Philbrick v. Azar II

UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF COLUMBIA

SAMUEL PHILBRICK, et al.,

Plaintiffs,
v. Civil Action No. 19-773 (JEB)
ALEX M. AZAR II, et al.,

Defendants.

MEMORANDUM OPINION

In November 2018, the Secretary of Health and Human Services approved the State of

New Hampshire’s proposal to impose work requirements on a significant share of its Medicaid

recipients. Under the proposal, most non-disabled Medicaid beneficiaries ages 19 to 64 would

be required to demonstrate that they have completed 100 hours of qualifying employment or

other “community-engagement” activities each month (or show that they satisfy an exemption)

or risk losing their health-care coverage. Four New Hampshire residents have challenged the

Secretary’s approval in this Court, arguing that it violates the Administrative Procedure Act and

the Constitution.

The issues presented in this case are all too familiar. In the past year or so, this Court has

resolved challenges to similar programs in Kentucky and Arkansas, each time finding the

Secretary’s approval deficient. See Stewart v. Azar, 366 F. Supp. 3d 125, 131 (D.D.C. 2019)

(Stewart II); Gresham v. Azar, 363 F. Supp. 3d 165, 169 (D.D.C. 2019); Stewart v. Azar, 313 F.

Supp. 3d 237, 243 (D.D.C. 2018) (Stewart I). The overriding shortcoming in the agency’s

decisions in those cases was its failure to adequately consider the requirements’ effects on

Medicaid coverage. Despite conceding that providing medical care to the needy is “Medicaid’s

1
core objective,” Gresham, 363 F. Supp. 3d at 176 (citation omitted), HHS did not “offer its own

estimates of coverage loss or grapple with comments in the administrative record projecting that

the proposal would lead a substantial number of residents to be disenrolled from Medicaid.” Id.

at 175 (cleaned up).

Plaintiffs argue that the Secretary’s approval of New Hampshire’s plan suffers from the

same deficiency and thus must meet the same fate. The Court concurs. On their face, these

work requirements are more exacting than Kentucky’s and Arkansas’s, mandating 100 monthly

hours — as opposed to 80 — of employment or other qualifying activities. They also encompass

a larger age range than in Arkansas, which applied the requirements only to persons 19 to 49.

Yet the agency has still not contended with the possibility that the project would cause a

substantial number of persons to lose their health-care coverage. That omission is particularly

startling in light of information before the Secretary about the initial effects of Arkansas’s

markedly similar project — namely, that more than 80% of persons subject to the requirements

had reported no compliance information for the initial months, and nearly 16,900 people had lost

coverage. The agency’s rejoinders — that the requirements advance other asserted purposes of

Medicaid, such as the health and financial independence of beneficiaries and the fiscal

sustainability of the safety net — are identical to those this Court rejected with respect to HHS’s

2018 approval of Kentucky’s program. Perhaps seeing the writing on the wall, the Government

conceded at oral argument that its reasoning was deficient in these respects under the analysis in

the Court’s prior Opinions.

In short, we have all seen this movie before. The Secretary has significant discretion to

approve demonstration projects that promote the objectives of the Medicaid Act, and it is not for

the Court to second guess his policy decisions or substitute its judgment for his. “But courts

2
retain a role, and an important one, in ensuring that agencies have engaged in reasoned

decisionmaking.” Judulang v. Holder, 565 U.S. 42, 53 (2011). At the heart of this review is an

assessment of “whether the decision was based on a consideration of the relevant factors.” Id.

(citation omitted). For the fourth time, HHS has fallen short of this fundamental administrative-

law requirement. The Court will, accordingly, grant summary judgment to Plaintiffs and vacate

the Secretary’s approval of New Hampshire’s community-engagement requirements.

I. BACKGROUND

The Court begins with a now-familiar overview of the relevant history and provisions of

the Medicaid Act. It then turns to New Hampshire’s challenged plan before concluding with the

procedural history of this case.

The Medicaid Act

Since 1965, the federal government and the states have worked together to provide

medical assistance to certain vulnerable populations under Title XIX of the Social Security Act,

commonly known as Medicaid. See 42 U.S.C. § 1396-1. The Centers for Medicare and

Medicaid Services (CMS), a federal agency within the Department of Health and Human

Services, has primary responsibility for overseeing Medicaid programs. Under the cooperative

federal-state arrangement, participating states submit their “plans for medical assistance” to the

Secretary of HHS. Id. To receive federal funding, those plans — along with any material

changes to them — must be “approved by the Secretary.” Id.; see also 42 C.F.R. § 430.12(c).

Currently, all states have chosen to participate in the program.

To be approved, state plans must comply with certain minimum parameters set out in the

Medicaid Act. See 42 U.S.C. § 1396a (listing 86 separate requirements). One such provision

requires state plans to “mak[e] medical assistance available” to certain low-income individuals.

3
Id. § 1396a(a)(10)(A). Until recently, that group included pregnant women, children, and their

families; some foster children; the elderly; and people with certain disabilities. Id. In 2010,

however, Congress enacted the Patient Protection and Affordable Care Act (ACA), colloquially

known as Obamacare, “to increase the number of Americans covered by health insurance.” Nat’l

Fed’n of Indep. Business v. Sebelius, 567 U.S. 519, 538 (2012). Of relevance here, that statute

required participating states to expand Medicaid coverage to additional low-income adults under

65 who did not previously qualify. See 42 U.S.C. § 1396a(a)(10)(A)(i)(VIII).

Generally, a state must cover all qualified individuals or forfeit its federal Medicaid

funding. Id. § 1396a(a)(10)(B); id. § 1396c. That was originally so for the ACA expansion

population as well. Id. § 1396c. In NFIB, however, the Supreme Court held that Congress could

not, consistent with the Spending Clause of the Constitution, condition a state’s entire Medicaid

funds on its agreeing to the expansion. See 567 U.S. at 584–85. As a result, states could choose

not to cover the new population and lose no more than the funds that would have been

appropriated for that group. Id. at 587. If the state, conversely, does decide to provide coverage,

those individuals would become part of its mandatory population. Id. at 585–87 (explaining that

Congress may “offer[] funds under the Affordable Care Act to expand the availability of health

care, and require[] that States accepting such funds comply with the conditions on their use”). In

that instance, the state must afford the expansion group “full benefits” — i.e., it must provide

“medical assistance for all services covered under the State plan” that are substantially

equivalent “in amount, duration, or scope . . . to the medical assistance available for [other]

individual[s]” covered under the Act. See 42 U.S.C. § 1396d(y)(2)(B); 42 C.F.R.

§ 433.204(a)(2).

4
 The Medicaid Act, in addition to defining who is entitled to coverage, also ensures what

coverage those enrolled individuals receive. Under § 1396a, states must cover certain basic

medical services, see 42 U.S.C. §§ 1396a(a)(10)(A), 1396d(a), and the statute limits the amount

and type of premiums, deductions, or other cost-sharing charges that a state can impose on such

care. Id. § 1396a(a)(14); see also id. § 1396o. Other provisions require states to provide up to

three months of retroactive coverage once a beneficiary enrolls, id. § 1396a(a)(34), and to ensure

that recipients receive all “necessary transportation . . . to and from providers.” 42 C.F.R.

§ 431.53. Finally, states must “provide such safeguards as may be necessary to assure that

eligibility” and services “will be provided, in a manner consistent with simplicity of

administration and the best interests of the recipients.” 42 U.S.C. § 1396a(a)(19).

Both before and after the passage of the ACA, a state accepting federal Medicaid funds is

not entirely locked in; instead, if it wishes to deviate from certain of the Act’s requirements, it

can seek a waiver from the Secretary of HHS. See 42 U.S.C. § 1315. In particular, Section 1115

of the Social Security Act allows the Secretary to approve “experimental, pilot, or demonstration

project[s] which, in [his] judgment . . . , [are] likely to assist in promoting the [Act’s]

objectives.” 42 U.S.C. § 1315(a). As conceived, experimental projects were “expected to be

selectively approved by the Department and to be those which are designed to improve the

techniques of administering assistance.” S. Rep. No. 1589, 87th Cong., 2d Sess. 19, reprinted in

1962 U.S.C.C.A.N. 1943, 1962. Once the Secretary has greenlighted such a project, he can then

waive compliance with the requirements of § 1396a “to the extent and for the period . . .

necessary to enable [the] State . . . to carry out such project.” Id. § 1315(a)(1).

While the ultimate decision whether to grant § 1115 approval rests with the Secretary, his

discretion is not boundless. Before HHS can act on a waiver application, the state “must provide

5
at least a 30-day public notice[-]and[-]comment period” regarding the proposed program and

hold at least two hearings at least 20 days before submitting the application. See 42 C.F.R.

§§ 431.408(a)(1), (3). Once a state completes those prerequisites, it then sends an application to

CMS. Id. § 431.412 (listing application requirements). After the agency notifies the state that it

has received the waiver application, a federal 30-day public-notice period commences, and the

agency must wait at least 45 days before rendering a final decision. Id. §§ 431.416(b), (e)(1).

Factual Background

1. New Hampshire Granite Advantage

In 2014, New Hampshire, like many states, expanded Medicaid under the ACA to

previously uninsured adults whose income is 133 percent of the federal poverty line or less. See

AR 17; AR 1949. More than 53,000 individuals have received coverage as a result, helping to

reduce the State’s uninsured rate by 45 percent. Id. at 4384. Since 2015, the State has covered

this population through Section 1115 demonstration projects that deviate from traditional

Medicaid delivery mechanisms — first adopting a premium-assistance model and later shifting

to a managed-care system. Id. at 4379. While New Hampshire has had an interest in work

requirements dating back to 2016, id. at 99, it proposed to amend its demonstration to add the

work and community-engagement requirements under consideration in this suit in 2018. Id. at

4377.

As proposed, the project — now called Granite Advantage — requires most non-disabled

adults aged 19 to 64 to complete 100 hours per month of employment or other community

activities. Id. at 4. Certain categories of beneficiaries are exempt, including caregivers for a

dependent child, pregnant women, and the medically frail. Id. at 5. If a beneficiary does not

demonstrate compliance with the work requirements in a particular month, she will be sent a

6
notice stating that her Medicaid will be terminated the following month if she does not make up

the hours or show that she qualifies for an exemption. Id. Once a beneficiary’s coverage is

suspended, it can be reactivated by completing 100 hours of qualifying activities or obtaining an

exemption. Id. at 5, 7. Separately, New Hampshire requested as part of these amendments that

HHS allow the State to eliminate all retroactive coverage. Id. at 4377.

The Secretary approved the amendments on November 30, 2018, explaining that they

promoted the purposes of the Medicaid Act because they would improve the “health and

wellness” of beneficiaries and enhance the “fiscal sustainability of the Medicaid program.” Id. at

1–2. With respect to commenters’ concerns that some beneficiaries would lose coverage, the

agency responded that “the demonstration will provide coverage to individuals that the state is

not required to cover” — namely, the ACA expansion population. Id. at 10. Indeed, because

“the state plans to end its current coverage of the new adult group” in the event the project were

not approved, HHS says, Granite Advantage necessarily increases coverage. Id. at 6, 10. The

agency further explained that the requirements were “not designed to encourage” coverage loss

and are “intended to [be] achievable,” citing certain exemptions and safeguards that are meant to

reduce the likelihood of persons improperly losing their Medicaid. Id. at 10–11.

While the new requirements could have been implemented under this approval beginning

January 1, 2019, id. at 1, they have still not been put into full effect. New Hampshire, after

several initial delays, required beneficiaries to submit qualifying hours or proof of an exemption

this past June. See ECF No. 1 (Complaint), ¶ 10. Under that timeframe, persons who did not

satisfy the reporting obligations would lose their coverage on August 1. Id. As of July 8, 2019,

however, approximately 17,000 non-exempt beneficiaries (out of about 25,000 total) had not

reported any compliance information to the New Hampshire Department of Health and Human

7
Services. See ECF No. 44-2 (Jeffrey A. Meyers Letter, July 8, 2019) at 3. Citing this

consideration and emphasizing the difficulty the State has had in communicating with persons

subject to the community-engagement requirements, the Department announced that it was

further delaying implementation until September 30, 2019. See ECF No. 44 (Notice) at 2.

Under the new implementation plan, Medicaid beneficiaries who do not report compliance with

the requirements would lose coverage beginning December 1. Around the same time, the New

Hampshire Legislature amended the program in several respects, including by expanding the

scope of the exemptions. The State explained that it plans to seek reapproval of such

amendments from CMS over the next several months. See Oral Argument Transcript

(Provisional) at 3, 17.

2. Other CMS Approvals

New Hampshire is not the only state that has been interested in work requirements. As

noted at the start, CMS has approved similar proposals submitted by Kentucky and Arkansas,

each of which has been challenged and struck down in this Court. Kentucky’s program — called

Kentucky HEALTH — mirrors New Hampshire’s in many respects. As relevant here, it requires

non-exempt adults aged 19 to 64 who receive coverage through the expansion to complete and

report 80 hours per month of qualifying activities, such as employment, education, or job

training. See Stewart I, 313 F. Supp. 3d at 246. The failure to do so or to report an exemption

results in the termination of Medicaid coverage. Id. at 246–47.

Before the requirements took effect in the Commonwealth, several Medicaid recipients

sought judicial review of HHS’s approval. Id. at 248. They argued, among other things, that the

agency had failed to adequately explain why Kentucky HEALTH promoted the objectives of

Medicaid and that the approval of the project exceeded the Secretary’s statutory authority. The

8
Court agreed with Plaintiffs in one central and dispositive respect: “[T]he Secretary never

adequately considered whether Kentucky HEALTH would in fact help the state furnish medical

assistance to its citizens, a central objective of Medicaid.” Id. at 243. It therefore vacated the

Secretary’s approval and remanded the matter to the agency for further consideration. Id. at

273–74. HHS subsequently reopened the comment period and reapproved Kentucky’s project on

November 20, 2018. The agency reasoned, along substantially similar lines as it did ten days

later when it approved New Hampshire’s project, that Kentucky HEALTH advanced the

Medicaid Act’s objectives because it would 1) promote the health and financial independence of

beneficiaries, a justification the Court had found wanting in the first round, 2) increase coverage

because it allows Kentucky to cover the expansion population when it would not do so

otherwise, and 3) advance the fiscal sustainability of the state’s Medicaid program. See Stewart

II, 366 F. Supp. 3d at 138. Believing these justifications still unsatisfactory, the Bluegrass State

plaintiffs returned to this Court, which again concurred. Concluding that the agency’s previous

rationales fared no better and that its new explanation still failed to grapple with the possibility of

coverage loss, the Court vacated the approval. Id. at 138–39.

Arkansas’s project, named the Arkansas Works Amendments, followed a similar,

although abbreviated, path. The State proposed to require most able-bodied beneficiaries in the

expansion population aged 19 to 49 to complete 80 hours of qualifying employment or other

activities. See Gresham, 363 F. Supp. 3d at 172. Non-exempt individuals who did not report

sufficient qualifying hours for three consecutive months in a calendar year would be disenrolled

from Medicaid for the remainder of that year. Id. The Secretary approved the requirements on

March 5, 2018, and their roll-out was staged through 2018 and early 2019. Id. During the first

six months after implementation, however, “only a small percentage of the persons required to

9
report compliance . . . actually did so” — in October 2018, only 12.3% reported any kind of

qualifying activities — and more than 16,900 persons lost Medicaid coverage for some period of

time as a result. Id. Several beneficiaries challenged the program under the APA, and, finding

the Secretary’s explanation deficient for the same reasons as in its first Kentucky decision, the

Court vacated his approval. Id. at 175. The Court’s decisions as to both of those cases are now

on appeal before the D.C. Circuit. See Case Nos. 19-5094, 19-5095, 19-5096, 19-5097. No oral

argument date has yet been set.

Procedural History

Believing with Shakespeare that what’s past is prologue, four New Hampshire residents

filed this lawsuit on March 20, 2019. Like the plaintiffs in Arkansas and Kentucky, they assert

that the Secretary’s approval of the proposed community-engagement requirements violates the

APA and the Constitution. Because it was designated as related to Stewart and Gresham, the

case was directed to this Court. See ECF Nos. 2–3. The State of New Hampshire has since

intervened as a Defendant, and numerous amici have also weighed in. Dueling Cross-Motions

for Summary Judgment are now ripe, and the Court held a hearing on July 23, 2019.

II. LEGAL STANDARD

The parties have cross-moved for summary judgment on the administrative record. The

summary-judgment standard set forth in Federal Rule of Civil Procedure 56(c), therefore, “does

not apply because of the limited role of a court in reviewing the administrative record.” Sierra

Club v. Mainella, 459 F. Supp. 2d 76, 89 (D.D.C. 2006); see also Bloch v. Powell, 227 F. Supp.

2d 25, 30 (D.D.C. 2002), aff’d, 348 F.3d 1060 (D.C. Cir. 2003). “[T]he 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.” Sierra Club, 459 F. Supp. 2d. at 90 (quotation

10
marks and citation omitted). “Summary judgment is the proper mechanism for deciding, as a

matter of law, whether an agency action is supported by the administrative record and consistent

with the [Administrative Procedure Act] standard of review.” Loma Linda Univ. Med. Ctr. v.

Sebelius, 684 F. Supp. 2d 42, 52 (D.D.C. 2010) (citation omitted).

The Administrative Procedure Act “sets forth the full extent of judicial authority to

review executive agency action for procedural correctness.” FCC v. Fox Television Stations,

Inc., 556 U.S. 502, 513 (2009). It requires courts to “hold unlawful and set aside agency action,

findings, and conclusions” that are “arbitrary, capricious, an abuse of discretion, or otherwise not

in accordance with law.” 5 U.S.C. § 706(2). Agency action is arbitrary and capricious if, for

example, the agency “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 of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463 U.S.

29, 43 (1983).

In other words, an agency is required to “examine the relevant data and articulate a

satisfactory explanation for its action including a rational connection between the facts found and

the choice made.” Id. at 43 (quoting Burlington Truck Lines, Inc. v. United States, 371 U.S. 156,

168 (1962)) (internal quotation marks omitted). Courts, accordingly, “do not defer to the

agency’s conclusory or unsupported suppositions,” United Techs. Corp. v. Dep’t of Def., 601

F.3d 557, 562 (D.C. Cir. 2010) (quoting McDonnell Douglas Corp. v. Dep’t of the Air Force,

375 F.3d 1182, 1187 (D.C. Cir. 2004)), and “agency ‘litigating positions’ are not entitled to

deference when they are merely [agency] counsel’s ‘post hoc rationalizations’ for agency action,

advanced for the first time in the reviewing court.” Martin v. Occupational Safety & Health

11
Review Comm’n, 499 U.S. 144, 156 (1991) (citation omitted). Although a reviewing court “may

not supply a reasoned basis for the agency’s action that the agency itself has not given,” a

decision that is not fully explained may nevertheless be upheld “if the agency’s path may

reasonably be discerned.” Bowman Transp., Inc. v. Arkansas-Best Freight System, Inc., 419

U.S. 281, 285–86 (1974) (citation omitted).

III. ANALYSIS

Just as their predecessors did, Plaintiffs here challenge the Secretary’s approval of New

Hampshire’s demonstration project on a number of different grounds, including that it was

arbitrary and capricious, in excess of statutory authority, and in violation of the Take Care

Clause of the U.S. Constitution. As in its three previous Opinions, the Court need only address

the first to resolve this case. Before turning to that issue, however, it begins with jurisdiction.

Jurisdiction

While Defendants largely do not contest whether subject-matter jurisdiction exists in this

case, the Court has an independent duty to assure as much. See Steel Co. v. Citizens for a Better

Environment, 523 U.S. 83, 94–95 (1998). Two potential hurdles stand across Plaintiffs’ path.

The first is whether the Secretary’s decision is “committed to agency discretion by law” and is

therefore unreviewable under the APA. See 5 U.S.C. § 701(a)(2). HHS insisted in Stewart I that

its determination to approve a Section 1115 demonstration project fell within this exception to

the general presumption that administrative action is judicially reviewable. After a lengthy

discussion, the Court agreed with “every court which has considered the issue” and found that

the Secretary’s approval was “subject to APA review.” 313 F. Supp. 3d at 256 (quoting Beno v.

Shalala, 30 F.3d 1057, 1067 & n.24 (9th Cir. 1994)). Apart from a couple of conclusory

12
sentences in their Opposition and Reply, Defendants no longer press this objection, and the Court

sees no reason to depart from its prior decision finding HHS’s decision reviewable.

The second jurisdictional question is whether Plaintiffs have standing to bring this suit.

Article III of the Constitution limits the Court’s jurisdiction to “cases” or “controversies.” The

standing doctrine enforces this requirement, assuring that courts only decide actual disputes

between parties with personal stakes in the outcome. See Clinton v. City of New York, 524 U.S.

417, 429 (1998). To establish standing, Plaintiffs must show that they have suffered (or will

suffer in the future) a concrete injury that is both fairly traceable to the challenged conduct and

likely to be redressed by a favorable judicial decision. See Lujan v. Defenders of Wildlife, 504

U.S. 555, 590 (1992). While Defendants disputed the basis of the plaintiffs’ standing in Stewart

I, in their briefing this time around, they generally leave their powder dry. Upon review, the

Court can see why: at least one Plaintiff clearly has standing.

Only a few sentences are needed to show as much. Plaintiff Ian Ludders is a forty-year-

old Medicaid recipient who will be subject to New Hampshire’s community-engagement

requirements. See ECF No. 1 (Compl.), ¶¶ 146–56; ECF No. 19-3 (Declaration of Ian Ludders);

see also ECF 37-2 (Declaration of Henry Lipman) at 2. Because he maintains a “subsistence

lifestyle that prioritizes living off the land” and his work is agricultural and thus largely seasonal,

he does not expect that he will be able to comply with the new requirements for multiple months

out of the year. See Compl., ¶¶ 147–56. Taken together, these facts satisfy each of the three

standing requirements: 1) there is a substantial risk that Ludders will lose his Medicaid coverage,

thereby injuring him; 2) this risk is traceable to the Secretary’s approval of the requirements; and

3) a judicial decision vacating them is likely to prevent the future injury. It may well be that

Plaintiffs apart from Ludders also have standing, but there is no need to delve into that issue

13
here. See Comcast Corp. v. FCC, 579 F.3d 1, 6 (D.C. Cir. 2009) (“[I]f one party has standing in

an action, a court need not reach the issue of the standing of other parties when it makes no

difference to the merits of the case.”).

Defendants separately suggest that none of the Plaintiffs has standing to challenge one

aspect of the Secretary’s approval: his decision to allow New Hampshire to eliminate retroactive

Medicaid coverage. See ECF No. 30 (HHS Cross-Mot.) at 24. As the Court explained in its

prior Opinions, however, it is appropriate to “examine[] the approval of the project as a whole,”

given the nature of Plaintiffs’ claim and the administrative action under review. Stewart I, 313

F. Supp. 3d at 253. That makes it unnecessary to decide whether any Plaintiff has standing with

respect to this particular component of the Secretary’s approval.

Merits

Turning to the merits, Plaintiffs principally submit that the Secretary’s approval of the

Granite Advantage project is arbitrary and capricious because it did not adequately consider the

effects of the demonstration project on Medicaid coverage. The Court, as discussed above,

found this argument persuasive in each of its prior three decisions, and it continues to do so. The

critical issues are thus whether the reasoning in the Secretary’s approval letter meaningfully

differs from the previous three, or whether attributes particular to New Hampshire’s community-

engagement program, as discussed in the approval, suggest coverage-loss concerns will be less

significant.

Defendants’ briefing does not attempt to distinguish the approval letter or the program

from those in Stewart I, Stewart II, and Gresham; indeed, it marches through its arguments

barely acknowledging that the Court has decided these precise issues before and adversely to

HHS. At oral argument, however, the Government effectively conceded that the Secretary’s

14
reasoning in this case cannot be distinguished from his explanations in the prior ones. See Tr. at

2, 10. The Court likewise finds the records to be indistinguishable. As discussed in more detail

below, CMS’s approval letter mirrors the one in Stewart II, with numerous key paragraphs

matching it word for word. And New Hampshire’s proposed project presents, if anything,

greater coverage-loss concerns than Kentucky’s and Arkansas’s, given the hours requirement and

the age range to whom it applies.

The Court’s analysis unfolds in two parts. First, it summarizes the now-familiar view

that the core objective of the Medicaid Act is to furnish health-care coverage to the needy and

explains why the Secretary failed to adequately consider that objective here. Turning to

Defendants’ counterarguments — nearly all of which were addressed at length in Stewart II —

the Court will offer an abbreviated restatement of why the agency’s consideration of other

Medicaid Act objectives does not remedy this deficiency.

1. Coverage as Objective of the Medicaid Act

The Secretary, as outlined above, can only approve demonstration projects that are

“likely to assist in promoting the objectives” of the Medicaid Act. See 42 U.S.C. § 1315(a).

Before greenlighting a project, he must therefore identify the objectives of the Act and explain

why the demonstration is likely to promote them. The Court has assumed that the Secretary’s

interpretation of those objectives is entitled to Chevron deference. See Gresham, 363 F. Supp.

3d at 176. That is, in reviewing his understanding, the Court must first ask whether “Congress

has directly spoken to the precise question at issue” and, if not, whether “the agency’s answer is

based on a permissible construction of the statute.” Chevron U.S.A., Inc. v. Nat’l Res. Def.

Council, Inc., 467 U.S. 837, 842–43 (1984). Because all parties agree that “Medicaid’s core

objective” is to “furnish[] medical assistance” to persons who cannot afford it, according such

15
deference would have no practical significance with respect to this objective. See ECF No. 38

(HHS Reply) at 5; ECF No. 35 (Plaintiffs’ Reply) at 6; see also Stewart I, 313 F. Supp. 3d at

260–62. The Court will not repeat its discussion of why the provision of medical assistance to

beneficiaries — both recipients of traditional Medicaid and members of the expansion population

— is the central purpose of the Act, but instead directs interested readers to its prior Opinions.

See Gresham, 363 F. Supp. 3d at 176; Stewart II, 366 F. Supp. 3d at 138; Stewart I, 313 F. Supp.

3d at 260–62.

Having correctly identified the provision of Medicaid coverage as a core objective, the

agency was required to reasonably explain whether New Hampshire’s proposed community-

engagement requirements would advance or impede that goal. In other words, “the Secretary

needed to consider whether the demonstration project would be likely to cause recipients to lose

coverage and whether it could cause others to gain coverage.” Gresham, 363 F. Supp. 3d at 177.

He once again neglected to do so on both counts.

a. Risk to Coverage

Before approving a proposed demonstration, the Secretary must address whether it

creates a risk that beneficiaries will lose their Medicaid coverage. Unlike in Stewart I and

Gresham, here the agency at least mentioned the possibility of coverage loss in its approval.

That is step one. But “stating that a factor was considered . . . is not a substitute for considering

it,” Getty v. Fed. Savs. & Loan Ins. Corp., 805 F.2d 1050, 1055 (D.C. Cir. 1986); he must

“adequately analyze . . . the consequences” of his actions. Am. Wild Horse Pres. Campaign v.

Perdue, 873 F.3d 914, 932 (D.C. Cir. 2017). Here the Secretary fell well short.

For starters, he “never provided a bottom-line estimate of how many people would lose

Medicaid” with Granite Advantage in place. Stewart II, 366 F. Supp. 3d at 140 (quoting Stewart

16
I, 313 F. Supp. 3d at 262). In its proposal, New Hampshire estimated that the project would have

no material effect on Medicaid enrollment, though it also hinted that coverage would otherwise

have expanded given population growth. See AR 4386. The many commenters who addressed

the issue unanimously agreed that coverage loss would be substantial. The Kaiser Family

Foundation, for example, projected an enrollment loss between 6 and 17 percent, corresponding

to between 2600 and 7500 people losing Medicaid. See AR 2532, 4384 (applying disenrollment

rates to 44,000 non-frail adults subject to Granite Advantage). Appendix A documents

numerous comments and studies projecting the same or greater coverage losses. See, e.g., AR

1949, 1953, 2132, 2208, 2210, 2238, 2241. When the Secretary approved the project, he also

had the benefit of data from the first several months of Arkansas’s comparable project, which

presented a stark picture. See AR 2731–47. In that time, only 12.3% of non-exempt persons

“reported any kind of qualifying activity,” and “16,900 individuals ha[d] lost Medicaid coverage

for some period of time” as a result. Gresham, 363 F. Supp. 3d at 172 (emphasis added).

Commenters explained that results were likely to be similar in New Hampshire. See, e.g., AR

2563, 2963. Indeed, because the State’s requirements are more stringent than Arkansas’s in key

respects — e.g., requiring 100 rather than 80 hours per month of activities and applying to adults

aged 19 to 64 rather than 19 to 49 — many projected that Granite Advantage would lead to more

coverage loss than Arkansas’s program. See, e.g., AR 2258, 2542, 2575, 2586.

What does the Secretary think about all this? Does he concur with New Hampshire’s

apparent view that coverage loss is going to be minimal, or does he agree with the commenters

that it is likely to be substantial? Are the coverage losses in Arkansas likely to be replicated in

New Hampshire? We have no idea, since the approval letter offers no hints. While Defendants

may well be correct that HHS does not need to provide a precise numeric estimate of coverage

17
loss, it can hardly be disputed that the agency needs to address the magnitude of that loss. That

is particularly so where the comments uniformly assert — and the record evidence from similar

programs strongly suggests — that the loss will be substantial. The Secretary’s “failure to

address” this “salient factor” renders his decision arbitrary and capricious. See Humane Soc’y of

United States v. Zinke, 865 F.3d 585, 606–07 (D.C. Cir. 2017).

HHS generally conceded at oral argument that the Secretary did not consider coverage to

a greater degree in this case than in Stewart II. See Tr. at 10–11. In their briefing, however,

Defendants offer several responses worthy of discussion. The agency first insists that it did

consider the risk of coverage loss because it mentioned that possibility at several points in its

approval letter. For example, the Secretary explained that “[t]he community engagement

requirements may impact overall coverage levels if the individuals subject to the requirements

choose not to comply with them.” AR 6. This and statements like it butter few parsnips, for

acknowledging the possibility of coverage loss is not the same as analyzing that possibility. See

Getty, 805 F.2d at 1055. The letter gives no indication, as mentioned above, about the

seriousness of the loss; for all the reader gleans, the project could expel 75% of prior Medicaid

beneficiaries. It also neglects the possibility, likewise discussed by numerous commenters, that

many beneficiaries will lose coverage merely because they are not able to satisfy the

administrative burden associated with reporting their compliance. See AR 1484, 1489, 2132,

2241, 3406, 4564.

The Secretary also offers a second, more substantive response to concerns about coverage

loss. He says that Granite Health was designed “to minimize coverage loss due to

noncompliance,” AR 11, citing the exemptions built into the project, the procedural safeguards

intended to prevent improper disenrollment, and the otherwise achievable nature of the

18
requirements. Id. at 7–11. The State of New Hampshire likewise cites the broader intentions of

the program and the corresponding protections put in place to prevent coverage loss. See ECF

No. 37 (NH Reply) at 8. To state the obvious, however, that a project is intended to avoid

coverage loss does not mean that it will do so. Similar intentions existed and corresponding

protections were put in place in Kentucky, but the Commonwealth projected a coverage loss

equivalent to 95,000 people losing Medicaid for one year. Same with Arkansas, yet it found that

nearly 17,000 lost coverage at some point in the first six months alone. The commenters in this

case explained as much, since most of the exemptions or safeguards CMS mentions were baked

into their comments about the likelihood of significant coverage loss. See, e.g., AR 1484, 2132,

2241. Ultimately, the agency’s explanation here comes up short — just as it did in the previous

three cases — because it does not address whether and how these design attributes bear on the

actual magnitude of coverage loss. See Stewart II, 366 F. Supp. 3d at 142–43.

While Defendants offer two remaining arguments as to why they have adequately

considered coverage, both are more properly addressed in subsequent sections. Briefly, they

assert that the project will not actually decrease coverage because it allows the State to cover a

population it would not otherwise cover — namely, the ACA expansion group. The Court will

analyze this argument in the ensuing section on whether the Secretary adequately considered if

the project would promote coverage. See infra Section III.B.1.b. Defendants also maintain that

any coverage loss is outweighed by the project’s promotion of other purposes of Medicaid,

including health, financial independence, and fiscal sustainability. This point — which does not

directly bear on whether the Secretary has ever grappled with coverage — will be addressed

below in Section III.B.2.

19
 b. Promote Coverage

The Secretary offers two possible reasons why Granite Advantage might improve the

state’s ability to furnish medical assistance to the needy. The first can be dispatched quickly. He

says that the elimination of retroactive eligibility may “encourage beneficiaries to obtain and

maintain health coverage, even when they are healthy.” AR 12. Setting aside the equivocal

nature of this assertion and the numerous comments to the contrary in the record (AR 1479–80,

4565), this possibility has nothing to do with the coverage loss the community-engagement

requirements might cause. To the extent the Secretary believes that the elimination of retroactive

coverage might (counterintuitively) increase coverage, he needs to weigh the promotion side of

the scale against the risk-of-loss side when approving the project. That did not happen. Indeed,

such a calculus would be difficult to undertake given the agency’s failure, discussed at length

above, to characterize the magnitude of coverage loss presented by the community-engagement

requirements.

Defendants’ second argument about coverage promotion is one the Court addressed at

length in Stewart II: because the State will “simply de-expand Medicaid” if Granite Advantage is

not approved, any coverage provided to the expansion population through the demonstration is

properly understood as increasing Medicaid coverage. See 366 F. Supp. 3d at 153; see also AR

10. In other words, “[a] demonstration that shrinks coverage may thus be coverage promoting

for the purposes of § 1115 as long as the state threatens that if the demonstration is not approved,

it will discontinue coverage entirely.” Stewart II, 366 F. Supp. 3d at 153; see HHS Cross-Mot. at

17–18. In Stewart II, the Court addressed this point as a variant of the agency’s fiscal-

sustainability rationale because the ostensible reason for Kentucky to de-expand absent approval

20
was its fiscal situation. In doing so, however, the Court noted that the argument “does not

depend on a state’s being in a fiscally precarious position because it does not take into account

the reason the state wants to discontinue participating in the Medicaid program.” Stewart II, 366

F. Supp. 3d at 154.

This case proves the point. New Hampshire has not justified its desire to de-expand

absent approval of the community-engagement requirements on fiscal concerns; indeed, the State

explained at oral argument that it was not facing such budgetary woes. See Tr. (Complete

Transcript Pending). It ultimately does not matter how their argument is characterized; “[it] is

both inconsistent with the Medicaid Act and arbitrary and capricious.” Id. at 153. Before briefly

restating why that is the case, the Court notes that HHS seems largely to have abandoned any

robust form of this argument on appeal. See Stewart v. Azar, Case No. 19-5095, Appellant Brief

at 37 (noting only that “it is permissible for HHS to take the optional character of the coverage

into account when considering such applications”).

As the Court explained in Stewart II, this position is unpersuasive for three interrelated

reasons. To start, Defendants incorrectly assume that “a state has additional discretion to

diminish or condition eligibility for the expansion — as opposed to the traditional —

population.” 366 F. Supp. 3d at 153. They appear to divine this principle from the Supreme

Court’s decision in NFIB, where it found Congress’s decision to require states to expand

Medicaid unduly coercive under the Spending Clause. See 567 U.S. at 583–85. But the remedy

for this constitutional problem was simply to prevent the Secretary from withdrawing “existing

Medicaid funds for failure to comply with the requirements set out in the expansion.” Id. at 585.

After the decision, states were thus left with a choice: accept ACA funds and “comply with the

conditions on their use” or decline ACA funds and keep prior federal Medicaid appropriations.

21
Id.; see also id. at 586 (explaining that decision does not “affect the Secretary’s ability to

withdraw funds provided under the Affordable Care Act if a [s]tate that has chosen to participate

in the expansion fails to comply with the requirements of that Act”). Since the Act otherwise

places the expansion population on the same plane as the traditional population, states have no

“additional discretion” in how to comply with Medicaid requirements as to the expansion group.

See Stewart II, 366 F. Supp. 3d at 153.

While Defendants thus err in believing that their treatment of the expansion population is

undergirded with any greater discretion than their administration of any other part of the

Medicaid program, their argument about flexibility vis-à-vis the expansion population are

ultimately a red herring. That is because the entire Medicaid program is optional for states. The

Court does not see why — if Defendants are correct that threats to terminate the expansion

program can supply the baseline for the Secretary’s § 1115 review — that argument would not

be equally good as applied to traditional Medicaid. Id. at 153. Their argument must thus posit

that any § 1115 program that maintains any coverage for any set of individuals promotes the

objectives of the Medicaid Act as long as the state threatens to terminate all of Medicaid in the

absence of waiver approval. The second problem with this position, then, is that it has no

limiting principle. Under the Secretary’s reasoning, states may threaten to de-expand, or indeed

do away with all of, Medicaid if he does not approve whatever waiver of whatever Medicaid

requirements they wish to obtain. The Secretary could then always approve those waivers, no

matter how few people remain on Medicaid thereafter, because “any waiver would be coverage

promoting compared to a world in which the state offers no coverage at all.” Id. at 154. This

reading of the Act would give HHS practically unbridled discretion to implement the Medicaid

Act as “an à la carte exercise, picking and choosing which of Congress’s mandates it wishes to

22
implement.” Id. at 153–54. Apart from the potential constitutional concerns such an

interpretation would raise, cf. Clinton, 524 U.S. at 440–47 (1998), it clearly constitutes “an

impermissible construction of the statute . . . because [it] is utterly unreasonable in [its] breadth.”

Aid Ass’n for Lutherans v. U.S. Postal Serv., 321 F.3d 1166, 1178 (D.C. Cir. 2003); see also

Agape Church, Inc. v. FCC, 738 F.3d 397, 410 (D.C. Cir. 2013).

The third and final reason to reject this reading is perhaps the most important: it is

inconsistent with the text of § 1115. The statute requires the Secretary to evaluate whether the

project “is likely to assist in promoting the objectives” of the Act. See 42 U.S.C. § 1315.

Against what baseline is he supposed to evaluate the project? The structure of the waiver

provision assumes the implementation of the Act. It confirms that the relevant baseline is

whether the waiver will promote the objectives of the Act as compared to compliance with the

statute’s requirements, “not as compared with a hypothetical future universe” where the Act has

no force. Stewart II, 366 F. Supp. 3d at 154. This is so because the overarching provision

authorizing these waivers stipulates that, if the Secretary makes a judgment that a demonstration

promotes the objectives of the Act, he may then waive compliance with certain of its provisions

“to the extent and for the period . . . necessary” to carry out the project. See 42 U.S.C. § 1315(a),

(a)(1). That is, the provision contemplates a limited waiver from compliance with the Act’s

provisions. Particularly in light of what the Court has discussed above, understanding the

baseline as such is the only way this provision makes sense.

For these reasons, the Secretary cannot escape his obligation to consider whether Granite

Advantage poses a risk to coverage or is likely to increase coverage by emphasizing the optional

nature of the expansion or by citing New Hampshire’s plan to de-expand absent project approval.

23
 2. Other Objectives of the Medicaid Act

Defendants argue that, regardless of whether the Secretary properly considered Granite

Advantage’s effects on coverage, he reasonably approved the project on the ground that it is

likely to advance several other Medicaid objectives — namely, the health and financial

independence of beneficiaries and the fiscal sustainability of the safety net. HHS relied on those

same objectives when it approved Kentucky’s project the second time, and the Court addressed

them at length in Stewart II. Indeed, the approval letter reviewed in that decision — which

preceded New Hampshire’s by only ten days — tracks the one here practically verbatim with

respect to these objectives. Compare AR 1–6 with Kentucky II Approval Letter at 1–6. The

agency acknowledged as much during oral argument. See Tr. at 14–15. For the same reasons

the Court found that discussion wanting before, it finds it unpersuasive here.

a. Health

The Secretary asserts that “Granite Advantage is [] independently justified because the

Secretary found that it was likely to improve the health of Medicaid recipients.” Def. Cross-

Motion at 16. Recognizing that this Court has found that health is not a freestanding objective of

the Medicaid Act, the agency persists in the contrary view on the ground that “there is little

intrinsic value in paying for services if those services are not advancing the health and wellness

of individuals receiving them.” Id. This position does not change the Court’s mind. Assuming

the Secretary’s interpretation is entitled to Chevron deference, it fails at step two because it falls

outside “the bounds of reasonableness.” Abbott Labs v. Young, 920 F.2d 984, 988 (D.C. Cir.

1990). Medicaid, both as enacted and as later expanded by the ACA, reflects Congress’s desire

to “mak[e] healthcare more affordable” for “needy populations.” Stewart II, 366 F. Supp. 3d at

144. Congress therefore designed a scheme “to address not health generally but the provision of

24
care to needy populations.” Id. The Secretary cannot “extrapolate the objectives of the statute to

a higher level of generality and pursue that aim in the way he prefers.” Id. (citing Waterkeeper

Alliance v. EPA, 853 F.3d 527, 535 (D.C. Cir. 2017)).

An example outside of the health-care context helps illustrate the problem with

Defendants’ interpretive leap. Say that Congress passed an education statute designed to

encourage local control over education because it believed that decentralization was the best way

to improve the quality of education. Implementing this view, the statute appropriates money to

states that give local school districts control over their own management and curricula.

Disagreeing with Congress’s view about local control, however, the Secretary of Education

decides she would like to fund experimental state programs that require districts to adhere to

strict national education standards. She argues that this advances the purposes of the statute

because, in her expertise, national standards improve education, and education must have been

the statute’s goal since “there is little intrinsic value in paying” for school expenses “if those

services are not advancing” students’ education. See AR 1. This species of argument, of course,

would never fly, yet it is indistinguishable from CMS’s in this case, and it exposes the weakness

of HHS’s position. By defining a statute’s purposes up a level of generality, the Secretary can

justify actions as consistent with the law even if they are clearly at odds with it. This scenario

underscores that, as the Court explained in Stewart I, a statute’s objectives are often bound up

with the way Congress sought to solve a particular problem. See 313 F. Supp. 3d at 266–67.

Ignoring as much grants largely unbounded discretion to agencies, whose exercise of that

discretion can veer far afield from anything resembling the statute Congress wrote. Id. at 267–

68.

25
 In any event, even assuming Defendants were correct that health is a freestanding

objective of the Act, the agency fell short “because it did not consider the health benefits of the

project relative to its harms to the health of those who might lose their coverage.” Stewart II,

366 F. Supp. 3d at 145. As with the second Kentucky approval, commenters here made clear

that these health effects would be significant. See, e.g., AR 2131–32, 2223–24, 2242–43. The

Secretary, moreover, needed to weigh the net effects on health against the effects on coverage

more generally. Here he stumbled as well. His neglect to address these considerations is

independently fatal to Defendants’ argument that the project is supported by its effects on health

alone.

b. Financial Independence

Defendants have previously justified the Secretary’s approval on his judgment that

community-engagement requirements like those in Granite Advantage improve beneficiaries’

financial independence. See AR 4–5; see also Stewart II, 366 F. Supp. 3d at 145–46. This

argument does not appear with much force in the Government’s papers here. To the extent

Defendants still press it, the Court rejects it for the reasons discussed at length in Stewart I and

Stewart II — namely, that financial independence is not an independent objective of the Act and

that the Secretary has not in any event adequately assessed “the benefits of self-sufficiency” and

weighed them “against the consequences of coverage loss.” Stewart II, 366 F. Supp. 3d at 148.

c. Fiscal Sustainability

The agency’s principal argument this time around in favor of upholding the Secretary’s

decision is the same as it was in Stewart II: that he reasonably concluded that the project would

allow “New Hampshire to stretch its limited Medicaid resources.” AR 6. Granite Advantage

appears to enhance the fiscal sustainability of the State’s safety net, according to HHS, because

26
beneficiaries who were not previously working may transition to commercial coverage and

become healthier. In Stewart II, the Court agreed that fiscal sustainability was a valid

consideration in a Section 1115 project, but it found the Secretary’s explanation for why the

project addressed that concern to be arbitrary and capricious. See 366 F. Supp. 3d at 149–52. As

the agency recognized at oral argument, HHS’s explanation in approving Granite Advantage,

which is practically identical to what it said in Stewart II, likewise cannot clear the bar. See Tr.

at 14–15.

To start, HHS made no finding that Granite Advantage would save the state “any amount

of money or otherwise make the program more sustainable in some way.” Id. at 149. With

respect to savings, the Court is not suggesting that the Secretary “must quantify some exact

amount . . . , but he must make some finding that supports his conclusion that the project”

addresses New Hampshire’s fiscal concerns. Id. at 149–50. Two considerations make this

analysis especially important in this case.

First, the State has represented that it neither intends for the demonstration to reduce

costs nor expects it to do so. At oral argument, it explained that New Hampshire is not

encountering the same fiscal concerns as Kentucky with respect to its Medicaid program and that

reducing health costs is not in fact an objective of this demonstration project. See Tr. (Full

Transcript Pending). Consistent with that position, New Hampshire’s waiver application

projected “that spending growth in the future [under Granite Advantage] will be consistent with

standard growth rates experienced in the past.” AR 4399. The glaring disconnect between the

Secretary’s position and New Hampshire’s raises substantial questions about how the agency

came to believe the program would improve the State’s fiscal circumstances, underscoring the

need for reasoned analysis of this issue.

27
 Second, the record in this case contains substantial reasons to doubt whether the program

will save any money given administrative costs and the possible rise in uncompensated care that

would accrue to the State. See AR 1480, 1949, 2206, 2241–43, 2534, 2710–11; cf. 42 U.S.C.

§ 1396d(y)(1) (providing that federal government shall pay between 90 and 100 percent of costs

to expansion population). While the Secretary may well have arrived at a different conclusion,

he needed to explain how he got there in light of the nearly uniform evidence going the other

direction.

The agency does propose several potential mechanisms by which the program could save

the State money. But those do not advance the ball because they are conclusory and unsupported

by any evidence in the record. The Secretary never explained, for example, why he thinks that

the program will transition beneficiaries to commercial coverage, given the consistent evidence

before him that nearly all Medicaid recipients are already working, unable to work, or able to

find only low-paying jobs that do not offer or lead to commercial coverage. See AR 1949,

2209–10, 2225, 2435; see also id. at 3587 (explaining that work requirements only expected to

apply to 6% of able and non-working beneficiaries and that work opportunities for them are

limited). That the community-engagement requirements can be met through education or

volunteer activities, rather than employment, gives more reason to wonder why the Secretary

thought the program would expand access to commercial coverage. Likewise, the agency has

not explained the sort of health benefits it expected would accrue to beneficiaries as a result of

the new requirements and how such benefits would save the State money.

Apart from these failings, the Secretary’s fiscal-sustainability discussion suffers from

another key flaw: it did not “compare the benefit[s]” to the State’s safety net “to the

consequences for coverage.” Stewart II, 366 F. Supp. 3d at 150. The D.C. Circuit’s decision in

28
Pharmaceutical Research & Manufacturers of America v. Thompson, 362 F.3d 817 (D.C. Cir.

2004), helps illustrate how the agency fell short in this respect. There, the court upheld HHS’s

decision to impose a minor restriction on beneficiaries’ access to prescription drugs to try to

prevent borderline populations from becoming Medicaid eligible. Id. at 825. In doing so, the

D.C. Circuit discussed at length whether the agency had adequately considered the burden on

Medicaid recipients and reasonably explained why such imposition was necessary under the

circumstances. Id. at 825–26. This is just the kind of analysis the Secretary never conducted

here. He neither addressed the magnitude of coverage loss nor weighed that concern against the

asserted fiscal-sustainability benefits.

This analysis is essential given the admonitions in Thompson and the Supreme Court’s

decision in Pharmaceutical Research & Manufacturers of America v. Walsh, 538 U.S. 644

(2003) (plurality), that projects imposing significant burdens on Medicaid recipients may not be

consistent with the Act’s purposes. Id. at 664–65 (“The fact that the [Program] . . . provid[es]

benefits to needy persons and . . . curtail[s] the State’s Medicaid costs . . . would not provide a

sufficient basis for upholding the program if it severely curtailed Medicaid recipients’ access to

prescription drugs.”); Thompson, 362 F.3d at 826 (relying on “the absence of any demonstrable

significant impediment to Medicaid services”). As explained in Stewart II, “That there are limits

on the extent to which fiscal sustainability can justify cuts like those outlined in these cases

makes sense. Most cuts to Medicaid services would reduce the cost of Medicaid and thus

advance the sustainability of the program to some extent. But it would be nonsensical to

conclude that any cut therefore always promotes the Act’s objectives.” 366 F. Supp. 3d at 152.

In short, the Court finds that the Secretary’s assertion that Granite Advantage advances

the purposes of Medicaid because it would improve the fiscal sustainability of New Hampshire’s

29
Medicaid program is arbitrary and capricious, especially where the State at argument generally

disclaimed such motivation. In so concluding, the Court is not questioning the agency’s

predictive judgments or evaluating the evidence before it on this issue. It is simply looking for

what the APA requires: a reasoned explanation that considers the factors relevant to the agency’s

decision.

d. Data Collection

The Secretary has one more arrow in his quiver. He suggests that Granite Advantage

advances the objectives of Medicaid regardless of what effect it has on beneficiary health and

coverage because it helps the State and agency collect useful data for future policymaking

purposes. See Def. Cross-Mot. at 22–3; AR 12. This one holds no water. As a textual matter,

the Secretary is authorized to approve only those projects “likely to assist in promoting the

objectives” of the Medicaid Act. See 42 U.S.C. § 1315(a). But no one is suggesting with a

straight face that a purpose of the Medicaid Act is to collect data. A demonstration can hardly be

justified, therefore, solely on data-collection grounds. The practical consequences of the

Government’s suggestion are also alarming. If experimentation alone could justify a project,

then demonstrations with dire consequences for Medicaid beneficiaries could be approved just

for the Government to gather information. Recognizing these difficulties, HHS disclaimed any

such position at oral argument, confirming that any demonstration project must do more than

collect data: it must also advance the purposes of the Medicaid Act, including the core objective

of providing medical assistance to the needy. See Tr. at 14.

30
 Remedy

That leaves only consideration of the proper remedy. Three sets of issues require

attention.

First, the Court must decide whether, having arrived at the conclusions outlined above, it

should issue its decision now or instead wait until November, nearer in time to when the

community-engagement requirements are set to take effect. HHS, at oral argument, maintained

that the Court should stay its pen because vacating the program will substantially disrupt New

Hampshire’s implementation and outreach efforts and because the Court of Appeals may issue a

decision in the Kentucky and Arkansas cases in the interim. See Tr. at 4–6. Neither factor,

however, counsels in favor of delay. As to the former, the State of New Hampshire, which is

presumably best situated to understand the consequences of a timely decision, asks the Court to

issue its Opinion now rather than waiting until November. It explained at oral argument that a

decision would provide further clarity to the State while it compiles an amended waiver

application and considers how to implement its program moving forward. Id. (Complete

Transcript Pending). Citing the uncertainty attending their circumstances and those of other

Medicaid beneficiaries, Plaintiffs agree that the Court should not wait to act. As to the latter,

while it is possible that the Court of Appeals will issue its decision between now and November,

it is also possible that the Circuit’s decision will post-date the December 1 implementation. At

this point, no oral-argument date has been set. Either way, that consideration does not outweigh

New Hampshire’s and Plaintiffs’ interests in seeking a timelier decision.

Second, the Court must consider whether to remand this matter to the agency without

vacating its underlying approval of New Hampshire’s demonstration. When a court concludes

that agency action is unlawful, “the practice of the court is ordinarily to vacate the rule.” Ill.

31
Pub. Telecomms. Ass’n v. FCC, 123 F.3d 693, 693 (D.C. Cir. 1997); Sierra Club v. Van

Antwerp, 719 F. Supp. 2d 77, 78 (D.D.C. 2010) (“[B]oth the Supreme Court and the D.C. Circuit

Court have held that remand, along with vacatur, is the presumptively appropriate remedy for a

violation of the APA.”). Remand without vacatur may be appropriate, however, depending on

the “seriousness of the order’s deficiencies (and thus the extent of doubt whether the agency

chose correctly) and the disruptive consequences of an interim change.” Allied-Signal, Inc. v.

U.S. Nuclear Regulatory Comm’n, 988 F.2d 146, 150–51 (D.C. Cir. 1993) (citation omitted); see

also Standing Rock Sioux Tribe v. U.S. Army Corps of Engineers, 282 F. Supp. 3d 91, 103

(D.D.C. 2017) (declining to vacate when agency “largely complied” with statute and could likely

substantiate prior conclusions on remand).

In Stewart I, Stewart II, and Gresham, the Court concluded that both factors supported

vacatur. The Secretary’s failure to consider an objective of Medicaid is a “major shortcoming”

going “to the heart” of his decisions. See 313 F. Supp. 3d at 273. As to Kentucky HEALTH,

vacatur was not overly disruptive because the project had “yet to take effect” and the plaintiffs

could suffer “serious harm[s]” were Kentucky HEALTH allowed to be implemented pending

further proceedings. Id.; see also Stewart II, 366 F. Supp. 3d at 156. And as to Arkansas, while

vacatur would concededly have been disruptive given that the program had already begun in

part, the seriousness of this disruption — which was largely administrative in nature — had to be

“balanced against the harms that Plaintiffs and persons like them will experience if the program

remains in effect.” Gresham, 363 F. Supp. 3d at 184.

Defendants do not dispute that vacatur is appropriate under this analysis. See HHS

Cross-Mot. at 27–29 (declining to argue that Allied-Signal factors counsel against vacatur). And

for good reason. Like the prior three approvals, this one suffers from a significant deficiency:

32
the failure to address a central factor in its decision. And vacatur will not be terribly disruptive,

given that New Hampshire has not fully implemented the community-engagement requirements

— indeed, the State has continued to make legislative and executive tweaks to the program since

CMS approved it in late 2018.

Third, HHS argues that rather than vacating the November 2018 approval as a whole, this

Court should tailor any relief solely to the four Plaintiffs and the aspects of the program that they

have successfully challenged. Id. at 27–29. The Court is no more persuaded by this contention

than it was in Stewart II. See 366 F. Supp. 3d at 155. In an APA case, the “ordinary result” of

the Court’s finding an agency action unlawful is to vacate that action — not to judicially re-write

what the agency did so that it somehow does not apply to a narrow group of people or so that it

persists piecemeal. See Harmon v. Thornburgh, 878 F.2d 484, 495 n.21 (D.C. Cir. 1989). The

Court of Appeals, accordingly, has explained that “if the plaintiff prevails” in challenging an

agency action under the APA, “the result is that the rule is invalidated, not simply that the court

forbids its application to a particular individual.” Nat’l Min. Ass’n v. U.S. Army Corps of

Engineers, 145 F.3d 1399, 1409 (D.C. Cir. 1998) (quoting Lujan v. Nat’l Wildlife Fed., 497 U.S.

871, 913 (1990) (Blackmun, J., dissenting)).

Defendants offer no reason to deviate from the settled manner in which courts accord

relief in APA cases; indeed, they do not cite a single APA case in which relief has been granted

in this manner. Plaintiffs have standing to challenge and have successfully challenged the

Secretary’s November 30, 2018, approval of the amendments to Granite Advantage. The proper

relief, under these circumstances, is vacatur of that action.

IV. CONCLUSION

For the foregoing reasons, the Court will grant Plaintiffs’ Motion for Summary Judgment

33
and deny Defendants’ Cross-Motion and Motion to Dismiss. A separate Order consistent with

this Opinion will issue this day.

/s/ James E. Boasberg
JAMES E. BOASBERG
United States District Judge

Date: July 29, 2019

APPENDIX A

Comments on Coverage Consequences of Community-Engagement Requirements
AR 1484 (Disability Rights Center) (“Red tape and paperwork requirements have been shown
to reduce enrollment in Medicaid and individuals coping with serious mental illness or
physical impairments may face particular difficulty meeting these requirements.”); AR 1489
(American Cancer Society Cancer Action Network) (“If work and community engagement is
required as a condition of eligibility, many cancer patients, recent survivors, and those with
other chronic illnesses could find that they are ineligible for the lifesaving care and treatment
services provided through the state’s Medicaid program . . . The increase in administrative
requirements for enrollees to attest to their working status would likely decrease the number of
individuals with Medicaid coverage, regardless of whether they are exempt.”); AR 1949
(Rights and Democracy NH) (“[I]mposing work requirements for people who need Medicaid
Expansion to gain access to health care will likely cause many people to lose coverage and
increase levels of uncompensated care.”); AR 1953 (Cystic Fibrosis Foundation) (“[M]any
people may have trouble complying with new eligibility requirements and for someone with
[cystic fibrosis], this could result in a life-threatening gap in coverage.”); AR 1956 (NH
Community Behavioral Health Association) (“We fear that the work requirement has the very
real potential to jeopardize care for individuals with mental illness if they lose their Medicaid
coverage. . . . Asking people struggling with mental illness to document their work by keeping
track of every week’s pay stubs is an onerous requirement.”); AR 2132 (American Diabetes
Association) (“New Hampshire’s proposal to limit access to Medicaid services through the
implementation of work requirements will decrease access to care for low-income New
Hampshire residents with and at risk for diabetes and increase state health care costs.”); AR
2201 (Lung Cancer Alliance) (citing estimates that similar Iowa waiver caused that “more than
3,000 beneficiaries” to lose coverage and “become uninsured”); AR 2208 (Center on Budget
and Policy Priorities & Georgetown University Center for Children and Families) (“Kaiser
conservatively estimates that between 25 and 50 percent of such enrollees are at risk of losing
coverage under work requirements. New Hampshire’s work requirement is especially onerous,
because just one month of non-compliance can lead to loss of coverage.”); AR 2224–25
(American Heart Association) (asserting that work requirements will reduce “access to
healthcare services both in the short and long term” for people with cardiovascular disease);
AR 2238 (National Council on Aging) (“Extending work requirements would also particularly
hurt the rural residents of NH,” who make up 48% of NH’s Medicaid population, because they

34
are less likely to have access to the internet or transportation and therefore “risk losing
coverage.”); AR 2241 (Center for Law and Social Policy) (“The evidence from other programs
with similar requirements is that in spite of official exemptions, in practice, individuals with
disabilities are often not exempted and are more likely to lose benefits.”); AR 2434 (CLASP
Volatile Job Schedules and Access to Public Benefits Report) (finding that tying benefits to
work requirements, especially a specific number of hours, can be problematic because “more
than 40 percent of early career hourly-workers (ages 26 to 32) receive one week or less
advance notice of their job schedules. Half of these workers have no input into their schedules
and three-quarters experience fluctuations in the number of hours they work, with hours
varying by more than eight hours per week on average.”); AR 2530 (Kaiser Family Foundation
Issue Brief) (considering national effects of a Medicaid work requirement and concluding that,
“[o]verall, among the 23.5 million non-SSI, non-dual, nonelderly Medicaid adults,
disenrollment ranges from 1.4 million to 4.0 million under the scenarios considered”); AR
2563 (New Futures) (stating that results of Arkansas’s similar program will likely reflect those
of NH’s program, and “[i]n the first month of implementation of the Arkansas Works work
requirement (June 2018) fewer than six percent of the nearly 8,000 Medicaid enrollees who
did not declare an exemption were able to satisfy the reporting requirement”); AR 2575
(Families USA) (“The coverage losses [caused by the work requirement] will result in an
increase in the state’s uninsured population, lost health care access, and worse health for low-
income adults in New Hampshire.”); AR 2696 (National Health Law Program) (“All evidence
indicates that the work requirement will lead to substantial numbers of individuals losing
Medicaid coverage. In Kentucky, which proposed a similar work requirement, researchers
have estimated that between 45,000 to 103,000 individuals could lose coverage due to the
work requirement alone.”); AR 2963 (New Hampshire Fiscal Policy Institute) (“Work
requirements implemented in Arkansas for certain populations starting in June 2018 . . .
provide an initial indication that similar requirements may reduce enrollment significantly in
New Hampshire.”); AR 3372 (Kaiser Family Foundation) (“Under the Medicaid work
requirement programs, the population subject to Medicaid work requirements may have access
to only low-wage, unstable, or low-quality jobs to meet the weekly/monthly hours
requirement[;] . . . ‘[p]olicies that promote job growth without giving attention to the overall
adequacy of the jobs may undermine health.’”); AR 3406 (Urban Institute) (“The red tape
associated with work requirements can cause people to lose access to vital supports even when
they are working or should be exempt from the requirements.”); AR 3584 (Philip Rocco, PhD)
(citing a study about “a Florida welfare reform experiment whose benefits were conditioned on
workforce participation [and] had a 16 percent higher mortality rate than comparable
recipients of welfare who were not subject to work stipulations”); AR 3644 (New Hampshire
Legal Assistance) (“All evidence indicates that the work requirement will lead to substantial
numbers of individuals losing Medicaid coverage.”); AR 4563–64 (Leukemia & Lymphoma
Society) (“New Hampshire’s proposal to extend its work requirement will perpetuate a return
to increased bureaucracy and paperwork and, in turn, coverage losses.”).

35