New Mexico Health Connections v. HHS

946 F.3d 1138

Opinion

PUBLISH FILED United States Court of Appeals UNITED STATES COURT OF APPEALS Tenth Circuit

FOR THE TENTH CIRCUIT December 31, 2019

_________________________________

Elisabeth A. Shumaker

Clerk of Court

NEW MEXICO HEALTH CONNECTIONS, a New Mexico nonprofit corporation,

Plaintiff - Appellee, v. No. 18-2186

UNITED STATES DEPARTMENT OF HEALTH & HUMAN SERVICES; CENTERS FOR MEDICARE AND MEDICAID SERVICES; ALEX M. AZAR, II, Secretary of the United States Department of Health and Human Services, in his official capacity; SEEMA VERMA, Administrator for the Centers for Medicare and Medicaid Services, in her official capacity,

Defendants - Appellants.

------------------------------------------

AMERICA'S HEALTH INSURANCE PLANS; BLUE CROSS BLUE SHIELD ASSOCIATION,

Amicus-Curiae.

_________________________________

Appeal from the United States District Court for the District of New Mexico (D.C. No. 1:16-CV-00878-JB-JHR)

_________________________________

Joshua Revesz, U.S. Department of Justice, Washington, D.C. (Joseph H. Hunt, Assistant Attorney General, John C. Anderson, United States Attorney, Alisa B. Klein, U.S. Department of Justice, Washington, D.C.; Robert P. Charrow, General Counsel, Kelly M. Cleary, Deputy General Counsel, H. Antony Lim, Jullia Callahan Bradley, Attorneys, U.S. Department of Health & Human Services, Washington D.C., with him on the briefs), for Defendants – Appellants.

Barak A. Bassman, Pepper Hamilton LLP, Philadelphia, Pennsylvania (Sara B. Richman, Leah Greenberg Katz, Pepper Hamilton LLP, Philadelphia, Pennsylvania; Marc D. Machlin, Pepper Hamilton LLP, Washington, D.C.; Nancy R. Long, Long, Komer & Associates, P.A., Santa Fe, New Mexico, with him on the brief), for Plaintiff – Appellee.

Julie Simon Miller, Thomas M. Palumbo, America’s Health Insurance Plans, Washington, D.C.; W. Scott Nehs, Blue Cross Blue Shield Association, Chicago, Illinois; Pratik A. Shah, Z.W. Julius Chen, Akin Gump Strauss Hauer & Feld LLP, Washington, D.C., filed an amicus curiae brief on behalf of Amici Curiae.

_________________________________

Before LUCERO, HARTZ, and MATHESON, Circuit Judges.

_________________________________

MATHESON, Circuit Judge.

_________________________________

In 2010, Congress passed the Patient Protection and Affordable Care Act (“ACA”) to “increase the number of Americans covered by health insurance and decrease the cost of health care.” Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519, 538 (2012); see ACA, Pub. L. No. 111-148, 124 Stat. 119 (2010) (codified primarily in title 42 of U.S.C.). Among its reforms, the ACA required private health insurers to provide coverage for individuals regardless of their gender or health status, including preexisting conditions. See 42 U.S.C. §§ 300gg-3, 300gg-4. It also

3

established “[h]ealth [b]enefit [e]xchanges” where individuals and small groups can purchase health insurance. Id. § 18031(b)(1). 1 Congress anticipated these reforms might hamper the ability of insurers to predict health care costs and to price health insurance premiums as more individuals sought health insurance. See Standards Related to Reinsurance, Risk Corridors and Risk Adjustment, 77 Fed. Reg. 17,220, 17,221 (Mar. 23, 2012) (codified at 45 C.F.R. pt. 153) (“Stabilization Rule”). It also anticipated insurers might refuse to provide insurance plans on the exchanges if they could not reasonably estimate their potential costs. See id. 2 To spread the risk of enrolling people who might need more health care than others, Congress established a risk adjustment program for the individual and small group health insurance markets. See 42 U.S.C. § 18063. 3 The program transfers

1 Health benefit exchanges “make available qualified health plans to qualified individuals and qualified employers” in a state. 42 U.S.C. § 18031(d)(2)(A).

2 Health insurance plans are “a defined set of benefits” offered by a health insurer to individuals seeking coverage of health care expenses. U.S. Dep’t of Health & Human Servs., Glossary of Terms (Nov. 18, 2019), https://perma.cc/GZ4D-DLHR. An insurer can offer multiple plans.

3 The “‘individual market’ means the market for health insurance coverage offered to individuals other than in connection with a group health plan.” 42 U.S.C. § 18024(a)(2). The “small group market” means “the health insurance market under which individuals obtain health insurance coverage . . . through . . . a small employer.” Id. § 18024(a)(3).

4

funds from plans with healthier enrollees to plans with sicker enrollees. A goal of the program is to discourage insurers from avoiding enrollment of sicker enrollees. 4 Congress tasked the Department of Health and Human Services (“HHS”) with designing and implementing this risk adjustment program with the states. Id. § 18063(b). HHS developed a formula to calculate how much each insurer would be charged or paid in each state. The formula relied on the “statewide average premium”—the average of all applicable premiums insureds pay to health insurers in a state—to calculate charges and payments.

Plaintiff-Appellee New Mexico Health Connections (“NMHC”), an insurer that was required to pay charges under the program, sued the HHS Defendants-Appellants 5 under the Administrative Procedure Act (“APA”). NMHC alleged that HHS’s use of the

4 In addition to the permanent risk adjustment program, the ACA established two other stabilization programs—risk corridors and reinsurance—which ran from 2014 to 2016. See 42 U.S.C. §§ 18061, 18062. “These [three] programs [were] designed to provide consumers with affordable health insurance coverage, to reduce incentives for health insurance issuers to avoid enrolling sicker people, and to stabilize premiums. . . .” Ctrs. for Medicare & Medicaid Servs., Premium Stabilization Programs (Nov. 18, 2019), https://perma.cc/U7S9-8PRU. Only the risk adjustment program is at issue here.

5 NMHC sued HHS; Centers for Medicare and Medicaid Services (“CMS”);

Alex M. Azar II, Secretary of HHS; and Seema Verma, Administrator of CMS. CMS is a part of HHS and administers the risk adjustment program. We refer to these Defendants-Appellants collectively as “HHS.”

5

statewide average premium to calculate charges and payments in New Mexico from 2014 through 2018 was arbitrary and capricious. 6 The district court granted summary judgment to NMHC, holding that HHS violated the APA by failing to explain why the agency chose to use the statewide average premium in its program. See N.M. Health Connections v. U.S. Dep’t of Health & Human Servs. (NMHC I), 312 F. Supp. 3d 1164, 1207-13 (D.N.M. 2018). The court faulted HHS for “erroneously read[ing] the ACA’s risk adjustment provisions to require” budget neutrality, which “infect[ed] [HHS’s] analysis of the relative merits of using a state’s average premium when calculating risk adjustment transfers instead of using a plan’s own premium.” Id. at 1209. It remanded to the agency and vacated the 2014, 2015, 2016, 2017, and 2018 rules that implemented the program. After the district court denied HHS’s motion to alter or amend judgment under Federal Rule of Civil Procedure 59(e), HHS appealed.

Exercising jurisdiction under 28 U.S.C. § 1291:

1. We hold NMHC’s claims regarding the 2017 and 2018 rules are moot, so we remand to the district court to vacate its judgment on those claims and dismiss them as moot.

2. We reverse the district court’s grant of summary judgment to NMHC as to the 2014, 2015, and 2016 rules. HHS

6 A health insurance premium is the “periodic payment (e.g. monthly, quarterly) . . . paid by a policyholder for insurance coverage.” U.S. Dep’t of Health & Human Servs., Glossary of Terms (Nov. 18, 2019), https://perma.cc/GZ4D-DLHR. We refer interchangeably to individuals who are covered by a health plan as “enrollees” and “insureds.”

6

acted reasonably in explaining why it used the statewide average premium in the formula.

Because we reverse the district court on its summary judgment ruling in favor of

NMHC, we need not address the denial of HHS’s Rule 59(e) motion.

I. BACKGROUND

This section describes the ACA’s risk adjustment program and summarizes the factual and procedural history of the case.

A. The ACA’s Risk Adjustment Program Purpose Congress included the risk adjustment program in the ACA to stabilize health insurance premiums, encourage health insurers to provide plans on the exchanges, and discourage insurers from eluding enrollment of sicker individuals. See 2014 Final Rule, 78 Fed. Reg. 15,410, 15,411 (Mar. 11, 2013), amended by 2014 Final Rule, 78 Fed. Reg. 65,046 (Oct. 20, 2013). The roles of “premiums” and “risk” in the health insurance market inform this purpose.

Enrollees pay premiums to receive health insurance coverage. Insurers set their premiums “based on the anticipated revenue needs for their enrolled population.” Ctr. for Consumer Info. & Ins. Oversight, Risk Adjustment Implementation Issues, 13 (Sept. 12, 2011) (“2011 White Paper”). Premiums differ from one plan to another for various reasons, including different estimated health care needs of each plan’s enrollees and the potential costs for those needs. See John Kautter et. al., Affordable Care Act Risk Adjustment: Overview, Context, and Challenges, 4 Medicare & Medicaid Res. Rev. 1, 5

7

(2014) (“Kautter Article”). Premiums also reflect a plan’s benefits and efficiency. See id. at 3.

In setting insurance premiums, health insurers consider the risk of loss they might face from providing coverage to their enrollees. Risk is the probability that an insured event will occur, requiring the insurer to pay for it. 7 Among other sources of risk, the ACA reforms have exposed insurers to risk of financial loss due to adverse selection, which occurs when individuals who anticipate high health care needs are more likely to purchase coverage than those who anticipate low health care needs. See Stabilization Rule, 77 Fed. Reg. at 17,221. This can result in “a health plan having higher costs than anticipated.” Id.

Before the ACA, insurers could limit their risk by adjusting premiums based on the age, gender, and health status of their enrollees. See Katherine M. Kehres, Cong. Research Serv., R45334, The Patient Protection and Affordable Care Act’s Risk Adjustment Program: Frequently Asked Questions ii (Oct. 4, 2018) (“CRS Report”). They also “could deny coverage if an individual represented too much risk.” Id. at 3. The ACA changed that. It forbids insurers from refusing to cover individuals with preexisting conditions and from “set[ting] premiums based on

7 “In health insurance parlance, ‘risk’ sometimes means simply ‘cost,’ with the implication that the realization of healthcare cost for an individual is uncertain.” Thomas G. McGuire & Richard C. van Kleef, Risk Sharing, in Risk Adjustment, Risk Sharing, and Premium Regulation in Health Insurance Markets 105 (Thomas G. McGuire & Richard C. van Kleef eds., 2018).

8

gender or health status.” Id. at 1. An insurer “that enrolls a larger proportion of sicker (i.e., high-risk) enrollees than other plans in the market would [therefore] need to charge” a higher premium to be financially viable. Id. at 7.

The risk adjustment program aims to (1) reduce an insurer’s incentive to enroll only low-risk individuals, (2) encourage insurers to stay in the market, and (3) enable insurers to set premiums based on plan design and benefits rather than the health risk of enrollees in the plan. See id. at ii, 1, 7. It seeks to mitigate the impact of these reforms by subsidizing certain insurers for covering high-risk individuals without compensating them for other plan differences included in the price of their premiums. See Kautter Article at 3; Purva H. Rawal, The Affordable Care Act 161 (2016). HHS devised a formula that calculates “payment transfers . . . to help cover [plans’] actual risk exposure beyond the premiums the plans” can charge under the ACA. 2014 Final Rule, 78 Fed. Reg. at 15,430.

Statutory Basis Section 1343 of the ACA established the risk adjustment program. See id. at 15,415. Codified at 42 U.S.C. § 18063, the statute provides:

(a) In general (1) Low actuarial risk plans

Using the criteria and methods developed under subsection (b), each State shall assess a charge on health plans and health insurance issuers (with respect to health insurance coverage) described in subsection (c) if the actuarial risk of the enrollees of such plans or coverage for a year is less

9

than the average actuarial risk of all enrollees in all plans or coverage in such State for such year that are not self-

insured group health plans (which are subject to the provisions of the Employee Retirement Income Security Act of 1974).

(2) High actuarial risk plans

Using the criteria and methods developed under subsection (b), each State shall provide a payment to health plans and health insurance issuers (with respect to health insurance coverage) described in subsection (c) if the actuarial risk of the enrollees of such plans or coverage for a year is greater than the average actuarial risk of all enrollees in all plans and coverage in such State for such year that are not self-insured group health plans (which are subject to the provisions of the Employee Retirement Income Security Act of 1974).

(b) Criteria and methods

The Secretary, in consultation with States, shall establish criteria and methods to be used in carrying out the risk adjustment activities under this section. The Secretary may utilize criteria and methods similar to the criteria and methods utilized under part C or D of title XVIII of the Social Security Act. Such criteria and methods shall be included in the standards and requirements the Secretary prescribes under section 18041 of this title.

(c) Scope

A health plan or a health insurance issuer is described in this subsection if such health plan or health insurance issuer provides coverage in the individual or small group market within the State. This subsection shall not apply to a grandfathered health plan or the issuer of a grandfathered health plan with respect to that plan.

10

The ACA directed the Secretary of HHS, in consultation with the states, to “establish criteria and methods” to implement this program. 42 U.S.C. § 18063(b). A state can carry out its own approved program or HHS will do so on its behalf. See 45 C.F.R. § 153.310(a). Only Massachusetts has managed its own program, but it ceased doing so in the 2017 benefit year. HHS has always managed New Mexico’s program and currently operates the program in all states.

The statute does not authorize any legislatively appropriated funding for this program. Compare 42 U.S.C. § 18063 (not authorizing appropriations for ACA’s risk adjustment program) with id. § 18042(g) (authorizing appropriations in ACA to provide loans to qualified health insurance issuers).

Mechanics HHS implemented the risk adjustment program on behalf of states through rules promulgated in separate notice-and-comment proceedings for the 2014, 2015, 2016, 2017, and 2018 benefit years. 8 Each succeeding rule employed the same methodology as the previous rules. See 2015 Final Rule, 79 Fed. Reg. 13,744, 13,753 (Mar. 11, 2014) (“We proposed to use the [2014] methodology in 2015 . . . .”); 2016 Final Rule, 80 Fed. Reg. 10,750, 10,760 (Feb. 27, 2015), corrected by 2016 Final Rule, 80 Fed. Reg. 38,652 (July 7, 2015) (“We proposed to continue to

8 As used in the risk adjustment program, a “‘benefit year’ means a calendar year for which a health plan provides coverage for health benefits.” 45 C.F.R. § 155.20.

11

use the same risk adjustment methodology finalized in the 2014 [rule] . . . .”); 2017 Final Rule, 81 Fed. Reg. 12,204, 12,217 (Mar. 8, 2016) (same); 2018 Final Rule, 81 Fed. Reg. 94,058, 94,100 (Dec. 22, 2016) (“The payment transfer formula is unchanged from what was finalized in the 2014 [rule] . . . .”).

To implement the program, HHS took the following steps for each benefit year:

1. Published a proposed rule approximately one year before the applicable benefit year and, after a public comment period, published the final rule a few months later so insurers could rely on it to price their plans. See Supp.

App. at 330; 45 C.F.R. § 153.100; id. § 153.320(b).

2. Collected enrollment and claims data from insurers. See 45 C.F.R. § 153.70; see also CRS Report at 7.

3. Calculated a “risk score,” which estimated the cost of each enrollee in each plan based on that person’s age, sex, and diagnoses. See, e.g., Ctr. for Consumer Info. & Ins.

Oversight, HHS-Operated Risk Adjustment Methodology Meeting, 5-6 (Mar. 24, 2016) (“2016 White Paper”); 9 CRS Report at 7.

4. Averaged the individual risk scores to calculate a “plan liability” risk score, which estimated an insurer’s costs for each of its plans. See 2016 White Paper at 5-6, 11; see also Kautter Article at 8; CRS Report at 7. HHS applied this step to each plan offered by an insurer. 10

9 The 2016 White Paper is not included in the appendices on appeal. We consider it, however, because it was included in the administrative record before the district court. See Dist. Ct. Doc. 25, Ex. A at 4.

10 The risk adjustment program calculates charges and payments on a plan-byplan basis. See 42 U.S.C. § 18063. Because a health insurer could offer multiple plans, it could be paid for some plans and charged for others. See, e.g., Ctrs. for

12

5. Calculated how much plans would be charged or paid by applying to each plan a payment transfer formula, which included an estimated premium based on the “plan liability” risk score and other plan-specific cost factors (such as benefit design). See 2014 Final Rule, 78 Fed.

Reg. at 15,417, 15,431; 2016 White Paper at 6, 11; see also Kautter Article at 8.

6. Published the risk adjustment transfers, collected charges from insurers that covered healthier enrollees, and paid insurers who covered sicker enrollees. See 45 C.F.R.

§ 153.310; see also CRS Report at 8.

Payment Transfer Formula and Statewide Average Premium The payment transfer formula referred to in step 5 is the focus of this case. 11 The basic concept is that a plan’s transfer is the difference between (1) its estimated premium

Medicare & Medicaid Servs., Summary Report on Permanent Risk Adjustment Transfers for the 2018 Benefit Year, 13-33 (June 28, 2019), https://perma.cc/J9D9- 9G7S.

11 As first published in the 2014 Final Rule (and later incorporated in the 2015, 2016, 2017, and 2018 rules), the formula was:

2014 Final Rule, 78 Fed. Reg. at 15,431.

13

given the health risk of its enrollees and (2) its estimated premium without the risk. See 2014 Final Rule, 78 Fed. Reg. at 15,430; see also Gregory C. Pope et al., Risk Transfer Formula for Individual and Small Group Markets Under the Affordable Care Act, 4 Medicare & Medicaid Res. Rev. 3 (2014) (“Pope Article”). “Both of these premium estimates are based on the [s]tate average premium.” 2014 Final Rule, 78 Fed. Reg. at 15,430. The 2014 Final Rule distilled the formula as follows:

See id. at 15,431.

If the transfer number is positive, a plan receives a payment. See 2016 White Paper at 11. If it is negative, a plan is charged. See id. “Transfers are intended to bridge the gap between these two premium estimates.” 2014 Final Rule, 78 Fed. Reg. at 15,430. The transfers compensate health insurance plans for covering less healthy enrollees while preserving permissible premium differences. 2016 White Paper at 11.

The risk adjustment rules made modest changes over the years. See 2016 White Paper at 31-34. For the purposes of the APA challenge here, the program did not materially change from 2014 to 2018. We therefore rely on the description of the program stated in the 2014 Final Rule.

14

HHS performed these calculations for every plan offered by every insurer in the state and then used the results to determine which insurers would be charged and how much, and which insurers would be paid and how much. See 2014 Final Rule, 78 Fed. Reg. at 15,431-32. The total amount charged minus the total paid by HHS “net[ted] to zero.” Id. at 15,516.

NMHC challenged only one aspect of this program: HHS’s use of the statewide average premium rather than each plan’s own premium in its payment transfer formula. See Aplee. Br. at 19. The former is an average of all the applicable health insurance premiums in a state. See 2014 Proposed Rule, 77 Fed. Reg. 73,118, 73,126 (Dec. 7, 2012); 2014 Final Rule, 78 Fed. Reg. at 15,431-32. The latter would be the premiums that the plans set themselves. HHS uses the statewide average premium in the formula to convert the prior calculations into a dollar amount that equals the transfer charge or payment. See 2016 White Paper at 81.

B. Agency and District Court Proceedings We present in chronological order the agency and district court proceedings leading to this appeal. As this presentation shows, the agency and court proceedings overlapped towards the end.

We describe:

1. HHS’s preparation for and its conduct of the notice-andcomment proceedings that led to the 2014 risk adjustment rule;

15

2. the agency’s notice-and-comment proceedings that led to the 2015, 2016, and 2017 rules;

3. NMHC’s 2016 district court complaint;

4. the agency’s proceedings that led to the 2018 rule;

5. NMHC and HHS’s cross-motions for summary judgment and the district court’s ruling on those motions;

6. HHS’s Rule 59(e) motion challenging the district court’s summary judgment order;

7. the agency’s proceedings, in response to the district court’s summary judgment order, that led to modifications to the 2017 and 2018 rules; and

8. the district court’s denial of HHS’s Rule 59(e) motion.

Notice-and-Comment Proceedings for the 2014 Rule HHS began its work on the risk adjustment program shortly after the ACA’s enactment. Before developing the details of the program, HHS promulgated a rule establishing general standards for the stabilization programs, including the risk adjustment program. 12 See Stabilization Rule, 77 Fed. Reg. at 17,220-52.

12 In the rulemaking process, HHS discussed the general features of the risk adjustment program, noting “risk adjustment is designed as a budget neutral activity.” Standards Related to Reinsurance, Risk Corridors and Risk Adjustment, 76 Fed. Reg. 41,930, 41,938 (July 15, 2011) (“Proposed Stabilization Rule”). No commenter addressed the adequacy of HHS’s explanation of either budget neutrality or the statewide average premium in this rulemaking proceeding. See id. at 41,937-40; Stabilization Rule, 77 Fed. Reg. at 17,220-52.

16

In late 2011, the Center for Consumer Information and Insurance Oversight (“CCIIO”), an agency within HHS charged with helping to implement ACA reforms, published a white paper on risk adjustment concepts to “begin the consultation process around the development of the [f]ederally-certified risk adjustment methodology developed by HHS and provide context for individuals to submit comments.” 2011 White Paper at 3; see 2014 Proposed Rule, 77 Fed. Reg. at 73,122. In a bulletin published in early 2012 (“2012 Bulletin”), HHS outlined its plan for implementing the program on behalf of a state. See 2014 Proposed Rule, 77 Fed. Reg. at 73,122. In mid-2012, HHS also hosted a public meeting to discuss the plan. See id.

On December 7, 2012, HHS proposed the first risk adjustment rule for benefit year 2014. 2014 Proposed Rule, 77 Fed. Reg. at 73,118-218. It solicited and received comments for the rule and published the final rule on March 11, 2013 (“the 2014 rule”). See 2014 Final Rule, 78 Fed. Reg. at 15,410-541.

We describe the foregoing in greater depth below, focusing on HHS’s decision to use the statewide average premium in a risk adjustment program where total charges would equal total payments—that is, budget neutrality.

a. 2011 White Paper

In the 2011 White Paper, CCIIO considered four alternatives to establish a “baseline premium” for the payment transfer formula: “weighted state average premiums,” “weighted rating area average premiums,” “actuarial value-adjusted

17

weighted average premiums,” and “plans’ own premiums.” 2011 White Paper at 14. 13 Because NMHC argued only in favor of using the plans’ own premiums, we focus on HHS’s choice to use the statewide average premium over the plans’ own premiums.

CCIIO noted in the 2011 White Paper that using the statewide average premium “would result in balanced payments and charges” and “could be calculated with or without adjustment.” Id. It expressed concern that using the plans’ own premiums would “create disincentives” for some plans because “the amount of charges and payments would be affected by each plan’s premiums.” Id. It explained that “[f]or plans with a sicker than average risk mix, a lower premium plan would receive less in payments than a higher premium plan, even if the two plans have the same risk level.” Id. CCIIO feared this “could create disincentives for high-risk plans to operate efficiently or set lower prices.” Id. It also noted additional calculations would be needed for all options “except for the [s]tate average,” id. at 15, meaning that using the statewide average premium would automatically make transfers net to zero whereas using plans’ own premiums would require additional adjustments.

13 The parties included only an excerpt of the 2011 White Paper in their appendices. We consider the entire 2011 White Paper because the agency relied on it in promulgating its 2014 rule and it was included in the administrative record below. See Dist. Ct. Doc. 25, Ex. A at 2.

18

CCIIO sought comments on each of the four alternatives and whether it should consider other alternatives. See id. at 16-17. It also provided an appendix with “examples to examine implications of the approaches to calculate and balance payments and charges in greater detail.” Id. at 16, 29-56.

b. 2012 Bulletin HHS described its key findings about operating the risk adjustment program in a 2012 bulletin—an informational guidance document. U.S. Dep’t of Health & Human Servs., Bulletin on the Risk Adjustment Program: Proposed Operations by the Department of Health and Human Services, 3-5 (May 1, 2012). It emphasized how permissible plan differences affect plan premiums. Id. HHS said it would need to control these different factors to ensure “payment transfers compensate for liability differences due to health status” not for other permissible differences. Id. at 5. HHS also noted its payment methodology “need[ed] to be balanced” and acknowledged this budget-neutral design “has significant implications for the equation that is used for calculating payments.” Id.

c. 2014 Rule In its proposed 2014 rule, HHS explained its decision to use the statewide average premium in the formula in a section titled “Rationales for a Transfer Methodology Based on State Average Premiums.” See 2014 Proposed Rule, 77 Fed. Reg. at 73,139. It said the statewide average premium “provides a straightforward and predictable benchmark for estimating transfers.” Id. Using the plans’ own

19

premiums “could introduce differences in premiums across plans that were not consistent with features of the plan.” Id. HHS thus sought to balance predictability with “preserving premium differences.” Id.

No commenter challenged HHS’s decision to use the statewide average premium for the 2014 rule. See 2014 Final Rule, 78 Fed. Reg. at 15,430-32. Some comments supported HHS’s choice to do so. See 2014 Final Rule, 78 Fed. Reg. at 15,432. For example, HHS noted, “We received a number of comments in support of our proposal to use the [s]tate average premium as the basis for risk adjustment transfers. One commenter suggested that use of a plan’s own premium may cause unintended distortions in the transfer formula.” Id. In its responses to those comments, HHS said

The goal of the payment transfer formula is, to the extent possible, to promote risk-neutral premiums. We agree with commenters that use of a plan’s own premium may cause unintended distortions in transfers. . . . We are finalizing our proposal to base the payment transfer formula on the [s]tate average premium.

Id. No commenter challenged HHS’s decision to design its program as budget neutral such that transfer payments “net to zero.” See 2014 Proposed Rule, 77 Fed. Reg. at 73,139; 2014 Final Rule, 78 Fed. Reg. at 15,417-36.

The risk adjustment program in the 2014 rule thus used the statewide average premium as the baseline in the formula. The rule provided that transfer payments and charges would be budget neutral, meaning HHS would not provide any additional

20

funds to pay for the transfers. The program developed in the 2014 rule became the blueprint HHS followed in promulgating the rules for the ensuing benefit years.

Notice-and-Comment Proceedings for the 2015, 2016, and 2017 Rules HHS adopted a similar risk adjustment program in its 2015, 2016, and 2017 rules. Because the rules used the statewide average premium in the payment transfer formula and built upon prior proceedings, the 2015, 2016, and 2017 rules relied on the same reasoning as the 2014 rule. See 2015 Final Rule, 79 Fed. Reg. at 13,753 (“We proposed to use the [2014] methodology in 2015 . . . .”); 2016 Final Rule, 80 Fed. Reg. at 10,760 (“We proposed to continue to use the same risk adjustment methodology finalized in the 2014 [rule] . . . .”); 2017 Final Rule, 81 Fed. Reg. at 12,217 (same).

a. 2015 Rule On December 2, 2013, HHS published a proposed rule for the 2015 benefit year. 2015 Proposed Rule, 78 Fed. Reg. 72,322, 72,322-92 (Dec. 2, 2013). On March 11, 2014, it published the final rule for 2015 (“the 2015 rule”). 2015 Final Rule, 79 Fed. Reg. at 13,744-843. HHS included no additional reasoning about its decision to use the statewide average premium or its budget-neutral design, see 2015 Proposed Rule, 78 Fed. Reg. at 72,322-92; 2015 Final Rule, 79 Fed. Reg. at 13,744- 843, and no commenter challenged either, see 2015 Final Rule, 79 Fed. Reg. at 13,744-843.

21

b. 2016 Rule On November 26, 2014, HHS published a proposed rule for the 2016 benefit year. 2016 Proposed Rule, 79 Fed. Reg. 70,674, 70,674-760 (Nov. 26, 2014). On February 27, 2015, it published the final rule for 2016 (“the 2016 rule”). 2016 Final Rule, 80 Fed. Reg. at 10,750-877. As with the 2015 rule, HHS included no additional reasoning about its decision to use the statewide average premium or its budget-neutral design, see 2016 Proposed Rule, 79 Fed. Reg. at 70,674-760; 2016 Final Rule, 80 Fed. Reg. at 10,750-877, and no commenter challenged either, see 2016 Final Rule, 80 Fed. Reg. at 10,750-877.

c. 2017 Rule On December 2, 2015, HHS published a proposed rule for the 2017 benefit year. 2017 Proposed Rule, 80 Fed. Reg. 75,488, 75,488-588 (Dec. 2, 2015). HHS published the final rule for 2017 on March 8, 2016 (“the 2017 rule”). 2017 Final Rule, 81 Fed. Reg. at 12,204-352. For the first time, HHS received comments opposing its use of the statewide average premium. It reported

Some commenters opposed the use of the statewide average premium because it disadvantages issuers with below average premiums. Commenters requested that 2014 and later risk adjustment transfers for all plans with below average premiums in a [s]tate be calculated using the plans’ own average premium amount or average claims cost, so that efficient plans are not penalized using the [s]tatewide average premium.

Id. at 12,230. HHS responded

22

We did not propose changes to the transfer formula, and therefore, are not addressing comments that are outside the scope of this rulemaking. We may be able to evaluate geographic differences in the future if we obtain enrolleelevel data for future recalibrations—a topic that we also intend to discuss in the White Paper and at the March 31, 2016 risk adjustment conference.

Id.

On March 24, 2016, HHS circulated the aforementioned white paper, which assessed the formula. See 2016 White Paper at 79-101. In it, CCIIO noted it used the statewide average premium both as “a premium and a cost scaling factor.” Id. at 83. It explained

Over the long run, the [s]tatewide average premium is expected to equal the [s]tatewide average cost (including allowable loading for administrative costs, surplus, and profit). The [s]tatewide premium is therefore simultaneously a premium and cost scaling factor. The [s]tatewide average premium embeds an average level of efficiency. All plans receive a risk adjustment payment or charge sufficient for a plan with average efficiency.

Two other reasons that transfers are scaled by the [s]tatewide average premium, as opposed to, for example, the plan’s own premium, are:

• Using the [s]tatewide average premium minimizes issuers’ ability to manipulate their transfers by adjusting their own plan premiums.

• Scaling all transfers to the same premium, combined with the assumption that the factors affecting premium requirements and allowable revenue have a multiplicative relationship, obviates any further adjustment of payments and charges to ensure that

23

risk adjustment transfers for the entire market sum to zero.

Id.

NMHC’s District Court Complaint On July 29, 2016, NMHC sued HHS, alleging that the agency violated § 1343 of the ACA and § 706 of the APA. Dist. Ct. Doc. 1 at 43-45. 14 NMHC alleged HHS’s use of the statewide average premium in its risk adjustment program in 2014, 2015, 2016, and 2017 was “arbitrary, capricious, and unlawful.” Id. at 45. It asserted that HHS “flout[ed] Congressional intent and the express mandate of the Risk Adjustment statute.” Id. NMHC filed its amended complaint on January 12, 2017, adding a challenge to the 2018 rule. See Supp. App. at 68-69.

Notice-and-Comment Proceedings for the 2018 Rule On September 6, 2016, HHS published a proposed rule for the 2018 benefit year. 2018 Proposed Rule, 81 Fed. Reg. 61,456, 61,456-536 (Sept. 6, 2016). Because the 2018 rule incorporated and built upon previous proceedings, it relied on the same reasoning as the prior rules. See 2018 Final Rule, 81 Fed. Reg. at 94,100

14 Other lawsuits challenging the risk adjustment program include Minuteman Health Inc. v. U.S. Dep’t of Health & Human Servs., 291 F. Supp. 3d 174, 202 (D. Mass. 2018) (holding HHS’s use of the statewide average premium was not arbitrary and capricious), appeal not filed; Evergreen Health Coop., Inc. v. U.S. Dep’t of Health & Human Servs., No. 16-2039 (D. Md. Jan. 31, 2017) (entering stipulated dismissal of challenge to risk adjustment program); and Ommen v. United States, No. 17-957 (Fed. Cl. July 25, 2019) (staying risk adjustment challenge pending the Supreme Court’s resolution of Moda Health Plan, Inc. v. United States, No. 18-1028 (Oct. 25, 2019)).

24

(“The payment transfer formula is unchanged from what was finalized in the 2014 [rule] . . . except with an adjustment to remove a portion of the administrative costs from the [s]tatewide average premium . . . .”).

On December 22, 2016, HHS published the final rule for 2018 (“the 2018 rule”). Id. at 94,058-183. HHS noted “[a] few commenters requested that HHS use a plan’s own actual average premium instead of the [s]tatewide average premium in the transfer formula.” Id. at 94,100. HHS responded

We have considered the use of a plan’s own premium instead of the [s]tatewide average premium. However, our analysis determined that this approach is likely to lead to substantial volatility in transfer results and even higher transfer charges for low-risk low-premium plans. Under such an approach, high-risk, high-premium plans would require even greater transfer payments; thus, low-risk, low-

premium plans would be required to pay in an even higher percentage of their plan-specific premiums in risk adjustment transfer charges. In other words, the use of a plan’s own premium does not reduce risk adjustment charges for low-cost and low-risk issuers, given the budget neutrality of the risk adjustment program.

Id.

For the first time, HHS received comments challenging the budget-neutral

design of the program. HHS reported

A few commenters noted that the budget neutrality of the risk adjustment program leads to inadequate compensation for enrollees’ risk and recommended a non-budget neutral risk adjustment program as with Medicare Advantage.

Id. at 94,101. It responded

25

In the absence of additional funding for the HHS-operated risk adjustment program, we continue to calculate risk adjustment transfers in a budget neutral manner and note that Medicare Part D risk adjustment transfers are also calculated in a budget neutral manner.

Id.

Cross-Motions for Summary Judgment and District Court Order In mid-2017, both NMHC and HHS moved for summary judgment. In its motion, NMHC made several arguments, including that HHS’s use of the statewide average premium to calculate risk adjustment transfer payments was arbitrary and capricious. It asserted it was charged $6,666,798 in transfer charges for the 2014 benefit year, representing 21.5 percent of its premiums. NMHC further asserted it was charged $14,569,495.74 for the 2015 benefit year, representing 14.7 percent of its premiums. NMHC alleged these fees were debilitating because industry margins were “typically, at best, a razor thin 2% [to] 3%.” Supp. App. at 258.

In its cross motion for summary judgment, HHS contended its methodology was “eminently reasonable” and “easily satisfie[d] the APA’s standard of review.” Id. at 335.

The district court granted summary judgment for NMHC, finding that HHS’s use of the statewide average premium was arbitrary and capricious. NMHC I, 312 F. Supp. 3d at 1170-72. 15 It noted that, “even though both [NMHC] and HHS ostensibly

15 NMHC and HHS advanced multiple arguments. The district court rejected all of NMHC’s claims except the one we address on appeal. The court concluded:

26

filed motions for summary judgment,” “district courts reviewing agency action do not determine whether a genuine dispute as to any material fact exists, and instead engage in a substantive review of the record to determine if the agency considered relevant factors or articulated a reasoned basis for its conclusions.” Id. at 1171 (quotations and citations omitted).

In reviewing the APA challenge, the district court stated the issue was “whether incorporating statewide average premiums in . . . [the] risk[]adjustment [program] is contrary to law or arbitrary and capricious.” Id. at 1170. It held that “HHS’[s] use of statewide average premiums in its risk adjustment [program] is not contrary to law, but is arbitrary and capricious.” Id. The court said, “HHS assumed, erroneously, that the ACA requires risk adjustment to be budget neutral, and all of HHS’[s] reasons for using the statewide average premium rely on that budget neutrality assumption.” Id. at 1202. It noted HHS had justified using the statewide average premium to achieve “budget neutrality and predictability.” Id. at 1209. The court said that this assumption “infect[ed]” HHS’s “analysis of the relative merits of

(1) the APA waived sovereign immunity for all claims presented; (2) HHS’s approach to predicting costs for enrollees (hierarchal condition category (“HCC”) “and non-HCC eligible enrollees”) was not arbitrary and capricious; (3) HHS’s decisions regarding partial-year enrollees and the use of prescription drug data in its risk adjustment model were not arbitrary and capricious; and (4) HHS’s payment transfer formula did not ban bronze health insurance plans. NMHC I, 312 F. Supp. 3d at 1171. We do not address these rulings because NMHC did not appeal the judgment.

27

using a state’s average premium . . . instead of a plan’s own premium,” including HHS’s explanation that the statewide average premium was more predictable. Id.

The district court vacated HHS’s 2014, 2015, 2016, 2017, and 2018 rules. Id.

at 1170-71. It entered judgment and remanded the case to the agency for further proceedings. Id.

HHS’s Rule 59(e) Motion HHS moved to alter or amend the judgment under Federal Rule of Civil Procedure 59(e). First, it argued NMHC had waived any argument that HHS acted arbitrarily and capriciously regarding the budget-neutral design of the program because no commenter had challenged budget neutrality in the 2014, 2015, 2016, or 2017 rulemaking proceedings. HHS also said it sufficiently explained its budget- neutral design when it received comments to the 2018 rule. Second, HHS contended it had not interpreted the ACA to mandate budget neutrality but rather had designed the payment transfer formula to be budget neutral because no appropriations from Congress were available for the program. Third, HHS argued that vacating the rules was not required under the APA and would be manifestly unjust.

Notice-and-Comment Proceedings for the Modified 2017 and 2018 Rules While this Rule 59(e) motion was pending, HHS attempted to adjust its 2017 and 2018 rules in response to the district court’s summary judgment order. On July 7, 2018, HHS suspended risk adjustment transfers for 2017 based on the uncertainty created by this case. See N.M. Health Connections v. U.S. Dep’t of Health & Human

28

Servs. (NMHC II), 340 F. Supp. 3d 1112, 1141 (D.N.M. 2018). HHS then notified the district court that it was formulating a new interim final rule for 2017. See Dist. Ct. Doc. 78 at 1-2.

On July 25, 2018, HHS sent a letter to the district court announcing it had issued a new final rule on the risk adjustment methodology for the 2017 benefit year (“the new 2017 rule”). Dist. Ct. Doc. 81 at 1. Although the new rule adopted the same methodology as prior rules, HHS “provide[d] additional explanation of the agency’s use of [the] statewide average premium in the risk adjustment payment transfer formula, as well as the risk adjustment program’s budget neutral design.” Id. at 1.

In the new 2017 rule, HHS explained it chose to use the statewide average premium to “support[] the overall goal of the risk adjustment program to encourage issuers to” set premium rates “for the average risk in the applicable state market risk pool.” New 2017 Final Rule, 83 Fed. Reg. 36,456, 36,457 (July 30, 2018). HHS also noted that using the statewide average premium “avoids the creation of incentives for issuers to operate less efficiently, set higher prices, develop benefit designs or create marketing strategies to avoid high risk enrollees.” Id.

HHS also explained that it chose to use a budget-neutral design for the program for several reasons. First, “Congress designed the risk adjustment program to be implemented and operated by states if they choose to do so,” and HHS could not require the states to provide funding. Id. at 36,458. Second, it noted the ACA

29

“neither authorized nor appropriated additional funding for risk adjustment payments beyond the amount of charges paid in.” Id. “Thus, as a practical matter, Congress did not give HHS discretion to implement a program that was not budget neutral.” Id. Third, “if HHS had elected to adopt a . . . methodology that was contingent on appropriations from Congress . . . that would have created uncertainty for issuers in the amount of risk adjustment payments they could expect.” Id. After issuing the new rule, HHS implemented transfer payments for the 2017 benefit year. Id. at 36,459.

On August 8, 2018, HHS sent a letter to the district court announcing it had issued a new proposed rule for the 2018 benefit year. Dist. Ct. Doc. 84 at 1; see New 2018 Proposed Rule, 83 Fed. Reg. 39,644, 39,644-48 (Aug. 10, 2018). HHS noted the new rule would respond to the district court’s summary judgment decision by providing additional explanation of the agency’s use of the statewide average premium and the program’s budget-neutral design. Dist. Ct. Doc. 84 at 1-2. HHS said it had sought comments on the issues. Id. at 1. It also argued its proceedings on the 2017 and 2018 rules did not moot its Rule 59(e) motion regarding the 2014, 2015, and 2016 rules. Id. at 2. 16

16 The new 2018 final rule was promulgated on December 10, 2018, after the district court’s denial of the 59(e) motion (“the new 2018 rule”). New 2018 Final Rule, 83 Fed. Reg. 63,419, 63,419-28 (Dec. 10, 2018). HHS had received comments related to its use of the statewide average premium and the program’s budget-neutral design. Id. at 63,422-27. Its responses to those comments contained explanations similar to those outlined in its new 2017 rule. See id.

30

On August 13, 2018, NMHC filed a letter with the court, arguing HHS’s motion on Rule 59(e) “is now entirely moot as to 2017 and 2018.” Dist. Ct. Doc. 85 at 1.

Denial of Rule 59(e) Motion The district court denied HHS’s Rule 59(e) motion. NMHC II, 340 F. Supp.

3d at 1121. First, it found HHS had sua sponte considered using the statewide average premium over alternatives, so the absence of comments challenging budget neutrality did not forfeit the issue. Id. at 1168. Second, it reiterated that the payment transfer formula was arbitrary and capricious because HHS had insufficiently explained the budget-neutral design of its program. Id. at 1165-66. The court claimed HHS could have relied on Congress’s “lump sum” appropriations to subsidize a non-budget-neutral formula program. Id. at 1174. Third, the court maintained vacatur was the proper remedy under the APA. Id. at 1175-84. After the denial of HHS’s Rule 59(e) motion, the agency timely appealed.

II. JURISDICTION

We address two jurisdictional issues. We conclude (A) we have appellate jurisdiction because the administrative remand rule does not apply, and (B) NMHC’s challenges to the 2017 and 2018 rules are moot.

A. Appellate Jurisdiction We must determine whether the administrative remand rule precludes our appellate jurisdiction. Under that rule, “[a] remand by a district court to an

31

administrative agency for further proceedings is ordinarily not appealable because it is not a final decision.” W. Energy All. v. Salazar, 709 F.3d 1040, 1047 (10th Cir. 2013) (quotations omitted). Here, the district court’s summary judgment order determined HHS had violated the APA because its implementation of the risk adjustment program was arbitrary and capricious. It remanded that issue to HHS. This court ordered the parties to address whether we lack jurisdiction under the administrative remand rule in their briefs. Although neither party contests our appellate jurisdiction, “federal courts are under an independent obligation to examine their own jurisdiction.” United States v. Hays, 515 U.S. 737, 742 (1995).

“[W]e exercise jurisdiction over final decisions of the federal district courts pursuant to 28 U.S.C. § 1291.” Am. Wild Horse Pres. Campaign v. Jewell, 847 F.3d 1174, 1183 (10th Cir. 2016) (quotations omitted). “A final decision is one ‘that ends the litigation on the merits and leaves nothing for the court to do but execute the judgment.’” W. Energy All., 709 F.3d at 1047 (quotations omitted). In determining whether the district court’s order was a final decision under the administrative remand rule, “this court considers the nature of the agency action as well as the nature of the district court’s order.” Cherokee Nation v. Bernhardt, 936 F.3d 1142, 1151 (10th Cir. 2019) (quotations omitted).

As to the nature of the agency action, we consider whether it was “essentially adjudicatory, essentially legislative, or some nonadversarial action such as grant of a license.” Am. Wild Horse Pres. Campaign, 847 F.3d at 1184 (quotations omitted).

32

Adjudicatory action affects individuals in their individual capacity. W. Energy All., 709 F.3d at 1048. Legislative action “affects the rights of individuals in the abstract and must be applied in a further proceeding before the legal position of any particular individual will be definitively” affected. Id. (quotations omitted). “[O]ur precedent indicates that we view the remand rule as most appropriate in adjudicative contexts.” Am. Wild Horse Pres. Campaign, 847 F.3d at 1184 (quotations omitted).

As to the district court’s order, we consider its character, “including whether it returns an action to . . . the agency for further proceedings.” Id. (alterations and citations omitted). If the district court’s order is not a “remand in the typical sense,” the administrative remand rule is inapplicable. W. Energy All., 709 F.3d at 1047 (quotations omitted).

We conclude the district court’s summary judgment order was a final, appealable decision and the administrative remand rule does not apply. First, HHS acted in a legislative capacity in promulgating the rules. See N.M. ex rel. Richardson v. Bureau of Land Mgmt., 565 F.3d 683, 698 (10th Cir. 2009) (holding agency action was final decision where agency proceeding was “quasi-legislative”). Second, although the district court’s order remanded to HHS to correct deficiencies, the 2014, 2015, and 2016 transfer payments have already occurred, so the order “does not share the features of a typical remand.” Am. Wild Horse Pres. Campaign, 847 F.3d at 1184 (holding agency action was final decision where action based on the challenged rule had already taken place and remand was to “correct procedural deficiencies” in rule

33

analysis). And, as discussed below, NMHC’s challenges to the 2017 and 2018 rules are moot.

The district court’s disposition was not an administrative remand but a final order. We have jurisdiction under 28 U.S.C. § 1291. 17 B. NMHC’s Challenges to the 2017 and 2018 Rules Are Moot NMHC asserts its “challenges to the old 2017 and 2018 [rules] are moot.” Aplee.

Br. at 45. 18 We agree. “We have no subject-matter jurisdiction if a case is moot.” Rio

17 Even if the district court’s disposition was an administrative remand, the order was final and appealable under the “practical finality exception” to the administrative remand rule. “[W]e employ a two-pronged test for applying the practical finality rule: the issue must be important and it must be urgent.” W. Energy All., 709 F.3d at 1049-50 (quotations omitted). “If the test is met, we then follow a balancing approach . . . .” Id. at 1050 (quotations omitted). “In practice, we have applied the practical finality rule when it is necessary to ensure that we can review important legal questions which a remand may make effectively unreviewable, because administrative agencies may be barred from seeking district court (and thus circuit court) review of their own administrative decisions.” Miami Tribe of Okla. v. United States, 656 F.3d 1129, 1140 (10th Cir. 2011) (quotations omitted). Here, the issue is important, and it is urgent to review this case on appeal because any additional delay could have a significant effect on the national health insurance market. The importance and urgency factors outweigh the inconvenience and cost of piecemeal review. Moreover, HHS might be foreclosed from future appellate review because it could not challenge changes to its own program based on the district court’s holding. See id.

18 In a Rule 28(j) letter, NMHC adds it “will not be challenging the new 2018 rule as it has suffered no injury.” Aplee. 28(j) Letter, Doc. 10662343 at 2. At oral argument, HHS argued NMHC’s challenges to the 2017 and 2018 rules are not moot but provided no explanation. Or. Arg. at 3:08-3:22. Nor did it argue mootness in its briefing. Aplt. Reply Br. at 17. Although NMHC raised mootness below after the new 2017 and 2018 rules were promulgated, the district court did not address mootness in its order denying HHS’s Rule 59(e) motion. See NMHC II, 340 F. Supp. 3d at 1120-84.

34

Grande Silvery Minnow v. Bureau of Reclamation, 601 F.3d 1096, 1109 (10th Cir. 2010). We review questions of mootness de novo as “squarely a legal determination.” Brown v. Buhman, 822 F.3d 1151, 1168 (10th Cir. 2016) (quotations omitted).

Mootness Article III of the Constitution permits federal courts to decide only “Cases” or “Controversies.” U.S. Const. art. III, § 2; see Hollingsworth v. Perry, 570 U.S. 693, 704 (2013). “This case-or-controversy requirement subsists through all stages of federal judicial proceedings, trial and appellate.” Spencer v. Kemna, 523 U.S. 1, 7 (1998) (quotations omitted). At the appellate stage, it “requires a party seeking relief to have suffered, or be threatened with, an actual injury traceable to the appellee and likely to be redressed by a favorable judicial decision by the appeals court.” United States v. Vera-Flores, 496 F.3d 1177, 1180 (10th Cir. 2007) (alterations and quotations omitted).

“A case becomes moot—and therefore no longer a ‘Case’ or ‘Controversy’ for purposes of Article III—when the issues presented are no longer live or the parties lack a legally cognizable interest in the outcome.” Already, LLC v. Nike, Inc., 568 U.S. 85, 91 (2013) (quotations omitted). “If an intervening circumstance deprives the plaintiff of a personal stake in the outcome of the lawsuit, at any point during litigation, the action can no longer proceed and must be dismissed as moot.” Brown, 822 F.3d at 1165 (quoting Campbell-Ewald Co. v. Gomez, 136 S. Ct. 663, 669 (2016)). “Corrective action by an agency is one type of subsequent development that can moot a previously justiciable issue.” Nat. Res. Def. Council, Inc. v. U.S. Nuclear Reg. Comm’n, 680 F.2d 810, 814

35

(D.C. Cir. 1982). An agency’s action moots an appeal “[b]y eliminating the issues upon which this case is based.” Hayes v. Osage Mins. Council, 699 F. App’x 799, 803 (10th Cir. 2017) (quoting Wyoming v. U.S. Dep’t of Agric., 414 F.3d 1207, 1212 (10th Cir. 2005)). 19 Courts recognize two exceptions to mootness. First, a case or controversy is not moot if the dispute is “capable of repetition, yet evading review.” Brown, 822 F.3d at 1166 (quotations omitted). This exception “applies where (1) the challenged action is in its duration too short to be fully litigated prior to cessation or expiration, and (2) there is a reasonable expectation that the same complaining party will be subject to the same action again.” Id. (quoting Fed. Election Comm’n v. Wis. Right to Life, Inc., 551 U.S. 449, 462 (2007)).

Second, “voluntary cessation of challenged conduct does not ordinarily render a case moot because a dismissal for mootness would permit a resumption of the challenged conduct as soon as the case is dismissed.” Id. (quoting Knox v. Serv. Emps. Int’l Union, Local 1000, 567 U.S. 298, 307 (2012)). This exception is designed to counteract gamesmanship, such as “a defendant ceasing illegal action long enough to render a lawsuit moot” before “resuming the illegal conduct.” Ind v.

19 Although not precedential, we find the reasoning of the unpublished decisions cited in this opinion instructive. See 10th Cir. R. 32.1 (“Unpublished decisions are not precedential, but may be cited for their persuasive value.”); Fed. R. App. P. 32.1 (same).

36

Colo. Dep’t of Corrs., 801 F.3d 1209, 1214 (10th Cir. 2015) (quotations omitted). It does not apply, however, if it is “absolutely clear the allegedly wrongful behavior could not reasonably be expected to recur.” Brown, 822 F.3d at 1166 (quotations omitted). Government “self-correction . . . provides a secure foundation for mootness so long as it seems genuine.” 13C Charles Alan Wright, Arthur R. Miller, & Edward H. Cooper, Fed. Prac. & Proc. Juris. § 3533.7 (3rd ed. 2008); see also Rio Grande Silvery Minnow, 601 F.3d at 1117-18 (holding “withdrawal or alteration of administrative policies” satisfies concerns regarding voluntary cessation (quotations omitted)).

Analysis Although HHS appealed the district court’s summary judgment order as to the 2014, 2015, 2016, 2017, and 2018 risk adjustment rules, NMHC’s challenges to the 2017 and 2018 rules became moot when HHS modified them. NMHC has not challenged the modified rules.

While its Rule 59(e) motion was pending, HHS (1) issued an emergency new risk adjustment final rule for the 2017 benefit year and (2) conducted rulemaking proceedings for a new risk adjustment rule for the 2018 benefit year. See New 2017 Final Rule, 83 Fed. Reg. at 36,456; New 2018 Proposed Rule, 83 Fed. Reg. at 39,644; New 2018 Final Rule, 84 Fed. Reg. at 63,419. Both proceedings modified and superseded the original 2017 and 2018 rules. See id. In its new 2017 and 2018 rules, HHS included explanations about why it used the statewide average premium in its transfer formula and why it

37

designed its program to be budget neutral. See New 2017 Final Rule, 83 Fed. Reg. at 36,457-58; New 2018 Final Rule, 83 Fed. Reg. at 63,420-27.

These actions cured the defects identified in the district court’s order for the original 2017 and 2018 rules. Because the new 2017 and 2018 rules supersede the old ones, 20 any decision by this court regarding the original 2017 and 2018 rules “would be wholly without effect in the real world.” Rio Grande Silvery Minnow, 601 F.3d at 1112; see also Wyoming, 414 F.3d at 1212 (noting “the alleged procedural deficiencies of the . . . [r]ule are now irrelevant because the replacement rule was promulgated in a new and separate rulemaking process”). NMHC’s challenges to the original 2017 and 2018 rules are therefore moot. 21

20 See New 2017 Final Rule, 83 Fed. Reg. at 36,457 (“This final rule adopts the HHS-operated risk adjustment methodology previously published [in the original 2017 Final Rule] for the 2017 benefit year with an additional explanation regarding the use of the statewide average premium and the budget-neutral nature of the program.”); New 2018 Final Rule, 83 Fed. Reg. at 63,419 (same as to 2018 benefit year).

21 NMHC has not challenged these curative actions or the new rules, and, as noted above, considers its case regarding the original 2017 and 2018 rules moot. NMHC filed a separate challenge to the new 2017 rule. See Aplee. Br. at 46; Complaint, N.M. Health Connections v. U.S. Dep’t of Health & Human Servs., No. 18-773 (D.N.M. Aug. 13, 2018), ECF No. 1. That case has been stayed by agreement of the parties pending this appeal. See Aplee. Br. at 46; Joint Status Report, N.M. Health Connections v. U.S. Dep’t of Health & Human Servs., No. 18-773 (D.N.M. Jan. 28, 2019), ECF No. 31. In its brief on appeal, NMHC noted it “has not yet decided whether it will seek to amend its complaint in the second action to challenge the new 2018 rule,” but it “did submit a 35-page comment, attaching 86 exhibits, in response to the proposed new 2018 rule.” Aplee. Br. at 46 n.11.

38

Neither exception to the mootness doctrine applies here. Mooting this case would not run afoul of the “capable of repetition” exception because any future failure to explain the agency rulemaking would allow NMHC ample time and opportunity to challenge HHS’s action. HHS is also unlikely to subject NMHC to a purportedly unexplained decision to use the statewide average premium or budget neutrality in future rules given the agency’s extensive explanation on those matters in the new 2017 and 2018 rules and the agency’s practice on each successive risk adjustment rule to build on and incorporate prior ones. The additional explanations provided in the new 2017 and 2018 rules will likely be incorporated into future rules relying on the same formula.

Nor does the “voluntary cessation” exception apply. Mooting NMHC’s challenges to the 2017 and 2018 rules would not offer HHS an opportunity to resume “the challenged conduct.” Brown, 822 F.3d at 1166. HHS already has replaced the original rules with these new ones and has added explanation to satisfy the district court. It cannot revert to the original rules without a new proceeding, see 5 U.S.C. § 553, which is unlikely as the new rules replace the original ones and the agency followed them to implement the program for 2017 and 2018. See Rio Grande Silvery Minnow, 601 F.3d at 1117-18. Finally, because we also continue to address NMHC’s challenges to the 2014, 2015, and 2016 rules, any concern about “gamesmanship” is unwarranted.

39

Because NMHC’s challenges to the 2017 and 2018 rules are moot, we do not address them in our following merits discussion. We remand NMHC’s challenges to the 2017 and 2018 rules to the district court with instructions to vacate its judgment as to those rules for lack of jurisdiction.

III. NMHC’S CHALLENGE TO THE 2014, 2015, AND 2016 RULES The district court vacated and remanded HHS’s 2014, 2015, 2016, 2017, and 2018 risk adjustment rules. Due to our mootness determination, we limit our review to the 2014, 2015, and 2016 rules. We reverse the district court’s summary judgment ruling in favor of NMHC regarding these rules. As a result, we need not address whether the district court erred in ordering vacatur or in denying HHS’s Rule 59(e) motion.

A. Legal Background

Under the APA, a reviewing court shall “hold unlawful and set aside agency action . . . found to be arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A).

In reviewing an APA challenge to agency action, a district court acts as an appellate court. See S. Utah Wilderness All. v. Dabney, 222 F.3d 819, 823 n.4 (10th Cir. 2000). The court employs summary judgment to “decid[e], as a matter of law, whether the agency action is supported by the administrative record and otherwise consistent with the APA standard of review.” Calloway v. Harvey, 590 F. Supp. 2d 29, 36 (D.D.C. 2008) (quotations omitted); see Olenhouse v. Commodity Credit

40

Corp., 42 F.3d 1560, 1576 (10th Cir. 1994) (“[T]he district court itself must examine the administrative record and itself must find and identify facts that support the agency’s action.”). 22 On appeal to this court, we review “de novo a district court’s decision in an APA case,” Biodiversity Conservation All. v. Jiron, 762 F.3d 1036, 1059 (10th Cir. 2014), and consider “the administrative record directly,” Stand Up for California! v. U.S. Dep’t of Interior, 879 F.3d 1177, 1181 (D.C. Cir. 2018). “In reviewing the agency’s action, we must render an independent decision using the same standard of review applicable to the [d]istrict [c]ourt[’s]” review. Olenhouse, 42 F.3d at 1580; see City of Colo. Springs v. Solis, 589 F.3d 1121, 1131 n.2 (10th Cir. 2009) (focusing on Department of Labor decision rather than district court framing in APA challenge on appeal). Our review is limited to the administrative record, including “all materials compiled by the agency that were before the agency at the time the decision was made.” James Madison Ltd. v. Ludwig, 82 F.3d 1085, 1095 (D.C. Cir. 1996) (quotations omitted).

Although we afford no deference to the district court’s decision, we must do so as to HHS’s actions. See W. Watersheds Project v. Bureau of Land Mgmt., 721 F.3d

22 See also 10A Charles Alan Wright & Arthur R. Miller, Fed. Prac. & Proc.

Civ. § 2725 (4th ed. 2019) (“[W]hen the court is being asked to review an administrative record and apply legal standards to that record, summary judgment is an appropriate vehicle for deciding the case.”).

41

1264, 1273 (10th Cir. 2013). “Our inquiry under the APA must be thorough, but the standard of review is very deferential to the agency.” Id. (quotations omitted). This deference means we may set aside an agency action only if it violates the APA. See id. HHS’s actions are entitled to a presumption of regularity, and NMHC bears the burden of persuasion to show HHS acted arbitrarily and capriciously. See Biodiversity Conservation All., 762 F.3d at 1060.

An agency’s action is arbitrary and capricious where the agency

(1) entirely fail[s] to consider an important aspect of the problem, (2) offer[s] 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 of the product of agency expertise, (3) fail[s] to base its decision on consideration of the relevant factors, or (4)

ma[kes] a clear error of judgment.

W. Watersheds Project, 721 F.3d at 1273 (quotations omitted).

“One of the basic procedural requirements of administrative rulemaking is that an agency must give adequate reasons for its decisions.” Encino Motorcars, LLC v. Navarro, 136 S. Ct. 2117, 2125 (2016). “Our scope of review is ‘narrow’: we determine only whether the [agency] examined ‘the relevant data’ and articulated ‘a satisfactory explanation’ for [its] decision, ‘including a rational connection between the facts found and the choice made.’” Dep’t of Commerce v. New York, 139 S. Ct. 2551, 2569 (2019) (quoting Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983)). We “must confine ourselves to ensuring that [the agency] remained within the bounds of reasoned decisionmaking.” Id. (quotations

42

omitted). “But where the agency has failed to provide even that minimal level of analysis, its action is arbitrary and capricious . . . .” Encino Motorcars, 136 S. Ct. at 2125.

B. NMHC’s Challenge to the Statewide Average Premium The district court decided HHS was arbitrary and capricious because it did not explain why the risk adjustment program would be budget neutral. On appeal, as in its complaint, NMHC more broadly attacks HHS’s rationale for the statewide average premium, which, if meritorious, would be an alternative ground to affirm. In the following discussion, we therefore determine that HHS did not act arbitrarily and capriciously in using the statewide average premium in its formula for the 2014, 2015, and 2016 rules. We further address why HHS was not arbitrary and capricious for failing to explain why it chose a budget-neutral design.

Statewide Average Premium NMHC challenged “HHS’s decision to use the statewide average premium, as opposed to an insurer’s own premium,” in the risk adjustment program. Aplee. Br. at 20-21. The district court concluded that “HHS’[s] risk adjustment regulations— specifically their use of the statewide average premium—are arbitrary and capricious.” NMHC I, 312 F. Supp. 3d at 1205; see also NMHC II, 340 F. Supp. 3d at 1166 (noting district court “concluded that HHS did not adequately explain its decision to employ state average premiums”). NMHC contends that “HHS points to nothing in the administrative record to support its use of the statewide average

43

premium, and instead relies on impermissible post hoc arguments. . . .” Aplee. Br. at 23. We disagree. The administrative record is replete with reasoned explanations for why HHS chose to use the statewide average premium in its formula.

As discussed above, HHS researched how to implement the risk adjustment program before promulgating the 2014 rule. In its 2011 White Paper, CCIIO identified four alternatives for the formula, including the statewide average premium and the plans’ own premiums. 2011 White Paper at 14. 23 It described the benefits and drawbacks of each alternative and invited the public to comment on them. Id. For example, CCIIO noted:

Option 1a: Weighted [s]tate average premiums. This approach would calculate the baseline premium according to the enrollment-weighted average premium in the [s]tate.

The [s]tate average could be calculated with or without adjustment for actuarial value of plans. Using a [s]tate average (without actuarial value adjustment) would result in balanced payments and charges, because the [s]tate average is a single dollar amount for all plans, and plan risk scores average to 1.0.

...

Option 2: Plan’s own premiums. This approach would use each plan’s own premiums as the baseline premium.

Relative to the prior options, charges would be lowest for low premium, low-risk plans under this approach, and payments would be highest for high risk, high premium plans. In this approach, the amount of charges and

23 As previously noted, because NMHC argued only in favor of using the plans’ own premiums, we focus on HHS’s choice to use the statewide average premium over the plans’ own premiums.

44

payments would be affected by each plan’s premium. For plans with a sicker than average risk mix, a lower premium plan would receive less in payments than a higher premium plan, even if the two plans have the same risk level. This could create disincentives for high-risk plans to operate efficiently or set lower prices. Conversely, among two plans with the same healthier than average risk mix, a lower premium plan would have lower charges, potentially creating incentives for low-risk plans to operate more efficiently and/or set lower premiums.

Id. at 14-15. It further noted:

When payments are greater than charges, a low risk plan with low premiums would be charged less if the baseline premium is the plan’s own premiums and payments are reduced to charges, as compared to what the plan would be charged if the baseline premium is the [s]tate average premium or the baseline premium is the plan’s own premiums with charges increased to payments.

Conversely, a high risk plan with high premiums would receive higher payments if the baseline premium is the plan’s own premium and charges are increased to payments, as compared to the payments the plan would receive if the baseline premium is the [s]tate average premium, or the baseline premium is the plan’s own premiums with payments decreased to charges.

Id. at 16.

When HHS proposed using the statewide average premium in its 2014 proposed rule, it relied on research from the 2011 White Paper. See 2014 Proposed Rule, 77 Fed. Reg. at 73,139. In a section titled “Rationales for a Transfer Methodology Based on State Average Premiums,” HHS explained:

Risk adjustment transfers are intended to reduce the impact of risk selection on premiums while preserving premium differences related to other cost factors, such as the

45

actuarial value, local patterns of utilization and care delivery, local differences in the cost of doing business, and, within limits established by the Affordable Care Act, the age of the enrollee. Risk adjustment payments would be fully funded by the charges that are collected from plans with lower risk enrollees (that is, transfers within a [s]tate would net to zero).

In the [2011] Risk Adjustment White Paper, we presented several approaches for calculating risk adjustment transfers using the [s]tate average premium and plans’ own premiums. The approaches that used plans’ own premiums resulted in unbalanced payment transfers, requiring a balancing adjustment to yield transfers that net to zero. These examples also demonstrated that the balancing adjustments could introduce differences in premiums across plans that were not consistent with features of the plan (for example, AV or differences in costs and utilization patterns across rating areas). A balancing adjustment would likely vary from year to year, and could add uncertainty to the rate development process (that is, plan actuaries would need to factor the uncertainty of the balancing adjustment into their transfer estimates).

Therefore, we propose a payment transfer formula that is based on the [s]tate average premium for the applicable market, as described in section III.B.3.a. of this proposed rule. The [s]tate average premium provides a straightforward and predictable benchmark for estimating transfers. As shown in the examples in the [2011] Risk Adjustment White Paper, transfers net to zero when the [s]tate average premium is used as the basis for calculating transfers.

Plan premiums differ from the [s]tate average premium due to a variety of factors, such as differences in cost-sharing structure or regional differences in utilization and unit costs. The proposed payment transfer formula applies a set of cost factor adjustments to the [s]tate average premium so that it will better reflect plan liability. These adjustments to the [s]tate average premium result in transfers that compensate plans for liability differences associated with risk selection,

46

while preserving premium differences related to the other cost factors described above.

Id.

In the 2014 final rule, HHS noted it received comments in support of its decision to use the statewide average premium. It said:

Comment: We received a number of comments in support of our proposal to use the [s]tate average premium as the basis for risk adjustment transfers. One commenter suggested that use of a plan’s own premium may cause unintended distortions in the transfer formula. One commenter suggested that we use net claims, or approximate net claims by using 90 percent of the [s]tate average premium, as the basis for risk adjustment transfers.

Response: The goal of the payment transfer formula is, to the extent possible, to promote risk-neutral premiums.

We agree with commenters that use of a plan’s own premium may cause unintended distortions in transfers. We also believe that both claims and administrative costs include elements of risk selection, and therefore, that transfers should be based on the entire premium. We are finalizing our proposal to base the payment transfer formula on the [s]tate average premium.

2014 Final Rule, 78 Fed. Reg. at 15,432.

In sum, HHS explained in the rulemaking proceeding that it chose the statewide average premium over alternatives such as the plans’ own premiums to (1) “reduce the impact of risk selection on premiums while preserving premium differences related to other cost factors,” 2014 Proposed Rule, 77 Fed. Reg. at 73,139; (2) achieve “a straightforward and predictable benchmark for estimating transfers” each year, id.; (3) “promote risk-neutral premiums,” 2014 Final Rule, 78

47

Fed. Reg. at 15,432; (4) avert “caus[ing] unintended distortions in transfers,” id.; see also 2011 White Paper at 14 (using plans’ own premiums “could create disincentives for high-risk plans to operate efficiently or set lower prices”); and (5) avoid disproportionately distributing costs to insurers when using balancing adjustments, 2011 White Paper at 16. A sixth explanation stated that using the statewide average premium facilitates budget neutrality, making transfers “net to zero” without additional balancing adjustments. 2014 Proposed Rule, 77 Fed. Reg. at 73,139; see 2011 White Paper at 14. 24 HHS thus acted reasonably in choosing to use the statewide average premium.

See Dep’t of Commerce, 139 S. Ct. at 2570 (“[T]he choice between reasonable policy alternatives in the face of uncertainty was the [agency’s] to make.”). It “considered the relevant factors, weighed risks and benefits, and articulated [several] satisfactory explanation[s] for [its] decision” in the administrative record. Id. No commenter challenged HHS’s decision to use the statewide average premium in the 2014, 2015, or 2016 rules.

24 These explanations overlap in the administrative record. HHS relied on similar reasons for the 2015 and 2016 rules. See, e.g., 2015 Final Rule, 79 Fed. Reg. at 13,753 (“We proposed to use the [2014] methodology in 2015 . . . .”); 2016 Final Rule, 80 Fed. Reg. at 10,760 (“We proposed to continue to use the same risk adjustment methodology finalized in the 2014 [rule] . . . .”).

48

District Court and Budget Neutrality Although the agency provided reasoned explanations for using the statewide average premium, the district court thought the agency should further have explained the budget-neutral design of the program because it “infect[ed]” HHS’s rationales for using the statewide average premium. NMHC I, 312 F. Supp. 3d at 1209. NMHC makes similar arguments in its brief on appeal. We view this issue differently and give HHS the deference the APA requires. See W. Watersheds Project, 721 F.3d at 1273; 5 U.S.C. § 706(A)(2). HHS was not arbitrary and capricious in using a budget- neutral design for its risk adjustment program in the 2014, 2015, and 2016 rules. 25

25 In general, an issue must have been raised before an agency for a party to seek judicial review of agency action on that issue. See United States v. L.A. Tucker Truck Lines, Inc., 344 U.S. 33, 36-37 (1952); Ark Initiative v. U.S. Forest Serv., 660 F.3d 1256, 1261 (10th Cir. 2011). HHS argues that NMHC failed to exhaust any challenge to budget neutrality. See Aplt. Br. at 22-23. It points out that no commenter objected to the budget-neutral design until the 2018 benefit year. See id. at 22. Courts may require issue exhaustion even when, as in this case, neither a statute nor a regulation requires it. See Advocates for Highway & Auto Safety v. Fed. Motor Carrier Safety Admin., 429 F.3d 1136, 1148-49 (D.C. Cir. 2005). In this context, issue exhaustion is not a jurisdictional prerequisite to judicial review. See id. at 1148; see also Avocados Plus Inc. v. Veneman, 370 F.3d 1243, 1248 (D.C. Cir. 2004).

HHS first raised its issue exhaustion defense in its Rule 59(e) motion, which may have been too late under cases like Eaton v. Pacheco, 931 F.3d 1009, 1028 (10th Cir. 2019) (rejecting argument raised “for the first time” in a Rule 59(e) motion). It could have raised the defense earlier in response to NMHC’s discussion in its summary judgment motion that “there is no statutory requirement that risk adjustment be budget neutral.” Dist. Ct. Doc. 33 at 22. But this discussion was part of a broader argument about the statewide average premium. And HHS may not have expected the district court to rely so heavily on budget neutrality for its summary judgment ruling. So it is not clear whether HHS should be faulted for its lateblooming issue-exhaustion defense.

49

As an initial matter, the district court found HHS was arbitrary and capricious for failing to explain “whether budget neutrality was sound public policy.” NMHC I, 312 F. Supp. 3d at 1210. But in its complaint, NMHC challenged HHS’s use of the statewide average premium, not HHS’s budget-neutral design of the program. See Dist. Ct. Doc. 33 at 22. And on appeal, NMHC continues to challenge HHS’s “decision to use the statewide average premium instead of each issuer’s own premium.” Aplee. Br. at 19-22. Rather than address whether HHS adequately explained why it chose the statewide average premium, the court addressed whether HHS provided a rationale for the budget-neutral design. Apart from this discrepancy, HHS was not arbitrary and capricious in using a budget-neutral design.

First, the district court determined “HHS’[s] justifications for using the statewide average premium instead of a plan’s own premium all assume that the ACA requires risk adjustment to be budget neutral, which is not correct.” NMHC I, 312 F. Supp. 3d at 1205. As the district court noted, the ACA neither requires nor forbids budget neutrality. Id. at 1209; see 42 U.S.C. § 18063. But the administrative record

We think the better course is to address the merits of the district court’s analysis of budget neutrality. In addition to arguing issue exhaustion, HHS also contends it complied with the APA in designing the risk adjustment program based on budget neutrality. See Aplt. Br. at 26-34. Because we disagree with the district court’s analysis of budget neutrality, we need not resolve the matter of issue exhaustion. The allegedly unexhausted issue fails on the merits. See Lincoln v. BNSF Ry. Co., 900 F.3d 1166, 1199-1200 (10th Cir. 2018) (considering merits of unexhausted claims on appeal where EEOC exhaustion was non-jurisdictional). Any resolution of HHS’s issue-exhaustion argument thus would not affect the outcome of this appeal.

50

shows HHS would have designed its program to be budget neutral even if it had used the plans’ own premiums rather than the statewide average premium in its formula. See 2011 White Paper at 15 (“Since payment and charge transfers will be budget neutral, a method is needed to balance them if payments are greater than charges or vice versa.”).

Second, the district court determined HHS could have justified using budget neutrality as a “worthy policy goal” but failed to make “such a determination in the record.” NMHC I, 312 F. Supp. 3d at 1205. But budget neutrality was not a policy goal. It was the product of funding constraints. The statute that created the program lacked any funding authorization. 26 NMHC has never pointed to an explicit congressional appropriation for the program. 27 The absence of an appropriation was

26 Compare 42 U.S.C. § 18063 (not authorizing appropriations for ACA’s risk adjustment program) with id. § 1395w-115(a) (establishing “budget authority in advance of appropriations Acts” for risk-adjusted payments under Medicare Part D) and id. § 18042(g) (authorizing appropriations in ACA to provide loans to qualified health insurers).

27 NMHC argues HHS could have used funds appropriated for CMS’s “other responsibilities.” See Aplee. Br. at 30; see, e.g., Dep’t of Health & Human Servs. Appropriations Act, 2014, Pub. L. No. 113-76, 128 Stat. 5, 374-75. But, as HHS explains, to enable health insurers to plan with more certainty, it announced the risk adjustment rules before these funds were appropriated. See Aplt. Br. at 30-31; Aplt. Reply Br. at 8; see also Minuteman, 291 F. Supp. 3d at 202-03. HHS further points out that Congress authorized the states to implement the program and did not authorize funding for the states. See 42 U.S.C. § 18063.

NMHC also argues that “user fees” were available for the program. See Aplee.

Br. at 30; Oral Arg. at 26:20-27:13. Although fees were used to cover HHS’s administrative costs for the exchanges and the risk adjustment program, they were not used for the payments made to high-risk insurers. See, e.g., 2014 Final Rule, 78

51

a matter of public record. See, e.g., Dep’t of Health & Human Servs. Appropriations Act, 2014, Pub. L. No. 113-76, 128 Stat. 5, 363-88; Dep’t of Health & Human Servs. Appropriations Act, 2015, Pub. L. No. 113-235, 128 Stat. 2130, 2466-93; Dep’t of Health & Human Servs. Appropriations Act, 2016, Pub. L. No. 114-113, 129 Stat. 2242, 2600-27. This is likely why no one questioned the budget-neutral design during notice-and-comment rulemaking until 2018. 28 HHS did not violate the APA when it designed the risk adjustment program as budget neutral. It lacked funding to do otherwise. The APA’s requirement that an agency explain its decision applies when the agency exercises its discretion. See State Farm, 463 U.S. at 48 (“[A]n agency must cogently explain why it has exercised its discretion in a given manner . . . .” (emphasis added)); see also United States v. Magnesium Corp. of Am., 616 F.3d 1129, 1144 (10th Cir. 2010) (same). Here, there was no exercise of discretion to explain. In hindsight, perhaps the agency should have explained more fully why it developed a program in which the risk adjustment

Fed. Reg. at 15,412, 15,416-17 (“[T]he risk adjustment user fee [is collected] for the sole purpose of funding HHS’s costs for operating the [f]ederal risk adjustment program . . . .”). To the extent NMHC contends that the charges imposed on low-risk insurers can be characterized as user fees, using them to make transfer payments is compatible with a budget neutral program.

28 At least one commenter challenged the budget-neutral design of the program in 2018. See 2018 Final Rule, 81 Fed. Reg. at 94,101. We do not consider comments made in the notice-and-comment proceedings for the 2018 rule because NMHC’s challenge to that rule is moot.

52

charges and payments netted to zero, such as the rationale for budget neutrality that HHS articulated for the new 2018 rule. But it hardly follows that the agency acted unreasonably. HHS’s actions did not violate the arbitrary and capricious standard.

To the extent the district court’s opinion can be read as criticizing HHS’s reasons apart from budget neutrality, see NMHC I, 312 F. Supp. 3d at 1211-12, our previous analysis addresses why the agency was not arbitrary and capricious in choosing the statewide average premium. 29 Courts cannot “second[]guess” an agency’s rulemaking decision when it provided “reasons for [its] chosen course of action.” Dep’t of Commerce, 139 S. Ct. at 2571. HHS was not arbitrary or capricious in choosing to use the statewide average premium in its formula.

IV. CONCLUSION

We remand NMHC’s challenges to the 2017 and 2018 rules to the district court with instructions to vacate its judgment as to those rules for lack of jurisdiction due to mootness and to dismiss those claims without prejudice. We reverse the

29 The district court said HHS did not adequately explain why the statewide average premium is better than the plans’ premiums to determine risk and encourage plans to enroll sicker individuals. See NMHC I, 312 F. Supp. 3d at 1211-12. NMHC makes similar arguments on appeal. See Aplee. Br. at 32-37. But HHS explained that use of the statewide average premium avoids the need for a “balancing adjustment,” which would contribute to uncertainty and premium variation unrelated to risk. See 2014 Proposed Rule, 77 Fed. Reg. at 73,139. HHS further explained that using the plans’ premiums to determine transfer payments would allow plans to seek higher payments by operating less efficiently rather than accepting high-risk enrollees. See 2011 White Paper at 14; see also 2014 Final Rule, 78 Fed. Reg. at 15,432. Unlike the district court, we defer to the agency’s explanations.

53

district court’s summary judgment ruling for NMHC as to the 2014, 2015, and 2016 rules. 30

30 Because we reverse the district court’s ruling that HHS violated the APA, we need not address the court’s decision to remand with vacatur or the court’s denial of HHS’s Rule 59(e) motion.