Association of Proprietary Colleges v. Duncan et al
Filing
64
OPINION: For the foregoing reasons, the Association of Proprietary Colleges' motion for summary judgment [DI 22] is denied. The Department of Educations cross-motion for summary judgment dismissing the complaint [DI 32] is granted. The Clerk of Court shall enter judgment and close the case. SO ORDERED. (Signed by Judge Lewis A. Kaplan on 5/27/2015) (kl)
UNITED STATES DISTRICT COURT
SOUTHERN DISTRICT OF NEW YORK
------------------------------------------x
ASSOCIATION OF PROPRIETARY COLLEGES,
Plaintiff,
–against–
14-cv-8838 (LAK)
ARNE DUNCAN, in his official capacity as Secretary
of the Department of Education, et al.,
Defendants.
------------------------------------------x
OPINION
Appearances:
Michael D. Hays
Jonathon Glass
Alyssa T. Saunders
William J. Schwartz
Michael A. Blasie
COOLEY LLP
Attorneys for Plaintiff
John D. Clopper
Assistant United States Attorney
PREET BHARARA
UNITED STATES ATTORNEY
Marcia Berman
Michelle Bennett
Greg Dworkowitz
United States Department of Justice
Attorneys for Defendants
LEWIS A. KAPLAN, District Judge.
Colleges and universities operated for profit play “an important role in serving
traditionally underrepresented populations of students” such as those who are older, poorer, or less
well educated, in addition to those who are women, Black, Hispanic, or single parents.1 Many of
these institutions offer classes online, as well as in-person classes at convenient times and locations.2
For these reasons, among others, the for-profit educational sector “has experienced tremendous
growth in recent years,”3 and there is no dispute that for-profit colleges and universities make
education more accessible to millions of students across the country.
And yet, despite these not insignificant benefits, the surging popularity of proprietary
institutions during the recent recession exposed, in the view of the U.S. Department of Education
(“DOE” or the “Department”), some “troubling outcomes and practices” in the industry.4 For-profit
schools often charge higher tuition than comparable public institutions—in some cases up to four
times higher—and students enrolled at proprietary colleges take on correspondingly more
educational debt than their peers at public schools.5 Attendees of for-profit colleges generally are
1
Program Integrity: Gainful Employment, 79 Fed. Reg. 64,890, 65,032 (Oct. 31, 2014) (to
be codified at 34 C.F.R. pts. 600, 668).
2
See id.
3
Id.
4
Id.
5
See id. at 65,032-65,033; see also STAFF OF S. COMM. ON HEALTH, EDUC., LABOR AND
PENSIONS, 112TH CONG., REP. ON FOR PROFIT HIGHER EDUCATION: THE FAILURE TO
SAFEGUARD THE FEDERAL INVESTMENT AND ENSURE STUDENT SUCCESS 35 (Comm. Print
2012) [hereinafter, REP. ON FOR PROFIT HIGHER EDUCATION].
2
less likely than attendees of traditional schools to graduate or otherwise complete their programs and
more likely to be unemployed.6 They make less money than students who attend not-for-profit
schools, and their career earning potential is lower.7 It is no surprise, then, that those who attend forprofit schools have substantially higher short- and long-term default rates on their student loan debt
than their peers at comparable not-for-profit schools.8
These prospects notwithstanding, prospective students continue to enroll at for-profit
colleges and universities in droves, lured in many cases by what DOE views as “deceptive or
otherwise questionable information about graduation rates, job placement, or expected earnings.”9
Proprietary schools are businesses; unlike public and non-profit schools, they are operated by private
companies whose continued profitability depends in large part on enrollment growth. Accordingly,
for-profit colleges, having incentives as they do to enroll as many students as possible, often devote
Not only do students attending for-profit schools take on more debt than students who
attend community colleges or public four-year universities, but more students who attend
for-profit schools take on debt. Indeed, up to 96 percent of students who attend proprietary
institutions take out student loans. See REP. ON FOR PROFIT HIGHER EDUCATION 112-13.
By contrast, about 13 percent of community college students and 48 percent of four-year
public university students take on student loan debt to finance their educations. See Tamar
Lewin, Senate Committee Report on For-Profit Colleges Condemns Costs and Practices,
N.Y. TIMES, July 30, 2012, at A12.
6
See 79 Fed. Reg. at 65,033-34; REP. ON FOR PROFIT HIGHER EDUCATION 72-75.
7
See 79 Fed. Reg. at 65,033-34; REP. ON FOR PROFIT HIGHER EDUCATION 72-75.
8
See 79 Fed. Reg. at 65,032-33; REP. ON FOR PROFIT HIGHER EDUCATION 112-17.
DOE found that about 19 percent of borrowers who attend for-profit schools default on their
federal student loans within three years of entering repayment. See 79 Fed. Reg. at 65,033.
That contrasts with borrowers who attend public institutions, where the figure is about 13
percent. See id.
9
Id. at 65,034 (internal quotation marks omitted).
3
substantial resources to recruiting and marketing, spending more money on those pursuits than they
do on academic instruction or on student support services.10 Those recruiting and marketing
practices sometimes are suspect; there is evidence that proprietary institutions often engage in
“aggressive sales practices” that provide “misleading information” to prospective students.11 And
as enrollment at for-profit colleges and universities has increased, some such schools have increased
their profit margins by setting tuition using “sophisticated market strategies designed to maximize
revenue without regard to the poor academic and employment outcomes faced by students.”12
To finance their educations, students enrolled at for-profit trade and vocational schools
are eligible, by virtue of the Higher Education Act of 196513 (“HEA”) and the National Vocational
Student Loan Insurance Act of 1965,14 for federal financial aid if, among other things, the institutions
they attend “prepare students for gainful employment in a recognized occupation.”15 Concerned
about the “growing evidence that many for-profit programs may not be preparing students for
careers as well as comparable programs at public institutions” and about the ability of students who
attend proprietary institutions to repay their considerable educational debt,16 DOE has promulgated
10
Id. at 65,033.
11
Id.
12
REP. ON FOR PROFIT HIGHER EDUCATION 35.
13
20 U.S.C. §§ 1001-1107.
14
20 U.S.C. §§ 961-996.
15
20 U.S.C. § 1002(b)(1)(A)(i), (c)(1)(A); see also 20 U.S.C. § 1088(b)(1)(A)(I).
16
79 Fed. Reg. at 65,033.
4
regulations (the “GE Rules”) designed to curtail federal financial aid to those students who choose
programs whose graduates regularly are un- or under-prepared for “gainful employment in a
recognized occupation” and have “unaffordable levels of loan debt in relation to their earnings.”17
These regulations, scheduled to go into effect July 1, 2015, establish a framework by
which DOE will evaluate programs subject to the “gainful employment” provisions of the HEA—that
is, “nearly all educational programs at for-profit institutions . . . , as well as non-degree programs
at public and private non-profit institutions such as community colleges”—with the ultimate goal
of confining eligibility for federal financial aid to students attending programs that satisfy criteria
intended to measure the ability of those programs to prepare their students for gainful employment.18
The Association of Proprietary Colleges (“APC”), a collection of 23 degree-granting
for-profit institutions with 34 campuses in New York State, here challenges the GE Rules. APC
contends principally that the GE Rules (1) deprive its members of procedural due process; (2)
exceed DOE’s statutory authority under the HEA; and (3) are arbitrary and capricious in violation
of the Administrative Procedure Act (“APA”). The matter is before the Court on cross-motions for
summary judgment.
17
Id. at 64,890.
18
See id.
5
I.
The Statutory and Regulatory Scheme
Every year, Congress provides hundreds of billions of dollars through loan and grant
programs to help students finance their post-secondary education.19 These programs, which were
established under the HEA, are administered by DOE which, in fiscal year 2014, paid almost $134
billion in federal aid to nearly 13 million students enrolled at more than 6,100 schools across the
country.20 For proprietary colleges and universities—unlike for most public and private non-profit
institutions of higher education, which draw their funding primarily from taxes and/or
endowments—this federal student aid is a significant source of revenue.21 In the 2009-2010
academic year, for example, students attending for-profit institutions received $32 billion in federal
loans and grants authorized under the HEA—money students used to pay their educational expenses
and which represented, by some estimates, nearly 80 percent of proprietary colleges’ revenue that
year.22 In fact, for-profit programs rely so heavily on federal financial aid that the HEA prohibits
them from deriving more than 90 percent of their revenue from DOE sources.23
19
See Ass’n of Private Colls. & Univs. v. Duncan, 870 F. Supp. 2d 133, 137 (D.D.C. 2012)
[hereinafter, APCU].
20
FED. STUDENT AID, U.S. DEP’T OF EDUC., ANNUAL REPORT FY 2014, at 81, 100 (2014).
21
See 79 Fed. Reg. at 64,957; REP. ON FOR PROFIT HIGHER EDUCATION 24.
22
See REP. ON FOR PROFIT HIGHER EDUCATION 24-25; Lewin, supra note 5.
23
See 20 U.S.C. § 1094(a)(24).
6
Students, of course, must repay their loans, but when they are unable to do so—which
happens with alarming frequency among attendees of proprietary institutions24—the costs of their
unpaid loans are borne not by the schools they attended but by U.S. taxpayers. In the absence of
regulation, then, schools could “receive the benefit of accepting tuition payments from students
receiving federal financial aid, regardless of whether those students are ultimately able to repay their
loans.”25 Congress therefore included in the HEA various requirements that schools must satisfy
to remain eligible for ongoing participation in federal student aid programs. The statutory provision
at issue here—which affects only for-profit institutions and public or non-profit post-secondary
vocational institutions—is one such requirement. It limits eligibility for federal financial aid to
students attending programs that “prepare students for gainful employment in a recognized
occupation.”26
24
For-profit college students comprise about 13 percent of student loan borrowers but 47
percent of defaulters. See REP. ON FOR PROFIT HIGHER EDUCATION 13.
25
Ass’n of Private Sector Colls. & Univs. v. Duncan, 681 F.3d 427, 435 (D.C. Cir. 2012).
26
20 U.S.C. § 1002(b)(1)(A)(i), (c)(1)(A); see also 20 U.S.C. § 1088(b)(1)(A)(i).
7
A.
Statutory Background27
As originally enacted, the HEA rendered only those students attending “public or
other non-profit” institutions eligible for federal financial aid.28 The National Vocational Loan
Insurance Act of 1965 worked in tandem with the HEA and extended eligibility to students at forprofit schools but limited it to those attending programs of “postsecondary vocational or technical
education designed to fit individuals for useful employment in recognized occupations.”29 When
Congress merged the two statutes in 1968, it retained their separate definitions of eligibility—schools
previously subject to the HEA became “institution[s] of higher education,” while those previously
subject to the National Vocational Loan Insurance Act of 1965 became “vocational schools”—and
made all programs—public, non-profit, and for-profit—eligible to participate in federal financial
aid programs under Title IV of the HEA, subject, in the case of for-profit institutions, to the “useful
employment” language noted above.30
In 1992, Congress revised the eligibility rules, splitting “vocational schools” into (1)
“‘proprietary institution[s] of higher education,’” i.e., for-profit schools, and (2) “‘post-secondary
vocational institution[s],’” i.e., non-degree training and vocational programs at public and non-profit
27
The court in Association of Private Colleges & Universities v. Duncan, 870 F. Supp. 2d 133
(D.D.C. 2012), took a more comprehensive look at the statutory history of the relevant
provisions of the HEA and the National Vocational Student Loan Insurance Act of 1965.
See id. at 138-41. This Court need not repeat that thorough analysis here.
28
Id. at 138-39 (internal quotation marks omitted).
29
Id. at 139 (internal quotation marks omitted).
30
See Higher Education Amendments of 1968, Pub. L. No. 90-575, § 116(a), 82 Stat. 1014;
see also APCU, 870 F. Supp. 2d at 140.
8
schools, and requiring that all such institutions “‘prepare students for gainful employment in a
recognized occupation.’”31 It is DOE’s interpretation of this language, which remains unchanged
in the current version of the HEA,32 that APC now challenges.
B.
The 2011 GE Rules
The GE Rules at issue here are not the first trip around this track. Consistent with
its broad authority to issue regulations implementing HEA programs, DOE promulgated gainful
employment rules in 2011 similar in some ways to those at issue here (the “2011 Rules”).33 Among
other things, the 2011 Rules would have assessed whether various post-secondary programs provide
training that leads to gainful employment by applying three tests—two based on debt-to-income
ratios measuring graduates’ ability to repay their loans and one based on repayment rates designed
to measure whether all attendees (i.e., not just graduates) actually were repaying their student loan
debts.34
31
APCU, 870 F. Supp. 2d at 140 (quoting Higher Education Amendments of 1992, Pub. L.
No. 102-325, § 481, 106 Stat. 448, and 20 U.S.C. § 1088(b)(1), (c)(1) (1994)).
Whereas the National Vocational Loan Insurance Act of 1965 (and the HEA as it existed
between 1968 and 1992) required for-profit schools to prepare students for “useful
employment,” the HEA has since 1992 required such schools, among others, to prepare
students for “gainful employment.”
32
See 20 U.S.C. § 1002(b)(1)(A)(i), (c)(1)(A); see also 20 U.S.C. § 1088(b)(1)(A)(i).
33
See Program Integrity: Gainful Employment—Debt Measures, 76 Fed. Reg. 34,386 (June
13, 2011) (promulgating 34 C.F.R. § 668.7); see also 20 U.S.C. § 1221e-3.
34
See APCU, 870 F. Supp. 2d at 141-42.
9
DOE set a minimum standard for the debt repayment rate and maximum standards
for the two debt-to-income ratios—which compared debt payments to each of (1) annual earnings,
and (2) discretionary income, and which were the same in virtually all other meaningful respects as
the tests now at issue35—and set up a regime by which a program “fail[ed]” if its final figures fell
above (or below) all three standards.36 Failing programs were rendered “ineligible” for continued
participation in federal student aid programs under the HEA if they fell above (or below) all three
standards for three out of the four most recent fiscal years.37
The Association of Private Colleges and Universities, an organization not unlike
APC, challenged the 2011 Rules in the federal district court in Washington, D.C., arguing, among
other things and as APC does here, that the debt repayment rate and debt-to-income ratios exceeded
DOE’s statutory authority and that they were arbitrary and capricious in violation of the APA.38
Judge Contreras’s 2012 opinion in Association of Private Colleges & Universities v.
Duncan39 had something in it for both sides. It found that the statutory phrase “gainful employment”
was ambiguous, and that the 2011 Rules were “a reasonable interpretation” of that “ambiguous
35
See id.
The current debt-to-income ratios are not identical to those set forth in the 2011 Rules, and
the Court will address the relevant distinctions in due course.
36
76 Fed. Reg. at 34,452; see also APCU, 870 F. Supp. 2d at 144.
37
76 Fed. Reg. at 34,452; see also APCU, 870 F. Supp. 2d at 144.
38
APCU, 870 F. Supp. 2d at 145, 149.
39
870 F. Supp. 2d 133 (D.D.C. 2012).
10
statutory command.”40 It then all but upheld the debt-to-income ratios under the APA on the theory
that they “were the product of a ‘rational connection between the facts found and the choice made.’”41
Nonetheless, Judge Contreras concluded that the debt repayment rate violated the
APA’s prohibition on arbitrary and capricious agency action because it “was not based upon any
facts at all.”42 Having decided that the debt repayment rate test could not be severed from the
(reasoned and lawful) debt-to-income ratios, Judge Contreras vacated both tests and sent most of
the 2011 Rules back to DOE.43
C.
The Current GE Rules
DOE then commenced a new rulemaking that produced the GE Rules here at issue.
It took its cues from Judge Contreras and jettisoned the debt repayment metric of the 2011 Rules, and
the GE Rules now rely substantially on similar versions of the two debt-to-income ratios that Judge
Contreras found were “based upon expert studies and industry practice—objective criteria upon
which the Department could reasonably rely” to evaluate whether certain vocationally-oriented
programs are complying with the HEA by “prepar[ing] students for gainful employment in a
recognized occupation.”
40
Id. at 146, 149.
41
Id. at 153-54 (quoting Motor Vehicle Mfrs. Ass’n of U.S. v. State Farm Mut. Auto. Ins. Co.,
463 U.S. 29, 43 (1983)).
42
Id. at 154.
43
See id.
11
These two debt-to-income ratios, also called debt-to-earnings ratios (“D/E rates”),
are, then, the driving force behind the GE Rules. Broadly speaking, they measure graduates’
educational debt loads as percentages of their annual earnings and of their discretionary income as
a way of assessing students’ ability to repay their loans.44 The D/E rates are calculated using data
provided by schools and by the Social Security Administration, and they measure programs in light
of their graduates’ mean or median figures.45
For a program to satisfy the requirements of the GE Rules, its graduates—measured
in two-year cohorts—need meet only a certain threshold on one of the two D/E rates. A program
“passes” if its graduates’ average annual loan payments are less than or equal to 20 percent of
discretionary income or eight percent of annual earnings. A program “fails” if its graduates’ average
annual loan payments are more than 30 percent of discretionary income and 12 percent of annual
earnings.46 Programs that neither pass nor fail—that is, programs that do not pass but which have
at least one D/E rate in between the pass/fail numbers—are “in the zone.” A program becomes
ineligible to participate in federal aid programs under Title IV of the HEA if it fails for any two of
three consecutive years or if it fails or is “in the zone” for four consecutive years.47
44
See 34 C.F.R. § 668.404; 79 Fed. Reg. at 64,950.
45
See 34 C.F.R. §§ 668.404-.405.
46
According to DOE, the “the passing thresholds” for the D/E rates—i.e., the eight and 12
percent figures for debt-to-annual earnings and debt-to-discretionary income,
respectively—are based on “industry practices and expert recommendations.” 79 Fed. Reg.
at 64,919.
47
See 34 C.F.R. § 668.403.
12
The central matter before the Court is whether the GE Rules, which represent DOE’s
interpretation of the statutory phrase “prepare students for gainful employment in a recognized
occupation,” withstand scrutiny under the Constitution and under various principles of
administrative law and statutory construction.
II.
Standard of Review
Where, as here, “a party seeks review of agency action under the APA, the district
judge sits as an appellate tribunal,” and “[t]he entire case on review is a question of law.”48 And
while the usual summary judgment standard under Federal Rule of Civil Procedure 56 does not
apply in such cases,49 summary judgment nonetheless is “generally appropriate,” as “[t]he question
whether an agency’s decision is arbitrary and capricious . . . is a legal issue amenable to summary
disposition.”50 Indeed, “district courts reviewing agency action under the APA’s arbitrary and
capricious standard do not resolve factual issues.”51 Instead, they address legal questions in deciding
whether the agency acted arbitrarily, capriciously or in some other way that violates 5 U.S.C. § 706.
48
Am. Bioscience, Inc. v. Thompson, 269 F.3d 1077, 1083 (D.C. Cir. 2001) (internal quotation
marks omitted); see also Just Bagels Mfg. v. Mayorkas, 900 F. Supp. 2d 363, 372 n.7
(S.D.N.Y. 2012) (quoting Thompson).
49
See UPMC Mercy v. Sebelius, 793 F. Supp. 2d 62, 67 (D.D.C. 2011).
50
Noroozi v. Napolitano, 905 F. Supp. 2d 535, 541 (S.D.N.Y. 2012) (alteration and ellipsis
in original) (internal quotation marks omitted).
51
James Madison Ltd. ex rel. Hecht v. Ludwig, 82 F.3d 1085, 1096 (D.C. Cir. 1996).
13
III.
Discussion
APC characterizes the two D/E rates as “novel, untested, and complex financial
calculations” and the GE Rules as “economically non-sensical regulations that good programs
cannot pass.”52 It accuses DOE of promulgating the regulations “in a fit of anti-proprietary school
zealotry,” and contends that some of the regulatory requirements are so vague that “institutions are
left entirely in the dark with respect to what the GE Rules will actually require.”53 Ultimately, APC
argues, “the GE Rules simply reflect[] the Department’s anti-proprietary school bias” and are
designed specifically to shutter for-profit colleges.54
APC challenges the GE Rules three ways. It asserts that they (1) violate the Due
Process Clause; (2) exceed DOE’s statutory authority under the Higher Education Act of 1965; and
(3) run afoul of the APA prohibition on arbitrary and capricious agency rulemaking.55 The Court
addresses each of APC’s arguments in turn. As we shall see, if either side is guilty of any “fit of . . .
zealotry” or “bias,” it is not DOE.
52
Compl. ¶¶ 22, 72. Elsewhere, APC calls the GE Rules “methodologically unsound.”
53
Id. ¶¶ 22, 92-93.
54
Id. ¶ 90.
55
APC argues also that the GE Rules require institutions to report information regarding
students’ private, non-federal loans and that this provision violates 20 U.S.C. § 1015c,
which prohibits the creation or maintenance of federal databases containing personally
identifiable student data. See id. ¶¶ 126-129.
14
A.
Due Process56
APC’s constitutional argument is predicated on the idea that institutions of higher
education that currently are eligible to participate in federal funding programs under Title IV of the
HEA, for-profit colleges among them, “have a property and liberty interest in continued . . .
eligibility that entitles them to protection under the Due Process Clause of the United States
Constitution.”57 It contends, among other things, that the GE Rules create a regulatory framework
by which schools may be stripped of their HEA eligibility based on evidence—in this case, earnings
data provided to DOE by the Social Security Administration—that the schools never get to see.
APC argues also (1) that the Social Security Administration earnings data on which DOE and the
GE Rules rely is flawed; (2) that the GE Rules are impermissibly retroactive because they determine
compliance based in part on data for periods that precede the rulemaking; and (3) that the GE rules
are so vague and ambiguous that for-profit colleges “cannot reasonably determine what is expected
of them.”58
56
The Court is mindful of the doctrine of constitutional avoidance and of the “general principle
that dispositive issues of statutory and local law are to be treated before reaching
constitutional issues.” Wolston v. Reader’s Digest Ass’n, 443 U.S. 157, 160 n.2 (1979).
However, because the Court rejects each of APC’s arguments and therefore must reach the
constitutional question before it sooner or later, it chooses to do so sooner for organizational
and conceptual purposes.
57
Compl. ¶ 94. APC elsewhere refers to this alleged constitutionally protected interest as a
“vested right in the Title IX eligibility of its educational programs.” Id. ¶ 104.
58
Id. ¶ 112.
15
1.
Justiciability59
Before proceeding to APC’s due process challenge, the Court pauses to consider a
threshold question: whether, as DOE argues, APC’s due process claim is unripe for adjudication
based on the theory that APC has not alleged an imminent injury but rather merely speculated as to
the scope and effect of the pending regulations.
“To be justiciable, plaintiffs’ claims must be ripe for federal review.”60 Ripeness,
fundamentally, is a question of timing; the doctrine’s “basic rationale is to prevent the courts,
through premature adjudication, from entangling themselves in abstract disagreements.”61 Any
59
Neither party has raised the issue of whether APC has standing to bring this action.
“Because the standing issue goes to this Court’s subject matter jurisdiction,” however, “it
can be raised sua sponte.” Cent. States Se. & Sw. Areas Health & Welfare Fund v. MerckMedco Managed Care LLC, 433 F.3d 181, 198 (2d Cir. 2005).
“When an association asserts standing solely as the representative of its members, it ‘must
allege that its members, or any one of them, are suffering immediate or threatened injury
as a result of the challenged action of the sort that would make out a justiciable case had the
members themselves brought suit.’” Disability Advocates, Inc. v. N.Y. Coal. for Quality
Assisted Living, Inc., 675 F.3d 149, 156-57 (2d Cir. 2012) (quoting Warth v. Seldin, 422
U.S. 490, 511 (1975)). Under Supreme Court precedent, “an association has standing to
bring suit on behalf of its members when: (a) its members would otherwise have standing
to sue in their own right; (b) the interests it seeks to protect are germane to the
organization's purpose; and (c) neither the claim asserted nor the relief requested requires
the participation of individual members in the lawsuit.” Hunt v. Wash. State Apple Adver.
Comm’n, 432 U.S. 333, 343 (1977). The Court is satisfied that this case falls squarely
within the three-part test set forth in Hunt. APC has standing to bring this action.
60
Thomas v. City of N.Y., 143 F.3d 31, 34 (2d Cir. 1998).
61
Thomas v. Union Carbide Agric. Prods. Co., 473 U.S. 568, 580 (1985) (internal quotation
marks omitted).
16
analysis of ripeness must consider two factors: (1) “the hardship to the parties of withholding court
consideration,” and (2) “the fitness of the issues for judicial decision.”62
While facial challenges seeking pre-enforcement review of agency regulations are
warranted only in “limited situations,” and “principally when an individual would, in the absence
of court review, be faced with a choice between risking likely criminal prosecution entailing serious
consequences, or forgoing potentially lawful behavior,”63 there are other considerations that favor
deciding this case now. The “parties agree that the issue tendered is a purely legal one.”64 Both
sides have moved for summary judgment, “and no claim is made here that further administrative
proceedings are contemplated.”65 Perhaps more importantly, the GE Rules are almost certainly a
“final agency action” under the meaning of the APA.66 There is “no hint that this regulation is
informal . . . or tentative,” and it is scheduled to go into effect, absent judicial intervention, in a
matter of weeks.67 Finally, the impact of the GE Rules on APC and its member institutions is
62
Abbott Labs. v. Gardner, 387 U.S. 136, 149 (1967), abrogated on other grounds by
Califano v. Sanders, 430 U.S. 99 (1977).
63
Thomas, 143 F.3d at 35.
64
Abbott Labs., 387 U.S. at 149.
65
Id.
66
5 U.S.C. § 704; see also Abbott Labs, 387 U.S. at 149; CBS v. United States, 316 U.S. 407,
418-19 (1942) (holding that where regulations “have the force of law before their sanctions
are invoked as well as after,” and where “the expected conformity to them causes injury
cognizable by a court of equity,” they are “appropriately the subject of attack”) .
67
Abbott Labs, 387 U.S. at 151 (citations omitted). By contrast, a pre-enforcement challenge
“directed at possibilities and proposals only, not at a concrete plan which has been formally
promulgated and brought into operation” may indeed be unripe for review. Isaacs v.
Bowen, 865 F.2d 468, 477 (2d Cir. 1989). That, of course, is not the case at hand.
17
“sufficiently direct and immediate as to render the issue appropriate for judicial review at this
stage.”68 As in Abbott Laboratories v. Gardner, the regulations at issue here “purport to give an
authoritative interpretation of a statutory provision that has a direct effect on the day-to-day business
of” many for-profit vocationally-oriented institutions.69 The Court concludes that this case is ripe
for review.
2.
Procedural Due Process
The Constitution imposes “constraints,” ordinarily in the form of notice and a predeprivation hearing, on “governmental decisions which deprive individuals of ‘liberty’ or ‘property’
interests within the meaning of the Due Process Clause” of the Fifth and Fourteenth Amendments.70
However, because the requirements of procedural due process “apply only to the deprivation of
interests encompassed by the [Constitution’s] protection of liberty and property” and because “the
range of interests protected by procedural due process is not infinite,”71 the “first inquiry in every
due process challenge is whether the plaintiff has been deprived of a protected interest in ‘property’
or ‘liberty.’”72 Only upon a finding that the plaintiff has been deprived of such an interest do courts
68
Abbott Labs., 387 U.S. at 152.
69
Id.
70
Mathews v. Eldridge, 424 U.S. 319, 332 (1976); see also Bd. of Regents of State Colls. v.
Roth, 408 U.S. 564, 569 (1972); Barrows v. Burwell, 777 F.3d 106, 113 (2d Cir. 2015).
71
Roth, 408 U.S. at 569-70 (emphasis added).
72
Am. Mfrs. Mut. Ins. Co. v. Sullivan, 526 U.S. 40, 59 (1999).
18
look to see if the government’s procedures comport with due process.73 In evaluating whether a
particular interest is constitutionally protected, courts look “to the nature of the interest at stake.”74
a.
Property Interest
The Due Process Clause’s “procedural protection of property is a safeguard of the
security of interests that a person has already acquired in specific benefits.”75 And while such
interests may take many forms, “not all benefits programs create constitutional property interests.”76
Nor does a property interest “exist solely because of the importance of the benefit to the recipient.”77
Indeed, “[t]o have a property interest in a benefit, a person clearly must have more than an abstract
need or desire for it. He must have more than a unilateral expectation of it. He must, instead, have
a legitimate claim of entitlement to it.”78
To determine whether a given regime yields a “legitimate claim of entitlement” to
benefits, courts consider “whether the statutes and regulations governing the distribution of benefits,”
state or federal, “‘meaningfully channel[] official discretion by mandating a defined administrative
73
See id.
74
Roth, 408 U.S. at 570-71.
75
Id. at 576.
76
Kapps v. Wing, 404 F.3d 105, 113 (2d Cir. 2005).
77
Kelly Kare, Ltd. v. O’Rourke, 930 F.2d 170, 175 (2d Cir. 1991).
78
Roth, 408 U.S. at 577; see also Barrows, 777 F.3d at 113; Kapps, 404 F.3d at 113.
19
outcome.’”79 Where they do—that is, where the statutory or regulatory scheme “sets fixed eligibility
criteria for the receipt of . . . benefits”—a property interest often will be found,”80 and “any
procedures that the state establishes for the revocation of that interest must comport with federal
procedural due process requirements.”81 On the other hand, where “the statute, regulation, or contract
in issue vests in the state significant discretion over the continued conferral of [a] benefit, it will be
the rare case that the recipient will be able to establish an entitlement to that benefit.82 Thus, “[w]here
the administrative scheme does not require a certain outcome, but merely authorizes particular
actions and remedies, the scheme does not create ‘entitlements’ that receive constitutional
protection” under the Due Process Clause.83
The Supreme Court regularly has found constitutionally protected property interests
in the continued receipt of benefits flowing from social welfare programs where eligibility for those
79
Barrows, 777 F.3d at 113 (quoting Sealed v. Sealed, 332 F.3d 51, 56 (2d Cir. 2003)); see
also Kapps, 404 F.3d at 113.
Property interests, of course, are not created by the Constitution. Instead, “they are created
and their dimensions are defined by existing rules or understandings that stem from an
independent source such as state law.” Cleveland Bd. of Educ. v. Loudermill, 470 U.S. 532,
538 (1985) (internal quotation marks omitted).
80
Kapps, 404 F.3d at 113-14; see also Town of Castle Rock v. Gonzales, 545 U.S. 748, 756
(2005) (“[A] benefit is not a protected entitlement if government officials may grant or deny
it in their discretion.”); Natale v. Town of Ridgefield, 170 F.3d 258, 263 (2d Cir. 1999).
81
Kelly Kare, 930 F.2d at 175.
82
Id.
83
Sealed, 332 F.3d at 56.
20
benefits is defined by a statutory and/or administrative framework.84 So, too, it usually has found
that public school officials have a protected property interest in their continued employment.85 But
it has deemed other purported property interests—including, for example, a woman’s interest in
continued police enforcement of a restraining order—unworthy of constitutional protection.86
While the Second Circuit on more than one occasion has said that a healthcare
provider lacks a constitutionally protected property interest in continued participation in a Medicaid
program,87 neither it nor the Supreme Court has addressed the question whether a school has a
recognizable property interest in continued eligibility to participate in programs that provide federal
aid money to students. At least two other circuits have, however, and their answers to the question
differ. The D.C. Circuit has said that “schools have no ‘vested right’ to future eligibility to
participate in [federal student loan] programs,” regardless of whether those schools “expected to be
84
See, e.g., Mathews, 424 U.S. 319 (Social Security disability benefits); Goldberg v. Kelly,
397 U.S. 254 (1970) (welfare benefits).
85
See, e.g., Connell v. Higginbotham, 403 U.S. 207 (1971) (teacher with clearly implied
promise of continued employment); Slochower v. Bd. of Higher Educ., 350 U.S. 551 (1956)
(tenured college professor); Wieman v. Updegraff, 344 U.S. 183 (1952) (staff dismissed
during terms of contract). But see Roth, 408 U.S. at 578 (holding that a non-tenure track
professor who was not rehired at the expiration of a one-year contract lacked a protected
property interest in continued employment).
86
See Town of Castle Rock, 545 U.S. at 766-67.
87
See Kelly Kare, 930 F.2d at 175-76; Plaza Health Labs., Inc. v. Perales, 878 F.2d 577, 582
(2d Cir. 1989).
According to the Second Circuit, while healthcare providers have a property interest in their
status as providers qualified to participate in Medicaid programs, see Patchogue Nursing
Ctr. v. Bowen, 797 F.2d 1137 (2d Cir. 1986); Case v. Weinberger, 523 F.2d 602 (2d Cir.
1975), they do not have an interest in continued participation in such programs. See Kelly
Kare, 930 F.2d at 175-76; Plaza Health Labs., 878 F.2d at 582.
21
eligible in the future,” because the schools were not “promised or even led to believe that there
[would] be no background changes in the federal law governing the [federal student loan]
program.”88 The Seventh Circuit, by contrast, has said that schools have a constitutionally protected
property interest in “certification or eligibility status” because the “eligibility standards in the HEA
provide the sort of substantive predicate the courts have required before finding that property
interests exist.”89
This Court agrees with the D.C. Circuit for a number of reasons.
First, though they are heavily dependent upon federal financial aid, proprietary
colleges are not, of course, direct beneficiaries of the funding programs administered under the
HEA. Rather, they are no more than third party beneficiaries; they receive the federal dollars on
which they rely only if, and when, their students do. As “indirect beneficiaries of government
programs,” APC and its member institutions have no procedural due process claim.90
88
Ass’n of Accredited Cosmetology Sch. v. Alexander, 979 F.2d 859, 864 (D.C. Cir. 1992).
89
Cont’l Training Servs, Inc. v. Cavazos, 893 F.2d 877, 893 (7th Cir. 1990) (internal
quotation marks omitted).
90
O’Bannon v. Town Court Nursing Ctr., 447 U.S. 773, 787 (1980); see also Dumas v. Kipp,
90 F.3d 386, 392 (9th Cir. 1996) (“Procedural due process protections do not extend to
those who suffer indirect harm from government action. Thus, indirect beneficiaries of
government programs have no due process rights.” (citation omitted)).
The Court recognizes that the Seventh Circuit said in dictum in Continental Training
Services, Inc. v. Cavazos that “an indirect beneficiary of student aid programs” had a
property interest in its continued eligibility status under the HEA. 893 F.2d at 893. But
Cavazos does not even mention O’Bannon, and the case on which Cavazos relies, Northlake
Community Hospital v. United States, 654 F.2d 1234, 1242-43 (7th Cir. 1981), contradicts
neither the Supreme Court in O’Bannon nor this Court here. The reasoning of Cavazos
therefore is unpersuasive to this Court.
22
Second, the HEA does not “channel[] official discretion by mandating a defined
administrative outcome.”91 Rather, it gives DOE latitude to decide whether to confer Title IV
funding eligibility on a particular program. For example, a school must demonstrate to DOE that it
“is capable of adequately administering” a federal funding program under Title IV of the HEA if it
wishes to “begin and to continue to participate” in such a program.92 In evaluating whether a
particular school has the requisite administrative capability, DOE considers many factors, including
whether the institution designates a “capable individual to be responsible for administering all the
Title IV, HEA programs in which [the institution] participates.”93 In addition to considering relevant
certification requirements, DOE “may consider other factors in determining whether an individual
is capable, including, but not limited to, the individual’s successful completion of Title IV, HEA
program training . . . and previous experience and documented success in administering the Title IV,
HEA programs properly.”94 Further, administrative capability determinations may turn also on
whether the relevant institution “[a]dministers Title IV, HEA programs with adequate checks and
balances in its system of internal controls.”95 Under the HEA, then, DOE retains “significant
91
Barrows, 777 F.3d at 113 (internal quotation marks omitted).
92
34 C.F.R. § 668.16.
93
Id. § 668.16(b)(1).
94
Id. (emphasis added).
95
Id. § 668.16(c)(1) (emphasis added).
23
discretion” to make Title IV eligibility determinations.96 The statutory scheme does not “require a
certain outcome” and thus “does not create ‘entitlements’ that receive constitutional protection.”97
Third, 20 U.S.C. § 1094(c)(1)(F) does not create a constitutionally protected property
interest. That statute provides, among other things, that DOE regulations may limit, suspend, or
terminate an eligible institution’s participation in a Title IV funding program “after reasonable
notice and opportunity for hearing.”98 But the very next subsection, which governs “emergency
action[s]” and allows DOE to “withhold funds from [an] institution or its students” upon notice to
the school, says nothing about a prior hearing.99 Thus, “the combination of rights reserved” by
Congress, which allows DOE to withhold funds from proprietary colleges “immediately in certain
circumstances, casts doubt on whether the [schools’] interest in continuing as [eligible institutions],
either indefinitely or for any period without interruption, is a property right that is protected by due
process.”100
In sum, this Court holds that proprietary colleges do not have a “vested right” to
continued eligibility to participate in Title IV federal funding programs under the HEA, and as a
96
Kelly Kare, 930 F.2d at 175.
97
Sealed, 332 F.3d at 56.
98
20 U.S.C. § 1094(c)(1)(F).
99
See id. § 1094(c)(1)(G).
While subsection G states that DOE “shall provide the institution an opportunity to show
cause, if it so requests, that the emergency action is unwarranted,” it says absolutely nothing
about a pre-deprivation hearing.
100
Plaza Health, 878 F.2d at 582; see also San Juan City Coll. v. United States, 391 F.3d 1357,
1364 (Fed. Cir. 2004).
24
consequence, that APC has not alleged adequately the deprivation of a constitutionally protected
property interest.
b.
Liberty Interest
The Supreme Court “has not attempted to define with exactness the liberty”
guaranteed by the Due Process Clause,101 but it is “now a commonplace that an entity may have a
liberty interest in its good name, and if its reputation is besmirched by governmental action, it may
be entitled to a name-clearing hearing.”102 However, while a private party may not be subjected to
“public embarrassment and ridicule” by the state without receiving due process,103 an interest in “good
reputation alone, apart from a more tangible interest, is not a liberty or property interest sufficient
to invoke the procedural protections of the Due Process Clause.”104 Rather, harm to one’s reputation
triggers constitutional protections only if that harm “is coupled with the deprivation of a more
101
Roth, 408 U.S. at 572.
102
Kelly Kare, 930 F.2d at 177; see also Wisconsin v. Constantineau, 400 U.S. 433, 437 (1971)
(“Where a person’s good name, reputation, honor, or integrity is at stake because of what
the government is doing to him, notice and an opportunity to be heard are essential.”).
103
Constantineau, 400 U.S. at 436.
104
Patterson v. City of Utica, 370 F.3d 322, 329-30 (2d Cir. 2004); see also Paul v. Davis, 424
U.S. 693, 711-12 (1976); White Plains Towing Corp. v. Patterson, 991 F.2d 1049, 1062 (2d
Cir. 1993) (“Injury to reputation alone, even when inflicted by a state official, does not
deprive an individual of a liberty or property interest protected by due process, and . . . the
defamation of a former public employee after the termination of the employment
relationship does not trigger due process rights.” (internal quotation marks, ellipsis, and
citation omitted)).
25
tangible interest.”105 Often this “more tangible interest” is government employment,106 but it need
not be. Procedural due process protections “extend[] as well to defamation occurring in the course
of ‘termination of some other legal right o[r] status.’”107 The Second Circuit has referred to due
process claims predicated, as this one is, on loss of reputation (“stigma”) coupled with the deprivation
of some “more tangible interest” (“plus”), without adequate process, as “stigma-plus” claims.108
Though APC does not frame it as such, its liberty interest argument amounts to a
stigma-plus claim based on the reputational and financial harm it says its members will suffer if the
GE Rules go into effect as written. Based on the Seventh Circuit’s opinion in Cavazos,109 APC
contends that “there can be little serious doubt” that the GE Rules will cause its members “financial
detriment” based on proprietary colleges’ “heav[y] dependen[ce] upon federal funding,” and that
the “publicity surrounding” those rules—not to mention some of their provisions—“poses a serious
threat” to for-profit colleges’ “reputational interest.”110
This argument misunderstands the law. APC has neither explained, persuasively or
otherwise, how the GE Rules are “capable of being proved false,” nor shown that they are
105
Patterson, 370 F.3d at 330 (describing a “stigma-plus claim”).
106
See, e.g., id. (citing cases).
107
White Plains Towing Corp., 991 F.2d at 1063 (second alteration in original) (quoting Easton
v. Sundram, 947 F.2d 1011, 1016 (2d Cir. 1991)).
108
See, e.g., Velez v. Levy, 401 F.3d 75, 87 (2d Cir. 2005); Patterson, 370 F.3d at 330;
DiBlasio v. Novello, 344 F.3d 292, 302 (2d Cir. 2003).
109
893 F.2d 877 (7th Cir. 1990).
110
Id. at 892.
26
“derogatory” in any meaningful way.111 And a statement that a school is ineligible to continue to
participate in federal funding programs is not derogatory if it is true. APC thus has not shown how
the GE Rules would stigmatize its member institutions or otherwise cause them “reputational” harm.
Moreover, even if APC could meet the “stigma” element of the stigma-plus test, it surely cannot meet
the “plus” element. The “financial detriment” that APC says its members will suffer due to the GE
Rules is linked inextricably to the alleged stigma: proprietary colleges will suffer reputational harm,
APC says, when they have to advise students that their federal funding may be at risk, causing those
students to transfer or otherwise leave the school, taking their aid money—and for-profit schools’
lifeblood—with them. Financial loss accompanies most, if not all, defamation charges. And yet,
“financial harm . . . caused by government imposed stigma does not transform an interest in
reputation into a liberty interest.”112 Indeed, a “generalized allegation of financial harm is
insufficient to meet the required ‘tangible injury’ element for a [due process] action based on injury
to one’s reputation.”113
APC has not shown that proprietary colleges have a constitutionally protected liberty
interest in their continued eligibility to participate in federal funding programs under the HEA.
111
Sadallah v. City of Utica, 383 F.3d 34, 38 (2d Cir. 2004) (Sotomayor, J.).
112
Mosrie v. Barry, 718 F.2d 1151, 1158 (D.C. Cir. 1983); see also Strum v. Clark, 835 F.2d
1009, 1012-13 (3d Cir. 1987).
113
Cherry v. Jorling, 31 F. Supp. 2d 258, 265-66 (W.D.N.Y. 1998).
27
c.
The Process Provided
Even if its members had a constitutionally protected property and/or liberty interest
in continued eligibility to participate in federal funding programs under Title IV of the HEA, APC’s
due process claim would fail because the GE Rules afford affected schools all the process that is
constitutionally due. A brief overview of the eligibility determination process—which involves DOE,
the Social Security Administration, and the institutions themselves—illustrates why.
To calculate the D/E rates, DOE uses a multi-step process that involves participation
by both the Social Security Administration and the institutions themselves.114 After DOE compiles
a list of students who completed the relevant program during the appropriate time—a list it compiles
using information provided by the institution—it gives the school an opportunity to correct that list.115
DOE then obtains annual earnings data for the students on the list from the Social Security
Administration and uses that to generate draft D/E rates, which it provides to the school.116 In so
doing, DOE provides the institution with “the mean and median annual earnings” data as well as “the
individual student loan information used to calculate the rates, including the loan debt that was used
in the calculation for each student.”117 DOE allows the institution then to “challenge the accuracy
of the loan debt information” (but not the annual earnings data provided by the Social Security
114
See 34 C.F.R. § 668.405.
115
See id. §§ 668.404(e), .405(b)-(c), .411.
116
See id. § 668.405(d)-(e).
117
See id. § 668.405(e)(3)(i).
28
Administration) before calculating, and informing the institution of, the final D/E rates.118 Upon
publication of the final D/E rates, “an institution may file an alternate earnings appeal to request
recalculation of the program’s most recent final D/E rates.”119 The basis for such an appeal may be
“alternate earnings from an institutional survey . . . or from a State-sponsored data system.”120
APC makes three primary arguments why the GE Rules—and, in particular, the
regulations governing calculation of the D/E rates—deprive its member institutions of due process.
First, it accuses DOE of denying schools access to the individual student earnings data that underlie
the D/E rates calculation. This lack of access, APC says, renders for-profit colleges (and other
affected institutions) unable meaningfully to challenge an ineligibility determination. Second, it
asserts that the GE Rules are flawed because the Social Security Administration earnings data on
which they rely are “systematically inaccurate” due to “improper[]” calculations.121 Third, APC
asserts that the GE Rules are “impermissibly retroactive” because they depend, at least to some
extent, upon data that precede the rules’ enactment date.
“The touchstone of due process, of course, is the requirement that a person in
jeopardy of serious loss [be given] notice of the case against him and opportunity to meet it.”122 The
118
See id. § 668.405(f)-(g).
119
See id. § 668.406.
120
See id. § 668.406(b)(1).
121
Specifically, APC takes issue with the Social Security Administration’s decision to impute
zero earnings to those individuals for whom data are lacking, and with its failure accurately
to “capture self-employed or tip earnings of many proprietary institution graduates.”
122
Spinelli v. City of N.Y., 579 F.3d 160, 169 (2d Cir. 2009) (alteration in original) (internal
quotation marks omitted).
29
question of what process is due in any given situation, however, is circumstance-dependent, and
“will depend on appropriate accommodation of the competing interests involved.”123 Indeed “due
process is flexible and calls for such procedural protections as the particular situation demands;”124
“[t]here is no rigid formula that determines the constitutional sufficiency of the process employed
in connection with any given deprivation of a protected interest.”125 Accordingly, the question
whether the process provided in a given situation is constitutionally sufficient “requires analysis of
the governmental and private interests that are affected.”126 In conducting this analysis, courts
consider three factors: (1) “the private interest that will be affected by the official action;” (2) “the
risk of an erroneous deprivation of such interest through the procedures used, and the probable
value, if any, of additional or substitute procedural safeguards;” and (3) “the Government’s interest,
including the function involved and the fiscal and administrative burdens that the additional or
substitute procedural requirement would entail.”127 Having considered these factors, the Court
concludes that the GE Rules provide constitutionally sufficient process.
The Court first considers the substantiality of APC’s members’ interest. While forprofit colleges have become heavily reliant on federal student aid, that reliance is of their own
creation, not of necessity. Simply put, proprietary institutions’ purported interest in remaining
123
Logan v. Zimmerman Brush Co., 455 U.S. 422, 434 (1982) (footnote and internal quotation
marks omitted).
124
Morrissey v. Brewer, 408 U.S. 471, 481 (1972).
125
DiMichele v. Greenburgh Cent. Sch. Dist. No. 7, 167 F.3d 784, 791 (2d Cir. 1999).
126
Mathews, 424 U.S. at 334.
127
Id. at 335.
30
eligible to participate in federal student aid programs under Title IV does not hold a candle to the
private interests traditionally afforded strong protections under the Mathews test, such as the
continued receipt of welfare and unemployment benefits, the retention of one’s job, or even freedom
from imprisonment.128 For-profit colleges would not be in such dire straits were their programs
deemed ineligible under the GE Rules. Most proprietary colleges offer multiple programs, and an
ineligibility determination for one has no bearing on the eligibility of others. Further, a program
deemed ineligible is ineligible only for Title IV purposes; it remains free to enroll students able to
finance their education in other ways. There may be a case in which a determination that a thirdparty beneficiary is ineligible to continue to participate in a benefits scheme constitutes a deprivation
of the indirect beneficiary’s “livelihood,” such that there is a strong private interest in procedural
due process protections, but this is not that case.129
By contrast, DOE has a strong interest in ensuring that students—who are, after all,
the direct (and Congress’s intended) beneficiaries of Title IV federal aid programs—attend schools
that prepare them adequately for careers sufficient for them to repay their taxpayer-financed student
loans. As noted above, when students are unable to repay their debt, the cost of their unpaid loans
is borne by U.S. taxpayers, making DOE’s asserted interest in ensuring that “students are not unduly
128
See, e.g., Turner v. Rogers, 131 S. Ct. 2507, 2510-11 (2011) (freedom from imprisonment);
Gilbert v. Homar, 520 U.S. 924 (1997) (job retention); Fusari v. Steinberg, 419 U.S. 379
(1975) (unemployment benefits); Goldberg, 397 U.S. 254 (welfare benefits).
129
FDIC v. Mallen, 486 U.S. 230, 243 (1988) (noting that the Supreme Court has “repeatedly
recognized the severity of depriving someone of his or her livelihood”).
31
burdened with debt” one and the same with its interest in “safeguard[ing] the Federal [i.e., taxpayer]
investment in the [federal student aid] program.”130
Finally, there is no greater risk of an erroneous eligibility determination under the
GE Rules than there would be if there were “additional or substitute procedural safeguards”—in this
case, primarily, giving proprietary colleges’ access to the Social Security Administration earnings
data that underlie the D/E rates. It is, of course, worth noting that the GE Rules give institutions a
number of opportunities to participate in the process, thus guarding against the risk of an unsound
ineligibility determination. The institutions submit the data used to create the list of students used
to calculate the D/E rates. They have an opportunity to correct that list. They receive mean and
median annual earnings data, as well as individual student loan data, for the students on that list at
the same time they are notified of the draft D/E rates. They may challenge the loan data before DOE
calculates the final D/E rates and, once the final rates are published, they may challenge an
unfavorable determination by filing an appeal based on alternate earnings figures.131
In light of all that, APC has not shown that the marginal benefit of providing its
members with individualized earnings data would decrease any risk of an erroneous eligibility
determination. For example, it contends that the Social Security Administration data is unreliable
because it depends, at least in part, on earnings reported to the IRS and “many people do not report
all of their income.” Even if that were true, and APC has offered scant evidence to show that it is,
it would not matter. APC has not shown how providing its members access to the individualized
Social Security Administration earnings data would cure any problems associated with people who
130
79 Fed. Reg. at 64,891.
131
See 34 C.F.R. §§ 668.405-.406.
32
misrepresent their income to the IRS. If the point is that for-profit colleges could cross-check their
own records of graduates’ incomes with the Social Security Administration figures, APC has offered
no proof that people who under report their income to the IRS do not under report their income to
their alma maters (in the unlikely event that they report it to their alma maters at all). Furthermore,
a school that doubts the accuracy of the mean and/or median annual earnings data for its students
may appeal those figures if it can cite to others showing different results.
In the last analysis, APC seeks a glut of pre-eligibility-determination process for its
member institutions. And while the Supreme Court has required rigorous pre-deprivation process
on occasion,132 it has held also that the Constitution does not always demand it. In Mathews v.
Eldridge,133 for example, the Court held that Social Security disability benefits could be terminated
without a pre-deprivation hearing, that due process required no “oral presentation” of any kind, and
that various pre-deprivation “written submissions” satisfied all that the Constitution commanded.134
The private interest at stake here—for-profit colleges’ continued eligibility to participate in federal
student loan programs under the HEA—unquestionably is weaker than the private interest at stake
in Mathews, the continued receipt of disability benefits. As the Seventh Circuit said in a case upon
which APC relies heavily: “If predeprivation written submissions suffice for disability claimants,
then it is hard to argue that a provider of services, whose interest is less, should be entitled to more
132
See, e.g., Goldberg, 397 U.S. at 264.
133
424 U.S. 319 (1976).
134
Id. at 345-47, 349.
33
process.”135 The GE Rules give affected institutions at least three opportunities to make submissions
to DOE, and two of those are pre-eligibility determination. The procedures contained in the GE
Rules are constitutionally sufficient.
3.
Retroactivity
Before addressing APC’s final due process argument—that the GE Rules are
unconstitutionally retroactive because the D/E rates rely, in part, on student loan debt incurred
before the Rules go into effect—it is helpful to explain how the D/E rates are determined.
DOE calculates the D/E rates for any given year using the average student loan debt
and average annual earnings of students who completed the relevant program during what the
regulations define as the “cohort period.”136 For most, but not all, programs subject to the GE Rules,
the “cohort period” is two years, and the relevant years for D/E rate calculation purposes are the
“third and fourth award years prior to the award year for which the D/E rates are calculated.”137 For
example, D/E rates for award year 2015-2016 are calculated using data from award years 2011-2012
and 2012-2013.138
135
Cavazos, 893 F.2d at 894.
136
34 C.F.R. § 668.404(b), (c); see also id. § 668.402.
137
Id. § 668.402.
On some occasions, DOE uses a four-year cohort period covering the “third, fourth, fifth,
and sixth award years prior to the award year for which the D/E rates are calculated.” See
id. On other occasions, the cohort period, whether two or four years, represents award years
as many as nine years prior to the award year for which the D/E rates are calculated. See
id. For purposes of this analysis, however, these distinctions are immaterial.
138
See id.
34
The GE Rules include also a five-, six-, or seven-year transition period (depending
on the length of the relevant program), that DOE says will “allow institutions to pass the D/E rates
measure by reducing the loan debt of currently enrolled students.”139 During the transition period,
DOE calculates “transitional draft D/E rates” for any program that fails or is in the zone based on
the traditional, cohort period-based D/E rates.140 The transitional draft D/E rates ignore the average
loan debt of the relevant cohort period and instead consider the average loan debt of the students
who completed the relevant program during the immediately preceding award year.141 “For any
award year for which [DOE] calculates transitional draft D/E rates for a program, [it] determines
the final D/E rates for the program based on the lower of the draft or transitional draft D/E rates,”
and a program may challenge or appeal the transitional draft D/E rates under the same procedures
by which it may challenge draft D/E rates.142
APC asserts that the GE Rules are “impermissibly retroactive” in that they “impair
(or terminate) an institution’s vested right in its Title IV eligibility” based on debt incurred as long
ago as 2005. It argues further that the transition period “provides no relief” because it does not cure
the “retroactivity problem.” These arguments, too, miss the mark.
139
79 Fed. Reg. at 64,891; see also 34 C.F.R. § 668.404(g).
140
See 34 C.F.R. § 668.404(g).
141
See id.
While the transitional D/E rates disregard the cohort period’s average student loan debt, they
do depend, like the traditional D/E rates, on the cohort period’s average annual earnings.
See id.
142
Id.; see also id. §§ 668.405, .406.
35
A regulation is unconstitutionally retroactive if it “alter[s] the past legal
consequences of past actions”143 or, put another way, if it “would impair rights a party possessed
when he acted, increase a party’s liability for past conduct, or impose new duties with respect to
transactions already completed.”144 Thus, whether a regulation “operates retroactively turns on
‘whether the new provision attaches new legal consequences to events completed before its
enactment.’”145 It is, however, “well-settled that ‘[a] statute is not rendered retroactive merely
because the facts or requisites upon which its subsequent action depends, or some of them, are
drawn from a time antecedent to the enactment.’”146 Nor is a statute impermissibly retroactive
simply because it “upsets expectations based in prior law.”147
APC is right that a regulation may not “take[] away or impair[] vested rights acquired
under existing laws.”148 But APC’s member institutions have no “vested right” to continued
eligibility to participate in Title IV funding programs.149 And even if they did, the regulations here
143
Bowen v. Georgetown Univ. Hosp., 488 U.S. 204, 219 (1988) (Scalia, J., concurring); see
also Guaylupo-Moya v. Gonzales, 423 F.3d 121, 129-30 (2d Cir. 2005).
144
Landgraf v. USI Film Prods., 511 U.S. 244, 280 (1994).
145
Centurion v. Holder, 755 F.3d 115, 121 (2d Cir. 2014) (quoting Landgraf, 511 U.S. at 270).
146
Morgan Guar. Trust Co. of N.Y. v. Republic of Palau, 971 F.2d 917, 921 (2d Cir. 1992)
(alteration in original) (quoting Reynolds v. United States, 292 U.S. 443, 449 (1934)).
147
Landgraf, 511 U.S. at 269.
148
INS v. St. Cyr, 533 U.S. 289, 321 (2001) (internal quotation marks omitted); see also Hizam
v. Kerry, 747 F.3d 102, 109 (2d Cir. 2014) (citing St. Cyr).
149
See Ass’n of Accredited Cosmetology Schs., 979 F.2d at 864 (“Member schools have no
‘vested right’ to future eligibility to participate in the [federal student loan] program.”).
36
at issue would not impair that right. The GE Rules have only future effect. They neither affect
programs’ past eligibility for Title IV funding, nor require schools to refund federal student aid
money received before the regulation’s effective date. Instead, they merely “look to the recent
performance of a program’s former students in order to determine whether that program will be
eligible, in the future, to receive Title IV funds.”150 During the transition period, moreover, the effect
of pre-enactment graduates’ student loan data may be reduced, in certain circumstances, at the request
of the program.
APC insists that the Supreme Court’s decision in Landgraf v. USI Film Products151
controls this case. But Landgraf does not sustain the weight that APC puts on it. Indeed, Landgraf
and Bowen v. Georgetown University Hospital152 illustrate the important distinction between
regulations that apply directly to past conduct and those that merely “aris[e] from conduct antedating
the [regulation’s] enactment.”153
In Bowen, for example, the Supreme Court invalidated a
Department of Health and Human Services regulation that would have “requir[ed] private hospitals
to refund Medicare payments for services rendered before promulgation of the rule.”154 By contrast,
students attending an institution deemed ineligible under the GE Rules would be unable to obtain
150
APCU, 870 F. Supp. 2d at 151 (emphasis added). Judge Contreras concluded that terminating
“schools’ future participation in the Title IV . . . programs based on their past track record”
does not “amount to retroactive application.” Id. (ellipsis in original).
151
511 U.S. 244 (1994).
152
488 U.S. 204 (1988).
153
Landgraf, 511 U.S. at 269.
154
Id. at 264-65 (emphasis added); see also Bowen, 488 U.S. at 207.
37
federal financial aid in the future; funds received before the regulation’s effective date would be
unaffected. Ultimately, a regulation that relies on data concerning the past to govern future conduct,
as this one does, is a far cry from one that reaches back to penalize past conduct, like the HHS
regulation at issue in Bowen. The GE Rules “may unsettle expectations and impose burdens on past
conduct,” but that does not make them unconstitutionally retroactive.155
This Court’s decision in Abreu v. Callahan156—which in relevant part held that the
Social Security Administration improperly had applied the Personal Responsibility and Work
Opportunity Reconciliation Act of 1996 (“Welfare Reform Act”),157 a federal law that disqualified
many lawful resident aliens from receiving supplemental security income benefits and food stamps,
to applicants who filed claims (or appeals of denied claims) for benefits before the statute’s effective
date158—does not warrant a different result.
The benefits at issue in Abreu provided “subsistence-level income for the most
disabled and financially needy members of our society.”159 Such benefits were akin to the welfare
assistance at issue in Goldberg v. Kelly,160 and they raised more substantial due process concerns
155
Landgraf, 511 U.S. at 269 n.24 (“If every time a man relied on existing law in arranging his
affairs, he were made secure against any change in legal rules, the whole body of our law
would be ossified forever.” (internal quotation marks omitted)).
156
971 F. Supp. 799 (S.D.N.Y. 1997).
157
Pub. L. No. 104-193, 110 Stat. 2105, 2262 (codified at 8 U.S.C. § 1612).
158
Abreu, 971 F. Supp. at 825.
159
Id. at 803.
160
397 U.S. 254 (1970). Indeed, the supplemental security income benefits in Abreu were
“given to persons on the very margin of subsistence.” Mathews, 424 U.S. at 340.
38
than the alleged “benefit” at issue here.161 Further, the question in Abreu was whether the Social
Security Administration constitutionally could interpret the Welfare Reform Act to deprive certain
beneficiaries of welfare-type benefits accrued, though not yet received, before the statute’s effective
date.162 Here, by contrast, the question is whether DOE may use historical data to inform future
eligibility determinations. Pre-enactment eligibility is unaffected entirely. Indeed, the Court cannot
conceive of an institution “accruing” eligibility to participate in a federal student aid program in the
future.163
B.
Statutory Authority
APC contends next that DOE acted “in excess of statutory . . . authority” under the
HEA when it promulgated the GE Rules,164 and that it has “transmogrified” the phrase “prepare
students for gainful employment in a recognized occupation” into “an unrecognizable behemoth.”
This is, in effect, a two-part claim—first, that the statutory phrase “gainful employment” has a plain
and unambiguous meaning: “a job that pays,” and, second, that even if the meaning of “gainful
161
See Abreu, 971 F. Supp. at 825 (“[Supplemental security income claimants] have more than
a mere expectation of receiving benefits. They have a legally protected right to accrued
preadjudication benefits . . . .”).
162
See id. at 824.
163
It is worth nothing also that the Court in Abreu did not address an issue that more closely
resembles the question presented by this action. The Welfare Reform Act required the
Social Security Administration to “redetermine” the eligibility of previously-eligible alien
beneficiaries and to terminate benefits for those who failed to satisfy the newly enacted
statutory requirements. See 8 U.S.C. § 1612(a)(D)(i); Abreu, 971 F. Supp. at 804.
164
5 U.S.C. § 706(2)(C).
39
employment” were ambiguous, DOE’s interpretation of that language would be unreasonable and
contrary to clear congressional intent.
Courts evaluate such arguments under the framework first established in Chevron
U.S.A. v. Natural Resources Defense Council.165 Under Chevron, a court reviewing “an agency’s
construction of the statute which it administers . . . is confronted with two questions.”166 The first
question, Chevron step one, asks “whether Congress has directly spoken to the precise question at
issue.”167 If it has—that is, if Congress’s intent is clear—“that is the end of the matter; for the court,
as well as the agency, must give effect to the unambiguously expressed intent of Congress.”168 If
congressional intent is unclear, however, courts address the second question, Chevron step two,
which asks “whether the agency’s [interpretation] is based on a permissible construction of the
statute.”169 Courts should defer to the agency’s statutory interpretation as long as it is “reasonable,”170
165
467 U.S. 837 (1984).
166
Id. at 842.
167
Id.
168
Cohen v. JP Morgan Chase & Co., 498 F.3d 111, 116 (2d Cir. 2007) (internal quotation
marks omitted).
To ascertain congressional intent, courts look first to the statutory text and, if the text is
ambiguous, to traditional canons of statutory construction and, as a last resort, to legislative
history. See Daniel v. Am. Bd. of Emergency Med., 428 F.3d 408, 423 (2d Cir. 2005).
169
Chevron, 467 U.S. at 843.
170
Cohen, 498 F.3d at 116 (internal quotation marks omitted).
40
for “ambiguities in statutes within an agency’s jurisdiction to administer are delegations of authority
to the agency to fill the statutory gap in reasonable fashion.”171
APC’s statutory authority argument differs little, if at all, from one asserted by the
Association of Private Colleges and Universities in 2012. In his decision vacating the 2011 Rules,
Judge Contreras rejected that argument wholesale, concluding, among other things, (1) that “[t]here
is no unambiguous meaning of what makes employment ‘gainful’: the phrase [i.e., ‘gainful
employment’] need not mean ‘any job that pays’;” (2) that the structure, purpose, and legislative
history of the HEA do “not unambiguously foreclose the agency’s interpretation;” and (3) that “[t]he
gainful employment regulations are a reasonable interpretation of an ambiguous statutory
command.”172 The Court finds Judge Contreras’s analysis thorough, his application of Chevron
faithful to Supreme Court precedent, and his logic and reasoning persuasive. Accordingly, the Court
feels no need to replow old ground; instead, with respect to Chevron step one, it incorporates the
portions of Judge Contreras’s opinion reproduced below, with appropriate alterations:
“[APC] argues that ‘gainful employment’ unambiguously means ‘a job that pays,’
and that the Department’s attempt to define the phrase in terms of debt and income
therefore exceeds its statutory authority. [APC] points to contemporary dictionaries,
which define ‘gainful’ as [‘used to describe useful work that you are paid for’] or
‘providing an income.’ See, e.g., WEBSTER’S THIRD NEW INTERNATIONAL
DICTIONARY [928 (1965)]. It also notes that the phrase ‘gainful employment’ is used
many times in Title 20, see 20 U.S.C. §§ 1036(e)(1)(B)(ii), 1134c(a), 1135c(d)(2)
. . . , 1161g(d)(5)(B), 2[0]08(a), 4706(a), 5605(a)(2)(B)—and each of those uses,
[APC] argues, means ‘a job that pays.’ . . .
171
Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 980 (2005)
(“Filling these gaps . . . involves difficult policy choices that agencies are better equipped
to make than courts.”).
172
APCU, 870 F. Supp. 2d at 146, 149 (alterations and internal quotation marks omitted).
41
The Department replies that Congress did not provide a precise definition of what
it means to ‘prepare students for gainful employment in a recognized occupation.’
The Department cites to many contemporary dictionaries that define ‘gainful’ as
‘profitable’ or ‘lucrative,’ thereby implying (the Department argues) an excess of
returns over expenses. See, e.g., BLACK’S LAW DICTIONARY 807 (4th ed. 1951);
WEBSTER’S NEW INTERNATIONAL DICTIONARY 1026 (2d ed. 1958); NEW STANDARD
DICTIONARY 1000 (Funk & Wagnalls Co. 1946). It also argues that the operative
statutory phrase is not simply ‘gainful employment’ but rather ‘gainful employment
in a recognized occupation,’ and suggests that the fuller phrase connotes
employment in an established occupation—and therefore, presumably, a decently
paying one. It further argues that ‘gainful employment’ means something different
in the context of fellowships for graduate study, for example, the recipients [of]
which are generally barred from such employment, see, e.g., 20 U.S.C.
§ 1036(e)(1)(B)(ii), than in the context of training for entrance into a recognized
occupation[, thus distinguishing the phrase ‘gainful employment in a recognized
occupation’ from the mere use of ‘gainful employment’ in other provisions of the
HEA]. The Department therefore concludes that the phrase ‘gainful employment in
a recognized occupation’ is ambiguous and that in enacting it Congress delegated
interpretive authority to the Department, whose interpretation ought therefore to be
evaluated under step two of the Chevron analysis.
The court agrees. There is no unambiguous meaning of what makes employment
‘gainful’: the phrase need not mean ‘any job that pays.’ ‘Gainful employment’ does
not unambiguously encompass work for minimal gain, nor does it necessarily
describe the gross profits from a given activity rather than the net gains derived
therefrom. Moreover—and more importantly—the relevant statutory command is
that a given program ‘prepare students for gainful employment in a recognized
occupation.’ The Department’s regulations are an attempt to assess whether certain
programs in fact provide such preparation. See, e.g., [79 Fed.Reg. at 64,890 (“The
[D/E rates] define[] what it means to prepare students for gainful employment by
establishing measures that assess whether programs provide quality education and
training to their students that lead to earnings that will allow students to pay back
their student loan debts.”).] The real question, then, is not how much gain is enough
but rather how much preparation is enough. The Department has attempted to
answer that question by reference to the economic success of a program’s former
students. The statute does not ‘unambiguously foreclose[] the agency’s
interpretation,’ [Brand X, 545 U.S. at 982-83], because it does not tell the
Department how to determine which programs actually prepare their students and
which programs do not. ‘The power of an administrative agency to administer a
congressionally created . . . program necessarily requires the formulation of policy
and the making of rules to fill any gap left, implicitly or explicitly, by Congress.”
Chevron, 467 U.S. at 843 (ellipsis in original) [(internal quotation marks omitted)].
The means of determining whether a program ‘prepare[s] students for gainful
42
employment in a recognized occupation’ is a considerable gap, which the
Department has promulgated rules to fill.”173
That brings us to Chevron step two, which in this case requires the Court to determine
whether the GE Rules are “based on a permissible construction of the statute.”174
APC musters many arguments that DOE’s interpretation of the HEA is unreasonable.
It contends first that the GE Rules, and specifically the D/E rates, constitute an “‘elephant in a
mousehole’—a policy change so large that Congress could not have meant to authorize it in the
statutory language on which the Department relies.”175 Indeed, APC characterizes the GE Rules as
a “sea-change in Title IV eligibility criteria, affecting tens of thousands of educational programs,
millions of students, and billions of dollars” in federal student loan funds, and asserts that Congress
would not have hidden such a “drastic change” in the “vague terms” of the phrase “prepare students
for gainful employment in a recognized occupation.”
Here, too, the Court relies on Judge Contreras, who considered and rejected a
substantially identical argument in his 2012 opinion vacating the 2011 Rules:
“The [GE Rules] are a significant regulatory intervention, but they do not suggest
that the Department has found ‘an elephant in a mousehole.’ That phrase entered the
lexicon of administrative law in Whitman v. American Trucking Associations, Inc.,
173
See id. at 145-46 (some alterations in original) (emphasis added) (some citations,
alterations, and internal quotation marks omitted).
APC’s argument that the GE Rules improperly exclude non-Title IV students is
unpersuasive. First, the whole point of the GE Rules is to evaluate programs’ performance
for the purpose of continuing eligibility for Title IV funds. Second, and more importantly,
DOE is barred from collecting data on students who do not receive Title IV funds. See
Ass’n of Private Sector Colls. & Univs. v. Duncan, 930 F. Supp. 2d 210, 221 (D.D.C. 2013).
174
Chevron, 467 U.S. at 843.
175
APCU, 870 F. Supp. 2d at 147.
43
531 U.S. 457 (2001), when the Supreme Court said that “Congress . . . does not alter
the fundamental details of a regulatory scheme in vague terms or ancillary
provisions—it does not, one might say, hide elephants in mouseholes.” Id. at 468.
The Whitman Court cited to FDA v. Brown & Williamson Tobacco Corp., 529 U.S.
120 (2000), for that proposition; there, the Court held that Congress had not
delegated to the Food and Drug Administration the authority to regulate tobacco as
a ‘drug’ under the Food, Drug, and Cosmetic Act, 21 U.S.C. § 301 et seq.—a power
that the agency had always disavowed. The Court concluded that ‘Congress could
not have intended to delegate a decision of such economic and political significance
to an agency in so cryptic a fashion.’ Id. at 160. The Whitman Court also invoked
MCI Telecommunications Corp. v. AT&T Co., 512 U.S. 218 (1994), to establish that
elephants do not hide in mouseholes. In MCI, the Court held that the Federal
Communications Commission’s statutory authority to ‘modify any requirement’ did
not empower it to alter the requirement underpinning the rate regulation of common
carriers. Id. at 231 (“It is highly unlikely that Congress would leave the
determination of whether an industry will be entirely, or even substantially,
rate-regulated to agency discretion—and even more unlikely that it would achieve
that through such a subtle device as permission to ‘modify’ rate-filing
requirements.”). In Whitman itself, the Court concluded that since the power to
consider costs when setting air quality standards was an ‘elephant’—that is, a major
issue of public policy—and the statutory command to set those standards with ‘an
adequate margin of safety’ a ‘mousehole,’ it was ‘implausible that Congress would
give to the EPA through these modest words the power to determine whether
implementation costs should moderate national air quality standards.’ 531 U.S. at 468.
Neither the elephant nor the mousehole is present here. Although the Department’s
regulation is significant, it does not approach the scale of the elephantine
interventions described above. Nor is the statutory language the Department invokes
especially broad or obscure. Concerned about inadequate programs and
unscrupulous institutions, the Department has gone looking for rats in ratholes—as
the statute empowers it to do.”176
APC argues also that the GE Rules “conflict[] with the . . . statutory framework” of
the HEA by “usurp[ing] the traditional role of accrediting bodies,” like the New York Board of
Regents, in evaluating student achievement and in determining the acceptable educational standards
for proprietary institutions. Indeed, APC accuses DOE of “reaching down to the program level to
effectively dictate which programs schools can offer” and characterizes the GE Rules as “a complete
176
Id. at 147-48 (some citations, alterations, and internal quotation marks omitted).
44
re-write of New York State curricula”—a “re-write” that “supersede[s] the role of state agencies and
impose[s] debt metrics that ultimately dictate the content of educational programs” and that operates
in “direct[] conflict with the New York State Board of Regents’ standards.” Ultimately, APC
contends, DOE has “undermine[d] the [New York Board of] Regents’ established regulatory
standards by imposing metrics that disqualify” otherwise high-performing programs.
This argument is quite surprising, but not for its merit. It is surprising because it is
at best ill-conceived and at worst misleading.
The GE Rules say nothing whatsoever about the content of educational programs.
They are concerned only with the quality of those programs as measured by graduates’ ability to
repay their loans. Federal law does prevent DOE from dictating state schools’ curricula or otherwise
micro-managing their day-to-day operations,177 but it says nothing about prohibiting DOE from
conditioning schools’ eligibility for federal student aid on performance-based metrics measuring
graduates’ success. APC seeks to create a conflict with 20 U.S.C. § 1232a out of whole cloth.
APC contends also that the GE Rules, and what they are designed to measure, are not
a reasonable construction of the statutory requirement that programs “prepare students for gainful
employment in a recognized occupation.” But “the adequacy of a program’s preparation is difficult
to measure—and it is reasonable to consider students’ success in the job market as an indication of
whether those students were, in fact, adequately prepared. If ‘a program of training to prepare
177
See 20 U.S.C. § 1232a (prohibiting the federal government from, among other things,
exercising “any direction, supervision, or control over the curriculum, program of
instruction, administration, or personnel of any educational institution, school, or school
system”).
45
students for gainful employment’ does not in fact lead to jobs for any of its students, it is reasonable
to conclude that those students were not truly prepared.”178
Finally, APC criticizes DOE for its allegedly selective reliance on the legislative
history of the HEA, which APC says establishes that “Congress never intended to empower the
Department to deny eligibility [to participate in Title IV funding programs] based on [the
Department’s] conclusion regarding program quality.” APC argues, in effect, that the GE Rules
gave too much weight to the testimony of a single witness and not enough to various other sources,
including committee reports and the testimony of other, purportedly more persuasive witnesses.
Whatever its value in divining congressional intent, it is undeniable that “legislative
history is itself often murky, ambiguous, and contradictory.”179 Indeed, that is the case here. Portions
of the legislative history of the HEA, including congressional reports and witness testimony from
two people involved in drafting the legislation, support APC’s claim that “accreditation [and another
statutory requirement] fully resolved Congress’s concerns regarding program quality as an eligibility
factor.”180 But other portions do not.181 Taking heed, then, of the Supreme Court’s warning that
“[j]udicial investigation of legislative history has a tendency to become . . . an exercise in looking
178
APCU, 870 F. Supp. 2d at 147 (citation omitted).
179
Exxon Mobil Corp. v. Allapattah Servs., Inc., 545 U.S. 546, 568 (2005).
180
See, e.g., S. REP. NO. 89-758, at 12 (1965), reprinted in 1965 U.S.C.C.A.N. 3742, 3753
(discussing, among other things, the accreditation requirements of the HEA).
181
See, e.g., id. at 5-6 (addressing congressional concern about graduates’ earnings and asserting
that it would be “reasonable” to consider the pace at which student loans are repaid after
graduation when determining eligibility to participate in Title IV funding programs).
46
over a crowd and picking out your friends,”182 the Court concludes only that the legislative history
of the HEA sheds at best an inconsistent light on what Congress meant when it directed DOE to
provide Title IV funding only to schools that “prepare students for gainful employment in a
recognized occupation.”183 In cases such as these, however, ambiguity in the statute—the tie, as it
were—goes to the agency.184
The Court holds that the GE Rules are a reasonable interpretation of an ambiguous
statutory command.185 It turns now to APC’s argument that the GE Rules are arbitrary and
capricious in violation of the APA.
C.
Administrative Procedure Act
APC’s last argument is that the GE Rules violate the APA because they are
“arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law”186 and that
they are not supported by substantial evidence in the record.
182
Allapattah Servs., 545 U.S. at 568 (internal quotation marks omitted).
183
APC asserts a number of other arguments, regarding, among other things, an alleged
conflict between the GE Rules and the role APC says is reserved for various accrediting
bodies; alleged overreach by DOE in seeking indirectly to regulate tuition at GE programs;
and Congress’s alleged acquiescence to what APC characterizes as DOE’s “consistent[]”
interpretation of the phrase “gainful employment” as meaning “a job that pays.” The Court
has considered these arguments and concluded that each is without merit.
184
See Cohen, 498 F.3d at 116 (noting that Chevron step two “instructs [courts] to defer to an
agency’s interpretation of [a] statute, so long as it is reasonable.” (internal quotation marks
omitted)).
185
See APCU, 870 F. Supp. 2d at 149.
186
5 U.S.C. § 706(2)(A).
47
Judicial review of agency action is “deferential”187 and its scope “narrow.”188 Thus,
while judicial inquiry into agency decisionmaking must be “searching and careful,”189 a reviewing
court “may not ‘substitute its judgment for that of the agency.’”190 The critical question in such cases
is merely whether the record shows that the agency “examine[d] the relevant data and articulate[d]
a satisfactory explanation for its action.”191 Put another way, “the agency’s decision must reveal a
‘rational connection between the facts found and the choice made’”192 such that a reviewing court
may conclude that the agency’s action “was the product of reasoned decisionmaking.”193
Courts reviewing agency action “must assess, among other matters, whether the
decision was based on a consideration of the relevant factors and whether there has been a clear
error of judgment.”194 Agency action generally will be set aside “if the agency has relied on factors
187
Guertin v. United States, 743 F.3d 382, 385 (2d Cir. 2014) (internal quotation marks
omitted).
188
Nat’l Audubon Soc’y v. Hoffman, 132 F.3d 7, 14 (2d Cir. 1997).
189
Marsh v. Or. Natural Res. Council, 490 U.S. 360, 378 (1989) (internal quotation marks
omitted).
190
Natural Res. Def. Council v. EPA, 658 F.3d 200, 215 (2d Cir. 2011) (quoting Citizens to
Pres. Overton Park v. Volpe, 401 U.S. 402, 416 (1971), abrogated on other grounds by
Califano v. Sanders, 430 U.S. 99 (1977)).
191
State Farm, 463 U.S. at 43.
192
Natural Res. Def. Council v. FAA, 564 F.3d 549, 555 (2d Cir. 2009) (quoting State Farm,
463 U.S. at 43).
193
State Farm, 463 U.S. at 52.
194
Bechtel v. Admin. Review Bd., 710 F.3d 443, 446 (2d Cir. 2013) (quoting Judulang v.
Holder, 132 S. Ct. 476, 484 (2011)); see also State Farm, 463 U.S. at 43.
48
which Congress has not intended it to consider, entirely failed to consider an important aspect of the
problem, offered an explanation for its decision that runs counter to the evidence before the agency,
or is so implausible that it could not be ascribed to a difference in view or the product of agency
expertise.”195 Finally, while “an agency changing its course” is in some circumstances “obligated
to supply a reasoned analysis for the change,” agencies “must be given ample latitude to adapt their
rules and policies to the demands of changing circumstances.”196
APC asserts a litany of reasons why the GE Rules are arbitrary and capricious, and
its reasons broadly fall into three categories: (1) DOE relied on flawed or irrelevant data, failed to
consider relevant data, and employed unreliable and biased methods in formulating the GE Rules;
(2) the GE Rules will confuse the public and yield absurd results due to their disproportionate impact
on for-profit colleges; and (3) DOE’s “ideologically driven” GE Rules are inconsistent with—and
more restrictive than—an earlier version of those rules.
With respect to its first claim—that there are problems with the data DOE considered
(and chose not to consider) and with the methodology underlying the GE Rules—APC contends,
among other things, that there is “no evidence at all of the relationship between the D/E rates and
program quality” and, similarly, that there is no rational connection between the facts before DOE
and the GE Rules. According to APC, in fact, the GE Rules are the product not of sound research,
but of “result-driven statistical shenanigans”—shenanigans that APC says are evident in DOE’s
195
State Farm, 463 U.S. at 43; see also NRDC v. EPA, 658 F.3d at 215.
196
State Farm, 463 U.S. at 42 (internal quotation marks omitted); see also Brand X Internet
Servs., 545 U.S. at 981 (“Unexplained inconsistency is, at most, a reason for holding an
interpretation to be an arbitrary and capricious change from agency practice under the
Administrative Procedure Act.” (emphasis added)).
49
“flawed” regression analyses. Rather than measure program quality, APC says, the GE Rates
“largely measure student characteristics, such as family income or minority status.” In effect, APC
suggests, the programs most likely to fail the GE Rules are not those whose students are unprepared
for gainful employment in a recognized occupation, but rather those that enroll more minorities and/or
poor students. The (asserted) fact that the GE Rules disproportionately will punish the most
demographically diverse institutions rather than those institutions falling short of the HEA’s statutory
command, APC says, renders the regulations arbitrary and capricious.
Putting aside for the moment that Judge Contreras correctly (in this Court’s view)
rejected this very claim in his 2012 opinion with respect to the 2011 Rules,197 APC’s argument on
this point appears utterly to disregard the extensive statistical analyses underlying the GE Rules.
In response to commenters who asserted the same thing that APC now does—i.e., “that the proposed
regulations primarily measure student characteristics instead of program quality”—DOE conducted
a series of multivariate regression analyses before promulgating the final regulations.198 From these
analyses, DOE concluded, among other things, that “passing, zone, and failing programs have very
similar proportions of low-income, non-traditional, female, white, Black, and Hispanic students”
and that the GE Rules “do not primarily measure student demographics because indicators of many
student characteristics have similar passing rates across quartiles.”199 DOE’s “determination that
197
See APCU, 870 F. Supp. 2d at 150-51 (concluding that DOE’s “determination that the debt
measures appropriately measured whether a program prepared its students for gainful
employment in a recognized occupation rather than a program’s demographics” was “not
arbitrary” in light of, among other things, DOE’s performance of “a series of multivariate
regression analyses”).
198
See 79 Fed. Reg. at 65,043-52.
199
Id. at 65,045, 65,052.
50
the [GE Rules] appropriately measure[] whether a program prepare[s] its students for gainful
employment in a recognized occupation rather than a program’s demographics was therefore not
arbitrary.”200
APC finds fault also with what it characterizes as DOE’s failure to follow its own
Information Quality Guidelines—which APC says require DOE not only to use a “state of the art
methodology” but to have that methodology peer reviewed—and with its decision to calculate D/E
rates based on graduates’ income as soon as 18 months after graduation, when earnings generally
are lowest, and which APC says is too soon adequately to assess the lifetime value of an education.
APC’s argument that the GE Rules run afoul of DOE’s Information Quality
Guidelines misunderstands the purpose and limitations of those guidelines, which “are intended only
to improve the internal management of the Department” and “do not create any private right of
action to be used by any party against the government in a court of law or in an administrative
hearing.”201 Indeed, APC’s challenge is brought under the APA, which requires only that agencies
provide the public with notice of any proposed substantive rulemaking as well as an opportunity to
comment thereon.202 The APA says absolutely nothing about a requirement that a proposed
rulemaking be peer reviewed or based on a “state of the art methodology,” whatever that means.203
200
APCU, 870 F. Supp. 2d at 151.
201
U.S. DEP’T OF EDUC., INFORMATION QUALITY GUIDELINES 2 (2002), available at
http://www2.ed.gov/policy/gen/guid/iq/infoqualguide.pdf.
202
See 5 U.S.C. § 553(b), (c).
203
See Vt. Yankee Nuclear Power Corp. v. Natural Res. Def. Council, 435 U.S. 519, 524
(1978) (noting that 5 U.S.C. § 553 “established the maximum procedural requirements
which Congress was willing to have the courts impose upon agencies in conducting
51
The Court rejects also APC’s argument that DOE “arbitrarily picked the time period
for measuring earnings” and that it should have engineered the GE Rules to consider the value of
an education over decades rather than beginning to focus on earnings as soon as 18 months after
graduation. APC asserts that DOE “provides no justification” for its decision but it forgets to mention
that the Court may “uphold a decision of less than ideal clarity if the agency’s path may reasonably
be discerned.”204 Here, that path is easily discernible. Measuring income shortly after graduation,
DOE found, provides a clearer picture of the “current or recent performance of the program” and
is a far better indicator than lifetime earnings of whether a recent graduate is capable of paying off
student debt when that debt becomes due.205 Put another way, income earned in the first few years
after graduation is more obviously tied to one’s training than income earned decades after one leaves
school, as Judge Contreras found in dismissing this very argument in his 2012 opinion.206
APC takes issue also with the D/E rates’ eight percent annual earnings and 20 percent
discretionary income passing thresholds, which it says are arbitrary and irrational because they are
based on mortgage data that have little to no relevance to recent graduates.
rulemaking procedures”).
204
State Farm, 463 U.S. at 43 (internal quotation marks omitted).
205
79 Fed. Reg. at 64,931; see also id. at 64,914 (“[T]he return on investment from training
may well be experienced over a lifetime, but benefits ultimately available over a lifetime
may not accrue soon enough to enable the individual to repay the student loan debt under
and within the schedules available under the title IV, HEA programs.”).
206
See APCU, 870 F. Supp. 2d at 152 (“[T]he Department rationally concluded that considering
a significantly longer earnings window in calculating the debt-to-income tests could weaken
or sever the connection between earnings and education.” (internal quotation marks
omitted)).
52
This argument mischaracterizes the GE Rules and the findings on which they are
based. It is true that eight percent “has been a fairly common mortgage-underwriting standard.”207
It is equally true, however, that “the 8 percent cutoff has long been referred to as a limit for student
debt burden,” and DOE not only cites to at least four studies that have accepted this standard in the
context of student debt, but refers also to various states—and the National Association of Student
Financial Aid Administrators—that have established their own student loan debt guidelines based
on the eight percent threshold.208 Similarly, economists on which DOE relied in promulgating the
GE Rules “proposed a benchmark for a manageable debt level of not more than 20 percent of
discretionary income,” having concluded that a higher ratio would be “unreasonable under virtually
all circumstances.”209 The annual earnings and discretionary income thresholds, in other words,
“were based upon expert studies and industry practice—objective criteria upon which the
Department could reasonably rely.”210 Bearing this in mind, the Court logically cannot conclude that
DOE acted arbitrarily or capriciously in setting those thresholds.211
APC contests the annual earnings and discretionary income thresholds for another
reason: it says DOE “acted arbitrarily” when it lowered the thresholds for passing those metrics from
207
79 Fed. Reg. at 64,919.
208
Id. at 64,919 & nn.100-03.
209
Id. at 64,919.
210
APCU, 870 F. Supp. 2d at 153.
211
See id. at 153-54 (“The debt to income standards were the product of a ‘rational connection
between the facts found and the choice made.’” (quoting State Farm, 463 U.S. at 43)).
53
12 and 30 percent, respectively, in the 2011 Rules, to eight and 20 percent, respectively, in the current
GE Rules.212
This argument misses the mark for several reasons. First, it is not inherently
problematic for an agency to change its position. Indeed, “[a]n initial agency interpretation is not
instantly carved in stone. On the contrary, the agency . . . must consider varying interpretations and
the wisdom of its policy on a continuing basis, for example, in response to changed factual
circumstances.”213 Indeed, agencies “must be given ample latitude to adapt their rules and policies
to the demands of changing circumstances.”214 Second, what APC terms a “dramatic” change is
really a change in name only. The 2011 Rules never went into effect. They established no “settled
rule” to which institutions were expected to adhere,215 and institutions did not actually rely on them
by committing to a “‘settled course of behavior’” governed by those rules.216 Third, DOE did
provide a “reasoned explanation” for changing the passing thresholds from 12 and 30 percent in the
2011 Rules to eight and 20 percent in the current GE Rules.217 Upon receiving additional data
212
Compare 76 Fed. Reg. at 34,448 (outlining the 12 and 30 percent passing thresholds in the
2011 Rules), with 79 Fed. Reg. at 65,037 (outlining the eight and 20 percent passing
thresholds in the current GE Rules).
213
Brand X Internet Servs., 545 U.S. at 981 (ellipsis in original) (citation and internal quotation
marks omitted); see also id. at 1001 (noting that an agency is “free within the limits of
reasoned interpretation to change course if it adequately justifies the change”).
214
State Farm, 463 U.S. at 42 (internal quotation marks omitted).
215
Id. at 41-42 (internal quotation marks omitted).
216
See FCC v. Fox Television Stations, Inc., 556 U.S. 502, 550 (2009) (Breyer, J., dissenting)
(quoting State Farm, 463 U.S. at 41-42).
217
Brand X Internet Servs., 545 U.S. at 1000-01.
54
showing default and repayment rates for graduates from a wide range of programs—data it obtained
after promulgating the 2011 Rules but before drafting the current GE Rules—DOE determined that
programs whose graduates were averaging payments between 8-12 percent of annual earnings and
20-30 percent of discretionary income “are much more similar to their failing counterparts than their
passing counterparts.”218 These programs, DOE concluded, “may not all be resulting in egregious
levels of debt in relation to earnings,” but “still exhibit poor outcomes and unsustainable debt
levels.”219 Based on this additional data, it was neither arbitrary nor capricious for DOE to conclude
that “the stricter thresholds would more effectively identify poorly performing programs.”220 Nor
was it unreasonable for DOE to justify eliminating the statistical “buffer” included in the 2011 Rules
based on its conclusion that the “three-tier pass, zone, fail construction” in the GE Rules provides
sufficient protection against the possibility that “atypical factors,” such as an unrepresentative cohort
of students or a struggling economy, could doom an otherwise high-performing program.221
APC’s remaining APA arguments fall flat, many for reasons already discussed here
and/or provided by Judge Contreras in his opinion with respect to the 2011 Rules. For example,
APC offers a number of purportedly absurd results as proof that the GE Rules are arbitrary and
capricious. But APC’s challenge is facial, not as-applied, and the fact that APC can “point to a
hypothetical case in which the rule might lead to an arbitrary result does not render the rule
218
79 Fed. Reg. at 64,920.
219
Id.
220
Id.
221
See id.
55
‘arbitrary or capricious.’”222 As Judge Contreras said, APC’s “argument that the regulation may
produce absurd results in certain circumstances is better suited to an as-applied challenge arising out
of such circumstances.”223
APC’s argument that the GE Rules fail to adjust for economic cycles is not just a red
herring—the HEA says nothing about DOE considering fluctuations in the market—it is also untrue.
DOE found that “recessions have, on average, lasted 11.1 months since 1945,” while the GE Rules
give struggling programs multiple years to improve their results before they lose HEA eligibility.224
Moreover, the D/E rates are calculated as means and medians, which DOE found “mitigate the
effects of economic cycles by measuring central tendency and reducing the influence of students
who may have been most impacted by a downturn.”225
Also unpersuasive is APC’s argument that the GE Rules’ use of a single metric (D/E
rates) to calculate Title IV eligibility is arbitrary in light of the 2011 Rules’ use of two metrics (D/E
rates plus the repayment rate that Judge Contreras found was “not based upon any facts at all”).
Several points bear mentioning here. The D/E rates do consist of multiple metrics—one comparing
222
Am. Hosp. Ass’n v. NLRB, 499 U.S. 606, 619 (1991).
223
APCU, 870 F. Supp. 2d at 148.
The fact that, according to APC, “degree programs at most colleges cannot pass the D/E
Rates” is, even if true, of no moment. The relevant provision of the HEA speaks only to
vocational programs; DOE lacked the authority when promulgating the GE Rules “to
regulate other higher education institutions or programs, even if such institutions or
programs would not pass the accountability metrics.” 79 Fed. Reg. at 64,904.
224
79 Fed. Reg. at 64,926; see also APCU, 870 F. Supp. 2d at 151 (“[T[he fact that the debt
measures may perform differently at different points in the economic cycle does not make
them arbitrary on their face.”).
225
79 Fed. Reg. at 64,926.
56
debt to annual earnings and another comparing debt to discretionary income—and a program
“passes” if it satisfies the relevant threshold on either one. Moreover, in considering whether to
replace the voided repayment rate metric—which measured whether all attendees (i.e., not just
graduates) were actually repaying their student loan debts—DOE analyzed a number of options but
ultimately decided that “further study is necessary before we adopt [an] accountability metric that
would take into account the outcomes of students who do not complete a program.”226 If anything,
DOE would have risked violating the APA had it included one of these unproven metrics in the D/E
rates calculus. That it chose not to shows restraint and careful consideration, not arbitrariness or
capriciousness. Finally, while there is no repayment rate metric, or anything like it, in the current
GE Rules, DOE did consider repayment data in formulating the passing thresholds for the D/E
rates.227 In light of all this, the idea that DOE acted arbitrarily in relying on the D/E rates alone to
determine eligibility for Title IV funding purposes simply does not hold water.
The Court concludes, in all the circumstances, that the GE Rules—and the D/E rates
contained therein—were the product of “reasoned decisionmaking.”228
226
Id. at 64,915.
227
See id. at 64,920.
228
State Farm, 463 U.S. at 52.
The Court has already rejected two of APC’s other assertions: that the GE Rules (1)
“conflict with and undermine” the New York Board of Regents’ regulatory scheme, and (2)
improperly exclude non-Title IV students. See supra Part III.B. There is no reason to
repeat that analysis here.
Finally, the Court rejects APC’s argument that the GE Rules violate 20 U.S.C. § 1015c’s
prohibition on “the development, implementation, or maintenance of a . . . database of
personally identifiable information on individuals receiving” federal student aid. The
provisions of section 1015c do “not apply to a system (or a successor system) that (1) is
57
Conclusion
For the foregoing reasons, the Association of Proprietary Colleges’ motion for
summary judgment [DI 22] is denied. The Department of Education’s cross-motion for summary
judgment dismissing the complaint [DI 32] is granted. The Clerk of Court shall enter judgment and
close the case.
SO ORDERED.
Dated:
May 27, 2015
necessary for the operation of programs authorized by [Title IV of the HEA]; and (2) was
in use by the Secretary . . . as of the day before August 14, 2008.” 20 U.S.C. § 1015c(b).
DOE intends to store the relevant information in the National Student Loan Data System,
which has been in use since 1993 and thus fits well within section 1015c(b)’s exception.
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?