I. The ACA's Modern Federalism: Our Findings

The ACA's statutory architecture purposefully involved state policymaking in key features of the law. As drafted, the ACA had many aims, but its primary goal was universal health insurance coverage. To accomplish that goal, the law employs two key mechanisms: expansion of Medicaid eligibility to childless, non-elderly adults, and creation of more robust individual and small group health insurance markets through newly-devised “exchanges.”²

Medicaid expansion created a national baseline of eligibility but relied on states for implementation, as Medicaid always has done. The health insurance exchanges were designed to be state-run, with federal administration as a fallback. These pillars created a national baseline that was meant to cover individuals who had been long excluded from health insurance, but they were also designed to put states in the driver's seat for many policy choices within that national baseline.

Although many call the ACA's federal/state structure a model of “cooperative federalism,” we discovered that this modern health care federalism is significantly more complex than traditional cooperative federalism doctrine acknowledges. We found the ACA's federalism to be dynamic, adaptive, pragmatic, negotiated, and robust in both horizontal and vertical intergovernmental activity. States learned from and leveraged the successes of thought-leader states for gains in their own negotiations. For example, after Arkansas gained approval for the first demonstration waiver to use premium assistance to expand Medicaid eligibility, Iowa, Michigan, and Pennsylvania quickly moved to build on and “one up” Arkansas's successful negotiations, with each state gaining some similar and some different policy concessions from HHS. The federal government adapted each time with the knowledge that states were learning from one another and viewed each negotiation as setting the stage for the next state's demands.

Federal and state officials each brought different pragmatic goals to the bargaining table. While the Obama Administration took a long-term view of entrenching the law — and was willing to compromise policy ideals to achieve that goal — state officials operated with shorter time horizons and took advantage of the Administration's eagerness to leverage their own policy concessions. States have even returned to the bargaining table to seek further concessions when other states win new ones. A recent example that post-dates our study but demonstrates its durability is HHS's approval of Kentucky's waiver application implementing a work requirement for the expansion population, which was followed quickly by new waiver approvals — with slight variations — for Indiana, Arkansas, and New Hampshire. (The validity of the Kentucky waiver is currently being litigated.)³ Although the policy goals of the Trump Administration's HHS may be different, these dynamic negotiations continue the patterns of the first five years of implementation of the ACA.⁴

These ongoing negotiations have not resulted in a federal/state binary but rather have produced a variety of state-led and federally-led models as circumstances warrant. Indeed, the ACA's initial structural architecture has turned out to be a mere starting point for the allocation of policymaking power between the federal government and states. Over the years, pragmatic and creative hybrids of national and state-level solutions have emerged that have allowed the states to remain in control but also to take advantage of the kind of help for which the federal government has economies of scale — such as financial, administrative, and technical assistance. Whereas Congress designed the ACA with an “either/or” vision — that is, with one or 50 policy options in mind — the realities of implementation have revealed a sweet spot somewhere in the middle. For instance, while some states created their own exchanges, more than half used variants of the federal model. Several states copied the exchange models established in other states, in many cases using the same consultants, to avoid reinventing the wheel. As one high-ranking former federal official told us: “We don't need 50 of these things, but we might need eight.”⁵

Moreover, while federalism scholars obsess over common values such as cooperation, disobedience, variety, and autonomy, we found those values nearly meaningless in the context of the ACA. Some states expanded Medicaid from the beginning, others did not expand at all, and others negotiated waivers that rode on the coattails of prior states' successes. Which states were more autonomous or more cooperative? As another example, take the question of whether to operate a state-based exchange. Many states appeared to choose not to run their own exchanges, and a common representation of that choice is that “blue” states cooperated by establishing state-run exchanges and that “red” states rebelled by defaulting to a federally-run exchange out of resistance to the law. This binary is false and seriously oversimplified. For instance, Oregon began with a state-based exchange but then switched to the federal exchange due to technical difficulties.^{Reference Norris6} As a result, Oregon's exchange was structurally the same as Texas's. Did Oregon become more “uncooperative,” “sovereign,” or “autonomous” when it gave up on its state-based exchange and used the federal exchange? Was it as uncooperative, sovereign, or autonomous as Texas? Surely federalism has to be more than a question of attitude.

On the other side, and perhaps counterintuitively given repeated warnings of a federal takeover, we found that states exerted real sovereign power when they implemented the ACA themselves — they were not simply acting as administrators. States enacted hundreds of state laws, enacted new state governance structures, and controlled swaths of health policy due to their inclusion in the ACA's statutory architecture. The alternative — leaving states out of implementation entirely — would have given states no role in health policy whatsoever. It also would have done nothing to preserve the relevance of the state sovereign lawmaking apparatus — the role of state law and state regulation — as ACA implementation has done. Ironically, the states that suffered the greatest power losses are the ones that have refused to engage with implementation at all and so invited the federal government to take over their small group and nongroup insurance markets.

But to be clear, not every state that defaulted to a national exchange opened the door to a federal take-over. We found that some red states, eager to maintain policy control but needing political cover, worked behind the scenes with the federal government, taking advantage of the Obama Administration's eagerness to help, even as these states publically appeared to resist. The ACA was entrenched by these efforts, and states exerted the policy control they desired, but at the expense of political accountability and transparency. One official colorfully labeled this the “secret boyfriend model” of state-national relations: a relationship coveted by the states, but one that states were unwilling to admit publically for political reasons.⁷

Finally, intrastate governance has been a key feature of the ACA's federalism. Each state is an individual republic, so states' unique internal structures have shaped their ACA-related decisions. For instance, some states had preexisting insurance regulations that affected the design of their exchanges as well as laws implementing Medicaid that influenced their negotiations over expanding eligibility. State officials also differ from one another, even within the same state. Governors — dealing with a longer time horizon and more direct accountability to the variety of stakeholders in their states — bucked legislators in their own parties to take advantage of the ACA's benefits. Insurance commissioners worked with HHS to maintain policy control, even as their own states' governors and legislators wished to rebel. These distinctions underscore the diverse priorities of different members of state government and the different structures of the state governments themselves. These internal state dynamics have had a profound influence on national policy implementation and have been largely overlooked in the federalism literature.

II. Implications for Federalism and for Health Policy

Our study deconstructs federalism's commonly named attributes, including sovereignty, autonomy, cooperation, and variety, and illustrates that many common federalism questions are oversimplified, and perhaps unanswerable, in the context of a modern statutory scheme such as the ACA. More specifically, the study challenges the idea that any particular governance arrangement will be the exclusive producer of any particular set of policy values, including the values traditionally associated with “federalism.” As one example, we saw as much policy variation and experimentation within nationally-run exchanges as across state-run exchanges.

Whether the ACA's structural architecture and dependence on the states actually serve health policy is a harder question. Indeed, it is not even clear whether the statute's structural architecture was supposed to serve health policy in the first place, or to serve political expediency, or the structural end of federalism for federalism's own sake. In other words, we must interrogate whether federalism is an end in itself or a means to an end in health care.

In part, this opacity can be attributed to the fact that the core goals of the American health care system have never been established through widespread political or social agreement. Do we all deserve access to health care or do we only get the health care we can afford? This fundamental normative question about the role of the government in health policy has remained unanswered through each effort at health reform from the Nation's founding. The Congress that enacted the ACA, although it moved the needle significantly toward the idea of universal coverage, was likewise unwilling to go all the way toward a unified, fully national program and retained important aspects of existing private markets. It is likely that, politically, this incremental approach was the only way to enact major health reform at the time. But there is little indication that the specific governance structures used were the result of evidence-based health policy choices.

In the end, our data were clearer about the effect of the ACA's governance structure on enhancing state power in health care than about whether the ACA's reliance on state implementation positively affected common metrics of good health policy, such as cost, access, and quality. For example, the ACA's Medicaid expansion as drafted by Congress — which mandated uniform eligibility expansion — was nationalist in structure and would have increased access to care more efficiently than the current, more federalist, structure has done (created when the Supreme Court in NFIB gave states a choice), simply by covering millions more lives. In the exchange realm, the data is unclear as to whether states that ran their own exchange did better in terms of costs, broad networks, or quality than the national exchanges.

Looking forward, as this article went to press, calls for “single-payer” health reform — a uniform national insurance coverage plan — were on the rise and for the first time were gaining political traction. Should the conversation seriously turn to this structural question, the federalism implications of our ACA study have important lessons. First, we have seen how, even within a national governance structure, states can exert enormous power and produce great variety if the statute is designed to allow it. Thus, one question is whether moves toward more nationalist health reform delivery should put states on the front lines as the ACA did. Again, it depends on the goals. State participation can increase variability or not; it depends on statutory design. At the same time, we also saw a federal government that consistently seems to prefer the states' help to implement health reform; it does not appear to want to go it alone. Given these practical considerations, and the entrenched state bureaucracies and expertise in health care, it may be very difficult to eliminate the state role entirely. Nor are we certain it would be wise to do so without empirical examination of when state leadership in health policy produces the best results. One major takeaway of our study is that path-dependent health reform should stop. The next major effort must involve a conversation about desired outcomes, implementation capacity, and reforms grounded in governance structures that have proved successful in producing results.