The Yale Law Journal


The Past, Present, and Future of Section 1115: Learning from History to Improve the Medicaid-Waiver Regime Today

25 Mar 2019

abstract. This Essay explores the use and abuse of section 1115 waivers in the Medicaid program over time. While the original intent of these waivers in Medicaid and Aid to Families with Dependent Children was to allow for experimental demonstration projects to improve local program delivery, they have increasingly been used to accomplish statewide transformations of Medicaid without any experimental purpose. Instead of evidence-based problem solving, the waiver provision has opened the door to ideologically motivated cuts or preconditions on coverage. After exploring the history of the waiver program since its inception in the 1960s, this Essay argues that its critical flaw is federalism gone awry. In response, I argue that these waivers should be viewed through the lens of scientific management, that they should be treated similarly to traditional public health interventions, and that they should return to a more local scope.


Nearly seventeen thousand Medicaid enrollees in Arkansas have lost their coverage since June 2018 because their state—with approval from the Trump Administration—attached work requirements to their Medicaid coverage.1 Nothing in the Social Security Act explicitly authorizes states to require work as a condition of Medicaid enrollment. Instead, Arkansas has implemented this rule by exploiting a little-discussed but often-used waiver authority. Section 1115 allows the Secretary of Health and Human Services (HHS) to waive particular federal requirements of the Medicaid program so that states can conduct “experimental, pilot, or demonstration project[s].”2 It provides sweeping authority to states with few statutory limitations other than the requirement that the project receive the Secretary’s approval.3 Yet HHS has been a lenient gatekeeper; only two of these waivers have been outright rejected in the last decade out of more than sixty-five requests.4 This Essay argues that the result has been waivers, such as Arkansas’s, that use nominally experimental demonstrations to implement ideological, statewide policy change in Medicaid. It chronicles this trend, contending that it is a detrimental abuse of the waiver provision.

Previous analyses of section 1115 have focused primarily on (1) its original intent and its use in the Aid to Families with Dependent Children (AFDC) program;5 (2) its use exclusively in the Medicaid program (particularly within the last decade);6 or (3) its (potential) treatment in courts.7 By contrast, this Essay chronicles the history of section 1115 in both the AFDC and Medicaid programs, and it discusses how that broader history of the statute’s use provides lessons for advocates pushing for better waiver policies in Medicaid today. Further, this Essay uses that analysis to suggest meaningful policy improvements to the waiver regime, as opposed to simply shining a light on its abuses.

Part I of the Essay chronicles the use of these waivers since the 1960s and the ensuing health-care delivery problems they have produced. Part II discusses lessons that we can learn from that history, and Part III diagnoses these problematic waivers as a product of regulatory federalism gone awry. Finally, Part IV provides policy improvements to rein in waiver authority, to realign incentives to ensure that waivers promote what are actually local innovations, and to strengthen Medicaid for the future.

I. the evolution of section 1115

A. Origin and Congressional Intent

Congress enacted section 1115 via amendments to the Social Security Act in 1962.8 When Congress created Medicaid three years later, it subjected the program to the same provision.9 President Kennedy’s endorsement of the 1962 bill provided a clear vision for the waiver authority. First, he called for “imaginative” solutions to problems in welfare programs and suggested that the proposed amendments to the Social Security Act would “help make our welfare programs more flexible and adaptable to local needs.”10 Yet he did not put forward a vision of comprehensive welfare reform on the statewide level.11 Instead, President Kennedy wanted section 1115 to foster innovations that would allow public-assistance programs to effectively deal with small, localized issues in program delivery. Second, anticipating that these waivers could be abused to cut benefits, President Kennedy urged that the amendments instead be used to invest in poor populations.12 He observed that “[c]ommunities which have . . . attempted to save money through ruthless and arbitrary cutbacks in their welfare rolls have found their efforts to little avail. The root problems remained.”13 Investment and experimentation would lead to innovation, which in turn would generate solutions to these “root problems.”14

Congress and the Department of Health, Education, and Welfare (HEW)15 echoed President Kennedy’s vision. Legislative history indicates that Congress expected HEW to primarily waive the statewideness requirement,16 and it did not anticipate awarding many identical waivers to different states.17 In turn, HEW interpreted section 1115 as a means to increase eligibility for federal programs, provide more effective methods of program administration and caseworker training, allow for the purchase of previously unavailable services, and provide supplemental social services such as “home management.”18 The Senate commentary and HEW guidance further emphasized that the waivers were to be both limited in scope and focused on innovation. If these programs were to be truly “experimental,” as opposed to arbitrary exceptions to federal mandates, it would not be necessary to approve a waiver that tests the same intervention in both Indiana and Illinois unless HEW had reason to believe the affected populations were sufficiently different. Further, since an experiment requires a control group, adhering to the statewideness requirement would prevent meaningful experimentation in state programs.19

B. Early Waivers: 1962 to the Mid-1980s

The first decade of section 1115 waivers reflected the limited scope envisioned by HEW guidance. Twenty-five waivers established programs specifically for child development.20 Another sixty or so programs established social services experiments or methods of training caseworkers.21 Four states received waivers to implement demonstration projects offering home and community-based services (HCBS), as opposed to nursing home services, through Medicaid.22 These HCBS programs reduced the number of Medicaid recipients who needed institutional care in nursing homes,23 and spurred Congress to establish a specific HCBS waiver program.24 These waivers were adopted nearly nationwide.25 Several states also began providing Medicaid’s Early and Periodic Screening, Diagnostic, and Treatment (EPSDT) services in schools and day cares.26 Finally, the 1970s and 1980s saw numerous small waivers experimenting with managed care in Medicaid.27 All of these early Medicaid waivers focused on administration and local delivery of services, and in several cases informed future policy in accordance with the statute’s intentions.

Arizona’s 1982 managed-care waiver was the primary outlier during this period, and it foreshadowed the evolution of section 1115.28 Arizona was the last state to accept the Medicaid program, which it did through a waiver.29 Unlike the more limited, localized, experimental waivers recounted above, Arizona’s waiver was a comprehensive statewide waiver that implemented managed care statewide for all beneficiaries. The Arizona waiver was more likely a product of political bargaining to encourage the state’s participation in Medicaid rather than to foster innovation, which would set an unfortunate precedent for subsequent years.30

C. More Managed Care, AFDC, and the Pre-PRWORA Wave

Between the mid-1980s and the passage of the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA),31 the use of section 1115 took a sharp turn. The program’s direct application to Medicaid was essentially the same as the 1980s, and an increasing number of states introduced waivers to put more categories of people on Medicaid managed care plans.32 Yet, as this Section discusses, a more radical use of section 1115 waiver authority began in the AFDC program. While these waivers did not directly affect the Medicaid program, the strategies used by states in adopting AFDC waivers during this period foreshadowed many of the Medicaid waivers states are adopting today. In addition, the rule changes that accompanied the boom in AFDC waivers had significant implications for Medicaid waivers. Few other scholarly works have made an explicit connection between the use of section 1115 in Medicaid and AFDC.33

The political and economic climate provided the critical impetus for this new use of waiver authority. The early 1980s witnessed the ascendancy of fiscal conservatism, with its hostility toward welfare and public assistance. These ideas were further bolstered by an economic recession and the election of President Reagan, who promoted a number of rule changes that altered state incentives. Under the Reagan Administration, the Office of Management and Budget (OMB) became more involved in the waiver process, and implemented a strict budget neutrality rule mandating that section 1115 waivers be cost neutral for every year of the program.34 HHS also removed the requirement that the Institutional Review Boards (IRBs) approve section 1115 waivers, which the federal government had previously required because these “experiments” were conducted on human subjects under the government’s authority.35 These rule changes, in combination with economic conditions of the 1980s, created a perfect storm for conservative waiver approvals: state budgets suffered from the sluggish economy and a refusal to raise revenues,36 the elimination of IRB oversight reduced states’ incentives to adhere to sound experimental practices,37 and strict budget-neutrality requirements made it incredibly difficult to use demonstrations to expand coverage without cuts.38 Thus, states had strong incentives to adopt cost-cutting proposals that fit President Reagan’s politically potent, anti-entitlement policies. Instead of local experiments to form innovative policy, waivers were primed to be used as political tools to advance the Administration’s ideological views on public benefits generally.

The resulting mix of AFDC waivers—endorsed by President George H.W. Bush and later expanded by the Clinton Administration—were comprehensive and statewide, rather than targeted and local. They also cut benefits under the guise of “experiments” to incentivize work, and attempted to regulate the morals of beneficiaries.39 Accepted proposals included time-limited benefits, tighter work requirements, reduced benefits for parents whose children exhibited poor school attendance or performance, “family caps” that decreased assistance for each new child, fingerprinting requirements, and benefit reductions for those moving between states.40 While some liberalizing policies were also approved, such as more generous earnings criteria, limits on vehicle-asset prohibitions, and expanded transitional Medicaid coverage,41 on a net basis the enacted policies significantly cut welfare costs.42 The “success” by which these waivers were measured—removing people from the welfare rolls—came down to whether the waivers minimized expenditures rather than whether they maximized individual attainment of services. The contrast between the scope and effects of these 1980s-era waivers and earlier Medicaid waivers, such as those for HCBS, is striking.

While the early 1990s AFDC waivers were a precursor to PRWORA, they are not an example of “innovation” resulting in new, effective policymaking adopted on the national level. What the federal government learned from the experience of AFDC waivers was not that welfare cuts would incentivize work, but rather that states were seeking to “reform” welfare—and by that they meant to cut costs—whether the federal government acted or not. Susan Bennett and Kathleen Sullivan presciently observed in 1993 that “[b]y the time the lawmakers agree on a plan to reform AFDC, they may no longer recognize the AFDC program that they plan to reform.”43 An “innovation” that entails cutting welfare but does not improve health-benefits delivery does not fall within either President Kennedy’s or Congress’s original understanding of how these waivers were meant to operate.44

The final major development in this era was the Clinton Administration’s relaxation of cost-neutrality requirements in 1994.45 The new rule allowed for section 1115 waivers to be cost-neutral over the life of the program, instead of in each year of its implementation. This was significant for Medicaid because the costs of systematic, statewide reforms were primarily borne upfront. Hence, this further facilitated the adoption of managed care through section 1115.

D. President George W. Bush, the Health Insurance Flexibility and Accountability Demonstration Initiative, and Katrina Waivers

The second Bush Administration’s initial mark on section 1115 was the Health Insurance Flexibility and Accountability Demonstration Initiative (HIFA), a set of waivers that—without increasing available funds—gave states the “flexibility” to expand coverage to formerly ineligible populations.46 Rather than incentivizing expanded coverage, however, because of the lack of additional funding, the added flexibility simply prompted states to reduce benefits to some populations and to fund any expansions in coverage with increased cost-sharing provisions. When coverage was expanded, the expansion populations (such as parents at a slightly higher percentage of the poverty line) were often subject to the highest cost-sharing limitations and limitations on services like inpatient care or family planning.47 For example, Oregon’s HIFA waiver included a $250 co-payment for hospitalization and denials of service for failures to pay premiums.48 These waivers differed from the early 1990s AFDC waivers insofar as they were primarily motivated by the goal of expanding coverage and they did not lower costs for the states—at least facially.49 However, it is clear that the George W. Bush Administration did not push these waivers to test hypotheses and discover new innovations. By that time, states were essentially evaluating themselves, and they varied greatly in how seriously and rigorously they conducted their evaluations.50 In addition, federal comment periods for waivers had ceased.51 Some saw the HIFA program as a precursor, and then a response, to the Administration’s failed attempts to block grant Medicaid.52

The second defining feature of the Bush era was the use of section 1115 waivers to address the effects of natural disasters. These so-called “Katrina waivers,” used to combat the effects of the 2005 hurricane, comprised about thirty-five percent of all waivers approved under President Bush.53 The waivers dealt with the problem of newly uninsured evacuees receiving uncompensated care in states other than their home state by providing temporary Medicaid and State Children’s Health Insurance Program coverage to evacuees based on HHS-recommended income guidelines.54 Recognizing the legally suspect nature of the Katrina waivers, Congress formally ratified them in the Deficit Reduction Act of 2005.55

The expansion of section 1115 authority peaked in the George W. Bush Administration. While past waivers could at least claim some nominal experimental purpose, neither HIFA waivers nor Katrina waivers could be justified in the same fashion. The driving force behind these waivers was not a localized search for innovation, but a top-down implementation of new Medicaid policy at the behest of the executive branch.

E. The Affordable Care Act’s Medicaid Expansion Under Obama and Trump

The Medicaid expansion ushered in by the Affordable Care Act (ACA) and the Supreme Court’s subsequent decision making expansions optional for states created a new battleground for section 1115 waivers.56 Because expansion was optional, states tested how much the Administration was willing to bend in order to incentivize them to accept the coverage expansion. Several of the accepted proposals involved some form of privatization of coverage for childless adults. For example, HHS approved waivers in both Arkansas and Iowa that allowed expansion funds to go towards the purchase of private plans.57 There were also some waivers approved that had far more obvious experimental value. For example, Indiana’s Obama-era waiver included the provision of personal accounts that gained funds based on healthy behavior.58

The Trump Administration, on the other hand, has ushered in an era of section 1115 waivers that harkens back to the era of 1990s AFDC. Since the Republican Party’s efforts to reform Medicaid in Congress have repeatedly failed,59 some states have taken reform into their own hands. Multiple states that had long held out on Medicaid expansion (or reluctantly accepted it) have successfully requested waivers with draconian restrictions on Medicaid recipients. For example, six states have received waivers that allow them to make meeting work requirements a precondition for receiving Medicaid benefits.60 Wisconsin received a waiver with a mandatory “health risk assessment,” which will likely include questions about alcohol and illicit drug use.61 These waivers represent the same types of faux “experiments” that were implemented in AFDC nearly thirty years prior. The experimental value of these new waivers is arguably even more suspect than those from AFDC.62 People need to be healthy to work, not the other way around. The inevitable result of these waivers will be a reduction in eligibility, as those with health issues who cannot get exemptions or cannot produce documentation of employment will lose coverage.

II. lessons learned: patterns in the use of section 1115

Understanding the patterns in the use of section 1115 since its enactment is key to diagnosing the source of abuse and crafting potential reforms. We can learn three lessons from the brief history recounted above: (1) the scope of the statute has grown exponentially; (2) waivers foreshadow national attempts at reform; and (3) the motivation for waivers is not always experimentation and the quest for innovation described in the statute.

The first lesson from this brief history is that the scope of activities permitted by section 1115 has expanded dramatically. Local interventions, caseworker trainings, and administrative innovations no longer form the basis of waiver requests. New Medicaid waivers tend to be wholesale changes to states’ Medicaid regimes, including who gets coverage, what the benefits are, who provides them, and how they are paid for. Despite some lip service to evaluations and experimentation in the Obama Administration, these waivers are almost completely untethered from their original purpose: spurring local innovations in public assistance that can be scaled up.

Second, use of waivers by the states has repeatedly foreshadowed changes in national policy that were often unrelated to any program delivery improvements or efficiencies experienced in waiver states. The first of these developments was when small managed care demonstrations in the 1970s and 1980s gave way to comprehensive, statewide managed care, which in turn spurred the creation of separate, congressionally sanctioned managed-care waivers. While the developments in managed care appeared to adhere more closely to developments anticipated at section 1115’s creation, they proved to be exceptions rather than the rule. AFDC waivers instituting work requirements and other restrictions on cash assistance presaged PRWORA and its Temporary Assistance for Needy Families (TANF) block grant (a broad, federal level cutback in cash assistance that replaced AFDC).63 HIFA “flexibility” came before failed attempts at a Medicaid block grant, and Medicaid work requirement waivers today coincide with repeated attempts by Congress to cut and reshape the Medicaid program.64 The lesson for advocates is that state waiver proposals should be taken seriously because the promulgation of a transformative waiver proposal often foreshadows transformative national policies.

Finally, the history of section 1115 illustrates that the primary motivations for policy changes via waivers are not limited to the terms of the statute. The statute’s original purposes as set out by President Kennedy—innovating in the delivery of health care for the poor and improving health outcomes locally—have often been relegated to justify the policy desires of various institutional actors. Thus, waivers have been shaped by the priorities of the national political parties, the solvency of state budgets, OMB and its budget neutrality rules, the Supreme Court, and the policy preferences of the Chief Executive.

III. identifying the problem: laboratories of democracy gone awry

The three problematic patterns evident from the statute’s history have a common source—excessive deference to state policy preferences. As drafted, section 1115 fits squarely into the idea that states can serve as “laboratories of democracy,” first articulated by Justice Brandeis in his dissent in New State Ice Co. v. Liebmann.65 But the phrase is often incorrectly invoked by commentators to call for an unbridled deference to state or local actors, and to decry federal government intervention in state activities.66 In this view, federal restraints on state policymaking necessarily hampers state innovation. These commentators would view the use of section 1115 during the pre-PRWORA wave and now during the Medicaid expansion debate as an example of successful state policy bubbling up to the federal level. However, this view is incompatible with a correct reading of Justice Brandeis’s analysis. Further, the outcome of many of these waivers—fewer poor Americans eligible for benefits—and the systematic undermining of the statute’s goals illustrate that it is actually deference to states gone awry.

Justice Brandeis’s metaphor is revealing in this context because if the state “laboratories” are working correctly, they should produce evidence-based policies that move the nation towards a national consensus on the next new innovation in Medicaid. As explained by Alan Tarr, the laboratory metaphor pays homage to the theory of scientific management, or the search for “the One Best Way.”67 Under the late-nineteenth-century theory, economic competition needed to be supplemented with a rigorous, scientific theory of production in order to produce efficiency and innovation.68 Thus, a multitude of individuals acting alone is insufficient to achieve progress; these actors must instead operate under certain specific conditions to create a market that promotes innovation. The United States’ experience with section 1115 bears this out. Early demonstrations worked as scientific management would predict. For example, there has been widespread adoption of managed care and the greater provision of EPSDT services after they started as local interventions.69 Yet as waiver authority has expanded, the opposite has generally occurred. States have not formed a consensus on how to best administer the Medicaid program based on waiver innovations. Instead of reaching consensus and improving the program nationwide, section 1115 waivers have subjected the poor to vastly disparate treatment based on where they live and the goals of those at the reins of the Medicaid programs in their states.

Tellingly, commentators observed this trend during the pre-PRWORA wave, and history is repeating itself. Section 1115 waivers for AFDC can be better explained by racial, symbolic politics than the desire to discover innovations in the provision of public assistance.70 A state’s propensity for innovation appeared to have no impact on its likelihood of adopting a waiver.71 Instead, racial, economic, and political factors were predictive. States with lower revenues were more likely to adopt waivers with time limits on AFDC benefits.72 States with seventy to ninety percent African American caseloads were five to six times more likely to adopt a waiver than states with predominantly white caseloads.73 Finally, waiver adoption correlated with Republican control of the executive branch in states where the Christian right has a strong presence.74

Today, states are similarly not “innovating” with work requirements for Medicaid, and similar political and racial lines seem to be motivating waiver adoption. States that have either failed to adopt the Medicaid expansion or are requesting work requirements have been overwhelmingly controlled by Republican governors or legislatures.75 In addition, exemptions to work requirement proposals have thus far favored rural white beneficiaries over urban African Americans.76 There is an imperative, therefore, to learn from the patterns of past waiver adoption to prevent race and politics from impeding waivers that promise actual innovation (as opposed to simply cutting benefits to achieve costs savings), and to ensure that these “laboratories” are functioning as intended. The answer is an approach that rigorously structures experiments and operates under a system of oversight that guarantees proper implementation.

IV. a path forward: policies to improve the use of section 1115

Section 1115 is built on a theory of scientific management, popular at the time of its passage. Fixing section 1115 then requires a return to a vision of the statute that is more faithful to this theory. This section offers three potential reforms to this end: 1) treating waivers more like public health interventions; 2) returning to smaller, localized waiver proposals; and 3) combatting the perverse incentives of budget-neutrality rules.

A. Incorporate Concepts Underlying Other Public Health Interventions to the Regulation of Section 1115 Waivers

HHS regulations should require standards for section 1115 proposals that ensure rigorous analysis of their benefits, aligning them with the standards of other public health interventions. In a 2013 article, Centers for Disease Control and Prevention Director Thomas Frieden discussed six points that are vital to successful public health interventions.77 While these guideposts were designed in the context of more traditional public health endeavors, as opposed to public assistance programs, they provide a useful framework in the context of section 1115. Below, these concepts are adapted to the context of section 1115 waiver programs.

First, there must be a commitment to innovation. Innovations themselves do not have to be as limited as program goals. They can be methodological, evaluative, or operations-based. However, in all cases, a state must work towards innovation by designing its intervention with the purpose of building an evidence base. For every section 1115 waiver, HHS should ask and make public the specific evidence that the state expects the waiver to provide.

Second, states must outline their “technical package”— a detailed description of the program’s scalability and sustainability that prioritizes planning for program management and administration. HHS and OMB’s budget-neutrality rules have already mandated some sustainability in terms of cost. However, this emphasis on cost alone is misplaced; sufficient funding does not guarantee that programs are administrated or managed in an effective way. Funding should accompany a commensurate commitment of political and bureaucratic actors to administer programs. For example, how are states with new work requirements ensuring that beneficiaries do not lose their benefits due to red tape alone? The details have been slim, and Arkansas has been a disastrous test case.78 Effective administration is key to sustainable waiver programs. Scalability is also vital. If demonstrations are effective, states should have a plan in place at the outset to show the Centers for Medicare and Medicaid Services (CMS) and the public how the demonstration could be implemented into its Medicaid program via a State Plan Amendment.79 Note that this involves small-scale experimentation and a foundation of evidence before statewide approval.

Third, states need a plan for “managing performance.” This plan should have monitoring systems in place to understand how waiver programs are functioning in real time and it should outline how the state will react to any problems that arise. If an experiment is failing, how does the state find out and roll it back? HHS should mandate that states have dedicated staff and sufficient manpower to address the issues that may come with implementation.

Fourth, federal regulations should require states to involve local community groups and nonprofit entities in waiver project design. Frieden correctly points out that the involvement of a diverse group of civil actors improves public perception of programs, increases accountability, and fosters effective communication between national and local actors.80 Top-down interventions that begin with statewide waivers naturally have few ties to the communities they affect. Navigators, which are mostly nonprofit entities that assist individuals with signing up for health insurance under the Affordable Care Act, can serve as a model for this type of community involvement.81

Fifth, CMS and states need to emphasize effective communication with communities not only during public comment periods on both the federal and state level, but also as waivers are implemented. Naturally, more localized waivers will help spur more manageable communications from constituents about waiver performance. CMS should mandate that states hold hearings and forums throughout the demonstration project to hear community concerns.

Sixth, there must be political commitment to successful waivers, established through the good-faith implementation of the concepts introduced thus far. Bad-faith waivers, such as the work requirements and drug testing provisions described in Section III.E, can arise when external influences such as party politics and racial biases take precedence over attempts at innovation. HHS must commit to ensuring that waivers are implemented fairly and with the goal of innovation at their core, no matter the political party in the White House. While this will likely require further statutory or regulatory enactments that foreclose opportunities to enact waivers without true experimental value, measures can also be taken through more rigorous enforcement of existing law. For example, section 1557 of the ACA, which provides for nondiscrimination in federal health programs, should readily be enforced in waiver programs to ensure that they are not disproportionately harming protected classes.

These six points should form the basis for new HHS regulations outlining the requirements for section 1115 waiver programs. While specific language or regulatory enforcement mechanisms fall beyond the scope of this Essay, the ultimate goal should be a move towards meaningful standards that allow for explicit judicial review and increased federal scrutiny of state waiver programs once approved.

B. Return to Targeted, Localized Waivers

The history of section 1115 Medicaid waivers highlights the benefits of a return to more localized waivers. With the exception of statewide managed care in Arizona in 1982, the rise of statewide reforms through section 1115 occurred in AFDC in the early 1990s. This change caused significant problems in adhering to effective experimental design and monitoring.82 As intrusive cuts, caps, and limitations on AFDC were implemented at a statewide level, it became almost impossible to establish a control group of recipients to compare to the group receiving the innovative treatment under the waiver.83 Further, monitoring became increasingly complex as experiments required longitudinal data for massive populations, and because data on education and earnings was difficult to obtain.84 Finally, as waivers became larger and more complex, it became nearly impossible to disentangle the effects of various provisions on beneficiary well-being.85 It is unsurprising that meaningful, independent evaluations of waiver proposals were essentially phased out by the end of the 1990s.

The Oregon Health Insurance Experiment is instructive as to why localized waivers are effective. In 2008, Oregon decided to expand its Medicaid program, but held a lottery to determine new beneficiaries because of funding limitations.86 This created a natural randomized experiment, allowing researchers to analyze the effects of having Medicaid coverage. Researchers were able to cabin their study to the Portland metro area to limit logistical problems that often come with statewide data.87 They concluded that Medicaid coverage resulted in decreased diagnoses for depression, increased diagnosis of diabetes, greater use of preventative services, and the “near-elimination” of catastrophic medical expenses.88 The data is publicly available,89 and the experiment has been the basis of eleven other studies and scholarly works.90 While these types of randomized experiments are not easy to come by in health-policy work, they provide proof that thoughtful experimental design can allow states to meaningfully test hypotheses using local waiver proposals.

A return to localized section 1115 waivers requires no legislative or regulatory change. The blessing, and perhaps the curse, of the waiver provisions is that approval and rejection lie entirely in the discretion of the Secretary of Health and Human Services. However, more rigorous regulations concerning acceptable experimental design would be a welcome limit on this discretion. As the Oregon Health Insurance Experiment has shown, implementing Medicaid changes with proper experimental design is both possible and a boon for researchers seeking to establish an evidence base for the efficacy of Medicaid policies.

C. Change Budget Neutrality Rules, Focusing on Outcomes Instead of Dollars

OMB and HHS should relax and amend budget neutrality regulations even further by allowing states to receive additional funding conditioned on meeting specific waiver goals. These can be procedural goals, such as meeting certain experimental design milestones, or substantive outcomes-based goals, such as increasing utilization of specific services by beneficiaries. This Essay’s historical review of waiver proposals indicates that many of the most restrictive waiver eras, including the 1990s AFDC and HIFA waivers, were related to restrictions on spending at both the federal and state level. History shows that strict budget neutrality requirements do more harm than good. The proposed conditions-based framework incentivizes faithful program implementation, and disincentivizes disingenuous waiver proposals that are a mere front for cutting public assistance programs when faced with budgetary pressures.

Critics would likely decry this proposal as opening the door to unlimited federal spending on state programs. Yet, if waivers are small and local, experimentally rigorous, and allow funding increases only on a conditional basis, there should be little opportunity for states to overreach.


A historical review of the use of section 1115 waivers reveals key patterns of abuse and an urgent need for reform. Instead of working to establish an evidence base for new, innovative policies that will improve outcomes for Medicaid beneficiaries, waivers have been repeatedly used as a guise for cuts in benefits and reductions in coverage of the Medicaid program and AFDC. This nonexperimental approach is the antithesis of the scientific management theory of governance at the heart of section 1115.

Admittedly, some effective policies have been implemented through section 1115. New waiver programs continue to emerge in states as an inducement to expand Medicaid.91 Many have likely gained coverage as a result.92 However, like Ulysses, those seeking more waivers more consistent with section 1115’s purpose should tie themselves to the metaphorical mast in order to rein in the abuse of the statute. Reform efforts that knowingly leave the door open for such obvious abuse are simply bad governance. Instead, reformers that seek broader coverage should focus on achieving comprehensive federal healthcare93 as well as grassroots organizing at the state level to expand the Medicaid program in a manner that is faithful to program requirements.94 This approach best ensures that integrity of the Medicaid program is maintained moving forward.

A return to the original vision of section 1115 is long overdue. This Essay has offered common sense, achievable reforms to this end. Legislators and regulators should rein in the use of section 1115 waivers and restore them to their original purpose: creating meaningful innovations that improve outcomes for Medicaid recipients across the nation.

Anthony Albanese is a member of the Georgetown University Law Center J.D. class of 2019. He also holds a B.A. in Government and Economics from Georgetown University. His studies have focused on public-benefits law and policy, with a particular interest in Medicaid and the Affordable Care Act. His experience in health law includes Medicaid policy work at the National Health Law Program, drafting Medicare decisions at the U.S. Department of Health and Human Services Departmental Appeals Board, and providing direct representation to benefits recipients at Georgetown’s Health Justice Alliance Law Clinic. Anthony will spend the next year as a fellow at Legal Aid of North Carolina working to build out the state’s medical-legal partnership in Raleigh. Thank you to Zohaib Chida, Simon Brewer, and the Yale Law Journal staff for their detailed edits and helpful comments.