Within the American system of shared power among institutions, the executive branch has played an increasingly prominent policy role relative to Congress. The vast administrative discretion wielded by the executive branch has elevated the power of the president. Republican and Democratic presidents alike have employed an arsenal of administrative tools to pursue their policy goals: high-level appointments, administrative rule making, executive orders, proclamations, memoranda, guidance documents, directives, dear colleague letters, signing statements, reorganizations, funding decisions, and more. Presidents Obama and Trump employed most of these tools in an effort to shape the implementation and outcomes of the Affordable Care Act (ACA) during its first decade. This article focuses on the Obama and Trump administrations' use of comprehensive waivers to shape ACA implementation. The Obama administration had mixed success using waivers to convince Republican states to expand Medicaid. Compared to Obama, the Trump administration has found it harder to accomplish its policy goals through waivers, but if the courts support the Trump administration's work requirement and 1332 waiver initiatives, it would enable the president to use waivers to achieve an ever broader set of goals, including program retrenchment.
Within the American system of shared power among institutions, the executive branch has played an increasingly prominent policy role relative to Congress. The vast administrative discretion wielded by the executive branch has elevated the power of the president. The rise of an aggressive, partisan, multifaceted administrative presidency fueled especially by Ronald Reagan in the 1980s, has meant that significant policy shifts often occur without congressional approval. Republican and Democratic presidents alike have employed an arsenal of administrative tools to pursue their policy goals (Thompson, Wong, and Rabe forthcoming). These include high-level appointments, administrative rulemaking, executive orders, proclamations, memoranda, guidance documents, directives, dear colleague letters, signing statements, reorganizations, funding decisions, and more (Howell 2003; Lewis 2008; Nathan 1983). Presidents Obama and Trump employed most of these tools in an effort to shape the implementation and outcomes of the Affordable Care Act (ACA) during its first decade. Our focus is on a tool particularly important in the context of federalism: program waivers. This article opens by briefly reviewing the role of waivers in health policy generally and with respect to the ACA more specifically. The next two sections then assay how the Obama administration sought to use waivers as a catalyst for the Medicaid expansion, and how the Trump administration employed this tool in an effort to depress Medicaid enrollments. We then address the strategic posture of the two presidential administrations toward the ACA's section 1332 waivers, which aim to provide states with opportunities to pursue alternative approaches to health reform. A final section assesses the degree to which the waiver strategies of the Obama and Trump administrations facilitated achievement of their respective policy goals.
Program Waivers as a Policy Tool
Waivers are congressional delegations of authority to the federal executive branch to permit selective deviations from the law. The use of waivers in health care was infrequent before President Clinton. Since that time, states have submitted and had approved hundreds of waivers, many of which allowed them to transform their Medicaid programs. States have obtained more focused waivers under section 1915 of the Medicaid statute in order to increase the proportion of long-term services and supports offered in the home and community rather than in institutions.
We focus on waivers that are broader in scope, more likely to kindle partisan differences and to be seen by presidents and state policy makers as vehicles for major policy action. In this, regard, waivers anchored in section 1115 of the Social Security Act, approved by Congress in 1962, give the federal executive branch substantial authority to experiment with alternative state approaches to program delivery. These waivers are supposed to generate policy learning by requiring that demonstrations be formally evaluated. Between the 1990s and the adoption of the ACA in 2010, 1115 waivers permitted states to shift most Medicaid enrollees from fee for service to managed care, expand Medicaid coverage to new cohorts of low-income adults, and reshape their health care infrastructures. About 80% of states had demonstration waivers approved or pending during the ACA's initial decade, with about one-third of federal Medicaid expenditures supporting waiver-based activities (GAO 2019a). These waivers figured prominently in the administrative strategies toward the ACA of both Presidents Obama and Trump.
The ACA added another major waiver to the tool box of the administrative presidency. Section 1332 of the ACA authorizes the Department of Health and Human Services and the Department of the Treasury to waive the law's insurance exchange provisions so that states may provide alternative approaches to expanding high-quality coverage in the individual and small-group health insurance markets. It requires state waiver proposals to offer insurance as comprehensive and affordable as ACA plans, and not to decrease the number of people who have coverage. The waivers must be authorized by state legislation and approved by the governor; they must not increase the federal deficit. Supporters of the ACA saw the 1332 waivers as an opportunity to explore other approaches to universal coverage. Officials in Vermont, for instance, planned to use a 1332 waiver to promote single-payer reform (McDonough 2015). But a president seeking to undermine the ACA may also turn to innovation waivers as a tool.
Different uses of demonstration and innovation waivers were on display during the ACA's first decade. President Obama used 1115 waivers as a platform for negotiating the ACA's Medicaid expansion in states that would otherwise resist such action. In dealing with the 1332 waivers he sought to establish guardrails that would make it hard for states to water down the ACA's insurance standards. The Trump administration's actions with respect to the demonstration and innovation waivers sought to substantially undermine the ACA. In the case of the Medicaid expansion population, the Trump administration invited states to submit 1115 waivers imposing work requirements and other administrative burdens on enrollees that would cause many of them to lose coverage. As for 1332 waivers, the Trump administration encouraged states to submit proposals that might not meet the ACA's requirements to cover people with preexisting conditions, provide the law's 10 essential health benefits, or cap the out-of-pocket expenses of enrollees.
Obama: Demonstration Waivers as Catalysts for Medicaid Expansion
The ACA's goal of expanding health insurance coverage depended heavily on Medicaid. The law mandated that, with certain exceptions, all nonelderly, nondisabled people with incomes up to 138% of the poverty line would qualify for Medicaid. It called for the federal government to pay the tab for the newly eligible for 3 years starting in 2014. In 2017, this federal match would decline, leveling off at 90% in 2020. States that refused to comply with this mandate risked having funding for their existing Medicaid program reduced. But in National Federation of Independent Business v. Sebelius, the Supreme Court in 2012 ruled that the federal government could not eliminate funding for existing Medicaid programs if states refused to join the ACA's Medicaid expansion. This decision effectively converted the expansion from a federal mandate to a state option and presented the Obama administration with the challenge of persuading resistant states to expand the program.
The partisan divisions over the ACA meant that the administrative presidency, rather than the career bureaucracy, led the negotiations between the federal and state governments. Both administrations saw the implementation and success (or failure) of the ACA as closely linked to broader party goals, so the White House wanted more direct involvement than was the case for many Medicaid waivers. In doing so, Obama faced the problem that Republicans dominated most state governments: in 2013, Republicans controlled the governorship and both houses of the legislature in 24 states; 6 of the 12 states with divided governments had Republican governors. National party leaders and related partisan networks urged these policy makers to shun more pragmatic policy and administrative considerations in weighing their responses to the ACA and instead join national partisan efforts to derail the law's implementation.
In waiver negotiations with governors, the Obama administration stood firm on certain matters. Several Republican governors pushed the Centers for Medicare and Medicaid Services (CMS) to accept a partial expansion of Medicaid, such as covering those up to 100% of the poverty line rather than 138%. In December 2012, however, the Obama administration announced that it would not approve partial measures. The White House gambled that, in doing so, the ACA's generous federal match, as well as lobbying pressures from providers and others, would eventually encourage most if not all states to expand Medicaid. Indeed, the administration did not merely hope that these groups would pressure state governments to expand Medicaid; it encouraged their mobilization.
Meanwhile, the Obama administration used section 1115 waiver authority to facilitate expansion. This included granting waivers to supportive states to jump-start implementation of the law prior to 2014. In late 2010, for instance, CMS approved a Medicaid waiver for California called “Bridge to Reform,” which provided federal matching funds to counties in the state that got a head start implementing the ACA's Medicaid provisions. The Obama administration also sought to use waiver renewal processes to incentivize Medicaid expansion. It strove to phase out waivers in states such as Florida that had diverted Medicaid monies into subsidies for hospitals rather than insurance for individuals. It saw no need to continue the hospital pools when these states could address problems of uncompensated care by expanding Medicaid.
But the primary Medicaid expansion strategy of the Obama administration centered on encouraging state Republican policy makers to submit new waiver proposals. The interest of conservative policy makers in market-oriented waivers gave the Obama administration an opening. The waivers pursued by these states sought to push Medicaid more fully toward a model that stressed private insurance, competition among providers, individual choice, consumer empowerment, and personal responsibility. The market-oriented model in varying degrees emphasizes two broad themes.
One theme is individual choice and personal responsibility. This variant stresses that Medicaid enrollees will behave more responsibly if they have some “skin in the game” when making health care decisions, such as by paying premiums, having copays for services, or managing health savings accounts. It seeks to discourage “inappropriate” care through greater cost sharing when enrollees use hospital emergency departments for nonurgent services. It also uses economic incentives to reward enrollees if they engage in certain desired health care behaviors (e.g., waiving premiums if they get an annual wellness exam) or other activities such as employment training or work. Under Obama, Indiana, Michigan, and Montana received waivers emphasizing individual choice and personal responsibility as a condition for expanding Medicaid.
A second theme of some market-oriented waivers is premium assistance. This approach involves the use of Medicaid monies to purchase coverage from private insurance companies on the ACA's exchanges. The Obama administration approved waivers in Arkansas, Iowa, New Hampshire, and Pennsylvania (where a newly elected Democratic governor terminated the waiver in 2015) that stressed this kind of assistance. In varying degrees, these waivers allow individuals in the Medicaid expansion cohort to seek coverage on the exchanges rather than sign up for the state's Medicaid program. The Obama administration faced the challenge that insuring people on the exchanges would likely cost more than enrolling them in Medicaid, which violates the federal requirement that demonstration waivers be budget neutral. Following in the tradition of earlier presidents, the Obama administration finessed this problem by accepting generous assumptions from the states about the cost savings their waivers would produce.
In using waivers as carrots for Medicaid expansion, the Obama administration did not grant all state requests. It resisted efforts by Republican governors in Arizona, Indiana, Kentucky, Ohio, and Pennsylvania to impose work requirements on Medicaid applicants. In other cases, the Obama administration rejected enrollee cost sharing that it considered excessive; it declined some state measures that would increase the administrative burdens of Medicaid enrollment processes. The use of waivers as well as other policy dynamics led to slow but steady progress in increasing the number of Medicaid-expansion states. By the time Obama left office, 31 states and the District of Columbia had expanded Medicaid.
Trump: Demonstration Waivers as Catalysts for Enrollment Erosion
Like Obama, the Trump administration viewed 1115 waivers as a tool to pursue its policy goals without congressional action. This became particularly important after the Republican-controlled Congress failed to adopt a “repeal-and-replace” law in 2017. Even before that, in March Tom Price, secretary of health and human services at the time, sent a letter to all governors encouraging new 1115 waiver submissions (Price and Verma 2017). The letter endorsed premiums and other enrollee cost sharing, as well as fees penalizing enrollees who use hospital emergency departments for nonurgent care. It opened the door to state initiatives that would reverse Obama administration efforts to increase Medicaid take-up rates. For instance, it welcomed state restrictions on “presumptive eligibility,” which health care providers had used to obtain Medicaid funding for uninsured patients who appeared to qualify for Medicaid but had not yet formally applied. The Trump administration's waiver themes emphasized work mandates, greater premiums, more extensive reporting, and a gaggle of other requirements for the able-bodied, nonelderly adults targeted by the ACA's Medicaid expansion. These requirements would greatly increase the administrative burdens on Medicaid applicants and beneficiaries and generally depress program enrollments (Moynihan, Herd, and Harvey 2015).
In November, Seema Verma, who had been confirmed as administrator of CMS, more fully developed these themes. Speaking at a meeting of state Medicaid directors, Verma (2017) promised a “new day” for Medicaid, stressing that the ACA had caused the program to stray from its core mission of helping society's most vulnerable members. Doing so had stretched “the safety net for some of our most fragile populations, many of whom are on waiting lists for critical home-care services while states enroll millions of newly eligible, able-bodied adults” at an enhanced match rate of 90%. Having portrayed the expansion population as undeserving and a threat to more meritorious cohorts of Medicaid enrollees, Verma went on to endorse state flexibility to impose work requirements on able-bodied adults. In her view, these requirements would help these enrollees “break the chains of poverty” and surmount “the soft bigotry of low expectations consistently espoused by the prior administration.” Amplifying this theme, CMS prepared a new website that no longer included “increase and strengthen overall coverage of low-income individuals” as a Medicaid waiver goal (Cohen 2017). On January 11, 2018, CMS sent a letter to state Medicaid directors justifying work requirements on grounds that they would “promote better mental, physical and emotional health” for beneficiaries (CMS 2018).
As of December 2019, 10 states (Arizona, Arkansas, Indiana, Kentucky, Michigan, New Hampshire, Ohio, South Carolina, Utah, and Wisconsin) had obtained CMS approval for work requirement waivers, with another 8 awaiting the agency's decision on their proposals. (After the 2019 elections, Democratic governors in Kentucky and Virginia announced their intention to withdraw their respective work waivers.) All but 2 of the 18 states with approved or pending waivers had a Republican governor at the time of waiver submission; all but two had voted for Trump in the 2016 presidential election. Waiver proposals varied, but they typically mandated that a specified cohort of nonelderly, able-bodied adults be employed for at least 20 hours per week. The work requirements primarily targeted those gaining eligibility under the ACA's Medicaid expansion. But such nonexpansion states as South Carolina and Wisconsin also won approval to impose them. “Work” generally encompassed such activities as employment, education or training, and job search. The waivers also tended to allow enrollees to meet the requirement through various forms of community engagement, such as volunteering with a nonprofit group.
While state work requirements would sap Medicaid enrollments, the degree of decline also depended on other administrative burdens embedded in the waivers. Reporting requirements comprised one such burden. These requirements become more burdensome if they must be done more frequently, call for more submitted information, or limit communication channels for enrollee reporting. Several of the approved work waivers featured burdensome reporting requirements. Arkansas, for instance, required able-bodied enrollees aged 19–49 to report the status of their work efforts on a monthly basis and limited the venues for communicating this information. Though many enrollees had no or limited computer skills, they had to report electronically through a special state web portal. If enrollees did not have access to the internet at home or through mobile devices, county offices would have portals available during work hours. To use the portal, an enrollee needed an e-mail address, a password unique to the portal, and a reference number that Medicaid officials had sent in a multipage letter. Once linked to the portal, enrollees had to click through several different screens to report their work and community engagement activities.
The requirement for electronic filing promised to pare the administrative costs of the work initiative. Although CMS does not take administrative costs into account in determining whether a waiver proposal is budget neutral, these costs tend to be substantial. Arkansas had incentive to rely exclusively on an electronic portal because the federal government covers 90% of state administrative costs for pertinent information system development while subsidizing only half of most other administrative costs (GAO 2019b).
This reporting structure precipitated declines in Medicaid enrollments. The Arkansas waiver stipulated that beneficiaries who failed to comply with work requirements for 3 months would be locked out of coverage for the remainder of the calendar year. In September, 3 months after waiver implementation began, state officials announced that over 4,300 people would lose coverage for failing to report. By the end of 2018, nearly 17,000 had lost coverage (Rudowitz, Musumeci, and Hall 2018).
Lock-out periods in many of the approved waivers also place burdens on enrollees. These periods refer to the length of time a beneficiary loses coverage for failure to comply with Medicaid requirements. For example, the Kentucky waiver established a 6–month disenrollment for beneficiaries who did not promptly report changes in income or employment status. Former enrollees could shorten the lock-out period by completing a financial or health literacy course.
Some of the approved waivers also impose new financial burdens on enrollees through cost-sharing provisions. Medicaid law permits some enrollee cost sharing but generally prohibits states from charging premiums to beneficiaries with incomes below 150% of the poverty line. Some of the approved waivers departed from that norm. For instance, Kentucky's waiver required enrollees who were at 0–25% of the poverty line to pay a monthly premium of $1; those in the 25–50% range had to pay $4 monthly.
CMS was not responsive to all initiatives embedded in the state waiver proposals. It did not, for instance, sign off on Wisconsin's proposal to impose mandatory drug screening for poor, childless, able-bodied Medicaid applicants ages 19–64, nor did it accept that state's proposal to establish a 4–year time limit on eligibility for this Medicaid cohort. CMS also turned down several state proposals to reduce Medicaid eligibility for the expansion population from 138% of the poverty line to 100%. While showing some restraint with these and other requests, however, the Trump administration's waiver approvals clearly threatened to erode Medicaid enrollments, and litigation to block these initiatives soon emerged.
Opponents filed suits over 2 years to derail the work requirement waivers, initially in Kentucky and Arkansas and then in New Hampshire, Indiana, and Michigan. In January 2018 three advocacy groups, the National Health Law Program, the Kentucky Equal Justice Center, and the Southern Poverty Law Center, sued Kentucky on behalf of 15 state residents (Stewart v. Azar, 308 F. Supp. 3rd, 239 [US Dist. 2018]). The suit alleged that the Trump waiver aimed to “comprehensively transform” and “re-write the Medicaid Act” while bypassing Congress. The plaintiffs attacked multiple waiver provisions as “arbitrary and capricious.” They claimed that CMS had violated the Administrative Procedure Act by announcing its new orientation toward Medicaid waivers via a letter to state officials rather than the formal rule-making process. The plaintiffs charged that the new waiver policy also violated the president's constitutional duty to “take care that the laws be faithfully executed”(First Am. Class Action Compl. for Declaratory and Injunctive Relief, Stewart v. Azar, 308 F. Supp. 3rd [US Dist. 2018] [No. 18–152 (JEB)]). The Kentucky challenge along with the Arkansas and New Hampshire suits were heard separately by Judge James Boasberg (an Obama appointee) in a district court of the DC Circuit. In a series of decisions halting implementation of three of the waivers (with the suits against Indiana and Michigan pending), Boasberg did not explicitly address whether Medicaid law permitted work requirements or other provisions. Instead, he vacated the waivers on grounds that they were likely to produce enrollment declines inconsistent with the coverage goals embedded in the Medicaid statute. He noted that the Kentucky waiver proposal had projected a substantial enrollment decline. Moreover, one study estimated that 600,000–800,000 enrollees would lose coverage if nine states with approved work-requirement waivers implemented them (Ku and Brantley 2019). As a result of Boasberg's rulings, the implementation of the work requirement waivers ground to a halt by the end of 2019. The Trump administration unsuccessfully appealed these decisions to the DC Circuit and the case might eventually land on the docket of the Supreme Court. If Trump officials ultimately prevail in the courts, it would greatly enhance the utility of waivers as a policy tool for the administrative presidency.
Meanwhile, the Trump administration announced its intention to consider state demonstration waiver proposals that would block-grant Medicaid rather than preserve it as an entitlement. Depending on their specifics, these waivers could also erode Medicaid coverage. In early 2020, CMS issued guidelines that invited states to submit block grant proposals.
The 1332 Innovation Waivers: Weakening the Guardrails
The ACA authorized states as of 2017 to propose section 1332 waivers to pursue innovative alternatives to achieving the law's goals. The Trump administration's initial use of these waivers departed from its more general sabotage efforts (Thompson, Gusmano, and Shinohara 2018). In his letter to governors in early March 2017, Secretary Price invited states to submit proposals that would slow the rate of premium growth and improve the exchange risk pools. The letter pointed to “reinsurance” as one way to accomplish this. This approach offered subsidies to insurance companies for covering people with more costly medical needs, allowing the exchanges to set lower premiums that would presumably help them attract healthier enrollees. In mid-2017, the agency approved a 1332 waiver proposal from Alaska that allowed the state to use federal funds to defray the cost of a reinsurance program. The lower premiums that resulted would reduce the tax credits that the federal government provided to exchange enrollees with incomes below 400% of the poverty line. By late 2019, 11 more states (Colorado, Delaware, Maine, Maryland, Minnesota, Montana, New Jersey, North Dakota, Oregon, Rhode Island, and Wisconsin) received CMS blessings to use reinsurance. The successful proposals came from relatively liberal states, with 8 of the 12 having voted for Hillary Clinton in 2016. We cannot explain why the Trump administration abandoned its overall strategy of sabotaging the marketplaces to buttress them through reinsurance waivers. It may be, however, that principles of federalism mattered, where the administration's declared commitment to greater state flexibility overrode its desire to undermine the exchanges.
Subsequent 1332 actions by the Trump administration were more consistent with sabotage. As noted earlier, the Obama administration had established “guardrails” for these waivers in 2015 to protect the ACA from erosion. The guardrails insisted that state alternatives provide coverage at least as comprehensive and as affordable as the ACA. They also stipulated that the waiver insure at least a comparable number of state residents and not increase the federal deficit. In October 2018, CMS issued new guidelines to the states to replace the Obama guardrails. The new guidelines changed the name of the initiatives to “State Relief and Empowerment Waivers.” They prioritized access to coverage over the quality (“comprehensiveness and content”) of insurance in reviewing state proposals. The guidance encouraged state proposals that would ask the federal government to subsidize enrollment in non-compliant-ACA products, such as short-term and association health plans. These plans would not necessarily cover the ACA's 10 essential health benefits or provide affordable access to care for those with preexisting conditions. They heightened the risk that healthier people would exit the exchanges to purchase cheaper, lower-quality coverage, worsening the exchange risk pools (Cohen 2018). The Obama guardrails had also insisted that the legislature and the governor sign off on a waiver proposal. The Trump administration relaxed this requirement and made it easier for a governor to submit the waiver application without specific legislative authorization.
The degree to which the new CMS guidelines would fuel flight from the ACA-compliant exchanges to lower-quality plans (what Democrats call “junk insurance”) depended on state willingness to pursue such waivers. Six months later, no state had requested waivers in response to the new section 1332 guidance. Concerned by this lack of interest, CMS Administrator Verma in late April 2019 implored states to take advantage of the new waiver flexibility (Lotven 2019). Finally, Republican Governor Brian Kemp of Georgia announced in early November that his state would accept the CMS invitation. Among other things, the Georgia proposed 1332 waiver would allow individuals with incomes below 400% of poverty to use their federal tax credits to subsidize the purchase of noncompliant ACA coverage, such as short-term health plans (Cohen 2019). Whether the Georgia proposal, when submitted and formally approved, will survive a certain court challenge remained an open question as 2020 dawned.
Waivers have helped fuel the growth in executive branch discretion to shape who gets what in health care. The professional bureaucratic complex, consisting of state and federal officials with similar educations and work experiences, at times dominates the exercise of this discretion (e.g., as with section 1915 waivers). As demonstrated in this article, however, decisions concerning waivers can also become part and parcel of the administrative presidency and swayed more by partisan preferences. Both the Obama and Trump administrations strove to use comprehensive waivers to shape ACA implementation.
The Obama administration experienced mixed success in employing this tool. Its deployment of market-oriented demonstration waivers helped entice seven otherwise reluctant states to participate in the Medicaid expansion at all or to do so earlier than they otherwise would have. This “reluctant” waiver cohort comprised over a fifth of all Medicaid expansion states by the time Obama left office. This participatory outcome proved durable at least through the first 3 years of the Trump administration, as none of the states had withdrawn from the expansion. The Obama administration won this victory without making major concessions on work requirements, cost sharing, or partial expansions that certain states had sought. As for the 1332 waivers, the Obama administration forged guardrails that would make it difficult for states to erode the ACA through innovation proposals. But waivers hardly proved to be an elixir for the Obama administrative presidency. Most Republican-dominated states evinced little or no interest in negotiating demonstration waivers leading to Medicaid expansion. Other Obama efforts to incentivize Medicaid expansions by cutting back on waivers allowing special hospital pools for uncompensated care in nonparticipating states did not have the desired effects. Moreover, the Trump administration promptly reversed the policy. And, Trump officials moved apace to weaken the 1332 guardrails that Obama officials had established to protect the ACA from erosion.
Compared to Obama, the Trump administration has found it harder to accomplish its policy goals through waivers. Its biggest success has been encouraging the diffusion of reinsurance waivers, which have damped down premium increases and generally bolstered the exchanges. Beyond that, two of its major waiver initiatives foundered. First, despite widespread interest in work requirement waivers among conservative states and CMS willingness to approve them, the federal courts have blocked their implementation. Second, the Trump administration's weakening of Obama-era 1332 guardrails to invite states to promote cheaper noncompliant ACA insurance in their marketplaces had, with the exception of Georgia, triggered minimal state interest by the end of 2019. Of course, this portrait of the “Trump administrative presidency stymied” might over time morph into a picture of “Trump triumphant.” If the courts support the Trump administration's work requirement and 1332 waiver initiatives, it would signal their approval of a highly expansive reading of pertinent statutes by the executive branch. It would raise the question of what, if any, congressional statutory limits apply to what can be done through waivers. It would enable the president to use waivers to achieve an ever broader set of goals, including program retrenchment.