This analysis is part of the USC-Brookings Schaeffer Initiative for Health Policy, which is a partnership between Economic Studies at Brookings and the University of Southern California Schaeffer Center for Health Policy & Economics. The Initiative aims to inform the national health care debate with rigorous, evidence-based analysis leading to practical recommendations using the collaborative strengths of USC and Brookings.
In late December, Georgia submitted a proposal to the federal government to reform the state’s health insurance market under a Section 1332 waiver. The waiver proposes, among other things, replacing the premium tax credit under the ACA with a new state-operated financial assistance program that can be used to purchase coverage with far higher cost-sharing than would otherwise be allowed under the ACA. Further, the state would cap the total amount of financial assistance it provides and would put potential consumers on a “waitlist” once the cap is reached.
A waiver structured in this way does not meet the requirements laid out in Section 1332 that waivers must provide coverage that is just as affordable and comprehensive to just as many people as would have it absent the waiver, all without increasing the federal deficit. And the analysis accompanying Georgia’s waiver includes errors both obvious and subtle that mean that the waiver’s actual effects are different than the state claims. This makes it impossible for the numbers to add up or for the analysis to satisfy the requirements of Section 1332. The waiver also fails other procedural tests. HHS and Treasury (the Departments) cannot legally approve this waiver, and if they attempted to do so, a plaintiff suing to overturn it would be expected to prevail. Other states considering following Georgia’s lead should be aware of these flaws before starting down this legally fraught path.
For a PDF version of the report, click here.
What Georgia’s waiver does
Georgia’s 1332 proposal contains two “phases.” The first phase, taking effect for plan year 2021, would implement a reinsurance program. The second phase would begin in 2022 and would overhaul the state’s individual health insurance market, including the Health Insurance Marketplace, as well as the financial assistance available to consumers buying individual market coverage. This analysis focuses primarily on the second phase of the waiver; the reinsurance program of the first phase is similar to other states’ 1332 waivers and poses no serious concerns, while the second phase proposes a novel set of policies that is not only unlawful, but also would undermine the ACA.
In the problematic second phase of the waiver, Georgia would implement four related policies:
- First, the waiver would eliminate the Health Insurance Marketplace and instead require consumers to enroll in coverage through private websites or call centers, operated by insurance companies or private vendors called web-brokers. The state would offer an eligibility tool that these private entities would use to verify consumer eligibility, but no public-facing, government-operated enrollment mechanism would be available. Vendors would be permitted to promote non-standard health products (like short-term plans and fixed indemnity coverage) alongside traditional health plans.
- Second, the state would allow the sale of two new types of plan that do not comply with all ACA requirements, which the state refers to as “non-QHPs” (or non-Qualified Health Plans). It would allow consumers to purchase “copper plans” with an actuarial value of 50%, which would have higher cost-sharing than would otherwise be allowed under the ACA. A simple plan design at this actuarial value would require a deductible and out-of-pocket limit equal to $13,500 per person – more than 60% greater than the ACA allows. At the same time, the waiver would eliminate the requirement that QHP issuers offer silver and gold plans.
- The state also contemplates allowing “disease management plans” that would be designed to meet the needs of people with chronic conditions. Of course, carriers can offer such plans under current law. The state indicates that under the waiver, disease management plans will have “flexibility” in meeting the essential health benefits requirements of the ACA. The state provides little detail on what this flexibility might entail, other than noting that the plans will still be required to cover all ten essential health benefits. The state suggests it might seek approval to offer other types of non-QHP plans in the future.
- Third, the waiver would eliminate the ACA’s premium tax credit (PTC) and cost-sharing reductions (CSR) for people who buy coverage in the individual market and replace them with a state-operated financial assistance program. The state would receive the federal money that would have been spent on these programs (net of any federal cost of the waiver) as a “pass-through” payment and supplement it with a small amount of additional state funding. A consumer eligible for assistance could use their state premium subsidy to buy either a traditional bronze, silver, gold, or platinum plan, or a non-QHP copper plan or disease management plan.
- The state financial assistance would be calculated the same way that financial assistance is calculated under current law. That is, PTC would be based on the second-lowest cost silver plan available to the consumer, and the consumer’s contribution would be set on a sliding scale based on income. The waiver proposal indicates that the state might change its methodology for calculating financial assistance in the future but would seek additional federal approval to do so.
- Finally, the state proposes a cap on the financial assistance provided. The state would make only a limited and specified amount of financial assistance available; once it is on track to spend that full amount, consumers would be placed on a waitlist. Consumers on the waitlist would not be able to enroll in coverage with financial assistance and would be forced to either pay the full premium or remain uninsured. (For context, in Georgia in 2019, 88% of Marketplace enrollees received financial assistance, and advance PTC payments covered an average of 88% of the premium among those receiving them, so the overwhelming majority of affected consumers would be expected to remain uninsured.)
The legal vulnerabilities of the waiver
Georgia’s proposed 1332 waiver is plagued by legal problems large and small. Even before considering the particularities of Georgia’s submission, a waiver structured in the way the state proposes will fail to satisfy the statutory requirement that it must provide coverage that is as affordable as coverage under the ACA. Further, the specific proposal submitted by the state is likely to violate the requirement that a waiver provide coverage to as many people as would have it absent the waiver, and suffers from additional weakness.
- By facilitating the purchase of copper plans with much higher cost-sharing than ACA-compliant plans, the waiver violates the statutory instruction to “provide coverage and cost sharing protections against excessive out-of-pocket spending that are at least as affordable” as the ACA.
- The waiver also fails under the (unlawful) interpretation of 1332 advanced by the Departments that a waiver should be judged based on the coverage that is “made available” to people. This is because, among other things, the waiver’s budget cap means that affordable coverage would in fact not be made available to otherwise-eligible consumers.
- Georgia’s analysis of its own policy proposal includes several objective errors that make its funding insufficient – even under its own modeling assumptions – by about $200 million per year.
- The analysis also makes a variety of unrealistic assumptions about consumer behavior, failing to consider a wide variety of reasons that premiums would rise or coverage drop under the waiver. More realistic assumptions suggest that the waiver would fail to satisfy the coverage guardrail or the state’s budget cap would be hit, or both.
- The submission fails to meet basic procedural requirements for 1332 waivers, including the requirement to provide an adequate actuarial and economic analysis and a meaningful state comment period pre-submission.
- The conceptual core of the waiver – allowing plans to impose much higher cost-sharing than otherwise permitted under the ACA – arguably requires waiving provisions of law that the Departments are not authorized to waive under Section 1332.
Each of these issues is considered in turn in the analysis that follows.
Georgia’s waiver violates the statutory affordability guardrail
Under Georgia’s waiver, consumers would newly be permitted to use financial assistance to buy copper plans with much higher cost-sharing than is allowed under the ACA. These copper plans would have an actuarial value of 50% (as opposed to the 60% AV required under the ACA), and out-of-pocket maximums of $13,500 per person or more (as opposed to the $8,150 cap under the ACA). This change will clearly cause the waiver to run afoul of Section 1332’s affordability guardrail.
We can begin with the statutory language:
The Secretary [of HHS or Treasury] may grant a request for a waiver… only if the Secretary determines that the State plan will provide coverage and cost sharing protections against excessive out-of-pocket spending that are at least as affordable as the provisions of [Title I of the ACA] would provide.
The waiver fails this test. A central feature of the waiver would provide coverage with cost-sharing that is less affordable than the plans allowed under the ACA. Enrollment in less affordable plans is not a second- or third-order impact of the state’s waiver plan that requires sophisticated analysis and modeling to understand; it is the intent of Georgia’s plan. The Departments cannot lawfully conclude a waiver intended to provide coverage less affordable than the ACA will provide coverage “at least as affordable” as the ACA.
The state may assert that because these low AV copper plans have premiums that are lower than premiums for other plans they still should be considered affordable under the statute, but the language of Section 1332 clearly forecloses this claim. The statute defines affordability not simply as a general balancing of the costs of coverage, but to require a determination of whether the waiver’s coverage will provide “protections against excessive out-of-pocket spending” that are equivalent to the ACA. But it is exactly that protection that the state’s waiver seeks to erode by allowing consumers to enroll in plans that expose them to out-of-pocket costs at least $5,000 greater than otherwise permitted.
The Departments have previously considered whether a state could use a 1332 waiver to offer copper plans, and concluded that the plain language of the statute precludes that policy, when they issued (now-revoked) 2015 guidance on the 1332 guardrails. As the Departments said then, in a straightforward interpretation of the statute, “Waivers that reduce the number of people with insurance coverage that provides both an actuarial value equal to or greater than 60 percent and an out-of-pocket maximum that complies with Section 1302(c)(1) of the ACA, would fail” the affordability requirement. A waiver that makes copper plans available with financial assistance must, under any plausible set of assumptions, decrease the number of people with coverage that meets this standard. Some people who would otherwise buy higher AV plans will elect to buy copper plans, and there is no plausible pathway by which the waiver would induce an equal or greater number of people who would otherwise be uninsured to buy ACA-compliant plans. Indeed, the state’s own analysis indicates that 33% of existing subsidized consumers in bronze plans would switch to copper plans.
Georgia’s waiver would also fail the affordability guardrail when considering the coverage “made available”
The state may hope to satisfy the affordability guardrail by relying on unlawful 1332 guidance released in 2018. In that guidance, the Departments purport to interpret the affordability guardrail to require that the affordability analysis will “focus on the nature of coverage that is made available to state residents (access to coverage), rather than on the coverage that residents actually purchase.” That is, the state may claim that because the waiver would allow consumers to buy bronze or other higher AV coverage if they choose, it satisfies this “made available” test because consumers have “access” to more affordable plans. The 2018 guidance is unlawful on both procedural and substantive grounds, but even under this weaker standard the state’s waiver would necessarily fail the affordability guardrail, for two distinct reasons.
Gold plans are unlikely to remain available
First, Georgia’s proposal waives the requirement for insurers to offer plans in the gold and silver tiers. The state’s analysis seems to assume that issuers would continue to offer higher AV plans roughly as they do today despite this change. But the evidence suggests they would not. Under the ACA, issuers in the Marketplace are permitted to offer bronze, silver, gold, and platinum plans, and they are required to offer silver and gold plans. Notably, today platinum plans have disappeared for many consumers – in 2020 80% of Georgia counties (and 95% of all counties served by HealthCare.gov) have no platinum plans available to consumers. Further, issuers have periodically sought relief from the requirement to offer gold plans, making clear that many issuers do not want to offer high AV plans and will only do so if required. Georgia’s proposal would waive the requirement that issuers offer silver and gold plans, and nothing else in the waiver offers any inducement for issuers to offer gold plans. Under these circumstances, gold plans would be expected to largely disappear from the market, just as platinum plans have under current law, so consumers would no longer have the higher cost-sharing protection of gold plans “made available” to them. Indeed, the only possible reason to approve a waiver of the requirement to offer silver and gold plans is to make them less available, so this cannot satisfy the guardrail.
The existence of a cap means affordable coverage is not available to all eligible residents
The second problem is more fundamental to the structure of the waiver. Recall that Georgia’s waiver converts the ACA’s entitlement to financial assistance for all eligible consumers into a capped state-run program under which the state will only provide financial assistance up to a budgetary cap. According to the state, the cap (which includes all federal pass-through funds and a small amount of additional state funding) is sufficient to provide financial assistance to about 359,000 consumers – 25,570 more than the 333,500 who are projected to receive it under current law – assuming benchmark premiums decrease by 1.9%. After hitting the cap, the state will place consumers on a waitlist – with no additional financial assistance provided. The state says it does not expect to hit this cap, because it expects fewer additional subsidized people to enroll than could fit below the cap.
But the very existence of this cap violates the (unlawful) “made available” test articulated by the federal government in its 2018 guidance. Under current law, affordable coverage is “made available” to all eligible consumers, including the roughly 330,000 people actually enrolled in coverage with financial assistance and the over 700,000 residents who the application says are eligible for financial assistance but have not enrolled. But under a capped plan, affordable coverage is only “made available” to the number who can fit below the cap, which is roughly equivalent to the much smaller number of people who are actually enrolled today. That is, under the state’s analysis, current-law financial assistance makes affordable coverage available to over 1 million people, while Georgia’s waiver makes affordable coverage available to only 359,000.
And, indeed, the 2018 guidance is explicit that the guardrail analysis under these tests should compare the “number” of people to whom affordable coverage is made available under the waiver and under current law:
The Departments will consider the affordability requirement to be met in a state plan that will provide consumers access to coverage options that are at least as affordable and comprehensive as the coverage options provided without the waiver, to at least a comparable number of people as would have had access to such coverage absent the waiver.
A capped waiver, then, cannot be said to provide access to affordable coverage to the same number of people who have access under current law – even if the state claims not to expect the cap to be triggered. That said, the following sections demonstrate that there is a substantial likelihood that the state will, in fact, hit the cap, meaning that there will be hundreds of thousands of people for whom affordable coverage is not made available.
The actuarial analysis includes clear objective errors that hide a deficit of around $200 million per year
The waiver as currently designed does not include sufficient funds to be implemented as claimed or to satisfy the guardrails. The state simply has less money committed than it projects relative to its costs, which makes it impossible to satisfy the coverage guardrail under the cap. Specifically, there are three reasons that federal pass-through funding would be lower and state costs higher than the state projects.
The state fails to account for about $150 million per year in federal employer mandate penalty lost due to waiving the premium tax credit.
The state entirely neglects the impact to the federal government of forgone employer mandate payments; those forgone amounts affect deficit neutrality and therefore would be subtracted from the pass-through amounts. Under the employer mandate rules, employers may owe a penalty only when their employees receive the premium tax credit (PTC). Because Georgia consumers will receive state assistance in place of the federal PTC, employers will generally not owe penalties with respect to employees who are Georgia residents. This will reduce federal revenue by about $150 million in the first year of the waiver alone – about as much as Georgia’s entire proposed contribution to the new subsidies.
The state wrongly assumes that HealthCare.gov can fully defray $100 million per year in lost user fee revenue
In its calculations of deficit neutrality and pass-through funding, the state assumes the federal government will achieve program savings equal to 100% of the user fee revenue that it will no longer receive once Georgia withdraws from HealthCare.gov: over $100 million per year. Substantial program savings are likely given reduced marginal costs for customer service and processing of enrollment-related documents. But operating HealthCare.gov also comes with substantial fixed costs to maintain the website, support extensive IT systems, and develop necessary new functionality. The relative shares of fixed versus marginal costs are not known, but even a relatively small fixed share could reduce pass-through funding by additional tens of millions of dollars per year.
The state ignores the costs of eliminating subsidy reconciliation, which are likely tens of millions of dollars per year
The structure of the state subsidy would likely increase the cost of covering each enrollee and therefore make it harder to stay under the cap. The application does not address whether state financial assistance will be “reconciled” at the end of the year, as the federal PTC is, but the complexity of standing up a reconciled credit and the lack of funding designated to do so suggest that Georgia’s subsidy probably would not be reconciled. If instead Georgia’s subsidies are based solely on income projected with the help of available data, the state is likely to spend more per enrollee than the federal government would have spent on PTC due to the failure to recoup excess advance payments. For example, IRS data show that for tax year 2017, Georgia taxpayers received about $58 million in additional PTC through reconciliation, while those who received excess APTC paid back about $109 million – a net recovery through reconciliation of about $50 million.
The application suggests Georgia will attempt to use more up-to-date information to improve the accuracy of upfront eligibility determinations. But even with substantial improvement, eliminating reconciliation could still cost in the tens of millions of dollars per year.
Correcting these modeling errors would reveal a $200 million annual shortfall
Taken together, these errors indicate a shortfall of about $200 million in plan year 2022 alone. Given the deficit neutrality constraint, that means a $200 million reduction in subsidies in that year – a cut of more than 8% compared to what Georgia promises. By comparison, Georgia estimates that its waiver has a $39 million funding buffer beyond its cap. In other words, not only will Georgia hit the cap, it will fall $160 million short of the amount needed to meet its own coverage projections.
Georgia’s waiver analysis makes unexplained and implausible assumptions
The objective flaws described above could potentially be addressed through an infusion of additional state dollars (though adding $200 million per year would nearly double Georgia’s total contribution to the waiver – something there’s no indication Georgia will do). But the analysis also includes a number of implausible or baseless assumptions that undermine the analysis even if this shortfall were addressed.
Georgia asserts that its plan will increase coverage in 2022 by 35,000 people – with 23,000 receiving subsidies and 12,000 unsubsidized – and also reduce premiums by 1.9%. The precision of these figures is crucial to satisfying the coverage guardrail and avoiding the budget cap, given the tension inherent in the waiver. But a closer look at the application indicates there these figures rely on a number of unsupported and sometimes implausible assumptions.
Georgia’s coverage assumptions are not supported
Georgia claims that 25,000 of its 35,000 additional enrollees will be currently uninsured people who enroll due to “increased accessed [sic] and web-broker marketing.” The other 10,000 would enroll due to the “availability of lower premium Copper Plans.” The application concedes that these figures are not the product of any modeling – they are assumptions made by the state. For example, the “support” given for the 35,000 figure is that “Roughly 35,000 people left the individual market between 2017 – 2019. Modeled the impact of 35,000 re-joining.”
Especially unsupported is the assumption of enrollment growth of 25,000 resulting from increased access and web-broker marketing. Web-brokers and direct enrollment are already major channels for enrollment in ACA plans; they are not a creation of this waiver. Indeed, the Georgia plan does not provide for any new enrollment avenues that are not allowed under current law; it merely eliminates the most established and popular one – HealthCare.gov. There is no reason to believe that eliminating an existing channel without creating anything new will lead to increased enrollment. Similarly, there is no reason to believe that increased marketing would occur or that it would lead to enrollment gains of this magnitude, because under current law, brokers have an incentive to attract business.
On the contrary, there are reasons to fear lower levels of enrollment when private entities assume full responsibility. Removing the option to enroll via HealthCare.gov and requiring consumers to find a new enrollment channel, along with potential administrative difficulties in enrolling in coverage in entirely new system, would likely depress enrollment. Some states that have shifted from a state Marketplace to a federal one or vice versa have seen lower levels of enrollment in the year of the transition – and the enrollment channel disruption Georgia contemplates here is far greater than the disruption experienced in those cases. Consumers shifting to non-single risk pool coverage (e.g. short-term plans sold by web-brokers) or consumers dropping coverage entirely if gold plans are not available could exacerbate these impacts.
There’s also reason to fear a less-than-smooth transition to the state’s back-end enrollment tool. The state provides essentially no information on how the administrative transition would work in practice. The state allocates a small amount of funding for their own administrative functions, but offers no budgeting to suggest that these resources will be sufficient to manage the transformation. And indeed, while some states have been able to stand up State-based Marketplaces (SBMs) fairly quickly, that work is based on off-the-shelf technology developed over many years, while this waiver would require an entirely new system for calculating eligibility and presenting different types of plans.
Given that the reinsurance program is estimated to increase enrollment by only about 1,500 people, these challenges could easily lead to a net reduction in coverage under the waiver. Indeed, a reduction in coverage seems more plausible than the state’s claim that 25,000 additional people will enroll via systems that are available under current law.
Georgia’s assumptions about unsubsidized enrollment are also unfounded
Georgia’s claim that the waiver can increase coverage within its spending cap is predicated on a large increase in consumers enrolling without subsidies. Specifically, of the 10,000 people projected to newly enroll due to the availability of lower-cost copper plans, Georgia assumes 92% would be unsubsidized and only 8% would receive subsidies. This is an extraordinary claim given that 88% of current Marketplace enrollees receive subsidies and over half of Georgia’s uninsured are subsidy eligible. But the application provides no reasonable explanation for this claim. It notes that “[b]ecause this population [of 10,000 new enrollees] is likely more price sensitive, the analysis assumed the increased enrollment reflects the makeup of the population who left between 2017 – 2019, of which 92% were unsubsidized.” But there is no reason to think the two groups will actually be similar.
Georgia’s claims about premiums changes are problematic
Georgia projects that the individual market changes in the waiver will reduce premiums by 1.9%. While the state concedes that copper plans will create adverse selection that will increase benchmark premiums by a small amount, it projects that other factors will eclipse this small increase and lower premiums overall. There are several omissions in their consideration of these issues.
First, the premium calculation is closely tied to the assumed enrollment figures. Georgia notes that the premium change “is primarily driven by an improvement in the overall market risk score due to new members entering the market whose health status (risk scores) are better than the current.” Given that the figures for new members entering the market are implausible for the reasons discussed above, there is reason to doubt these premium impacts will occur.
Second, there is reason to believe that the enrollment system will contribute to adverse selection in ways Georgia does not seem to contemplate. To the extent the waiver anticipates allowing web-brokers to do anything differently than they do today, it is to allow web-brokers to sell plans that are not a part of the single risk pool (like short-term limited duration and fixed indemnity plans) alongside other plans. But this would be expected to cause adverse selection against the single risk pool – an impact that the state does not consider but that would be expected to increase benchmark plan premiums. In addition, if more consumers shift from higher AV plans down to copper plans than the state projects, benchmark plan premiums will also increase.
Finally, if, as discussed above, administrative challenges make enrollment more difficult, it is healthy people who are most likely to give up in frustration. Similarly, eliminating the government-operated Health Insurance Marketplace, and relying on private vendors and insurance companies to operate the enrollment functions could also cause adverse selection and higher premiums as healthier individuals do not bother to seek out a new channel.
More accurate assumptions about pass-through, enrollment, and premiums could result in the waiver failing under the coverage guardrail
As the discussion above shows, there are a number of different ways the state’s waiver could fail under the coverage guardrail once the analytic deficits noted above are corrected. That is, it is likely the state will fail to satisfy the statutory requirement that it “provide coverage to at least a comparable number of its residents” as the ACA.
First, the analysis above notes that pass-through must be reduced by about $200 million (roughly 8%) compared to the state’s current assumptions. If that funding is not replaced with state dollars, then the state will have far less funding under its budget cap – indeed, their own analysis shows that a $200 million reduction would leave them with less funding than they need to provide financial assistance to everyone who receives it under current law. If the state were to hit its budget cap when providing financial assistance to fewer people than current law, then those denied assistance would almost certainly remain uninsured – meaning the state’s waiver would cover fewer people. (The waiver would also fail the affordability guardrail, as discussed above).
Second, higher premiums can generate a coverage guardrail violation, even if the $200 million in missing pass-through funding was replaced. Higher premiums would mean higher subsidies per covered person, which could result in the state hitting its budget cap while enrolling fewer people than projected. Under their projections, the state’s budget cap allows them to provide financial assistance to 25,570 more people than current law. But if benchmark premiums were 6% higher than they would be without the waiver, that would likely cause the state to hit the cap while enrolling fewer people than would be enrolled under current law. Accounting for the analytic gaps noted above could suggest premium changes of that magnitude – similarly violating the coverage (and affordability) guardrails.
Finally, correcting for the implausible assumptions about enrollment and giving proper consideration to the factors that could depress enrollment, it could also be the case that the state simply enrolls fewer people than current law – even without hitting its budget cap – since fewer consumers successfully navigate the enrollment process.
Clear guardrail violations
To summarize, a fair reading of either the text of the statute, or even of the administration’s 2018 guidance that unlawfully attempts to loosen the statutory guardrails, makes clear that Georgia’s waiver is unapprovable. Without needing to consider any detailed analysis of the state plan, the waiver violates the affordability guardrail by driving consumers toward less affordable copper plans, by repealing the requirement that issuers offer gold plans, and by using a budget cap that prevents the state from claiming that affordable plans are available to all eligible residents. Further, the waiver analysis includes clear objective errors – like ignoring lost employer mandate revenue and overstating HealthCare.gov savings – that make it impossible for the numbers to add up as claimed. Finally, even a cursory examination of the factors likely to influence consumer behavior and premiums make clear that the assumptions employed in the analysis are too optimistic, potentially resulting in lower enrollment and higher premiums than projected, triggering the cap and violating the affordability, and potentially the coverage, guardrail.
In the next section we turn to procedural problems that underscore that approval of this waiver would be unlawful.
Shortcoming of the economic and actuarial analysis
As discussed above, the Georgia application’s economic and actuarial analysis is characterized by numerous serious analytic gaps that render it insufficient to satisfy the regulatory requirements for a complete application. To briefly restate the shortcomings:
- The analysis fails to consider the impact of repealing the requirement that issuers offer silver and gold plans, which would be expected to lead to erosion in the availability of gold plans – an obvious impact that the state ignores.
- The waiver modeling asserts with no evidence that 25,000 additional consumers will enroll because of increased access to coverage and marketing by web-brokers. Indeed, this naked assumption is the core of all claims that the state makes about its waiver. As noted above, this assumption seems implausible. The waiver creates no new methods for accessing coverage. Web-brokers already exist, and Georgia’s waiver gives them no new tools to seek out consumers nor additional incentive to conduct outreach beyond what they do today. To the contrary, it is far more likely that the state will lose some existing consumers by eliminating HealthCare.gov as an option and by undertaking an enrollment platform transition – which has been the experience of states that have transitioned enrollment technology in recent years.
- The waiver makes similar unsupported and implausible assumptions about the share of new enrollees who would be unsubsidized. Taken together, these shortcomings eliminate any credibility in the application’s projections of premiums changes, which are driven by enrollment changes.
- The waiver’s analysis entirely ignores the fact that consumers may elect plans outside of the single risk pool (like short-term limited duration plans), despite the fact that the only new opportunity the waiver provides to web-brokers is the ability to market those plans alongside other plans.
- The state’s funding assumptions neglect the impact of reduced federal employer mandate revenue – a factor that, when properly included, would reduce the state’s pass-through funding by approximately $150 million in plan year 2022 alone.
- The actuarial analysis assumes that HealthCare.gov can entirely defray about $100 million per year in lost user fee revenue through program savings. This is almost surely wrong, as HealthCare.gov has substantial fixed costs. As a result, the state is likely to receive tens of millions of dollars less per year in pass-through funding than it projects.
- It does not appear that state financial assistance will undergo “reconciliation,” but the analysis does not consider how much this will increase the cost of state financial assistance. Available figures suggest it could be tens of millions of dollars per year.
For all of these reasons, the state’s analysis does not provide a reasonably accurate picture of the impact of the waiver and therefore fails to satisfy the requirements for actuarial and economic analysis supporting the waiver application.
No meaningful state or federal comment period
The statute and its implementing regulations require “a process for public notice and comment at the State level, including public hearings” and “a process for providing public notice and comment after the application is received by the Secretary,” both of which must be “sufficient to ensure a meaningful level of public input.” A meaningful state comment period has not occurred, and the document submitted to the federal government does not allow for a meaningful federal comment period, either.
The state’s process to date does not constitute a proper state comment period. The state conducted public comment on an entirely different waiver proposal. The initial draft proposed to offer Georgia consumers plans that did not cover all of the ACA’s essential health benefits (EHB) – but those plans would have had cost-sharing, actuarial values, and an out-of-pocket maximum unchanged from current law. Commenters objected to the exclusion of EHB from Georgia plans, and the state reacted by eliminating its proposal to allow plans to exclude EHB. But the state has never sought comment on a proposal to allow plans to increase their deductibles and out-of-pocket maximums by $5,000. These are different proposals, and comments on the former cannot be considered meaningful with respect to the latter.
The document submitted to the federal government also contains major flaws that prevent it from serving as the basis for meaningful federal comments. Given the weaknesses of the analysis noted above, commenters lack accurate information about how this waiver would impact premiums, coverage, and affordability. In addition, fundamental information about how the waiver would actually operate is not available. The state fails to provide information about how it would operate its financial assistance program. It provides very little details – and no budget – for how it would manage an enormous takeover of technical responsibilities related to enrollment, disenrollment, and financial assistance. The budget cap is a critical feature of the waiver and yet the state’s discussion of the cap is limited to noting its existence and suggesting the state has a “plan” – without describing how that plan would actually work. For example, the proposal offers no information on how the state would determine when a cap is needed and how it would conduct the enrollment waitlist. This kind of information is necessary for a comment period to be considered “meaningful.”
The waiver also provides essentially no detail on the operation of “disease management plans.” The state seeks a waiver of Section 1302(b) – the requirement to cover essential health benefits – for the disease management plans, but then notes that the plans will still be required to cover all ten EHB and will simply have “flexibility” in meeting the EHB requirements. How that flexibility differs from current policy is unspecified, other than an oblique reference in the actuarial analysis suggesting that there will be “services the insurer cuts back on” in disease management plans. The state’s initial 1332 proposal to waive EHB in its market has been abandoned (likely because it would have caused premiums to increase by more than 50%, resulting in 110,000 people losing coverage). The state’s discussion of disease management plans seems to imply that this waiver of EHB will be more modest than the original proposal – but the state has provided no information that would allow commenters to understand that assumption.
Waiver is not complete
For the same reasons that the waiver document cannot serve as the basis for meaningful public comment, it also cannot be considered complete. The gaps in the analysis and the absence of discussion of major operational considerations mean the waiver cannot be declared complete by the federal government.
More specifically, the regulations require a specifically enumerated list of data elements that the state has not provided. The state is directed to provide “information on the age, income, health expenses and current health insurance status of the relevant State population; the number of employers by number of employees and whether the employer offers insurance; cross-tabulations of these variables; and an explanation of data sources and quality.” No information on age, income, or health expenses appears in the waiver, cross-tabulations are not available, and data sources and quality are discussed only in passing. Indeed, the state’s failure to provide these somewhat ministerial elements is perhaps reflective of the fact that the state may not, in fact, have conducted the sort of analysis that the statute and regulations contemplate, relying instead on general assumptions rather than a more careful analysis of the waiver’s impacts.
No authorizing state legislation
Section 1332 requires that a state seeking a waiver enact (and include in the application) a state law “that provides for State actions under a waiver under this section, including the implementation of the State plan.” Georgia has seemingly passed no such law. It has passed a general statute – the Patients First Act – that authorizes the governor to apply for a waiver and to implement waivers “in a manner consistent with state and federal law.” The problem is that without further legislation, implementing the waiver is not consistent with state law. For example, the state has not enacted legislation to authorize creation of the state eligibility tool. Without such authorization, implementing the waiver is not seemingly “consistent with state…law.”
The state may argue that the Patients First Act provides blanket authority for the governor to implement anything in a waiver. This would be an extraordinary grant of authority, and it is difficult to imagine that the state legislature intended it. But this argument is also undermined by the rest of the authorizing statute. In particular, another section of the Patients First Act authorizes the state to seek a section 1115 waiver to increase Medicaid eligibility up to 100% of the federal poverty line. The legislation then makes abundantly clear that the governor has authority to implement that change without seeking additional authority: “[U]pon approval of the waiver, the [Georgia health department] shall be authorized to take all necessary steps to implement the terms and conditions of the waiver without any further legislative action” (emphasis added). In short, Georgia’s legislature knows how to be clear that a waiver may be implemented without additional legislation. The fact that the 1332 section of the bill does not include such clear language indicates a different legislative intent in this case. As a result, the state has not satisfied the requirement for legislation to implement the state plan.
And, indeed, in waivers like the one contemplated here – where implementation of the waiver would result in residents being denied benefits to which they are entitled under federal law – state legislation authorizing the deprivation of benefits plays an especially important role in ensuring accountability for the waiver policy.
Copper plans may require waiver of provisions that 1332 cannot reach
Finally, under Section 1332, the Departments are provided authority to waive only certain enumerated provisions of the Affordable Care Act. Implementing Georgia’s proposal would arguably require the Departments to waive provisions of law that cannot be waived under Section 1332. This is because the ACA imposes its requirements related to health plan cost-sharing in two separate provisions of the ACA, one of which can be waived under 1332 and one of which cannot.
Requirements related to health plan cost-sharing appear in ACA Section 1302, which can be waived by a 1332 waiver, and Public Health Service Act Section 2707 (codifying ACA Section 1201), which cannot be waived. ACA Section 1302(a) is definitional – it defines a constellation of health plan protections as the “essential health benefits package.” One component of this package is the requirement that a plan “limits cost-sharing for such coverage in accordance with” limits on total out-of-pocket spending as described in 1302(c). Section 1302(c) caps total out-of-pocket spending at $8,150 for plan year 2020. Public Health Service Act Section 2707(a) requires that all health plans in the individual and small group market “shall ensure that such coverage includes the essential health benefits package required under section 1302(a).” Taken together, this means that Section 1302(a) defines a package of reforms to include a cap on total cost-sharing, while Section 2707(a) actually imposes that requirement on health plans.
The copper plans in Georgia’s waiver would exceed the limits of section 1302 significantly – they will have an out-of-pocket maximum that exposes consumers to at least $5,000 more in spending than otherwise permitted. That is why Georgia has asked that the Departments waive the cost-sharing cap of Section 1302(c). But Georgia cannot seek a waiver of Section 2707(a), which requires health plans to offer a “package” that complies with the Section 1302(c) limit.
Therefore, even if Georgia receives a waiver of Section 1302(c), the requirement of Section 2707(a) is arguably still binding on all individual market health plans, including the new copper plans. In that case, it would be unlawful for copper plans to be sold because they exceed the $8,150 limit.
This same logic applies to the requirements of Section 1302(d), which requires plans to have an actuarial value of 60, 70, 80, or 90%. Section 1302(a) defines the “essential health benefits package” to also include complying with the actuarial value requirement of 1302(d), and that requirement is therefore imposed on all plans via Section 2707(a). But copper plans would have an actuarial value of 50%, and would therefore violate this requirement of Section 2707 as well.
The Departments may argue that their authority to waive Section 1302 effectively modifies how Section 2707 operates in the state. In assessing the plausibility of this claim, one should note that the statutory text expressly denies the Departments the authority to waive Section 2707, and there are generally reasons to be skeptical of the agency’s attempt to expand the scope of their own waiver authority.
An approval of the waiver is not likely to hold up in federal court
For all of the reasons enumerated above, the Departments lack authority to approve Georgia’s waiver proposal. It violates the statutory guardrails, the requirement to provide an accurate and adequate actuarial analysis, several other procedural requirements, and even the Administration’s own weakened interpretation of the guardrails. Approval would exceed the Departments’ authority under Section 1332. If the Departments were to approve the waiver, a lawsuit against the federal government and Georgia would be expected to prevail.
Establishing standing to challenge the Georgia waiver is fairly straightforward. If approved, the waiver will result in some Georgia consumers enrolling in plans with higher deductibles than otherwise allowed under federal law. That is not a speculative claim; rather, it is a core function of the waiver. Georgia health care providers will therefore face higher uncompensated care burdens associated with uncollected revenue below the deductible. Some of today’s uncompensated care is associated with consumers failing to pay amounts under current-law deductibles ; increasing those deductibles by another $5,000 will drive uncompensated care burdens higher.
Consumers also face concrete and particularized injury associated with the budget cap. For example, any Georgia enrollee receiving financial assistance under current law has reason to fear she will lose financial assistance under a cap. Further, Georgia residents who frequently churn in and out of employer-based health coverage are likely to need to enroll in coverage with financial assistance mid-year, at a time when the state is most likely to have hit their budget cap, generating an especially significant source of concern. And Georgia consumers face similar harm from the potential for gold plans to become unavailable.
Having established standing, challengers to the waiver could demonstrate that Departmental approval of the waiver is ultra vires (that is, it exceeds the authority committed to the agency) and arbitrary and capricious:
- Section 1332 grants the Department authority to approve a waiver only if it provides “protections against excessive out-of-pocket spending that are at least as affordable” as the ACA. Georgia’s waiver increases consumer exposure to out-of-pocket spending by $5,000 and facially fails to satisfy this requirement.
- The Administration’s unlawful attempt to interpret 1332 to require only an analysis of the number of consumers to whom affordable coverage is “made available” is no help to Georgia or the federal government. The waiver’s budget cap means current law makes affordable coverage available to at least 400,000 more Georgians than the waiver, so the waiver cannot be approved on those grounds.
- A conclusion that Georgia’s waiver satisfies the coverage guardrail (i.e. the requirement that a waiver provide coverage to as many people as would have it absent the waiver) “entirely fail[s] to consider an important aspect of the problem” and is therefore arbitrary and capricious under the test articulated by the Supreme Court in Motor Vehicle Manufacturers Association v. State Farm Mutual Automobile Insurance Company. The analysis includes multiple unfounded and implausible assumptions about behavior in response to the waiver. The claims in the existing administrative record regarding the coverage guardrail rely on the unsupported assertion that eliminating the HealthCare.gov enrollment channel without creating any new options for consumers will somehow generate 25,000 new enrollees, a paradigmatic example of an assertion “so implausible that it could not be ascribed to a difference in view.”
- Georgia’s waiver has not received a meaningful state or federal comment period as required under the statute and its implementing regulations and does not comply with the Departments’ own requirements for waiver completeness.
- The Departments do not have authority to waive Section 2707 of the Public Health Service Act requiring that individual market health plans comply with the ACA’s out-of-pocket limit and actuarial value requirements, so the sale of copper plans may be unlawful even if the waiver is approved.
In sum, if the Departments were to approve this waiver, the federal government and Georgia should be bracing for a grueling and expensive legal battle that challengers would be expected to win.
 The reinsurance phase of the waiver is expected to decrease premiums and slightly increase enrollment, consistent with other states’ experiences. The Departments generally evaluate waivers as a package, and one could imagine an analysis under which a reinsurance policy would generate “gains” against the 1332 guardrails that the state could “spend” in pursuing policies that would otherwise cause “losses” under the guardrails. However, in this case the reinsurance program is not expected to meaningfully ameliorate the legal obstacles to approval described below. Georgia projects that the reinsurance waiver will increase enrollment by about 1,500 people – not enough to overcome the potential coverage losses from the second phase. And the waiver’s affordability problems focus on cost-sharing, while the reinsurance program changes only premiums. One could also argue that, since the state proposes that its reinsurance program be both approved and implemented separately from the individual market changes, the two policies should be evaluated separately, but such a claim is not necessary to show that the waiver cannot be approved.
 Using CMS’s current actuarial value calculator, modified to allow inputs that are otherwise prohibited because they violate the ACA provisions the state seeks to waive, a plan with a deductible and out-of-pocket limit of $13,500 has an AV of 50%. Plans with lower deductibles can also achieve a 50% AV, but only by increasing the out-of-pocket limit; an $11,000 deductible coupled with 20% coinsurance requires a $16,000 out-of-pocket limit to achieve a 50% AV. That is, roughly $13,500 is the smallest possible out-of-pocket maximum that can achieve an AV of about 50%.
 For a discussion of the myriad ways in which the 2018 guidance violates the statute, see Joel McElvain, The Administration’s Recent Guidance on State Innovation Waivers under the Affordable Care Act Likely Violates the Act’s Statutory Guardrails, Yale J. Reg., December 11, 2018, https://www.yalejreg.com/nc/the-administrations-recent-guidance-on-state-innovation-waivers-under-the-affordable-care-act-likely-violates-the-acts-statutory-guardrails-by-joel-mcelvain. For a discussion of why the guidance is procedurally invalid; see Christen Linke Young, The Trump Administration Side-Stepped Rulemaking Processes on the ACA’s State Innovation Waivers, The Brookings Institution, November 28, 2018, https://www.brookings.edu/blog/usc-brookings-schaeffer-on-health-policy/2018/11/28/the-trump-administration-side-stepped-rulemaking-processes-on-the-acas-state-innovation-waivers-and-it-could-make-their-new-section-1332-guidance-invalid.
 See ACA § 1301(a)(1)(C)(ii).
 The absence of gold plans may also create an affordability problem under the statutory language, without considering the unlawful “made available” language. Some consumers may also see reduced choice among silver or even bronze plans, but because the state-law equivalent of CSRs are available only in silver plans, it seems likely that issuers would continue to offer some silver plans.
 See Actuarial Analysis Table 2.7.
 Specifically, the application twice notes that Georgia has 1.4 million uninsured residents, of whom more than 50% are eligible for financial assistance. See pages 1 and 18.
 Kaiser Family Foundation figures based on the 2017 American Community Survey put the current number of uninsured Georgia residents who are eligible for Marketplace financial assistance lower, at 435,000 (instead of over 700,000). Even using the lower figure, less than half of those eligible for financial assistance under current policy could receive it under the waiver. See Kaiser Family Foundation, Distribution of Eligibility for ACA Health Coverage Among those Remaining Uninsured as of 2017, https://www.kff.org/health-reform/state-indicator/distribution-of-eligibility-for-aca-coverage-among-the-remaining-uninsured/.
 This figure is derived by allocating the U.S. Treasury Department’s projection of employer mandate payments for fiscal year 2023 by each state’s share of 2019 APTC recipients. See Office of Management and Budget, Supplemental Materials: Receipts XLS, https://www.whitehouse.gov/omb/supplemental-materials/; and Centers for Medicare & Medicaid Services, 2019 Marketplace Open Enrollment Period Public Use Files, https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Marketplace-Products/2019_Open_Enrollment.
 The Trump Administration has confirmed that states seeking to waive the PTC through a Section 1332 waiver must account for lost employer mandate revenue. See [CMS Letter to Iowa Commissioner Doug Ommen, Oct. 19, 2017.] (“[B]ecause individuals in Iowa will not receive premium tax credits, employers of Iowa residents will not be liable for employer shared responsibility payments with respect to their Iowa employees under IRC section 4980H…To prevent the waiver from increasing the Federal deficit, the Departments would provide passthrough funding equal to the estimated amount of premium tax credits that would have been paid on behalf of Iowa residents had the State not received such a waiver, less the estimated reduction in…employer shared responsibility payments [and other adjustments].”)
 See Actuarial Analysis Table 5.26.
 See Actuarial Analysis Table 5.11.
 Indeed, the only information that the state provides is that would use the Georgia Gateway system, a piece of technology that has caused errors and backlogs in Medicaid processing. See Andy Miller, Enrollment Drops, Eligibility Snags, Cyber-Errors in GA. Benefit Programs Spark Concern, WABE, https://www.wabe.org/enrollment-drops-eligibility-snags-cyber-errors-in-ga-benefit-programs-spark-concern/.
 See Actuarial Analysis at 30.
 See Waiver at page 18-19.
 It is noteworthy that the proposal submitted to the federal government does not provide a single example of the deductibles and out-of-pocket maximums that the waiver would require – despite the fact that those dollar figures can be easily calculated – making it more difficult for consumers to understand what this waiver would do.
 See Actuarial Analysis at 32.
 45 CFR 155.1308(f)(4)(iii)(A).
 One of us (Levitis) wrote previously that these Section 1302 requirements may “formally” be waived under a Section 1332 waiver but that the comprehensiveness and affordability guardrails significantly constrain flexibility to loosen these rules. This issue has not been addressed in regulations or by the Courts. See Jason Levitis, Changes to State Innovation Waivers in the Senate Health Bill Undermine Coverage and Open the Door to Misuse of Federal Funds, The Brookings Institution, June 23, 2017, https://www.brookings.edu/blog/usc-brookings-schaeffer-on-health-policy/2017/06/23/changes-to-state-innovation-waivers-in-the-senate-health-bill-undermine-coverage-and-open-the-door-to-misuse-of-federal-funds/.
 Jeff Lagasse, Benefit Design, Higher Deductibles Will Increase Bad Debt for Hospitals, Healthcare Finance, November 22, 2019, https://www.healthcarefinancenews.com/news/benefit-design-higher-deductibles-will-increase-bad-debt-hospitals.