Indiana Health Law Review Volume XV 2018 Number 1 ARTICLES THE NEW HEALTH CARE FEDERALISM ON THE GROUND ABBE R. GLUCK & NICOLE HUBERFELD * I. INTRODUCTION The central question posed by this symposium concerns the methods of evaluating health policy. Our empirical work on that topic centers on American health care federalism, the relationship between the federal and state governments in the realm of U.S. health care policy and regulation. We recently completed a five-year study tracking implementation of the Patient Protection and Affordable Care Act (ACA) from 2012-2017. The study focused on the two key pillars of1 the ACA, which also happen to be the most state-centered aspects of the law. Those pillars are the expansion of Medicaid—the U.S. health insurance program for low-income individuals, originally established in 1965—and the implementation of the ACA’s new insurance marketplaces, known as “health insurance exchanges.” Our results shed light on what we call the “new health care federalism”—federalism in the modern era of nationally enacted health laws that preserve key roles for state leadership. At the same time, our study reveals the theoretical and empirical challenges of quantitatively evaluating health care federalism at all. What is it really? Does it exist? It is successful? Why do we choose federalism-oriented models in the first place? The full study and its implications are presented in a forthcoming article in the Stanford Law Review . Our purpose here is to offer a more accessible2 snapshot; to discuss the methodological challenges we faced; and to offer in more detail than the longer article one aspect of our methodology: interviews of approximately twenty high ranking former state and federal officials who were at the forefront of the first years of ACA implementation. What we heard from * Professor of Law and Faculty Director, Solomon Center for Health Law and Policy, Yale Law School; Professor of Health Law, Ethics & Human Rights, Boston University School of Public Health & Professor of Law, Boston University School of Law. We are deeply thankful to our interviewees for their insights, candor, and time, and to our colleagues in the Indiana Health Law Program for the invitation. 1. We began the work of studying implementation with colleagues at the University of Pennsylvania, Tom Baker and Theodore Ruger, in a project we detail in the Stanford Law Review. We are indebted to them for their contributions. 2. Abbe R. Gluck & Nicole Huberfeld, What Is Federalism in Health Care For?, 70 STAN. L. REV. _ (forthcoming 2018). http://doi.org/10.18060/3911.0041 2 INDIANA HEALTH LAW REVIEW [Vol. 15:1 those officials corroborates our empirical data and substantiates our conclusions about the defining characteristics, from a federalism perspective, of the ACA’s implementation in the states. Of course, implementation is ongoing and the ACA’s politics remain in flux. Much of what we observed in our study continues to apply to the current context. The implementation process in the states has been 1) dynamic; 2) pragmatic; 3) negotiated; and 4) and marked by intrastate politics. There have been waves of engagement and estrangement between states and the federal government, and state decisions to participate in the ACA’s programs have not been binary, on/off choices. States have moved surprisingly fluidly between models of implementation, and some engaged only partially. States and the federal Department of Health and Human Services (HHS) have each been extremely creative, devising new ways—some not contemplated by the statute’s drafters—to regulate together in a manner that would be politically palatable for Republican governors who concluded that implementing the ACA was in the state’s best interest but at the same time did not wish to appear treasonous to their party by “cooperating” with President Obama. Intergovernmental negotiations have been constant: states have watched other states extract concessions and generous deals from the federal government, then piggybacked on and one-upped those state efforts in their own negotiations. Intrastate politics have also been a central feature throughout, as have state-to-state differences that are often glossed over by federalism theorists who discuss the states as monolithic blocs. Governors often had different (usually more pro-ACA) policy objectives than state legislators. State insurance commissioners also worked around resistant state officials, including some governors, to implement the statute. At the same time, the ACA’s implementation raises deeper questions about health care federalism itself—questions that relate to the methodological question posed by this symposium about evaluating health policy. What, exactly, is health care federalism for? In the context of the ACA, it sometimes appears that health care federalism is not about improving health outcomes at all but rather is “federalism for federalism’s sake”—federalism to advance political or constitutional values, such as reserving power to the states in the interest of sovereignty and balance of power—regardless of the effect on health care coverage, cost, quality, or other standard measures of health care policy success. We cannot evaluate federalism—whether it exists, whether it is working—without knowing what its goals are. As it turns out, we were unable to achieve our goal of quantifying the extent of federalism under the ACA precisely because of this theoretical haziness. We tried, for instance, to measure how “cooperative” the states were—in line with federalism scholars’ preoccupation with the idea of “cooperative federalism”—only to find that concept largely meaningless in this complex context. Some states tried to be dutiful soldiers of the ACA but failed to run their own programs; other states rebelled and refused to run their own programs from the beginning. The federal government stepped in for both. Were such states equally “cooperative”? Were they equally “autonomous”? The same challenges appeared in the context of other metrics, like sovereignty. Were the states that accepted federal funds and expanded Medicaid more or less sovereign than states 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 3 that refused? Were the states that bargained for innovative waivers to expand Medicaid in new ways—for instance, privatizing some of the program—more sovereign than both, or were they weaker capitulators than the states that resisted completely? These questions, we discovered, were impossible to answer. We were left with the sense that federalism functions as both means and end in health care. But tension exists between those functions. We can say more assuredly that the ACA’s federalism served state power than we can say that its federalism produced better health policy outcomes, which was corroborated by our interviewees’ insights. The ACA’s implementation, like other federal programs before it, reveals that traditional federalism policy goals—like policy diversity and experimentation—can be served as easily by nationalist governance structures as by state-centered ones. Consistent with the focus of this symposium3 on “measuring” health care, we believe our study poses a major challenge for the4 field: what is health care federalism for, and on what metrics should its success be evaluated? II. BACKGROUND Both by virtue of its statutory design and the way it has been interpreted by courts and administrators, the ACA’s implementation has been a joint effort between the federal government and states. The ACA as drafted had many aims, but its primary goal was universal health insurance coverage. It aimed to accomplish that goal through two key mechanisms: expansion of Medicaid eligibility to childless adults and creation of more robust individual insurance markets through newly-created exchanges. A. The ACA’s Federalism Design The ACA adopts a federalism-oriented structure in its two main components, namely Medicaid and the exchanges. But the federalism design of each of those pillars is not the same. And it looks different today than it did when the statute was drafted. Medicaid, the public health care program for low-income individuals that is jointly run by states and the federal government, has always been held out as a quintessential example of federalism in administration. Medicaid is a centerpiece of the ACA and was supposed to be significantly expanded by it to address the absence of health insurance options for low income individuals. As drafted, the ACA modified Medicaid’s limited enrollment for the “deserving poor” (which in many states, at the time the ACA passed, did not include low-income childless adults) by expanding eligibility to all adults under age sixty-five with income up 3. See Abbe R. Gluck, Federalism from Federal Statutes: Health Reform, Medicaid, and the Old-Fashioned Federalists’ Gamble, 81 FORDHAM L. REV. 1749 (2013) (collecting sources on how experimentation has come more easily from state programs within federal law than outside it). 4. Indiana Health Law Review Symposium: Ideology Meets Reality: What Works and What Doesn’t Work in Health Law and Policy (Oct. 21, 2016). 4 INDIANA HEALTH LAW REVIEW [Vol. 15:1 to 138% of the federal poverty level (FPL). 5 With respect to the exchanges, these new marketplaces were designed to be state-run. States were given the right of first refusal to operate their own exchanges—something conservative Senators demanded—and the federal government was to be a fallback for states that did not do so. The exchanges serve several functions for the private side of the market, particularly, individuals who do not receive insurance at work but rather must find it for themselves. These new marketplaces act as health insurance clearinghouses. They make the sale of insurance more transparent and efficient, allowing individuals and small groups to compare and purchase health plans that all meet certain basic coverage and quality criteria. They also administer federal tax credits to subsidize the purchase of insurance for individuals earning between 100-400% of the FPL. 6 As implemented, however, the ACA did not fully effectuate the policy goals of its federalism design. The Supreme Court’s 2012 ruling on the constitutional challenges to the ACA, NFIB v. Sebelius, largely upheld the law but took issue with the mandatory nature of the Medicaid expansion. Finding the expansion unconstitutionally coercive for states, the Court reinterpreted the expansion as optional. The decision dramatically changed the federalism dynamics of7 Medicaid implementation, giving states enormous leverage to negotiate for concessions with a federal administration eager to expand the Medicaid program’s eligibility so as to achieve the goal of universal insurance coverage.8 Exchange implementation also did not proceed as envisioned from a federalism perspective. The statutory design of the insurance exchange provisions was supposed to be the mirror image of the Medicaid expansion. The Medicaid expansion was drafted to be universal, a significant federal intervention in historic state choices regarding Medicaid eligibility. The exchange provisions, on the9 5. Patient Protection and Affordable Care Act, Pub. L. No. 111-148, 124 Stat. 119 § 2001(a)(1) (2010) (noting that the HCERA added a 5% income disregard to the 133% of FPL income level delineated in the ACA, raising eligibility to 138% of FPL). See also Health Care and Education Reconciliation Act, Pub. L. No. 111-152, 124 Stat. 1029 § 1004(b) (2010)). 6. See Pub. L. No. 111-148, tit. II. 7. Nat’l Fed’n of Indep. Bus. v. Sebelius, 132 S. Ct. 2566 (2012). Most of the public focus during the case was on the challenge to the insurance mandate—the requirement in the ACA that all individuals become insured or pay a tax. The mandate, which at the time this article went to press had been repealed by the 2017 tax bill, helped to sustain private insurance markets in the face of the ACA’s new requirements for the insurance industry. The ACA requires insurers to cover individuals at essentially equal rates and regardless of medical risk. These and other major changes to the industry’s business model required a source of new funding, as the industry could no longer profit by discriminating on the basis of health and betting on individual risk. The mandate’s purpose was to add new healthy customers to the insurance risk pool. In NFIB, the Court upheld the mandate as a valid exercise of the tax power. 8. In reality “near universal” because undocumented immigrants were deliberately excluded from the law’s insurance coverage reforms. Nicole Huberfeld, The Universality of Medicaid at Fifty, 15 YALE J. OF HEALTH POLICY, L. & ETHICS 67, 68 n. 7 (2015). 9. Nicole Huberfeld, Federalizing Medicaid, 14 U. PA. J. CONST. L. 431, 444 (2011). 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 5 other hand, were designed to be state-led. In the name of federalism, as noted, Senate conservatives had rejected the House’s preference for a new nationally-run exchange, and the ACA enacted the Senate’s preference for state-led exchanges.10 It was widely expected that most states—especially Republican states—would run their own exchanges, rather than allow the federal government to take over their insurance markets. The ACA provided a fallback if they did not: the federal government would operate the exchanges in those states. As in the case of the Medicaid expansion, however, the ACA as implemented saw a dramatic change to the expected federalism structure of the exchanges. The Republican Party intensely politicized ACA implementation in ways unforeseen at the time of the enactment. As relevant to the exchanges, refusing to implement a state exchange became an act of political resistance to the Obama administration and a litmus test for party loyalty. The ironic result was that the very states that had worried most about federal takeover of insurance markets were the same states that refused to implement the law, thereby triggering the federal fallback provisions, and allowing the creation of a national, federally-run exchange in those states. Thus, whereas the ACA was intended to strengthen the federal role in Medicaid and minimize it in the private insurance context, the actual implementation flipped this structure on its head. The Court’s decision in NFIB reinvigorated state power to determine whether and how they would participate in Medicaid, and that attitude bled into the exchange implementation. The ACA built states into its structure for many reasons. Some were at least partly political. For example, giving the states leadership over the exchanges made the ACA’s federal intrusion into the realm of insurance (traditionally a state function) more politically palatable to Senate conservatives. Some reasons were more pragmatic, not least that the federal government did not have the capacity to implement the ACA alone and had an expert body of health administrators waiting in the states, which long have controlled most aspects of health care and insurance regulation. Some reasons were ostensibly policy driven, including a belief in the benefit of local policy variation and experimentation. The flipped federalism dynamics that emerged in implementation complicated these matters and enabled the fluid, negotiated federalism that ultimately has characterized the law. B. Measurement Methodology We devised a three-track strategy in our effort to measure the ACA’s federalism. First, we collaborated with colleagues at the University of Pennsylvania who were developing a quantitative database to track the development of the ACA’s insurance markets. We added a series of criteria on the Medicaid expansion to that project, as well as a series of new federalism- 10. Compare America’s Affordable Health Choices Act of 2009, H.R. 3200, 111th Cong. tit. II.A. (as reported by House, Oct. 14, 2009), with Patient Protection and Affordable Care Act, H.R. 3590, 111th Cong. tit. I.D.II. (as passed by Senate, Dec. 24, 2009). 6 INDIANA HEALTH LAW REVIEW [Vol. 15:1 related questions in the insurance arena. Second, using publicly available11 sources, including government materials and media, we independently catalogued the details of the implementation of Medicaid and the exchanges across the states. We tracked factors ranging from program design, to political party in office, to the legal means—law, executive order, etc.—by which the new programs were implemented in any individual state. Finally, to corroborate that the federalism implications we drew from the data were also perceived in the “real world,” we interviewed the implementers themselves—current and former state and federal officials who ranged from state governors to insurance commissioners to high- ranking members of the Obama Administration. We also interviewed leaders in major healthcare nonprofit and trade groups that were known to be working closely with state and federal officials on implementation. All interviews took place in 2016, were confidential, and were conducted by phone over the course of an hour. As noted, our efforts to quantify “federalism” as part of the University of Pennsylvania project were unsuccessful. But those failures were illuminating themselves, and in large part were the result of observations we drew from our qualitative work. We found ourselves unable to quantitatively measure, for example, the extent to which states retained sovereignty because, in looking at ACA implementation, it became clear that so many different potential weights of sovereignty might be used. Take for instance, the importance of state law. The centrality of the state lawmaking apparatus is a key feature of state sovereignty. Our qualitative study illustrates that states passed tens, if not hundreds, of state laws and regulations to implement the ACA. In contrast, states that resisted implementation simultaneously invited in the federal government to use federal law to design and implement the insurance markets within those states. Which one of these is more sovereign? How can numerical weights be assigned? We faced the same challenges when we attempted to assign weights for “cooperative federalism.” For example, Texas refused to operate its own insurance exchange. Oregon tried to operate its own but failed due to technological challenges and fell back on the federal government. As a result, both states have national exchanges—exchanges that are identical from the standpoint of federalism structure—even though Oregon tried to be a cooperative partner with the federal government. Would both measure equally on a scale of cooperative federalism? Could federalism possibly be just a matter of attitude? Thus, realizing the terrain was nuanced and lacking in defined parameters, the interviews helped to substantiate the descriptive story. We asked the interviewees questions on topics ranging from how much autonomy the ACA gives states, to questions about differences between state and federally run structures; policy variations; state interactions with HHS; and state interactions with one another. 11. HIX Projects, PENN LDI (last visited Sept. 24, 2017), https://ldi.upenn.edu/hix/projects [https://perma.cc/RP33-E54G]. The UPenn HIX 2.0 Project, which is no longer active, aimed to construct quantitatively coded datasets to support research on the impact of variations in state health law and policy choices on changes in outcome measures of significance, such as the rate of uninsured citizens, the number of insurers active in a state market, and health insurance prices. 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 7 Our interviewees consistently told us that the topics about which we inquired were central to their experience implementing the law. As one very high-level former federal official said during our interview: “federalism is everything I did.”12 III. KEY THEMES MADE TANGIBLE Our forthcoming longer article details the four major themes that emerged from our study of ACA implementation. In brief, we found the ACA’s federalism to be 1) dynamic; 2) pragmatic; 3) negotiated; and 4) marked by intrastate politics. It does not operate as an on/off switch. States have moved between state and federal exchange models. Many announced initial resistance to Medicaid expansion only to opt in down the line. These moves, especially in the Medicaid context, were marked by states watching and leveraging the successes of other states for their own gains in negotiations with the federal government. Implementation across both Medicaid and the exchanges also was marked by dramatic state-federal creativity in developing hybrid implementation models that the ACA itself did not envision and that we have not seen theorized before. For example, HHS facilitated exchange models that were neither state based nor fully federal in order to accommodate states that wished to retain control over their insurance market without looking like cooperative traitors to the Republican Party. HHS also conceded on Medicaid waivers more than could have been foreseen, for example allowing states to privatize more of the Medicaid market or to adopt punitive measures to prevent emergency department misuse, to convince reticent states to expand. The implementation details also underscore the importance of intrastate governance, even within a robust federal statutory structure. Each state is a republic of its own, yet “the states” often are discussed as if they are a bloc. State actors have different goals. Governors have different positions from legislators, whose ideas are different from state Medicaid and insurance commissioners, even when they are all of the same political party. The ACA implementation prompted some governors to buck legislators in their own party to take advantage of the Medicaid expansion for their citizens, underscoring state officials’ diverse priorities. In the insurance context, some state insurance commissioners, convinced that keeping the exchanges in the state rather than ceding control to the federal government, worked directly with HHS to get around their own governors’ resistance to operating ACA insurance markets. These workarounds offer a perfect example of the famous concept of “picket fence federalism” on the ground: state and federal administrators in the same area working together on shared goals, often skipping over the preferences of generalist (including more senior) members of their own governments. 12. Telephone Interview with former executive branch federal health care official #5 (Oct. 6, 2016 11:00 a.m.). 8 INDIANA HEALTH LAW REVIEW [Vol. 15:1 A. Federalism as the Central Consideration As one of the most important officials in the Obama Administration on Medicaid policy told us: “Federalism is central to how Medicaid works, it is not off to the side..” All of the officials we spoke with made clear how dramatically13 the NFIB decision had changed the dynamics of implementation. As one state official put it, whereas the “ACA was designed to be an incremental federalization of health care” —in the sense that it was intended to be a federal14 (national and universal) expansion of Medicaid and enlargement of federal control over state insurance markets—NFIB and the ACA’s politics made state negotiations a much more central dynamic of implementation. Even as initially drafted, the officials told us, the federal government knew that the ACA could not have been effectively implemented without the states. As one federal official said: If the legislation was purely federal and federally funded [it] would absolutely have been easier; but the way it played out was more complicated than it had to be because of the Supreme Court decision about Medicaid. That really complicated everything, created extra work and complexity for the federal government. The fact that so many states opted out was not expected in legislation or by the agency, and it produced one of most complicated programs we ever worked on.15 The state policy makers we interviewed consistently praised two key Obama Administration officials, HHS Secretary Kathleen Sebelius and Medicaid Director Cindy Mann, for their state-centered mindsets. “Cindy Mann was a unique leader,” one told us, “she engaged states and moved them with her; she established state operation calls, she reached out to states to get people talking.”16 Another reported: “Cindy Mann was doggedly determined to protect the Medicaid program to make sure it was strong and flourished. She understood the back and forth and the need to let states be as autonomous as she could. It was political and financial realities balanced with [the] highest possible standards.”17 We heard similar praise for Secretary Sebelius. For example: “Sebelius was a governor and really understood the governors’ frustrations. The state legislators’ misunderstandings were clear to her. . . . I was awed at how Sebelius treated every state relationship as so important. Negotiations on waivers were taken so seriously.” One former governor put it this way:18 Sebelius is a hero, she did what she could to get the ACA done. Her background as a governor made it so she knew what states were dealing 13. Id. 14. Telephone Interview with state policy organization officers #1, 2, 3 & 4 (June 6, 2016, 1:00 p.m.). 15. Telephone Interview with former executive branch federal health care official #1 (June 21, 2016, 10:00 a.m.). 16. Telephone Interview, supra note 14. 17. Id. 18. Telephone Interview, supra note 15. 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 9 with every day. She was very attuned to how states function, we had a great working relationship. That attitude of cooperation filtered down to her underlings.19 B. Medicaid We detail in our longer article what we observed, through our tracking data, to be four distinct waves of Medicaid expansion, each marked by different features. The most eager states made moves toward expansion before NFIB interrupted the ACA’s implementation. After NFIB, states began a period of watching and waiting. Some states wanted to implement beyond the ACA’s baseline; others wanted to negotiate their own version of expansion; but all wanted to see what HHS would allow given the ACA's still-existing statutory parameters. Many prominent negotiations centered on states’ efforts to obtain Section 1115 demonstration waiver approvals, through which HHS permits states to deviate from the Medicaid Act so long as the purposes of the Act are furthered and the proposed plan appears budget neutral. Demonstration waivers are20 unique to the proposing state and can facilitate—especially in ACA-resistant states—a sense that “Obamacare” is being rejected and the state has instead implemented its own version of health care reform. 21 1. A Primary Goal of Entrenchment: Negotiations and Horizontal Federalism The interviews confirmed our observations about the four waves of Medicaid expansion. They also confirmed the observation we drew from our data about the tension between the federal governments’ motivation to expand Medicaid and the substance of what that expansion would look like. The federal officials we interviewed emphasized the pull of the ACA’s “goal of universal coverage” on the one hand, and the pragmatic view, on the other, on the part of the federal implementers that the critical short-term goal was to get expansion under way. The view was that short-term malleability would serve the long term goal of universality of health care coverage. In retrospect, it seems those officials were22 correct. As of Fall 2017, the entrenchment of the Medicaid expansion (in whatever form) and its importance to low-income and vulnerable populations across many red states has proved a powerful barrier against Republican efforts under the Trump Administration to “repeal and replace” the ACA. This is not to say the Obama Administration drew no lines. As detailed in the longer article, while more concessions were made in terms of allowing privatization of the Medicaid market and other Medicaid features than would have been predicted, the Obama Administration never capitulated to allow work requirements (something that changed as this article went to press).23 19. Telephone Interview with former governor (Aug. 4, 2016, 12:00 p.m.). 20. 42 U.S.C. § 1315 (2014). 21. Telephone Interview with former federal executive branch health care officials #2, 3 & 4 (Aug. 5, 2016, 12:30 p.m.). 22. Huberfeld, The Universality of Medicaid at Fifty, supra note 8. 23. See, e.g., Thomas Price & Seema Verma, Letter to States, DEP’T OF HEALTH & HUM. 10 INDIANA HEALTH LAW REVIEW [Vol. 15:1 The key theme here, in addition to policy entrenchment, was negotiation. “Medicaid waiver negotiations were intensely a federalism-in-action kind of discussion,” one official said. “They became very one on one, governor versus CMS as to what they wanted, needed, and what the guardrails are.” Several24 interviewees substantiated our perception that the federal government was under intense pressure to make these deals. “Everyone was watching,” one official told us, adding that federal negotiators felt they “could not fail.” 25 Other officials emphasized the ways in which states were watching one another, eager to piggyback on concessions received by earlier states and to do better for themselves. Our data confirm this account. For example, in a post-NFIB wave of Medicaid expansion through demonstration waiver, Arkansas, Iowa, Michigan, and Pennsylvania negotiated with HHS to expand in rapid succession. The pattern of following one another’s successes is not hard to discern: Arkansas openly promoted its negotiations with HHS, which involved the first premium assistance expansion model. Iowa announced interest in a waiver soon after26 Arkansas did and benefited from Arkansas’s application by negotiating for27 SERVS., (March 2017), https://www.hhs.gov/sites/default/files/sec-price-admin-verma-ltr.pdf [https://perma.cc/27PB-BRST]; About Section 1115 Demonstrations, Medicaid.gov, ht tps: / /www.medicaid.gov/medicaid/sect ion-1115-demo/about-1115/index.html [https://perma.cc/JN8S-LXMQ] (last visited Dec. 14, 2017) (describing CMS’s new goal to “[s]upport coordinated strategies to address certain health determinants that promote upward mobility, greater independence, and improved quality of life among individuals”, which translates to work requirements); Letter from Brian Neale, Director of the Center for Medicaid & CHIP Services, Department of Health and Human Services, SMD: 18-002 RE: Opportunities to Promote Work and Community Engagement Among Medicaid Beneficiaries (Jan. 11, 2018), at h t tps : / /www.medicaid .gov/ federal -po l icy-gu idance/downlo ad s / smd 1 8 00 2 .pdf [https://perma.cc/TFP3-V5XG]. 24. Telephone Interview, supra note 12. 25. Telephone Interview, supra note 15. 26. See Robert Pear, States Urged to Expand Medicaid with Private Insurance, N.Y. TIMES (Mar. 21, 2013), http://www.nytimes.com/2013/03/22/us/politics/states-urged-to-expand-medicaid- with-private-insurance.html?pagewanted=all&module=Search&mabReward=relbias% 3Ar%2C%7B%222%22%3A%22RI%3A16%22%2C%221%22%3A%22RI%3A7%22%7D [http://perma.cc/RUV6-DXGY]. 27. Iowa announced openness to negotiating a waiver on February 26, 2013, see Rod Boshart, Branstad to seek federal IowaCare waiver, SIOUX CITY JOURNAL (Feb. 26, 2013), http://siouxcityjournal.com/news/state-and-regional/iowa/branstad-to-seek-federal-iowacare- waiver/article_a5a6b879-66b8-5850-81b7-f6e1364d1296.html [https://perma.cc/B8PA-YXZZ]; and intent to pursue a waiver on March 4, 2013, see Mike Wiser, Branstad releases Medicaid expansion alternative, SIOUX CITY JOURNAL (Mar. 4, 2013), http://siouxcityjournal.com/ news/local/state-and-regional/branstad-releases-medicaid-expansion-alternative/article_c0ab2c08- c443-5f0f-9135-79058fc88bcb.html [https://perma.cc/E2RP-DJGC]. Arkansas publicized successful conversations with HHS to seek a waiver on February 26, 2013. See David Ramsey, UPDATE Medicaid game-changer, ARK. TIMES (Feb. 26, 2013), https://m.arktimes.com/arkansas/ blogs/Post?basename=medicaid-game-changer-feds-approve-putting-entire-expansion-population- 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 11 premium assistance plus other concessions, including enforceable premium payments for individuals earning more than 100% of the federal poverty level, healthy behavior rewards, and copayments for non-emergency use of the emergency department, all of which HHS approved. Michigan initiated28 expansion waiver negotiations shortly before Arkansas’s waiver was formalized, and like Iowa, it sought concessions for cost sharing and healthy behavior incentives, which HHS approved. Michigan also created health savings accounts for enrollees’ cost sharing requirements, which Arkansas later proposed in an29 amendment to its original waiver. HHS approved Michigan’s waiver application30 a few weeks after approving Iowa’s. Soon thereafter, Pennsylvania engaged in negotiations with HHS, originally seeking an Arkansas-style premium assistance expansion but later switching to use managed care networks for the newly eligible population—like Iowa’s waiver. As one official said:31 States absolutely said I want that plus, or with a variation, just approve my waiver. Anything you give to another state you have to assume the next state will come in and ask for that and more. There was a sense on-exchange&day=26&id=ArkansasBlog&month=02&year=2013 [https://perma.cc/HW4P-V3J7]. 28. See THE HENRY J. KAISER FAMILY FOUND., Medicaid Expansion in Iowa (Nov. 20, 2015), http://kff.org/medicaid/fact-sheet/medicaid-expansion-in-iowa/ [http://perma.cc/B6DZ- 94NZ] (explaining that, although Iowa initially required people earning 101-138% of FPL to enroll in its exchange, low insurer participation led the state to offer an option for premium assistance rather than require it). 29. Letter from CMS Admin. Marilyn Tavenner, to Stephen Finton, Dir. of Michigan Medical Services Admin. (Dec. 30, 2013) (approving waiver request and state plan amendment for Medicaid expansion, including cost sharing for beneficiaries over 100% of FPL and MI Health Accounts for maintaining cost sharing funds), http://www.michigan.gov/documents/snyder/ Heal thy_Mich igan _ 1 1 1 5 _ Demo n st r a t i on_Approval_12302013_443466_7.pdf [http://perma.cc/V4AN-PMB6]. 30. Michelle Andrews, Arkansas Weighs Plan to Make Some Medicaid Enrollees Fund Savings Accounts, KAISER HEALTH NEWS (July 22, 2014), http://www.kaiserhealthnews.org/ Stories/2014/July/22/Michelle-Andrews-on-Arkansas-plan-for-Medicaid-savings- accounts.aspx?utm_campaign=KHN%3A+Daily+Health+Policy+Report&utm_source=hs_emai l & u t m _ m e d i u m = e m a i l & u t m _ c o n t e n t = 1 3 5 4 3 0 0 6 & _ h s e n c = p 2 A N q t z - 9 j f n E w 3 T 7 J Y v r m T w h 3 f Z 4 R N S m x k 2 9 1 2 - 7 m l 9 Z X M a B 1 0 V E k L 4 W c 5 L K Q - EAMm1OpJJgjJDVNwibEDTSodynr7CpWDXvsDw&_hsmi=13543006 [http://perma.cc/6QWR- CJXG]. See also ARK. ST. LEG., 2015 Changes to Arkansas’s Private Option: Independence Accounts and Cost Sharing, (June 17, 2014), http://www.arkleg.state.ar.us/assembly/2013/ Meeting%20Attachments/830/I12552/Exh%20%20D%20-PrivateOption-2015.pdf. [http://perma.cc/BEX5-WZ5C]. 31. Under a new governor (Tom Wolf), Pennsylvania abandoned its expansion waiver instead using expansion envisioned by the ACA. See Letter from Marilyn Tavenner, to Beverly Mackereth, Sec’y, Pa. Dep’t of Pub. Welfare, Approving Pennsylvania’s application for a five-year demonstration project (Aug. 28, 2014), https://11042-presscdn-0-63-pagely.netdna-ssl.com/wp- content/uploads/indres/082814healthypaapproval.pdf [http://perma.cc/53X3-KFQA]. 12 INDIANA HEALTH LAW REVIEW [Vol. 15:1 among states that the Administration was so eager that it was best to be last in line.32 From the state perspective, we heard: “On expansion, the biggest question is what the feds will allow. This played out in Arkansas, Pennsylvania, etc. States are finding out how much they can get. . . . Cost sharing is expensive to administer,” continued the official. “There is no evidence that jobs will make any difference. But there is a political need to drive for these things.” A state official agreed:33 Because of the Supreme Court, the Obama Administration bent over backward to get states to expand Medicaid, permitting demonstrations in some states that may not be beneficial to beneficiaries but that allowed some Republican governors to expand when they would not have otherwise . . . [HHS Secretaries] Burwell and Sebelius had the attitude to cover people, even with certain limitations.34 The horizontal state dynamic was not only about one-upping earlier states that had won concessions. We were told that governors share tactics with one another in places where they can talk behind closed doors, such as the National Governors’ Association meeting, and learned from each other’s successes and failures in implementing the ACA. Other interviewees noted that the officials35 in charge of the Arkansas Medicaid expansion were very open in sharing their negotiating tactics with other states. Even federal officials were part of the36 horizontal dynamic. Cindy Mann was singled out by our interviewees for establishing mechanisms for states to share information about Medicaid expansion. Non-government organizations played critical roles, too, in helping to share information across states. Horizontal federalism suffused the Medicaid expansion. 2. Accountability: Pragmatic Solutions, Masking Cooperation, and Intrastate Dynamics Differences across and within states emerged as another key theme that shaped the implementation. States have different mechanisms for implementing a program like Medicaid, which also drove how each implemented the ACA. As one federal administrator deeply involved in waiver negotiations observed: “state expansion decisions reflect huge interaction with internal state rules; there are glaring institutional differences between states.” 37 While differences between states are important, intrastate differences are just as important, and these kinds of differences have been seriously underplayed in the federalism literature. The ACA’s story makes clear how important it is to 32. Telephone Interview, supra note 12. 33. Telephone Interview with health policy nonprofit officer #1 (a former state official) (Aug. 1, 2016, 10:00 a.m.). 34. Telephone Interview, supra note 19. 35. Id. 36. Telephone Interview, supra note 33. 37. Telephone Interview, supra note 21. 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 13 understand the differences across intrastate actors and how those differences shape interaction with the federal government. Governors of both political parties emerged as key partners of the federal government when it came to Medicaid expansion. Recognizing the impact of resisting the Medicaid expansion on state budgets, citizens, and health care businesses, many governors moved to expand even where legislatures of the same party refused. As one former governor put it: The Governor represents the entire state and is there to put a vision out for the state and to implement it and to propose legislation to implement. The legislator is from a small district and we can’t get them all going in same direction. The Governor has to be the leader of the state, he has to be the one to advocate for a vision for moving a state forward. The legislature is a more reactionary type of body, each legislator comes forth with his own bills but usually on small topics rather than big policy measures.38 A member of a key nonprofit group that assisted with implementation agreed, noting: “The big change was the sense that whatever happens in DC, governors have to run a state and are more practically minded. They have to do a budget, people need coverage, hospitals need money, etc.”39 In some instances, these realignments were fraught. For instance, Kentucky Governor Steve Beshear implemented Medicaid expansion through an existing Kentucky law that commanded Medicaid funds to be maximized. The40 legislature argued he could not expand in this manner (administratively and without additional legislative action), but state courts sided with the governor, allowing expansion to proceed. Similarly, Ohio Governor John Kasich asked the state Controlling Board (a state commission that facilitates use of federal funds outside the legislative budgeting process) to approve use of available federal funds for Medicaid expansion. The Ohio legislature had refused to pass a budget41 that included expansion but was bypassed by the Controlling Board in conjunction with the Governor, and state courts upheld the expansion. In Arizona, Governor Jan Brewer would not approve the termination of the legislature’s session until it approved Medicaid expansion, strong-arming rather than working around the legislature. And in Maine, the legislature continuously passed42 Medicaid expansion legislation, only to be thwarted by Governor LePage’s vetoes, until the state became the first to expand it by voter referendum. 43 38. Telephone Interview, supra note 19. 39. Telephone Interview, supra note 33. 40. Ky. Rev. Stat. Ann. § 205.520(3). 41. Meeting Minutes of the Controlling Board, Ohio Office of Budget and Management (Oct. 21, 2013), https://ecb.ohio.gov/public/MeetingsandAgendas.aspx [https://perma.cc/2ANH- 38AP]. 42. See Mary K. Reinhart, Brewer signs into law Arizona’s Medicaid program, THE REPUBLIC/AZCENTRAL.COM (June 18, 2013), http://www.azcentral.com/news/politics/articles/ 20130617brewer-signs-law-arizona-medicaid-program.html [https://perma.cc/46J9-6MSH]. 43. Patrick Whittle, Maine OKs Medicaid expansion in first-of-its-kind referendum, AP 14 INDIANA HEALTH LAW REVIEW [Vol. 15:1 At the same time, even governors who pushed implementation—in both parties—were not always eager to take the credit for cooperating with the Obama Administration. What emerged was a fascinating story about accountability that becomes especially salient in the exchange context, which we detail below. We were told time and again that governors were under pressure to execute waiver deals and come up with versions of Medicaid that did not look like “Obamacare,” which, we were told, was a “curse word.” Further to that point has been the44 practice of renaming and rebranding the state’s version of Medicaid with state- associated names like “TennCare” or “Husky Health.” States have long used such names for their Medicaid programs; they both highlight how state-centered Medicaid is, but also conveniently mask these programs’ status as participants in a major federal program.45 One former federal official closely involved with the Medicaid expansion described it as follows: “The idea was Obamacare is a bad word, we’re not doing Obamacare, it’s ‘healthy blank,’ with the state name inserted. This involved rebranding Medicaid. Charging premiums also falls in this category: it looks more like private coverage, but offers political cover.” Another agreed: “There is a46 political dynamic to the desirability of 1115 [waivers]. It makes governors’ offices sound like they are doing something different and has unique political appeal. States love to say they want to evaluate the waiver to sound innovative.”47 This accountability story is interesting for at least two related reasons. On one hand, the rebranding and ownership of state versions of Medicaid as programs that the public views as associated with the states and not the federal government contributes to a story of how state implementation of federal statutes can further state autonomy and a sense of state sovereignty. On the other hand, these kinds of efforts can confuse the public and mask political responsibility when programs fail or are undesirable. The seven Justices who voted to strike down the mandatory Medicaid expansion in NFIB focused especially on this argument about obfuscated accountability. We return to this topic in the next part.48 C. Exchanges Most of the same themes that developed in the Medicaid context played out in parallel when it came to the exchanges. As noted, the exchanges developed quite differently from what was expected when the ACA was passed. As with other statutes that give states the “right of first refusal” to implement a federal (Nov. 8, 2017), https://www.apnews.com/cf007502a8dc421990c4d8dfcde19235/Maine-OKs- Medicaid-expansion-in-first-of-its-kind-referendum [https://perma.cc/ZAR8-ZE2C]. 44. Telephone Interview, supra note 19. 45. See, e.g., Sarah Kliff, Kentuckians only hate Obamacare if you call it Obamacare, VOX (May 12, 2014), https://www.vox.com/2014/5/12/5709866/kentuckians-only-hate-obamacare-if- you-call-it-obamacare [https://perma.cc/4BS6-SY4D]. 46. Telephone Interview, supra note 21. 47. Id. 48. Nat’l Fed’n of Indep. Bus. v. Sebelius, 567 U.S. 519 (2012). 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 15 program at the state level, the ACA’s drafters assumed many states would prefer to run their own exchanges. Motivation for doing so includes the federalism value of local variation—being able to tailor the program to the needs of the particular state—but also the view that a federal statute encroaches less on state domains when states control statutory implementation of a federal law. This point was49 emphasized by Republicans early in the ACA implementation; one Republican official said letting the federal government operate a state’s insurance exchange was a “Trojan horse” that would pave the way to a full-scale federal takeover.50 But the hot politics of the ACA reversed the usual pattern. The paradoxical result was that the most anti-ACA states were the same states inviting the federal government to take over their insurance markets. And yet, the story on the ground is much more nuanced. Simultaneous with the public resistance and in direct tension with it, many red states quietly have worked under the radar to maintain some policy control, even where the federal government was running the exchange platform in those states. As we detail below, many states have kept these decisions purposefully secret to avoid looking like traitors to the Republican Party. Moreover, whereas on the surface a blue state/red state divide seems to exist in counting which states created their own exchanges, in reality the state-versus- federal-exchange choice has been a landscape as dynamic and adaptive as the Medicaid expansion. For example, Idaho started as a federal exchange because of difficulties with launching an exchange but soon transitioned to a state-based exchange by piggybacking on efforts of leader states. Maryland nearly failed in51 its effort to roll out a state exchange but managed to remain a state-based exchange rather than moving to the federal platform by adopting a state-based platform developed by other states, including Connecticut, which was actively trying to sell its model to other states. Oregon, as noted, started with its own52 exchange but moved to a federally-supported platform due to technical difficulties. Utah split its exchange, putting individuals on a federal exchange and the small group market on a state-based exchange. Kentucky was rhetorically opposed to the ACA at the national political level and yet implemented one of the most successful state-run exchanges under Governor Beshear, until his53 49. See generally Abbe R. Gluck, Intrastatutory Federalism and Statutory Interpretation: State Implementation of Federal Law in Health Reform and Beyond, 121 YALE L. J. 534 (2011). 50. Douglas Holtz-Eakin, Yes to State Exchanges, NAT’L REVIEW (Dec. 6, 2012, 12:00 PM), http://www.nationalreview.com/article/334956/yes-state-exchanges-douglas-holtz-eakin [https://perma.cc/D3UW-2WVT]. 51. Louise Norris, Idaho health insurance marketplace: history and news of the state’s exchange, HEALTHINSURANCE.ORG (Dec. 15, 2017), https://www.healthinsurance.org/idaho-state- health-insurance-exchange/ [https://perma.cc/3VNG-VRC9]. 52. Jenna Johnson, Maryland Looks to Connecticut for Health Exchange Answers, WASH. POST (May 31, 2014), https://www.washingtonpost.com/local/md-politics/maryland-looks-to- connecticut-for-health-exchange-answers/2014/05/31/481a7b9c-db83-11e3-8009- 71de85b9c527_story.html [https://perma.cc/HFS9-TFD3]. 53. See, e.g., Glenn Kessler, What Did Mitch McConnell Mean When He Suggested the 16 INDIANA HEALTH LAW REVIEW [Vol. 15:1 successor closed the state-run exchange and switched to the federal platform. One official emphasized how difficult it was to tell the extent of state cooperation merely from the formal structure it had adopted: “Across the 25 pure [federal-exchange] states it looks pretty much the same, almost completely the same from the outside, but behind the scenes decisions are more shared than it might appear. Like how much plan management is done by the insurance department; it is not obvious that a state is doing plan management from looking at it from the pure federal exchange level.” States had great flexibility in both54 types of exchanges. But “autonomy is not quite the word,” said one of the more senior federal officials responsible for exchange implementation. “The ACA set55 up new relationships with states.” Another official thought that state-run56 exchanges did indeed have “a lot of autonomy.”57 As in the case of the Medicaid expansion, there was enormous pressure on the federal administrators not to fail, to do whatever was needed to get the exchanges up and running, regardless of structure, as was detailed by a senior federal official: The perception was that HHS really wanted states to succeed but it was also dealing with its own extreme challenges. The relationship with all states was critical, many people up to highest level held painstaking negotiations with almost every state, but there were different sets of sensitivities and resources for each state, so it was enormously complex. We were blindsided by so many states refusing to establish their own exchanges. Every day we were negotiating with a state trying to help, to make them come along. Some states might have been hostile publically but not necessarily behind the scenes. HHS was trying to treat each state as an autonomous entity with a unique situation. HHS tried to be flexible.58 We also saw the theme of horizontality emerge in the context of the exchange implementation but differently than in the Medicaid context. States here were not trying to one-up one another’s victories; rather, they worked together and learned from one another. Some of this cooperation was facilitated by formal networks empowered within the ACA itself, such as the National Association of Insurance Kentucky State Exchange Was ‘Unconnected’ to Obamacare?, WASH. POST (May 29, 2014), https://www.washingtonpost.com/news/fact-checker/wp/2014/05/29/what-did-mitch-mcconnell- mean -wh en-h e- sugges t ed - th e-ken tu cky-s t a t e -exch an ge-was-unconnected- to - obamacare/?utm_term=.00de39db49d3 [https://perma.cc/7CYC-2NP3]; Say Again, Senator, ACA Unkynected?, LEXINGTON HERALD-LEADER (May 28, 2014) http://www.kentucky.com/opinion/ editorials/article44490792.html [https://perma.cc/BJP3-DMNM]. 54. Telephone Interviews, supra note 14. 55. Telephone Interview, supra note 15. 56. Id. 57. Telephone Interview with health policy nonprofit officer #1 (a former state official) and #2 (Aug. 1, 2016, 10:00 a.m.). 58. Telephone Interview, supra note 12. 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 17 Commissioners, which was given an explicit role in the statute. Others were59 more informal, such as a technical assistance network facilitated by the Robert Wood Johnson Foundation, Commonwealth Fund, and Milbank. A high level60 official from another major health care nonprofit said she essentially “was running a state to state technical assistance program.” 61 Still other forms of horizontal cooperation emerged from the early appearance of “leader states”—states that had successfully implemented exchanges and were serving as models for others. Connecticut’s entrepreneurial exchange leader told us that Connecticut officials “realized they had invented a better mousetrap,” and that they could “package their services and expertise and make them available to other states.” Connecticut exchange leaders promoted their system as giving62 other states “the benefits of a state-based marketplace without the headaches of building or staffing it,” and argued that they can “package [their] services and expertise and make them available to other states, promoting collaboration and avoiding a duplication of effort.” Perhaps most interestingly, serious horizontal63 state networking was facilitated by the federal government itself. State and federal officials we interviewed commended CMS for facilitating “learning collaboratives” to enable state success in implementation. These collaboratives,64 which have helped states avoid the unnecessary cost and time that would be involved with reinventing the wheel, were not written into the ACA but rather are a pragmatic invention by its administrators. 1. Accountability and Creativity Redux: Intrastate Politics and Hybrids As in the case of the Medicaid expansion, conflicts among intrastate actors were major obstacles to the establishment of some state-based exchanges. Three states’ executives—Kentucky, New York, and Rhode Island—worked an end run around recalcitrant legislatures and created state-based exchanges through executive authority. In other states, governors and insurance commissioners,65 eager to retain control over insurance but facing the wrath of their political party, sought creative workarounds to mask their cooperation with the federal government. HHS was an active partner in these pragmatic negotiations. HHS produced two new insurance-exchange structures not initially envisioned by the 59. See Patient Protection and Affordable Care Act of 2010, Pub. L. 111–148, § 1311(c)(1)(F), 1321(a)(2), 1333(a)(1), 1341(b)(1). See also Christine H. Monahan, Safeguarding State Interests in Health Insurance Exchange Establishment, 21 CONN. INS. L. J. 376 (2015) (describing some of the NAIC’s efforts). 60. See, e.g., St. Health Reform Assistance Network, ROBERT WOOD JOHNSON FOUND., http://statenetwork.org/ [https://perma.cc/P927-NG7M] (last visited Oct. 11, 2017). 61. Telephone Interview, supra note 58. 62. Kevin Counihan, CEO, Access Health CT, Panelist at the Yale Law School Health Insurance Exchange Implementation Conference, Cooperative Federalism Panel (Feb. 8, 2014). 63. Robert Pear, Connecticut Plans to Market Health Exchange Expertise, N.Y. TIMES, (Feb. 24, 2014), http://www.nytimes.com/2014/02/25/us/connecticut-plans-to-market-health-exchange- expertise.html [https://perma.cc/FJ26-D9GF]. 64. Telephone Interview, supra note 12. 65. Id. 18 INDIANA HEALTH LAW REVIEW [Vol. 15:1 law to accommodate states seeking cover for cooperation. HHS developed these new concepts in administrative guidance documents, calling them “hybrid” and “partnership” exchanges. 66 These new types of exchanges blend state and federal management functions and come in many different forms. As one high-level federal interviewee told us, HHS was able to get seven states—Arkansas, Delaware, Illinois, Iowa, Michigan, New Hampshire, and West Virginia—to come on board by giving them more than the binary state-versus-federal choice set forth in the ACA. These states wanted to remain in control over policy but still use the federal government for as much technical and practical support as necessary. In our longer article, we call these hybrids “federalism by necessity.” Some states needed more political cover. One problem was rhetorical. The original version of the blended exchange model was called “partnership,” and some states did not want to appear to be in “partnership” with the Obama Administration. Another problem was the intrastate political arena. Some67 insurance commissioners and other lower state officials wanted to retain control over state insurance markets, even as some governors and legislatures insisted on public resistance. Seven states could not adopt the hybrid model for that reason.68 HHS adapted in response by announcing a new hybrid option called state “plan management” (ditching the politically toxic term “partnership”), which the seven holdout states adopted. Plan management exchanges do not require formal gubernatorial approval but rather require only informal communications between the federal government and state insurance commissioners, thereby allowing state insurance officials to get around resistant state capitols. As one former federal official said: Policies developed because things were not working as the legislation intended. The sense with exchanges was that [HHS] had to be flexible and pragmatic. It was let’s do what we have to do to make this work. If that meant that the exchange got done, then that was far better than having a state give up. We had to make exchanges and the whole system 66. See Timothy Jost, Implementing Health Reform: Moving Forward On Exchanges, HEALTH AFF. BLOG (Jan. 4, 2013), https://www.healthaffairs.org/action/showDoPubSecure?doi= 10.1377%2Fhblog20130104.026901&format=full [https://perma.cc/S9SA-L6QJ] (explaining partnership exchange guidance issued by HHS); see also CTR. FOR CONSUMER INFO. & INS. OVERSIGHT, Regulations & Guidance, CTRS. FOR MEDICARE & MEDICAID SERVS., https://www.cms.gov/CCIIO/Resources/Regulations-and-Guidance/index.html#Health Insurance Marketplaces [https://perma.cc/HJG6-HM58] (last visited 12/14/17). 67. Cf. Elizabeth Weeks Leonard, Rhetorical Federalism: The Value of State-Based Dissent to Federal Health Reform, 39 HOFSTRA L. REV. 111, 162 (2011) (“[R]hetorical federalism acknowledges that federalism arguments have political salience aside from earnest concerns about the federal structure.”). 68. Kansas, Maine, Montana, Nebraska, Ohio, South Dakota, and Virginia were in this group. See Christine H. Monahan, Safeguarding State Interests in Health Insurance Exchange Establishment, 21 CONN. INS. L. J. 376, 423–24 (2015). 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 19 work. It was not going to work if certain states or insurers weren’t playing.69 The same official continued: “If a state was publicly anti-ACA, behind the scenes they might ask for help. Even members of Congress who were publically opposed asked the Secretary for help.” A former state official observed:70 “Administrators’ overarching goals may be different than political goals.”71 Scholars of federalism will recognize in this story the phenomenon of “picket fence federalism.” That term is used to describe when administrators at different levels of government may more closely identify with one another in furtherance of their policy goals than they do with their particular sovereign.72 The accountability implications of this story are fascinating. In some cases, state actors were evading accountability for their decisions from even superior state actors, not to mention the public. One former state official told us that similar issues abounded with the federally run exchanges, noting that states were making decisions “behind the scenes” even in exchanges where states had73 ostensibly ceded control to the federal government. 2. The “Secret Boyfriend” Model of State-Federal Relations The workarounds were pervasive. One senior Obama Administration official offered the most colorful metaphor for the relationship that emerged from them that we heard—“the secret boyfriend.” Political cover is a real thing. . . . We internally joked that some health insurance exchanges were the secret boyfriend model. They couldn’t show it in public that the state partnered with the feds. It is happening in Medicaid too, some governors want to use special terminology to gain political cover for Medicaid expansion.74 The “secret-boyfriend” model of federal-state relations suggests a relationship one covets but is afraid to show publicly. We suspect this model appears in other domains of law, not just health care. But we have never seen it detailed before, or so aptly described. In the context of the ACA, the secret-boyfriend model was an important development that gave states a path to participate in the ACA without appearing to do so, allowing them to save face while also helping their citizenry. IV. CONCLUSIONS Our interviewees’ answers were so consistent that we felt no need to interview more key players after a certain point. No matter which person, which 69. Telephone Interview, supra note 15. 70. Id. 71. Telephone Interview, supra note 33. 72. See Roderick M. Hills, Jr., The Eleventh Amendment as Curb on Bureaucratic Power, 53 STAN. L. REV. 1225, 1227-28 (2001). 73. Telephone Interview, supra note 14. 74. Telephone Interview, supra note 21. https://doi.org/10.2307/1229541 20 INDIANA HEALTH LAW REVIEW [Vol. 15:1 affiliation, or which role they played, our interviewees corroborated our observations that federalism was the focal point of the implementation of the ACA and that the ACA’s federalism implementation has been marked by dynamism, creativity, pragmatism, negotiations, and horizontality. And yet, the ACA’s implementation also reveals how difficult it is to “measure” federalism. It seems to be everywhere, but it is hard to pinpoint exactly what it is, what it is for, and if it has been successful. In large part, that difficultly stems from the fact that many of the conventional federalism metrics and values are far too simplistic for a modern federal statutory scheme that invites state participation for a variety of purposes. As we detail in the longer article, the most commonly-touted federalism values—cooperation, autonomy, sovereignty, and even policy variation—can be seen in the ACA across all of the different kinds of structures, whether state or federal-led. Policy variation, for instance, occurred as much within federally-run exchanges as within state-run exchanges. Autonomy was exercised as much in Medicaid expanding states, as in Medicaid waiver- seeking states, as in Medicaid-resisting states. Cooperation, as we have shown, is a meaningless concept. So, to the question posed by this symposium about evaluating policy in health care, with respect to federalism what we can say is: “It depends.” It seems much clearer to us, as elaborated in the longer piece, that the ACA’s federalism design has advanced state power than that it has achieved good health policy outcomes. If federalism is best understood as a constitutional/structural value—state power as an end in itself—the ACA, especially as implemented, seems to advance it. But a different way we commonly understand federalism is as a means to an end; in other words, why protect and advance state power if it does not serve the public good? When it comes to that perspective, it is not clear that the ACA’s state-75 centered components were necessary to achieve good health policy, or even beneficial for it. After all, leaving the Medicaid expansion to the states has been at least partially responsible for preventing the goal of universal coverage from being achieved. State-run exchanges did not necessarily do better, in terms of access, choice or cost, than federally run exchanges. When we asked interviewees whether they thought outright national control would be simpler, the answer was often yes. But when we asked them if it could76 be possible, the answer was usually no, for both political and practical reasons. And among state officials, regardless of party, a strong sense exists that state 75. See JAMES MADISON, THE FEDERALIST PAPERS NO. 45 (C. Rossiter ed., 1961) (“It is too early for politicians to presume on our forgetting that the public good, the real welfare of the great body of the people, is the supreme object to be pursued; and that no form of government whatever has any other value than as it may be fitted for the attainment of this object. Were the plan of the convention adverse to the public happiness, my voice would be, Reject the plan. Were the Union itself inconsistent with the public happiness, it would be, Abolish the Union. In like manner, as far as the sovereignty of the States cannot be reconciled to the happiness of the people, the voice of every good citizen must be, Let the former be sacrificed to the latter.”) 76. Telephone Interview, supra note 12; Telephone Interview, supra note 15; Telephone Interview, supra note 57. 2018] THE NEW HEALTH CARE FEDERALISM ON THE GROUND 21 control over health policy is a good thing that needs to be maintained—not only for structural reasons but because they are convinced local control leads to better outcomes, a conclusion our study calls into question and about which we think further empirical work is sorely needed. And so, federalism continues to be baked into health care reform, both as an end and as a means, even as those two inputs often conflict.. As one former federal official observed: “it’s not an accident that we keep coming back to cooperative federalism.”77 Our challenge to the health law and policy academy, including ourselves, is to tackle this question of what role federalism is supposed to play in health care and whether it is actually playing that role. Federalism scholarship is often criticized for its tendency toward the theoretical, floating free of real-world detail. Health law scholarship, on the other hand, tends toward “real world” research—a level of gritty detail often unappealing to constitutional theorists—that nods at federalism but does not deeply evaluate its role in health care. We have attempted to join the two worlds through the interviews relayed in this essay and in our other work. The two vectors of federalism and health policy can work at cross purposes. But we also found that the new health care federalism can make a reform effort work—indeed, make it deeply resilient—even in the face of wholly unexpected political and legal hurdles. 77. Telephone Interview, supra note 21.