetheless, there could be long-lasting consequences of an outlier event eliminating large numbers of higher education programs that will be needed after the recession is over and unemployment declines.^[99]

Used as an example, the Great Recession was highly instructive, and we cannot assume that similar recessions will not occur again in the future. Not only did the Great Recession create downward pressure on wages, it also ushered in wide-spread credential inflation such that jobs that once required only a high school diploma now required a bachelor's degree simply because employers were using degrees as a filter to screen large numbers of resumes.^[100]

The Department does not believe that any studies used to make and support our decision to rescind the GE regulations were misinterpreted. The Abel and Dietz study was used to support the point that during the high unemployment of the Great Recession, credential inflation may have resulted in graduates taking jobs with earnings much lower than expected simply because other unemployed individuals with higher level credentials were plentiful. The study also points to the fact that job placement rates may have been skewed during the recession because credentials that may have technically qualified a person for a job were not sufficient enough to compete with other applicants. For example, while executive assistant jobs in the past did not require a college credential, a Burning Glass study of job postings showed that while only about a third of current executive assistants had a college credential, two-thirds of current job postings for executive assistants required at least a bachelor's degree.^[101] Credential inflation could have a significant impact on job placement rates reported by institutions since it can take years for institutions to gain approval to raise the credential level of their programs.

The Department understands the concerns about the lack of information regarding the methodology that underlies the CNN Money article.^[102] The article was included in the NPRM for the purpose of illustrating the point that economic recessions impact graduates of all institutions, not just GE programs. Even without relying on the CNN article, however, we still believe that the D/E rates calculation has numerous flaws and sources of error for reasons explained elsewhere in this document.

The Department notes that bachelor's degree programs are included as GE programs if they are offered by proprietary institutions. In fact, the largest enrollments in the proprietary sector are at online institutions that offer degrees through the doctorate level, all of which are considered to be GE programs. During the Great Recession, there were many factors that made it harder for students to get jobs, or that required them to obtain a higher degree than would otherwise be expected. All of this had a negative impact on earnings and potentially the D/E rates of some programs.

Now that the economy has recovered and unemployment is low, it is reasonable to expect that the lack of access to workers with sufficient education and credentials could hold organizations back from growth they could otherwise support. The Department believes that it is dangerous and inappropriate for it to use two words in the HEA to create an approach to institutional accountability, that could potentially be used to manipulate the higher education marketplace. We think consumers should make those decisions for themselves, aided by information the Department plans to make available through the College Scorecard.

Changes: None.

Geographic Disparities

Comments: One commenter stated that pay disparities based on location and geography would impact a program's D/E rates but would be beyond the institution's control.

On the other hand, another commenter stated that the Department has conducted no analysis to demonstrate that there is a connection between geography and D/E rates.

Discussion: A review of published GE earnings data, if sorted by program, show that earnings differ widely among both community colleges and proprietary institutions (for certificate programs offered by both institutions), with some community college graduates earning more than proprietary graduates in some instances, and proprietary graduates earning more than community college graduates in others. A close examination of these data reveal that geography could be responsible for earnings differences.^[103] For example, not a single cosmetology program in Oregon passed the D/E rates measure, whereas almost all programs in Maryland passed.^[104] While programs in Puerto Rico resulted in the lowest earnings among all GE programs, nearly all passed the D/E rates measure because of the significant subsidies that public institutions receive. It therefore appears that geography can, in fact, have an impact on wages.

In some instances, it may be difficult to fully appreciate the impact of geography on D/E rates because large, national institutions may have, in addition to a main campus in one state, additional locations in multiple States. Yet program outcomes are reported in aggregate and attributed to the main campus at its location.

The Department of Labor's ONET database provides evidence that geography has an impact on earnings. For each occupation, ONET lists wages by State, and those data make it clear that many occupations have prevailing wages that differ from one State or region of the country to another. For example, the ONET page for cosmetologists provides wage data by State showing that cosmetologists in Alaska earn more than the U.S. average, whereas cosmetologists in Mississippi earn less than the U.S. average.^[105]

Therefore, we believe the evidence is substantial that even within a given occupation, salaries can vary from one geographic region of the country to another, and yet the D/E rates measure fails to take those differences into account. This is another example of why a bright-line standard is inappropriate and invalid since the D/E rates calculation does not control for general differences in wages across States. Note that when calculating the Estimated Family Contribution, FSA considers differences in taxes and the cost of living across States. That the Department didn't similarly build in a correction factor for differences in prevailing wages from one State to the next in calculating D/E rates was an unfortunate omission with potentially devastating impacts on students.

Changes: None.

Cohort Sizes

Comments: Some commenters expressed concerns that the small size of some program cohorts could result in year-to-year fluctuations in D/E rates due to the career decision or performance of a single student, whereas the impact of a single student's career decision or performance would not have a noticeable impact on larger cohorts.

Discussion: The Department agrees with the commenters that cohort sizes can have an impact on year-to-year changes in outcomes, since smaller cohorts can be significantly impacted by the decision of just a small number of graduates to work part time or to take time out of the workforce. This means that year-over-year outcomes could differ, even if there are no changes in program content or quality. Given the large number of low-enrollment GE programs, a single student's earnings or career choices could have a significant impact on outcomes for a number of programs and institutions.

We agree that this is yet another weakness of the D/E rates methodology and appreciate the commenters for bringing it to our attention.

Changes: None.

Influence of Student Demographics

Comments: One commenter stated that D/E rates can be affected by the percentage of adult students enrolled in a GE program because of their higher loan limits. The commenter recommended either reporting D/E rates separately for independent and dependent students or capping the amount of independent student borrowing at a lower level, rather than rescinding the GE regulations.

Many commenters supported the proposed rescission of the 2014 Rule due to the impact that various types of employment have on their programs' D/E rates. For example, one commenter stated the 2014 Rule hurts students who are on State assistance due to health issues but want to prepare for a new occupation that could accommodate their individual health needs and allow them to work, even if they cannot work full time. The commenter opined that educating such students would unfairly affect that program's metrics. Another commenter stated that the GE regulations create a disincentive to enroll students with the greatest financial need since they would be most likely to borrow to pay for the education, and the level of a student's borrowing is beyond the institution's or program's control. One commenter noted that much of the total borrowing by students is used for living expenses and not tuition and fees. Another commenter stated that students who are pregnant or have young families may unfairly and negatively impact a program's D/E rates, because their focus may be on their family rather than on finding a job with high earnings.

One commenter noted that the proposed regulations contradict the statement in the 2014 Rule that the GE regulations “do not disproportionately negatively affect programs serving minorities, economically disadvantaged students, first-generation college students, women, and other underserved groups of students.”

A few commenters objected to the Department's assertion that title IV eligibility based on D/E rates creates unnecessary barriers for institutions or programs that serve larger portions of women and minority students. One commenter asserted that the NPRM misrepresents the experiences of historically disadvantaged groups, including in its suggestions regarding women and students of color. The commenter contended that rescission of the 2014 Rule will exacerbate gender-based and race-based disparities in wealth, income, and employment.

Another commenter stated that the NPRM falsely asserts that the 2014 Rule limits postsecondary access based on geographic, racial, and gender considerations. The commenter contended that many proprietary institutions have a track record of enrolling disproportionate numbers of minorities, lower-income individuals, and single mothers, not because of a lack of accessible options elsewhere, but rather because the programs successfully target underserved communities and low-information consumers.

One commenter stated that the College Board chart used to show inherent earnings differences linked to race, gender, and family socioeconomic status relies on Current Population Survey data that is not limited to those students who graduated from gainful employment programs and received Federal financial aid. The commenter claimed that the Department provided no real analysis as to how the data in this chart should be interpreted or applied to the rescission of the GE regulations, while an earlier version of the report was used in 2014 to reflect the point that higher education provides returns for students overall.

One commenter provided citations from NCES and the Brookings Institution—cited elsewhere in this document—to refute the Department's assertion that student demographics and socioeconomic status play a significant role in determining student outcomes, and suggested that these data similarly refute our claim that student demographics rather than program quality could be responsible for GE outcomes.

Discussion: The Department agrees that the percentage of independent students enrolled in a program could impact the calculation of D/E rates because of the higher loan limits Congress has provided to those students. Congress has established student loan limits at $31,000 for dependent students and $57,500 for independent students, recognizing that independent students are less likely to receive financial assistance from parents and are more likely to have higher housing and dependent care costs than dependent students.^[106] Because borrowing limits are based not just on the cost of tuition, fees, and books, but also include housing, transportation, and dependent care expenses, independent students may rely on student loans to offset lost wages and pay costs of living during periods of postsecondary enrollment.

The Department wishes to point out that the amount of debt utilized for calculating the debt portion of the D/E rates is the lower of mean/median debt or total direct educational costs—tuition, fees, books, supplies, and equipment—so that loans taken for non-direct expenses may be excluded from the calculation. Still, adults with higher borrowing limits who borrow to generate a credit balance must first borrow enough to pay all of the direct costs of education since the credit balance is generated only after those other expenses are paid.

As described earlier, independent students borrow more frequently and at higher levels than dependent students.^[107] Therefore, institutions that serve higher proportions of independent students will likely have higher student loan medians and averages. Proprietary institutions serve a disproportionate number of independent students (80% vs. 59% and 36%), as compared to community colleges or four-year public institutions, which will impact their D/E rates.^[108]

The 2015 follow-up survey to the 2003-04 Beginning Postsecondary Survey shows that after twelve years of loan repayment, independent students across all institutional sectors still owed between 78.1 percent (average) and 96 percent (median) of their original loan balance.^[109] The 1994 follow-up survey of the 1989-90 BPS showed that independent learners are less likely to complete their programs, especially if they also have dependents other than a spouse, enroll part time, or work full time while in school.^[110] Clearly student age is one factor that impacts both borrowing levels and completion rates.

While one commenter recommended that a separate D/E rate be calculated for independent students, since the Department is rescinding the GE regulations for the reasons discussed elsewhere, this distinction is no longer necessary.

The Department agrees with commenters about the negative, unintended consequences that the 2014 Rule could have on the lives of students and on the national economy. As noted in the NPRM, and elsewhere in this document, the Department is aware that some students take time out of employment or elect part-time work over full-time work to care for children, care for other family members, manage a personal health condition, start a business, or pursue other personal lifestyle choices.^[111] The Department concurs that students who may not want to or be able to work full time should not be denied an educational opportunity.

The Department also agrees with commenters who expressed concern that the GE regulations could deter programs from enrolling students with high financial need, minority students, or women because they are more likely to borrow more and to have greater challenges in earning equal pay to men and non-minority students who complete similar programs. Thus, these students could make it more difficult for the institutions' programs to pass the D/E rates measure, regardless of program quality.^[112]

According to the Census Bureau, real median earnings differ by race, with Asians ($81,331) and whites ($68,145) earning more than Hispanics ($50,486) or African Americans ($40,258), and with males ($44,408) earning more than females ($31,610).^[113] While these data are not limited to students who participate in GE programs, we believe it is likely that the disparities that exist among the population at large are also reflected in the subpopulation of students who enroll in GE programs, and may even be greater.

Moreover, programs serving women who are pregnant or who have young children are less likely to pass the D/E rates measure because women with children under the age of 6 are more likely to leave the workforce in order to care for children. According to the Census Bureau, in 2017, among married couples with children under the age of 6, 36 percent rely solely on the husband's income to support the family.^[114] In such a case, the D/E rates for the program from which the wife graduated would be negatively impacted by zero earnings for that graduate, even though she is part of a household with sufficient income to support her decision to leave the workforce.^[115] Therefore, two programs of equal quality could have significantly different outcomes under the D/E rates measure simply because one serves a higher proportion of married female students with children than the other.

Almost four million families with a female head of household and no husband present live below the poverty level, whereas only 793,000 families with a male head of household and no wife present live below the poverty level.^[116] In 2018, 30 percent of households with children under the age of 18 are led by a single mother.^[117] These data also have implications on student loan repayment rates since a borrower in an income-driven repayment plan will have a monthly payment based on a percentage of discretionary income, which varies by the number of people in a family. Therefore, a borrower who is a parent may have a smaller portion of income available for student loan payments, potentially resulting in negative amortization of their loans.

College Board data confirm that achievement gap disparities exist between men and women and between children from wealthier families and children of low-income families.^[118] Additionally, a 2017 report released by NCES confirmed the persistence of achievement gaps between non-minority students and minority students.^[119] Therefore, if programs are incentivized to serve more advantaged students to ensure better D/E rates outcomes, they would likely follow the lead of more selective non-profit institutions that enroll small proportions of low-income, minority, and non-traditional students.

The Department has not analyzed participation in GE programs by students with health conditions that preclude them from working full time, but any student who works less than full time will earn wages that reduce the mean and potentially the median earnings used for the D/E calculation. Therefore, the Department agrees with the commenter who suggested that programs may be less interested in serving students with chronic health conditions or disabilities, since doing so could reduce mean or median earnings among a cohort of completers.

The Department wishes to clarify that in the 2014 Rule, it stated that “student characteristics do not overly (emphasis added) influence the performance of programs in the D/E rates measure.” ^[120] However, the Department acknowledges that this statement was based on an incomplete analysis of the data available to the Department and considered only students enrolled in GE programs without controlling for other variables that may have impacted GE outcomes. NCES data confirm the impact of student characteristics on outcomes, and the Department erred in ignoring those findings when making this claim in the 2014 Rule.^[121] Moreover, a review of the final GE data reported in 2017 confirms that programs that prepare students for occupations that are dominated by males rarely fail the D/E rates measure, whereas occupations dominated by women are represented disproportionately. This would suggest that gender does have a larger impact on D/E rates than the Department originally anticipated.

When full student populations are analyzed, such as through the Beginning Postsecondary Survey, we see over and over again that student characteristics have a considerable impact on student outcomes.^[122] It was misleading for the Department to make a statement in the 2014 Rule that does not accurately reflect the consistent findings of the National Center for Education Statistics, which conclude that student demographics and characteristics have a considerable impact on student outcomes.

The Department disagrees with the commenter who said that College Board data showing disparities in earnings based on gender, race, or ethnicity does not apply to the GE regulations because these data are not limited to students who complete GE programs or students who receive financial aid. The point of sharing the College Board data was to illustrate that pay disparities exist among women and minorities across the population, which supports our assertion that programs with larger proportions of women and minorities may achieve poorer outcomes under the D/E rates measure. It is unlikely that students who complete GE programs are not subjected to the same gender and race pay disparities that exist across the general population.

The Department agrees with commenters that historical and continuing discrimination has unfairly depressed the earnings of historically disadvantaged groups. We did not mean to suggest that women and minorities wish to earn less money or select occupations in order to earn less. We simply were making a statement of fact, which is that women and minorities still earn less than non-Hispanic whites and men, even when they graduate from the same institutions. We applaud first generation college students, women, and minorities who wish to leverage their own hard work and opportunities to give back to their communities by working in occupations that have high societal value, even if these jobs pay low wages.

In the NPRM, we were simply pointing out that nationally, women and minorities enroll in majors associated with lower wages than those selected, on average, by white males, and that the GE regulations could reduce the number of options available to women and minorities despite their interest in pursuing certain careers and the benefits that those individuals and occupations provide to society because occupations that pay lower wages are more likely to fail the D/E rates measure. Although some institutions have implemented differential pricing so that students pay tuition based on the program in which they enroll, many institutions do not offer different tuition levels for different majors. Unfortunately, the earnings gap between female-dominated and male-dominated occupations persist, making it more likely that programs serving mostly women will fail the D/E rates measure.

The Department does not agree with the commenter that by continuing the GE regulations, women will benefit since the programs that failed the D/E rates measure were far more likely to serve female students rather than male students. Eliminating programs that predominately serve women, and that prepare large numbers of them for rewarding occupations, is not the solution to the lack of pay equity in this country. While the commenter may be implying that women who are shut out of healthcare and childcare occupations, for example, will be more likely to pursue higher earning occupations, such as computer science or advanced manufacturing, there are no data to support that conclusion. Instead, women who lack access to the academic programs of interest to them may be reluctant to pursue higher education.

The Department disagrees with commenters who suggested that by rescinding the GE regulations, the Department will exacerbate gender-based and race-based disparities in wealth, income, and employment. Since many GE programs serve high proportions of women and minorities, sanctions that would eliminate these programs could reduce postsecondary opportunities, thereby contributing to the earnings and opportunity gap.

The Department agrees that proprietary institutions serve a disproportionate share of underserved communities, and that this could be as much the result of nefarious targeted marketing efforts ^[123] as it is the result of bona fide efforts to serve a population of students not served by traditional institutions. We have seen no national effort on the part of traditional four-year institutions to serve, en masse, the population of students who have been served by community colleges and proprietary institutions.

While the Department shares the commenter's concern about exploitative practices, many proprietary institutions employ pedagogical strategies—such as block scheduling, predetermined course sequences, year-round scheduling, and accelerated completion pathways—that may be more appealing to non-traditional students.^[124]

The Department has not analyzed the racial or ethnic demographics of students served by programs that failed the 2015 D/E calculations. However, given that a large number of programs that failed the D/E rates measure, or that were discontinued by institutions that expected they would fail the D/E rates measure in the future, were medical assisting and related programs, or cosmetology programs—both female-dominated professions—it seems clear that women will be impacted more significantly by program closures than men. Also, given the high percentage of Pell grant recipients enrolled in programs with failing 2015 D/E rates, there is evidence that program closures would have a disproportionate impact on low-income students. Programs that serve high-income students would not fail the D/E rates measure because those students are far less likely to take student loans and, in addition, are more likely to receive financial support from parents during the early years of repayment.^[125]

The Department continues to believe that the GE regulations could significantly disadvantage institutions or programs that serve these already underserved communities, further reducing the educational options available to them.

The data are clear that proprietary institutions serve higher proportions of non-traditional and low-income students, as demonstrated by the fact that nearly 87 percent of students enrolled at proprietary institutions are Pell eligible, as opposed to 45 percent at community colleges and even lower percentages at public or private four-year institutions.^[126]

As College Scorecard expands to the program-level for all categories (GE and non-GE) of title IV programs, it will be important to keep in mind student demographics when comparing outcomes, including among open-enrollment institutions that typically serve higher proportions of low-income and minority students. Many of these institutions attract low-income populations to increase enrollment, but the Department believes that most also do it to fulfill their mission to improve educational opportunities for all students.

The Department does not disagree that low-income and minority students have poorer educational and employment outcomes, and it does not disagree that proprietary institutions serve large proportions of these students than any other institutional sector. Compared to public two-years, public four-years, and private non-profits, proprietary institutions serve greater numbers of females, minorities, financially independent, and single parents.^[127] The Department encourages more selective institutions to do a better job of serving this population of students, but recognizes the unique opportunities provided by institutions that are designed to serve the needs of non-traditional students and may be more aware of their unique challenges and needs.

Changes: None.

Role of Tuition in Determining D/E Rates

Comments: One commenter noted that the GE regulations do not prohibit institutions from lowering tuition, which would also increase a program's chances of passing the D/E rates measure. The commenter suggested that focusing on cost is one way to avoid the impacts that macroeconomic trends have on earnings.

Several disagreed with conclusions they believe were drawn in the NPRM regarding program cost relative to value. These commenters suggested the Department focused only on one half of the D/E rates calculation to make its point, and that it inaccurately suggested that a program of higher cost is necessarily of higher quality. One commenter stated that “a program that has low costs but results in higher earnings to students obviously has higher quality than one that has high costs and low earnings.” This commenter suggested that the Department's assertion reflects a rampant fallacy in higher education that a higher cost program is a higher quality program.

Another commenter stated that the Department seems to be skeptical that program costs and earnings are reliable measures of success. Multiple commenters disagreed with the Department's contention that high-quality GE programs could potentially fail the D/E rates measure, because it costs more to provide high-quality education in certain fields or disciplines.

One commenter specifically mentioned that community colleges provide high-quality GE programs despite their low tuition and fees.

Discussion: The Department agrees that the GE regulations do not prohibit an institution from lowering tuition for a program, and that doing so could favorably impact GE outcomes. And the Department agrees that just because a program is higher cost, it is not necessarily higher quality. However, in some instances the higher cost is associated with better equipment and facilities, more highly qualified faculty, better quality or more plentiful supplies, and more abundant or convenient student support services. In some instances, if an institution were forced to lower its prices, it would be unable to provide the unique learning environment or well-equipped facilities that distinguish the institution.

The Department commends community colleges for the tireless and vitally important work they do. However, as pointed out by the CSU Sacramento report,^[128] as well as data collected by the Department through IPEDS, many community colleges have small or shrinking CTE programs and may not be able to meet workforce needs or accommodate adult learners who may prefer accelerated scheduling, more personalized support services, smaller campus environments, more frequent program start dates, and predetermined course schedules that are more common among proprietary institutions.^[129]

A review of 2015 GE data reveal that in some instances, graduates of proprietary institutions enjoy significantly higher earnings than graduates of community college programs, which may indicate that the higher cost program might be a higher quality program, or that the institution has valuable partnerships with employers or has better job placement services.^[130] As Cellini pointed out, despite several limitations of the data she used, students who earn a cosmetology certificate at a proprietary institution are more likely to earn higher wages, perhaps due to the affiliation of some proprietary institutions with high-end salons.^[131] At the same time, the graduates of many proprietary institutions achieve lower earnings gains than the graduates of other institutions, including community colleges or four-year institutions. And similarly, even among programs with the same CIP code, the GE data illustrate that there are vast earnings differences among community colleges and among proprietary institutions.

Students may find that public colleges offer smaller numbers of CTE programs than private or proprietary institutions. Nationally, the largest community college majors are liberal arts or general studies, which could signal that the majority of students are interested in transferring to a four-year program or that vocational programs are limited. In other instances, entry-level CTE programs might be offered only through the institution's non-credit or continuing education programs. These programs are not eligible for title IV funding and do not result in academic credit, which can disadvantage students who wish to continue their education and earn a college degree.

The Department is concerned that at many public colleges, students who are enrolled in pre-professional programs have nowhere to turn if they are not admitted to the professional program of interest. For example, many students enroll at a two- or four-year institution with the goal of studying nursing, physical therapy (or physical therapy assistant), or occupational therapy (or occupational therapy assistant); however, these programs are often highly competitive, and the majority of applicants are not admitted. The absence of other allied health options at some institutions may require those who are not admitted to professional programs to either pursue a general studies major or to transfer to another institution that offers a larger number of related programs that enable a student to stay in their field of interest even if it means pursuing a different occupation in that field.

The Department encourages institutions to work hard to reduce costs, encourages states to continue subsidizing higher education to reduce the price of public institutions, and encourages employers to provide more generous education benefits to reduce out-of-pocket costs to students. As stated earlier, public institutions offer lower tuition and fees because of the public subsidies they receive from state and local governments. However, at some public institutions out-of-state students who may be more academically gifted or who pay higher tuition and fees take priority over lower-income or less prepared in-state students because out-of-state students are perceived as being necessary to improve the institution's finances and reputation.^[132]

Research shows that the administrative costs for CTE programs are typically higher because of the need for specialized facilities, expensive equipment or supplies, smaller class sizes (due to space and/or safety concerns), and the higher cost of faculty with advanced technical skills.^[133] And as pointed out by Shulock, Lewis, and Tan, community colleges often reduce the number of CTE programs or the number of enrollment slots in the CTE programs they administer when budgets are tight.

As already discussed, the largest community college major is general studies or liberal arts, which according to Holzer and Baum has no market value for the majority of students who earn this degree and then do not transfer to complete a four-year degree. It is, therefore, difficult to know whether a general studies program is a worthwhile investment, if a student's goal is to earn a two-year degree that will lead to a higher paying job. A students may be better off paying more to attend an institution that increases the likelihood that the student will be able to enroll in an occupationally-focused program, or will be more likely to complete their program, than attending the lower tuition school if doing so limits the student's opportunity to pursue occupational education.

In conducting the current rulemaking effort, the Department considered tuition and fees charged by all institutions since our goal was to expand the accountability and transparency framework to include all institutions. Nearly all private institutions charge higher tuition and fees than public institutions, and a growing number of students who enroll at public institutions attend an institution outside of their own state. Out-of-state tuition at public institutions mirrors the tuition charged at private institutions. Negotiators representing private, non-profit institutions made it clear that D/E rates will differ between private and public institutions due to differences in the level of public subsidies an institution receives. An institution's geographic location, campus facilities, and engagement in research and graduate education could impact the tuition and fees that students are charged. The Department sought through rulemaking a data-driven solution that could be applied to all institutions of higher education to better inform students and families about likely costs, borrowing, and earnings.

Over the years, policymakers of both major political parties have admonished institutions to lower their costs, but proposals that would impose federally mandated price controls have never gained sufficient support to become law.^[134] For example, in order to help families make better decisions about where to enroll and how much to borrow, Congress proposed in the College Access and Affordability Act of 2005 the creation of a College Affordability Index (CAI) which would have identified institutions whose tuition increases outpaced inflation. In the House Report 109-231 at 159, Congress stated that the CAI: “simply ask[s] that an institution of higher education provide additional information to allow for a clear and informed decision by consumers. If a student decides to attend an institution that increases tuition and fees that exceed the College Affordability Index, they do so fully aware and educated. It is the Committee's position that the Federal government does not currently have the authority to dictate tuition and fee rates for institutions of higher education. . . . The provisions in the bill simply serve as a means by which additional information can be provided to students and their families so that they can make informed and educated decisions about their postsecondary education options.” ^[135]

Therefore, the Department believes that creating a system of sanctions that are so closely linked to the tuition and fees a college charges would exceed the Department's current authority and run counter to the authorities laid out by Congress to inform decisions, but not dictate what prices a college can charge. As a result, the Department continues to believe that a program could fail the D/E rates measure not because it is of poor quality or because it is over-priced relative to the cost of delivering the program, but instead because the cost of educational delivery is high or because an institution does not receive public subsidies.

Changes: None.

Challenges in Predicting Future Earnings

Comments: One commenter urged the Department to apply any outcomes metrics equitably to all institutions, rather than singling out or discriminating against one type of institution. The commenter also urged the Department to use simple, easy to understand formulas and to keep in mind that it is impossible for colleges to predict future changes in the economy or career areas.

Discussion: The Department agrees, as we discussed earlier in this document, that the widespread problem of student loan debt makes it important to apply the same transparency and accountability metrics to all institutions. We also agree that we should avoid the use of complex formulas or those that allow the Department to manipulate outcomes by defining variables that are inconsistent with the requirements of student loan repayment programs. The Department agrees with the commenter that because the GE regulations do not calculate D/E rates until years after a student is admitted—sometimes as many as nine years after a student enrolls in a bachelor's degree program—an institution must be able to predict macro-economic conditions, future earnings, and various other factors that influence employment and earnings well in to the future in order to establish a price that will guarantee passing D/E rates, a nearly impossible task. Institutions that receive generous taxpayer subsidies can reduce the price students pay such that graduates pass almost any earnings test, but taxpayers also deserve to know if the price they are paying for a student's tuition is justified by the outcomes students achieve. The Department has determined that the best way to establish an equitable and meaningful transparency framework is by reporting debt and earnings income for all types of title IV programs to the public so that a market-based accountability system can flourish.

Changes: None.

Impact of the 90/10 Rule

Comments: One commenter expressed concern that the 2014 Rule may be in tension with the 90/10 requirement. The commenter believed logic from the Department or others indicating the 2014 Rule could encourage schools to reduce tuition is faulty because it puts schools at risk of noncompliance with the 90/10 rule without giving these schools tools necessary to reduce student borrowing.

Many commenters argued that some colleges use aggressive marketing and recruiting to target veterans and service members in an effort to supplement title IV funds with GI Bill funds because the latter do not count against institutions for purposes of 90/10 rule compliance.

Another commenter mentioned law enforcement investigations and actions regarding proprietary institutions. Three of the investigations specifically reference court cases where some institutions were under investigation for misrepresenting their compliance with the 90/10 rule.

Some commenters, who were in favor of rescinding the regulations, argued that they do not treat all educational institutions the same. One commenter argued that public institutions are afforded much more leniency in the same industry, and that these public universities and community colleges are already being given a strategic advantage of not being accountable to metrics such as retention, placement, and 90/10.

Discussion: Schools that misrepresent their compliance with 90/10 are in violation of the Department's regulations, regardless of whether we rescind the GE regulations. The Department strongly believes these institutions should be held accountable and takes action against schools out of compliance with 90/10—as is required by law—including loss of title IV participation.

The Department appreciates comments that point out the upward pressure that the 90/10 rule places on tuition costs at proprietary institutions and demonstrate the perverse incentives these regulations create that are not helpful to students. Because of the statutory requirement that proprietary institutions generate at least 10 percent of their revenue from non-title IV sources, coupled with the inability for an institution to establish lower student loan borrowing limits or to deny a student the right to borrow, an institution serving large majorities of low-income students will find it challenging to pass the 90/10 requirement if they lower tuition well beneath federally established borrowing limits.

Also, since independent students have higher borrowing limits than dependent students, and since the title IV loan programs enable students to borrow enough to pay for living expenses, an institution may be unable to prevent students from borrowing a more reasonable amount and working to pay some of the costs in cash because doing so will interfere with the student's ability to receive a credit balance to use for rent, food, and other costs of living. Since borrowing limits are based not just on tuition and fees, but also include housing, food, dependent care, and transportation, lowering tuition may not have a dramatic impact on borrowing. Even among community college borrowers where tuition is low, the average debt is $13,830, which shows the impact of non-tuition costs on student borrowing.^[136]

Therefore, the Department believes that providing program-level debt and earnings information for all categories (GE and non-GE) of title IV participating programs is the best way to help all students make better informed decisions.

Although certainly there may be instances in which veterans were targeted to help meet the 90/10 requirement, it is inappropriate to suggest that schools serving thousands of veterans are somehow not delivering on their promises or providing opportunities veterans want and need. Some institutions that “target” veterans do so because they provide unique program opportunities, student services, or adult learning environments better suited to the needs of veterans.

Some proprietary institutions are more attractive to veterans than other institutions because they are designed around the needs of adult learners, serve large populations of veterans who share certain values and life experiences, provide additional training to faculty on the unique needs of veteran students, are more likely to accept credits earned from other institutions, and they are more likely to give credit for skills learned during military service. Student veterans made tremendous sacrifices to earn their GI Bill benefits and should be able to use their benefits to attend any school that works well for them. The Department appreciates the comments on 90/10; however, that rule is not the subject of this rulemaking.

Changes: None.

Reporting and Compliance Burdens for GE Programs

Comments: Several commenters expressed concern that if the Department chose to expand GE-like requirements to include all institutions, it would add significant reporting and compliance burden to all institutions. Some commenters expressed a desire to limit the applicability of the GE regulations to the programs covered by the definition of “institution of higher education” in section 102 of the HEA.

One commenter discussed other Department requirements that institutions are already subject to, such as enrollment reporting and requested the Department carefully consider the implications of expanding disclosure requirements to all title IV-eligible programs.

Several commenters discussed how the reporting burden from the 2014 Rule took away resources from efforts that would actually improve student outcomes.

Other commenters described the problems that would be presented by the requirement to directly distribute disclosures to prospective students by specified procedures at the correct stage of the matriculation process and to maintain all the records to document compliance. Commenters also expressed concerns about protecting student privacy and managing data associated with the records retention requirements.

On the other hand, other commenters stated that burden reduction was not a sufficient reason to justify the proposed regulatory changes.

One commenter stated that the Department misrepresents the stance of the American Association of Community Colleges (AACC) in relation to the burden associated with the reporting and disclosure requirements of the GE regulations and that community colleges have been supportive of the GE regulations.

Several commenters stated that they thought efforts to reduce regulatory burden should be made while also maintaining sanctions for poorly performing programs or while maintaining the GE regulations.

Several commenters affirmed that meeting disclosure requirements using the standardized GE Disclosure Template posted to individual program web pages presented a much greater administrative burden than was reflected in the 2014 Rule's Regulatory Impact Analysis.

Some commenters described how the burden from GE reporting requirements impacted student services at their school, with one commenter stating that it slowed down responsiveness to student and business needs at community colleges. Another commenter described services that were impacted by resources needed to fulfill GE reporting requirements, explaining that resources were taken away from activities that would help students achieve gainful employment such as providing student counseling and making efforts that would assist students with completion.

Some commenters pointed out that the costs of compliance are reflected in higher program costs passed on to students and taxpayers. Another commenter emphasized the need for the Department to carefully consider costs when establishing any future disclosure framework.

One commenter indicated that it would be unlikely for institutions to save much money from the reduced administrative burden from the regulatory change. The commenter also indicated that it would be unlikely that any savings passed to students would be enough to change student decision-making. The commenter expressed concern that removing the extra costs would provide proprietary institutions with a wider profit margin to operate and would encourage expansion.

Multiple commenters stated that the Department should encourage maximum transparency by requiring all programs at all institutions to disclose the same information so that students could have a baseline in which to compare information.

Some commenters suggested that the Department should publish information from data that it already has access to, sparing institutions from having to meet additional reporting requirements.

Some commenters emphasized that program disclosures should be easy to find.

Some of these commenters expressed concern that the direct distribution requirement in the GE regulations would take away ease and flexibility that students need in the application process and that students may be overwhelmed by disclosures.

Some commenters expressed concern regarding inconsistencies in the way that job placement rates are determined and reported under the GE disclosure requirements. Several commenters suggested that the Department standardize the methodology for calculating in-field job placement rates the same way that accreditors have done.

Many commenters expressed the desire to see fair and consistent disclosures allowing students to make apples-to-apples comparisons among programs. Several commenters explained the difficulty of manually gathering GE reporting data, such as job placement rates, as is required by the 2014 Rule. One commenter stated that they were not confident in the reliability of data calculated by thousands of institutions according to their own interpretations of the 2014 Rule, especially with regard to the definitions and calculations of job placement rates. Multiple commenters emphasized the importance of avoiding disclosure of metrics such as job placement rates that are not comparable due to differences in State and accreditor definitions.

Others were opposed to requiring GE-style disclosures of all institutions but did agree that there is a need for greater transparency which could be achieved by the Department through the College Scorecard.

One commenter would prefer that any net price disclosures focus on tuition and fees, independent of living expenses.

One commenter stated that the Department had not adequately explained why direct disclosures should not be provided to prospective and enrolled students or included in promotional and advertising materials.

Discussion: The Department thanks the commenters for sharing their insight into how the GE regulations are affecting schools and their ability to serve students. The Department's decision to rescind the GE regulations will enable institutions to redirect resources to other institutional functions and priorities. We strongly encourage institutions to do so. The Department agrees with the commenter who stated that proprietary institutions could use the cost savings generated from rescinding GE to increase their profit margin, but that is true of any institution that has GE programs. The Department sincerely hopes that institutions apply the savings generated to education and student services, but it acknowledges that it cannot control how institutions utilize cost savings.

In addition to reducing the cumbersome reporting burden associated with the reporting provisions of the GE regulations, by rescinding the regulations, institutions will no longer be required to engage in the direct distribution of disclosures or maintain records to prove that students receive those disclosures.

The Department agrees with the commenter who pointed out that it can be difficult to find GE disclosures on many websites. In our own efforts to review GE disclosures, we found that many of them are more than one or two clicks away from the program page, and some are not even referenced on the program pages, but instead are under a separate page for institutional research or consumer information. The College Scorecard, focusing on tuition and fees, will provide “one stop shopping” to students and families seeking information about institutions and programs, and it will allow the student to select multiple campuses and programs for the purpose of comparing information on the same screen.

The Department acknowledges that the AACC has been generally supportive of the concept of the GE regulations; however, they have not spoken favorably about the administrative burden the regulations have placed on their own members. Due to taxpayer subsidies, which reduce the price students pay, their programs will likely pass the D/E rates measure even if earnings or program quality are very low. In fact, the Department points to this as one of the reasons why the D/E rates measure is not an accurate indicator of quality since programs with exceptionally low earnings will pass the measure as long as those programs continue to be subsidized by taxpayers.

In addition, given the small number of community college GE programs that met the minimum cohort size, the Department agrees that the burden of reporting was not justified by the information provided. For many programs, D/E rates were not issued because of small cohort sizes and many data items on the GE Disclosure Template output would appear as “not applicable” because a group contained fewer than 10 students. Of the 18,184 GE programs offered by non-profit institutions in 2017-18, only 3,708 have cohort sizes of 10 or more. This means that relatively few GE programs offered by non-profit institutions would be subject to the D/E rates measure or disclosure requirements, but it also means that there are relatively few opportunities for students to engage in occupationally focused education at non-profit institutions. This fact may be the single most important clue as to why proprietary institutions have become increasingly attractive to students seeking occupational education and credentials. A program that graduates less than 10 students per year is obviously quite small, either because of enrollment caps that the institution or its accreditor places on the program or because students at the institution are largely unaware that the program exists. Clearly, the majority of GE programs accommodate a very small group of students as table 1-1 previously showed, which may suggest that the programs available at non-profit institutions simply do not provide the supply of enrollment opportunities that meet student or workplace demand.

The Department notes that AACC states in its comments that “implementing the gainful employment regulation has been hugely burdensome for community colleges” and that “any future GE regimen must be extremely sensitive to cost.” ^[137] Therefore, we do not believe that we have misrepresented the position of AACC regarding the reporting and disclosure burden. We agree that the GE regulations have been overly burdensome to schools and to the Department, and that all regulations should be sensitive to cost and burden. By rescinding the GE regulations, the cost and burden associated with GE reporting has been permanently removed.

The Department did not receive quantitative estimates of costs associated with changing web architecture or updating GE disclosures on institutional websites each year, so we cannot comment on whether the burden estimates in the 2014 Rule were accurate or not. Because the Department is rescinding the GE regulations, institutions will no longer be required to post disclosures of program outcomes on their websites. The Department will now provide outcomes data to all students using the College Scorecard, or its successor, which has the advantage of reducing the burden on institutions and allowing students to more easily compare outcomes among the institutions and programs available to them.

The Department thanks the commenters for their feedback and points out that the Senate Task Force on Higher Education Regulations similarly pointed to the GE regulations as being particularly burdensome regulations that outstrip legislative requirements and intent.^[138] Administering the GE regulations, particularly alternate earnings appeals, has also turned out to be much more burdensome to the Department than was originally anticipated.

Although, the Department has changed disclosure templates in an effort to make them user friendly, we are not convinced that the GE disclosures are useful to students. Consumer testing has revealed that students mostly want to know how students like them have done in the program.^[139]

In developing any future transparency framework, the Department will focus on using administrative data sets and Department-developed data tools to minimize burden on institutions and to allow students to compare all of the institutions and programs they are considering by accessing a single website. This website will be accessible to individuals with disabilities, in accordance with section 508 of the Rehabilitation Act. This will ensure that students with disabilities will be able to use the website tools and have equal access to the data that are available to all other students.

The Department agrees that as a result of differences in definitions by States and accreditors, including not only differences in how jobs are defined but also in which students are to be included in or excluded from the measurement cohort, the job placement rates reported in current GE disclosures are not comparable. In addition, the results of a 2013 Technical Review Panel highlighted that job placement determinations are highly subjective and error prone, since there is no reliable data source available to institutions for the purpose of determining or verifying job placements. Until a reliable data source is available for determining job placements, the Department believes that earnings data is the most reliable information that can be made available to students to give them a sense of graduate earnings, even if those data do not specify the precise type of job graduates have secured.

The Department agrees with the commenter that the Department should encourage maximum transparency by ensuring that institutions provide the same information to all students and prospective students. The Department has determined that an expanded College Scorecard, or its successor, not direct disclosures to students, is the appropriate way to share this information, and plans to do so by adding program-level outcomes data for completers of as many title IV programs as possible without compromising student privacy. Although the Department does not require regulatory changes to implement or modify the College Scorecard, we appreciate the many comments we received in response to the NPRM and will consider them as we plan our Scorecard modifications.

Changes: None.

Scorecard

The Department is not required to engage in rulemaking in order to make changes to the College Scorecard. Therefore, the following section of this final rule is not subject to the APA or the requirements of rulemaking. However, because we believe that the Scorecard is a critical tool to improving transparency and informing a market-based accountability system, we sought feedback from the public regarding recommended content for the Scorecard. We are providing a summary of the comments and our responses to better inform the public, but we are not creating regulations related to the College Scorecard.

Comments: Many commenters supported the Department's efforts to expand the College Scorecard to include program level data. One commenter stated that placing the information in a central location will be more effective than allowing institutions to comply with disclosure requirements by placing them in obscure sections of their websites. Another commenter supported moving all consumer data to the College Scorecard.

Several commenters had questions or concerns regarding College Scorecard data. Some commenters expressed concerns that College Scorecard data are based only on undergraduate students and that this results in inaccurate data for many institutions.

One commenter expressed concern that small cohorts are not excluded from the calculation and that the data may contain discrepancies between cohorts and methodologies used for each of the metrics or rates provided. The commenter gave the examples of such discrepancies, including their belief that: Debt amounts are based only on students with Federal loans, but earnings information is based on all students attending the institution; debt includes debt for indirect costs in addition to direct expenses; some metrics are based on completers only while others include all students; and retention and graduation rates are based on first-time, full-time students only, which is not representative of the current student population. The commenter then expressed concerns that students will not know that the outcomes data are based on different student cohorts.

Many commenters stated that they would like to see the Department's data collection efforts expanded beyond first-time, full-time students. Given the increase in part-time students, transfer students, and students who stop-out for various reasons, some commenters pointed out that by including only first-time, full-time students, the majority of students at some institutions are excluded from the data.

One commenter requested that the Department develop a mechanism that would authorize institutions to forward student data to the Department of the Treasury so that Treasury can disclose to the Department information about the earnings of all program completers and not just those who participated in title IV programs.

One commenter stated that calculators and other financial management tools that can be customized to an individual student's situation provide better information than mandatory standardized disclosures on program pages. Another commenter suggested that the Department publish a calculator allowing students to understand debt, the application of compound interest, and the expected income of a career choice.

Some commenters stated that although they value transparency and are encouraged by the Department's aims to provide more relevant information via an online portal, they believe that there is no replacement for in-person disclosures, which ensure that a student receives information and has an opportunity to ask questions and understand metrics being provided.

Several commenters expressed that they were skeptical that institutions would provide accurate information on institutional disclosures, and these commenters were concerned that institutions would put the disclosures in obscure portions of their website.

Several commenters supported the idea of adding a link to the College Scorecard from institutional program pages. One commenter suggested that the Department create a standardized icon for hyperlinking to the data disclosure portal, mandate that schools use it on their websites and set requirements for its size and prominence. Other commenters suggested that the Department require links to Department data on school websites. One commenter stated that such a link should only be to the main College Scorecard page and that requiring specific links based on program would cause undue burden.

One commenter stated that the centralized Scorecard approach would be less burdensome than updating websites and catalogs. Another advocated for measurements to be added to a national website and require that the link should be included in Admissions paperwork, Free Application for Federal Student Aid (FAFSA) documents and student catalogs.

One commenter recommended that the Outcome Measures Survey for 200 percent of time to completion be used to calculate the graduation rate data and then made recommendations for how to augment the IPEDS data collection.

Many commenters stated that disclosures should be part of the PPAs for all schools, and that all participating schools should be required to link to College Scorecard or a similar national website containing standardized disclosures. Commenters stated that such disclosures would be easy for students to use and would result in meaningful comparisons. Another commenter pointed out that disclosure requirements exist for other large transactions, such as buying a car, and students need this information when making life-impacting decisions. The commenter thought it was especially important that disclosure requirements be applied to programs subject to the 2014 Rule given past history of predatory practices at some schools.

Many commenters discussed items that they thought should be included in any upcoming disclosure framework, including: Whether a program meets State requirements for graduates to obtain licensure in the field; information about programmatic accreditation requirements, program costs, and program size; data on program outcomes such as completion rates and withdrawal rates; earnings data for program graduates after a set period of time in the job market; the percentage of students who complete the program or transfer out within 100/150/200 percent of the normal time to complete; the percentage of Pell recipients who complete the program or transfer out within 100/150/200 percent of the normal time to complete; institution-level success rates parsed out by credential level; the percentage of program graduates earning above a particular income threshold after a set period of time in the job market; and the percentage of students receiving Pell grants.

One commenter expressed concerns that the Department had not discussed any plans to include other data in the College Scorecard, such as: Primary occupation for which a program is designed to prepare students, program length, completion and withdrawal rates, loan repayment rates, program costs, percentage of title IV or private student loan borrowers enrolled in a program, median loan debt, mean or median earnings, program cohort default rates, or State licensure information, which are disclosure items covered under the GE regulations.

One commenter stated that the Department needed to provide a rationale for the decision to not continue each item required for disclosure under the 2014 Rule.

Some commenters listed questions that they would want answered if the Department establishes disclosures via the College Scorecard or other means. These questions included: How the Department will gather the information for the centralized data portal; what requirements there would be to submit data to the centralized data portal; what format the information would need to be disclosed in; how frequently information would need to be submitted to the Department; whether the Department would make it possible to submit data more frequently to ensure that the best possible data are available to students; whether the data would be disclosed on a rolling basis or with deadline requirements; how the College Scorecard or other website would indicate missing information; what enforcement mechanism might be used and how it would work; how institutions would have access to monitor and update disclosure information; what privacy controls would be used; what evidence institutions would be required to provide to support their disclosures and whether those documents would be viewable by the general public; how the Department would explain the data collection period used; what action the Department would take if it found during an audit that an institution misrepresented disclosure information; whether the Department would regularly review which data items would be disclosed for usefulness to students and; what role stakeholders would play in such a review process.

Several commenters stated that an informational solution alone, was not adequate protection for students. Some of these commenters believed that relying solely on the College Scorecard places the burden on students to find and interpret information on programs. One commenter stated that no evidence supports the conclusion that publishing more outcome data will lead to better decision making on the part of students and that most college students would not use the information anyway. One commenter cited research that indicated that upper-income students were more likely to use Federal data in their college decision-making process.^[140]

One commenter noted that the College Scorecard is not implemented through regulation and, therefore, is not a good disclosure tool to expand for programmatic disclosure purposes. Another stated that the College Scorecard will not be as effective as a disclosure template and will not lead to loss of eligibility or include a direct warning from an institution to a student considering a poor-performing program. Another commenter questioned the Department's assertion in the NPRM that the College Scorecard will provide more accurate and reliable data than the GE Disclosure Template. Finally, several commenters expressed concerns that the College Scorecard will not be enough to dissuade students from enrolling in a program if high pressure sales tactics, advertisements, commission-based compensation, and “pain points” are used in recruiting tactics.

Another commenter asked how the Department will balance the need for data with privacy protections in cases of programs with less than ten students. One commenter asked whether the Department will relax privacy protections if it provides program-level data through the College Scorecard. Without doing so, any disclosures through the College Scorecard would still not have program-level data for programs with fewer than ten completers. Several commenters suggested various metrics for inclusion in the College Scorecard, while others noted that privacy laws will prevent students from getting a truly clear picture of programmatic outcomes.

One commenter suggested differentiating earnings between those who complete and those who do not complete. Another commenter pointed out that the College Scorecard does not provide information on a programmatic level and instead provides information at the institution level. One commenter expressed concerns that the College Scorecard cannot be updated with program-level data soon. The commenter then stated that the Department should clarify if it intends to keep the same time horizon of six to ten years after entering schools, whether it will disaggregate earnings for completers and non-completers, and whether it will group very small majors in similar content areas to ensure it is able to produce data covering as many students as possible. Finally, the commenter suggested that the Department conduct consumer testing, consider holding a technical review panel with behavioral economists, designers, and other experts, and construct a data download tool for users who wish to access files with the data in smaller chunks than the current large zip file.

One commenter requested that the Department make sure that the reporting accurately accounts for the enrollment patterns of community college students who may take longer than the traditional time to complete. Another commenter expressed concerns that because most of the key College Scorecard data are based on title IV recipients, information would be made available for a minority of community college students, as fewer than four out of ten community college students receive any Federal financial student aid. The commenter went on to state that this minority of students is unrepresentative of the larger population of community college students—title IV aid recipients are generally less affluent and likelier to have worse outcomes than their better-resourced colleagues.

Many commenters pointed out that cosmetology schools and other certificate programs are not included in the current College Scorecard. One commenter asked that if the College Scorecard approach is adopted, that cosmetology schools should be included in a sensible way or be exempted from the requirement. Additionally, the commenter contended that program-level earnings data will not be representative of the income made by graduates because many completers work part-time, are building businesses, or fail to include tips in their reported earnings. One commenter asked that the Department hold off on requiring certificate programs from having to include a link to the College Scorecard until it contains data regarding certificate programs.

One commenter suggested that the Department adopt language in the College Scorecard that addresses occupational circumstances and geographic differences that have the potential to impact the accuracy and validity of the data. Another commenter suggested that the Department provide earnings information only for program completers, which differs from the current College Scorecard because the earnings information encompasses both completers and non-completers. The commenter argued that the purpose of the College Scorecard's earnings data is to inform students of what they may expect to earn if they complete a given program and that including non-completers' earnings is confusing. One commenter suggested incorporating a risk-adjusted model for presenting data based upon variables such as socioeconomic demographics and geographical location of students and the institution.

Another commenter expressed concerns that including self-reported data on the College Scorecard would invite misrepresentation.

One commenter suggested reporting median earnings of graduates by program. Another commenter suggested integrating analytic insights derived from unique, consumer-level data maintained by other sources. Another commenter suggested using the Credential Transparency Description Language schema in the College Scorecard and providing the data on the institution's website.

Some commenters stated that they did not believe it necessary for the Department to require institutions to publish information such as net price, program size, completion rates, and accreditation and licensing requirements because this information could be added to an FAQ page published to the College Scorecard site so that students could ask the schools the questions if they so choose.

One commenter expressed concern that the College Scorecard website would not include all of the information a student might need to effectively select a school. The commenter explained that disclosures are more effective when they are produced by government regulators to further policy goals rather than from an institution whose goal is to limit liability.

One commenter stated that the Department has not negotiated in good faith, because the Department has not committed to update the College Scorecard with program-level data.

Several commenters expressed concern that increasing the profile of the College Scorecard would increase burden on institutions since there would be more reporting requirements for an expanded College Scorecard. One commenter stated that requiring individual programs to track and disclose information such as programmatic outcomes, program size, completion rates, and net price would result in costs that the institutions would then pass on to students in the form of higher tuition and fees. Several commenters expressed concern over whether students would know where to find program-level information on the College Scorecard after it was added and how to interpret the information. One commenter expressed concern that there is currently no law or regulation requiring that the program-level information be added to the College Scorecard.

Discussion: The Department very much appreciates the suggestions, ideas, and potential inclusions and exclusions in the future College Scorecard, or similar tool. The Department continues to believe that the best way to create a transparency and market-based accountability system that serves all students is by expanding the College Scorecard to include program-level outcomes data for all categories (GE and non-GE) of title IV participating programs, so that students can make informed decisions regardless of which programs or institutions they are considering. The Department is also working towards providing more information to students and parents about the level of Parent PLUS borrowing. Only when parent borrowing is included can students fully understand the level of borrowing in which families engage at a particular institution. This also provides families with more complete and meaningful expectations of educational costs and students and parents should be aware of this when making enrollment decisions.

Parents in the later years of their career may be less able to manage student loan repayment than their children who have an entire career ahead of them, yet borrowing limits on Parent PLUS loans are exceedingly high regardless of the parent's income, which could have dire results as parents near their retirement years.^[141] We intend to list Parent PLUS debt separate from student debt, but nonetheless believe that it is an important addition to consider in the expanded College Scorecard.

The Department notes that several negotiators recommended that if earnings are to be reported by the Department, those earnings should be considered at 5 or 10 years post-graduation, since earnings in the early years after completion may not reflect the true earnings gains that individuals will realize from their college credential. The Department agrees that earnings at the 5- and 10-year mark, or within a similar timeframe, will provide more meaningful information about a borrower's likelihood to repay his or her loans throughout the standard repayment period. The three- and four-year earnings data currently used to calculate D/E rates were an aspect of the GE regulations that made it an unreliable proxy for program quality since it is not unusual for a graduate to take a few years to hit their career stride, especially if they enter the job market during a time of high unemployment.

Therefore, the Department intends to integrate earnings data closer to the suggested 5- and 10-year earnings data into the expanded College Scorecard. However, since the Department does not have program-level data prior to 2014-15, it will report shorter-term earnings during the first year of Scorecard expansion, and will increase the number of years following graduation that are captured in the data until it reaches the target post-completion metric.

Because students who do not complete the program will not benefit from the full program or curriculum, it is inappropriate to include the earnings of non-completers in the determination of program outcomes. While we encourage institutions to take action to increase program completion rates, the Department recognizes that there are many factors that influence a student's decision or ability to persist and complete the program. Since the HEA is designed to increase access, and since loans are made available to all students regardless of their level of academic preparedness, institutions that adhere to open-enrollment admissions policies and institutions that are minimally selective will likely have lower completion rates than highly selective institutions that serve mostly students who are economically-advantaged, traditionally-aged, and academically well-prepared for college-level work. It is not appropriate to penalize institutions because they take on the difficult work of serving high risk students.

The Department is sympathetic to the concern that by including only title IV participating students, some institutions will not have a representative sample of students included in the earnings calculation and the populations on which earnings are reported are likely to be lower earners. The Department agrees that students from socioeconomically disadvantaged backgrounds tend to have lower earnings in the early years after graduation. However, the Department is permitted to collect data only on title IV participants, unless Congress passes legislation to lift the current data collection prohibitions. Both debt and earnings data presented in the Scorecard will be limited to title IV participating students; however, the Department will work to help students understand why earnings data are being reported for a different cohort for students (i.e., those who graduated 5 or 10 years ago) than the cohort for which median borrowing levels are reported (the most recent cohort of graduates for which data are available). Since college costs can change dramatically over time, we believe that median debt from the most recent cohort of graduates will more closely approximate what a current or prospective student might need to borrow, whereas the amount a student borrowed many years ago may not be meaningful if the tuition and fees are considerably higher now or the demographics of students served have shifted over time (such as because the institution has become more or less selective over time).

The Department does not believe it has the authority to include in its MOU with the Department of Treasury a request for institutions to provide Social Security numbers for non-title IV participants in order to include their earnings data in the Scorecard. We will continue to explore what options, if any, might be available to us so that non-title IV students can be included in Scorecard.

The Department agrees that calculators and financial management tools can be useful to students. Already, the Department has debt calculators on the FSA website, and as the Department launches the NextGen Financial Services Environment, it will include additional borrower education opportunities. We will explore ways to connect those tools to the College Scorecard so that students can manipulate data from the Scorecard as part of their exploration.

The Department is not suggesting that the College Scorecard replace person-to-person meetings or conversations between campus staff and prospective students and does not intend for the College Scorecard to replace those interactions. We do believe, however, that students who have access to the Scorecard, and who receive Scorecard information as they complete their FAFSA, will be able to identify which institutions they may want to attend and to enable outcomes comparisons between institutions that serve demographically matched populations or that support similar educational missions. Our goal is to go beyond a passive website and to connect Scorecard to the MyStudentAid mobile app so that Scorecard data becomes part of the experience and not an ancillary tool that students may or may not utilize.

While the Department encourages all institutions to post links to the Scorecard on their institutional websites and likes the idea of developing a recognizable icon so that students know where to find the link, we are not including those requirements here. We believe that by linking the College Scorecard to electronic or mobile FAFSA completion, and by providing Scorecard data in an API format so that others, such as Google, can develop new ways to make these data available to consumers, more students will interact with these data and have the opportunity to use them in their personal decision-making process.

The Department agrees that if institutions are left on their own to calculate and disclose their own outcomes, the data may be less accurate and reliable since different data sources could be used to produce those data, since human error could be introduced, and since dishonest institutions could misrepresent the truth. However, it must be noted that IPEDS data are similarly self-reported, and the Department has often pointed out its concern about the likely presence of errors in those data. Still, IPEDS reporting is the best data available to the Department, and we believe that as those data become more readily available to students for use in enrollment decision-making, institutions will be incentivized to further assure the accuracy of those data.

Still, the Department believes that the best way to provide accurate and comparable data to students and parents is to expand the College Scorecard to provide program-level outcomes data for title IV participating programs at all credential levels and regardless of institutional type. We agree with the commenter who stated that a centralized tool like the College Scorecard will be easier to update than websites and catalogs.

We appreciate the commenter who suggested that Outcome Measures Survey data be included in Scorecard, which has more comprehensive graduation rate information including rates for non-first time and part-time students, and the Department will take this recommendation under advisement as it develops the expanded Scorecard.

The Department acknowledges that disclosures are often made available to consumers making large financial transactions. We nonetheless believe that the College Scorecard is the optimal way to share information to student and to ensure that comparable data are made available to students and parents. The Department will explore the possibility of separating debt and earnings data for Pell and non-Pell students at the program-level by examining to what extent these data can be made available while maintaining student privacy.

As for concerns about data privacy, the Department notes that it receives earnings data in aggregate, not at the student level. Therefore, there was no potential for a breach of privacy regarding earnings. The Department has no plans of changing this policy and rescinding the regulation will not change any students' privacy safeguards, regardless of the size of the program in question.

The Department will continue to include information about institutional costs on the College Scorecard and will explore the feasibility of including program-level cost data. The Department has also explored calculating program-level completion rates for title-IV students but believes there will be challenges to creating entry cohorts because students can transfer from program to program within an institution, which makes it difficult to determine which students to include in an entry cohort. The Department is also exploring ways to provide information on program size to help students understand how competitive it might be to be admitted to, how many different class sections will be available, and how likely it is that the program is actually offered each semester. This will also help to reduce the use of tactics that lure a student to an institution and then redirect that student to a different program. The Department is concerned that some institutions may be advertising highly sought programs in order to attract students, but once students enroll at the institution, they then find that the program either is not enrolling more students, has entrance requirements substantially more rigorous than entrance requirements to the institution, or has a long waiting list, at which point the institution may then encourage them to enroll in a different program, such as a general studies program or a lower-level applied program. By publishing program size, students may get important clues about the likelihood of their program of choice being available to them. It may also help explain why proprietary institutions have entered into markets where the uninformed believe a community college is meeting career and technical training needs simply because they list having a program in their catalog.

The Department will consider the usefulness of IPEDS completion rate data to the Scorecard and appreciates the recommendations regarding the 100/150/200 percent completion rates. The Department does not have access to data that provides accurate information about the primary occupations for which a program prepares a student, and in non-CTE programs, it is difficult to determine what does or does not constitute a primary occupation. Therefore, we will likely not include information about primary occupations on the College Scorecard. Similarly, current plans do not include job placement rates because we do not have access to accurate data on this. Our goal is to encourage accreditors and states to stop relying on subjective, and error prone job placement rate determinations to evaluate program outcomes, and to instead encourage the use of College Scorecard earnings data to more accurately inform students about the earnings of prior graduates.

The Department is planning to include program-level information such as median debt, loan repayment rates, monthly payment associated with that debt, and cohort default rates in the Scorecard, although initially some of those data points may be calculated at the institution level rather than the program level. The Department does not have plans to include information about private loans in the College Scorecard, since we do not have access to those data without requiring institutions and students to report additional data to the Department.

The Department believes it has provided sufficient rationale for not including every element of the 2014 Rule disclosures in the expanded College Scorecard. However, we have described more generally throughout this document, and in this and the earlier section about GE disclosures, why we will no longer be requiring GE disclosures. Since our goal is to develop a transparency framework that can be applied to all categories (GE and non-GE) of title IV programs, we are concerned that such disclosures could be too burdensome to large institutions that offer hundreds of programs. Therefore, we will not require any institutions to post GE-type disclosures as a result of this final rule.

The Department plans to begin with annual updates to the College Scorecard and will consider whether more frequent updates are appropriate. College Scorecard will continue to adhere to the Department's privacy standards and suppress values with small cohort sizes and will consider aggregating data from multiple years if necessary, to achieve larger cohort sizes. The Department plans to engage in consumer testing of the College Scorecard.

We hope that more students will use the College Scorecard since we have mechanisms to disseminate data to students through the mobile app and other NextGen FSA tools. We also believe that by providing data in API format, other developers will find novel and innovative ways of making data available to students in a user-friendly format and in ways the Department is unlikely to explore with its own limited resources.

We agree that the College Scorecard will not prevent high pressure sales tactics or pain point recruiting, but it will provide information that makes it difficult for institutions to misrepresent the truth about their outcomes. By rescinding this rule, we are making no changes to the incentive compensation regulations; therefore, we are not proposing any changes to prohibitions on commission-based compensation.

We will work towards expanding the College Scorecard to include programs-level metrics, including for certificate programs, undergraduate programs, graduate programs and professional programs. The Department is not currently planning to separate total debt from debt associated with tuition and fees; however, we will continue to consider the request to do so.

The Department plans to continue providing institution level information to help students understand the impact of variables, such as geographic differences, on outcomes. In addition, other contextual information, such as institutional selectivity or percent of Pell recipients to help students compare similar institutions. We will consider ways in which we might interact with other databases, such as credit bureau data or student outcomes data.

The Department has negotiated in good faith and has committed to updating and expanding the College Scorecard. Since we are still developing the tool and are not required to publish regulations in order to produce the College Scorecard, we will not commit to all of the particulars of its content in this final regulation. However, we will consider the recommendations we received through the public comments as we update and expand the College Scorecard. The Department will continue to enforce disclosure and reporting requirements that remain part of the PPA. In addition, the Department will continue to be mindful of the reporting burdens placed upon institutions for all reporting or disclosure requirements.

Certification of GE Programs

Comments: One commenter stated that institutions of higher education should be required to certify programs that lead to careers with State licensure requirements actually meet those State licensure standards.

Discussion: The Department considered disclosures related to licensure and certification, as well as accreditation, as part of its Accreditation and Innovation negotiated rulemaking package and, therefore, will not include regulations related to disclosures of this information in this rulemaking.

Changes: None.

Continued Implementation of the GE Regulations Prior to Rescission

Comments: One commenter representing a coalition of members of advocacy groups stated that until a rescission of the 2014 Rule is effective, the Department is obligated to follow the law as it exists but has failed to do so.

Alternately, two commenters requested that the Department suspend any further requirements to comply with the GE regulations, including the GE data reporting requirements, publication, or revisions to the disclosure template, and requirements to submit appeals information.

Discussion: The GE regulations remain in effect until this regulation is final and the 2014 Rule is rescinded. However, the Department does not have access to the SSA earnings data necessary to calculate future D/E rates. As a result, the Department cannot take action to remove programs from title IV participation since no program will have failed the D/E rates measure for two out of three consecutive years or had a combination of fail and zone rates for four consecutive years. The Department will produce a modified disclosure template that institutions must use to disclose information, as prescribed by the GE regulations.

Changes: None.

Rulemaking Process

Comments: One commenter stated that the Department did not conduct a reasoned rulemaking since it has proposed to eliminate all sanctions. One commenter stated that the proposed regulations are arbitrary and capricious, because the Department failed to justify its regulatory choices. Specifically, the commenter referred to the removal of the sanctions for poor-performing programs and the removal of disclosures to students about program outcomes. The commenter stated that Executive Order 12866 was not followed because the Department did not issue a regulation where the benefits of the new policy outweigh the costs. The commenter also stated that the Department has not presented rigorous analysis and evidence to support its claims.

A commenter stated that the Department did not negotiate in good faith because it refused to hold a fourth session of negotiations after tentative consensus on the proposal was reached.

One commenter accused the Department of ignoring and disregarding years of public input on GE matters.

One commenter provided an appendix in which he quoted from the 2014 NPRM but did not provide a comment to explain its inclusion. The commenter also provided research by Libassi and Miller about how the GE regulations reduce loan forgiveness costs, but again did not provide any explanation as to its inclusion.^[142]

Discussion: The Department disagrees with the commenter who asserted that the Department is advancing a policy where the risks outweigh the benefits. Throughout the NPRM, and this document, we have provided sufficient evidence that the benefits of the final regulation—including ensuring that all students are free to choose the school and program of their choice—outweigh the risks. In fact, we have been clear that by expanding the College Scorecard to improve program-level outcomes data for all title IV-participating programs, we will expand the benefits of transparency to all students and not just those who seek enrollment in a GE program. The Department also disagrees with the commenter who said that it did not provide rigorous analysis to support its position. The Department has provided a more than rigorous review of data that was not considered in connection with the 2014 Rule and disagrees with earlier claims.

The Department disagrees with the suggestion that it did not conduct a good faith, open, and reasoned rulemaking. The Department proposed the removal of sanctions at the first negotiating session, explaining that the numerous sources of error in the D/E rates measure make it an invalid proxy for program quality. Nonetheless, when a negotiator proposed the use of one-to-one debt-to-earnings ratios that would be more easily understood by students, the Department supported this approach and voted favorably.

Although the Department hoped for consensus among the members of the negotiating committee, it was not reached. A number of negotiators, including representatives of non-profit institutions, discussed the many reasons why sanctions are not appropriate based on the inaccuracies of the D/E rates measure as a proxy for quality since the rates may be influenced by many factors outside of the institution's control. The Department believes it is inappropriate to sanction institutions and eliminate opportunities for students based on metrics that are influenced by factors outside of the control of institutions, such as student loan interest rates.

The Department also disagrees with the assertion that a program that fails the D/E rates measure is automatically and necessarily a poor performing program. As noted in the NPRM, there are a plethora of factors that influence a program's D/E rates. As such, the Department does not believe that failing the D/E rates measure is an accurate indicator that the program is a poor performing program. In addition, given the number of passing programs that have associated earnings below the poverty level, the Department does not believe that passing the D/E rates measure indicates that the program is a good program or that students are benefiting themselves by completing it.

The Department also believes that stewardship of taxpayer dollars includes providing information that allows taxpayers to understand not only the number of dollars at risk through the student loan program, but the number of dollars that are directed through State and local appropriations to programs that yield low earnings. Students also have the right to know, regardless of whether they pay cash, use other forms of credit, or use Federal student loans to pay for their programs, if doing so is likely to generate financial benefits. Employers similarly should be able to review program outcomes before spending their hard-earned dollars to provide employee education and professional development. Therefore, the Department believes that its decision to use the College Scorecard or its successor as the mechanism to increase transparency and inform a market-based accountability system that continues to honor student choice is reasonable. The Department recognizes that students select institutions and programs, including GE and non-GE programs, for many different reasons, of which future earnings may be only one of many deciding factors.

Even without currently having access to all program-level data for non-GE programs, as stated elsewhere, the Department believes that the benefits of rescinding the GE regulations outweigh the potential costs, since GE programs represent just a small portion of title IV programs available to students. In order to ensure that all students make better informed enrollment and borrowing decisions, a comprehensive approach is required. Because the Department does not yet have access to program-level data, we cannot accurately estimate savings associated with reduced enrollments in undergraduate and graduate programs across all institutional sectors as a result of unimpressive outcomes.

The Department's review of the outstanding student loan portfolio has provided ample evidence that the problem of borrowing more than a student can repay in 10 years extends well beyond proprietary institutions and includes institutions from all sectors. According to Jason Delisle and Alex Holt, income-driven repayment programs actually provide disproportional advantage to higher income students, which is not the population for whom IDR programs were designed.^[143] Student loan non-repayment poses considerable costs to taxpayers, regardless of which institutions are the source of loans in non-repayment. While the Department did not approve of a fourth negotiating session, we believe we engaged in a good faith effort to negotiate and reach consensus. The Department does not believe that there was tentative consensus on the proposal during the third session or that a fourth session would have brought the group closer to consensus. To the contrary, the Department made considerable compromises in order to arrive at consensus, but it was clear by the end of the third session that consensus would not be achieved. Also, a number of negotiators expressed opposition to the idea of adding another session. There were several negotiators who made it clear that they would never concur with any regulation that did not include program sanctions and one negotiator stated that he would never agree to a regulation without first knowing which programs would pass or fail, so that he could be sure that only the truly “bad” programs would fail, since some “good” programs could fail if the formula was not properly designed.

The Department believes that it is not appropriate to evaluate the validity of a methodology by reviewing the results to see if they align with a more subjective view of which programs should pass or fail. Either the methodology is valid, or it is not, and while it would be helpful to know which and how many programs would be impacted by a valid methodology, those results are not what determine the accuracy of the methodology. The Department acknowledges that it was able to provide only very limited data to negotiators and could not provide earnings data for non-GE programs since the Department was unable to obtain additional earnings data from SSA. However, neither negotiators nor the Department could identify a new accountability metric that is supported by research and appropriately controls for factors that impact student debt or program earnings. Further, additional data were not needed to develop the methodology. Rather, additional data would have only enabled negotiators to determine which programs would be on the “right” side of the formula.

The Department negotiated in good faith, including putting forth a proposal during the third session that deviated significantly from our original proposal and took into account many of the suggestions made by negotiators. However, even with all of those changes, consensus was not reached. From the time that the negotiated rulemaking committee was announced, negotiators knew that the Department was planning to hold three negotiating sessions. Three sessions provided ample opportunity to fully discuss the issues and determine whether consensus could be reached.

Discussion has continued about the GE regulations since the first rulemaking effort commenced in 2010, and that discussion continued through a second rulemaking effort and this current negotiated rulemaking and public comment. The Department does not believe that uniform consensus about the validity of the GE regulations has ever been achieved, and it notes that there has been vociferous disagreement among those who support and those who oppose the 2014 Rule.

More recently, we have been unable to enter into an updated MOU with SSA, which means that we are unable to obtain earnings data to continue calculating D/E rates. Therefore, the Department has no choice other than to cease D/E calculations and reporting using the methodology defined by the GE regulations. Most importantly, the GE regulations cannot be expanded to include all title IV programs. The Department has determined that the 2014 Rule is fundamentally flawed and does not provide a reliable methodology for identifying poorly performing programs and, therefore, should not serve as the basis for high stakes sanctions that negatively impact institutions and students.

Changes: None.

Information Quality Act (IQA)

Comments: A commenter stated that the NPRM relied upon “inaccurate, misleading, and unsourced information in violation of the Information Quality Act.” Additionally, the commenter stated that the Department did not meet the clear standards set forth in both the ED Guidelines related to the IQA and the IQA itself because the data and research cited lacked objectivity since the NPRM was filled with examples of information that was not supported by sources, do not stand for the proposition cited, failed to explain the methodology used, or were not accompanied by information that allows an external user to understand clearly the analysis and be able to reproduce it, or understand the steps involved in producing it.

Discussion: The Department separately addresses each of the specific comments and requests related to compliance with the IQA below.

Changes: None.

Comments: A commenter questions the Department's statement “The first D/E rates were published in 2017, and the Department's analysis of those rates raises concerns about the validity of the metric, and how it affects opportunities for Americans to prepare for high-demand occupations in the healthcare, hospitality, and personal services industries, among others.” The commenter stated that this assertion fails to clearly describe the research study approach or data collection technique, fails to clearly identify data sources, fails to confirm and document the reliability of the data and acknowledge any shortcomings or explicit errors, fails to undergo peer review, and fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department is referring to data tables published on the Department's website, based upon the methodology described in the 2014 Rule.^[144] Our statement in the NPRM was based upon our analysis of the data in the published D/E rates data table, as discussed above in the Geographic Disparities and the D/E Thresholds and Sanctions sections.

Changes: None.

Comments: A commenter questioned the Department's statement “In promulgating the 2011 and 2014 regulations, the Department cited as justification for the eight percent D/E rates threshold a research paper published in 2006 by Baum and Schwartz that described the eight percent threshold as a commonly used mortgage eligibility standard. However, the Baum Schwartz paper makes clear that the eight percent mortgage eligibility standard `has no particular merit or justification' when proposed as a benchmark for manageable student loan debt. Upon further review, we believe that the recognition by Baum and Schwartz that the eight percent mortgage eligibility standard `has no particular merit or justification' when proposed as a benchmark for manageable student loan debt is more significant than the Department previously acknowledged and raises questions about the reasonableness of the eight percent threshold as a critical, high-stakes test of purported program performance.” The commenter states that the Department fails to present conclusions that are strongly supported by the data, which has been highlighted recently by Sandy Baum, the co-author of the 2006 study cited by the Department, who stated that “the Department of Education has misrepresented my research, creating a misleading impression of evidence-based policymaking. The Department cites my work as evidence that the GE standard is based on an inappropriate metric, but the paper cited in fact presents evidence that would support making the GE rules stronger.” The commenter further asserts that “[the Department is] correct that we were skeptical of [the eight percent] standard for determining affordable payments for individual borrowers, but incorrect in using that skepticism to defend repealing the rule. In fact, our examination of a range of evidence about reasonable debt burdens for students would best be interpreted as supporting a stricter standard.”

Discussion: The Department is aware of and respects Ms. Baum's opinion that the 2014 Rule should not be rescinded. However, that does not change the fact that in their earlier paper, Baum's and Schwartz's state that the eight percent mortgage eligibility standard has “no particular merit or justification” as a benchmark for manageable student loan debt. Since this paper was cited in the 2014 Rule as the source of the eight percent threshold, it is relevant that even the authors of the paper are skeptical of the merit of the 8 percent threshold as a student debt standard. It is not only appropriate, but essential, that the Department points out that upon a more careful reading of the paper, we realize that the paper does not support the eight percent threshold, but instead clearly refutes it for the purpose of establishing manageable student loan debt. As for the notion that the Baum Schwartz paper supported a stricter standard, the commenter did state that the 2014 Rule was too permissive, but did not provide a specific threshold for what the number should be and the negotiating committee similarly was unable to identify a reliable threshold for the D/E rates measure.

Changes: None.

Comments: Several commenters expressed the opinion that research and evidence cited in the NPRM was misinterpreted by the Department or used selectively in an attempt to mislead. One commenter specifically asserted that the NPRM cites evidence in a way that leads to factual errors, does not attempt to justify key choices, and ignores hundreds of pages of evidence in favor of citations that have no bearing on the claims asserted. Another commenter offered that the 2014 Rule is based on extensive research and evidence, which the NPRM fails to adequately refute, showing that some GE programs were accepting Federal financial aid dollars and enrolling students while consistently failing to train and prepare those students for employment.

Discussion: The Department disagrees with the commenter's interpretation of the data provided in the NPRM. We continue to believe that the NPRM included adequate justification for its conclusion that the D/E rates measure is an unreliable proxy for program quality for all of the reasons described, including that the Department's selection of an amortization term that could significantly skew pass or fail rates, and the Department's selection of a 10-, 15-, or 20-year amortization term that does not align with the amortization terms provided by Congress and the Department through its various extended and income-based repayment programs.

Similarly, the Department has provided sufficient evidence to support its position that while program quality could have an impact on earnings, so too could a variety of other factors outside of the institution's control, including discriminatory practices that have resulted in persistent earnings gaps between men and women, between individuals from underrepresented minority groups and whites; geographic differences in prevailing wages; difference in prevailing wages from one occupation to the next; micro- and macro-economic conditions; and other factors.

Changes: None.

Comments: One commenter disagreed with the Department's statement that, “Research published subsequent to the promulgation of the GE regulations adds to the Department's concern about the validity of using D/E rates to determine whether or not a program should be allowed to continue to participate in title IV programs.” The commenter believed that the Department failed to identify data sources, including whether a source is peer-reviewed and scientific evidence-based, failed to confirm and document the reliability of the data and acknowledge any shortcomings or explicit errors, and failed to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department has used well-respected, peer-reviewed references to substantiate its reasons throughout these final regulations for believing that D/E rates could be influenced by a number of factors other than program quality. As such, the D/E rates measure is scientifically invalid because it fails to control or account for the confounding variables that could influence the relationship between the independent (program quality) and dependent variable (D/E rates) or render the relationship between the independent and dependent variables as merely correlative, not causal.

Changes: None.

Comments: One commenter disagreed with the Department's assertion that “the highest quality programs could fail the D/E rates measures simply because it costs more to deliver the highest quality program and as a result the debt level is higher.” The commenter stated that the Department “Fails to identify data sources and fails to be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: As stated above, where a higher quality program requires better facilities, more highly qualified instructors, procurement of expensive supplies, small student-to-teacher ratios, and specialized equipment to provide high-quality education, someone must pay the cost. Although taxpayers may pay some of these costs on behalf of students enrolled at public institutions, private institutions typically pass all or most of these costs on to students, which results in high tuition. However, there is no correlation between the cost to deliver a high-quality education and wages paid to program graduates. The Department cites research from CSU Sacramento that serves as evidence that high quality career and technical education programs can be more than four times as expensive to run as general studies programs.^[145]

Changes: None.

Comments: One commenter disagreed with the Department's statement that, “Other research findings suggest that D/E rates-based eligibility creates unnecessary barriers for institutions or programs that serve larger proportions of women and minority students. Another commenter claimed that studies demonstrated that rescinding the 2014 Rule could exacerbate gender and race wage gaps. Such research indicates that even with a college education, women and minorities, on average, earn less than white men who also have a college degree, and in many cases, less than white men who do not have a college degree.” The commenter went on to state that the Department fails to draw upon peer-reviewed sources, fails to acknowledge any shortcomings or explicit errors in the data, fails to present conclusions that are strongly supported by the data. The commenter stated that the source cited by the Department does not draw the same conclusion as the Department reached. For example, the cited table appears to relate to graduates of bachelor's degree programs, and not gainful employment programs. The commenter also states that the statement fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department emphasizes that bachelor's degree programs are included as GE programs if offered by proprietary institutions. Moreover, the NPRM cites data provided by the College Board that points to disparities in earnings between men and women and people of color. The College Board is a reliable and trusted source of data, and its publications undergo rigorous peer review prior to publication. The citation provided links to the College Board's report and data tables, which are robust, and which include information about data sources and methodology used.

The data sourced from the U.S. Census Bureau's Current Population Survey which calculated median earnings based on race/ethnicity, gender and educational level, includes disaggregated earnings based on other characteristics, such as having less than a high school diploma, a high school diploma, some college, no degree, associate degree, bachelor's degree, and advanced degree. While this research did not address GE programs specifically, the point is that there are general earnings disparities based on race and gender. Programs that serve large proportions of women and minorities, therefore, would likely post lower earnings than programs of similar quality primarily serving whites and males, simply because of wage advantages certain groups have had for centuries. The Department agrees that our statement is an extrapolation of the data provided, but this extrapolation is well reasoned and supported by other research. Given that proprietary institutions serve the largest proportions of women and minority students, and that some GE programs (such as those in medical assisting, massage therapy, and cosmetology) serve much larger proportions of female students, it is likely that student demographics will impact earnings among these programs. This is not an unreasonable extrapolation to make, since the impact of gender and race on earnings is well-documented and the subject of considerable policy discussion and public debate.

Changes: None.

Comments: A commenter has concerns about the Department's statement “[D]ue to a number of concerns with the calculation and relevance of the debt level included in the rates[,] we do not believe that the D/E rates measure achieves a level of accuracy that it should [to] alone determine whether or not a program can participate in title IV programs.” The commenter states that the Department fails to clearly describe the research study approach, fails to identify data sources, fails to confirm and document the reliability of the data, fails to undergo peer review, fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: As was discussed during the 2014 negotiations and continuing through the more recent negotiations, public hearings, and public comment, the debt metric can change significantly depending upon the amortization term used, interest rates and congressionally determined student loan lending limits. No research is needed to show that a student in a 20-year repayment plan will pay a lower monthly and annual payment than one in a 10-year repayment plan as this is a well understood mathematical fact. Since REPAYE created an opportunity for all students to qualify for a 20- to 25-year repayment term, depending upon their credential level attainment, it is unreasonable to use a 10- or 15-year amortization period to calculate the annual cost of student loan repayment just because GE programs tend to serve a larger proportion of non-traditional students. Even if using a 10-year repayment term was justified for certificate or associate degree programs, which we do not believe is the case, there is no possible justification that borrowers in bachelor's programs should be evaluated based on a 15-year amortization period whereas students who complete the same credentials at non-profit and private institutions can qualify for 20-, 25-, or even 30-year repayment terms based on the level of their degree and the amount they owe. The Department sees no basis for such a double standard.

The Department does not believe it is appropriate to use REPAYE as the tool to help some students manage a debt load disproportionate to their earnings, imposing no sanctions on the institutions that led the borrower to this position, while penalizing other institutions by eliminating a program because the students who need income driven repayment assistance happened to graduate from a school that pays taxes rather than consuming direct taxpayer subsidies. The 2015 REPAYE regulations, coupled with the gainful employment rule, established a double standard that sanctions proprietary institutions if their graduates need income driven repayment programs to repay their loans, and promises graduates of non-profit institutions income-based repayment and loan forgiveness in return for irresponsibly borrowing.

Changes: None.

Comments: One commenter has concerns with the Department's statement “[I]ncreased availability of [income-driven] repayment plans with longer repayment timelines is inconsistent with the repayment assumptions reflected in the shorter amortization periods used for the D/E rates calculation in the GE regulation.” The commenter states that the Department fails to rely upon peer-reviewed, scientific evidence-based research, fails to identify data sources, fails to confirm and document the reliability of the data, fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: This comment is a statement of fact, which is substantiated by information provided on the Federal Student Aid website.^[146]

Changes: None.

Comments: One commenter raised issues about the Department's statement “[A] program's D/E rates can be negatively affected by the fact that it enrolls a large number of adult students who have higher Federal borrowing limits, thus higher debt levels, and may be more likely than a traditionally aged student to seek part-time work after graduation in order to balance family and work responsibilities.” The commenter continued that the Department fails to rely upon peer-reviewed, scientific evidence-based research, fails to identify data sources, and fails to confirm and document the reliability of the data.

Discussion: It is a statement of fact that independent students have higher Federal loan borrowing limits, because Congress has established those higher limits for independent students (which include students over the age of 25, graduate students, married students, and students with dependents).^[147] Independent students can borrow up to $57,500 for undergraduate studies whereas dependent students can borrow only $31,000. Simple mathematics explain that if a larger proportion of students can borrow $57,500 rather than $31,000 to complete a bachelor's degree, the median debt level will be higher at an institution that serves a large portion of independent students than dependent students.^[148] As Baum points out in her 2015 publication, 70 percent of students who hold student loan debt of $50,000 or more are independent students. This is not a surprising fact since it is only those students who have borrowing limits over $50,000. These datasets are derived from NCES data reports and were compiled by Sandy Baum.

Therefore, it is not surprising that institutions serving larger proportions of independent students will have higher median borrowing levels, and since proprietary institutions serve the highest portion of independent students, it is not unreasonable that these institutions would have higher median debt levels, which they do.

Data reported by Pew proves that the percentage of college graduates who work part-time rather than full-time increased from 15 percent in 2000 to 23 percent in 2011. We have addressed concerns about data regarding adult students working part-time and the gender gap in earnings earlier in these final regulations. Research provided by the Center for American Progress substantiates that even among college graduates, women tend to earn less than men, in part because they tend to select lower paying majors and in part because of time spent out of the workforce raising children.^[149] The Pew Research Center confirms that a higher percentage of women take time out of their career or work part-time because of child-rearing responsibilities.^[150]

Changes: None.

Comments: One commenter raised issues about the Department's statement “[I]t is the cost of administering the program that determines the cost of tuition and fees.” The commenter continued that the Department fails to rely upon peer-reviewed, scientific evidence-based research, fails to identify data sources, fails to confirm and document the reliability of the data, fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department did not state that it is the cost of administering academic programs that determines tuition and fees. To the contrary, the Department made clear in the NPRM that at most non-profit institutions, direct taxpayer appropriations and tuition surpluses generated from the low-cost programs the institution administers are used to offset the financial demands of higher cost programs. In this case, the cost of administering the program does not directly drive the cost of tuition and fees. Were that the case, liberal arts programs would charge lower tuition and fees than laboratory science and clinical health sciences programs—which is not the case at most non-profit institutions. Instead, what the NPRM said is that in some cases, the cost of tuition and fees is driven by the higher cost of administering some programs. The Shulock, Lewis and Tan study provides peer reviewed research to support this position.^[151]

Changes: None.

Comments: One commenter raised concerns about the Department's statement “Programs that serve large proportions of adult learners may have very different outcomes from those that serve large proportions of traditionally aged learners.” The commenter continued that the Department fails to rely upon peer-reviewed, scientific evidence-based research, fails to identify data sources, fails to confirm and document the reliability of the data, fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department offers as evidence to support the statement made in the NPRM data from the NCES Study of Persistence and Attainment of Nontraditional Students.^[152] NCES is a reliable and trusted source of higher education data.

Changes: None.

Comments: One commenter raised issues about the Department's statement “Data discussed during the third session of the most recent negotiated rulemaking demonstrated that even a small change in student loan interest rates could shift many programs from a `passing' status to `failing,' or vice versa, even if nothing changed about the programs' content or student outcomes.” The commenter continued that the Department fails to clearly describe the research study approach and data collection technique, fails to identify data sources, fails to confirm and document the reliability of the data, fails to undergo peer review, fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department points the commenter to our website, where data provided by the negotiator during the third negotiating session show the change in outcomes based on a small shift in interest rates.^[153] The negotiator is an economist at Columbia University, Cornell University, and the Urban Institute, and is thus a trusted source of data. However, any loan amortization table will show that when interest rates change, payments on debt increase. Again, this is a basic mathematical fact that requires no statistical study or peer review to be proven true.

Changes: None.

Comments: One commenter challenged the Department's statement “There is significant variation in methodologies used by institutions to determine and report infield job placement rates, which could mislead students into choosing a lower performing program that simply appears to be higher performing because a less rigorous methodology was employed to calculate in-field job placement rates.” The commenter continued by stating the Department fails to clearly describe the research study approach and data collection technique, fails to clearly identify data source, fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department cited in the NPRM the findings of the Technical Review Panel (TRP), convened in response to the 2011 GE regulations to address the confusion created by multiple job placement rate definitions. This TRP is a trusted source, as is the external research that was retained to provide background research on job placement rates.^[154]

Changes: None.

Comments: One commenter raised concerns about the Department's statement “The Department also believes that it underestimated the burden associated with distributing the disclosures directly to prospective students. A negotiator representing financial aid officials confirmed our concerns, stating that large campuses, such as community colleges that serve tens of thousands of students and are in contact with many more prospective students, would not be able to, for example, distribute paper or electronic disclosures to all the prospective students in contact with the institution.” The commenter continued that the Department fails to draw upon peer-reviewed, scientific-evidence based research and fails to confirm and document the reliability of the data.

Discussion: The Department continues to assert that the negotiator who made this statement is a reliable authority on the burden institutions would face if required to distribute disclosures. The point of having negotiators is to consider the opinions of experts in the field. However, the Department did not require the negotiator to provide data to substantiate her claim. Nonetheless, while the Department cited regulatory burden as a contributing factor to its decision to rescind the GE regulations, it was not the primary reasons for making this decision. The primary reason for rescinding the GE regulations, as stated earlier, is evidence that the D/E rates measure is not a reliable proxy for quality since many factors other than quality can impact both the debt and earnings elements of the equation.

Changes: None.

Comments: One commenter raised concerns about the Department's statement “The Department believes that the best way to provide disclosures to students is through a data tool that is populated with data that comes directly from the Department, and that allows prospective students to compare all institutions through a single portal, ensuring that important consumer information is available to students while minimizing institutional burden.” The commenter continued that the Department fails to draw upon peer-reviewed, scientific evidence-based research and fails to identify data sources. Specifically, in the 2014 Rule, the Department stated that it “would conduct consumer testing” to determine how to make student disclosures as meaningful as possible. The NPRM fails to acknowledge whether such testing occurred, including the results of that testing. The NPRM also fails to state any other basis for the Department's conclusions.”

Discussion: The Department did conduct consumer testing on the disclosure template after the 2014 Rule went into effect, the results of which proved that disclosures are typically very confusing to students, that the results presented are frequently misinterpreted, and that in general, students find disclosures most meaningful when they provide information about the students included in the disclosures, including what course loads the students were taking.^[155] The Department points to a number of commenters who said that the current GE disclosures can be difficult to find on institutional websites, which the Department has found to be the case in its own attempts to identify GE disclosures when reviewing websites. In addition, the Department points to statutory requirements for the College Navigator which emphasize the importance of using a standardized data tool to provide comparable data to students and that allow students to compare multiple institutions.^[156]

Changes: None.

Comments: One commenter raised issues about the Department's statement “[T]he Department does not believe it is appropriate to attach punitive actions to program-level outcomes published by some programs but not others. In addition, the Department believes that it is more useful to students and parents to publish actual median earnings and debt data rather than to utilize a complicated equation to calculate D/E rates that students and parents may not understand and that cannot be directly compared with the debt and earnings outcomes published by non-GE programs.” The commenter continued that the Department fails to draw upon peer-reviewed, scientific evidence-based research and fails to identify data sources.

Discussion: Elsewhere in this document, the Department has provided adequate support for its assertion that the D/E rates measure is not sufficiently accurate or reliable to serve as the sole determinant of punitive action against a program or institution. The Department conducted significant consumer testing prior to the launch of the College Scorecard to better understand which data are most relevant to students and parents and will continue to conduct consumer testing. However, the Department is committed to providing data that can reduce the reporting burden to institutions while still providing additional information to students.

Changes: None.

Comments: One commenter challenged the Department's statement “The Department has reviewed additional research findings, including those published by the Department in follow-up to the Beginning Postsecondary Survey of 1994, and determined that student demographics and socioeconomic status play a significant role in determining student outcomes.” The commenter continued that the Department fails to identify data sources. Specifically, the website cited by the Department links to the Beginning Postsecondary Survey of 1994's findings, and not the “additional research” mentioned by the Department, including the Department's own “follow-up.” Additionally, the Department fails to confirm and document the reliability of the data, and fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department misstated the name of the reference from which it drew data regarding outcomes of non-traditional students. The NPRM should have said that “The Department has reviewed additional research findings, including the 1994 follow-up on 1989-90 Beginning Postsecondary Survey, which determined that student demographics and socioeconomic status play a significant role in determining student outcomes.” Other research reviewed included publications by the American Association of Colleges and Universities on the needs of adult learners,^[157] a publication about Adult Learners in Higher Education produced by the U.S. Department of Labor ^[158] and another research study that focused specifically on the needs of adult learners enrolled in online programs.^[159]

Changes: None.

Comments: One commenter raised issues with the Department's statement “The GE regulation failed to take into account the abundance of research that links student outcomes with a variety of socioeconomic and demographic risk factors.” The commenter continued that the Department fails to identify data sources and fails to confirm and document the reliability of the data.

Discussion: This sentence refers to the same NCES study referenced in the NPRM and above.

Changes: None.

Comments: One commenter raised concerns about the Department's statement that “the GE regulation underestimated the cost of delivering a program and practices within occupations that may skew reported earnings. According to Delisle and Cooper, because public institutions receive State and local taxpayer subsidies, even if a for-profit institution and a public institution have similar overall expenditures (costs) and graduate earnings (returns on investment), the for-profit institution will be more likely to fail the GE rule, since more of its costs are reflected in student debt. Non-profit, private institutions also, in general, charge higher tuition and have students who take on additional debt, including enrolling in majors that yield societal benefits, but not wages commensurate with the cost of the institution.” The commenter stated that the study mentioned did not support the conclusion that the GE regulations underestimated the cost of delivering a program and the NPRM failed to identify the data sources.

Discussion: The Department relied on the Delisle and Cooper's research and analysis to substantiate that public institutions are often able to charge less for enrollment than private and proprietary institutions because they receive direct appropriations from a State or local government, are not required to purchase or rent their primary campus buildings or land, and enjoy substantial tax benefits. As such, they can charge the student a lower price for a program that has similar overall expenditures as another program sponsored by a private institution that does not receive direct subsidies, have endowment holdings, or benefit from preferential tax treatment. Specifically, Delisle and Cooper state that “[o]ne shortcoming of the 2014 Rule is that it does not take into account society's full investment in credentials produced by public institutions of higher education.” ^[160] As noted in their research, the data sources used by Delisle and Cooper were Department GE Data and data from IPEDS.

Changes: None.

Comments: A commenter raised concerns about the Department's statement “In the case of cosmetology programs, State licensure requirements and the high costs of delivering programs that require specialized facilities and expensive consumable supplies may make these programs expensive to operate, which may be why many public institutions do not offer them. In addition, graduates of cosmetology programs generally must build up their businesses over time, even if they rent a chair or are hired to work in a busy salon.” The commenter continued that the Department fails to identify data sources and fails to confirm and document the reliability of the data.

Discussion: Our statement was intended to give further examples of ways that cosmetology programs have been challenged in implementing the GE regulations. The Department received these comments from multiple commenters in connection with the 2014 Rule, as well as this rulemaking, and heard these arguments from negotiators and speakers at negotiations and other public forums.

It is unclear why public institutions do not operate cosmetology programs in greater numbers, but NCES data point to the limited number of enrollments in cosmetology programs among public colleges and universities. It is well known that cosmetologists typically must build their own clientele, even when working in a salon owned by another operator, and that tip income is an important part of the total earnings of cosmetologists. As a blog posted by a cosmetology program explains, if an individual does not make an effort to get clients, the individual may “have to sit around for hours waiting for a client to walk in and this is likely to affect your income. On the other hand, if you have reliable repeat customers, you can make sure that you have a steady stream of income throughout the year.” ^[161]

Changes: None.

Comments: One commenter raised concerns with the Department's statement “[S]ince a great deal of cosmetology income comes from tips, which many individuals fail to accurately report to the Internal Revenue Service, mean and median earnings figures produced by the Internal Revenue Service underrepresent the true earnings of many workers in this field in a way that institutions cannot control.” The commenter continued that the Department fails to present conclusions that are strongly supported by the data. The commenter noted that the Internal Revenue Service (IRS) tax gap study cited by the Department does not support the Department's specific conclusions about cosmetology graduates as it is from 2012 and covers tax year 2006 only. Additionally, the commenter stated that the Department failed to confirm and document the reliability of the data.

Discussion: Throughout the 2014 and 2018 negotiations, as well as between those negotiations, the Department has heard from cosmetology programs and their representatives on this matter. These stakeholders have regularly informed the Department that cosmetologists regularly under-report their earnings and hide a portion of their tipped earnings. In the 2014 Rule, the Department admitted that individuals who work in barbering, cosmetology, food service, or web design may under report their income (79 FR 64955) and hoped that the alternate earnings appeal would provide an opportunity to correct earnings in those fields for the purpose of the D/E rates.^[162] However, the Department lost a lawsuit filed by the American Association of Cosmetology Schools (AACS) and is no longer able to deny earnings appeals based on the failure of institutions to meet the survey response rates dictated by the 2014 Rule.

Changes: None.

Comments: One commenter raised concerns about the Department's statement “While the GE regulations include an alternate earnings appeals process for programs to collect data directly from graduates, the process for developing such an appeal has proven to be more difficult to navigate than the Department originally realized. The Department has reviewed earnings appeal submissions for completeness and considered response rates on a case-by-case basis since the response rate threshold requirements were set aside in the AACS litigation. Through this process, the Department has corroborated claims from institutions that the survey response requirements of the earnings appeals methodology are burdensome given that program graduates are not required to report their earnings to their institution or to the Department, and there is no mechanism in place for institutions to track students after they complete the program. The process of Departmental review of individual appeals has been time-consuming and resource-intensive, with great variations in the format and completeness of appeals packages.” The commenter continued that the Department fails to present conclusions that are strongly supported by the data. The commenter notes that despite asserting that the alternate appeals process is “time-consuming and resource-intensive, with great variations in the format and completeness of appeals packages,” the Department then “estimates that it would take Department staff [only] 10 hours per appeal to evaluate the information submitted.” Additionally, the commenter states that the Department fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: The Department has received numerous inquiries about how to file an appeal, and the inquirers have expressed confusion, frustration, and have described excessive burden on their institutions (especially small institutions) in filing an appeal. Additionally, this has come up multiple times at public hearings, in comments received, and at the negotiations themselves. Institutions have had difficulty gathering the earnings information for their appeal because there is no formal mechanism in place for students to report their income to their programs. Even at 10 hours per appeal, the Department has insufficient resources to review appeals in a timely manner. Of the 326 appeals submitted in response to the 2014 earnings data, the Department has completed the review and rendered a decision on only 101 of those claims. Rescinding the regulations will mitigate the flaw in the D/E rates measure that is associated with underreported income or earnings appeals.

Changes: None.

Comments: One commenter raised concerns about the Department's statement “We believe that the analysis and assumptions with respect to earnings underlying the GE regulation is flawed.” The commenter continued that the Department fails to draw upon peer-reviewed, scientific evidence-based research and fails to confirm and document the reliability of the data.

Discussion: The Department has provided sufficient evidence to support the conclusion that the D/E rates measure is a flawed metric. As noted earlier, the Department is referring to a claim made in the 2014 Rule that graduates of many GE programs were earning less than those of the average high school dropouts.

Upon further review of the Department of Labor data used to make this claim, the Department has determined that the claim was inaccurate. First, the Department did not differentiate between program completers and program drop-outs in calculating earnings outcomes, which is inappropriate because program drop-outs will not reap the full benefits of the program. In addition, the figure used to represent the earnings of high school dropouts was derived by multiplying a weekly earnings figure by 52, assuming that all high school dropouts will work a full 52 weeks or benefit from paid vacation or sick leave during some of that time. However, the BLS report on Contingent Workers shows that individuals without a high school diploma are more likely to be part of the contingent workforce than the non-contingent workforce, meaning that they are more likely to have employment that is not expected to last or that is described as temporary.^[163] Therefore, calculating earnings for high school drops outs based on an assumption that high school drop outs work 52 weeks per year inflates the likely earnings of high school drop outs. Yet, in addition to not differentiating between program completers and program drop-outs, the inflated figure that assumed all workers work 52 weeks per year was compared to SSA earnings data for GE program graduates that included individuals working full-time, part-time, individuals who are self-employed, and those who may not report some or all of their earned income.

It is illogical that students would earn less after completing a postsecondary program than they would have had they not completed high school. Even if the postsecondary education provides zero earnings gains, the program graduate should earn a wage comparable with that of high school dropouts. Therefore, this conclusion defies logic, and was the result of a poorly designed comparison.

Changes: None.

Comments: One commenter raised issues with the Department's “Table 1—Number and Percentage of GE 2015 Programs That Would Pass, Fail, or Fall into the Zone Using Different Interest Rates.” The commenter stated that the Department fails to clearly describe the research study approach and data collection technique, fails to identify data sources, fails to confirm and document the reliability of the data, fails to undergo peer review, and fails to “be accompanied by supporting documentation that allows an external user to understand clearly the information and be able to reproduce it, or understand the steps involved in producing it.”

Discussion: “Table 1—Number and Percentage of GE 2015 Programs That Would Pass, Fail, or Fall into the Zone Using Different Interest Rates” from the NPRM illustrates how a change in interest rates would change the results of the 2015 GE rates, altering the number of programs that would pass, fail, or fall into the zone based on debt and earnings data published in 2015. Although the impact of a change in interest rates on the debt portion of the D/E calculation is obvious, these data were provided by a negotiator who is an economist at Columbia and Cornell Universities and the Urban Institute, and who was one of the designers of the College Scorecard during the Obama Administration. Although he built his own model to calculate the impact of changing interest rates, the source of the underlying debt and earnings data was provided by the Department in the data files provided along with the 2015 GE results.

Changes: None.

Comments: Several researchers submitted a joint comment opposing the rescission of the 2014 Rule. They argued that the rescission is arbitrary and capricious, because it ignores both the benefits of the 2014 Rule and the data analysis supporting the 2014 Rule. The commenters noted that Congress had reason to require that for-profit programs be subject to increased supervision. They cite a post on the Federal Reserve Bank of New York's blog that states that attending a four-year private for-profit college is the strongest predictor of default, even more so than dropping out. They cited evidence that students who attend for-profit institutions are 50 percent more likely to default on a student loan than students who attend community colleges. The commenters also argued that a rise in enrollment in the for-profit sector corresponded with reports of fraud, low earnings, high debt, and a disproportionate amount of student loan defaults. They cited an example that stated that, of the 10 percent of institutions with the lowest repayment rates, 70 percent were for-profit institutions. They argued that because poor outcomes are concentrated in for-profit programs, the 2014 Rule is justified.

Discussion: The Department does not disagree with the findings cited by some commenters, including the Federal Reserve Bank of New York's blog, but instead calls attention to the fact that these outcomes may be the result of the demographics of the students served rather than the quality of the educational program. A National Bureau of Economic Research (NBER) study of student loan repayment rates makes clear that race, financial dependency status and parental wealth transfer are the strongest predictors of default and non-repayment.^[164] Further, the Department's own research found that being over 25, having a child, being a single parent, and working full-time while in college are each factors that increase the risk of non-completion, and that the more risk factors a student demonstrates, the less likely the student is to complete the program and repay loans.^[165] Given that proprietary institutions serve a population of students that include a much higher percentage of Pell eligible, non-traditional and minority students, the results of these research papers are not surprising. The Department agrees with these researchers that non-profit institutions must do more to serve this population of students so that they enjoy the benefits of taxpayer subsidized tuition.

As discussed earlier, the majority of students enrolled in proprietary institutions is enrolled in bachelor's or graduate degree programs, not associate degree programs, making comparisons with community colleges irrelevant. In addition, since most proprietary institutions have open-enrollment policies, they cannot be compared directly with most public four-year institutions, that do not typically have open enrollment policies. These institutions are unique and serve a high-risk population. If other institutions are not willing to serve them, the question must be asked about whether or not these individuals should have the opportunity to go to college. The Department agrees that for many of these students, a work-based learning opportunity or a shorter-term training program could provide a more cost-effective option. However, apprenticeship programs are not open-enrollment opportunities, and many have considerable academic entrance requirements, including performance on mathematics tests. In addition, there are not enough of these opportunities to serve all interested participants.

It may be convenient to ignore the many confounding variables that impact student outcomes, and to ignore that the demographics of students enrolled at proprietary institutions are quite different than those of public or private non-profit two- and four-year schools, but the Department cannot ignore those facts, which our own data, published in 2017, substantiates.^[166]

The Department believes that more must be done to improve outcomes for high-risk students, and more options must be made available to students for whom college is not the best or preferred option, but in the meantime, the conclusion that institutional quality is the cause for lower outcomes is not substantiated by fact. There is clearly a crisis among minority students, with predictions for defaults among African American students to reach 70 percent in the next 20 years.^[167] It is true that defaults are higher among African Americans as compared to other demographics. It is also true that African Americans attend proprietary institutions in higher proportions than other demographics.

But the question is one of cause and effect. Do African American students default at higher rates because they attend proprietary institutions, or are default rates among proprietary institutions higher because these institutions are more likely to serve African-American students? We simply do not currently know.

We are not persuaded by the data commenters cited because the studies did not suppress or control for the many confounding variables that influence student outcomes, nor did they rely on carefully constructed matched comparison groups to better isolate the impact of the institution's tax status on student outcomes. These papers also fail to consider the unique structure of proprietary institutions that enable many of them to offer both associate degrees and bachelor's degrees—making them unlike typical public community colleges or typical four-year institutions. In addition, comparisons are further complicated by the number of proprietary institutions that offer online education, which is well-known to have results that are very different than those achieved by ground-based institutions.^[168]

The Department is not suggesting that all proprietary institutions offer high-quality opportunities, or that these institutions should not be held accountable for the outcomes their students achieve. Instead, the Department understands that evaluating college outcomes is an incredibly complicated undertaking, and even with all of the data available to Department researchers, it has been impossible to develop a methodology that allows us to accurately and reliably assess program quality or to make scientifically valid claims of causality between program quality and student outcomes. For that reason, the Department has determined that sanctions limited to a small percentage of institutions and programs—while ignoring other programs whose graduates similarly default on loans or find themselves in a negative amortization repayment situation—are an inappropriate remedy.

Changes: None.

Comments: Commenters also noted that students enrolled in programs that close generally re-enroll in nearby non-profit or public institutions and that shifting aid to better performing institutions will result in positive impacts for students. They also cited evidence that, after enrollment in for profit programs declined in California, local community colleges increased their capacity. They argued that in light of these examples, the 2014 Rule would not reduce college access for students but would rather direct them into programs that are more beneficial in the long term.

Discussion: The California study referenced by the commenter is limited to students who were enrolled at proprietary institutions in that State. Given the large public community college and university system in California, it is not surprising that students closed out of one option in that State found their way to another. However, the Department has recently provided automatic closed school loan discharges for over 15,000 students whose institution closed, and three years later still had not enrolled at another institution. This provides more convincing evidence to us that some students find it harder than others to find a new program. Also, research produced by CSU Sacramento suggests that even among those who find a new home at a lower cost community college, they are likely to be ushered into a general studies program which may result in lower debt, but has no market value unless the student transfers and completes a four-year degree.

In the same way that the Department does not require students seeking a liberal arts education to pursue that degree at the lowest cost institution available, the Department similarly does not require that students interested in occupationally focused education pursue the lowest cost option available.

Moreover, it is entirely unclear whether a student is better off attending a lower cost institution if the only program option available to them is a general studies program, which has little or no market value, rather than a CTE program, which might yield better results.^[169] A 2014 study by CSU Sacramento shows that as enrollments increased in the California Community College system during the Great Recession, there was a decrease in enrollment slots in career and technical programs since more students could be served in lower-cost general studies programs.^[170] Even so, it is not the Department's role under the HEA to evaluate program quality—as accreditors are charged with that responsibility. Nor does the HEA require students to attend the lowest cost institution available or enroll in the program generating the highest earnings. Students enrolled in CTE-focused programs are guaranteed by section 102 of the HEA to have equal access to title IV programs and benefits. The GE regulations deny students interested in CTE-focused programs the same rights as students who enroll in traditional, liberal arts programs.

Changes: None.

Comments: As further justification for the 2014 Rule, commenters stated that there has been a dramatic increase in the number of borrowers who leave school with high debt and low earnings. In one study, a researcher noted that many such programs left students earning less than they did before entering their program. Another study found that the average change in earnings 5 to 6 years post-attendance for over 1.4 million students attending GE programs between 2006 and 2008 was negative for students at for-profit certificate, associates, and bachelor's degree programs. It also found that earnings gains for students in for-profit certificate programs were much lower than for students who attended public institutions even after for controlling for student characteristics. They also stated that at institutions with high D/E rates, students of all income types had poor outcomes, suggesting that the characteristics of the institution are responsible for the poor outcomes. This study also compared students at for-profit certificate programs to demographically similar students who never attended college and found no earnings gains in attendance, suggesting that these students would have been better off choosing not to obtain a postsecondary credential.

Another study cited by the commenters controlled for differences in students' background and characteristics and found that earnings outcomes for students at for-profit programs are typically lower than, or at best equal, to lower-cost programs at public institutions. They cited two studies that found that the poor outcomes of students attending for-profit programs remain even after controlling for family income, race, age, and academic preparation.

Discussion: The Department contends that institutions with high D/E rates exist across all sectors of higher education.^[171] It makes sense that the change in earnings for 2006-2008 program graduates would be negative since this coincides with the Great Recession, which had a more dramatic impact on low-income and minorities than it did on wealthier, white individuals.^[172] In addition, it is impossible for the researcher in the cited studies to have assembled demographically matched comparison groups since the data required to do this is not publicly available.^[173]

The Department notes that several of these studies are based on the unauthorized use of a dataset that was made available by a former Department of Treasury employee to himself and a limited number of outside, like-minded researchers. The Department has been unable to review the data files that were removed from Department of Treasury, since the combined Education-Treasury datafiles were not made available to the Department of Education, to confirm their accuracy or completeness, or to ensure that the data were not manipulated by the person who removed those data from government safekeeping. The Department questions the reliability of research results that are based upon the unauthorized use and the unauthorized release of a dataset since other researchers, including Department of Education researchers, are unable to replicate the calculations to confirm the validity of the methodology or the accuracy of the conclusions.

Regardless, the Department believes that the D/E rates measure is a flawed metric that inflates a borrower's monthly or annual repayment obligation above that which is required by the law and does not accurately distinguish between high-quality and low-quality programs.

Changes: None.

Comments: Commenters criticized the Department's efforts to analyze relevant data related to the NPRM's assertions that, if the D/E rates measure was applied to all degree programs, it would show poor outcomes across all sectors. They argued that if the Department believes this to be the case, it should calculate D/E rates for all programs using available data in NSLDS and with SSA and prove that this is the case. They also criticized the Department's reliance on institutional-level College Scorecard data in lieu of more specific NSLDS data during the negotiated rulemaking process. They further argued that in the absence of such data, the Department has a responsibility to protect students where it has the authority to do so.

Discussion: The Department was unable to obtain SSA earnings data during this rulemaking and continues to be unable to obtain those data. The IRS continues to be willing to provide data for our College Scorecard effort, but § 668.405 of the GE regulations does not allow the use of IRS data to calculate D/E rates. The Department does not currently have program-level earnings data for programs other than GE programs. The Department fulfilled as many data requests as possible, but outdated systems, prohibitions on student unit records, and the inability to get additional earnings data from SSA made it impossible to fulfill all of the requests. However, the Department has access to sufficient data to determine that the D/E rates measure is influenced by a variety of variables other than quality, and that the debt calculation methodology is inconsistent with loan repayment programs available to students. That is sufficient evidence to support our decision to rescind the GE regulations.

Changes: None.

Comments: Commenters disagreed with the statement that for-profit programs would have better D/E rates but for student characteristics outside the institution's control. They argued that it is easy to control for these characteristics and produce adjusted D/E rates, but that the Department had not done so. They believe that such an adjustment would not result in significant numbers of failing programs passing the D/E rates measure. On the point that D/E rates are sensitive to economic conditions, the commenters stated that the Department could use multiple cohorts of rates across institutions to show how changes in the local economy affect D/E rates. They also state that even in large recessions there are not large declines of employed workers and that wages usually do not fall. They argued that because of this, it is likely that only a small number of programs that would have otherwise passed would fail solely due to a recession. They also disagreed with our conclusion in the NPRM that D/E rates are flawed because they are sensitive to tuition and interest rates. These commenters stated this is a desirable outcome because high interest rates and tuition reduce either the government's return on investment or the ability of borrowers to repay.

Discussion: The Department has not been able to develop a methodology to accurately control for or repress confounding variables, such as student demographic characteristics, to isolate the impact of institutional quality on student outcomes, more accurately attribute student outcomes to a single variable, such as institutional quality. In the past, the Department has performed single variant analysis to identify non-traditional student characteristics that increase the risk of non-completion or student loan defaults. However, the Department has not performed multi-variant analysis to develop an algorithm that would allow it to isolate independent variables and examine causal relationships between those variables and student outcomes.

In addition, the negotiators were unable to recommend or reach a consensus on such a methodology. Therefore, the Department is rescinding the 2014 Rule that relies on the flawed D/E rates measure to impose sanctions on institutions and remove them from title IV participation.

Changes: None.

Comments: Commenters argued that while disclosures are beneficial, a disclosure-only regime is unlikely to result in the same benefits that the 2014 Rule provides. As evidence, the commenters cited a study that the College Scorecard had small impacts overall on college application behavior and none in less affluent high schools, households with low parental education, and underserved groups. They also noted that similar studies find little impact of informational disclosures on enrollment behavior, but they provided suggestions on how to improve disclosures. They also stated that removing the disclosure requirements prior to enrollment is a mistake.

Discussion: The Department disagrees with the commenters who state that removing the disclosure requirements prior to enrollment is a mistake and has provided ample explanation above for our disagreement. The Department agrees that disclosures have not been informative to students, especially when comparable information is not provided for all institutions or programs. However, the Department is pursuing a number of options for making College Scorecard data readily available to students, such as through the MyStudentAid mobile app. In addition, the Department believes that an online tool that allows students to compare multiple institutions or programs on a single screen is more user friendly than trying to find disclosures in each institution's or program's web page. Perhaps ease of use will promote increased utilization of important program-level data.

Perhaps one of the most important features of the College Scorecard is that it provides downloadable data files that can be used by researchers, consumer advocacy groups, and technology companies to develop new data tools that are user-friendly and easily accessible to students and parents. Data tools may prove to be more effective in informing student decisions, especially if third parties help students digest and interpret those data, that traditional paper disclosures could.

Changes: None.

Comments: Commenters stated that the Department has not provided enough evidence that the administrative burden is higher than expected or so high as to outweigh the benefits of the 2014 Rule to students. They pointed out that simple adjustments to the D/E rates calculation would reduce burden by allowing the Department to calculate D/E rates using administrative data instead of institutional reporting, although it may not be advisable to do so.

Discussion: The Department disagrees that it has not provided enough evidence that the administrative burden of the GE regulations was higher than expected. In addition, negotiators representing institutions not subject to the GE regulations were adamant that it would be too burdensome for them if we expanded the scope of the 2014 Rule to cover all programs. While simple adjustments to the D/E rates might reduce the administrative burden to institutions, there is no evidence that such adjustments would improve the accuracy and validity of the D/E rates measure.

Changes: None.

Appendix A

	2017 Gainful employment disclosures	Current scorecard	Expanded scorecard
	Gainful employment programs	All undergraduate institutions	All title IV programs
Completion	Percent of students graduating on time for each program	Institution level data that includes the percentage of first-time, full-time undergraduate students who graduated within 150 percent of the published credential length. Students may also view and can select part-time, full-time, transfer, and first-time institution level graduation rates	Same as current Scorecard plus: Expanded Scorecard could include total awards conferred at the program level.
Cost	Program costs (in-state, out-of-state, books and supplies, off-campus room and board, etc.)	Institution level net price for first-time, full-time undergraduate students who received TIV Federal financial student aid. For public schools, this includes only in-state tuition costs	Same as current Scorecard.
Debt	Percent of students who borrow money to pay for the program	Institution level data on the percent of undergraduate students who borrow TIV Federal student loan	Same as current Scorecard, plus: Program level total number of title IV borrowers who complete the program.
	Median debt of TIV Federal financial aid recipients who completed for each program. Median debt includes private, institutional and TIV Federal student loan debt	Institution level data on median TIV Federal student loan debt of undergraduate borrowers who completed. Does not include Parent PLUS	Same as current Scorecard, plus: Program level median TIV Federal student loan debt among completers who borrowed to attend college. Future expanded Scorecard could add median debt among Parent PLUS borrowers who borrowed on behalf of a student in the program and median Grad PLUS debt for graduate and professional programs.
	Estimated monthly loan payment of the median private, institutional and TIV Federal student loan debt for TIV Federal financial aid recipients who completed for each program	Institution level data on the estimated monthly payment of the median TIV Federal student loan debt for TIV Federal financial aid undergraduate borrowers who completed	Same as current Scorecard, plus: Program level estimated monthly payment of the median TIV Federal student loan debt for TIV Federal financial aid borrowers who completed. Future Scorecard could include median monthly payment for Parent PLUS borrowers.
Earnings	Median earnings two- and three-years post-completion of TIV Federal financial aid recipients who completed for each program	Institution level data on median earnings of TIV federal financial aid recipients, 10 years after they began their enrollment	Same as current Scorecard, plus: Program level data on median earnings of TIV Federal financial aid recipients who completed some number of years after completion (number of years not yet determined, but likely at 1, 5, and 10 years after completion).
Job Placement	Job placement rates for students who completed reported to the relevant accreditor and/or state for each program	None	None.
	Fields that employ students who complete for each program	None	Link to relevant occupational information such as O*NET.
Licensure Requirements	Licensure requirements—at least in the state in which the institution is located	None	The consensus achieved during the recent Accreditation and Innovation Negotiated Rulemaking directs all institutions to disclose to students enrolled in programs that lead to occupational licensing whether the program does or does not prepare a student for licensure requirements in the state in which the student is located, or if the institution does not know, and how a student could find this information if he or she relocates. (This will not be on Scorecard.)
Warning	Programs that fail the D/E rates test include a warning that students may not be able to use Federal financial aid for that program in the future	None	None.
Student Demographics (Institution level)	No	Yes	Same.
SAT/ACT Test Scores (Institution level)	No	Yes	Same.
Most popular academic programs	No	Yes	Same.
Institutional type	No	Yes	Same.
Institutional size	No	Yes	Same.
Geographic location	No	Yes	Same.
Institutional control (public, private, proprietary)	No	Yes	Same.
Link to FAFSA	No	Yes	Same.
Link to data about GI Bill benefits	No	Yes	Same.
Net price calculator	No	Yes	Same.
Note: This proposed list provides potential data that the Department plans to include in its expanded College Scorecard or other educational data tools. As a result, this proposed list is provided for informational purposes and is subject to change without notice.

Regulatory Impact Analysis (RIA)

Under Executive Order 12866, the Office of Management and Budget (OMB) must determine whether this regulatory action is “significant” and, therefore, subject to the requirements of the Executive Order and subject to review by OMB. Section 3(f) of Executive Order 12866 defines a “significant regulatory action” as an action likely to result in a rule that may—

(1) Have an annual effect on the economy of $100 million or more, or adversely affect a sector of the economy, productivity, competition, jobs, the environment, public health or safety, or State, local, or Tribal governments or communities in a material way (also referred to as an “economically significant” rule);

(2) Create serious inconsistency or otherwise interfere with an action taken or planned by another agency;

(3) Materially alter the budgetary impacts of entitlement grants, user fees, or loan programs or the rights and obligations of recipients thereof; or

(4) Raise novel legal or policy issues arising out of legal mandates, the President's priorities, or the principles stated in the Executive order.

This final regulatory action will have an annual effect on the economy of more than $100 million because elimination of the ineligibility provision of the GE regulations impacts transfers among borrowers, institutions, and the Federal Government and elimination of paperwork requirements decreases costs. Therefore, this final action is “economically significant” and subject to review by OMB under section 3(f)(1) of Executive Order 12866. Notwithstanding this determination, we have assessed the potential costs and benefits, both quantitative and qualitative, of this final regulatory action and have determined that the benefits justify the costs.

Under Executive Order 13771, for each new regulation that the Department proposes for notice and comment or otherwise promulgates that is a significant regulatory action under Executive Order 12866 and that imposes total costs greater than zero, it must identify two deregulatory actions. These regulations are a deregulatory action under E.O. 13771 and are estimated to yield $160 million in annualized cost savings at a 7 percent discount rate, discounted to a 2016 equivalent, over a perpetual time horizon.

We have also reviewed these regulations under Executive Order 13563, which supplements and explicitly reaffirms the principles, structures, and definitions governing regulatory review established in Executive Order 12866. To the extent permitted by law, Executive Order 13563 requires that an agency—

(1) Propose or adopt regulations only on a reasoned determination that their benefits justify their costs (recognizing that some benefits and costs are difficult to quantify);

(2) Tailor its regulations to impose the least burden on society, consistent with obtaining regulatory objectives and taking into account—among other things and to the extent practicable—the costs of cumulative regulations;

(3) In choosing among alternative regulatory approaches, select those approaches that maximize net benefits (including potential economic, environmental, public health and safety, and other advantages; distributive impacts; and equity);

(4) To the extent feasible, specify performance objectives, rather than the behavior or manner of compliance a regulated entity must adopt; and

(5) Identify and assess available alternatives to direct regulation, including economic incentives—such as user fees or marketable permits—to encourage the desired behavior, or provide information that enables the public to make choices.

Executive Order 13563 also requires an agency “to use the best available techniques to quantify anticipated present and future benefits and costs as accurately as possible.” The Office of Information and Regulatory Affairs of OMB has emphasized that these techniques may include “identifying changing future compliance costs that might result from technological innovation or anticipated behavioral changes.”

We are issuing these final regulations only on a reasoned determination that their benefits justify their costs. Based on the analysis that follows, the Department believes that these final regulations are consistent with the principles in Executive Order 13563.

We also have determined that this regulatory action does not unduly interfere with State, local, and tribal governments in the exercise of their governmental functions.

In accordance with OMB circular A-4, we compare the final regulations to the 2014 Rule. In this regulatory impact analysis, we discuss the need for regulatory action, the potential costs and benefits, net budget impacts, assumptions, limitations, and data sources, as well as regulatory alternatives we considered.

As further detailed in the Net Budget Impacts section, this final regulatory action has an annual effect on the economy at the 7 percent discount rate of approximately $518 million in increased transfers among borrowers, institutions, and the Federal government primarily related to the elimination of the ineligibility provision of the GE regulations. This figure does not take into account that a number of large proprietary chains have closed since the 2014 Rule was promulgated, nor the fact that college enrollments have declined dramatically since 2014—especially at proprietary institutions—meaning that with or without the GE regulations, there are significantly fewer GE programs available to students and students likely to enroll in the programs that remain available than when the 2014 Rule was developed. Therefore, transfers to borrowers and institutions may be lower than anticipated by the Net Budget Impact statement.

In addition, our analysis does not include any reductions in transfers to students and institutions that may result from the market-based accountability system that the expanded College Scorecard will enable. Even in the absence of sanctions or loss of eligibility, programs that yield unfavorable outcomes may be significantly less attractive to students who, prior to expansion of the Scorecard, may have been misled by more generalized claims about the earnings advantage of a college degree.^[174] In general, college enrollments have dropped significantly since 2014, and in particular, enrollments at proprietary institutions have decreased markedly since 2014, due in part to the significant public campaign against those institutions and to the well-publicized closure of Corinthian Colleges. According to the National Student Clearinghouse Research Centers, declines in enrollments at proprietary institutions have been sharper than declines in other sectors:^[175]

Semester	Percent enrollment decline relative to previous year at 4-year, for-profit institutions (%)
Fall 2014	−0.4
Spring 2015	−4.9
Fall 2015	−13.7
Spring 2016	−9.3
Fall 2016	−14.5
Spring 2017	−10.1

As noted in the Net Budget Impacts section of this RIA, this enrollment decline may reflect institutional response to the 2014 Rule or other factors such as the sensitivity of non-traditional student enrollment to economic conditions. Therefore, it is possible that the cost of eliminating the 2014 Rule to taxpayers is lower than the estimate provided in our Regulatory Impact Statement.

We estimate $209 million in benefits due to reduced burden from eliminating paperwork requirements. Additionally, we estimate $593 million at a 7 percent discount rate in annualized increased transfers to Pell Grant recipients and borrowers. This economic estimate was produced by comparing the regulation to the PB2020 budget. The required Accounting Statement is included in the Net Budget Impacts section.

Elsewhere, under Paperwork Reduction Act of 1995, we identify and explain burdens specifically associated with information collection requirements.

1. Need for Regulatory Action

A number of factors compel the Department to take this regulatory action including concerns about the validity of the D/E metric and the integration of factors in the D/E equation, such as repayment terms, that are inconsistent with requirements of the student loan program. In addition, the Department has recognized that by providing consumer information on only a small portion of higher education programs, it fails in providing information that consumers can use to compare all programs available to them, and that enables all students to make informed decisions. The Department believes that in the 2014 GE regulation it underestimated the burden associated with this regulation and ignored the conclusions of a technical review panel that made clear how unreliable, subjective and inaccurate job placement reporting is in the absence of standardized definitions, reliable data sources and a single calculation methodology. The Department attempted to resolve the current challenges associated with job placement rate reporting, but the technical review panel assembled failed to do so. Therefore, it is inappropriate for the Department to require institutions to publicly report job placement rates knowing that direct comparisons between institutions could easily mislead consumers since different institutions are required to calculate these rates in different ways. Also, the Department's 2014 burden estimate did not include an assessment of burden on the government.

Perhaps most importantly, now that the Department is aware that the majority of student borrowers are not repaying their loans using a standard 10 year repayment plan, and many are in income driven repayment plans that lead to negative amortization, it is imperative to implement a transparency framework that provides comparable information to all students and parents to inform the enrollment and borrowing decisions of all consumers. The Department has determined that a more effective and comprehensive solution to the problem of student loan under-repayment is the expansion of the College Scorecard to provide program-level debt and earnings data for all title IV eligible academic programs. Such a transparency framework will support a market-based accountability system that respects consumer choice while enabling more informed decision-making. In addition, by using administrative data rather than requiring institutions to report and review additional data, the College Scorecard will ensure that consumers are provided with information that is consistent, accurate and reliable. It will also enable consumers to more easily compare outcomes among the institutions and programs available to them and reduce costly reporting burden to institutions.

As cited earlier in these final regulations, the Department's determination that only 24 percent of loans in the current $1.2 trillion Direct Loan portfolio are paying down at least a dollar of principal points to the need for a more comprehensive transparency and accountability framework. The Department considered through rulemaking how it might apply GE-like requirements to all institutions by amending the regulations for the Program Participation Agreement; however, negotiators could not agree on which, if any, of the metrics, thresholds, or disclosure requirements included in the GE regulations should be applied to all title IV participating institutions.

Upon further review of studies published subsequent to the 2014 Rule as well as our review of the research paper that originally led to the Department's decision to use an 8 percent D/E rate as the “passing” score led the Department to the conclusion that the D/E methodology was fundamentally flawed, as were the thresholds for ending a school's title IV participation.^[176] In addition, the Department's decision to use its regulatory authority to create a sweeping new student loan repayment program, the REPAYE program, provided the Department with an opportunity to revisit student debt management opportunities and establish new student loan repayment levels and terms. The choices made in establishing the repayment term for REPAYE render the amortization term used for GE calculations of debt-to-earnings inappropriate and obsolete. The GE regulations essentially held GE programs to a student loan repayment standard that no student would be held to by law or regulation. At a minimum, the Department would have needed to adjust the D/E calculation to adopt the amortization terms of REPAYE since any borrower could elect to enter into REPAYE repayment, a program that eliminates an income test for eligibility. However, this adjustment would not solve for the other problems with the validity of the D/E calculation.

The Department's review of the only set of D/E data published to date also reveals the serious weaknesses of the GE methodology since programs with very low earnings passed the D/E rate simply because taxpayers were providing significant financial support to those programs. These data call into question whether taxpayers should continue to subsidize these programs, and also highlight that direct subsidies are every bit a risk to taxpayer investments that do not yield benefits as are student loans that cannot be repaid. While having lower debt is certainly better for students, the Department must weigh the impact of having debt with the impact of achieving higher earnings. From a student perspective, higher earnings may be preferable to lower debt, especially since Congress and the Department have created student loan repayment management programs to help students repay their loans. In some cases, the amount of Federal debt a student could accumulate (due to limits imposed on undergraduate borrowing) would be offset by added earnings (relative to programs in the same field that resulted in lower earnings) just a few years into the student loan amortization period. The GE data made it clear to the Department that there is wide earnings variability among programs within all sectors (non-profit, public, and for-profit), and the Department can no longer assume that this variability accurately reflects differences in program quality. This variability could also be the result of geographic differences in prevailing wages, demographic and socioeconomic differences in student populations, and salary differences from one occupational field to the next. Since the Department is not satisfied that the D/E rates are a reliable or accurate proxy for program quality, the Department is not justified in its use of those data as the determinant for applying sanctions to institutions or eliminating them from title IV participation.

The Department recognizes that some GE programs have inferior outcomes to others, that proprietary institutions like almost all non-public institutions charge higher tuition than public institutions, that earlier comparisons between proprietary institutions and community colleges are misleading since the majority of students enrolled in proprietary institutions are enrolled in four-year programs, and that students who attend proprietary institutions, in general, default at higher rates. However, as pointed out by a recent Brown Center study, proprietary institutions also serve a much higher proportion of high-risk students, low-income and minority students, and students over the age of 25 who by law have significantly higher borrowing limits, than non-profit institutions, which may explain differences in observed outcomes. The Brown Center study also pointed to challenges in comparing data from non-profit institutions and proprietary institutions since non-profit institutions rarely offer both 2-year and 4-year degrees, whereas many proprietary institutions offer both, making comparisons between these institutions and community colleges improper and inaccurate.^[177] A more informative and appropriate comparison between proprietary institutions and non-profit institutions, especially with regard to cost and student debt, would need to include non-profit, private 4-year institutions, since the lack of public subsidies makes their cost structure more similar to many proprietary institutions than two-year or four-year public institutions (except for out-of-State students who receive fewer benefits of taxpayer subsidies and therefore pay a higher cost). Institutional comparisons must also take into account institutional selectivity and student demographics because student borrowing behaviors and earnings outcomes are influenced by many factors other than program quality.

Finally, since the SSA has not renewed the MOU with the Department to provide future earnings data, the Department cannot calculate or report future D/E rates. At a minimum the Department would have had to consider different data sources as part of its rulemaking effort, but at the time of rulemaking, it was not yet apparent that SSA would not provide additional earnings data. Therefore, the Department did not seek comment on the risks or benefits of utilizing Census or IRS data to determine earnings, or the impact of the use of those earnings on the validity of the D/E rates calculation or the comparison between D/E rates based on SSA data and the rates that would be calculated using IRS or Census data. Unable to get the data needed to make those determinations, the Department decided to rescind the 2014 Rule and develop a new tool—the expanded College Scorecard—to implement a transparency framework for GE and non-GE programs that will enable a more robust market-based accountability system to thrive.

2. Summary of Comments and Changes From the NPRM

The Department is making no changes from the NPRM. Comments received by the Department relative to the regulatory impact analysis are summarized and discussed below.

Summary: Commenters stated that the Department failed to discuss regulatory alternatives that it considered. Commenters offered alternatives for the Department to consider as discussed earlier in the document.

Discussion: We thank the commenter for identifying that we inadvertently omitted the Regulatory Alternatives Considered section from the NPRM prior to publication. We have included it in this final rule.

Comments: Commenters stated that the NPRM ignored research showing that students are likely to find and attend another institution if a GE program closes because of sanctions or other adverse actions against a for-profit institution.^[178]

Discussion: The Department agrees that in California, where the study was conducted, there are many choices of two-year colleges that may enable students to find a new program at a public institution if their GE program closes. However, the study does not demonstrate that students were able to find a similar CTE or applied program when moving to the community college. If those students moved from an applied program at a proprietary institution to a general studies or liberal arts program at a two year college (the largest majors at most community colleges nationally according to NCES data), they may not be better off since Holzer and Baum have determined that these programs have no market value to students who do not complete a four-year degree at another institution.^[179] Nonetheless, the Department has always assumed a high level of transfers related to gainful employment disclosures and institutional closures. As noted in the Net Budget Impacts section, the estimates in the PB2020 baseline for the impact on Pell Grants derive from the assumptions about students who would not pursue their education in response to programs' gainful employment results. These assumptions ranged from 5 percent stopping for the first disclosure of a zone result to 20 percent for a second failure.^[180] The Department believes this is consistent with the high degree of transfers reflected in the research cited by the commenters. Additionally, even if the percentage of students who lose access to programs is small, the Department maintains that there are significant consequences to students whose educational plans are disrupted by gainful employment related transfers. As recent experience with institutional closures demonstrates, having to find an alternative program that fits with the other restrictions in students' lives is a stressful process. Not all programs, especially those with specific equipment or other resource requirements, are immediately available for students whose programs would be ineligible for Federal aid. Students may be delayed in pursuing their education or may choose another field, both outcomes that could reduce their earnings potential.

Comments: Several commenters contended that the Department raised questions about the GE regulations without acknowledging the extensive public record on GE topics, ignored evidence compiled through years of analysis and study, and failed to acknowledge its own factual findings on economic benefits and educational value. The commenters stated the Department did not rely upon its own data or research to formulate its policy.

Discussion: The Department considered an abundance of data, including a number of studies that did not exist at the time the Department promulgated the 2014 GE regulation, and NCES data produced by the Department, when trying to develop a methodology for expanding the GE transparency and accountability framework to include all title IV participating programs. While there is an abundance of research comparing proprietary college outcomes with non-profit college outcomes, these studies all have omissions and limitations that make it unclear whether inferior outcomes, where they exist, are the result of program quality or other factors, such as student demographics. These studies also often times compare proprietary colleges with community colleges even though many proprietary institutions offer four-year programs, which makes comparisons with community colleges inappropriate. There is a dearth of research on the low student loan repayment rates across the entire student loan portfolio. The Department recognizes the need to create a transparency and accountability framework that includes all title IV programs and institutions since the problem of student loan over-borrowing and under-repayment impacts all sectors of higher education. However, the Department identified a number of flaws in the D/E rates methodology and thresholds, and excessive burden associated with GE disclosures, making it clear that expanding the components of the GE regulations to all institutions could not be supported by data. The Department believes that in order for consumers to be able to compare their options, all programs they are considering must be subjected to the same analysis and students must have access to comparable data.

The Department did consider data available to it when deciding to rescind the 2014 Rule. In particular, it considered that the data and research presented in conjunction with the 2014 Rule did not support the use of an 8 percent threshold for differentiating between passing and zone or failing programs since the research used to justify the 8 percent threshold specifically pointed out that the 8 percent threshold—a mortgage standard—would not be justified for use in establishing student loan limits.

The 2014 Rule also ignored the role of taxpayer subsidies in allowing programs that generate very low earnings to pass the D/E rates measure. This could give students the inaccurate impression that if a program passes the D/E rates measure, it is high quality and will yield strong outcomes. However, the Department's review of the D/E rates published in 2017 showed that a number of programs that yield earnings below the poverty rate for a family of four passed the test simply because the taxpayer, rather than the student, took on the larger burden of paying for the program. We do not believe that we should mask low earning programs simply by suggesting that if the taxpayer continues to pay for these programs, somehow students benefit.

Given the Department's realization that a sizable percentage of loans in the outstanding student loan portfolio are not shrinking due to student payments, a more comprehensive strategy is required. The GE regulations cannot be expanded to include all programs, and the Department's negotiated rulemaking did not result in consensus on a methodology for applying sanctions or requiring disclosures of all institutions that could be supported by research or justify the potential cost of the added burden or the loss of program options to students. Applying the GE regulations to all institutions could have profound negative impacts on all private institutions, regardless of whether they are non-profit or proprietary, since the absence of direct appropriations naturally pushes the cost burden to students. The Department now believes it is better to use administrative data to provide comparable debt, earnings, default and repayment information across all programs to consumers and taxpayers. Since the Department could not get earnings data for all students in all title IV programs to support this rulemaking effort, the Department is unable to test the impact of applying GE-like metrics to all title IV programs, and would be impetuous to apply GE-like metrics to all title IV programs absent such test data given the sweeping impact that such an action could have.

Comments: Commenters stated that the Department's discussion of costs and benefits in the RIA section of the NPRM did not acknowledge the loss of competitive advantage that institutions face if the GE regulations are rescinded because a program with good D/E rates could market that their rates are good and attract more students versus nearby institutions with poor D/E rates. Meanwhile, other commenters submitted data analyses countering these claims.

Discussion: After reviewing the published GE rates produced in 2017, the Department does not believe that passing D/E rates should be viewed by consumers as the mark of a “good” program since a number of programs that generated lower earnings than failing programs passed the test simply because the taxpayer heavily subsidized the program. The Department is concerned about the false effect that the D/E rates measure could have on a program's or institution's reputation, and that students could be misled to enroll in a program that generates lower earnings without fully understanding the long-term impact of that decision on earnings across a lifetime.

The Department agrees that there may be positive reputational effects lost as a result of rescinding the GE regulations; however, the Department believes that some of these positive reputational effects were inappropriate and harmful since taxpayer generosity rather than program quality is responsible for those outcomes. However, those programs that enjoyed earned positive reputational effects will see them continue as the College Scorecard will provide debt and earnings data for all programs. This may improve the reputational effects for a larger number of deserving programs and institutions.

Comments: Commenters stated that the Department did not consider in the NPRM the full costs of the rescission of the 2014 Rule, including costs that accrue to students with high debt in failing programs and to taxpayers when students default. Commenters further stated that controlling for demographics, location, and major field of study, students in proprietary GE certificate programs earned $2,100 less annually than students in non-profit GE certificate programs.

Commenters also expressed concern that, in rescinding the GE regulations, the Department has failed to consider the cost to borrowers that are not gainfully employed and who may default as a result of unsustainable debt. Commenters cited research and stated that these borrowers would be saddled with capitalized interest and high collection fees, which would require them to pay more per month than borrowers in good standing.^[181]

Discussion: The Department agrees that student loan debt is costly to students and undermines the earnings benefits that many students would otherwise enjoy. However, this problem is not limited to students who enrolled at proprietary institutions. This is a widespread problem that needs a solution that includes all title IV participating programs. The Department agrees that taxpayers need to understand the risks and benefits associated with investing in higher education, but we believe that includes the money that taxpayers invest directly in higher education, including through direct appropriations and State student aid and scholarship programs. Those dollars were ignored in the methodology selected for the 2014 Rule, which was a major shortcoming of the regulation.

The Department has reviewed the research showing that students who complete certificate programs at proprietary institutions earn around $2,100 less per year than those who complete certificate programs at non-profit institutions. However, certificate programs represent only a proportion of higher education programs and it is not clear that those results would persist if the study were expanded to include all degree programs. Also, the research on certificate programs attempted to conduct matched comparison group studies, but it did not accomplish that goal since broad comparisons based on student age and zip codes were used to establish comparison groups, and factors other than that are critical to identifying student matched comparison groups. Even within a single zip code there can be considerable socioeconomic diversity. The study also did not compare outcomes between particular kinds of certificates for particular occupations, meaning that the outcomes could be the result of more students at non-profit institutions pursuing certificates in IT, practical nursing, or the traditional trades, as opposed to more students at proprietary institutions pursuing certificates in allied health professions (other than nursing) or cosmetology. Schools with larger proportions of students in IT and nursing certificate programs will certainly post higher average earnings than those with larger proportions of students in other certificate programs, and yet State nursing boards and accreditors may disallow those institutions to offer programs in higher wage occupations. However, when the study compared earnings outcomes among graduates of certificate programs in cosmetology, it turned out that graduates of proprietary cosmetology programs had higher earnings than graduates of community college cosmetology programs. Therefore, we must interpret the results of the study with caution.

We must also understand that students may have limited options due to location or scheduling convenience, so we need to understand not only whether a student has better earnings potential if she completes a certificate program at a community college versus a proprietary institution, but if she would suffer from lower employability or earnings if in the absence of the proprietary program, the student was unable to complete a career and technical education program at all, or if in the absence of an opportunity to enroll in a certificate program at the community college, she could enroll only in a general studies program. Chances of completing the program could be lower and the market value of doing so could be null. So, we need to also compare the outcomes of general studies programs at community colleges with the outcomes of CTE programs at proprietary institutions since the number of community college GE programs with less than 10 students suggests that only small numbers of students have access to those programs. The largest major at most community colleges is general studies or liberal arts. Therefore, it may not be relevant to compare the outcomes of a proprietary and a non-profit certificate program if the student who enrolls at the non-profit institution is more likely to be ushered into a general studies or liberal arts program than the equivalent certificate program.

The Department does not disagree that the cost of college is a serious concern, but that concern extends well beyond proprietary institutions. The Department is not ignoring that a higher proportion of students at proprietary institutions take on more debt than at community colleges; however, given the size of many community colleges, a lower percent does not translate into fewer students (in whole numbers) taking on debt or defaulting on loans. Total student loan portfolio analysis proves that over-borrowing and under-repayment extends far beyond students who enrolled at proprietary institutions.

The Department is taking a new approach to reducing defaults across the portfolio by implementing better student loan origination and servicing information and support through our Next Generation Financial Services Environment. The Department also believes that by providing comparable information about all programs, enrollment reductions in poor performing programs in all sectors could generate substantial savings.

In the near term, transfers to students and institutions could increase since failing D/E rates will not eliminate the participation of certain programs. However, we have never been able to predict the macro-economic impact of those closures over time. In addition, over the longer-term, the Department believes that the expanded College Scorecard will result in greater savings to students and taxpayers when consumers have earnings and debt data for all title IV programs and can make better choices as a result.

The Department also wishes to point out that macro-economic conditions may have a greater impact on higher education costs and savings to students and taxpayers since college enrollments, in general, have been reduced significantly, especially among students over the age of 24.

Comments: Commenters stated that the Department could use data from the National Student Loan Database (NSLDS) and compute consistently measured D/E rates across all programs and not rely on institutional-level data from the College Scorecard which uses different definitions and is not a reliable cross-sector comparison of programs. Additionally, this NSLDS data could be used to substantiate the Department's claim that whether programs pass or fail the D/E rates measure is unduly affected by the enrollment of disadvantaged students. This was presented for the 2014 Rule.

Discussion: The Department made NSLDS data available during the negotiated rulemaking sessions.^[182] It should be noted that the earnings data obtained from SSA was anonymous and in the aggregate, so there was no way to disaggregate earnings data to test the impact of disadvantaged students on rates as the commenter describes. The Department currently does not have program-level data for non-GE programs, as it requires obtaining data from a different department.

If the commenter is referring to estimates provided in the 2011 GE regulations, the Department wishes to point out that those estimates included title IV and non-title IV programs, since, at the time, IPEDS was the only source of program-level data and it included a larger number of programs.

The Department believes that the commenter misunderstands the use of the expanded College Scorecard, which is not to take data from the Scorecard to calculate D/E rates but is instead to use the Scorecard to provide program-level debt and earnings data for GE and non-GE programs. We agree that the current Scorecard would not inform D/E rates calculations since the current Scorecard includes all students, not just completers, and provides institution-level data only. The expanded Scorecard will report program-level median debt and earnings data for GE and non-GE programs at all credential levels. The Department plans to rely on the IRS, rather than SSA as was the case in the GE regulations, to provide aggregate earnings data and NSLDS will continue to serve as the data source for debt data. Since the GE regulations apply only to GE programs, and the full GE regulations cannot be applied to non-GE programs, the only way to provide cross-sector comparisons based on comparable data is by eliminating the GE regulations and developing a new transparency tool that can be applied to all title IV programs. The College Scorecard will serve as that tool.

The Department is currently considering ways to develop risk-adjusted outcomes metrics that leverage the power of regression techniques to control for differences in student-level risk factors such as age, socioeconomic status, or high school preparation when comparing student outcomes. In the meantime, we believe that by providing institution—level selectivity ratings and student demographics, we can begin to put outcomes in the context of differences in student demographics and institutional selectivity.

Comments: A commenter stated that during the first year of the D/E calculation GE programs declined from 39,000 to 27,000 programs indicating that failing programs dropped out.

Discussion: We were unable to replicate the findings the commenter referenced, and the commenter provided no documentation or data to support this assertion. In the 2014 Rule, the Department did report a total of 37,589 programs for which institutions reported enrollment in FY2010, of which 5,539 met the 30 completer threshold to be included in the 2012 D/E rates calculations.^[183] Several factors contribute to the decline in programs for 2008-09 from the first GE reporting reflected in the 2012 informational rates and the data presented for this regulation. As institutions became more familiar with the reporting requirements, they may have changed 6-digit OPEIDS, CIP codes or updated students' enrollment status, all of which could consolidate the number of programs reported. Some of the decline likely was in response to anticipated non-passing gainful employment results, but mergers and changes in program offerings occur on a regular basis for a variety of business reasons, especially when considering the small size of many of the programs captured in the GE reporting. Therefore, we do not agree with the commenter that the reduction in the number of programs is due exclusively to institutions' decisions to discontinue programs that would have failed. However, even in the absence of the GE regulation, when students are able to compare earnings and debt outcomes among all of their options, low-performing programs may suffer from such low enrollments that schools will discontinue them even in the absence of Department sanctions.

During negotiated rulemaking the Department provided^[184] Table 3.1 Program and Enrollment Counts during the second negotiated rulemaking session which included GE programs counts from the 2008-2009 thru 2015-2016 year, copied below in Table 3.

Table 3—Number of GE Programs and Enrollees by Award Year

Award year	Programs	Enrollment
2008-2009	27,611	2,787,260
2009-2010	30,674	3,613,730
2010-2011	32,908	3,892,590
2011-2012	34,252	3,767,430
2012-2013	35,075	3,515,210
2013-2014	35,905	3,326,340
2014-2015	35,399	3,077,970
2015-2016	32,970	2,529,190
Enrollment values rounded to the nearest 10.

The number of GE programs and enrollment in them changed over time, but do not show a decline from 39,000 to 27,000 programs. During the time period shown above, program count peaked in 2013-2014 and enrollment peaked in 2010-2011.

Comments: Commenters stated that during the one year that the 2014 Rule was implemented, results of the rule showed that 98 percent of over 800 programs that failed were offered by for-profit institutions. Commenters stated that risk-based compliance efforts appropriately target proprietary institutions. Commenters asserted that the Department relied on the premise that there are justifiable reasons to provide title IV funds to students enrolled in low-quality programs. Commenters claim that data show that the GE regulations affect institutional behavior with respect to zone and fail programs. Commenters also submit data analyses supporting expanding the application of the D/E rates measure to all programs at all institutions or rescinding it entirely.

Discussion: The table below is based on data the Department distributed ^[185] during the second session of negotiated rulemaking, February 2018 `Gainful Employment Data Analysis' section 6, table 3.2.

Table 4—Number and Percent of Programs That Failed GE

GE programs—all programs	Number		Percent and confidence interval
Sector	Fail	Total	Percent fail (%)	LCL (%)	UCL (%)
Public	1	2,493	0.04	−0.04	0.12
Private	24	476	5.04	3.08	7.01
Proprietary	878	5,681	15.46	14.52	16.40
Overall	903	8,650	10.44	9.79	11.08

GE programs—certificate only		Number		Percent and confidence interval
Sector	Certificate level	Fail	Total	Percent fail (%)	LCL (%)	UCL (%)
Public	Undergraduate	1	2,428	0.04	−0.04	0.12
Public	Post baccalaureate	0	17	0.00	0.00	0.00
Public	Graduate	0	48	0.00	0.00	0.00
Private	Undergraduate	21	405	5.19	3.03	7.34
Private	Post baccalaureate	0	27	0.00	0.00	0.00
Private	Graduate	3	44	6.82	−0.63	14.27
Proprietary	Undergraduate	196	3,260	6.01	5.20	6.83
Proprietary	Post baccalaureate	0	5	0.00	0.00	0.00
Proprietary	Graduate	2	23	8.70	−2.82	20.21
Overall Certificate Programs		223	6,257	3.56	3.10	4.02

We used the published data to produce the tables above, which compare GE programs by sector—public, private, and proprietary—and level-undergraduate, post baccalaureate, and graduate. Overall totals from the table show that there are 8,650 (Proprietary 65.7 percent, Private 28.8 percent Public 5.5 percent) total GE programs of which 903 or 10.44 percent failed the D/E rates measure. When significance tests are run at the sector level on this data at the 95 percent confidence interval producing lower (LCL) and upper (UCL) confidence limits, the three sectors appear to be significantly different because their confidence intervals do not overlap. However, these data contain non-comparable data in the reported totals because only degree programs are only counted as GE programs in the proprietary sector. When the proprietary data are subset to certificate-only, 198 programs of 3288 failed, resulting in 6.02 percent failing with a confidence interval ranging from 5.21 percent to 6.84 percent; this interval overlaps with that of private, non-profit institutions. Because there are no comparable data at the degree levels, a valid comparison is not possible with Department data.

The second part of the table subsets the data to certificate programs and further breaks down certificates by level. There were 6,257 GE certificate programs of which 223 or 3.56 percent failed the D/E rates measure. When degree programs are removed from proprietary programs (computed using addition), the resulting percentage of proprietary certificate programs failing is 6.02 percent (198/3288) with a confidence interval of 5.21 to 6.84 percent. This overlaps with the private, non-profit certificate confidence interval of 3.08 to 7.01 percent. Therefore, there is no statistical difference between private and proprietary certificate program GE failure rates. Further, we found no significant differences between the percentages of failing certificate programs at non-profit private and proprietary private institutions, regardless of level under examination. Public GE certificate programs had significantly lower failure rates than both private and proprietary GE certificate programs. However, as was pointed out earlier in this document, GE programs offered by taxpayer subsidized public institutions may have passed, despite very low earnings by program graduates, simply because taxpayers take on the largest portion of cost burden. While we agree that taxpayer support benefits students, the masking effect of direct appropriations reduces the accountability of publicly subsidized programs when they are producing sub-optimal earnings outcomes, which is disadvantageous to both students and taxpayers. In other words, a program that passes the D/E rates measure because of taxpayer funding may not impose overwhelming debt burden on students; however, those programs may reduce students' full earning potential and may be directing scarce taxpayer resources to low-performing programs rather than high performing programs.

Summary: Commenters stated that this regulatory action will cost taxpayers $5.3 billion over 10 years.

Discussion: Comments related to the cost of the regulations are addressed in the Net Budget Impacts section of this document.

Comments: Commenters requested information relative to the budget estimate. Commenters requested the Department clarify the assumptions it used to produce its estimate and incorporate the effect of changed institutional behavior. Commenters also requested that the effects of rescission on default rate and resulting costs to borrowers, society, and the economy be reflected in the budget estimate. Commenters requested modifications to the budget estimate to adjust for IDR, loan forgiveness, and default.

Discussion: Comments related to the cost of the regulation are addressed in the Net Budget Impacts section of this document.

Comments: Commenters stated that the Department did not justify the rescission of the discretionary D/E rate. Other commenters provided evidence to support its rescission.

Discussion: The Department clearly stated in the NPRM that neither it nor non-Federal negotiators could identify a D/E metric that was sufficiently valid and accurate to serve as a high-stakes quality test or to become a new, non-congressionally mandated, eligibility criteria for title IV participation. Regardless of whether gross income or discretionary income forms the basis of the D/E rates calculation, the methodology is inaccurate and fails to control for the many other factors other than program quality that influence debt and earnings.

Comments: Commenters stated the Department failed to comply with E.O. 12291 because it did not estimate either the number of or dollar impact to students or institutions nor did it match costs to benefits. A commenter asserted that the RIA failed to show why rescission is beneficial.

Discussion: Executive Order 12291 was revoked by Executive Order 12866 on September 30, 1993. Further, the monetized estimates in the Regulatory Impact Analysis are based on the budget estimates, which can be found in the Net Budget Impacts section. Other impacts, including expected burdens and benefits are discussed in the Costs, Benefits, and Transfers and Paperwork Reduction Act of 1995 sections. The Department believes it is in compliance with Executive Order 12866.

Comments: Commenters asserted that the regulatory text does not support the transparency argument from E.O. 13777 because the regulatory text does not include disclosures.

Discussion: The Department agrees with the commenter and has revised its Need for Regulatory Action.

3. Analysis of Costs and Benefits

These regulations affect prospective and current students; institutions with GE programs participating in the title IV, HEA programs; and the Federal government. The Department expects institutions and the Federal government to benefit as this action eliminates reporting, administrative costs, and sanctions. As detailed earlier, pursuant to this regulatory action, the Department removes the GE regulations and adopts no new ones.

3.1 Students

Based on 2015-16 Department data from the National Student Loan Data System (NSLDS), about 520,000 students would be affected annually by the rescission of the GE regulation. The Department estimates this rescission will result in both costs and benefits to students, including the costs and benefits associated with continued enrollment in zone and failing GE programs and the benefit of eliminating paperwork burden.

Eliminating sanctions against institutions based on the D/E rates measure will impact students. Under the GE regulations, if a GE program became ineligible to participate in the title IV, HEA programs, its students would not be able to receive title IV aid to enroll in that program. Because D/E rates have been calculated under the GE regulations for only one year, no programs have lost title IV, HEA eligibility. However, 2,050 programs were identified as failing programs or programs in the zone based on their 2015 GE rates and would have been at risk of losing eligibility under the GE regulation. NSLDS data from 2015-16 shows 329,250 students were enrolled in zone GE programs and 189,920 students were enrolled in failing programs (about 520,000 total). These students will not lose access to title IV Federal financial aid at their initially chosen program. As further explained in the Net Budget Impacts section, the Department estimates that there will be an annual increase in Direct Loan and Pell grant transfers from the Federal government to students of $593 million at the 7 percent discount rate when compared to the GE regulations under PB2020.

There are further costs and benefits to students who continue enrollment in a program that would have been in the zone or failing under the GE regulations, which the Department was unable to monetize because the actual outcome for these students is unknown. This includes the impact that students will not lose access to title IV aid for those programs, which is a benefit of continued financial aid but could also be a cost if the investment is not as fruitful as it might be at a similar nearby program. What the Department is unable to determine for the purpose of these costs estimates is what number of students displaced from a GE program that loses title IV eligibility will be able to find a similar program at another institution or will enroll in a non-applied program, a different applied program of study, or a general studies program that yields even poorer outcomes. However, given that the large majority of GE programs have less than 10 students suggests that a significant number of students who lose access to a GE program will end up in a community college general studies program, where we do not have D/E outcomes data to inform our analysis. Other impacts relate to whether students would have transferred, found alternate funding, or discontinued postsecondary education as a result of their program losing title IV eligibility under the GE regulation. As a result of the rescission, students would not face this stressful choice, which could be seen as a benefit of continued postsecondary education and not having to transfer institutions, but also a potential cost of completing a program that may be judged less favorably than a similar program at a nearby institution.

The Department will also discontinue GE information collections, which is detailed further in the Paperwork Reduction Act of 1995 section of this preamble. Two of these information collections impact students—OMB control number 1845-0123 and OMB control number 1845-0107. By removing these collections, the regulations will reduce burden on students by 2,167,129 hours annually. The burden associated with these information collections is attributed to students being required to read warning notices and certify that they received them. Therefore, using an individual hourly rate of $16.30,^[186] the benefit due to reduced burden for students is $35,324,203 annually (2,167,129 hours per year * $16.30 per hour).

With the elimination of the disclosures and the ineligibility sanction that would have removed students' program choices, students, their parents, and other interested members of the public will have to seek out the information that interests them about programs they are considering. Affordability and earnings associated with institutions and programs continues to be an area of interest. The College Scorecard is one source of comparative data, but others are available, so students will have the opportunity to incorporate the information into their decisions and rely on their own judgement in choosing a program based on a variety of factors.

To the extent non-passing programs remain accessible with the rescission of the 2014 Rule, some students may choose sub-optimal programs. Whatever the reason, these programs have demonstrated a lower return on the student's investment, either through higher upfront costs, reduced earnings, or both. As some commenters have noted, this could lead to greater difficulty in repaying loans, increasing the use of income-driven repayment plans or risking defaults and the associated stress, increased costs, and reduced spending and investment on other priorities. These regulations emphasize choice and access for all students, and we encourage students to make informed enrollment decisions regardless of which institutions or programs they are considering, and regardless of whether the institution is proprietary, non-profit, or public.

3.2 Institutions

Based on 2015 GE program rates from the National Student Loan Data System (NSLDS), about 2,600 institutions will be affected annually by the removal of the GE regulation. These institutions will have a reduced paperwork burden and no longer be subject to potential GE sanctions that caused loss of title IV eligibility. The table below shows the distribution of institutions administering GE programs by sector.

Table [1]—Institutions With 2015 GE Programs ¹⁸⁷

Type	Institutions		Programs
Public	865	33%	2,493	29%
Private	206	8%	476	5%
Proprietary	1,546	59%	5,681	66%
Total	2,617		8,650

All 2,617 institutions with GE programs will benefit from the elimination of GE reporting requirements. As discussed further in the Paperwork Reduction Act of 1995 section of this preamble, reduction in burden associated with removing the GE regulatory information collections for institutions is 4,758,499 hours. Institutions would benefit from these proposed changes, which would reduce their costs by $173,923,138 annually using the hourly rate of $36.55.^[188]

There are 778 institutions administering 2,050 zone or failing GE programs that will benefit because they no longer will be subject to sanctions that would result in the loss of title IV eligibility. As further explained in the Net Budget Impacts section, the Department estimates this change will increase Pell grant and Direct Loan transfers from students to institutions by $518 million annually under the 7 percent discount rate when compared to PB2019. Although the Department was unable to monetize this impact, institutions further benefit from the elimination of the need to appeal failing or zone D/E rates. The table below shows the distribution of institutions with zone and failing programs by institutional type, which represents 24 percent of the 8,650 2015 GE programs and 30 percent of the 2,617 institutions with GE programs.

Table [2]—Institutions With 2015 GE Zone or Failing Programs ¹⁸⁹

Type	Institutions	Zone programs	Failing programs	Zone or failing programs
Public	9	9		9
Private	34	68	21	89
Proprietary	735	1,165	787	1,952
Total	778	1,242	808	2,050

Table [3] shows the most frequent types of programs with failing or zone D/E rates. Cosmetology undergraduate certificate programs had the most programs in the zone or failing categories, which represented 40 percent of all of these programs. The proportion of programs in zone or fail shown in the table below ranged from 17 to 82 percent. These programs and their institutions would be most significantly affected by the proposed removal of GE sanctions as they would continue to be eligible to participate in title IV, HEA programs.

Table [3]—Zone or Failing 2015 GE Programs by Frequency of Program Types ¹⁹⁰

CIP	Credential level	Zone	Fail	Zone or fail	All programs
Cosmetology/Cosmetologist, General	Undergraduate Certificate	270	91	361	895
Medical/Clinical Assistant	Associates Degree	35	56	91	119
Medical/Clinical Assistant	Undergraduate Certificate	78	12	90	424
Massage Therapy/Therapeutic Massage	Undergraduate Certificate	43	4	47	270
Business Administration and Management, General	Associates Degree	24	22	46	74
Legal Assistant/Paralegal	Associates Degree	20	25	45	58
Barbering/Barber	Undergraduate Certificate	22	16	38	96
Graphic Design	Associates Degree	16	17	33	45
Criminal Justice/Safety Studies	Associates Degree	20	11	31	41
Massage Therapy/Therapeutic Massage	Associates Degree	8	19	27	33
All other programs		706	535	1,241	6,595
Total		1,242	808	2,050	8,650

While programs with non-passing results will benefit from avoiding ineligibility and potentially reputational contagion to other programs at the institution that performed better, programs with passing results could lose the benefit of their comparatively strong performance, although the Department believes that comparatively strong performance will be revealed through program-level College Scorecard outcomes as well. Consistently strong earnings or low costs would likely be an attractive draw for students in a given region or field of study, as long as the low-cost program is available to students and offers the same scheduling flexibility, convenience, and student support services as the higher-cost program offered. While there will not be an established standard to be categorized as passing, the Department does believe that programs with strong outcomes could still gain from their strong performance. Presumably, if a large percentage of programs at their institutions do well on gainful employment measures, the earnings, debt levels, and other items reported in the College Scorecard will be strong compared to their peers with similar offerings. As information and analytical tools become more accessible, the Department believes the lost potential reputational benefit from gainful employment can be replaced.

3.3 Federal Government

Under the proposed regulations, the Federal government will benefit from reduced administrative burden associated with removing provisions in the GE regulations and from discontinuing information collections. As discussed in the Net Budget Impacts section, the Federal government will incur annual costs to fund more Pell Grants and title IV loans, including the costs of income-driven repayment plans and defaults.

Reduced administrative burden due to the proposed regulatory changes will result from elimination of sending completer lists to institutions, adjudicating completer list corrections, adjudicating challenges, and adjudicating alternate earnings appeals. Under the GE regulations, the Department estimated about 500 Notices of Intent to Appeal, and each one took Department staff about 10 hours to evaluate. Using the hourly rate of a GS-13 Step 1 in the Washington, DC area of $46.46,^[191] the estimated benefit due to reduced costs from eliminating earnings appeals is $232,300 annually (500 earnings appeals * 10 hours per appeal * $46.46 per hour). Similarly, the Department sent out 31,018 program completer lists to institutions annually, which took about 40 hours total to complete. Using the hourly rate of a GS-14 Step 1 in the Washington, DC area of $54.91,^[192] the estimated benefit due to reduced costs from eliminating sending completer lists is $2,196 annually (40 * 54.91). Likewise, the Department processed 90,318 completer list corrections and adjudicated 2,894 challenges. The Department estimates it took Department staff 1,420 hours total to make completer list corrections. Similarly, the Department estimates it took $1,500,000 in contractor support and 1,400 hours of Federal staff time total to adjudicate the challenges. Using the hourly rate of a GS-13 step 1 in the Washington, DC area of $46.46, the estimated benefit due to reduced costs from eliminating completer lists, corrections, and challenges is $1,631,017 ($1,500,000 contractor support + (1,420 + 1,400) staff hours * $46.46 per hour).

Additionally, the Department will rescind information collections with OMB control numbers 1845-0121, 1845-0122, and 1845-0123. This will result in a Federal government benefit due to reduced contractor costs of $23,099,946 annually. Therefore, the Department estimates an annual benefit due to reduced administrative costs under the regulations of $24,965,459 ($232,300 + $2,196 + $1,631,017 + $23,099,946).

Finally, the Department will also incur increased budget costs due to increased transfers of Pell Grants and title IV loans, as discussed further in the Net Budget Impacts section. The estimated annualized costs of increased Pell Grants and title IV loans from eliminating the GE regulations is approximately $518 to $527 million at 7 percent and 3 percent discount rates, respectively.

4. Net Budget Impacts

The Department received a number of comments related to its estimated net budget impact for the regulations proposed in the NPRM that rescinded the current GE regulation. In particular, some commenters presented analysis of the potential effect on defaults and loan forgiveness as a cost of the regulation not accounted for in the Department's analysis. One such commenter's analysis modeled IDR usage at gainful employment programs using the debt and earnings data published for gainful employment programs and found that many borrowers in non-passing programs would qualify for IDR plans and their payments under REPAYE would be $1.5 billion less than under a 10-year standard plan on a net present value basis.^[193] The Department appreciates the analysis presented and acknowledges that there are potential interactions between gainful employment, student program choice, repayment outcomes, and other factors that could affect the estimates presented. Other commenters noted the effect of the current gainful employment regulations on institutional behavior, noting that institutions closed or revised programs anticipated not to pass the gainful employment measures and the loss of this deterrent should be factored into the Department's estimates.^[194] However, the Department never attributed any savings to default reductions or decreased loan forgiveness in relation to the 2014 GE Regulations. The increased volume in the 2-year proprietary risk group estimated from rescinding the gainful employment regulations, as described in the NPRM and reiterated below, is subject to the relatively high default and income-driven repayment plan assumptions. Therefore, we do not anticipate a significant change in those areas from these final regulations.

As indicated in the NPRM published August 14, 2018, The Department proposes to remove the GE regulations, which include provisions for GE programs' loss of title IV, HEA program eligibility based on performance on the D/E rates measure. In estimating the impact of the GE regulations at the time they were developed and in subsequent budget estimates, the Department attributed some savings in the Pell Grant program based on the assumption that some students, including prospective students, would drop out of postsecondary education as their programs became ineligible or imminently approached ineligibility.

This assumption has remained in the baseline estimates for the Pell Grant program, with an average of approximately 123,000 dropouts annually over the 10-year budget window from FY2019 to FY2028. By applying the estimated average Pell Grant per recipient for proprietary institutions ($4,468) for 2019 to 2028 in the PB2020 Pell Baseline, the estimated net budget impact of the GE regulations in the PB2020 Pell baseline is approximately $−5.2 billion. As was indicated in the Primary Student Response assumption in the 2014 Rule,^[195] much of this impact was expected to come from the warning that a program could lose eligibility in the next year. If we attribute all of the dropout effect to loss of eligibility, it would generate a maximum estimated Federal net budget impact of the final regulations of $5.2 billion in costs by removing the GE regulations from the PB2020 Pell Grant baseline.

The Department also estimated an impact of warnings and ineligibility on Federal student loans in the analysis for the 2014 Rule, that, due to negative subsidy rates for PLUS and Unsubsidized loans at the time, offset the savings in Pell Grants by $695 million.^[196] The effect of the GE regulations is not specifically identified in the PB2020 baseline, but it is one of several factors reflected in declining loan volume estimates. The development of GE regulations since the first negotiated rulemaking on the subject was announced on May 26, 2009, has coincided with demographic and economic trends that significantly influence postsecondary enrollment, especially in career-oriented programs classified as GE programs under the GE regulation. Enrollment and aid awarded have both declined substantially from peak amounts in 2010 and 2011.

As classified under the GE regulations, GE programs serve non-traditional students who may be more responsive to immediate economic trends in making postsecondary education decisions. Non-consolidated title IV loans volume disbursed at proprietary institutions declined 48 percent between AY2010-11 and AY2016-17, compared to a 6 percent decline at public institutions, and a 1 percent increase at private institutions. The average annual loan volume change from AY2010-11 to AY2016-17 was −10 percent at proprietary institutions, −1 percent at public institutions, and 0.2 percent at private institutions. If we attribute all of the excess decline at proprietary institutions to the potential loss of eligibility under the GE regulations and increase estimated volume in the 2-year proprietary risk group that has the highest subsidy rate in the PB2020 baseline by the difference in the average annual change (12 percent for subsidized and unsubsidized loans and 9 percent for PLUS), then the estimated net budget impact of the removal of the ineligibility sanction in the final regulations on the Direct Loan program is a cost of $1.04 billion.

Therefore, the total estimated net budget impact from the final regulations is $6.2 billion cost in increased transfers from the Federal government to Pell Grant recipients and student loan borrowers and subsequently to institutions, primarily from the elimination of the ineligibility provision of the GE regulation. As in all previous estimates related to Gainful Employment regulations, the estimated effects are associated with borrowers who could no longer enroll in a GE program that loses title IV eligibility and would not enroll in a different program that passes the D/E rates measure, but would instead opt out of a postsecondary education experience. Some commenters submitted research analyzing how CDR-related sanctions in the 1990s resulted in small declines in the aggregate enrollment.^[197] Other commenters have suggested that 10 percent of students would not enroll in a different program. The transfer rates estimated for the 2014 Rule which ranged from 5 percent for a first zone result to 20 percent for potential ineligibility were in line with the high transfer rate suggested by the commenters. Given the potential for several programs to become ineligible in the same timeframe and for the loss of eligibility to affect grant and loan programs, the Department believes the transfer and dropout rates it used in developing the GE estimates that are now being rescinded are reasonable. The long-term impact to the student and the government of the decision to pursue no postsecondary education could be significant but cannot be estimated for the purpose of this analysis, which does not include long-term macro-economic impacts, such as long-term tax revenue impacts of a workforce with less education.

This is a maximum net budget impact and could be offset by student and institutional behavior in response to disclosures in the College Scorecard and other resources. In the 2014 GE rule, the Department stated: “The costs of program changes in response to the regulations are difficult to quantify generally as they would vary significantly by institution and ultimately depend on institutional behavior.” ^[198] In these final regulations, we follow pervious Department practice where we do not attribute a significant budget impact to disclosure requirements absent substantial evidence that such information will change borrower or institutional behavior.

Other factors that could affect these estimates include recent institutional closures, particularly of proprietary institutions whose programs would have been subject to the gainful employment measures. Depending upon where the students who would have attended those programs in the future decide to go instead, the amount of Pell Grants or loans they receive may vary and their earnings and repayment outcomes could also change. The budget impact associated with the rescission of the gainful employment rule would also be affected if significant closures continue and those students pursue programs not subject to the 2014 Rule or leave postsecondary education altogether.

5. Accounting Statement

As required by OMB Circular A-4 we have prepared an accounting statement showing the classification of the expenditures associated with this final rule (see Table 4). This table provides our best estimate of the changes in annual monetized transfers as a result of the final rule. The estimated reduced reporting and disclosure burden equals the $−209 million annual paperwork burden calculated in the Paperwork Reduction Act of 1995 section (and also appearing on page 65004 of the regulatory impact analysis accompanying the 2014 Rule). The annualization of the paperwork burden differs from the 2014 Rule as the annualization of the paperwork burden for that rule assumed the same pattern as the 2011 rule that featured multiple years of data being reported in the first year with a significant decline in burden in subsequent years.

Table [4]—Accounting Statement: Classification of Estimated Expenditures

[In millions]

Category	Benefits
Discount rate	7%	3%
Reduced reporting and disclosure burden for institutions with GE programs under the GE regulation	$209.3	$209.3

Category	Costs
Discount rate	7%	3%
Reduced market information about gainful employment programs; offset by development of College Scorecard for wider range of programs	Unquantified.

Category	Transfers
Discount rate	7%	3%
Increased transfers to Pell Grant recipients and student loan borrowers from elimination of ineligibility provision of GE regulation	$593	$608

6. Regulatory Alternatives Considered

In response to comments received and the Department's further internal consideration of these final regulations, the Department reviewed and considered various changes to the final regulations detailed in this document. The changes made in response to comments are described in the Analysis of Comments and Changes section of this preamble. We summarize below the major proposals that we considered but which we ultimately declined to implement in these regulations.

In particular, the Department extensively reviewed outcome metrics, institutional accountability, sanctions, data disclosure, data appeals, and warning provisions in deciding to rescind the GE regulations. In developing these final regulations, the Department considered the budgetary impact, administrative burden, and effectiveness of the options it considered.

Table [5]—Summary of Alternatives

Topic	Baseline	Alternatives	NPRM proposal	Final regs.
Universe of Coverage	GE Programs	None; GE Programs; all programs at all institutions (IHEs); all programs at all IHEs except graduate programs; and all programs at all IHEs except professional dental, and veterinary	None	None.
Disclosures: Calculations and posting location	IHEs calculate and post on their website using a Department-provided template	None; IHEs calculate and post on their website using a Department-provided template; IHEs and Department calculate and IHEs post on program homepage in any format; Department calculates and posts all disclosures on program-level College Scorecard and IHEs post link to College Scorecard on program homepage; and Department calculates and posts all disclosures on program-level College Scorecard and IHEs post mean debt, mean earnings, and a link to College Scorecard on program homepage	None	None.
Occupational licensure requirements	List States where licensure is required and indicate whether program meets requirements	None; List States where licensure is required and indicate whether program meets requirements; For State in which institution is located, indicate whether the program meets any certification requirements and list other States for which the institution is aware the program meets certification requirements; and List States where program meets requirements	None	None.
Cohort lists and challenges	Lists by Department, challenges available to IHEs	None; Lists by Department, challenges available to IHEs; Lists by Department, no challenges;	None	None.
Earnings appeals	Available to IHE and adjudicated by Department	None; and Available to IHE and adjudicated by Department	None	None.
Sanctions	Automatic loss of title IV eligibility in certain circumstances	None; and Automatic loss of title IV eligibility in certain circumstances	None	None.
Warnings	Required in certain circumstances	None; and Required in certain circumstances	None	None.

6.1 Baseline

We use the 2014 Rule as the baseline. Under the GE regulations, institutions must certify that each of their GE programs meets State and Federal licensure, certification, and accreditation requirements. Also, to maintain title IV, HEA program eligibility, GE programs must meet minimum standards under the D/E rates measure. Programs must issue warnings to their students if they could lose their title IV, HEA program eligibility based on their next year's D/E rates.

Institutions are required to disclose a program's student outcomes and information such as costs, earnings, debt, and completion rates, and whether the program leads to licensure on the program's home page. Institutions compute these statistics and enter them into the Department's GE Disclosure Template. Then, the institution posts the template on its website.

6.2 Summary of the Final Regulations

The Department's final regulations rescind the 2014 Rule.

6.3 Discussion of Alternatives

During negotiated rulemaking, the Department considered expanding the universe of institutions and programs to which the regulations would apply. This would have expanded the burden on institutions compared to the baseline. Various alternatives considered would have affected slightly different groups of institutions by excluding special populations. The final regulations rescind the GE regulations and therefore remove the institutional burden associated with it. Under various universe options, cohort lists would have been created; further, the Department did consider permitting and not permitting challenges to those lists. Ultimately, the lists are eliminated and also the need to challenge them because no cohorts are created under the rescission.

The Department considered multiple options regarding which metrics to disclose, which entity bears the burden of computing them, and how to disseminate them to students and the public. One option has the Department computing all metrics administratively and publishing them on its College Scorecard and requiring institutions to post a link to the Scorecard on their program pages. Another option shared burden for metric computation by requiring institutions to compute some and the Department to compute the rest administratively; we considered either having institutions develop their own format for posting the data on their websites or providing them a general format to follow, including links to the College Scorecard. Metrics of specific concern included earnings and the appeals thereof as well as occupational licensure requirements. The Department considered eliminating the appeals process to reduce burden on institutions and the Department and allow for smaller cohort sizes, keeping the appeals process to allow institutions to contest earnings reported to the IRS but thereby causing increased burden to the institution and also to the Department, and replacing the appeals process with secondary metrics like repayment rate thereby increasing burden on the Department to compute extra metrics but to a much smaller amount than adjudicating alternate earnings appeals. Ultimately, the Department chose to rescind these regulations; without regulating it, the Department plans to expand its College Scorecard in order to report data at the program level in the future. In accordance with Executive Order 13864, this would accomplish the presidential mandates both to increase transparency and also to deregulation.

Finally, the Department considered alternative sanctions scenarios. One option was to make no change relative to the baseline, while another made the sanction discretionary. Further, the Department considered options for when and how to deliver warnings to students when a program is zone or failing. Some options discussed included delivering warnings only by email or only posting on the institution's website. Other options included only providing the warning upon matriculation whereas others would have required a reminder annually. Under rescission, the sanctions and associated warnings are eliminated.

7. Regulatory Flexibility Act (RFA) Certification

The U.S. Small Business Administration (SBA) Size Standards define proprietary institutions as small businesses if they are independently owned and operated, are not dominant in their field of operation, and have total annual revenue below $7,000,000. Non-profit institutions are defined as small entities if they are independently owned and operated and not dominant in their field of operation. Public institutions are defined as small organizations if they are operated by a government overseeing a population below 50,000.

The Department lacks data to identify which public and private, non-profit institutions qualify as small based on the SBA definition. Given the data limitations and to establish a common definition across all sectors of postsecondary institutions, the Department uses its proposed data driven definitions for “small institutions” (Full-time enrollment of 500 or less for a two-year institution or less than two-year institution and 1,000 or less for four-year institutions) in each sector (Docket ID ED-2018-OPE-0027) to certify the RFA impacts of this final rule. The basis of this size classification was described in the NPRM published in the Federal Register July 31, 2018 for the proposed borrower defense rule (83 FR 37242, 37302). The Department has discussed the proposed standard with the Chief Counsel for Advocacy of the Small Business Administration, and while no change has been finalized, the Department continues to believe this approach better reflects a common basis for determining size categories that is linked to the provision of educational services.

Table 5—Small Entities Under Enrollment Based Definition

Level	Type	Small	Total	Percent
2-year	Public	342	1,240	28
2-year	Private	219	259	85
2-year	Proprietary	2,147	2,463	87
4-year	Public	64	759	8
4-year	Private	799	1,672	48
4-year	Proprietary	425	558	76
Total		3,996	6,951	57

When an agency promulgates a final rule, the RFA requires the agency to “prepare a final regulatory flexibility analysis”.” (5 U.S.C. 604(a)). Section 605 of the RFA allows an agency to certify a rule, in lieu of preparing an analysis, if the final rule is not expected to have a significant economic impact on a substantial number of small entities.

These final regulations directly affect all institutions with GE programs participating in title IV aid. There were 2,617 institutions in the 2015 GE cohort, of which 1,357 are small entities.

The Department has determined that the impact on small entities affected by these final regulations would not be a significant burden and will generate savings for small institutions. For these 1,357 institutions, the effect of these final regulations would be to eliminate GE paperwork burden and potential loss of title IV eligibility. Across all institutions, the net result of the institutional disclosure changes is estimated savings of $209,247,341 annually. Using the 57 percent figure for small institutions in Table 5, the estimated savings of the disclosures in the proposed regulations for small institutions is $119.3 million annually. We believe that the economic impacts of the paperwork and title IV eligibility changes would be beneficial to small institutions. Accordingly, the Secretary hereby certifies that these final regulations would not have a significant economic impact on a substantial number of small entities.

8. Paperwork Reduction Act of 1995

As part of its continuing effort to reduce paperwork and respondent burden, the Department provides the general public and Federal agencies with an opportunity to comment on proposed or continuing, or the discontinuance of, collections of information in accordance with the PRA (44 U.S.C. 3506(c)(2)(A)). This helps ensure that: The public understands the Department's collection instructions, respondents can provide the requested data in the desired format, reporting burden (time and financial resources) is minimized, collection instruments are clearly understood, and the Department can properly assess the impact of collection requirements on respondents. Respondents also have the opportunity to comment on the Department's burden reduction estimates.

A Federal agency may not conduct or sponsor a collection of information unless OMB approves the collection under the PRA and the corresponding information collection instrument displays a currently valid OMB control number. Notwithstanding any other provision of law, no person is required to comply with, or is subject to penalty for failure to comply with, a collection of information if the collection instrument does not display a currently valid OMB control number.

Comments: One commenter asserted that the Department relied upon anecdote to support its claim of burden being higher than expected upon institutions of higher education regarding providing disclosures to students. The commenter stated that this claim was not substantiated in the Paperwork Reduction Act section of the NPRM. Further, the commenter argued that the Department made no effort to quantify or substantiate its anecdotally supported claims.

Discussion: As stated above, while administrative burden is not the only reason that the Department is rescinding the GE regulations, the Department believes that the regulations do impart reporting burdens upon institutions and that requiring all institutions to adhere to GE-like regulations would add considerable burden to institutions and, in turn, costs to students. However, the Department has determined that not only will expanding the College Scorecard provide more comprehensive and useful data to current and prospective students, but since the Department can populate the Scorecard using data schools already reported for other purposes, it will be less burdensome to institutions. Since the Department will provide all of the data, we can be sure it was calculated using the same formula, and that it has the same level of reliability.

Further, the final regulations will rescind the GE regulations. That action will eliminate the burden as assessed to the GE regulations in the following previously approved information collections. We will prepare Information Collection Requests, which will be published in the Federal Register upon the effective date of this final rule, to discontinue the currently approved information collections noted below.

Changes: None.

1845-0107—Gainful Employment Disclosure Template *

	Respondents	Burden hours eliminated
Individuals	−13,953,411	−1,116,272
For-profit institutions	−2,526	−1,798,489
Private Non-Profit Institutions	−318	−27,088
Public Institutions	−1,117	−176,311
Total	−13,957,372	−3,118,160

1845-0121—Gainful Employment Program—Subpart R—Cohort Default Rates

	Respondents and responses	Burden hours eliminated
For-profit institutions	−1,434	−5,201
Private Non-Profit Institutions	−47	−172
Public Institutions	−78	−283
Total	−1,559	−5,656

1845-0122—Gainful Employment Program—Subpart Q—Appeals for Debt to Earnings Rates

	Respondents	Responses	Burden hours eliminated
For-profit institutions	−388	−776	−23,377
Private Non-Profit Institutions	−6	−12	−362
Public Institutions	−2	−4	−121
Total	−396	−792	−23,860

1845-0123—Gainful Employment Program—Subpart Q—Regulations

	Respondents	Burden hours eliminated
Individuals	−11,793,035	−1,050,857
For-profit institutions	−28,018,705	−2,017,100
Private Non-Profit Institutions	−442,348	−76,032
Public Institutions	−2,049,488	−633,963
Total	−42,303,576	−3,777,952

The total burden hours and change in burden hours associated with each OMB Control number affected by the final rule follows:

Regulatory section	OMB control No.	Burden hours	Estimated cost $36.55/hour for institutions; $16.30/hour for individuals
§ 668.412	1845-0107	−3,118,160	−$91,364,240
§§ 668.504, 668.509, 668.510, 668.511, 668.512	1845-0121	−5,656	−206,727
§ 668.406	1845-0122	−23,860	−872,083
§§ 668.405, 668.410, 668.411, 668.413, 668.414	1845-0123	−3,777,952	−116,804,291
Total		−6,925,628	−209,247,341

Intergovernmental Review

These programs are not subject to Executive Order 12372 and the regulations in 34 CFR part 79.

Assessment of Educational Impact

In accordance with section 411 of GEPA, 20 U.S.C. 1221e-4, the Secretary particularly requests comments on whether the proposed regulations would require transmission of information that any other agency or authority of the United States gathers or makes available.

Accessible Format: Individuals with disabilities can obtain this document in an accessible format (e.g., braille, large print, audiotape, or compact disc) on request to the program contact person listed under FOR FURTHER INFORMATION CONTACT.

Electronic Access to This Document: The official version of this document is the document published in the Federal Register. You may access the official edition of the Federal Register and the Code of Federal Regulations at www.govinfo.gov. At this site you can view this document, as well as all other documents of this Department published in the Federal Register, in text or Adobe Portable Document Format (PDF). To use PDF, you must have Adobe Acrobat Reader, which is available free at the site.

You may also access documents of the Department published in the Federal Register by using the article search feature at: www.federalregister.gov. Specifically, through the advanced search feature at this site, you can limit your search to documents published by the Department.

(Catalog of Federal Domestic Assistance Number does not apply.)

List of Subjects

34 CFR Part 600

• Colleges and universities
• Foreign relations
• Grant programs-education
• Loan programs-education
• Reporting and recordkeeping requirements
• Selective Service System
• Student aid
• Vocational education

34 CFR Part 668

• Administrative practice and procedure
• Aliens
• Colleges and universities
• Consumer protection
• Grant programs-education
• Loan programs-education
• Reporting and recordkeeping requirements
• Selective Service System
• Student aid
• Vocational education

Dated: June 24, 2019.

Betsy DeVos,

Secretary of Education.

For the reasons discussed in the preamble, and under the authority at 20 U.S.C. 3474 and 20 U.S.C. 1221e-3, the Secretary of Education amends parts 600 and 668 of title 34 of the Code of Federal Regulations as follows:

PART 600—INSTITUTIONAL ELIGIBILITY UNDER THE HIGHER EDUCATION ACT OF 1965, AS AMENDED

1. The authority citation for part 600 continues to read as follows:

Authority: 20 U.S.C. 1001, 1002, 1003, 1088, 1091, 1094, 1099b, and 1099c, unless otherwise noted.

2. Section 600.10 is amended by revising paragraph (c)(1) and (2) to read as follows:

§ 600.10

Date, extent, duration, and consequence of eligibility.

* * * * *

(c) Educational programs. (1) An eligible institution that seeks to establish the eligibility of an educational program must—

(i) Pursuant to a requirement regarding additional programs included in the institution's program participation agreement under 34 CFR 668.14, obtain the Secretary's approval;

(ii) For a direct assessment program under 34 CFR 668.10, and for a comprehensive transition and postsecondary program under 34 CFR 668.232, obtain the Secretary's approval; and

(iii) For an undergraduate program that is at least 300 clock hours but less than 600 clock hours and does not admit as regular students only persons who have completed the equivalent of an associate degree under 34 CFR 668.8(d)(3), obtain the Secretary's approval.

(2) Except as provided under § 600.20(c), an eligible institution does not have to obtain the Secretary's approval to establish the eligibility of any program that is not described in paragraph (c)(1) of this section.

* * * * *

3. Section 600.21 is amended by revising the introductory text of paragraph (a)(11) to read as follows:

§ 600.21

Updating application information.

(a) * * *

(11) For any program that is required to provide training that prepares a student for gainful employment in a recognized occupation—

* * * * *

PART 668—STUDENT ASSISTANCE GENERAL PROVISIONS

4. The authority citation for part 668 continues to read as follows:

Authority: 20 U.S.C. 1001-1003, 1070a, 1070g, 1085, 1087b, 1087d, 1087e, 1088, 1091, 1092, 1094, 1099c, and 1099c-1, 1221e-3, and 3474; Pub. L. 111-256, 124 Stat. 2643; unless otherwise noted.

§ 668.6

[Removed and Reserved]

5. Remove and reserve § 668.6.

6. Section 668.8 is amended by revising paragraphs (d)(2)(iii) and (d)(3)(iii) to read as follows:

§ 668.8

Eligible program.

* * * * *

(d) * * *

(2) * * *

(iii) Provide training that prepares a student for gainful employment in a recognized occupation; and

(3) * * *

(iii) Provide undergraduate training that prepares a student for gainful employment in a recognized occupation;

* * * * *

Subpart Q—[Removed and Reserved]

7. Remove and reserve subpart Q, consisting of §§ 668.401 through 668.415.

Subpart R—[Removed and Reserved]

8. Remove and reserve subpart R, consisting of §§ 668.500 through 668.516.

Footnotes

1.  “Gainful Employment Information,” Federal Student Aid, studentaid.ed.gov/sa/about/data-center/school/ge?src=press-release.

2.  Note: The term “D/E rates measure” is used in the 2014 Rule. Although the Department views this term as redundant, we use it here for clarity and consistency.

3.  “Amended Information Exchange Agreement Between the Department of Education and the Social Security Administration for Aggregate Earnings Data, ED Agreement No. 10012, SSA IEA No. 325,” www.warren.senate.gov/​imo/​media/​doc/​ED%20Agreements1.pdf.

4.  79 FR 64908.

5.  Lance Lochner and Alexander Monge-Naranjo, “Default and Repayment Among Baccalaureate Degree Earners, National Bureau of Economic Research,” NBER working paper 19882, Revised, March 2014, www.nber.org/​papers/​w19882.

6.  Lance Lochner and Alexander Monge-Naranjo, “Default and Repayment Among Baccalaureate Degree Earners, National Bureau of Economic Research,” NBER working paper 19882, Revised, March 2014, www.nber.org/​papers/​w19882;​; see also: Government Accountability Office, “Proprietary Schools: Stronger Department of Education Oversight Needed to Help Ensure Only Eligible Students Receive Federal Student Aid,” August 2009, www.gao.gov/​new.items/​d09600.pdf.

7.  Stephanie Riegg Cellini and Rajeev Davolia, Different degrees of debt: Student borrowing in the for-profit, nonprofit, and public sectors. Brown Center on Education Policy at Brookings. June 2016.

8.  Stephanie Riegg Cellini and Rajeev Davolia, Different degrees of debt: Student borrowing in the for-profit, nonprofit, and public sectors. Brown Center on Education Policy at Brookings. June 2016.

9.  Sandy Baum and Martha Johnson. Student Debt: Who Borrows Most? What Lies Ahead? Urban Institute, April 2015, www.urban.org/​sites/​default/​files/​alfresco/​publication-pdfs/​2000191-Student-Debt-Who-Borrows-Most-What-Lies-Ahead.pdf.

10.  “Updated Data for College Scorecard and Financial Aid Shopping Sheet,” Published: January 13, 2017, ifap.ed.gov/eannouncements/011317UpdatedDataForCollegeScorecardFinaidShopSheet.html; Dancy, Kim and Ben Barrett, “Fewer Borrowers Are Repaying Their Loans Than Previously Thought,” New America, January 13, 2017, www.newamerica.org/​education-policy/​edcentral/​fewer-borrowers-are-repaying-their-loans-previously-thought/;;; Kelchen, Robert, “How Much Did A Coding Error Affect Student Loan Repayment Rates?” Personal Blog Post, January 13, 2017, robertkelchen.com/2017/01/13/how-much-did-a-coding-error-affect-student-loan-repayment-rates/.

11.  Paul Fain, “College Scorecard Screwup,” Inside Higher Ed, Published: January 16, 2017, www.insidehighered.com/​news/​2017/​01/​16/​feds-data-error-inflated-loan-repayment-rates-college-scorecard;​; see also: Robert Kelchen, Higher Education Accountability (Baltimore: John Hopkins University Press, 2018), 54-55.

12.  Dancy and Barrett, www.newamerica.org/​education-policy/​edcentral/​fewer-borrowers-are-repaying-their-loans-previously-thought/​.

13.  www.higheredtoday.org/​2018/​01/​12/​increasing-community-college-completion-rates-among-low-income-students/​/.

14.  “U.S. Secretary of Education Betsy DeVos Warns of Looming Crisis in Higher Education,” Published: November 27, 2018, www.ed.gov/​news/​press-releases/​us-secretary-education-betsy-devos-warns-looming-crisis-higher-education;​; Analysis of FSA Loan portfolio with NSLDS Q12018, Federal Reserve Economic Data (Credit card delinquencies average for all commercial banks).

15.  www.whitehouse.gov/​presidential-actions/​executive-order-improving-free-inquiry-transparency-accountability-colleges-universities/​

16.  Note: Agencies “obviously” have broad discretion when reconsidering a regulation. Clean Air Council v. Pruitt, 862 F.3d 1, 8 (D.C. Cir. 2017). As the Supreme Court has noted: “An initial agency interpretation is not instantly carved in stone,” rather an agency “must consider varying interpretations and the wisdom of its policy on a continuing basis.” Chevron U.S.A. Inc. v. NRDC, Inc., 467 U.S. 837, 864-865 (1984). Significantly, this is still true in cases where the agency's review is undertaken in response to a change in administrations. National Cable Telecommunications Ass'n. v. Brand X Internet Services, 545 U.S. 967, 981 (2005).

17.  Rajashri Chakrabarti, Nicole Gorton, Michelle Jiang, and Wilbert van der Klaauw, “Who is Likely to Default on Student Loans?” Liberty Street Economics, November 20, 2017, libertystreeteconomics.newyorkfed.org/2017/11/who-is-more-likely-to-default-on-student-loans.html.

18.  Scott-Clayton, Judith (2018). “What accounts for gaps in student loan default, and what happens after.” Brookings Evidence Speaks Reports, 2(57).

19.  Cellini, Stephanie Riegg (2010). “Financial Aid and For-Profit Colleges: Does Aid Encourage Entry?” Journal of Policy Analysis and Management. 29(3): 526-52.

20.  U.S. Department of Labor. July 2018. “Job Openings and Labor Turnover Summary.” www.bls.gov/​news.release/​jolts.nr0.htm.

21.  https://www.nytimes.com/​2017/​07/​17/​theater/​harvard-graduate-theater-art-paulus.html.

22.  Cooper, Preston and Jason D. Delisle, “Measuring Quality or Subsidy? How State Appropriations Rig the Federal Gainful Employment Test,” American Enterprise Institute, March 2017, www.luminafoundation.org/​files/​resources/​measuring-quality-or-subsidy.pdf.

23.  Ibid. Note: The authors also suggest that the application of the 2014 Rule to public institutions would also be insufficient. Since public institutions still benefit from direct appropriations, the uneven playing field would still exist and disadvantage some institutions over others.

24.  NASFAA Issue Brief, “Campus-Based Aid Allocation Formula,” January 2019, www.nasfaa.org/​issue_​brief_​campus-based_​aid.

25.  Sandy Baum and Martha Johnson. Student Debt: Who Borrows More? What Lies Ahead? Urban Institute, April 2015.

26.  Note: This is not to suggest that institutions have no role to play in establishing reasonable tuition and fee costs. Even so, many public institutions have tuition and fees dictated to them by State legislators and many private institutions establish tuition and fees based on the actual cost of providing the education as well as the many amenities today's consumers demand.

27.  Federal Student Aid, 2018.

28.  Cellini and Davolia, “Different degrees of debt: Student borrowing in the for-profit, nonprofit and public sectors. Brown Center on Education Policy at Brookings.”; Gilpin, G. A., Saunders, J., Stoddard, C., “Why has for-profit colleges' share of higher education expanded so rapidly? Estimating the responsiveness to labor market changes,” Economics of Education Review 45, 2015, scholarworks.montana.edu/xmlui/bitstream/handle/1/9186/Gilpin_EER_2015_A1b.pdf;sequence=1.

29.  Cellini and Turner; Note: (pg. 5): “For-profit schools may have better counseling compared to community colleges . . . the for-profit sector has been quicker to adopt online learning technologies for undergraduate education compared to less selective public colleges.”

30.  Gregory Gilpin, et al., “Why has for-profit colleges' share of higher education expanded so rapidly? Estimating the responsiveness to labor market changes,” Economics of Education Review 45 (April 2015): 53-63; See also: Grant McQueen, “Closing Doors: The Gainful Employment Rule as Over-Regulation of For-Profit Higher Education That Will Restrict Access to Higher Education for America's Poor,” Georgetown Journal on Poverty Law Policy, Volume XIX, Number 2, Spring 2012: “The for-profit higher education industry has filled a rapidly expanding demand for higher education in American society that public and non-profit institutions of higher education have not been able to meet.” (pg. 330)

31.  Ibid.; also see: Schneider, Mark, “Are Graduates from Public Universities Gainfully Employed? Analyzing Student Loan Debt and Gainful Employment,” American Enterprise Institute, 2014, www.aei.org/​publication/​are-graduates-from-public-universities-gainfully-employed-analyzing-student-loan-debt-and-gainful-employment/​.

32.  Jennifer Ma and Sandy Baum, Trends in Community Colleges: Enrollment, Prices, Student Debt, and Completion. College Board Research Brief, April 2016.

33.  Cellini, Stephanie and Nicholas Turner, “Gainfully Employed? Assessing the Employment and Earnings for For-Profit College Students Using Administrative Data,” National Bureau of Economic Research, January 2018, www.nber.org/​papers/​w22287.

34.  nces.ed.gov/pubs2018/2018410.pdf.

35.  www.jamesgmartin.center/​2017/​06/​colleges-allowed-limit-students-federal-loans/;;; www.washingtonpost.com/​news/​grade-point/​wp/​2016/​07/​01/​the-surprising-number-of-community-college-students-without-access-to-federal-student-loans/​?noredirect=​on​utm_​term=​.cd4dd8528001.

36.  www.higheredtoday.org/​2018/​01/​12/​increasing-community-college-completion-rates-among-low-income-students/​.

37.  Shulock, Nancy, Jodi Lewis, and Connie Tan, “Workforce Investments: State Strategies to Preserve Higher-Cost Career Education Programs in Community and Technical Colleges,” Institute for Higher Education Leadership and Policy, California State University, Sacramento, August 2013, eric.ed.gov/?id=ED574441.

38.  Ibid.

39.  Ibid.

40.  Carnevale, Anthony, et al., “Learning While Earning: The New Normal,” Center on Education and the Workforce, Georgetown University, 2015, 1gyhoq479ufd3yna29x7ubjn-wpengine.netdna-ssl.com/wp-content/uploads/Working-Learners-Report.pdf; see also: Holzer, Harry J. and Sandy Baum, Making College Work: Pathways to Success for Disadvantaged Students (Washington, DC: Brookings Institution Press, 2017).

41.  Holzer and Baum.

42.  Analysis of FSA Loan portfolio with NSLDS Q12018, Federal Reserve Economic Data (Credit card delinquencies average for all commercial banks).

43.  Federal Student Aid, “Gainful Employment Information,” studentaid.ed.gov/sa/about/data-center/school/ge?src=press-release.

44.  ONet OnLine, “Bright Outlook Occupations,” www.onetonline.org/​help/​bright/​. Note: “Bright Outlook” Occupations are defined as matching at least one of the following criteria: (1) Projected to grow faster than average (employment increase of 10 percent or more) over the period 2016-2026; or (2) projected to have 100,000 or more job openings over the period 2016-2026.

45.  Task Force Report at 14.

46.  FCC v. Fox Television Stations, Inc., 556 U.S. 502, 515 (2009).

47.  Id.

48.  Deming, David, Claudia Goldin, and Lawrence Katz, “For-Profit Colleges,” The Future of Children, Vol. 23, No 1, Spring 2013, scholar.harvard.edu/files/lkatz/files/foc_dgk_spring_2013.pdf.

49.  Cellini, Stephanie Riegg (2010). “Financial Aid and For-Profit Colleges: Does Aid Encourage Entry?” Journal of Policy Analysis and Management. 29(3): 526-52.

50.  20 U.S.C 1002(b).

51.  “Gainful Employment,” 79 FR 65035, October 31, 2014.

52.  www.govinfo.gov/​content/​pkg/​GPO-CRECB-1972-pt16/​pdf/​GPO-CRECB-1972-pt16-2-2.pdf.

53.  153 Cong. Rec. 13777 (2007).

54.  Chen, Grace, “The College Cost Reduction and Access Act of 2007,” Community College Review, August 7, 2018, www.communitycollegereview.com/​blog/​the-college-cost-reduction-and-access-act-of-2007.

55.  Alcee L. Hastings, “Over 100 Members Send Bipartisan Letter to President Obama Urging Withdrawal of `Gainful Employment' Regulation,” April 26, 2011, alceehastings.house.gov/news/documentsingle.aspx?DocumentID=327789.

56.  Amendment 241 to H.R. 1 (www.congress.gov/​amendment/​112th-congress/​house-amendment/​94), was offered by Chairman John Kline (R-MN2) in 2011, and passed by the full House of Representatives. Amendment 241 was not included in the final Consolidated Appropriations Act (www.congress.gov/​bill/​112th-congress/​house-bill/​2055/​amendments), www.congress.gov/​amendment/​112th-congress/​house-amendment/​94).

57.  Senate Task Force on Higher Education Regulations, “Recalibrating Regulation of Colleges and Universities,” www.help.senate.gov/​imo/​media/​Regulations_​Task_​Force_​Report_​2015_​FINAL.pdf,, pg. 13.

58.  Cooper and Delisle, www.luminafoundation.org/​files/​resources/​measuring-quality-or-subsidy.pdf.

59.  Paul Fain, “Nebraska AG Sues Bellevue Over Nursing Program,” Inside Higher Ed, February 28, 2019, www.insidehighered.com/​quicktakes/​2019/​02/​28/​nebraska-ag-sues-bellevue-over-nursing-program.

60.  David Kidwell, “The Graduates*,” Better Government Association, November 1, 2017, projects.bettergov.org/the-graduates/.

61.  Office of Public Affairs, “Maricopa County Community College District Agrees to Pay $4 Million for Alleged False Claims Related to Award of AmeriCorps Education Awards,” Office of Public Affairs, December 1, 2014, www.justice.gov/​opa/​pr/​maricopa-county-community-college-district-agrees-pay-4-million-alleged-false-claims-related.

62.  Paul Campos, “Served,” The New Republic, April 25, 2011, newrepublic.com/article/87251/law-school-employment-harvard-yale-georgetown.

63.  Institute for Women's Policy Research, “Fact Sheet: The Gender Wage Gap by Occupation 2017 and by Race and Ethnicity,” April 2018, iwpr.org/wp-content/uploads/2018/04/C467_2018-Occupational-Wage-Gap.pdf.

64.  Stirgus, Eric, “Rising Student Loan Debt a `crisis,' DeVos Says,” Atlanta Journal-Constitution, November 27, 2018, www.ajc.com/​news/​local-education/​rising-student-loan-debt-crisis-devos-says/​kgGX56hb11yhIpOyQURlSJ/;;; Analysis of FSA Loan portfolio with NSLDS Q12018, Federal Reserve Economic Data (Credit card delinquencies average for all commercial banks).

65.  Federal Student Aid, “Official Cohort Default Rates for Schools,” www2.ed.gov/​offices/​OSFAP/​defaultmanagement/​index.html.

66.  Delisle, Jason, “Costs and Risks in the Federal Student Loan Program,” American Enterprise Institute, January 30, 2018, www.help.senate.gov/​imo/​media/​doc/​Delisle.pdf.

67.  Cellini, Stephanie and Nicholas Turner, “Gainfully Employed? Assessing the Employment and Earnings for For-Profit College Students Using Administrative Data,” National Bureau of Economic Research, January, 2018, www.nber.org/​papers/​w22287.

68.  Chou, Tiffany, Adam Looney, and Tara Watson, “Measuring Loan Outcomes at Postsecondary Institutions: Cohort Repayment Rates as an Indicator of Student Success and Institutional Accountability,'' National Bureau of Economic Research, February, 2017, www.nber.org/​papers/​w23118.

69.  Looney, Adam 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, Fall 2015, www.brookings.edu/​wp-content/​uploads/​2016/​07/​PDFLooneyTextFallBPEA.pdf.

70.  Armona, Luis, Rajashri Chakrabarti, and Michael F. Lovenheim, “How Does For-Profit College Attendance Affect Student Loans, Default, and Labor Market Outcomes?” Federal Reserve Bank of New York Staff Report No. 811, September 2018, www.newyorkfed.org/​medialibrary/​media/​research/​staff_​reports/​sr811.pdf?​la=​en.

71.  Darolia, Rajeev, et al., “Do Employers Prefer Workers Who Attend For-Profit Colleges? Evidence from a Field Experiment,” RAND Corporation, 2014, onlinelibrary.wiley.com/doi/pdf/10.1002/pam.21863.

72.  Chakrabarti, Rajashri and Michelle Jiang, “Education's Role in Earnings Employment and Economic Mobility,” Liberty Street Economic Blog, Federal Reserve Bank of New York, September 5, 2018, libertystreeteconomics.newyorkfed.org/2018/09/educations-role-in-earnings-employment-and-economic-mobility.html.

73.  www.goodcall.com/​news/​employers-dont-prefer-for-profit-over-community-college-graduates-reveals-new-study-04948.

74.  www.usnews.com/​news/​articles/​2014/​10/​17/​employers-shy-away-from-online-for-profit-graduates.

75.  Cellini and Turner, www.nber.org/​papers/​w22287.

76.  Cellini, S.R. and Turner, N. (January 2018). `2018 Gainfully Employed? Assessing the Employment and Earnings of For-Profit College Students Using Administrative Data' National Bureau of Economic Research working paper series. Available at: www.nber.org/​papers/​w22287.

77.  www.bls.gov/​ooh/​healthcare/​medical-assistants.htm.

78.  nces.ed.gov/surveys/ctes/xls/P155_2013.xls.

79.  “A Look at the Shocking Student Loan Debt Statistics for 2018,” Published May 1, 2018, studentloanhero.com/student-loan-debt-statistics/.

80.  Zach Friedman, “Student Loan Debt Statistics in 2018: A $1.5 Trillion Crisis,” Published June 13, 2018, www.forbes.com/​sites/​zackfriedman/​2018/​06/​13/​student-loan-debt-statistics-2018/​#53efceb77310.

81.  www.census.gov/​library/​publications/​2018/​demo/​p60-263.html.

82.  www.census.gov/​data/​tables/​time-series/​demo/​families/​families.html,, table MC-1.

83.  Dirk Wittenveen and Paul Attewell, The Earnings Payoff from Attending a Selective College, Social Science Research 66 (2017), 154-169.

84.  Baum, Sandy and Saul Schwartz, “How Much Debt is Too Much? Defining Benchmarks for Manageable Student Debt,” The College Board, 2006, files.eric.ed.gov/fulltext/ED562688.pdf.

85.  “Student Assistance General Provisions, Federal Family Education Loan Program, and William D. Ford Federal Direct Loan Program,” 80 FR 67204, October 30, 2015.

86.  “Student Assistance General Provisions, Federal Family Education Loan Program, and William D. Ford Federal Direct Loan Program,” 80 FR 67205, October 30, 2015.

87.  Institucion: $1,984, Leston: $2,322, per 2015 GE Final Rates.

88.  The College Board, “Median Debt by Institution Type, 2013-2014,” Trends in Higher Education, trends.collegeboard.org/student-aid/figures-tables/median-debt-institution-type-2013-14.

89.  Ma, J., et al., “Education Pays 2016: The Benefits of Higher Education and Society,” College Board, trends.collegeboard.org/sites/default/files/education-pays-2016-full-report.pdf.

90.  79 FR 64995.

91.  American Association of Cosmetology Schools v. DeVos, 258 F.Supp.3d 50 (D.D.C. 2017).

92.  As the court stated in AACS v. DeVos: “by inexplicably requiring high response rates to submit state-sponsored or survey-based alternate earnings calculations, the DOE narrowly circumscribed the alternate-earnings appeal process, making it unfeasible for certain programs to appeal their designations.” Id. at 57.

93.  Id. at 74.

94.  The AACS court noted that the existence of penalties is “irrelevant” to the issue of undercounting income. Id. at 56.

95.  Id. at 74.

96.  Id.

97.  The Department notes that the 2014 Rule has been challenged numerous times in court proceedings, notably in Association of Private Sector Colleges and Universities v. Duncan, 640 Fed.Appx. 5 (D.C.C. 2016) and Association of Proprietary Colleges v. Duncan, 107 F. Supp.3d 332 (S.D.N.Y. 2015). The argument in these cases is nearly identical. The Department observes that in the Southern District of New York case, the court rejected APC's hypothetical “absurd” results because it was not an “as applied” challenge to the rule. 107 F.Supp.3d at 367. As a result, the court left the door open to a challenge arising out of an as-applied circumstance, such as the one made by AACS two years after the Southern District of New York's ruling, referenced above.

98.  Abel, Jaison Richard Dietz, “Underemployment in the Early Career of College Graduates Following the Great Recession”, National Bureau of Economic Research, September 2016, www.nber.org/​papers/​w22654.

99.  Note: The Court in APC v. Duncan (2015) stated that the Plaintiff's argument that the 2014 Rule failed to adjust for economic cycles was “just a red herring.” 107 F.Supp.3d at 368. The court agreed with the Department at the time that recessions lasted, on average, 11.1 months, while the GE regulations gave “struggling programs multiple years to improve their results before they lose HEA eligibility.” Id. The Department points out that the Great Recession lasted eighteen months. Importantly, the Center on Budget and Policy Priorities cited that while, technically, the recession lasted from December 2007 to June 2009, the unemployment rate did not fall to where it was at the start of the recession (5%) until late 2015. (CBPP, “Chart Book: The Legacy of the Great Recession,” May 7, 2019, www.cbpp.org/​research/​economy/​chart-book-the-legacy-of-the-great-recession.) Using that unemployment data—the metric that would have the most profound impact on D/E rates outcomes—the three-year window afforded to institutions in the 2014 Rule would come up desperately short of a jobless recovery that lasted eight years.

100.  Burning Glass Technologies, “Moving the Goalposts: How Demand for a Bachelor's Degree is Reshaping the Workforce,” September 2014, www.burning-glass.com/​wp-content/​uploads/​Moving_​the_​Goalposts.pdf. (“65% of postings for Executive Secretaries and Executive Assistants now call for a bachelor's degree. Only 19% of those currently employed in these roles have a B.A.”) (pg. 5)

101.  Burning Glass Technologies, www.burning-glass.com/​wp-content/​uploads/​Moving_​the_​Goalposts.pdf.

102.  Chris Isidore, “The Great Recession's Lost Generation,” CNN Money, May 17, 2011, money.cnn.com/2011/05/17/news/economy/recession_lost_generation/index.htm; cited on 34 FR 40172.

103.  https://studentaid.ed.gov/​sa/​about/​data-center/​school/​ge?​src=​press-release.

104.  11 out of 15 cosmetology programs in Maryland passed, while four were in the zone. No cosmetology program in Maryland had a failing score.

105.  www.onetonline.org/​link/​summary/​39-5012.00.

106.  Federal Student Aid, “Subsidized and Unsubsidized Loans,” studentaid.ed.gov/sa/types/loans/subsidized-unsubsidized#how-much.

107.  Baum, Sandy and Martha Johnson, “Student Debt: Who Borrows Most? What Lies Ahead?” Urban Institute, April 2015, www.urban.org/​sites/​default/​files/​alfresco/​publication-pdfs/​2000191-Student-Debt-Who-Borrows-Most-What-Lies-Ahead.pdf.

108.  Stephanie Riegg Cellini and Rajeev Davolia, Different degrees of debt: Student borrowing in the for-profit, nonprofit and public sectors. Brown Center on Education Policy at Brookings, June 2016.

109.  nces.ed.gov/pubs2018/2018410.pdf.

110.  nces.ed.gov/pubs/web/97578g.asp.

111.  Carnevale, Anthony, et al., “Learning While Earning: The New Normal,” Center on Education and the Workforce, Georgetown University, 2015, 1gyhoq479ufd3yna29x7ubjn-wpengine.netdna-ssl.com/wp-content/uploads/Working-Learners-Report.pdf.

112.  Guida, Anthony J. and David Figuli, “Higher Education's Gainful Employment and 90/10 Rules: Unintended “Scarlet Letters” for Minority, Low-Income, and Other At-Risk Students,” The University of Chicago Law Review, 2012, lawreview.uchicago.edu/publication/higher-education%E2%80%99s-gainful-employment-and-9010-rules-unintended-%E2%80%9Cscarlet-letters%E2%80%9D.

113.  www.census.gov/​library/​publications/​2018/​demo/​p60-263.html,, Table A1.

114.  www.census.gov/​data/​tables/​time-series/​demo/​families/​families.html,, Table SHP1.

115.  www.census.gov/​data/​tables/​time-series/​demo/​families/​families.html,, Table SHP1.

116.  www.census.gov/​data/​tables/​time-series/​demo/​families/​families.html,, Pov-table4.

117.  www.census.gov/​data/​tables/​time-series/​demo/​families/​families.html,, Figure FM-1.

118.  Jennifer Ma, et al., “Education Pays 2016: The Benefits of Higher Education for Individuals and Society,” CollegeBoard, trends.collegeboard.org/sites/default/files/education-pays-2016-full-report.pdf.

119.  nces.ed.gov/pubs2017/2017051.pdf.

120.  79 FR 64910.

121.  https://nces.ed.gov/​pubs/​web/​97578g.asp.

122.  nces.ed.gov/pubs/web/97578g.asp.

123.  Bonadies, Genevieve, et al., “For-Profit Schools' Predatory Practices and Students of Color: A Mission to Enroll Rather than Educate,” Harvard Law Review Blog, July 30, 2018, blog.harvardlawreview.org/for-profit-schools-predatory-practices-and-students-of-color-a-mission-to-enroll-rather-than-educate/.

124.  Sugar, Tom, “Boosting College Completion at Community Colleges: Time, Choice, Structure and the Significant Role of States,” Complete College of America, www2.ed.gov/​PDFDocs/​college-completion/​05-boosting-college-completion-at-community-colleges.pdf.

125.  Lochner and Monge-Naranjo, www.nber.org/​papers/​w19882.

126.  U.S. Department of Education, September 2015, NCES 2015-601. Trends in Pell Grants Receipt and the Characteristics of Pell Grant Recipients: Selected Years, 1999-2000 to 2011-12.

127.  Stephanie Riegg Cellini and Rajeev Davolia, Different degrees of debt: Student borrowing in the for-profit, nonprofit and public sectors. Brown Center on Education Policy at Brookings, June 2016.

128.  Shulock, N., Lewis, J., Tan, C. (2013). Workforce Investments: State Strategies to Preserve Higher-Cost Career Education Programs in Community and Technical Colleges. California State University: Sacramento. Institute for Higher Education Leadership Policy.

129.  Cellini and Turner, www.nber.org/​papers/​w22287. See: “For profit-schools may have better counseling compared to community colleges . . . for-profit sector has been quicker to adopt online learning technologies . . . for-profits respond to local labor market demand.” (pg. 5); Richard Kazis, et al., “Adult Learners in Higher Education: Barriers to Success and Strategies to Improve Results, Employment and Training Administration,” Occasional Paper 2007-03, March 2007. files.eric.ed.gov/fulltext/ED497801.pdf.

130.  studentaid.ed.gov/sa/about/data-center/school/ge.

131.  Cellini and Turner, www.nber.org/​papers/​w22287.

132.  www.jkcf.org/​research/​state-university-no-more-out-of-state-enrollment-and-the-growing-exclusion-of-high-achieving-low-income-students-at-public-flagship-universities/​.

133.  Shulock, Lewis, and Tan, eric.ed.gov/?id=ED574441.

134.  Committee on Education and the Workforce, Report 109-231, www.congress.gov/​109/​crpt/​hrpt231/​CRPT-109hrpt231.pdf.

135.  Ibid.

136.  Community College Review, “Average Community College Debt for Graduating Students,” www.communitycollegereview.com/​average-college-debt-stats/​national-data.

137.  Walter G. Bumphus and J. Noah Brown, American Association of Community Colleges and the Association of Community College Trustees Comments on the NPRM on Gainful Employment, (Docket ID ED-2018-OPE-0042), September 13, 2018, www.aacc.nche.edu/​wp-content/​uploads/​2018/​09/​GE_​nprm_​final_​comments_​AACC_​ACCT_​091318.pdf.

138.  www.help.senate.gov/​imo/​media/​Regulations_​Task_​Force_​Report_​2015_​FINAL.pdf. (pg. 29)

139.  Bozeman, Holly, Meaghan Mingo, and Molly Hershey-Arista, “Summary Report for the 2017 Gainful Employment Focus Group,” Westat, https://www2.ed.gov/​about/​offices/​list/​ope/​summaryrpt2017gefocus317.pdf.

140.  Hurwitz, Michael and Jonathan Smith, “Student Responsiveness to Earnings Data in the College Scorecard,” SSRN, September 1, 2017, papers.ssrn.com/sol3/papers.cfm?abstract_id=2768157.

141.  See: Andriotis, AnnaMaria, “Over 60, and Crushed by Student Loan Debt,” Wall Street Journal, February 2, 2019, www.wsj.com/​articles/​over-60-and-crushed-by-student-loan-debt-11549083631.

142.  Libassi, C.J. and Miller, B. (8 June 2017). How Gainful Employment Reduces the Government's Loan Forgiveness Costs. Center for American Progress.

143.  Delisle, Jason and Alex Holt, “Safety Net or Windfall? Examining Changes to Income-Based Repayment for Federal Student Loans,” New American Foundation, October 2012, static.newamerica.org/attachments/2332-safety-net-or-windfall/NAF_Income_Based_Repayment.18c8a688f03c4c628b6063755ff5dbaa.pdf.

144.  See: studentaid.ed.gov/sa/sites/default/files/GE-DMYR-2015-Final-Rates.xls and studentaid.ed.gov/sa/about/data-center/school/ge.

145.  Shulock, Lewis and Tan.

146.  See: studentaid.ed.gov/sa/repay-loans/understand/plans.

147.  See: studentaid.ed.gov/sa/fafsa/filling-out/dependency.

148.  www.urban.org/​sites/​default/​files/​alfresco/​publication-pdfs/​2000191-Student-Debt-Who-Borrows-Most-What-Lies-Ahead.pdf.

149.  cdn.americanprogress.org/wp-content/uploads/2016/09/06111119/HigherEdWageGap.pdf.

150.  www.pewsocialtrends.org/​2013/​12/​11/​10-findings-about-women-in-the-workplace/​.

151.  Shulock, N., Lewis, J., Tan, C. (2013). Workforce Investments: State Strategies to Preserve Higher-Cost Career Education Programs in Community and Technical Colleges. California State University: Sacramento. Institute for Higher Education Leadership Policy.

152.  nces.ed.gov/pubs/web/97578g.asp.

153.  See: “Minimum Earnings Necessary to Pass D/E, Various Measures,” Submitted by Jordan Matsudaira, www2.ed.gov/​policy/​highered/​reg/​hearulemaking/​2017/​gainfulemployment.html.

154.  nces.ed.gov/npec/data/Calculating_Placement_Rates_Background_Paper.pdf.

155.  Bozeman, Holly, and Meaghan Mingo, “Summary Report for the Gainful Employment Focus Groups,” Prepared for the U.S. Department of Education, February 10, 2016, www2.ed.gov/​about/​offices/​list/​ope/​summaryrptgefocus216.pdf. Note: Student also ranked the following as “most important”: job placement rate, annual earnings rate, and completion rates for full-time and part-time students.

156.  The Higher Education Opportunity Act of 2008, Public Law 110-315. 122 Stat. 3102.

157.  www.aacu.org/​publications-research/​periodicals/​research-adult-learners-supporting-needs-student-population-no.

158.  files.eric.ed.gov/fulltext/ED497801.pdf.

159.  eric.ed.gov/?id=ED468117.

160.  Delisle and Cooper, www.aei.org/​wp-content/​uploads/​2017/​03/​Measuring-Quality-or-Subsidy.pdf.

161.  www.evergreenbeauty.edu/​blog/​how-to-build-clientele-in-cosmetology/​.

162.  79 FR 64955.

163.  www.bls.gov/​spotlight/​2018/​contingent-workers/​home.htm.

164.  Lochner and Monge-Naranjo, www.nber.org/​papers/​w19882.

165.  nces.ed.gov/pubs/web/97578g.asp.

166.  Caren A. Arbeit and Sean A. Simone, “A Profile of the Enrollment Patterns and Demographic Characteristics of Undergraduates at For-Profit Institutions,” Stats in Brief, February 2017, nces.ed.gov/pubs2017/2017416.pdf.

167.  Judith Scott-Clayton, “The Looming Student Loan Default Crisis is Worse Than We Thought,” Brookings Institute, January 11, 2018, www.brookings.edu/​research/​the-looming-student-loan-default-crisis-is-worse-than-we-thought/​.

168.  See: Matthew J. Werhner, “A Comparison of the Performance of Online Versus Traditional On-Campus Earth Science Student on Identical Exams,” Journal of Geoscience Education, 2010, files.eric.ed.gov/fulltext/EJ1164616.pdf; Anna Ya Ni, “Comparing the Effectiveness of Classroom and Online Learning: Teaching Research Methods,” Journal of Public Affairs Education, 2013, w.naspaa.org/JPAEmessenger/Article/VOL19-2/03_Ni.pdf; Alsaaty, Falih, et al., “Traditional Versus Online Learning in Institutions of Higher Education: Minority Business Students' Perceptions,” Business and Management Research, 2016, www.sciedupress.com/​journal/​index.php/​bmr/​article/​view/​9597/​5817;​; Steven Stack, “Learning Outcomes in an Online vs Traditional Course,” International Journal for the Scholarship of Teaching and Learning, January 2015, digitalcommons.georgiasouthern.edu/cgi/viewcontent.cgi?article=1491context=ij-sotl; Caroline M. Hoxby, “The Returns to Online Postsecondary Education,” NBER, February 2017, www.nber.org/​papers/​w23193.

169.  Holzer and Baum, Making College Work: Pathways to Success for Disadvantaged Students, Brookings Institute, 2017.

170.  Shulock, Lewis and Tan, eric.ed.gov/?id=ED574441.

171.  Kelchen, Robert, “The Relationship Between Student Debt and Earnings,” Brookings Institute, Brown Center Chalkboard, September 23, 2016, https://www.brookings.edu/​blog/​brown-center-chalkboard/​2016/​09/​23/​the-relationship-between-student-debt-and-earnings/​.

172.  Paul Taylor, et al., “Wealth Gaps Rise to Record Highs Between Whites, Blacks, and Hispanics,” Pew Social Demographic Trends, July 26, 2011, www.pewresearch.org/​wp-content/​uploads/​sites/​3/​2011/​07/​SDT-Wealth-Report_​7-26-11_​FINAL.pdf.

173.  Note: Study referenced here used a data set that is of questionable quality and not publicly available. In addition, the study relied on the use of birthdates and zip codes, which is not sufficient to establish matched comparison groups, since people of the same age, living in the same zip code, can substantially differ in other ways.

174.  See blog.ed.gov/2011/12/in-america-education-is-still-the-great-equalizer/ and www.census.gov/​prod/​2002pubs/​p23-210.pdf.

175.  National Student Clearinghouse Term Enrollment Estimates, Spring 2017. National Student Clearinghouse Research Center. nscresearchcenter.org/wp-content/uploads/CurrentTermEnrollment-Spring2017.pdf.

176.  Note: Association of Proprietary Colleges v. Duncan (2015), suffers from this same limitation of not having access to studies conducting following the passage of the rule.

177.  Stephanie Riegg Cellini and Rajeev Davolia, Different degrees of debt: Student borrowing in the for-profit, nonprofit and public sectors. Brown Center on Education Policy at Brookings, June 2016.

178.  Cellini, S. R. “2018 Gainfully Employed? Assessing the Employment and Earnings of For-Profit College Students Using Administrative Data,” www.nber.org/​papers/​w22287;​; Cellini, S. R., Darolia, R., and Turner, N. (December 2016). “Where do students go when for-profit colleges lose federal aid?” National Bureau of Economic Research working paper series. Available at: www.nber.org/​papers/​w22967;​; and Blagg, K., Chingos, M. (2016). Choice Deserts: How Geography Limits the Potential Impact of Earnings Data on Higher Education. Urban Institute. Available at: www.urban.org/​sites/​default/​files/​publication/​86581/​choice_​deserts_​1.pdf).

179.  NCES, nces.ed.gov/pubs2017/2017051.pdf; Holzer and Baum, Making College Work: Pathways to Success for Disadvantaged Students.

180.  See Table 3.4: Student Response Assumptions, 79 FR 211 p. 65077. Available at www.govinfo.gov/​content/​pkg/​FR-2014-10-31/​pdf/​2014-25594.pdf.

181.  Cellini, S. R. `2018 Gainfully Employed? Assessing the Employment and Earnings of For-Profit College Students Using Administrative Data' www.nber.org/​papers/​w22287.

182.  U. S. Department of Education. (February 2018). Gainful employment: background data analysis. Available at: www2.ed.gov/​policy/​highered/​reg/​hearulemaking/​2017/​geprogramdata.docx.

183.  79 FR 211 p. 65037. Available at www.govinfo.gov/​content/​pkg/​FR-2014-10-31/​pdf/​2014-25594.pdf.

184.  U.S. Department of Education. (February 2018). Gainful employment: background data analysis. Available at: www2.ed.gov/​policy/​highered/​reg/​hearulemaking/​2017/​geprogramdata.docx.

185.  Ibid.

186.  PRA calculations based on recession of information collection requests associated with existing GE requirements and use the same wage rates as the 2014 GE rule. The $16.30 rate for students was the 2012 median weekly wage rate for high school diplomas of $652 divided by 40 hours. Available at http://www.bls.gov/​emp/​ep_​chart_​001.htm as accessed in January 2014.

187.  The count of programs includes programs that had preliminary rates calculated, but were not designated with an official pass, zone, or fail status due to reaccreditation and reinstatements of eligibility during the validation process of establishing D/E rates.

188.  PRA calculations based on recession of information collection requests associated with existing GE requirements and use the same wage rates as the 2014 GE rule. The $36.55 was calculate for the 2014 GE Rule based on an assumption that 75 percent of the work would be done by staff at a wage rate equivalent to information industries sales and office workers of $33.46 and 25 percent of the work would involve those paid the equivalent of Education Services—managers with a wage rate of $45.81. Wage rates taken from http://www.bls.gov/​ncs/​ect/​sp/​ecsuphst.pdf as accessed for calculation in January 2014.

189.  The count of programs includes programs that had preliminary rates calculated, but were not designated with an official pass, zone, or fail status due to reaccreditation and reinstatements of eligibility during the validation process of establishing D/E rates.

190.  The count of programs includes programs that had preliminary rates calculated, but were not designated with an official pass, zone, or fail status due to reaccreditation and reinstatements of eligibility during the validation process of establishing D/E rates.

191.  Salary Table 2018-DCB effective January 2018. Available at www.opm.gov/​policy-data-oversight/​pay-leave/​salaries-wages/​salary-tables/​pdf/​2018/​DCB_​h.pdf.

192.  Ibid.

193.  Center for American Progress, How Gainful Employment Reduces the Government's Loan Forgiveness Costs, June 18, 2017. Available at www.americanprogress.org/​issues/​education-postsecondary/​reports/​2017/​06/​08/​433531/​gainful-employment-reduces-governments-loan-forgiveness-costs/​.

194.  New America Foundation comments on GE Regulations, pp. 17-18 available at www.regulations.gov/​document?​D=​ED-2018-OPE-0042-13659.

195.  See 79 FR 211, Table 3.4: Student Response Assumptions, p. 65077, published October 31, 2014. Available at www.regulations.gov/​document?​D=​ED-2014-OPE-0039-2390. The dropout rate increased from 5 percent for a first zone result and 15 percent for a first failure to 20 percent for the fourth zone, second failure, or ineligibility.

196.  See 79 FR 211, pp. 65081-82, available at www.regulations.gov/​document?​D=​ED-2014-OPE-0039-2390.

197.  Stephanie R. Cellini, Rajeev Darolia, and Lesley J. Turner, Where Do Students Go When For-Profit Colleges Lose Federal Aid? NBER Working Paper No. 22967 December 2016 JEL No. H52, I22, I23, I28. Available at www.nber.org/​papers/​w22967.pdf. Finds a 3 percent decrease in overall enrollment within counties of Pell Grant recipients from sanctions on for-profit institutions.

198.  79 FR 65080.