Analyzing the Impact of State Policy Discretion on Risk, Enrollment, and Plan Offerings in the ACA Individual Health Insurance Markets

Abstract

Health insurance is a key enabler of improving population health by reducing fiscal barriers to care, improving access and reducing stress.1–4 The Affordable Care Act (ACA), passed in 2010, has been successful in reducing the percentage of people who do not have health insurance to 7.7% in 2023 from a peak of 16% being uninsured in 2010 when the law was enacted.5,6 The law is implemented by combined federal and state actions. The two major sources of new health insurance coverage are states volunteering to expand Medicaid eligibility to adults with incomes under 138% Federal Poverty Level (FPL), and the subsidized and regulated individual health insurance marketplaces. States have substantial discretion on how both of these programs area implemented.7,8 These state decisions may shift who is insured, what type of coverage individuals purchase and choices insurers offer to enrollees. Medicaid and the ACA individual health insurance markets have overlapping eligibility criteria. Individuals with incomes between 100% to 138% FPL are eligible for financial assistance from the federal government to purchase ACA plans if the state has not elected to expand Medicaid. Once a state has elected to expand Medicaid eligibility to 138% FPL, these individuals are transferred from the ACA individual market to Medicaid. Prior research has found that states which expand Medicaid experience lower ACA premiums after Medicaid Expansion for the benchmark silver plan.9 However, there is no research on the effects of Medicaid Expansion on risk scores, enrollment or paid premiums within the entire individual health insurance market. States are also able to choose how to regulate the ACA individual health insurance markets by regulating plan offerings and premiums. Some states like California and Massachusetts actively manage premiums and plan offerings.10–12 Other states have been more passive in regulating their insurance markets.10 Pricing regulations change the prices that consumers face and differences in prices may plausibly change the number of people who are insured in the ACA marketplaces as well as what plans individuals may choose. Prior research on state discretion on implementing the ACA individual markets have focused on political institution design, coalition building and clinical impacts of coverage expansions.13–17 Enactment of the ACA and its coverage expansions had been highly politicized and partisanized during the first six years of the coverage expansions.18,19 However, after the failure of an attempt to repeal the law in 2017, more states engaged with the individual health insurance market place and embraced the programmatic flexibilities that states possess. Sixteen states have enacted reinsurance programs with the intent to reduce gross premiums.20,21 Conservative and liberal states are adapting state based marketplaces. 22,23 Sixteen states have expanded Medicaid since 2014 for a total of forty-one states with expanded eligibility for Medicaid.24 Several states, including Texas, have altered their pricing regulations in an attempt to make ACA individual market health insurance more affordable to enrollees.25–27 This dissertation addresses three evidence gaps that will inform state and federal policymakers as they consider policies reforming and restructuring the US health insurance system:Aim 1: To estimate the association between a state’s decision to expand Medicaid to 138% FPL and a state’s ACA individual market characteristics including risk score, enrollment and average premiums paid. We used public use files that provided the annual risk adjustment reports for the ACA regulated individual and small group health insurance markets.28 We estimated, using a staggered implementation difference in difference approach with two way fixed effects, the impact of a state’s decision to expand Medicaid eligibility to 138% FPL and the level of predicted risk, enrollment and average premium in the state’s individual health insurance marketplace. We controlled for state-year policy variables of operation of a state-based marketplace and the presence of a reinsurance waiver. We estimated that expanding Medicaid was led to a 6.1% decrease in state PLRS, as well as a decrease of 28% member months but no change in average state per member per month premium paid. We conducted a placebo test using the small group market and found null results which strengthens our causal interpretation of the linkage between Medicaid Expansion and the individual health insurance market outcomes. Our results suggest that the Medicaid expansion eligible population has higher average risk than the remaining ACA enrollees and that changes in prices that other researchers have found for the benchmark plan are not reflective of enrollee choices.9 These results will inform state policy makers who wish to expand Medicaid or integrate Medicaid with the elements of the ACA marketplace. Aim 2: To evaluate the causal impacts on enrollment, and plan choice of the 2023 Texas policy transition to “premium alignment” in the ACA individual health insurance markets We aimed to causally estimate the impact of a new premium pricing policy known as “premium alignment,” in the state of Texas’s individual health insurance market, on enrollment, affordability and plan choice relative to national trends by using synthetic difference-in-differences methods.29 Synthetic difference-in-differences estimation creates a weighed matched counterfactual to treatment counties based on the observed outcomes in the pre-policy implementation period.29 We constructed a balanced panel from 2018 to 2023 of counties that were always on Healthcare.gov in states that had not adapted similar premium alignment policies. Our primary data was from the Qualified Health Plan Landscape Public Use Files and Open Enrollment Period Public Use files published by the Centers for Medicare and Medicaid Services. 30–44 In a sensitivity analysis we controlled for state-year policy changes including Medicaid Expansion and Section 1332 reinsurance waiver usage, and a rich set of county level demographic characteristics from the American Community Survey and found similar results.21,24,45 Permutation testing supports causal interpretation of results for enrollees with incomes between 200% to 400% FPL. We estimated that enrollment increased by 14.9% for enrollees with incomes between 201-250% FPL, 18.23% for enrollees with incomes between 251-300% FPL and 13.48% for enrollees with incomes between 301-400% FPL relative to the counterfactual control of states that did not enact premium alignment policies. We observed that the least expensive plans for both subsidized and unsubsidized buyers decreased in price and became more affordable, potentially increasing the availability of zero premium plans which have been shown to increase enrollment and retention in other contexts.46,47 Enrollees with incomes between 100 to 200% FPL decreased by 9.25% the purchase of enriched benefit, Cost-Sharing Reduction Silver plans as new trade-offs between premium affordability and financial protection became available.48 Our results are informing state policy makers on options to improve affordability and state health insurance coverage without the potential negative impacts of reinsurance.20 Aim 3: To estimate the effects of premium alignment policy changes on benefit design and plan offerings in the Texas individual health insurance markets We sought to establish a causal estimate of changes in plan offerings, specifically the percentage of insurers offering at least one Bronze plan in a county-year, and benefit design for Gold plans after Texas introduced its premium alignment policy. The ACA relies on a public-private partnership with federal funds partially paying private insurers to offer heavily subsidized health insurance policies.49 We identified the least expensive Gold plan for each county-year and then identified its actuarial value, a group level measure of benefit richness, the deductible for a single individual and the maximum out of pocket limit for a single individual for each plan-county-year triad using the Qualified Health Plan Landscape Public Use Files and Plan Attributes Public Use Files. 39–44,50–56 We then constructed a balanced panel of all counties that had continual healthcare.gov usage from 2018 to 2024, excluded states with similar premium alignment policies. We used a matched county-border discontinuity design with county-pair, state and year fixed effects to estimate Using the synthetic difference in difference methods, we estimated that premium alignment led to a 34.7 percentage point decrease probability of all insurers in a Texas county offering at least one low premium Bronze plan and an increase of $272.26 for the deductible for the least expensive Gold plan. Changes in the actuarial value and maximum out of pocket limit for the least expensive Gold plan were not significant. These results suggest that insurers are crafting plans that are less attractive to relatively healthy individuals. State and federal policymakers can use these findings to weigh the trade-offs between enrollment levels and benefit richness.