Court Description: MEMORANDUM OPINION. See text for details. Signed by Judge John D. Bates on 6/23/15. (lcjdb2)
Download PDF
ASSOCIATION OF PRIVATE SECTOR COLLEGES AND UNIVERSITIES v. DUNCAN et al Doc. 31 UNITED STATES DISTRICT COURT FOR THE DISTRICT OF COLUMBIA ASSOCIATION OF PRIVATE SECTOR COLLEGES & UNIVERSITIES, Plaintiff, v. Civil Action No. 14-1870 (JDB) ARNE DUNCAN, Secretary of the Department of Education, et al., Defendants. MEMORANDUM OPINION This case asks how the government should determine whether certain for-profit and vocational schools “prepare [their] students for gainful employment in a recognized occupation.” 20 U.S.C. §§ 1002(b)(1)(A)(i), 1002(c)(1)(A). The Department of Education proposes one answer: it plans to measure the average debt load of a program’s former students against their earnings. Only graduates who make enough money to service their educational debt, the thinking goes, are “gainful[ly] employ[ed].” But the Association of Private Sector Colleges and Universities approaches the question from a different angle. It believes that the only permissible measure for “gainful employment” is whether students get a job—any job—that pays. The Association thus argues that the Department’s debt-to-earnings test is both outside the scope of the Higher Education Act and arbitrary or capricious under the Administrative Procedure Act. But sustaining these (and other) claims under the deferential Chevron and APA standards presents too great a hurdle for the Association, and the Court will accordingly deny the Association’s motion for summary judgment and grant the Department’s cross-motion. 1 1 See Pl.’s Mot. for Summ. J. [ECF No. 13] (“Pl.’s Mot.”); Gov’t’s Cross-Mot. for Summ. J. & Opp’n to Pl.’s Mot. [ECF No. 17] (“Gov’t’s Cross-Mot”); Pl.’s Opp’n & Reply to Gov’t’s Cross-Mot. [ECF No. 20] (“Pl.’s Opp’n”); 1 BACKGROUND Title IV of the Higher Education Act of 1965 “provides billions of dollars [every year] through loan and grant programs to help students pay . . . for their postsecondary education.” Ass’n of Private Sector Colls. & Univs. v. Duncan, 681 F.3d 427, 433 (D.C. Cir. 2012). But there are boundaries to this billion-dollar industry—including limits on the kinds of schools whose students qualify for federal loans and grants. As this Circuit has explained, “[t]hese [statutory] requirements are intended to ensure that participating schools actually prepare their students for employment, such that those students can repay their loans.” Id. at 434. One such requirement sits at the center of this case. Students enrolled at “[p]roprietary institution[s] of higher education” (that is to say, for-profit schools) or “[p]ostsecondary vocational institution[s]” are only eligible for Title IV funding if the institutions’ programs “prepare [those] students for gainful employment in a recognized occupation.” 20 U.S.C. §§ 1002(b), 1002(c); see also id. § 1088(b)(1)(A)(i). Although Congress did not explain what it meant by “prepare” or “gainful employment” or “recognized occupation,” the Department has grappled with this language once before. After a notice-and-comment period ending in 2011, the Department promulgated a final rule designed “to improve disclosure of relevant information and to establish minimal measures for determining whether certain postsecondary educational programs lead to gainful employment in recognized occupations.” 76 Fed. Reg. 34,386, 34,386 (June 13, 2011). The rule established several tests to measure the adequacy of schools’ preparation efforts, including two different metrics for weighing the average student’s debt load against her income, as well as a separate metric to calculate whether a sufficient percentage of a school’s students were actually repaying their loans. See id. at 34,389. The 2011 rule also required schools “to report the information necessary for the Department to Gov’t’s Reply to Pl.’s Opp’n [ECF No. 24]; see also Chamber of Commerce Amicus Br. [ECF No. 15]; Air Force Sergeants Ass’n, et al. Amicus Br. [ECF No. 19]. 2 calculate [its new metrics] . . . and to disclose to prospective students various facts about [its] program[s].” Ass’n of Private Sector Colls. & Univs. v. Duncan, 870 F. Supp. 2d 133, 143 (D.D.C. 2012) (hereinafter “APSCU I”) (internal citation omitted). This rule never took effect, however. In 2012, the Association of Private Sector Colleges and Universities—a trade group representing 1,400 for-profit colleges which is also the plaintiff here, see Compl. [ECF No. 1] ¶ 2—brought suit against the Department, arguing that the proposed “gainful employment” regulations exceeded statutory authority and violated the APA, see APSCU I, 870 F. Supp. 2d at 144–45. And the district court agreed—at least in part. While the court held that the “gainful employment” provision was ambiguous, that the Department’s interpretation of the provision was reasonable, and that many aspects of the proposed regulations withstood arbitrary-and-capricious review, see id. at 149–53, it nonetheless vacated the bulk of the regulations because one of the Department’s debt metrics did not reflect reasoned decisionmaking, see id. at 154 (“Because the Department has not provided a reasonable explanation of [its debtrepayment] figure, the court must conclude that it was chosen arbitrarily.”). The court likewise rejected some aspects of the Department’s proposed reporting requirements, while retaining the rule’s disclosure requirements. See id. at 155–57; see also Ass’n of Private Sector Colls. & Univs. v. Duncan, 930 F. Supp. 2d 210, 218–21 (D.D.C. 2013) (hereinafter “APSCU II”) (denying a motion to amend APSCU I’s holdings concerning the reporting and disclosure requirements). The Department went back to the drawing board in 2014. It announced a proposed rulemaking in March of that year, see 79 Fed. Reg. 16,426, 16,426 (Mar. 25, 2014), and seven months later published its final rule, see 79 Fed. Reg. 64,890, 64,890 (Oct. 31, 2014). As in the previous go-round with these regulations, the Department concluded that the statute’s “gainful employment” clause was ambiguous. Id. at 64,893. Again, the Department interpreted this 3 language to mean “provid[ing] . . . training that lead[s] to earnings that will allow students to pay back their student loan debts.” Id. at 64,890. And again, the Department issued several regulations designed to give this interpretation some teeth, including (1) a modified regime for measuring the acceptable debt-to-earnings ratios for a program’s students, 2 and (2) revised disclosure, reporting, and certification requirements for programs subject to the regulations. See id. Beginning with the debt-to-earnings test: the 2014 regulations prescribe two metrics for measuring a program’s compliance with the “gainful employment” eligibility restrictions. The first is a debt-to-discretionary-income metric, which the Department calculates by dividing the median annual loan payment for a program’s students by those same students’ discretionary income (which is equal to “the higher of the mean or median annual earnings” of the students minus 150% of the poverty-guidelines figure). 34 C.F.R. § 668.404(a)(1). The second is a debtto-annual-income metric, which the Department calculates using a simpler equation—just dividing the median annual loan payment for a program’s students by the mean or median annual earnings of those students, whichever is greater. See id. § 668.404(a)(2). Two percentage scores thus result from the Department’s debt-to-earnings test, and those two scores determine whether a program is passing or failing—or neither. Specifically, the Department will pass a program if its median annual loan payment is less than or equal to either 20% of discretionary income or 8% of annual earnings. See id. § 668.403(c)(1). In other words, a program need not satisfy both metrics to receive a passing grade—just one will do. Failure, by contrast, requires a failing score under both metrics. A program fails only if its median annual 2 The new “gainful employment” metrics are quite similar to those established in the 2011 final rule. The biggest change between the 2011 regulations and those at issue here: the debt-repayment-rate test—which the APSCU I court determined was arbitrary and capricious—has been abandoned. See, e.g., 79 Fed. Reg. at 64,915 (“[W]e believe further study is necessary before we adopt [a debt-repayment-rate] or another accountability metric that would take into account the outcomes of students who do not complete a program.”). 4 loan payment is both more than 30% of discretionary income and more than 12% of annual earnings. See id. § 668.403(c)(2). Careful readers will notice a gap separating programs that pass from those that fail. But this is by design. A program that does not receive a passing score under either metric, and that has at least one metric that falls in the gap between passing and failing (e.g., median annual loan payments equal to 25% of discretionary income) is considered to be “in the zone”—a label akin to a warning shot. See id. § 668.403(c)(3). Where a program falls on this spectrum—passing, failing, or in the zone—comes with consequences. A program loses eligibility for all Title IV financial aid if it fails the debt-toearnings test for two out of three consecutive years, or if it has debt-to-earnings scores that are in the zone or failing for four consecutive years. See id. § 668.403(c)(4). And once a program is deemed “ineligible” for such federal funding, the program remains in that status for three years. See id. § 668.410(b)(2). Moreover, schools with programs on the verge of an ineligibility determination must affirmatively warn their students (and prospective students) of this fact. See id. § 668.410(a). Among other things, the warning must tell students that they are in danger of losing their federal grants and loans, that they “may have to find other ways . . . to pay for the[ir] program,” and that there might be other similar (and presumably less risky) programs available to them—even at different schools altogether. Id. § 668.410(a)(2). The regulations also include other disclosure, reporting, and certification requirements. For example, programs may be required to disclose to students information like “[t]he length of the program in calendar time,” “[t]he total cost of tuition and fees,” and “[t]he [job] placement rate for” students in the program. Id. § 668.412(a). Programs must also report to the Department various data points concerning each student, including—as relevant to this litigation—“[t]he total amount the student received from private education loans . . . for enrollment in the program,” “[t]he 5 total amount of institutional debt . . . the student owes . . . after completing or withdrawing from the program,” “[t]he total amount of tuition and fees assessed the student,” and “[t]he total amount . . . for books, supplies, and equipment included in the student’s” cost of attendance for each year. Id. § 668.411(a)(2). Finally, the regulations require a school with programs seeking federal student aid to certify that its financial-aid-eligible programs meet various accreditation and licensure standards, including not just federal standards regarding program content and quality, but individual state standards too. See id. § 668.414(d). It should come as no surprise that the Association is no happier with the 2014 version of these regulations than it was with the 2011 version. And it has rehashed its challenge to those earlier regulations here, moving for summary judgment on its claims that the regulations exceed the Department’s statutory authority, are arbitrary or capricious, or otherwise violate federal law. But the Department stands by its latest “gainful employment” regime, and it has filed a competing summary-judgment motion of its own. STANDARD OF REVIEW In the usual case, courts will review such cross-motions for summary judgment with an eye toward finding “genuine dispute[s] as to any material fact[s]” that might make summary judgment inappropriate. Fed. R. Civ. P. 56(a). But this is not the usual case. It is, instead, a case involving review of a final agency action under the APA, which limits the Court’s power of review beyond the Rule 56 standard. Under the APA, it is the agency’s role “to resolve factual issues to arrive at a decision that is supported by the administrative record, whereas the function of the district court is to determine whether or not as a matter of law the evidence in the administrative record permitted the agency to make the decision it did.” Univ. of Mass. v. Kappos, 903 F. Supp. 2d 77, 84 (D.D.C. 2012) (internal quotation marks omitted). Thus, “when a party seeks review of 6 agency action under the APA, the district judge sits as an appellate tribunal. The entire case on review is a question of law.” Am. Biosci., Inc. v. Thompson, 269 F.3d 1077, 1083 (D.C. Cir. 2001) (internal quotation marks and footnote omitted). And that question of law is: based on the record, did the agency act in an “arbitrary” or “capricious” manner or otherwise violate the law? 5 U.S.C. § 706(2); see also Rempfer v. Sharfstein, 583 F.3d 860, 865 (D.C. Cir. 2009). DISCUSSION I. STATUTORY INTERPRETATION Before answering this arbitrary-or-capricious question, the Court must first address a preliminary one—in truth, a series of them. Does “prepare students for gainful employment in a recognized occupation” have a plain meaning that the Department (and the Court) must simply implement? Or is this language ambiguous such that the Court should accept the Department’s interpretation—assuming, of course, that its interpretation is a reasonable one? See Chevron U.S.A. v. Natural Res. Def. Council, Inc., 467 U.S. 837, 842–43 (1984). The Association argues that “gainful employment” can mean only one thing: “a job that pays.” Pl.’s Mot. at 27 (emphasis omitted). The Department, for its part, disagrees, pointing out that both “gainful” and “employment” have multiple dictionary definitions that could give the statute various meanings— including the Department’s preferred meaning: a job that pays enough to cover the student’s educational debt. See Gov’t’s Cross-Mot. at 12. The Department has the better of this dispute, as more than one court has now recognized. See Ass’n of Proprietary Colls. v. Duncan, --- F. Supp. 3d ---, 2015 WL 3404190, at *19 (S.D.N.Y. May 27, 2015); APSCU I, 870 F. Supp. 2d at 146. 3 We start with the first step of the Chevron analysis, which asks “whether Congress has directly spoken to the precise question at issue.” 467 3 Throughout this Memorandum Opinion, this Court will frequently rely on the careful analysis previously provided by Judge Contreras and Judge Kaplan. 7 U.S. at 842. The answer is important: “[i]f the intent of Congress 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.” Id. at 842–43. But that is not the case here. For one thing, Congress opted not to define words like “gainful” in the Higher Education Act. And for another, the usual sources do not establish a single, ordinary meaning for this word. The duel of dictionaries, in fact, ends in stalemate. “Gainful” could mean (as the Association contends) “productive of gain” or “provid[es] an income.” Webster’s Third New Int’l Dictionary 928 (1965). But it could also mean (as the Department counters) “profitable” or “lucrative,” Black’s Law Dictionary 807 (4th ed. 1951)— which, of course, implies that a “gainful” job must not just pay, but instead must pay enough to exceed the job-holder’s expenditures. This back-and-forth is the very definition of “ambiguous.” See, e.g., Black’s Law Dictionary 79 (7th ed. 1999) (“An uncertainty of meaning or intention, as in a . . . statutory provision.”). Context only muddies the waters. Consider the word “employment,” which Congress also left undefined. Dictionaries explain that “employment” could mean “to make use of” or “to occupy . . . advantageously.” Webster’s New Collegiate Dictionary 373 (1975) (“employ”). These definitions suggest that payment is an unnecessary part of the “employment” concept, and thus that Congress might have used the “gainful” qualifier to distinguish those jobs that pay from those that do not. But “employment” can also mean “the act of hiring, implying a request and a contract for compensation,” Black’s Law Dictionary 618 (4th ed. 1951) (emphasis added), or “a job that pays wages or a salary,” Webster’s New Collegiate Dictionary 373 (emphasis added). If Congress adopted these definitions of “employment”—where payment is baked into the word—then “gainful” must mean something more than simply “paying.” After all, reading “gainful employment” as “a paying job that pays” is hardly consistent with the command that courts should 8 avoid interpreting statutes in ways that “render[] some words altogether redundant.” Gustafson v. Alloyd Co., 513 U.S. 561, 574 (1995). The point is: these words, read together, could reasonably mean different things; hence, they are ambiguous. One more ambiguity puts the first Chevron question to bed. However the Department prefers to mix-and-match definitions of “gainful” and “employment,” it is beyond dispute that the statute is silent regarding how, exactly, the Department should assess whether a program “prepare[s] students for” such employment. It must establish some kind of test. And perhaps, as the Association believes, the Department could assess higher-education programs through a simple yes/no exam: does the average former student have a paying job? If so, the program qualifies; if not, then it does not. But the statute does not “unambiguously foreclose[]” all other tests. Nat’l Cable & Telecomms. Ass’n v. FCC, 567 F.3d 659, 663 (D.C Cir. 2009). The statute, instead, leaves a policy gap, which it is the Department’s prerogative to fill, see Chevron, 467 U.S. at 843– 44—possibly through a binary, yes/no test, but just as possibly through more nuanced metrics that assess what jobs students land, how much those jobs pay, etc. For this reason, too, the Court concludes that the “gainful employment” provision—read in its entirety—is ambiguous. See, e.g., New York v. EPA, 413 F.3d 3, 23 (D.C. Cir. 2005) (“[T]he statute is silent or ambiguous on how to calculate baseline emissions . . . .” (emphasis added)). The Department’s reading of this provision also navigates Chevron’s second step. Under this step, “a court must defer to the agency’s interpretation of the ambiguous statutory term if it represents a reasonable accommodation of conflicting policies that were committed to the agency’s care by the statute.” Id. (internal quotation marks omitted). That is the case here. As explained, the Department’s interpretation is rooted in the text: “employment” can mean a paying job, and “gainful” can specify which of those paying jobs meet the mark (e.g., only those that pay enough 9 to be “profitable” or “lucrative”). But that only gets the Department halfway to a useful regulation. It still must decide how to measure “profit.” And the two-part debt-to-earnings test is one reasonable way to do just that—especially when the experts (Congress, the Department, etc.) are “concern[ed] that [the government’s] . . . resources w[ill] be used to provide financial aid to students who [are] unable to find jobs that would allow them to repay their loans.” 79 Fed. Reg. at 64,894. The Court is in no position to second-guess this reasonable policy choice. See Chevron, 467 U.S. at 844 (“[A] court may not substitute its own construction of a statutory provision for a reasonable interpretation made by the administrator of an agency.”). The Association sees things differently, but its arguments are not persuasive. It first notes that “gainful employment” is used elsewhere in other provisions of the Higher Education Act, and that the Department’s preferred reading of these words would not make sense in those contexts. See, e.g., Pl.’s Mot. at 28 n.16 (citing, for example, 20 U.S.C. § 1036(e)). But it is axiomatic that a term “may have a plain meaning in the context of a particular section” of a statute without having “the same meaning in all other sections and in all other contexts.” Robinson v. Shell Oil Co., 519 U.S. 337, 343 (1997). Robinson illustrates the point. In that case, the defendant argued that because the word “employee” meant “current employee” in sections of Title VII dealing with salary and promotions, “employee” must carry this single meaning throughout Title VII. Id. But the Supreme Court rejected that argument, reiterating, instead, that context matters. Perhaps “employee” means something different when talking about salary and promotions than it does when discussing reinstatement or hiring, the Court noted. See id. at 342–44. Or perhaps Congress included qualifying words near “employee” in some sections but not in others, which could cast the same or similar words in a different light. See id. at 344–45. 10 Both of Robinson’s contextual points apply here. Let’s assume that “gainful employment” unambiguously means “any job that pays” in a section of the Higher Education Act that tells schools they cannot award federally funded fellowships to individuals who already have other jobs. See 20 U.S.C. § 1036(e)(1)(B)(ii). Must those words mean the same thing in a different section of the Act that applies not to fellowships, but to for-profit and vocational schools that send students to a wide range of jobs? Not necessarily. And this conclusion is doubly warranted here, where the words “gainful employment” do not stand alone. The complete phrase is “prepare students for gainful employment in a recognized occupation.” 20 U.S.C. §§ 1002(b)(1)(A)(i), 1002(c)(1)(A) (emphasis added). But the reader will search in vain for these qualifying words in the Association’s cited provisions from elsewhere in the Act. See id. §§ 1036(e)(1)(B)(ii), 1034c(a), 1035c(d)(2), 1140(1)(B), 1140g(d)(3)(D), 1161g(d)(5)(B), 2008(a), 4706(a), 5605(a)(2)(B). Just as in Robinson, then, the “gainful employment” provision at the heart of this case remains ambiguous—whatever the rest of the Act might say. See APSCU I, 870 F. Supp. 2d at 146; cf. Sosa v. Alvarez-Machain, 542 U.S. 692, 711 n.9 (2004) (invoking “the usual rule that when the legislature uses certain language in one part of the statute and different language in another, the court assumes different meanings were intended” (internal quotation marks omitted)). The Association also appeals to other provisions of the Higher Education Act to argue that the Department’s reading of “gainful employment” irreconcilably conflicts with the larger structure and purpose of the Act. See Pl.’s Mot. at 31–34. But there is no conflict here, irreconcilable or otherwise. It may be true, for instance, that the Act was designed to “expand access to higher education,” id. at 31—but the Act does not countenance expansion at all costs. After all, it includes several explicit limits on the financial-aid eligibility of schools, including the provision at issue here. See, e.g., 20 U.S.C. § 1002(b)(1)(A)(i) (“gainful employment” provision); 11 see also id. § 1094(a)(24) (denying financial aid to for-profit schools that derive more than 90% of their revenue from Title IV funds). Other supposed “conflicts” are likewise not conflicts at all. For example, just because a provision of the Higher Education Act establishes debt metrics at the institution-level, see id. § 1085(m)(1), does not mean the Department lacks the power to establish separate program-level debt metrics. Indeed, a “program” level review is exactly what the “gainful employment” provision requires. See id. § 1002(b)(1)(A)(i) (asking whether schools “provide[] an eligible program of training to prepare students for gainful employment in a recognized occupation” (emphasis added)). And just because the Act empowers accrediting bodies to assess the inputs (for lack of a better word) provided to students at programs subject to the “gainful employment” provision, see id. § 1099b(a) (allowing accreditors to review “the quality of education or training offered”), does not forbid the Department from examining those programs’ outputs in terms of student earnings and debts. 4 Far from conflicting with the Higher Education Act, in short, the Department’s reading of the “gainful employment” provision actually complements it. 5 4 To put this into (slightly) more familiar terms: imagine that Arsenal, your favorite English soccer team, has just lost a match to one of its London rivals, Chelsea. And it was not just any loss. Arsenal was down four-nil by halftime and finished the match six goals behind. How should one assess the squad’s efforts for that game? One way, of course, is to look at the pre-match training, scouting, and fitness efforts of the team. These “inputs,” so to speak, might tell something about Arsenal’s quality. But they might not tell the whole story, and they are not easily accessible. As a fan (and not, let’s say, Arsène Wenger, Arsenal’s manager), you might not get to see the day-in, dayout work put in on Arsenal’s training pitch. But you can more easily judge the team’s efforts based on the team’s “outputs,” that is to say, its performance on the field. On the day, the team was terrible—it gave up an early goal and then unraveled, and the defense was generally disorganized. So it stands to reason that the squad’s preparation for this particular game was not quite what it needed to be. There is no irreconcilable conflict between these input/output perspectives: you, as a fan, are not usurping Wenger’s role as manager by reaching your own conclusion; and he, as the coach, is not forbidding fans from adopting any perspective but his own. The fans and Wenger, in other words, can exist together, with somewhat distinct means of assessing the team’s efforts. 5 One other supposed statutory conflict warrants mention. The Association notes that the Department is “barred . . . from interfering with school administration” and cannot “dictate [a school’s] tuition and fees.” Pl.’s Mot. at 20. But nothing in the 2014 regulations oversteps these bounds. The regulations say nothing about what schools must teach or how much money they can charge. They merely measure student outcomes based on a school’s independent decisions in these areas. Of course, schools may decide to “lower[] tuition and fees” in response to these regulations. 79 Fed. Reg. at 64,924. But they might also do other things: invest in job-placement services, change how they market programs to prospective students, ensure that students receive the necessary licenses for a post- 12 What about In re Academy for Jewish Education? 1994 WL 1026087, at *1 (Dep’t of Educ. Mar. 23, 1994). The Association leans on that case to argue that the Department has previously interpreted the “gainful employment” provision “without any consideration of students’ earnings or debt”—which may call into question a definition containing these elements now. Pl.’s Opp’n at 20. But this administrative adjudication cannot bear the weight the Association assigns it. There, the program in question failed to meet the definition of “gainful employment in a recognized occupation” because it was undisputed that “the training offered [was] not directed toward a specific occupation.” In re Acad. of Jewish Educ., 1994 WL 1026087, at *2. The program, in other words, failed at the threshold—it left students unprepared for any job, let alone a job that was “gainful” or “in a recognized occupation.” Given this failure, the Department had little reason to settle on a more nuanced definition for the full “gainful employment” provision— and (so far as the record reflects) it offered none. See id. (“It is difficult to objectively assess what, per se, prepares one for ‘gainful employment in a recognized occupation.’”). Without any prior definition to hold over the Department’s head, then, the Association’s appeal on this front comes to naught. See APSCU I, 870 F. Supp. 2d at 149–50; see also Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 515–16 (1994) (“Petitioner’s attempt to infer from that silence the existence of a contrary policy fails because the [agency’s] letter did not purport to be a comprehensive review of all conditions that might be placed on reimbursement of educational costs.”). 6 graduation job, etc. See id. at 65,032. None of this is “dictate[d]” by the Department’s interpretation of the “gainful employment” provision. Rather, schools get to decide—which is just as the Higher Education Act would have it. 6 A similar conclusion torpedoes another Association argument: that the Department’s preferred definition of “gainful employment in a recognized occupation” conflicts with other regulations using similar words. See Pl.’s Mot. at 29. The Association points to a regulation that requires certain schools to document the placement of individual students in “gainful employment” positions by providing things like tax returns or letters from the student’s employer. See 34 C.F.R. § 668.8(g)(2). But this does not close the book on the Department’s more nuanced definition here for at least two reasons. First, the cited regulation does not explicitly define the words that matter to this case, which makes it difficult to find any material “conflict” worthy of discussion. See id. § 668.8(b) (defining several words “[f]or purposes of this section” without including “gainful employment” in the list); see also Thomas Jefferson Univ., 512 U.S. at 515 (“While it is true that an agency’s interpretation of a statute or regulation that conflicts with a 13 The Association persists that the legislative history undermines the Department’s preferred definition. See Pl.’s Mot. at 34–35. It relies on, for example, the testimony of the “principal author” of the National Vocational Student Loan Insurance Act (a law that works in concert with the Higher Education Act), who explained that he “d[id] not intend that the Department . . . should . . . build barriers that will keep . . . students from attending the so-called privately owned schools if it is their desire to do so.” Pl.’s Mot. at 35 (internal quotation marks and citation omitted). As is so often the case with these sorts of arguments, however, the history of the Higher Education Act is crowded with evidence (and snippets of evidence), and each side can find its “friends.” Exxon Mobil Corp. v. Allapattah Servs., Inc., 545 U.S. 546, 568 (2005) (internal quotation marks omitted). Hence, the Department relies on the testimony of a Senate committee witness, who “posed two critical questions . . . : ‘If loans were made to these kinds of students, is it likely they could repay them following training? [And w]ould loan funds pay dividends in terms of benefits accruing from the training students received? It would seem that any discussion concerning this bill must address itself to these questions.’” Gov’t’s Cross-Mot. at 27 (quoting S. Rep. No. 89758, at 7). The legislative history thus only underscores the ambiguity of the “gainful employment” provision—it does not cure it. The Association next retreats to its back-up position: maybe the provision is ambiguous, but even assuming this is so, the Department’s debt-to-earnings test is outside the scope of that prior interpretation is entitled to considerably less deference than a consistently held agency view, that maxim does not apply here because petitioner fails to present persuasive evidence that the Secretary has interpreted the . . . provision in an inconsistent manner.” (internal quotation marks and citation omitted)). And second, even assuming some conflict in the abstract, the cited regulation is not on point. It explains how schools must calculate their students’ job-placement rates for purposes of a separate statutory requirement, see 20 U.S.C. § 1088(e)(2)(a); it does not speak to how the Department may measure whether schools “prepare students for gainful employment in a recognized occupation,” id. §§ 1002(b)(1)(A)(i), 1002(c)(1)(A) (emphasis added); see also APSCU I, 870 F. Supp. 2d at 150 (“Nor is the Department’s regulation requiring schools to report whether individual students have attained gainful employment counter to its attempt to use the debt measures to identify programs that offer adequate preparation for such employment.” (emphases in original)). 14 ambiguity. See Pl.’s Opp’n at 17–18. “‘It does not matter,’” the Association argues, “‘whether the word yellow is ambiguous when the agency has interpreted it to mean purple.’” Id. at 18 (quoting United States v. Home Concrete & Supply, LLC, 132 S. Ct. 1836, 1846 n.1 (2012) (Scalia, J., concurring in part and concurring in the judgment)) (some internal quotation marks omitted). But the Department did nothing of the sort. Rather, it determined that “gainful employment in a recognized occupation” means something like “profitable employment.” And then it reached the sensible conclusion that it should test the profitability of students’ employment by asking whether students earn enough to pay their bills. See 79 Fed. Reg. at 64,894. This is hardly transforming yellow into purple. It is, instead, like using a swatch to decide how yellow is yellow enough for your living room wall—because sometimes you just want a wall that’s “yellow,” not, let’s say, amber, maize, goldenrod, or (heavens no) chartreuse. In this sense, the Department’s test adds a touch of clarity to an ambiguous statutory command; it is not outside the scope of that ambiguity. In sum, then, the text, context, and legislative history of the Higher Education Act point in several directions, and the Association has offered no compelling reason why this Court’s answers to the Chevron step one (ambiguous?) and step two (reasonable?) inquiries should be any different from those given in other cases before other courts. This Court therefore responds in the same way as those others: the “gainful employment” provision is ambiguous, and the Department’s interpretation of that provision is reasonable. See Ass’n of Proprietary Colls., 2015 WL 3404190, at *19; APSCU I, 870 F. Supp. 2d at 146. II. APA ANALYSIS OF THE DEBT-TO-EARNINGS TEST That conclusion only scratches the surface of this case, however. And so the Court will dig further. The Association maintains that even if the Department has reasonably interpreted the ambiguous “gainful employment” command, it still arrived at that interpretation in an arbitrary 15 and capricious way. But the Association “carries a heavy burden indeed” regarding this claim. Wisc. Valley Improvement v. FERC, 236 F.3d 738, 745 (D.C. Cir. 2001) (internal quotation marks omitted). While “an agency must articulate a satisfactory explanation for its action[,] including a rational connection between the facts found and the choice made,” Owner-Operator Indep. Drivers Ass’n v. Fed. Motor Carrier Safety Admin., 494 F.3d 188, 203 (D.C. Cir. 2007) (internal quotation marks omitted), courts will “presume[] agency action to be valid,” Am. Wildlands v. Kempthorne, 530 F.3d 991, 997 (D.C. Cir. 2008) (internal quotation marks omitted). Indeed, courts will not set aside such action unless the agency, for instance, “relied on factors 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.” Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983). The Association cannot meet their “heavy burden” here. Far from acting arbitrarily or capriciously, the record shows that the Department engaged in a thorough rulemaking process before promulgating its debt-to-earnings regulations. It announced a proposed rulemaking in March 2014, see 79 Fed. Reg. at 16,426, and that announcement prompted “approximately 95,000 comments on the proposed regulations,” id. at 64,892. The Department fastidiously reviewed these comments, making changes to the proposed regulations based on some comments, and explaining its rationale for rejecting others. See id. (“The final regulations contain a number of changes from the [National Proposed Rulemaking].”). And at the other end of this process came a final rule issued in October 2014 that spans fully 213 pages of the Federal Register, that “cogently explain[s] why [the Department] has exercised its discretion in a given manner,” and that “enable[s the Court] to conclude that the [Department’s] action was the product of reasoned decisionmaking.” Owner- 16 Operator, 494 F.3d at 203 (internal quotation marks and citation omitted). The APA requires nothing more. See Ass’n of Proprietary Colls., 2015 WL 3404190, at *23 (upholding the 2014 debt-to-earnings test against a similar arbitrary-and-capricious challenge); see also APSCU I, 870 F. Supp. 2d at 149–54 (same for the 2011 debt-to-earnings test). This is not the result the Association wants to hear, however, and it throws a host of arbitrary-or-capricious arguments against the wall in hope of a different outcome. None of them stick. Its first toss: the Department failed to “offer[] a rational explanation of how debt and earnings metrics measure whether programs prepare students for gainful employment.” Pl.’s Mot. at 36. But this is plainly incorrect. If—as the Court has already held—the Department’s interpretation of “gainful employment” (i.e., a job that pays enough to cover the cost of student loans) is reasonable, then balancing former students’ debt against their earnings is a rational mechanism for assessing whether a program has, in fact, prepared those students for such employment. Moreover, the final rule is replete with explanations for the Department’s chosen metric. The rule explains that Congress was particularly concerned “that vocational and career training . . . should equip students to earn enough to repay their loans,” 79 Fed. Reg. at 64,913, and then describes how comparing debt to earnings can accurately predict students’ ability to repay these loans, see id. at 64,920–21. Of course, the Association might not agree with the Department’s explanations. But that alone does not make them irrational, arbitrary, or capricious. See Owner-Operator, 494 F.3d at 211 (“[Plaintiff] may disagree with this policy balance, but it does not reflect a failure to consider relevant factors.”). The Association’s second argument fares no better. It contends that the debt-to-earnings test must be arbitrary and capricious, because the test’s percentage scores turn on factors outside a school’s control, including student employment choices, external market forces, and the amount 17 of debt a student chooses to take on. See Pl.’s Mot. at 37–39. But several problems plague this argument—including its premise. As the final rule recognized, schools have some say in many of these (supposedly) “outside” factors, including how they market their programs, how much they charge for their programs, and how diligent they are about satisfying licensing and accreditation requirements for their programs. See 79 Fed. Reg. at 65,032 (explaining that such factors affect default rates). Additionally, the Department designed the debt-to-earnings regulations to account for—and mitigate—these externalities. For example, regarding the possibility of “unanticipated fluctuations in local labor market conditions,” the Department explained that the regulations’ “in the zone” category protects schools from such conditions, because it allows programs “to remain eligible [for financial aid] for up to four years”—which is long enough to survive “[m]ost economic downturns” that tend to last around eleven months. Id. at 64,926. Other aspects of the test offer additional protections. See, e.g., id. (“[M]eans and medians 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.”); id. at 64,918 (“[L]oan debt [is] capped [in the debt-toearnings test] at the amount charged for tuition and fees and books, supplies, and equipment, because those costs are within an institution’s control.”). Thus, the Department has already heard the Association’s complaints on this score, and the “court is ill-equipped to second-guess [the agency’s] . . . judgment under the guise of the APA’s arbitrary and capricious standard.” Cytori Therapeutics, Inc. v. FDA, 715 F.3d 922, 927 (D.C. Cir. 2013). Third, the Association protests that the debt-to-earnings test does not measure a program’s preparation efforts at all. Instead, it believes the test “really measure[s] student demographics,” including the racial and socio-economic makeup of a program’s student body. Pl.’s Mot. at 39. But again, the Department expressly considered this criticism, and it explained its rationale for 18 rejecting it. Indeed, in response to comments during the rulemaking process, the Department conducted “several regression analyses examining the relationship between demographic factors and program results under the [debt-to-earnings] rates.” 79 Fed. Reg. at 65,039; see also id. at 65,043 (“[T]he Department conducted additional analysis for the final regulations.”). After explaining its methodology and publishing the results, the Department summarized its findings: [T]he Department cannot conclude . . . that demographic characteristics are largely determinative of results . . . . Instead, we find a negative association between the proportion of low-income students and the annual earnings rate when controlling for other demographic and non-demographic factors, similar passing rates across all quartiles of low-income variables, and similar demographic profiles in passing, zone, and failing programs for almost all of the variables examined. These and other results of our analyses suggest that the regulation is not primarily measuring student demographics. Id. at 65,057. The Department therefore made extensive efforts to get to the bottom of this criticism, and this Court cannot fairly say that the agency acted arbitrarily in the face of it. See Ass’n of Proprietary Colls., 2015 WL 3404190, at *21 (“[The Association of Proprietary College’s] argument on this point appears utterly to disregard the extensive statistical analyses underlying the [“gainful employment” regulations].”); APSCU I, 870 F. Supp. 2d at 151 (similar). The Associations’ fourth argument also fails to persuade. It challenges the Department’s decision to apply “[a] [s]ingle [d]ebt [t]est” to measure “gainful employment”—which (the Association believes) is a reversal of the Department’s prior position. Pl.’s Mot. at 40. Yet this argument rests on shaky ground too. Although the APSCU I court noted that the Department’s debt-to-earnings test was “designed to work together” with its separate (but now abandoned) debtrepayment test, 870 F. Supp. 2d at 154, the Department never conceded that the debt-to-earnings test could not operate on its own. Instead, it said that “no single percentage . . . answers the question of how much students can borrow without risking repayment difficulties,” Gov’t’s Reply in APSCU I, 11-1314 [ECF No. 20] at 18 (emphasis added)—a statement wholly consistent with 19 its current debt-to-earnings test, which calculates two distinct percentages for students’ average debt load: a debt-to-annual earnings percentage and a debt-to-disposable income percentage. 7 What is more, the decision to drop the debt-repayment test was not an arbitrary one. It came, indeed, in response to commenters who argued that “further study [was] necessary before [the Department] adopt[s] [a debt-repayment test] or another accountability metric.” 79 Fed. Reg. at 64,915. Abandoning an aspect of a rule that commenters object to (and that a federal court has called arbitrary, see APSCU I, 870 F. Supp. 2d at 154) is not, itself, arbitrary. See Ass’n of Proprietary Colls., 2015 WL 3404190, at *23 (“If anything, [the Department] would have risked violating the APA had it included [the debt-repayment test in its] . . . calculus. That it chose not to shows restraint and careful consideration, not arbitrariness or capriciousness.”). 8 7 The Association notes that the 2011 regulations also used two different metrics in the debt-to-earnings test, and that the APSCU I court nevertheless found that the (rational) debt-to-earnings test and the (arbitrary) debtrepayment test were so “intertwined” that both had to be vacated due to the shortcomings of the latter. See Pl.’s Opp’n at 36 (citing APSCU I, 870 F. Supp. 2d at 154). But this does not change anything. The Association can point to nothing in the record—neither the record for the 2014 final rule, nor the record for the 2011 rule—that establishes a formal Department position that the debt-to-earnings test (with its two percentages) could not work on its own. Indeed, the record from the 2011 litigation suggests that the Department believed the opposite to be true. See Gov’t’s Reply in APSCU I at 18–19 (explaining that while “no single percentage [can] answer[] the question of how much students can borrow without risking repayment difficulties[,] *** the repayment test is just one of three debt measures; a program that fails this test can remain eligible by passing one of the other two tests [namely, the two debt-to-earnings metrics]” (emphasis added and internal quotation marks omitted)). Moreover, the APSCU I court does not get the final word on whether the debt-repayment test is severable from the rest of the rule. The point of the judiciary’s severability doctrine is to ensure “that agency policy is . . . made, in the first instance, by the agency itself—not by the courts, and not by agency counsel.” Harmon v. Thornburgh, 878 F.2d 484, 494 (D.C. Cir. 1989). It therefore makes good sense to vacate entire rules and remand them to the agency whenever there is “substantial doubt that a partial affirmance would comport with the [agency’s] intent.” Tel. & Data Sys., Inc. v. FCC, 19 F.3d 42, 50 (D.C. Cir. 1994). But the Department has now reconsidered its “gainful employment” rule, and it has determined—consistent with its prior litigation position—that two different metrics is enough. See Gov’t’s Reply in APSCU I at 18. This ought to calm whatever doubt existed at the time of APSCU I. 8 The Association hints at one other reversal-of-position argument. It contends that—for some fifty years— the Department’s position has been that “gainful employment” means something akin to “any job that pays.” See, e.g., Pl.’s Opp’n at 20–21. But—aside from the 2011 regulations, which were largely vacated after APSCU I—the Department has never issued a regulation interpreting the “gainful employment in a recognized occupation” language. It is therefore difficult to see how the Department has “reversed” or “changed” its reading of this provision. See APSCU I, 870 F. Supp. 2d at 149–50 (rejecting a similar argument). Moreover, just because the Department had not previously tried to interpret this provision, does not mean that it lacks the power to do so. “Authority actually granted by Congress . . . cannot evaporate through lack of administrative exercise.” FTC v. Bunte Bros., 312 U.S. 349, 352 (1941). This secondary argument therefore does little to advance the Association’s position. 20 Fifth, the Association labels as arbitrary the Department’s decision to rely on debt-toearnings percentages derived from the reported bills and incomes of recent graduates, because “[h]igher education is a lifelong investment that yields benefits not fully realized within the first few years after graduation.” Pl.’s Mot. at 41. “[L]ifelong” benefits, however, might not do much good when the bill collectors come calling today. And the Department realized exactly that when crafting its final rule. As the rule explains, “[b]orrowers are still responsible for managing debt payments, which begin shortly after they complete a program, even in the early stages of their career.” 79 Fed. Reg. at 64,914; see also id. (“[B]enefits ultimately available over a lifetime may not accrue soon enough to enable the individual to repay [her] student loan debt.”). This truth cuts to the heart of the problem: when graduates get low-paying jobs and then default on their student loans, nobody wins—not the government (which picks up the tab), and not the student (who may get back on her feet eventually, but who—in the meantime—may be denied credit, miss bill due dates, or even file for bankruptcy). See id. The Department’s train of thought therefore exhibits the hallmarks of rational—not arbitrary or capricious—decision-making. See, e.g., Ass’n of Proprietary Colls., 2015 WL 3404190, at *21. The same can be said in response to the Association’s sixth argument, which concerns the timeline by which the Department assumes students will typically repay their loans. See Pl.’s Mot. at 42–43. Under the rule, the Department calculates the annual loan payment for students based on the credential level they receive from a program (i.e., the Department calculates debt payments based on a ten-year amortization period for students receiving certificates, diplomas, and associate’s degrees, but it calculates those payments based on a fifteen- or twenty-year period for students receiving more advanced degrees). See 79 Fed. Reg. at 64,939. These periods survive scrutiny under the APA. “Agencies generally do not violate the . . . arbitrary-and-capricious 21 standard when they employ bright-line rules . . . , so long as those rules fall within a zone of reasonableness and are reasonably explained.” Emily’s List v. Fed. Election Comm’n, 581 F.3d 1, 22 n.20 (D.C. Cir. 2009). And the Department’s rules do (and are) just that. The “gainful employment” rule references over thirty years of repayment data, which suggests that a majority9 of student borrowers from two-year institutions are able to fully repay their loans within ten years, while borrowers from four-year institutions and graduate schools tend to need more time. See 79 Fed. Reg. at 64,939–40. Drawing amortization lines based on decades’ worth of data is reasonable. After all, “the line[s] had to be drawn somewhere,” and it is not this Court’s province—or the Association’s, for that matter—to “redraw the line[s] according to [its] own notions of what might be best.” Process Gas Consumers Grp. v. FERC, 712 F.2d 483, 488 (D.C. Cir. 1983). Seventh, the Association considers the Department’s passing and failing debt thresholds to be arbitrary, and it takes specific issue with the 8% threshold attached to the Department’s debtto-annual-earnings metric. See Pl.’s Mot. at 43–45. As the Association sees things, this number is inapt, because it is based on the acceptable debt load for individuals who also have a mortgage— something most recent graduates do not have to worry about. See id. at 43. But the Department is hardly alone in adopting the 8% figure as the tipping point for unbearably high student debt. As the final rule recognized, “the 8 percent cutoff has long been referred to as a limit for student debt burden. Several studies of student debt have accepted the 8 percent standard. [And] [s]ome State agencies have established guidelines based on this limit.” 79 Fed. Reg. at 64,919 (four footnotes citing studies omitted). Moreover, even experts who are critical of the 8% cutoff “acknowledge the widespread acceptance of [that figure] and conclude that . . . it is ‘not . . . unreasonable.’” Id. 9 One study showed that 58% of students at two-year institutions met the ten-year repayment timeline, while another study showed that 55% of such students met the same mark. See 79 Fed. Reg. at 64,939. Under either study, then, a majority of students needed ten years or less to repay their student loans. 22 (quoting Sandy Baum & Saul Schwartz, How Much Debt is Too Much? Defining Benchmarks for Managing Student Debt 3 (2006)). The Association’s arguments on this score therefore present, at best, “a battle of experts—a battle conducted in an arena that is off limits to APA judicial review.” Envtl. Def. v. U.S. Army Corps of Eng’rs., 515 F. Supp. 2d 69, 82 (D.D.C. 2007). 10 The Association’s eighth argument fares no better. It sees the underlying data used in the Department’s calculations as “[u]nreliable” and “[s]kewed.” Pl.’s Mot. at 45. Specifically, the Association complains that the Department’s data—borrowed from the Social Security Administration’s Master Earnings File—“provides a severely distorted (and understated) picture of earnings,” because some data is allegedly missing or inaccurate. Id. at 45, 46. Even accepting that there are problems with the Social Security data, however, the APA does not “demand the perfect at the expense of the achievable.” Am. Pub. Gas Ass’n v. Fed. Power Comm’n, 567 F.2d 1016, 1046 (D.C. Cir. 1977) (internal quotation marks omitted). Instead, “the accuracy of any particular [data] . . . must be evaluated by reference to the data that was available to the agency at the relevant time.” Baystate Med. Ctr. v. Leavitt, 545 F. Supp. 2d 20, 41 (D.D.C. 2008). Agencies, in other words, must use “the best data available,” id. at 44, and the Department met that standard here. It used the Social Security data—but only after determining that no better data existed, see 79 Fed. Reg. at 64,956; only after rejecting other possible sources of data as inadequate, see id. at 10 The Association raises one other challenge to the 8% threshold. It suggests that the Department has not adequately explained its downward shift in the passing threshold from 12% (under the old regime) to 8% (under the current one). See Pl.’s Mot. at 44–45. The Association calls this “abandon[ment]” of the 12% threshold arbitrary. But the final rule negates this argument as well. As the rule explains, the Department analyzed data obtained “after the 2011 Prior Rule,” which showed that programs exceeding the 8% threshold “still exhibit poor outcomes and unsustainable debt levels.” 79 Fed. Reg. at 64,920. In other words, the data showed that certain programs that would have passed under the old rule (e.g., programs whose students averaged a 10% debt-to-annual-earnings rate) were much more reminiscent of failing programs than passing ones. The Department has therefore sufficiently explained the rationale for its changed position in the face of new facts. This is not arbitrary. See Nat’l Cable & Telecomms. Ass’n v. Brand X Internet Servs., 545 U.S. 967, 981 (2005) (“An 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, or a change in administrations.” (internal quotation marks and citation omitted)). Additionally, the Department has not “abandon[ed]” the 12% figure. Instead, it uses that figure as the new threshold for “in the zone” programs. 23 64,941–42 (describing problems with alternative data from the Bureau of Labor Statistics); and only after answering commenters’ concerns regarding the data, see, e.g., id. at 64,950–59. Once again, then, the Department’s treatment of the Associations’ criticism demonstrates rationality— not arbitrariness. See Mt. Diablo Hosp. v. Shalala, 3 F.3d 1226, 1233 (9th Cir. 1993) (“The agency simply chose one imperfect database over another while seeking to develop data superior to either. This choice was rational.”). Ninth, the Association raises the specter of absurd results. See Pl.’s Mot. at 46–47. It hypothesizes, for example, that one Colorado medical school “likely will be in in the ‘zone’ or failing under the new rule” despite its “100% residency-placement rate.” Id. at 47 n.56. But this hypothetical does not render the “gainful employment” rule arbitrary and capricious. For one thing, this result remains just a specter. There has been no determination of any program’s debtto-earnings scores at this point—the regulations do not even go into effect until July 2015. Such hypothetical examples are not the stuff of ripe APA challenges. For another thing, a solitary hypothetical is irrelevant to facial challenges like this one. “To prevail in . . . a facial challenge, [a plaintiff] must establish that no set of circumstances exists under which the [regulations] would be valid.” Ass’n of Private Sector Colls. & Univs., 681 F.3d at 442 (internal quotation marks omitted). That is to say, “[t]he fact that [the Association] can point to a hypothetical case in which the rule might lead to an arbitrary result does not render the rule ‘arbitrary or capricious[,]’ [because] [t]his case is a challenge to the validity of the entire rule in all its applications.” Am. Hosp. Ass’n v. NLRB, 499 U.S. 606, 619 (1991) (emphasis added). The Association’s hypotheticals (however based-on-a-true-story they may be) are therefore unconvincing. The same can be said of the Associations’ tenth challenge. The Association detects arbitrariness in the final rule, because (it believes) the Department failed to address “[t]he immense 24 . . . practical consequences of [its] debt-to-earnings test,” including the (it says) enormous negative impact on “disadvantaged students” like those “who are older, women, Black, Hispanic, . . . with low incomes . . . [and] single parents.” Pl.’s Mot. at 47–48 (internal quotation marks omitted). But the final rule belies this claim. To start, the Department’s studies “show[ed] that passing, zone, and failing programs have very similar proportions of low-income, non-traditional, female, white, Black, and Hispanic students,” so there is no basis for the claim—on this record, at least— that any particular group of students will suffer special harm under these regulations. 79 Fed. Reg. at 65,045. And in any event, the Department gave all these alleged harms their due, producing an eighty-page Regulatory Impact Analysis that weighed the positive and negative implications of its debt-to-earnings test. See 79 Fed. Reg. at 65,024–103. The result? “[T]he benefits [of the regulations] justify the costs.” Id. at 64,993. The Court will not “quibble” with this “policy choice[]”; it will not, put differently, “substitute [its] judgment for that of the agency.” FCC v. Fox Television Stations, Inc., 556 U.S. 502, 530 (2009) (internal quotation marks omitted). Eleventh, the Association argues that the “gainful employment” rule is arbitrary because it retroactively sanctions programs for their past behavior. See Pl.’s Mot. at 50 (“[T]he Department’s debt-to-earnings test calculates the earnings of students who graduated before the regulations were finalized.”). But legally speaking (which, of course, is all that matters), this is not a retroactive rule. “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.” Reynolds v. United States, 292 U.S. 443, 449 (1934). That is to say, so long as the Department’s regulations do not alter “the past legal consequences of past actions,” they are not improperly retroactive. Nat’l Cable, 567 F.3d at 670 (emphasis added) (internal quotation marks omitted). And the debt-to-earnings test does no such thing. It does not, for example, retroactively take away 25 failing programs’ already-granted Title IV financial aid; rather, the rule affects only future eligibility for that aid (which at least initially will depend on past facts). Legally prospective regulations like this one are not arbitrary or capricious. See, e.g., Ass’n of Accredited Cosmetology Schs. v. Alexander, 979 F.2d 859, 865 (D.C. Cir. 1992) (“Instead of undoing past eligibility, . . . the Act and the regulations merely require the Department to look at schools’ past default rates in determining future eligibility for . . . program participation.”). 11 In a twelfth try at an arbitrary-or-capricious finding, the Association calls the punishments attached to the debt-to-earnings test “[o]verbroad,” because (as it reads the regulations) programs that lose Title IV funding eligibility “cannot offer any other program . . . that shares the same fourdigit Classification of Instructional Programs . . . code.” Pl.’s Mot. at 51 (emphasis omitted). The Association misreads the rule, however. In truth, just because one program within a four-digit CIP code fails the debt-to-earnings test (e.g., a school’s associate’s degree program in nursing), does not mean that all of a school’s existing programs that share that code become ineligible for Title IV funds. Instead, the prohibition on similar programs applies only to new programs; that is to say, a school cannot get a failing grade in its associate’s degree nursing program and then start a bachelor’s degree program in nursing. See 34 C.F.R. § 668.410(b)(2); 79 Fed. Reg. at 64,973. Painted in the correct light, then, there is nothing “overbroad” about this rule’s brush. The Department rationally concluded that the prohibition on new, similar programs was necessary in order to “deter institutions from making small changes to a program solely for the purpose of” 11 The Association raises a similar argument concerning so-called “secondary” retroactivity. A regulation can be secondarily retroactive if it “upsets expectations based in prior law.” Nat’l Petrochemical & Refiners Ass’n v. EPA, 630 F.3d 145, 159 (D.C. Cir. 2010) (internal quotation marks omitted). But there are no such expectations in this case. As the APSCU I court noted, “schools have no vested right to future eligibility to participate in Title IV programs.” 870 F. Supp. 2d at 151 (internal quotation marks and citation omitted). Moreover, “[b]ecause there were no prior rules that related to the gainful employment regulations at issue in this case, plaintiffs cannot claim they have incurred any past investment in reliance on any rule.” Id. at 152 (internal quotation marks and alterations omitted). 26 circumventing the Department’s aid-ineligibility determination. 79 Fed. Reg. at 64,991. There is, again, nothing irrational about this policy choice. Finally, the Association’s thirteenth argument is also unavailing. It contends that the regulations do not include any meaningful mechanism for challenging the Department’s debt-toearnings test results, and that this omission independently violates the APA. See Pl.’s Mot. at 52– 53. Not so. The Higher Education Act requires the Department to give only “reasonable notice and opportunity for hearing” to programs denied Title IV funding. 20 U.S.C. § 1094(c)(1)(F). These are not the magic words that trigger the APA’s formal adjudication provisions, see R.R. Comm’n of Texas v. United States, 765 F.2d 221, 227 (D.C. Cir. 1985) (holding that formal adjudication applies only where a “hearing on the record” is required by the governing law (emphasis added)), and thus the Department’s rule need only meet the standard for informal adjudication under the APA, see 5 U.S.C. § 555. That standard is no obstacle here, where the regulations allow schools to, for example, see the mean and median earnings figures the Department used to calculate debt-to-earnings rates, offer contrary earnings data based on the school’s own survey of graduates or other state-gathered data, and challenge the list of students the Department used in its calculations. See 34 C.F.R. § 668.405; id. § 668.406. This suffices. See, e.g., Sw. Airlines Co. v. Transp. Sec. Admin., 650 F.3d 752, 757 (D.C. Cir. 2011) (“In informal adjudications like these, agencies must satisfy only minimal procedural requirements.” (internal quotation marks and brackets omitted)). To recap: try as it might, the Association has not shown that the Department unreasonably interpreted an ambiguous statutory command, nor has it proven (despite at least a baker’s-dozen arguments) that the debt-to-earnings portion of the Department’s final “gainful employment” rule is arbitrary or capricious or otherwise in violation of the APA. The debt-to-earnings test therefore 27 survives this latest challenge, as it previously has. See Ass’n of Proprietary Colls., 2015 WL 3404190, at *23 (“The Court concludes, in all the circumstances, that the [“gainful employment” rules]—and the . . . rates contained therein—were the product of reasoned decisionmaking.” (internal quotation marks omitted)); APSCU I, 870 F. Supp. 2d at 153–54 (similar). III. ANALYSIS OF THE DISCLOSURE, REPORTING, & CERTIFICATION RULES The Association’s final set of challenges takes aim not at the debt-to-earnings test, but at the “gainful employment” rule’s various disclosure, reporting, and certification requirements. Each of these challenges also proves unsuccessful. A. Disclosure Requirements The regulations require that “[a]n institution . . . use the disclosure template provided by the [Department] to disclose information about each of its [“gainful employment”] programs to enrolled and prospective students.” 34 C.F.R. § 668.412(a). “That information may include, but is not limited to” details concerning the size (in students), length (in calendar time), and cost (in dollars) of the program. Id. The Association believes that these requirements are not authorized by any statute, that they violate the First Amendment, and that they are arbitrary or capricious. Can any of these claims succeed? The answer is no for both procedural and substantive reasons. On procedure: for starters, this challenge is not yet ripe for judicial review, as a quick glance at the regulations makes clear. The regulations say that the information subject to disclosure “may include, but is not limited to . . .” sixteen possible data points, id. (emphasis added), and the final rule further equivocates that the Department “do[es] not intend to include all of the disclosure items listed in [that regulation] on the disclosure template each year,” 79 Fed. Reg. at 64,976. The upshot is that the parties (and the Court) cannot yet know what information, exactly, programs will be required to provide, or 28 what problems, exactly, the disclosures will cause. This foggy state of affairs hardly lends itself to review of concrete harms subject to judicial remedy. See Texas v. United States, 523 U.S. 296, 300 (1998) (“A claim is not ripe for adjudication if it rests upon contingent future events that may not occur as anticipated, or indeed may not occur at all.” (internal quotation marks omitted)). Moreover, at least one of the Association’s claims runs headlong into an already-litigated issue. Recall that in APSCU I, the district court upheld the disclosure requirements imposed under the 2011 version of the “gainful employment” rule. In so doing, the court explicitly rejected the Association’s argument (virtually identical to the argument it makes here) that the Department lacked statutory authority to promulgate any such disclosure regulations: “[t]he disclosures mandated . . . fall comfortably within [the Department’s] regulatory power, and are therefore within the Department’s authority under the Higher Education Act.” 870 F. Supp. 2d at 156 (citing 20 U.S.C. §§ 1221e-3, 3474). The Association did not appeal this portion of APSCU I, yet it now hopes to relitigate that already-considered, already-decided issue. This it cannot do. See Yamaha Corp. of Am. v. United States, 961 F.2d 245, 254 (D.C. Cir. 1992) (“[O]nce an issue is raised and determined, it is the entire issue that is precluded, not just the particular arguments raised in support of it in the first case.”). Some of the Association’s complaints therefore fail from the outset. Alternatively, and in any event, a substantive review of the Association’s various claims comes to the same result. First, several statutes give the Department all the authority it needs to require these disclosures. Take the two statutes relied on by the APSCU I court: 20 U.S.C. § 1221e-3 and 20 U.S.C. § 3474. The former gives the Department broad authority “to make, promulgate, issue, rescind, and amend rules and regulations governing the manner of operation of, and governing the applicable programs administered by, the Department.” 20 U.S.C. § 1221e-3. And the latter authorizes “[t]he Secretary . . . to prescribe such rules and regulations as the 29 Secretary determines necessary or appropriate to administer and manage the functions of the Secretary or the Department.” Id. § 3474. These provisions fashion an awfully big umbrella, and it is no stretch to conclude that the 2014 disclosure regulations fall under it. The regulations, after all, are designed to give the public “access to meaningful and comparable information about student outcomes and the overall performance of [“gainful employment”] programs,” 79 Fed. Reg. at 64,891—a goal that surely advances the purposes of both the Higher Education Act and Title IV (i.e., a statute and programs “administered” and “manage[d]” by the Department). Second, the Association’s First Amendment claim falls short as well. It believes that requiring schools to reveal to students “uncertain, speculative estimates” concerning the total cost of completing their programs is impermissible compelled speech. See Pl.’s Mot. at 54. But there are two problems with this argument. To begin with, the Association has not offered a single piece of evidence to show that any of its members do not know the cost of their programs and instead can only estimate. The Association’s claimed injury thus falls on the speculative (and insufficient) side of the line. See, e.g., Clapper v. Amnesty Int’l USA, 133 S. Ct. 1138, 1147 (2013) (“[T]hreatened injury must be certainly impending to constitute injury in fact.” (internal quotation marks omitted)). Moreover, requiring commercial entities to disclose “purely factual and uncontroversial information” does not contravene the First Amendment. Zauderer v. Off. of Disciplinary Counsel, 471 U.S. 626, 651 (1985). That describes this case. The regulations require schools to disclose the “total cost” of things like tuition, fees, books, and supplies for its programs. 34 C.F.R. § 668.412(a)(7). But if a program does “not know precisely the cost[,] . . . the institution shall include a disclaimer advising that the data are estimates.” 79 Fed. Reg. at 64,977–78. Either option thus produces “factual” and “uncontroversial” disclosures: if the program knows its dollarsand-cents cost, it reveals it; if the program does not, it reveals its estimate along with a disclaimer 30 making clear that the estimate is, indeed, just that. Such compelled speech steers well clear of any First Amendment pitfalls. See Zauderer, 471 U.S. at 651 (contrasting a permissible “purely factual” disclosure requirement with “attempt[s] to prescribe what shall be orthodox in politics, nationalism, religion, or other matters of opinion” (internal quotation marks omitted)). 12 Third, the disclosure requirements are not arbitrary, despite the Associations’ protests to the contrary. It complains: (1) that the program-level disclosures conflict with another provision of the Higher Education Act that already requires disclosure of certain institution-level information, see Pl.’s Mot. at 55 (citing 20 U.S.C. § 1092(a)(1)(A), (E), (L)); (2) that programs must disclose default-rate information, even though “the Department abandoned [this rate] as a gainful-employment metric,” id. at 56; and (3) that “[t]he new rule’s requirements that schools provide non-English language warnings are . . . unduly vague and impossible to obey” because schools must follow this command “‘[t]o the extent practicable,’” id. at 55 n.77 (quoting 34 C.F.R. § 668.410(a)(4)). But these arguments turn out to be just three more strikes against the Association. To start, as has been explained, there is no “conflict” with the Act here; there is, instead, a gap—one the Department has chosen to fill with program-level disclosure requirements in response to another statutory command. See supra at 12; see also 79 Fed. Reg. at 64,979. Next, although the Department has opted against relying on default rates as a separate “gainful 12 The cases the Association cites in response are not much help. First up is Chemical Manufacturers Association, 1980 WL 29285 (D.D.C. Dec. 21, 1980). That case involved not the First Amendment, but the Freedom of Information Act. Specifically, the court analyzed whether cost estimates for proposed labelling requirements were deliberative and thus exempt from FOIA’s disclosure requirements. See id. at *3. That discussion therefore offers nothing useful about how this case should come out. The Association’s second case, R.J. Reynolds Tobacco Co. v. FDA, 696 F.3d 1205 (D.C. Cir. 2012), abrogated on other grounds by Am. Meat Inst. v. U.S. Dep’t of Agric., 760 F.3d 18, 22–23 (D.C. Cir. 2014) (en banc), is also inapt. There, government regulations required cigarette packages to display “graphic warning labels—which incorporate[d] . . . textual warnings, a corresponding graphic image, and [a] ‘1-800-QUIT-NOW’ cessation hotline number.” Id. at 1211. Obviously, such graphic, compelled speech is a far cry from the vanilla, estimated-cost disclosures at issue here. Moreover, the court in R.J. Reynolds acknowledged that the Zauderer standard applies not just to purely factual and uncontroversial information, but also to “‘accurate statement[s].’” Id. at 1216 (quoting Milavetz, Gallop & Milavetz, P.A. v. United States, 559 U.S. 229, 231 (2010)). The “total cost” estimates contemplated here certainly meet that description. 31 employment” metric, that information still has some non-arbitrary use. “Default rates,” as the rule explains, “are important information for students to consider as they decide where to pursue . . . their postsecondary education and whether or not to borrow to attend a particular program.” 79 Fed. Reg. at 64,915. Finally, the “to the extent practicable” language hardly sinks the rule. The Supreme Court, after all, has previously rejected a similar vagueness argument. See Sproles v. Binford, 286 U.S. 374, 393 (1932) (“‘Shortest practicable route’ is not an expression too vague to be understood.”). And in any event, the rule provides a roadmap for how programs can comply with the requirement—so it cannot be called “impossible.” See id. at 64,970 (describing “one simple test” for meeting the requirement). The disclosure requirements therefore survive. B. Reporting Requirements The same goes for the 2014 rule’s reporting requirements. The rule requires schools to report several pieces of information about students to the Department, which the Department will then use to calculate the schools’ debt-to-earnings scores. 13 Schools need not report information on all students, however, because the rule applies only to those “enrolled in a [“gainful employment”] program during an award year who received title IV, [Higher Education Act] program funds for enrolling in that program.” 34 C.F.R. § 668.411(a)(1). That is to say, schools need not collect and report information about students who do not receive such funds. Despite this limitation, the Association believes the requirements both exceed the Department’s statutory authority and violate a separate statute. Neither argument prevails. Start with the statutory-authority question. Federal law authorizes the Department to “collect data and information on applicable programs for the purpose of obtaining objective 13 Among the information that schools must report is “[t]he total amount the student received from private education loans . . . for enrollment in the program that the institution is, or should reasonably be, aware of.” 34 C.F.R. § 668.411(a)(2)(ii). More on this in a moment. 32 measurements of the effectiveness of such programs in achieving the intended purpose of such programs.” 20 U.S.C. § 1231a(3). And as explained, federal law also gives the Department the power to “carry out” programs like Title IV. Id. §§ 1221e-3, 3474. Surely, the reporting requirements at issue here fall within this broad authority. After all, the requirements supply the data necessary to feed the debt-to-earnings metrics, which are “objective measurements of the effectiveness of [“gainful employment”] programs.” Because these debt-to-earnings metrics survive, the Association’s statutory-authority argument must fail—as the Association all but admits. See Pl.’s Opp’n at 54 (“If the debt-to-earnings test or the disclosure rules are invalidated, the reporting rules must fall also.”). Nor do the reporting requirements violate 20 U.S.C. § 1015c. That statute prohibits the collection of certain personally identifiable information about students, unless the information is collected as part of “a system . . . that (1) is necessary for the operation of programs authorized [under Title IV]; and (2) was in use by the [Department as of August 13, 2008].” Both prongs of the exception apply here. The student information collected under the 2014 rule is a necessary prerequisite for both the debt-to-earnings test and the disclosure requirements. And the Department has explained that it will incorporate the reported information into the National Student Loan Data System, a pre-existing database that the Department is authorized to maintain. See 79 Fed. Reg. at 64,975–76; see also APSCU II, 930 F. Supp. 2d at 218 (“[T]he NSLDS was in use by the Department long before August 1[3], 2008. The Department is obviously not barred from maintaining it.” (internal citation omitted)). The Association disagrees with the latter conclusion. Borrowing from APSCU II, it argues that because the Department plans to collect private-loan information under the latest reporting regime, it has effectively transformed the NSLDS into something new. See Pl.’s Mot. at 56–57. 33 But neither the holding of APSCU II nor its rationale support that conclusion. Regarding the holding: the court acknowledged that “there [is] a point at which an existing database could be changed so substantially that it effectively be[comes] a new database.” 930 F. Supp. 2d at 218. But—when faced with a 2011 regime that (like the present rule) allowed the collection of privateloan data, and that (unlike the present rule) allowed collection of such data from non-Title IVfunding recipients, see id. at 214—where did the court draw the line? As Judge Contreras saw things, it was not the data the Department sought to collect that was the problem; it was, instead, who the data concerned. “The Department,” in other words, “could not create a . . . system of information on all students in gainful employment programs; nor can it graft such a system onto a pre-existing database of students who have applied for or received Title IV assistance.” Id. at 221 (emphasis added). This holding does not undermine the narrower reporting regime at issue here. See 34 C.F.R. § 668.411(a)(1) (excluding non-Title IV students from the reporting requirements). APSCU II’s rationale does the Association no favors either. As the court explained, “[b]efore the [2011] regulations . . . were promulgated, the NSLDS contained records on borrowers who have applied for and received loans under a variety of federal programs . . . . It did not contain any information about students who were not direct beneficiaries of Title IV programs and had not applied to benefit from those programs.” 930 F. Supp. 2d at 219. Given this pedigree, the court reasoned that any proposed expansion of the NSLDS beyond Title IV students would not fit with the database’s “‘overall purpose.’” Id. at 221. But history points in a different direction here. In fact, the NSLDS has long collected data that is similar to the private-loan information included in the 2014 rules. The database already stores information concerning, for example, “the amount and type of each . . . loan made” to a student, “the total amount of loans made to any borrower and the remaining balance of the loans,” “the lender, holder, and servicer of such loans,” and—perhaps 34 most tellingly—“information concerning other student financial assistance received by the borrower.” 20 U.S.C. § 1092b (emphasis added). Moreover, Congress made clear that the universe of collectible data is not limited to those items listed in the statute. See id. (“The information in the data system shall include (but is not limited to) . . .”). Unlike in APSCU II, then, the Department’s current expansion of the NSLDS is a reasonable (and hence permissible) one. See Chevron, 462 U.S. at 843; see also APSCU II, 930 F. Supp. 2d at 220–21 (asking whether the 2011 reporting regime was a reasonable interpretation of sections 1015c and 1092b). 14 C. Certification Requirements Leaving no stone unturned, the Association takes issue with the 2014 regulations’ certification requirements as well. The regulations require a school to certify that (among other things) “[e]ach eligible [“gainful employment”] program it offers is approved by a recognized accrediting agency,” “is programmatically accredited,” and “satisfies the applicable educational prerequisites for professional licensure . . . in [the state in which the institution is located].” 34 C.F.R. § 668.414(d). The Association argues that these requirements, too, exceed the Department’s statutory authority and are arbitrary and capricious to boot. Wrong again. In a new take on an old theme, the Association contends that the Department only has authority to require “institutions” (not “programs”) to certify that they meet “national” (not “state”) accreditation and licensure standards. In other words, “[b]y conditioning Title IV eligibility on program-level, state-specific accreditation, the Department has impermissibly substituted its 14 The Association suggests that it is arbitrary and capricious to exclude non-Title IV student data from the reporting requirements. See Pl.’s Mot. at 45 (“Nothing in the [Higher Education Act] supports ignoring students who receive no Title IV funds, and excluding them will distort schools’ debt-to-earnings ratios.”). But this is disingenuous; after arguing that including non-Title IV students would violate section 1015c in APSCU II, the Association now asserts that such students’ data must be included in order to avoid arbitrary or capricious results. The Association cannot have it both ways, and the Court declines to play this “gotcha” game. Moreover, as described previously, leaving this non-Title IV data out of the equation does not render the Department’s calculations arbitrary or capricious, as the Department is only required to use the “best data available”—and non-Title IV data collection is, as the court in APSCU II held, inconsistent with the Higher Education Act. 930 F. Supp. 2d at 221. 35 policy judgment for that of Congress.” Pl.’s Mot. at 59 (internal quotation marks and brackets omitted). But a familiar rejoinder rings true here as well: just because Congress elsewhere discussed institution- and national-level accreditations does not rob the Department of authority to mandate accreditations at the program- and state-level. See, e.g., 20 U.S.C. § 1002(b)(1)(A)(i) (describing “eligible program[s] of training . . .” (emphasis added)). Regardless, the Department has broad authority to carry out the purposes of Title IV—and thus, to ensure that Title IV funds only go to programs that actually prepare students for gainful employment in recognized occupations. See, e.g., id. §§ 1221e-3, 3474. Requiring schools to certify that their programs comply with Title IV’s funding limitations does not exceed that authority. One last question needs answering: are the certification requirements arbitrary or capricious? The Association says yes, reasoning first that the certifications are arbitrarily (even unconstitutionally) vague, “because they fail to define clearly which licensure requirements any given program must satisfy.” Pl.’s Mot. at 59 (emphasis added). But this gets both the law and the facts wrong. On the law: “[t]he vagueness doctrine does not require perfect clarity and precise guidance,” Act Now to Stop War and End Racism Coal. v. Dist. of Columbia, 905 F. Supp. 2d 317, 331 (D.D.C. 2012) (internal quotation marks omitted), and this is especially true where economic regulations are at play. “[B]usinesses,” after all, “face economic demands to plan behavior carefully, [and] can be expected to consult relevant legislation in advance of action.” Vill. of Hoffman Estates v. Flipside, Hoffman Estates, Inc., 455 U.S. 489, 498 (1982). Just so here. Indeed, a review of the facts reveals nothing vague about the rule’s licensure scheme. A school need only certify that its programs meet the licensing standards for occupations that the school chooses to list on its disclosure template, see 34 C.F.R. § 668.414(d)(3); 79 Fed. Reg. at 64,989—which is not an especially complicated or difficult-to-follow command. 36 And the answer to the arbitrary-or-capricious question remains no, notwithstanding the Association’s second argument: that the certifications “increase the risk that schools will be subject to conflicting requirements of multiple States.” Pl.’s Mot. at 59. This is an invention by the Association. The current regulatory scheme requires schools to be “legally authorized to provide an educational program . . . in the State in which the institution is physically located.” 34 C.F.R. §§ 600.4(a)(3), 600.5(a)(4), 600.6(a)(3); see also 79 Fed. Reg. at 64,992. Those regulations say nothing about authorizations in other states, and the mere possibility that the Department may someday amend its regulations to require multiple authorizations in multiple states (and that those authorizations could theoretically subject schools to conflicting commands) is at least two degrees of guesswork too many. See Texas, 523 U.S. at 300. 15 The certification requirements thus go the way of the disclosure and reporting requirements before them: they survive the Association’s challenge in full. CONCLUSION And that, as they say, is that. The Department’s “gainful employment” regulations— including the current debt-to-earnings test and disclosure, reporting, and certification requirements—survive this court challenge in their entirety, just as prior courts have concluded. The Court will therefore deny the Association’s motion for summary judgment and grant the Department’s cross-motion. A separate Order has issued on this date. /s/ JOHN D. BATES United States District Judge Dated: June 23, 2015 15 Perhaps recognizing this insurmountable problem, the Association declined to press this argument in its reply brief. See Pl.’s Opp’n at 56. 37