HEALTH POLICY CENTER RE S E AR C H RE P O R T The Effects of Medicare Buy-In Policies for Older Adults on Health Insurance Coverage and Health Care Spending Bowen Garrett Jessica Banthin Anuj Gangopadhyaya Matthew Buettgens Adele Shartzer John Holahan Diane Arnos December 2020 (corrected February 2021) AB O U T T HE U R BA N I NS T I T U TE The nonprofit Urban Institute is a leading research organization dedicated to developing evidence-based insights that improve people’s lives and strengthen communities. For 50 years, Urban has been the trusted source for rigorous analysis of complex social and economic issues; strategic advice to policymakers, philanthropists, and practitioners; and new, promising ideas that expand opportunities for all. Our work inspires effective decisions that advance fairness and enhance the well-being of people and places. Copyright © December 2020. Urban Institute. Permission is granted for reproduction of this file, with attribution to the Urban Institute. Cover image by Tim Meko. Contents Acknowledgments iv Executive Summary v Errata xi The Effects of Medicare Buy-In Policies for Older Adults on Health Insurance Coverage and Health Care Spending 1 Methods 3 Definition of the Base Medicare Buy-In Policy 3 Provider Payment Rates and Premiums under the Medicare Buy-In Policy 4 Variations on the Base Medicare Buy-In Policy 4 Findings for the Base Buy-In Policy Scenario 7 Changes in Health Insurance Coverage 7 Changes in Premiums 13 Changes in Health Care Spending for Individuals 17 Changes in Aggregate Health Care Spending 21 Effects of Alternative Buy-In Scenarios on Coverage and Aggregate Spending 24 Restrict Buy-In Eligibility to Those Ages 55 to 64 28 Restrict Geographic Variation in Premium Prices for the Buy-In Plan 28 Coordinate Premium Tax Credits for Households with Members Above and Below Age 50 29 Remove the ESI Firewall 29 Modernize Medicare Benefits 30 Charge the Medicare Premium 30 Enhance Premium Subsidies 31 Discussion 32 Appendix A. Methodology 38 Modeling Insurance Coverage Choices 38 Provider Payment Rates and Premiums under the Medicare Buy-In Policy 39 Notes 41 References 42 About the Authors 43 Statement of Independence 44 Acknowledgments This report was funded by the National Institute for Health Care Reform. We are grateful to them and to all our funders, who make it possible for Urban to advance its mission. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Funders do not determine research findings or the insights and recommendations of Urban experts. Further information on the Urban Institute’s funding principles is available at urban.org/fundingprinciples. The authors are grateful for comments and suggestions from Stephen Zuckerman and for editing by Rachel Kenney. IV ACKNOWLEDGMENTS Executive Summary A Medicare buy-in program would allow qualifying individuals currently ineligible for Medicare to purchase a Medicare-like health insurance plan. Two recent legislative proposals that would offer a Medicare buy-in option for adults not yet eligible for Medi- care are S. 470, or the Medicare at 50 Act, sponsored by Senator Deborah Stabenow, and H.R. 1346, the Medicare Buy-in and Health Care Stabilization Act of 2019, sponsored by Representative Brian Higgins. Both bills would offer adults ages 50 to 64 a new option to buy into Medicare, meaning they can purchase traditional Medicare benefits (Parts A, B, and D) or a Medicare Advantage plan. President-elect Joe Biden, in contrast, has proposed both lowering the age of Medicare eligibility to 60 and creating a new Medicare-like health plan that could be purchased without age restrictions. In this report, using the Urban Institute's Health Insurance Policy Simulation Model, we estimate the coverage and health care spending implications of a Medicare buy-in policy similar to legislation proposed by Stabenow and Higgins that targets adults ages 50 to 64. We also examine seven policy variations: restricting eligibility to those ages 55 to 64; setting a national average buy-in premium; coordinating the calculation of premium tax credits so families with mixed coverage pay up to a maximum percentage of income for all coverage; allowing people with an offer of affordable employer- sponsored insurance to buy in; offering more extensive modernized Medicare benefits, including an out- of-pocket maximum and coverage for dental, vision, and hearing services; setting the full buy-in premium to a lower level than that for Medicare Parts A (typically no premium), B, and D; and providing enhanced premium subsidies in both the Marketplace and the buy-in plan. In the base buy-in policy scenario we model, adults ages 50 to 64, whom we call older adults, can purchase a plan like traditional fee-for-service Medicare that covers hospital care (Part A), physician and outpatient services (Part B), and prescription drugs (Part D). The plan would have an overall actuarial value of 85 percent, which is comparable with fee-for-service Medicare. The plan would reimburse providers at Medicare payment rates, which are typically lower than payment rates paid in the current nongroup market. Enrollees in this plan would form their own risk pool and the plan would be self-supporting. All buy-in enrollees within a rating area would pay the same premium (i.e., would not face age rating within the 50–64 age group). Premiums would be adjusted to reflect geographic variation in health care prices and would adopt the same subsidy structure (premium tax credit and cost-sharing reductions) used in the Affordable Care Act (ACA). EXECUTIVE SUMMARY V We estimate the base buy-in policy would have the following effects if fully implemented in 2020:  About 2.1 million people ages 50 to 64 would enroll in the Medicare buy-in plan. Most buy-in enrollees (1.9 million) would shift from current Marketplace or nongroup coverage. Very few adults in this age group (176,000) who are uninsured under current law would gain coverage under the base buy-in policy.  The buy-in plan would disproportionately attract people with higher incomes. Nearly half of the buy-in enrollees (994,000) would have income above 400 percent of the federal poverty level, and 710,000 would have incomes between 200 percent and 400 percent of the federal poverty level. Income-based tax credits for Marketplace and buy-in coverage shield people with lower incomes from premium differences, so the lower costs of the buy-in plans would be more attractive to higher-income people.  The buy-in plan would also disproportionately attract older individuals within the 50–64 age group. Nearly half of buy-in enrollees (931,000) are ages 60 to 64, 505,000 are ages 50 to 54, and 653,000 are ages 55 to 59. Whereas ACA Marketplace premiums rise with increasing age— even within this older age group—the community-rated premiums for the buy-in plan would not vary with age, making the plan relatively more attractive for adults ages 60 and over.  By attracting a substantial portion of older adults from the current nongroup market, the buy-in policy could have indirect effects on the population under age 50. With the buy-in policy, 16,000 fewer adults under 50 would have nongroup coverage, and 110,000 more people would be uninsured. Full premiums for nongroup coverage increase from $4,905 to $5,096, with only a slight rise in out-of-pocket premiums after subsidies. Premiums rise in the existing nongroup market because, as actuaries have found, premiums exceed claims costs for older populations.  Among buy-in enrollees who switched from Marketplace or other nongroup coverage, annual premiums (single coverage, before any tax credits) would decline from $8,843 to $8,371, even as the average actuarial value of coverage rises from 65 percent to 85 percent. Out-of-pocket premiums would increase slightly from $5,579 to $5,651.  Some people who switch to the buy-in policy would pay more and others would pay less in out- of-pocket premiums than they did with nongroup coverage. Reductions in premiums would be concentrated among older people and those not already receiving premium subsides (i.e., those with higher incomes). VI EXECUTIVE SUMMARY  Among adults ages 50 to 64 who stay in the nongroup market after the buy-in policy is introduced, average premiums before subsidies would increase from $9,752 to $10,142. Out- of-pocket premiums would increase slightly from $2,073 to $2,101.  Total health care spending among those who switch from nongroup plans to the buy-in plan would decline from $9,789 to $8,944 because of the buy-in plan’s lower provider payment rates. If payment rates were not reduced, total health care spending for the buy-in policy would be $10,028 (higher than the current-law $9,789 because of induced demand for services given the buy-in plan’s higher actuarial value).  Out-of-pocket health spending among those who switch to the buy-in plan would fall from $3,414 to $1,528 because of the plan’s higher actuarial value, and out-of-pocket premiums would fall from $6,430 to $6,230. Thus, buy-in enrollees would pay less out-of-pocket and consume slightly more health care but at a lower total cost than they would in the current nongroup market.  On net, the base buy-in policy would reduce aggregate health care spending by about $1.8 billion (about 0.1 percent). This reflects a net decrease in spending of $2.8 billion (0.3 percent) for older adults and an increase in spending of about $1.1 billion (less than 0.1 percent) for people under age 50.  The net decline in health care spending for adults ages 50 to 64 would be primarily driven by reduced household spending. Though out-of-pocket health care spending for households would decline by $3.6 billion (2.9 percent) because buy-in coverage is more generous, total premium spending would increase by $1.4 billion (1.2 percent) because of the 176,000 people who would gain coverage. Federal spending would change very little ($95 million, or less than 0.1 percent), because increased spending on buy-in tax credits would be almost fully offset by reduced spending on Marketplace tax credits.  The net increase in health care spending for those under age 50 would mainly reflect increased premium spending by households of $656 million (0.4 percent) and an additional $631 million (0.2 percent) in federal spending for premium tax credits, both due to the buy-in increasing nongroup premiums relative to the current nongroup market. EXECUTIVE SUMMARY VII We estimate the alternative buy-in scenarios would have the following effects if fully implemented in 2020:  Limiting eligibility for the buy-in plan to those ages 55 to 64 would reduce enrollment in the buy-in plan from 2.1 million under the base policy to 1.5 million.  Setting a uniform national premium would have little effect on coverage and spending relative to the base buy-in policy.  Computing premium tax credits for families with a mix of health insurance coverage (e.g., with members eligible for the buy-in and members under age 50 in the ACA Marketplaces) together, rather than computing subsidies independently, would have little effect on coverage and spending because relatively few families would be affected by the policy change.  Eliminating the ESI firewall for those ages 50 to 64 would increase buy-in plan enrollment from 2.1 million in the base scenario to 2.5 million, with commensurate reductions in employer- sponsored insurance enrollment. This policy would result in larger net savings ($3.3 billion) relative to the base policy because more people would enroll in the buy-in plan with lower payment rates.  Modernizing the Medicare benefit by adding a maximum on out-of-pocket spending and some additional benefits, such as routine dental, vision, and hearing, would have a modest effect on coverage, increasing buy-in plan enrollment to 2.4 million (from 2.1 million in the base scenario). This alternative would also result in somewhat larger net savings ($2.2 billion).  Reducing buy-in plan premiums to levels comparable with those in traditional Medicare would increase enrollment in the plan to nearly 2.9 million and reduce the number of uninsured adults ages 50 to 64 by 227,000, compared with current law. This policy would require substantially higher federal spending of $10.1 billion but would reduce household spending by $14.3 billion. It is estimated to reduce total health care spending by $5.4 billion.  Enhancing premium subsidies in the buy-in policy and current nongroup market for all nonelderly people would not have a large effect on buy-in plan enrollment, but it would reduce the number of uninsured by nearly 3 million. Compared with current law, this policy would reduce the number of uninsured people ages 50 to 64 by nearly 700,000 and reduce the number of uninsured people under age 50 by 2.3 million. Though this policy would require an additional $15.3 billion in federal spending, spending by all other payers would fall. By taking advantage of lower payment rates for enrollees, the buy-in policy would reduce total health care spending by $3.0 billion. VIII EXECUTIVE SUMMARY The main finding from our analysis is that a Medicare buy-in policy’s potential to substantially expand health insurance coverage is limited given the subsidies already provided under the ACA. Buy-in enrollment does not exceed 3 million in any of our scenarios, including ones with lower premiums or much more generous subsidies than the ACA provides. What the buy-in policy primarily does in the scenarios we modeled is take advantage of lower provider payment rates and thereby increase coverage generosity and reduce out-of-pocket spending for beneficiaries. Buy-in policies can result in savings to national health spending overall because of lower provider payment rates. Increases in federal spending under more expansive policy scenarios (enhanced subsidies and reducing buy-in premiums to traditional Medicare levels) can shift dollars from private payers to public payers in a way that is roughly neutral for aggregate spending. Some savings from more expansive options also likely result from shifts from employer-sponsored insurance to Marketplace coverage, which has lower payment rates than typical large-group commercial insurance. EXECUTIVE SUMMARY IX Errata This report was revised February 1, 2021. On page vii, incorrect parenthetical percentages in the fourth bullet were corrected (the dollar values are unaffected). That bullet now reads as follows:  On net, the base buy-in policy would reduce aggregate health care spending by about $1.8 billion (about 0.1 percent). This reflects a net decrease in spending of $2.8 billion (0.3 percent) for older adults and an increase in spending of about $1.1 billion (less than 0.1 percent) for people under age 50. Related changes were made on page 21, where the text above was repeated as the first sentence of the first bullet. That sentence was deleted, and the parenthetical percentages were moved to the second sentence. The bullet on page 21 now reads as follows:  Overall, we estimate the base buy-in scenario would lower total health care spending by $1.8 billion (0.1 percent). That net decline would result from an estimated $2.8 billion (0.3 percent) in savings for those in the buy-in-eligible age range (second panel) and $1.1 billion (less than 0.1 percent) in increased total spending among those below age 50 (third panel). In addition, some text was changed on page 22. The first full bullet was edited to clarify that (a) the buy-in policy would have little net effect on federal spending for adults ages 50 to 64 and (b) the amount of that effect is $95 million, as shown in table 8. In the second bullet, “($631 billion)” was changed to “($631 million),” as shown in table 6. ERRATA XI The Effects of Medicare Buy-In Policies for Older Adults on Health Insurance Coverage and Health Care Spending A Medicare buy-in program would allow qualifying individuals currently ineligible for Medicare to purchase a Medicare-like health insurance plan. The buy-in would be administered as a distinct program but could take advantage of Medicare's premium structure, benefit design, or provider payment rates (NASI 2020). After comprehensive health reform failed to materialize during the Clinton administration, Medicare buy-in policies gained prominence as potential incremental reforms (Johnson, Moon, and Davidoff 2002). 1 Given the popularity of the Medicare program, borrowing elements from the existing program may be an appealing way to build upon the reforms made under the Affordable Care Act (ACA), short of more comprehensive, Medicare for All– type reforms. Two recent legislative proposals that would offer a Medicare buy-in option for adults currently ineligible for Medicare are the Medicare at 50 Act (S. 470), sponsored by Senator Deborah Stabenow, and the Medicare Buy-In and Health Care Stabilization Act of 2019 (H.R. 1346), sponsored by Representative Brian Higgins. Both bills would offer adults ages 50 to 64 a new option to buy into Medicare. Both would allow purchase of traditional Medicare benefits (Parts A, B, and D) and would alternatively allow purchase of a Medicare Advantage (MA) plan. Current sources of private and public health coverage would continue under both plans. Whereas S. 470 would allow buy-in enrollees to access premium subsidies comparable to those available in the ACA Marketplaces, H.R. 1346 would feature enhanced subsidies beyond what the ACA provides. Unless enrollees opt for an MA plan, traditional Medicare cost-sharing rules would apply; unlike the ACA or MA, traditional Medicare has no out-of-pocket maximum on cost sharing. In contrast to the two Medicare buy-in proposals, President- elect Biden has proposed lowering the age of Medicare eligibility to 60 and creating a new Medicare- like health plan that could be purchased without age restriction. 2 Abstracting from the details of specific proposals, here we estimate the coverage and health care spending implications of a Medicare buy-in policy targeting adults ages 50 to 64, along with several variations. The base Medicare buy-in policy scenario we model would allow adults ages 50 to 64 to purchase a plan like traditional fee-for-service (FFS) Medicare, covering hospital care (Part A), physician and outpatient services (Part B), and prescription drugs (Part D). The plan would have an overall actuarial value of 85 percent, which is comparable with FFS Medicare overall. It would also reimburse providers at Medicare payment rates, which are typically lower than payment rates paid in the current nongroup market. Age-eligible individuals with an offer of employer-sponsored coverage deemed affordable per ACA rules would not be eligible to buy in. Enrollees in the buy-in plan would form their own risk pool; the full premium (before any applicable subsidies) would be set at the level needed to pay for the covered benefits of those who enroll. In this sense, the plan would be self-supporting. Unlike in the Marketplaces, in which premiums increase with age on a prescribed curve, all enrollees would pay the same premium (i.e., face no age rating within the 50–64 age group). Premiums would be adjusted to reflect geographic variation in health care prices (as in H.R. 1346). The buy-in plan would adopt the same premium tax credit and cost-sharing reduction structure as used in the ACA (supplemental tax credits available in Massachusetts and California under the ACA would extend to the buy-in as well). We also consider seven variations to the base policy design: 1. Restricting the eligibility age range to 55 to 64 2. Setting a single national average buy-in premium (as in S. 470), rather than making geographic adjustment to account for health care price variation 3. Coordinating the calculation of premium tax credits so families with mixed coverage (e.g., one spouse with Medicare buy-in and the other under age 50 in a Marketplace plan) pay only up to an applicable percentage of income for both coverages 4. Removing the employer-sponsored insurance (ESI) "firewall" by allowing those with affordable ESI coverage or offers to buy in to the program 5. Offering more extensive modernized Medicare benefits that include an out-of-pocket limit and coverage for dental, vision, and hearing services while maintaining an 85 percent actuarial value 6. Setting the full buy-in premium to a much lower level, comparable with what Medicare enrollees pay for Parts A (typically no premium), B, and D 7. Providing enhanced premium subsidies in both the Marketplace and buy-in plan 2 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS Using the Urban Institute's Health Insurance Policy Simulation Model (HIPSM), we estimate the implications of the base buy-in policy and the policy variations on health insurance coverage and health care spending, including changes for adults ages 50 to 64, who are the target population for these policies, and any effects on the population under age 50. In the next sections, we provide an overview of our modeling approach and more details on the buy-in policy scenarios. Methods We produced the estimates for this report using HIPSM, a detailed microsimulation model of the health insurance system designed to estimate the cost and coverage effects of proposed health care policy options. HIPSM is based on two years of data from the 2012–13 American Community Survey, which provides a representative sample of families constituting more than 6 million individuals of all ages. The sample is reweighted to reflect more recent information on income and demographics and aged to future years using recent American Community Survey data and projections from the Urban Institute’s Mapping America’s Futures program. In this analysis, we use the 2020 version of the model, and our findings reflect the enrollment and spending effects if the buy-in policy were fully implemented in 2020 (Buettgens and Banthin 2020). Further information on how we modeled insurance choices and estimated changes in provider payments rates are available in the appendix. Definition of the Base Medicare Buy-In Policy We defined the base Medicare buy-in policy as a traditional insurance product, similar to current FFS Medicare, sponsored by the federal government. We simplified the plan, combining Parts A, B, and D, and assumed a unified deductible of $189 and a 15 percent coinsurance rate on all services. We set the deductible and coinsurance rates at amounts that would achieve an actuarial value of 85 percent among participants in the new plan within our model; this roughly equals the actuarial value of the FFS Medicare benefit under current law. 3 As is the case with the Medicare FFS benefit, the base policy has no maximum on out-of-pocket expenditures. Eligibility would be restricted to people ages 50 to 64 who did not have an affordable offer of ESI, according to the same rules in the Marketplace. Premiums would not vary by age among those eligible for the program (ages 50 to 64) but would vary by geographic region. Dependents under age 50 would not be eligible. We assume undocumented immigrants would also be ineligible for the buy-in, just as they are currently ineligible to purchase Marketplace coverage. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 3 Enrollees in this new Medicare option would form a risk pool separate from that for traditional Medicare enrollees. In addition, premiums for the Medicare buy-in plan are expected to cover the costs of the program, unlike the traditional Medicare program, which is funded by a payroll tax, general funds under mandatory spending rules, and beneficiary premiums. This means premiums would be calculated based on the expenses of the group of people purchasing the new Medicare buy-in option and set at levels that would cover the costs of the program (before subsidies to individuals). Under the base scenario, the annual premium for single coverage would average $8,371 in 2020. Provider Payment Rates and Premiums under the Medicare Buy-In Policy An important feature of the Medicare buy-in policy is its ability to reimburse providers at Medicare payment rates. Consequently, Medicare buy-in plan premiums would be lower than if the plan had to pay providers at higher commercial rates or had to negotiate rates in each market. Medicare buy-in premiums would also be lower because of lower administrative costs. We assume the administrative loading factor for Medicare buy-in would be 6 percent, less than half the 15 percent administrative loading factor we assume for private plans in the nongroup market. Variations on the Base Medicare Buy-In Policy As noted above, we also consider seven variations to the base policy design, which we describe further in table 1. The first alternative would restrict eligibility to those ages 55 to 64 but is otherwise identical to the base scenario. The second alternative scenario would eliminate local differences in pricing by charging a national average premium. As noted earlier, the premiums for the Medicare buy-in plan under the base scenario would vary by state and rating region. The third alternative assumes premium tax credit calculations would be coordinated across programs for families with enrollees in both the Medicare buy-in and the federal and state-based Marketplaces, applying a percentage-of-income cap to the sum of all premiums facing each family. This coordination could substantially reduce premiums for such mixed-enrollment families but may be difficult to implement. The fourth scenario would eliminate the ACA’s firewall that prevents people with an affordable offer of coverage from their employer from receiving premium tax credits in the Marketplace. In our base scenario, the firewall would apply to both the Marketplace and the Medicare buy-in option. Under 4 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS this alternative scenario, the firewall would be eliminated for the Medicare buy-in but remain in place for the Marketplace. Consequently, the number of people eligible for subsidized Medicare buy-in coverage who are disqualified from Marketplace premium tax credits would increase. The fifth scenario would modernize Medicare benefits by adding an out-of-pocket maximum of $8,200 for single coverage ($16,400 for family coverage) and some dental, vision, and hearing benefits to the basic Medicare buy-in plan while keeping the same actuarial value of 85 percent. The primary difference between this and the base buy-in scenario is the reduced risk of high out-of-pocket health care costs. This makes the buy-in more attractive for some. However, the reduction in risk is a relatively small factor in the utility function, so we would not expect a significant behavioral response to this change in the plan benefits relative to the base buy-in plan. The sixth policy alternative would change the Medicare buy-in premium tax credit structure from one that replicates the ACA Marketplace calculations to one that replicates current Medicare Parts B and D out-of-pocket premium calculations. (Most Medicare enrollees do not pay premiums for Part A, and we therefore assume Part A–equivalent premiums for the buy-in plan are $0.) For many enrollees, this policy scenario would increase the generosity of premium subsidies relative to what is available to them in the Marketplaces. This alternative would also expand eligibility for premium subsidies because it eliminates the affordable ESI offer test, which does not exist in the traditional Medicare program. 4 The final policy alternative differs from the others in that it would make a major change to both Marketplace and Medicare buy-in plans. To protect current marketplace enrollees under age 50 from potential premium increases resulting from the implementation of the Medicare buy-in policy, this scenario would increase the generosity of subsidies for both age groups. We enhance premium tax credits and cost-sharing reductions in both the Marketplaces and the buy-in in a way similar to that used in earlier work (Blumberg, Holahan, Buettgens, et al. 2019). This scenario would scale down the ACA’s applicable percentage of income to a maximum 8.5 percent and raise the minimum actuarial value to 80 percent. The increased affordability of subsidized Marketplace coverage would not only increase enrollment but lower the average health risk of enrollees under 50. Consequently, the buy-in would decrease nongroup premiums, rather than raising them. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 5 TABLE 1 Descriptions of Medicare Buy-In Scenarios for Adults Ages 50 to 64 Scenario Eligibility Benefits and costs Premiums Base  Ages 50–64  $198 unified  No variation by age  Firewall for those with deductible  Geographic variation to access to affordable  15 percent cost reflect differences in health ESI sharing for all care prices services  ACA PTC/CSR subsidies are  No OOP available to enrollees in the maximum buy-in plan; for mixed-age  85 percent AV households, buy-in and (to reflect silver Marketplace subsidies are Marketplace computed independently of plan) each other Alternatives 1. Age restriction  Restricted to ages 55–64 2. National premium  No geographic variation in premiums 3. Coordinated PTCs  Households with mixed coverage pay only up to a percentage of income for both Marketplace and buy-in coverage 4. Removing ESI  Eligibility extended to firewall those with affordable ESI coverage 5. Modernized  Marketplace-like Medicare benefits structure at same AV (85 percent) with OOP maximum and dental, vision, and hearing benefits added 6. Medicare premiums  Premiums set to Medicare Parts B and D levels 7. Enhanced subsidies  Increased premium subsidies for buy-in participants and Marketplace enrollees, including those under age 50 in the Marketplaces Source: Urban Institute. Notes: ESI = employer-sponsored insurance. AV = actuarial value. OOP = out-of-pocket. ACA = Affordable Care Act. PTC = premium tax credit. CSR = cost-sharing reduction. OOP = out-of-pocket. Alternate scenarios are identical to the base scenario except for the specified change; only one parameter changes per scenario. 6 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS Findings for the Base Buy-In Policy Scenario In this section, we describe how the base buy-in policy would affect health insurance coverage, premiums, health care spending for individuals, and aggregate health care spending. Changes in Health Insurance Coverage In table 2, we present how the distribution of health insurance coverage would shift, relative to current law, if the base buy-in policy were introduced and fully implemented in 2020. Under current law in 2020, among an estimated 62 million older adults ages 50 to 64,  38 million (61 percent) have coverage through an employer,  8.5 million (14 percent) have Medicaid coverage,  5.6 million (9 percent) are uninsured (including those with coverage through a plan that is noncompliant with minimum insurance requirements under the ACA),  5.8 million (9 percent) have private nongroup coverage (i.e., coverage through an ACA Marketplace plan or other private nongroup plan), and  4.2 million (7 percent) have other public coverage (i.e., Medicare coverage for nonelderly people with disabilities). About 2.1 million older adults would enroll in the Medicare buy-in plan, most shifting from current Marketplace or nongroup coverage. Of the 2.1 million people who would enroll in the base Medicare buy- in plan (table 2, column 2), an estimated 775,000 would purchase these plans at reduced price with premium tax credits (data not shown). Others would obtain coverage with higher actuarial value at a similar price. We estimate implementing the Medicare buy-in plan would result in very little change in coverage for those who have ESI, Medicaid, or other public coverage under current law. We also see introducing the buy-in policy would have a relatively small effect on the number of uninsured older adults; only 176,000 people gain coverage, representing a 3 percent reduction in the 5.6 million uninsured adults in this age group. 5 However, we estimate the introduction of the base buy-in policy would result in substantial changes to the nongroup insurance market; about 1.9 million people ages 50 to 64 who had purchased Marketplace plans or other nongroup coverage would no longer do so when the Medicare buy-in plan is available (a 33 percent decrease relative to current law), nearly all of which represents shifts to the buy-in plan. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 7 TABLE 2 Health Insurance Enrollment under Current Law and Base Medicare Buy-In Scenario, by Age Group, 2020 Base buy-in Current law scenario Change (thousands of (thousands of Thousands of people) people) people Percent Total nonelderly population Insured 244,346 244,413 67 0.03 ESI 151,117 151,018 -99 -0.07 Medicare buy-in — 2,089 2,089 — Marketplace or other nongroup 15,131 13,237 -1,894 -12.52 Medicaid 69,478 69,450 -29 -0.04 Other public 8,619 8,619 0 0.00 Uninsured 31,128 31,062 -67 -0.21 Total 275,474 275,474 0 0.00 Ages 50–64 Insured 56,523 56,700 176 0.31 ESI 38,053 38,021 -32 -0.08 Medicare buy-in — 2,089 2,089 — Marketplace or other nongroup 5,751 3,873 -1,878 -32.66 Medicaid/CHIP 8,523 8,521 -2 -0.03 Other public 4,196 4,196 0 0.00 Uninsured 5,593 5,416 -176 -3.15 Total 62,116 62,116 0 0.00 Under 50 Insured 187,822 187,713 -110 -0.06 ESI 113,064 112,997 -67 -0.06 Medicare buy-in — 0 0 — Marketplace or other nongroup 9,380 9,364 -16 -0.17 Medicaid/CHIP 60,955 60,929 -26 -0.04 Other public 4,423 4,423 0 0.00 Uninsured 25,536 25,645 110 0.43 Total 213,358 213,358 0 0.00 Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: ESI = employer-sponsored insurance. CHIP = Children’s Health Insurance Program. A dash indicates the column heading does not apply. The uninsured category includes all people without minimum essential coverage, including those with short-term, limited-duration plans. Other nongroup coverage includes ACA-compliant policies purchased outside the Marketplace. 8 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS Introducing the base buy-in policy would have small, indirect effects for those under age 50, including slightly reduced coverage. Table 2 also shows the buy-in policy’s possible indirect effects on coverage for those under 50. By attracting a substantial portion of older adults in the current nongroup market, the buy-in policy could affect the remaining nongroup risk pool. We estimate the buy-in policy would have a small but negative effect on coverage, increasing the number of uninsured under age 50 by an estimated 110,000. Most of this increase would be from a reduction in ESI coverage, as adults ages 50 and older enroll in the buy-in option and younger family members find alternative coverage sources, with some younger family members opting to remain uninsured. An estimated 16,000 of the nearly 9.4 million adults under age 50 with nongroup coverage under current law would lose such coverage because of higher premiums. We also estimate that full nongroup premiums for single coverage would increase from $4,905 to $5,096, on average, and out-of-pocket premiums would increase slightly from $2,054 to $2,094 (data not shown). Income-based tax credits for both Marketplace and buy-in coverage would shield people with lower incomes from premium differences and limit their enrollment in the buy-in plan. Figure 1 presents changes in Medicare buy-in enrollment, Marketplace or other nongroup coverage, and the number uninsured under the base buy-in scenario relative to current law; table 3 shows additional detail on coverage levels and changes by income group. Among people with incomes below 138 percent of FPL, about 122,000 would enroll in the buy-in plan, and most of this change can be attributed to a reduction in enrollment in Marketplace plans (81,000-enrollee decline). Also among this income group, the number of uninsured people would drop slightly by 31,000. We would expect very few individuals with incomes below 138 percent of FPL in states that expanded Medicaid under the ACA (expansion states) to transfer to the buy-in, because they can already access Medicaid coverage at no cost. 6 Indeed, we find that, among those with incomes below 138 percent of FPL, 59 percent of those enrolling in the buy-in plan would reside in states that have not expanded Medicaid under the ACA (nonexpansion states; data not shown). Among people with incomes below 138 percent of FPL, those living in nonexpansion states would take up the Medicare buy-in option at 2.5 times the rate of those living in expansion states (data not shown). MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 9 FIGURE 1 Change in Enrollment under Base Medicare Buy-In Scenario Relative to Current Law, by Income Group, 2020 Thousands of people < 138% of FPL Uninsured -31 Marketplace or other nongroup -81 Medicare buy-in 122 138–200% of FPL Uninsured -37 Marketplace or other nongroup -222 Medicare buy-in 263 200–400% of FPL Uninsured -49 Marketplace or other nongroup -651 Medicare buy-in 710 > 400% of FPL Uninsured -60 Marketplace or other nongroup -925 Medicare buy-in 994 URBAN INSTITUTE Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: FPL = federal poverty level. The federal poverty level for a household of two in 2020 is $17,240 per year. Values reflect totals for all coverage types in each income group. The uninsured category includes all people without minimum essential coverage, including those with short-term, limited-duration plans. Other nongroup coverage includes ACA-compliant policies purchased outside the Marketplace. 10 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS TABLE 3 Health Insurance Enrollment under Current Law and Base Medicare Buy-In Scenario among Adults Ages 50 to 64, by Income Group, 2020 Thousands of people Current law Base buy-in scenario Change < 138% of FPL Medicare buy-in — 122 122 Marketplace or other nongroup 643 562 -81 Uninsured 2,449 2,418 -31 All other coverage types 9,625 9,614 -11 Total 12,717 12,717 — 138%–200% of FPL Medicare buy-in — 263 263 Marketplace or other nongroup 1,541 1,319 -222 Uninsured 528 491 -37 All other coverage types 2,951 2,948 -3 Total 5,020 5,020 — 200%–400% of FPL Medicare buy-in — 710 710 Marketplace or other nongroup 1,944 1,293 -651 Uninsured 1,292 1,243 -49 All other coverage types 11,933 11,922 -10 Total 15,169 15,169 — > 400% of FPL Medicare buy-in — 994 994 Marketplace or other nongroup 1,623 699 -925 Uninsured 1,324 1,264 -60 All other coverage types 26,263 26,253 -10 Total 29,210 29,210 — Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: FPL = federal poverty level. A dash indicates the column heading does not apply. The federal poverty level for a household of two in 2020 is $17,240 per year. Values reflect totals for all coverage types in each income group. Uninsured category includes all people without minimum essential coverage, including those with short-term, limited-duration plans. Other nongroup coverage includes ACA-compliant policies purchased outside the Marketplace. Among adults ages 50 to 64 with incomes between 138 percent and 200 percent of FPL, we estimate about 263,000 would enroll in the base buy-in plan. We also estimate a similarly sized group of people in this income range (222,000) would no longer enroll in Marketplace or other nongroup plans under the base buy-in scenario. This income group is expected to predominantly qualify for premium tax credits for both the Marketplace and the buy-in, and, relative to those for people with incomes between 200 percent and 400 percent of FPL, the tax credits are large. Thus, transitions between Marketplace and buy-in coverage for this group would likely reflect the higher actuarial value of buy-in plans relative to that of Marketplace plans (i.e., the metal-tier-weighted average actuarial value). We investigate this possibility in the premiums section below. Lastly, the number of uninsured individuals with incomes between 138 percent and 200 percent of FPL would fall by 37,000 under the buy-in policy. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 11 Among people with incomes between 200 percent and 400 percent of FPL, about 710,000 would enroll in the Medicare buy-in plan under the base scenario. For people with incomes in this range, introduction of the plan would result in an estimated 651,000 leaving Marketplace plans or other nongroup plans and 49,000 gaining coverage. The buy-in plan would disproportionately attract people with higher incomes; nearly half would have incomes at or above 400 percent of FPL. As shown in figure 1 and table 3, we estimate nearly 1.0 million people with incomes above 400 percent of FPL would enroll in a Medicare buy-in plan. People in households earning incomes above 400 percent of FPL do not qualify for premium tax credits for Marketplace plans, nor would they qualify for premium tax credits under the base buy-in scenario. Being fully exposed to the premium costs, they would likely benefit most from the lower premiums resulting from the buy-in plan’s reduced provider payment rates. The buy-in plan is expected to reduce Marketplace and other nongroup enrollment by 925,000 and the number of uninsured people 60,000 among this higher-income group. The buy-in plan would also disproportionately attract older individuals within the 50–64 age group; nearly half of enrollees would be ages 60 to 64. Figure 2 presents, by age group, estimates of buy-in enrollment and changes in Marketplace or other nongroup coverage and uninsurance rates following the introduction of a Medicare buy-in plan. Adults ages 60 to 64 would make up the largest share of buy-in enrollment. Among people ages 50 to 54, 505,000 would enroll in the buy-in plan. Among those ages 55 to 59, 653,000 would enroll in the buy-in plan. Among those ages 60 to 64, 931,000 would enroll in the buy-in plan. As we explain below, the older age group would benefit most from the community rating within the buy-in plan. Across all three age groups, nearly as many people would drop Marketplace or other nongroup coverage as would enroll in a Medicare buy-in plan. Despite the differences in buy-in enrollment by age, reductions in the number of uninsured people would be roughly equal across all three age groups. 12 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS FIGURE 2 Change in Enrollment under Base Medicare Buy-In Scenario Relative to Current Law, by Age Group, 2020 Thousands of people 50–54 Uninsured -73 Marketplace or other nongroup -423 Medicare buy-in 505 55–59 Uninsured -52 Marketplace or other nongroup -590 Medicare buy-in 653 60–64 Uninsured -51 Marketplace or other nongroup -866 Medicare buy-in 931 URBAN INSTITUTE Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: Values reflect totals for all coverage types in each age group. The uninsured category includes all people without minimum essential coverage, including those with short-term, limited-duration plans. Other nongroup coverage includes ACA-compliant policies purchased outside the Marketplace. Changes in Premiums Premiums reflect expected health care costs and administrative costs of beneficiary groups (risk pools). One of the buy-in plan’s main features is that health care provider prices are set at Medicare payment rates, which are typically lower than payment rates of private nongroup plans, particularly in less densely populated areas. Payment rates can also be very low in more competitive areas in the Marketplaces (Blumberg, Holahan, McMorrow, et al. 2020). In isolation, these reduced payment rates would decrease buy-in plan premiums relative to those for private nongroup plans. Moreover, Marketplace plans permit three-to-one age rating on premiums, whereas buy-in plans would charge the same premiums for those ages 50 to 64 (benefitting those who are older and face higher risk). Finally, both buy-in and Marketplace premiums would be affected by the risk profiles of those who enroll in the buy-in plan versus those who choose to stay enrolled in the Marketplace. Predicting these enrollment changes based on expected health care costs is challenging; though buy-in plans have a higher actuarial value (85 percent) than Marketplace gold plans (approximately 80 percent, often the highest available MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 13 in the Marketplace because platinum plans are only available in a few markets), 7 they follow traditional Medicare in not including an out-of-pocket maximum, whereas Marketplace plans do. Thus, some enrollees expecting to use a significant amount of health care could benefit from remaining in the Marketplace. Given these wide-ranging factors, it is difficult to predict how the introduction of a buy-in plan will affect premiums in the Marketplace and which enrollees would likely therefore choose to enroll in the buy-in plan. Further complicating the analysis is actuarial evidence that older Marketplace enrollees (ages 50 and older) cross-subsidize younger enrollees, as reported in a Milliman study (Kotecki and Westrom 2020). That is, older Marketplace enrollees are in slightly better than average health for their age than younger enrollees, who are in slightly worse than average health for their age. The official age curve determines how Marketplace premiums increase by age and does not fully account for the differences in current enrollees’ actual costs by age. This relationship suggests that if sufficient numbers of older Marketplace enrollees leave the Marketplace, premiums for those who remain might increase. This is a somewhat unexpected result, but one also found by RAND researchers (Eibner et al. 2019). Among buy-in enrollees switching from Marketplace or other nongroup coverage under current law, total premiums for single coverage before tax credits would decline from $8,843 to $8,371, even as the actuarial value of the coverage would rise from 65 percent, on average, to 85 percent. In table 4, we explore how the introduction of a buy-in plan would affect (1) the average single-policy premiums in private nongroup plans (across all metal tiers) and (2) the average premiums for the buy-in for those ages 50 to 64 previously enrolled in private nongroup plans under current law. We present premiums separately for those opting to enroll in the buy-in plan and those opting to stay in private nongroup plans. We limit the analysis to single enrollees to isolate per capita premiums. Further, table 4 presents premium differences separately by age group and premium tax credit eligibility. Introducing the buy-in policy would lead 33 percent of people enrolled in private nongroup plans with single coverage under current law to switch to the buy-in plan. 8 For these people overall, we estimate the average current-law premium for a private nongroup plan is $8,843. The average full premium for the buy-in for these enrollees would be $8,371. Thus, the buy-in plan would reduce premiums for those switching from nongroup coverage because of the lower provider payment rates key to the buy-in plan. And buy-in coverage would be much richer, with an actuarial value of 85 percent, compared with the average 65 percent actuarial value of plans in the current nongroup market. For those switching to buy-in plans, the average current-law private nongroup out-of-pocket average premium is $5,579, compared with the slightly higher average premium of $5,651 for buy-in plans. 14 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS TABLE 4 Annual Premiums and Plan Actuarial Value for Single Coverage for Current-Law Nongroup Enrollees Ages 50 to 64, by Age, Subsidy Eligibility, and Whether Enrollees Switch from Nongroup to Medicare Buy-In Coverage, 2020 Switching from Nongroup to Staying in Nongroup Coverage Buy-In Coverage Current-law Buy-in policy Current law After reform nongroup after reform Total nonelderly population 2,137,000 1,058,000 Share taking up buy-in plans 33% Full single premium $9,752 $10,142 $8,843 $8,371 Out-of-pocket premium $2,073 $2,101 $5,579 $5,651 Plan AV 83% 83% 65% 85% Eligible for subsidy Ages 50–54 455,000 156,000 Share taking up buy-in plans 25% Full single premium $8,058 $8,366 $7,214 $8,619 Out-of-pocket premium $1,083 $1,076 $2,430 $2,905 Plan AV 87% 87% 75% 85% Ages 55–59 644,000 173,000 Share taking up buy-in plans 21% Full single premium $9,811 $10,206 $8,346 $8,628 Out-of-pocket premium $1,221 $1,209 $2,495 $3,300 Plan AV 85% 85% 72% 86% Ages 60–64 738,000 216,000 Share taking up buy-in plans 23% Full single premium $11,592 $12,056 $9,986 $8,584 Out-of-pocket premium $1,209 $1,195 $2,591 $3,637 Plan AV 85% 85% 69% 86% Ineligible for subsidy Ages 50–54 107,000 88,000 Share taking up buy-in plans 45% Full/out-of-pocket single premium $6,526 $6,790 $7,100 $7,720 Plan AV 67% 67% 57% 84% Ages 55–59 120,000 128,000 Share taking up buy-in plans 52% Full/out-of-pocket single premium $7,871 $8,157 $8,259 $8,058 Plan AV 68% 68% 57% 85% Ages 60–64 72,000 297,00 Share taking up buy-in plans 80% Full/out-of-pocket single premium $8,475 $8,735 $9,590 $8,267 Plan AV 72% 72% 58% 85% Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: AV = actuarial value. Population includes only those with single nongroup coverage under current law, excluding those in Basic Health Programs. Premiums and actuarial values are averaged over all metal-tier choices, including cost-sharing reductions. Among the population eligible for Medicare buy-in, their health plans’ actuarial values vary. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 15 Among adults ages 50 to 64 who stay in the nongroup market after the buy-in is introduced, average premiums before subsidies would increase from $9,752 to $10,142 for single enrollees, but out-of-pocket premiums would remain little changed. For those who would stay in the nongroup market after a buy-in policy is introduced, we estimate the average annual full premium of a private nongroup plan under current law is $9,752. For these people, the buy-in would raise their private nongroup plan premiums to $10,142. However, the ACA provides significant subsidies for premiums for adults with household incomes between 100 percent and 400 percent of FPL (138 percent to 400 percent of FPL in Medicaid expansion states), and consequently, very few people pay full premiums for coverage. For people remaining in nongroup coverage, the average out-of-pocket spending for private nongroup plans would not be greatly affected by the buy-in plan’s introduction ($2,073 under current law and $2,101 under the base buy-in). Notably, those who switch to the buy-in would pay much higher out-of-pocket premiums under current law than those staying in nongroup coverage. This suggests people choosing to enroll in a buy-in plan will likely have more limited eligibility for premium tax credits than those remaining in the private nongroup market, consistent with the finding above that buy-in enrollment would skew toward those with higher incomes. Some people switching to the buy-in would pay more and others would pay less in out-of-pocket premiums than they currently pay in the nongroup market, and reductions would be concentrated among older people and those not receiving premium subsidies. The second panel of table 4 presents differences in premiums by age group for single people eligible for premium tax credits. Across all three age groups, the estimated share of enrollees who switch from current nongroup coverage to the buy-in is roughly similar, ranging from 21 percent to 25 percent. Full premiums for single coverage under current law show a clear age gradient, reflecting the age-rating curve in the Marketplace. Out-of-pocket premiums show a relatively flat age gradient in the current nongroup market, however, because premium subsidies limit how much people must pay. Though out-of-pocket premiums would increase for people switching to buy-in plans, the actuarial value of their coverage would increase substantially. For those ages 60 to 64, for example, out-of-pocket premiums would increase from $2,591 to $3,637, while actuarial value would increase from 69 percent to 86 percent. The 17 percentage-point increase in actuarial value is worth roughly $1,700 in additional spending on medical care covered by insurance for the average enrollee ages 50 to 64. 9 It may be worth more to those ages 60 to 64. 16 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS Among people switching to the buy-in plan who are ineligible for premium subsidies in the current nongroup market, single buy-in enrollees ages 50 to 54 would pay higher premiums, and single enrollees ages 60 to 64 would pay lower premiums. The last panel of table 4 presents differences in premiums for single people ineligible for premium tax credits. A stark difference between this group and the subsidy-eligible group is the higher rate of switching to the buy-in. Forty-five percent, 52 percent, and 80 percent of subsidy- ineligible nongroup enrollees would enroll in the buy-in among those ages 50 to 54, 55 to 59, and 60 to 64. Whereas subsidy-ineligible enrollees are fully exposed to the ACA's age-rating curve in the current nongroup market (ranging from $7,100 to $9,590 for those who switch to the buy-in), buy-in premiums would be community rated and show very little difference by age group ($7,720 to $8,267). Consequently, those ages 60 to 64 would pay around $1,300 less in premiums under the buy-in, and those ages 50 to 54 would pay about $600 more. As with subsidy-eligible people, subsidy-ineligible people opting for the buy-in would gain substantial insurance protection in the form of increased actuarial value. For this group, the increases in actuarial value would be substantially greater than for those eligible for subsidies. Changes in Health Care Spending for Individuals Switching from the current nongroup market to the buy-in plan would affect individuals' total health care spending in two ways. First, because buy-in coverage would often be more generous (i.e., with lower cost sharing and higher actuarial value), it could induce people to use more health care services. Second, because provider payment rates under the buy-in play would typically be lower than they are in the current nongroup market, spending would also be lower under the buy-in plan for a given amount of utilization. In figure 3, we show how total health care spending per enrollee would change, on net, for those switching from nongroup coverage to the buy-in plan. We also decompose the change into the parts attributable to changes in utilization and differences in payment rates. For those switching to the buy-in plan, health care utilization would be slightly higher but at a substantially lower total cost. Overall, average total health care spending per person would decrease from $9,789 in the nongroup market to $8,944 in the buy-in plan, a net reduction of $845 (figure 3). However, this reduction reflects two partially offsetting effects: (1) an estimated increase in utilization valued at $239 at "standardized" payment rates (i.e., prevailing payment rates in the existing nongroup market) and (2) a reduction in per enrollee spending of $1,084, owing to lower provider payment rates under the buy-in plan. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 17 Those switching to the Medicare buy-in would mainly benefit from lower out-of-pocket spending. In table 5, we examine changes in health care spending for two groups: those who would stay in the existing nongroup market after the buy-in is introduced and those who would switch to the buy-in plan. The table shows total spending at standardized payment rates so differences reflect changes in utilization (as in figure 3). We also report out-of-pocket spending for health care and premiums that incorporate the buy-in policy’s lower payment rates, which lower costs for consumers. Heath care spending for those staying in the nongroup market would change only modestly, so we focus the discussion on effects for those who switch to a buy-in plan. 10 They make up 34 percent of the total number of adults ages 50 to 64 with nongroup coverage under current law (3.6 million staying plus 1.9 million switching). FIGURE 3 Health Spending per Enrollee Switching from Nongroup to Base Buy-In Coverage, 2020 URBAN INSTITUTE Source: Urban Institute Health Insurance Policy Simulation Model, 2020. 18 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS TABLE 5 Total Health Spending per Capita and Out-of-Pocket Costs among Current-Law Nongroup Enrollees Ages 50 to 64, by Age, Subsidy Eligibility, and Whether Enrollees Switch from Nongroup to Medicare Buy-In Coverage, 2020 Staying in Nongroup Switching from Nongroup to Coverage Buy-In Coverage Current law Buy-in policy Current law After reform nongroup after reform Total nonelderly population 3,634,000 1,859,000 Share taking up buy-in plans 34% Total health spending at standardized (Marketplace) payment rates $10,693 $10,693 $9,789 $10,028 Out-of-pocket health spending $2,489 $2,489 $3,414 $1,528 Out-of-pocket premium $2,593 $2,653 $6,430 $6,230 Standardized spending less out-of- pocket payments $5,611 $5,551 -$56 $2,271 Eligible for subsidy Income < 200% of FPL 1,594,000 181,000 Share taking up buy-in plans 10% Total health spending at standardized (Marketplace) payment rates $10,428 $10,428 $7,624 $7,918 Out-of-pocket health spending $1,281 $1,281 $1,039 $972 Out-of-pocket premium $681 $680 $1,261 $1,186 Standardized spending less out-of- pocket payments $8,466 $8,467 $5,324 $5,760 Income > 200% of FPL 1,170,000 516,000 Share taking up buy-in plans 31% Total health spending at standardized (Marketplace) payment rates $11,270 $11,270 $8,075 $8,456 Out-of-pocket health spending $3,456 $3,456 $2,808 $1,451 Out-of-pocket premium $1,593 $1,568 $2,772 $3,871 Standardized spending less out-of- pocket payments $6,221 $6,246 $2,496 $3,133 Ineligible for subsidy Income < 400% of FPL 174,000 263,000 Share taking up buy-in plans 60% Total health spending at standardized (Marketplace) payment rates $12,742 $12,742 $12,892 $13,196 Out-of-pocket health spending $3,396 $3,396 $3,761 $1,901 Out-of-pocket premium $7,839 $8,126 $8,771 $7,969 Standardized spending less out-of- pocket payments $1,507 $1,220 $360 $3,326 Income > 400% of FPL 697,000 899,000 Share taking up buy-in plans 56% Total health spending at standardized (Marketplace) payment rates $9,818 $9,818 $10,298 $10,428 Out-of-pocket health spending $3,398 $3,398 $4,138 $1,575 Out-of-pocket premium $7,339 $7,624 $8,883 $8,088 Standardized spending less out-of- pocket payments -$919 -$1,204 -$2,722 $765 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 19 Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: FPL = federal poverty level. The federal poverty level for a household of two in 2020 is $17,240 per year. Population includes only those with single nongroup coverage under current law, excluding those in Basic Health Programs. Premiums are averaged over all metal-tier choices, including cost-sharing reductions. Total health care spending at standardized payment rates would increase from $9,789 to $10,028 because of increased health care use. This results from (1) the higher actuarial value of buy-in plans relative to Marketplace plans and (2) reductions in consumer cost sharing that lead to modest increases in demand for health care. Out-of-pocket health care spending (consumer cost sharing) for those switching to the buy-in would fall from $3,414 to $1,528 because of the buy-in plan’s higher actuarial value. Out-of-pocket premiums would fall only slightly from $6,430 to $6,230. Thus, buy-in enrollees would pay less out-of-pocket and consume slightly more health care but at a lower total cost than they would in the current nongroup market. As a summary measure of the net benefit (or loss) individuals would experience with the introduction of a buy-in policy, we also compute the change in standardized total spending less out-of- pocket payments. This measure captures the net change in total health care utilization (the benefit) less the change in out-of-pocket premiums and health care cost sharing (the cost), thereby providing an indicator of the buy-in policy’s net benefit. Individuals’ standardized spending less out-of-pocket payments would increase from a $56 deficit to a $2,271 surplus, or a change in net benefit of $2,326. Table 5 shows how current nongroup market spending levels would differ for those who stay in the nongroup market or switch to a buy-in policy if it were introduced. Those who switch would spend somewhat less than under current law ($9,789 versus $10,693) but would pay substantially more out of pocket for health care ($3,414 versus $2,489) and premiums ($6,430 versus $2,593). This pattern suggests those with higher out-of-pocket spending would be more attracted to the buy-in plan. With their relatively high health care spending, those currently ineligible for premium tax credits would experience the largest net benefit from switching from nongroup coverage to the buy-in plan. Table 5 shows analogous changes in spending measures for four subgroups: those eligible for premium tax credits with income below 200 percent of FPL, those who are subsidy eligible with incomes at or above 200 percent of FPL, those ineligible for subsidies with incomes below 400 percent of FPL, and those ineligible for subsidies with incomes at or above 400 percent of FPL. Of the estimated 1.8 million adults ages 50 to 64 who are eligible for subsidies and have incomes below 200 percent of FPL, we estimate just 10 percent would enroll in a buy-in plan if one became available. Those switching to buy-in plans would generally have much lower total health care spending than those staying in nongroup coverage, but they would pay higher out-of-pocket premiums. This 20 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS indicates they would likely have more income and thus qualify for less in subsidies than those staying in nongroup coverage. For these people, moving to the buy-in plan would raise their standardized spending less out-of-pocket spending by $436 (from $5,324 to $5,760). For subsidy-eligible people with incomes above 200 percent of FPL, a larger share (31 percent) would switch to buy-in plans. In switching, they would pay less in cost sharing but more in out-of-pocket premiums and, on net, increase their standardized spending less out-of-pocket spending by $637 (from $2,496 to $3,133). People in the subsidy-ineligible groups would experience the largest gains from switching to the buy-in plan. For those with incomes below 400 percent of FPL, 60 percent would move to the buy-in plan. And 56 percent of those with incomes at or above 400 percent of FPL would move to the buy-in plan. Both income groups have high total spending levels (relative to the subsidy-eligible groups) and would experience substantial reductions in out-of-pocket spending given the buy-in plan’s higher actuarial value. For those with incomes below 400 percent of FPL, standardized spending less out-of- pocket payments would increase from $360 to $3,326 after switching to the buy-in plan. For those with incomes at or above 400 percent of FPL—the largest of the four subgroups, constituting about 900,000 people—standardized spending less out-of-pocket payments would increase from a $2,722 deficit to a $765 surplus. Changes in Aggregate Health Care Spending To understand the overall financial implications of the buy-in policy, including what it would take to publicly finance it, we estimate changes in total health care spending for all nonelderly individuals. We provide estimates by payer and separately for those ages 50 to 64 and those under age 50 (table 6). We find the following:  Overall, we estimate the base buy-in scenario would lower total health care spending by $1.8 billion (0.1 percent). That net decline would result from an estimated $2.8 billion (0.3 percent) in savings for those in the buy-in-eligible age range (second panel) and $1.1 billion (less than 0.1 percent) in increased total spending among those below age 50 (third panel).  The net decline in health care spending for adults ages 50 to 64 would be driven mainly by reduced household spending resulting from more generous buy-in coverage. Though out-of- pocket health care spending by households for those ages 50 to 64 would decline by $3.6 billion because of the more generous buy-in coverage, total premium spending would increase by $1.4 billion because of 176,000 people gaining coverage. Thus, total household spending would decline by about $2.2 billion. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 21  The buy-in policy would have little net effect on federal spending for adults ages 50 to 64 ($95 million, or an increase of less than 0.1 percent). We estimate that the buy-in plan would cost the federal government about $4.0 billion. However, this cost would be almost fully accounted for by redirecting premium tax credits from the Marketplace to the buy-in (an offset of $3.9 billion). The net result would be little net change in federal spending (an increase of $95 million, as shown in table 8) for those ages 50 to 64.  The net increase in health care spending for those under age 50 would mainly reflect increased premium spending by households ($656 million) and increased federal spending for premium tax credits ($631 million) as shown in table 6. Both effects would result from the modest increases in nongroup premiums in the current nongroup market under the buy-in policy. 22 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS TABLE 6 Health Care Spending for Adults Ages 50 to 64 under Current Law versus the Base Medicare Buy-In Scenario, by Age Group and Funding Source, 2020 Millions of dollars Household Federal Government Employer Providers’ OOP Medicare State Premium Uncompensated Premiums payments buy-in Other Government Contributions Care Total Total nonelderly population Current law 272,849 273,027 — 435,130 198,624 720,926 22,975 1,923,531 Base buy-in scenario 274,772 269,342 3,996 431,771 198,462 720,551 22,803 1,921,698 Difference 1,923 -3,685 3,996 -3,359 -162 -375 -172 -1,833 Ages 50–64 Current law 120,214 122,499 — 182,152 79,125 295,475 11,657 811,123 Base buy-in scenario 121,609 118,929 3,996 178,251 78,905 295,205 11,452 808,347 Difference 1,395 -3,570 3,996 -3,901 -221 -270 -206 -2,775 Under 50 Current law 166,518 175,121 — 306,068 146,897 458,231 12,922 1,265,757 Base buy-in scenario 167,174 174,978 — 306,699 146,977 458,088 12,983 1,266,899 Difference 656 -143 — 631 80 -143 61 1,143 Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: OOP = out-of-pocket; these payments are nonpremium spending on health care (i.e., cost sharing). A dash indicates the column heading does not apply. All other federal government spending is Medicaid, Marketplace premium tax credits and cost-sharing reductions, and uncompensated care. State government spending is Medicaid, Marketplace premium tax credits and cost-sharing reductions, and uncompensated care. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 23 Effects of Alternative Buy-In Scenarios on Coverage and Aggregate Spending In addition to our base buy-in scenario, we estimate seven alternative scenarios that change features of the base buy-in program in isolation. In this section, we first summarize the overall trends across the alternative scenarios and key findings and then describe in greater detail each alternative’s impact on coverage and spending. The discussion draws from table 7, which summarizes the coverage effects of all alternative scenarios relative to current law, and table 8, which presents impacts on spending by payer for all scenarios. Again, the effects of each scenario are estimated in isolation from the others. Most of the alternative buy-in scenarios result in small differences in overall coverage and spending, similar to the base buy-in scenario. Though net differences across most scenarios are small, alternative scenarios likely affect different segments of the population. For example, removing the ESI firewall would increase coverage among those who previously had offers of affordable ESI, coordinating premium tax credits would increase coverage among those with household members above and below the buy-in eligibility age group, and modernizing Medicare benefits to include out-of-pocket maximums (among other benefits) would incentivize those with higher expected health care costs to enter the buy- in program. Consequently, we predict that combining features of these alternatives could result in larger gains in coverage and differences in spending by payer. Comparatively, the alternative scenario with enhanced subsidies would result in larger gains in coverage than the base buy-in scenario and other alternatives by extending enhanced subsidies to both the Medicare-buy in and entire nongroup populations. Federal spending would also increase to a greater degree than in other scenarios assessed. 24 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS TABLE 7 Changes in Health Insurance Enrollment under Alternative Medicare Buy-In Scenarios Relative to Current Law, by Age Group, 2020 Thousands of people Alternative Medicare Buy-In Scenarios Eligibility Current Base restricted to National Coordinated No ESI Modernized Medicare Enhanced law buy-in ages 55–64 premium PTCs firewall benefits premiums subsidies Total nonelderly population Insured 244,346 67 -4 66 83 77 70 50 2,972 Employer 151,117 -99 -93 -98 -100 -445 -107 -156 -1,030 Medicare buy-in — 2,089 1,515 2,082 2,114 2,512 2,379 2,880 2,237 Marketplace or other nongroup 15,131 -1,894 -1,397 -1,890 -1,903 -1,962 -2,174 -2,645 1,794 Medicaid/CHIP 69,478 -29 -29 -29 -29 -29 -29 -29 -29 Medicare or other public 8,619 0 0 0 0 0 0 0 0 Uninsured 31,128 -67 4 -66 -83 -77 -70 -50 -2,972 Total 275,474 0 0 0 0 0 0 0 0 Ages 50–64 Insured 56,523 176 95 175 193 185 181 227 697 Employer 38,053 -32 -27 -31 -32 -377 -39 -89 -618 Medicare buy-in — 2,089 1,515 2,082 2,114 2,512 2,379 2,880 2,237 Marketplace or other nongroup 5,751 -1,878 -1,391 -1,874 -1,886 -1,948 -2,158 -2,561 -919 Medicaid 8,523 -2 -2 -2 -2 -2 -2 -2 -2 Medicare or other public 4,196 0 0 0 0 0 0 0 0 Uninsured 5,593 -176 -95 -175 -193 -185 -181 -227 -697 Total 62,116 0 0 0 0 0 0 0 0 Under 50 Insured 187,822 -110 -98 -109 -110 -108 -110 -178 2,275 Employer 113,064 -67 -66 -67 -67 -68 -68 -67 -412 Medicare buy-in — — — — — — — — — Marketplace or other nongroup 9,380 -16 -6 -16 -17 -14 -16 -84 2,714 Medicaid/CHIP 60,955 -26 -26 -26 -26 -26 -26 -26 -26 Medicare or other public 4,423 0 0 0 0 0 0 0 0 Uninsured 25,536 110 98 109 110 108 110 178 -2,275 Total 213,358 0 0 0 0 0 0 0 0 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 25 Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: PTCs = premium tax credits. ESI = employer-sponsored insurance. A dash indicates the table heading does not apply. The uninsured category includes all people without minimum essential coverage, including those with short-term, limited-duration plans. Other nongroup coverage includes ACA-compliant policies purchased outside the Marketplace. TABLE 8 Changes in Health Insurance Spending under Alternative Medicare Buy-In Scenarios Relative to Current Law, by Age Group, 2020 Millions of dollars Alternative Medicare Buy-In Scenarios Eligibility Current Base restricted to National Coordinated No ESI Modernized Medicare Enhanced law buy-in ages 55–64 premium PTCs firewall benefits premiums subsidies Total nonelderly population Household 545,876 -1,761 -2,359 -1,088 -1,658 -3,355 -1,715 -14,258 -7,997 Premiums 272,849 1,923 739 2,495 2,032 654 2,273 -8,941 -2,439 Other health care spending 273,027 -3,685 -3,098 -3,583 -3,690 -4,008 -3,989 -5,317 -5,557 Federal government 435,130 637 829 603 599 3,311 305 10,054 15,311 Medicare buy-in — 3,996 2,471 3,976 4,040 6,899 5,152 13,342 4,653 Other health care spending 435,130 -3,359 -1,642 -3,373 -3,441 -3,588 -4,846 -3,289 10,658 State government 198,624 -162 -63 -154 -174 -12 -178 -206 -2,077 Other Employers’ premium contributions 720,926 -375 -346 -364 -377 -3,047 -421 -785 -5,977 Providers’ uncompensated care 22,975 -172 -38 -174 -186 -190 -190 -213 -2,283 Total, all payers 1,923,531 -1,833 -1,977 -1,176 -1,796 -3,293 -2,199 -5,408 -3,022 Ages 50–64 Household 242,713 -2,175 -2,805 -1,489 -2,069 -3,717 -2,129 -14,757 -5,697 Premiums 120,214 1,395 178 1,978 1,505 176 1,746 -9,555 -1,143 Other health care spending 122,499 -3,570 -2,983 -3,467 -3,575 -3,894 -3,875 -5,202 -4,554 Federal government 182,152 96 143 32 67 2,896 -220 9,341 6,889 Medicare buy-in — 3,996 2,471 3,976 4,040 6,899 5,152 13,342 4,653 Other health care spending 182,152 -3,901 -2,329 -3,944 -3,972 -4,003 -5,371 -4,001 2,236 State government 79,125 -221 -106 -222 -233 -72 -246 -288 -1,141 Other 26 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS Alternative Medicare Buy-In Scenarios Eligibility Current Base restricted to National Coordinated No ESI Modernized Medicare Enhanced law buy-in ages 55–64 premium PTCs firewall benefits premiums subsidies Employers’ premium contributions 295,475 -270 -245 -260 -272 -2,943 -317 -682 -4,385 Providers’ uncompensated care 11,657 -206 -71 -207 -220 -223 -223 -246 -1,191 Total, all payers 811,123 -2,775 -3,085 -2,147 -2,727 -4,060 -3,134 -6,632 -5,526 Under 50 Household 303,163 414 446 401 411 363 413 499 -2,300 Premiums 152,635 529 561 516 526 477 527 615 -1,297 Other health care spending 150,528 -115 -115 -115 -115 -115 -114 -115 -1,003 Federal government 252,978 541 687 571 532 416 525 712 8,423 Medicare buy-in — — — — — — — — — Other health care spending 252,978 541 687 571 532 416 525 712 8,423 State government 119,498 59 43 69 59 60 67 82 -936 Other Employers’ premium contributions 425,451 -105 -101 -103 -105 -105 -104 -103 -1,591 Providers’ uncompensated care 11,318 33 33 33 33 33 33 33 -1,092 Total, all payers 1,112,408 942 1,108 971 931 766 935 1,224 2,503 Source: Urban Institute Health Insurance Policy Simulation Model, 2020. Notes: PTCs = premium tax credits. ESI = employer-sponsored insurance. A dash indicates the column heading does not apply. All other federal government spending is Medicaid, Marketplace premium tax credits and cost-sharing reductions, and uncompensated care. State government spending is Medicaid, Marketplace premium tax credits and cost-sharing reductions, and uncompensated care. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 27 Restrict Buy-In Eligibility to Those Ages 55 to 64 In this alternative scenario, eligibility for the buy-in program would be limited to those ages 55 to 64, rather than 50 to 64. With the narrower age range for eligibility, we find correspondingly smaller buy-in enrollment of 1.5 million people (compared with 2.1 million in the base scenario). As in the base scenario, enrollment in the Marketplace and other nongroup policies would decrease compared with current law (1.4 million fewer enrollees). The reduction would be almost entirely among those in the age range eligible for the buy-in. Across the total nonelderly population, the number of those with insurance coverage would slightly decrease, but we find coverage would increase among those ages 50 to 64 (95,000 more enrollees). Narrowing buy-in eligibility would result in an overall reduction in health care spending of about $2.0 billion relative to current law. Households’ spending on premium payments would increase (by $739 million), but other household spending for health care would fall by about $3.1 billion. The difference in premium payments relative to the base scenario suggests much of the premium increase in the base scenario would be concentrated among those ages 50 to 54. Spending by the federal government for premium tax credits and cost-sharing reductions for buy-in enrollees would be about $2.5 billion; this is about 62 percent of the cost to the federal government for the base buy-in scenario, consistent with smaller buy-in enrollment. The federal government would also save less in spending on premium tax credits from the nongroup market than in the base scenario. Fewer people going into the buy-in plan, with lower payment rates, would result in higher net federal spending in this scenario ($829 billion) than in the base scenario ($637 billion). Restrict Geographic Variation in Premium Prices for the Buy-In Plan In this scenario, the buy-in premium would not vary across states; instead, the program would have one national premium. We find this would result in a total increase in insurance coverage of 66,000 people, very similar to the increase in the base scenario. Other changes in insurance coverage type relative to current law would also be similar to those under the base buy-in scenario. Because this alternative scenario would eliminate geographic variation in the premium, its differences from the base scenario are more pronounced at the state level. For example, Arizona, Colorado, and Iowa would all see buy-in premiums fall about 9 percent and have the largest relative increases in buy-in enrollment, about 1 percent (data not shown). Arkansas and New Mexico would see 28 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS premium increases of 15 percent and 12 percent and enrollment declines of a little less than 1 percent (data not shown). Relative to the base buy-in scenario, eliminating geographic variation in the buy-in premium would make little difference to changes in federal spending, state spending, and spending by employers and providers. Household premium spending would be slightly higher (an additional $2.5 billion versus $1.9 billion in the base buy-in), however, suggesting some increases in premium spending are mitigated when premiums are adjusted to reflect geographic variation in prices. Coordinate Premium Tax Credits for Households with Members Above and Below Age 50 In the scenario that coordinates premium tax credits across the buy-in and Marketplace coverage (relevant for households with members above and below age 50), we find that overall enrollment in the buy-in would be higher than in the base scenario by about 25,000 enrollees (still rounding to 2.1 million total enrollment). Marketplace enrollment would decrease by 1.9 million on net, nearly all of which would occur among those ages 50 to 64. Overall, we estimate 83,000 additional people would gain insurance coverage relative to current law in this scenario, compared with 67,000 in the base policy. The reason coverage effects of this scenario are very similar to those in the base scenario is that only 2.5 percent of current Marketplace enrollees with premium tax credits (249,000 people) are in mixed-age families with incomes above 200 percent of FPL, and only a minority of these would enroll in the buy-in. As with changes in coverage, overall changes in spending under this scenario would be very similar to those under the base buy-in scenario. Remove the ESI Firewall In the scenario that extends buy-in eligibility to those with offers of affordable ESI coverage, we find 2.5 million people would enroll in the buy-in, about 400,000 more enrollees than under the base buy-in scenario. Compared with current law, 445,000 fewer nonelderly people would have employer coverage, most of whom would be ages 50 to 64 (377,000). This policy would reduce the number of uninsured people by 77,000, only a little more than the base policy. Removing the ESI firewall is associated with a nearly $4.1 billion decrease in total health care spending among those ages 50 to 64 and a $3.3 billion decrease in total spending overall. Thus, this policy’s savings to the overall health care system would be larger than those in the base policy. Notably, MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 29 decreased ESI enrollment would lead to a $3.0 billion reduction in employer premium contributions and a $3.4 billion reduction in household spending as households shift to buy-in coverage with lower provider payment rates. We estimate federal spending for the buy-in would total $6.9 billion under this scenario (about $3.0 billion more than the base scenario), whereas the offset on other federal spending ($3.6 billion) would be similar to that in the base policy. Consequently, total federal spending would increase by $3.3 billion, but the policy would achieve overall net savings. Modernize Medicare Benefits In this scenario, the buy-in plan would be enhanced to include an out-of-pocket maximum and dental, vision, and hearing benefits while maintaining an 85 percent actuarial value. Adding these benefits makes the Medicare buy-in plan superior to gold and silver Marketplace plans. This alternative scenario would have a modest effect on coverage, attracting 2.4 million enrollees into the buy-in plan (compared with 2.1 million in the base scenario). The increase in buy-in enrollment relative to the base policy is modest because the plan’s overall value is similar (the effect of the difference in risk profile is relatively small), and those who would benefit most from the buy-in would already have taken it under the base policy. Among those ages 50 to 64, this scenario would result in 2.2 million people leaving Marketplace or other nongroup coverage, versus 1.9 million under the base scenario. Close to 200,000 people would gain coverage. The modernized benefits could be particularly attractive to people with higher health care costs, who might prefer an out-of-pocket maximum. Offering a modernized Medicare benefits package with the same actuarial value is associated with a $2.2 billion total decrease in spending relative to current law. The net impact on household premiums would be similar to that in the base scenario (a $1.7 billion decrease), though household premium spending would be slightly higher, offset by a greater reduction in other household spending. Federal spending on the buy-in would total $5.2 billion, and other federal spending would decrease by $4.8 billion. Charge the Medicare Premium We also assess the impact of introducing a buy-in plan that would reduce household premiums to levels comparable with those for traditional Medicare (retaining the same 85 percent actuarial value). Whereas President-elect Biden has proposed lowering the age for Medicare eligibility to 60, the policy we model would operate as a separate risk pool from Medicare and be available to those ages 50 to 64, alongside ACA Marketplace coverage. 30 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS We find this policy would lead to a larger increase in buy-in enrollment relative to current law, with 2.9 million people enrolling in the buy-in plan (about 800,000 more enrollees than the base scenario). Using Medicare premiums would result in 2.6 million fewer people being enrolled in Marketplace plans. Though the overall change in insurance coverage is comparable with that under the base scenario, we find this approach would increase coverage among those ages 50 to 64 (40,000 more people than the base scenario), but a higher number of people younger than 50 would lose insurance coverage (about 68,000 additional uninsured people). This policy would require more federal investment in the buy-in ($13.3 billion), nearly $10 billion more than the base scenario. However, reductions in household spending on premiums ($8.9 billion) and out-of-pocket spending ($5.3 billion) would offset the increased federal spending and contribute to an overall reduction in health care spending of $5.4 billion compared with current law. Enhance Premium Subsidies The enhanced subsidies scenario would expand premium subsidies for those with incomes between 100 percent and 400 percent of FPL and extend fully subsidized premiums to those with incomes below 100 percent of FPL, a group lacking access to premium tax credits for Marketplace plans under current law. This alternative is a broader policy change than the other alternative buy-in scenarios because it would also bolster the premium tax credit subsidies and cost-sharing reductions available for Marketplace plans. Though enhanced subsidies could be implemented separately from the introduction of the buy-in plan, they could also work in tandem with the buy-in plan to boost overall coverage and mitigate any adverse premium effects on those under age 50. Overall, the enhanced subsidies policy would reduce the number of uninsured people by nearly 3 million. It would result in a similar number of people ages 50 to 64 enrolling in the buy-in plan (2.2 million), but with a much smaller reduction in enrollment in Marketplace or other nongroup plans. Thus, we find a larger net increase in insurance coverage in this scenario of 700,000 people ages 50 to 64. In this scenario, we also find a large increase in Marketplace enrollment for those under age 50; 2.7 million individuals would gain coverage. Thus, we find an even larger decrease in the number of uninsured people younger than 50 (2.3 million fewer uninsured), likely driven by availability of premium tax credits for those with incomes below 100 percent of FPL. This scenario would also lead to a large reduction in employer-based coverage (1 million fewer enrollees altogether), likely arising from people newly eligible for premium tax credits for Marketplace or buy-in plans switching from unaffordable employer-based coverage. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 31 This policy scenario would increase federal spending by $15.3 billion: $4.7 billion on the buy-in and $10.7 billion in additional Marketplace premium subsidies. Household spending on premiums and out- of-pocket expenses would decrease by $8 billion, and we find much larger decreases in state government spending, employer premium contributions, and uncompensated care spending than in other alternatives examined. Overall, we find the enhanced subsidies scenario is associated with a $3 billion decrease in total health care spending. Discussion In this paper, we examine how a Medicare buy-in option for adults ages 50 to 64 could be used to expand the number of insurance options available for people in that age group and possibly reduce the number of uninsured people and household, employer, and federal spending. The basic Medicare buy-in we analyze would offer a new health insurance choice to people ages 50 to 64, but they could also choose to stay with a Marketplace plan. We modeled this plan as one product with an actuarial value similar to existing traditional Medicare. We assume Medicare payment rates would be applied and administrative costs would be close to those of the existing program, but, like traditional Medicare, the product would have no cap on out-of-pocket costs. Without modification, a Medicare buy-in product based on traditional Medicare would offer the choice of enrolling in Parts A, B, or D, but we assume the buy-in option would likely morph into a single product combining Parts A, B, and D. A Medicare buy-in policy option was considered after the broader Clinton health reforms failed in the 1990s, but it seems to be a less critical policy option since implementation of the ACA. This is because the ACA has already considerably reduced uninsurance in this age group. We estimate 5.6 million adults ages 50 to 64 are currently uninsured. Making a product very similar to Medicare available to this age group and allowing them to buy into it using ACA subsidies would only reduce the number of uninsured people in this age group by 176,000, or 0.3 percent, in our base buy-in scenario. However, because of older adults enrolling in the buy-in, premiums would increase by a modest amount for younger individuals, 110,000 of whom would drop coverage. Thus, the basic Medicare buy-in would result in a net coverage gain of only 67,000 people. Of the 2.1 million people we estimate to take up the buy-in, most would already have had Marketplace coverage and would just be switching their coverage type. These people would account for 1.9 million of the Medicare buy-in enrollees. Those with incomes below 400 percent of FPL who are eligible for premium subsidies would face no additional cost from choosing the buy-in; they would be obtaining a product that provides better coverage (85 percent actuarial value from the buy-in versus, 32 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS say, 70 percent actuarial value for a Marketplace silver plan), even without a cap on out-of-pocket spending. People ineligible for subsidies would be able to buy a better product and benefit from the lower premiums resulting from the buy-in’s lower provider payment rates. Because of existing ACA subsidies and competition among Marketplace plans that has kept premiums low, the net cost to the federal government of offering the buy-in would be relatively small: $637 million in 2020, or only 0.15 percent of current federal spending for health care for the nonelderly population. The former reduces the marginal cost of the buy-in. The latter has brought down provider payment rates and thus premiums to levels consistent with Medicare payment rates in many markets, particularly in well-populated urban areas. The base Medicare buy-in would reduce household spending by $1.8 billion, or 0.32 percent, and health care spending by $1.8 billion, or 0.01 percent. Employer spending would also fall slightly. Though these results are small in aggregate, their effects on households are more significant. People choosing to stay in the Marketplace would be largely unaffected by the introduction of a Medicare buy-in plan. But those who would switch to the buy-in would be considerably better off: Their out-of-pocket health care spending would fall from $3,414 to $1,528 because of the higher actuarial value of the buy-in, and their out-of-pocket premiums would fall from $6,430 to $6,230. Those who switch would also pay less out of pocket and consume more health care, and at a lower cost than those who remain in the Marketplace. We also analyze seven alternative designs of the basic Medicare buy-in policy (table 1). Each modification would be implemented in isolation, meaning results (tables 7 and 8) are not cumulative. Limiting the age of eligibility, as in the first alternative, would have virtually no impact on overall coverage. Among the various other alternatives, only one—offering higher subsidies than those currently available through the ACA Marketplaces—would result in substantial reductions in the number of uninsured (almost 3 million). The other options would expand coverage relative to the base buy-in policy somewhat, but their overall coverage gains never exceeded 100,000; they reduce uninsurance more for people ages 50 to 64, but some people under 50 would lose coverage as their premiums increase. Put simply, the Medicare buy-in approach can increase coverage to some degree but is unlikely to significantly expand coverage without additional financial support. However, under some policy alternatives, some people switching from Marketplace or employer coverage to the Medicare buy-in would be better off because of the plan’s lower premiums and more generous coverage in terms of actuarial value. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 33 The alternative that would set buy-in premiums uniformly across all geographic areas but at self- financing levels would produce coverage gains comparable to those under the base scenario. When premiums are set nationally at self-financing levels, about 175,000 people ages 50 to 64 would gain coverage, but 109,000 people under age 50 would lose coverage. Federal costs would be slightly lower. Modernizing the Medicare benefit package to include an out-of-pocket cap and dental, vision, and hearing services would benefit many households, particularly those with high medical expenses. This option would increase buy-in enrollment by an estimated 300,000 people relative to the base scenario, and 70,000 people would gain coverage. Most of the new buy-in enrollees would be people moving from Marketplace coverage. In the base case, most who would retain Marketplace coverage would do so because of the low-income subsidies they already had and the out-of-pocket limit. The modernization would change this calculus by extending the out-of-pocket limit to the Medicare buy-in. Federal spending for the buy-in would increase, but spending on Marketplace coverage would decrease, resulting in little net change. Employer spending would be virtually unchanged. Overall health care spending would increase by $2.2 billion, compared with $1.8 billion in the base case, because more people would be in the buy-in plan, which would have higher actuarial value. The reduction in provider payment rates would temper the increase in overall spending. Eliminating the ESI firewall would allow workers to leave their employer plan if they find the Medicare buy-in an attractive alternative. Eliminating the ESI firewall would increase enrollment in the Medicare buy-in to 2.5 million, from 2.1 million in the base case. But, it would only add about 77,000 people to the insured population. The large increase in Medicare buy-in enrollment would be offset by reductions in ESI and other nongroup coverage. About 2 million of the buy-in enrollees would be switching from the Marketplace, and another 445,000 would be leaving their employer plans. Eliminating the firewall would reduce household expenses by $3.4 billion as individuals move to less costly buy-in coverage. The cost to the federal government would increase by $3.3 billion, about five times the $637 million they would spend in the base case. Employers would also see a substantial $3.0 billion in savings, compared with $375 million in the base case, as their enrollees leave to join the Medicare buy-in. In the base Medicare buy-in policy and all the alternatives scenarios, a family with one person eligible for the buy-in and one or more people who are not (e.g., because they are younger) would need to meet the full percentage-of-income requirement separately for the buy-in-eligible person and other people. This is the most literal interpretation of the existing proposals that do not address this issue. This would impose extra costs on the family beyond those envisioned in ACA. Therefore, we model an option that coordinates the calculation of premium tax credits so that a family, as a group, would only 34 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS have to pay a percentage of their income up to a single ACA-specified limit. In principle, this would make the Medicare buy-in more attractive to people in these families. But, compared with the base buy-in scenario, it would only lead to an increase in buy-in enrollment of 25,000 people and a net increase in overall coverage of about 16,000 people. The effects would be small, presumably because the younger groups are not in the nongroup market (e.g., have ESI). This approach would have little effect on federal spending. Reducing premiums to current Medicare levels would increase buy-in participation by around 800,000 people beyond the base scenario, almost all because of people switching from the Marketplace. This alternative would have no meaningful effect on coverage. The federal government would spend about $10 billion more than under the base scenario because current Parts B and D premiums are heavily subsidized. This additional federal spending would translate into lower spending among households ($14.3 billion), because Medicare premiums would be lower than the Marketplace premiums for most people. We only observe truly substantial gains in coverage when we consider enhanced subsidies that would be available to purchase either Marketplace or buy-in coverage. This alternative would reduce premiums significantly for both people in the Marketplace and those choosing the buy-in. Relative to the base case, the number of people enrolling in the buy-in would increase by 150,000, and the number enrolling in Marketplace plans would increase by 3.7 million. About 3.0 million people would gain coverage under this policy. Household spending would fall by $8.0 billion, compared with $1.8 billion in the base scenario. Of course, the higher subsidies would drive up federal spending from $637 million in the base buy-in to $15.3 billion. Employers would save $6.0 billion because of lower ESI enrollment, because fewer workers would need such coverage to remain insured. Overall health care spending would fall by $3.0 billion, as opposed to $1.8 billion in the base scenario, because so many more individuals would enroll in either the buy-in or the Marketplace, leaving more costly ESI. An unfortunate and largely unexplored effect of a Medicare buy-in policy, not fully addressed in our modeling, is how it could affect insurers’ willingness to participate in the Marketplaces. Under current law, the ACA permits three-to-one variation in premiums across age groups in the nongroup market. A Medicare buy-in would reduce the share of individuals ages 50 to 64 with Marketplace coverage. This could be a serious issue, because evidence shows older Americans’ claims costs are lower than the premiums insurers set, whereas the reverse is true for younger populations (Kotecki and Westrom 2020). So, contrary to initial expectations, older Marketplace enrollees are accounting for premiums that cross-subsidize younger enrollees’ premiums. A buy-in would take many individuals now providing cross-subsidies to younger populations out of the Marketplace. This would result in premiums for those MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 35 under age 50 increasing slightly (roughly 4 percent), thereby increasing federal subsidies. But another important result is that the Marketplaces would be much smaller. Individuals ages 50 to 64 account for 38 percent of Marketplace enrollment. Marketplaces would simply not be as attractive to many insurers as they are today. This could lead insurers to exit many Marketplaces, which drives up premiums and, in turn, could result in many more people ineligible for subsidies choosing the buy-in product. A Medicare buy-in would not eliminate the administrative complexity of Medicare if it uses the Parts A, B, and D structure. For modeling purposes, we use a base scenario that reflects current Medicare benefits and cost sharing but is abstracted from the three distinct parts of the program. Indeed, a Medicare buy-in plan might consider offering a choice of traditional Medicare or an MA plan, or only offer an MA plan. A major attraction of a Medicare buy-in to many of its proponents is that it would extend Medicare’s lower provider payment rates to a larger share of physician, hospital, and other services. This may not be essential in many urban Marketplaces that currently have payment rates similar to those in Medicare, because of the large number of competing insurers and the emergence of Medicaid plans in the Marketplace (Blumberg, Holahan, Wengle, et al. 2019; Holahan et al. 2019). Insurers aggressively compete for market share in these areas by keeping provider rates low, because enrollees can be price sensitive and the ACA bases tax subsidies on the second-lowest-cost plan in each market. On the other hand, the impact of the Medicare buy-in would be to apply Medicare rates to the remaining markets. Many of these markets have few insurers and relatively few providers. But policymakers may need to be cautious, because extending Medicare payment rates to buy-in enrollees could result in access problems; lower rates would lower provider revenues. How rates should be set in any expansion of the Medicare program is a serious issue, because reducing payment rates to Medicare levels could have serious adverse consequences for some providers. Some adjustments to payment rates may be necessary (e.g., setting rates some percentage above Medicare for certain specialties or types of hospitals). But this would increase the buy-in’s costs. Further, if payment adjustments are made for the buy-in population, providers would likely seek those same rates for existing Medicare enrollees. An enhanced subsidy schedule and eliminating the ESI firewall would offer significant savings to households and employers and yield the greatest buy-in enrollment. The modernized benefit policy would dramatically improve benefits but increase government spending, because the cost of added benefits would offset much of the savings from increased enrollment in an option with lower payment rates. If these three variations to the base policy were packaged together, the cumulative effects would be to increase the number of people enrolling in the buy-in and reduce household and employer 36 MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS spending, but with some increase in government expenditures. Considering an enhanced Medicare product combining these features was beyond the scope of this paper. We also did not examine allowing employers to directly buy in (Docteur et al. 2020). Allowing an employer buy-in would provide employers with a lower-cost insurance option, holding actuarial value constant, than most have today. An employer buy-in would be a major expansion compared with what we have modelled. It could be offered by an employer as a choice for employees, or employers could offer it as the only plan for workers ages 50 and over. Allowing employers to buy in could result in more employers offering coverage and more employees taking up coverage and thus could have a greater impact than what we show here. Not all employers would find this an attractive choice. Some would be satisfied with offering a product with lower actuarial value. Some employers would find the product unattractive because it lacks an out-of-pocket limit and would fear the impacts on their workers. Other employers may believe workers are better off with a product that would pay higher provider payment rates. A likely outcome would be that small employers with low-wage workers would be the most likely to find the buy-in attractive. The main overall take-away from our analysis is that the potential for a Medicare buy-in policy to substantially expand health insurance coverage is fairly limited given what the ACA already provides. Buy-in enrollment does not exceed 3 million in any of our scenarios. What the buy-in policy primarily does is take advantage of lower provider payment rates. Increases in federal spending under more expansive policy scenarios can shift dollars from private payers to public payers in a way that is roughly neutral for aggregate spending. Buy-in policies could result in savings to national health spending overall because of the lower payment rates. Some savings from more expansive options would also likely result from shifts from ESI to Marketplace coverage, which has lower payment rates than typical commercial insurance. MEDICARE BUY-IN POLICIES’ EFFECTS ON HEALTH COVERAGE AND COSTS 37 Appendix A. Methodology We produced the estimates for this report using the Health Insurance Policy Simulation Model (HIPSM), a detailed microsimulation model of the health insurance system designed to estimate the cost and coverage effects of proposed health care policy options. HIPSM is based on two years of data from the 2012–13 American Community Survey, which provides a representative sample of families constituting more than 6 million individuals of all ages. The sample is reweighted to reflect more recent information on income and demographics and aged to future years using recent American Community Survey data and projections from the Urban Institute’s Mapping America’s Futures program. In this analysis, we use the 2020 version of the model, and our findings reflect the enrollment and spending effects if the buy-in policy were fully implemented in 2020 (Buettgens and Banthin 2020). Modeling Insurance Coverage Choices In our simulation of current law, individuals and families face choices between several types of health coverage: ESI, Medicaid, the Children’s Health Insurance Program, Basic Health Programs, the ACA- compliant nongroup market, and the non-ACA-compliant short-term, limited-duration market. We simulate people’s choices between these options using an expected-utility framework; people choose the coverage option that maximizes their expected utility, which depends on their income, premiums, expected out-of-pocket costs, the variance in out-of-pocket costs across plans, and total health spending. The model can replicate real-world enrollment patterns by generating latent preference terms for the health coverage options available to a family. Latent preference terms represent factors that influence people’s coverage decisions that are not explicitly measured in the expected-utility function. An example is the breadth of a provider network; it is not measured explicitly but may influence people’s coverage decisions. Introducing the Medicare buy-in plan, a new insurance coverage option, would require people to choose between the new option and the existing private options (the current nongroup market; short- term, limited-duration plans; and ESI, if they have an offer). One of the most important factors in their decisions is the out-of-pocket premium. To compute premium tax credits for the Medicare buy-in option, we apply the ACA’s applicable percentage-of-income limit to the Medicare buy-in premium. People eligible for the buy-in would pay the same premiums as those for Marketplace silver plans. This makes the choice between the two options mostly about differences in plan benefits and actuarial value, rather than premiums. The new option is designed to be similar to ACA-compliant nongroup coverage, 38 APPENDIX so we do not assume any latent preferences would cause people to inherently prefer the Medicare buy- in option over nongroup coverage or vice versa (aside from the financial terms included in the expected- utility function). However, it is possible that such preferences may exist in the real world because of factors like the broad provider networks in the Medicare FFS program and the popularity of the name “Medicare.” The age restrictions on Medicare buy-in eligibility complicate family decisionmaking. In many families, some adults would be eligible for the Medicare buy-in plan, but other family members, such as younger spouses or children, would not. This is particularly important for the handling of premium tax credits. In our base buy-in scenario, we assume the Marketplaces compute tax credits for family members who choose to enroll in Marketplace coverage in the same way they do now. Family members choosing the Medicare buy-in, however, would have their tax credits computed separately by that program, meaning the family may end up paying more in premiums than if everyone in the family were enrolled in a single program. Specifically, the family could pay up to the applicable percentage of income to cover its Marketplace enrollees and an additional premium up to the applicable percentage of income to cover its Medicare buy-in enrollees, which could effectively double premiums. 11 In one alternative scenario, we show what would happen if a single cap were placed on the total sum of Marketplace and buy-in premiums at the applicable percentage of income. Provider Payment Rates and Premiums under the Medicare Buy-In Policy An important feature of the Medicare buy-in policy is the plan’s ability to reimburse providers at Medicare payment rates. Consequently, Medicare buy-in plan premiums would be lower than if plans had to pay providers at higher commercial rates or had to negotiate rates in each market. Medicare buy- in premiums would also be lower because of lower administrative costs. We assume the administrative loading factor for Medicare buy-in plans would be 6 percent, less than half the 15 administrative loading factor we assume for private plans in the nongroup market. Computing premiums for the Medicare buy-in plans takes two steps. The version of HIPSM used for this analysis is calibrated to replicate 2020 Marketplace premiums at the bronze, silver, and gold tiers in each premium rating region. Underlying these premiums is the assumption that they reflect provider payment rates paid by insurers for Marketplace enrollees in each rating region. We then estimate out- of-pocket spending and insured costs compatible with these premiums for all potential Medicare buy-in APPENDIX 39 enrollees. In the first step, we start with health costs in the gold plan, which has an 80 percent actuarial value and is closest to the actuarial value of the Medicare buy-in plan. We adjust the costs to reflect the base buy-in deductible and coinsurance, along with a moral-hazard effect. At the end of this step, we have estimated Medicare buy-in premiums at Marketplace rates, computed based on the resulting insured costs for simulated enrollees, plus a 6 percent administrative load. Unlike Marketplace premiums, buy-in premiums would not vary by age. In the second step, we calculate new, lower premiums to reflect moving from Marketplace to Medicare provider reimbursement levels. We estimate the impact of Medicare payment rates on the buy-in plan premiums using methods developed in previous research. 12 Our approach assumes prices paid to providers by insurers in highly competitive nongroup markets are approximately the same as Medicare payment rates. We take this approach because there are no nationally or state-representative sources of claims data for private nongroup insurers, and this lack of information means we cannot easily make a direct comparison of commercial and Medicare payment rates. We therefore proxy the geographic variation in nongroup provider payment rates using 2017 Marketplace premium data at the rating-region (substate) level. 13 We estimate reducing payment rates for physicians and hospitals to Medicare levels (all else equal) would decrease medical expenses by approximately 12 percent, on average, nationwide. Under this approach, we do not estimate any savings in prescription drug expenses. The differences in medical costs and resulting premiums would vary, however, by state and rating region. We compute the percent difference between a rating region’s 2020 premium and our proxy premium. We then apply the computed percent difference to a rating region’s actual benchmark premium to calculate the premium for the Medicare buy-in plan in 2020. Using this method, premium differences between Medicare buy-in and Marketplace plans would be smaller in more competitive markets and larger in less competitive ones. 40 APPENDIX Notes 1 John M. Broder, “Clinton Proposes Opening Medicare to Those 55 to 65,” New York Times, January 7, 1998, https://www.nytimes.com/1998/01/07/us/clinton-proposes-opening-medicare-to-those-55-to-65.html. 2 Selena Simmons-Duffin, “Comparing Biden’s and Trump’s Different Visions of Health Care,” NPR, October 22, 2020, https://www.npr.org/sections/health-shots/2020/10/22/926372475/comparing-bidens-and-trump-s- different-visions-for-health-care. 3 A Milliman analysis indicates the actuarial value of traditional Medicare is 83.3 percent, and the actuarial value of Medicare Advantage is 89.3 percent (Mike, Friedman, and Yilmaz 2019). 4 These premiums include surcharges for earned income. However, people opting into these buy-in plans are no longer eligible for premium tax credits available for Marketplace plans (which were also available to buy-in enrollees in all other buy-in scenarios). Instead, the federal government pays an implicit subsidy for the difference in health care spending and premium prices. Thus, individuals ineligible for premium tax credits would still benefit from this alternative if the premium amount is lower than their expected health care spending. 5 Throughout this report, “uninsured” refers to people who either lack insurance coverage or have insurance coverage that does not comply with ACA-established standards (e.g., short-term, limited-duration plans). 6 In Medicaid expansion states, those with incomes below 138 percent of FPL who qualify for Medicaid coverage do not qualify for premium tax credits for Marketplace plans. In nonexpansion states, those with incomes between 100 percent and 138 percent of FPL qualify for premium tax credits for Marketplace plans but have little to no eligibility for Medicaid coverage. Those with incomes below the FPL in nonexpansion states generally have neither Medicaid nor subsidized Marketplace coverage options. For these reasons, we expect (and find) a greater share of people in this low-income group who enroll in the buy-in plan and disenroll from Marketplace plans reside in nonexpansion states (data not shown). 7 “Marketplace Plan Selections by Metal Level,” Henry J. Kaiser Family Foundation, accessed November 18, 2020, https://www.kff.org/health-reform/state-indicator/marketplace-plan-selections-by-metal-level- 2/?currentTimeframe=1&sortModel=%7B%22colId%22:%22Location%22,%22sort%22:%22asc%22%7D. 8 Note the transitions reported in table 4 do not reflect overall transitions between Marketplace and buy-in plans, because this analysis is restricted to those with single coverage plans only. Including those with family coverage would complicate comparing premiums. 9 Figure 3 shows average spending for people ages 50 to 64 is roughly $10,000 per capita under standard payment rates. A plan with actuarial value of 69 percent covers about $6,900 of benefits for the average person. A plan with an actuarial value of 85 covers about $8,500 of benefits. The increase in actuarial value is worth roughly $1,700 for the average person ages 50 to 64 and more for a person ages 60 to 64. This value does not include the insurance protection, or the value of reduced risk. 10 By model construction, people who do not switch plans when the buy-in plan is introduced accrue the same total health spending when they stay in the same plan. 11 This is similar to the situation under current law; premiums for family members covered by the Children’s Health Insurance Program or Medicaid are not counted when Marketplace premium tax credits are computed. Children’s Health Insurance Program and Medicaid premiums are much lower, however, than Marketplace and Medicare buy-in premiums. 12 See the appendix in Blumberg, Holahan, McMorrow, et al. (2020). 13 See the appendix in Blumberg, Holahan, McMorrow, et al. (2020). NOTES 41 References Blumberg, Linda J., John Holahan, Matthew Buettgens, Anuj Gangopadhyaya, Bowen Garrett, Adele Shartzer, Michael Simpson, Robin Wang, Melissa M. Favreault, and Diane Arnos. 2019. From Incremental to Comprehensive Health Reform: How Various Reform Options Compare on Coverage and Costs. Washington, DC: Urban Institute. Blumberg, Linda J., John Holahan, Stacey McMorrow, and Michael Simpson. 2020. Estimating the Impact of a Public Option or Capping Provider Payment Rates. Washington, DC: Urban Institute. Blumberg, Linda J., John Holahan, Erik Wengle, and Caroline Elmendorf. 2019. “Is There Potential for Public Plans to Reduce Premiums of Competing Insurers?” Washington, DC: Urban Institute. Buettgens, Matthew, and Jessica Banthin. 2020. The Health Insurance Policy Simulation Model for 2020: Current-Law Baseline and Methodology. Washington, DC: Urban Institute. Docteur, Elizabeth, Renée M. Landers, Bethany Cole, Marilyn Moon, and Cori Uccello. 2020. Examining Approaches to Expand Medicare Eligibility: Key Design Options and Implications. Washington, DC: National Academy of Social Insurance. Eibner, Christine, Raffaele Vardavas, Sarah A. Nowak, Jodi L. Liu, and Preethi Rao. 2019. Medicare for 50-to-64-Year- Olds: Assessing the Effects of Allowing Older Adults to Buy Into the Medicare Program. Santa Monica, CA: RAND Corporation. Holahan, John, Linda J. Blumberg, Erik Wengle, and Caroline Elmendorf. “What’s Behind 2018 and 2019 Marketplace Insurer Participation and Pricing Decisions?” Washington, DC: Urban Institute. Johnson, Richard W., Marilyn Moon, and Amy J. Davidoff. 2002. A Medicare Buy-In for the Near Elderly: Design Issues and Potential Effects on Coverage. Menlo Park, CA: Henry J. Kaiser Family Foundation. Kotecki, Lindsy, and Stan Westrom. 2020. “A Case Study on the Actuarial Implications of a Medicare Advantage Buy-In Option for Older Adults.” Seattle: Milliman. Mike, David, Julia Friedman, and Gokce Yilmaz. 2019. Average Annual Beneficiary Health Care Costs for Medicare Coverage Options. Seattle: Milliman. 42 REFERENCES About the Authors Bowen Garrett is an economist and senior fellow in the Health Policy Center at the Urban Institute. Jessica Banthin is a senior fellow in the Health Policy Center. Anuj Gangopadhyaya is a research associate in the Health Policy Center. Matthew Buettgens is a senior fellow in the Health Policy Center. Adele Shartzer is a research associate in the Health Policy Center. John Holahan is an Institute fellow in the Health Policy Center. Diane Arnos is a research assistant in the Health Policy Center. ABOUT THE AUTHORS 43 STATEMENT OF INDEPENDENCE The Urban Institute strives to meet the highest standards of integrity and quality in its research and analyses and in the evidence-based policy recommendations offered by its researchers and experts. We believe that operating consistent with the values of independence, rigor, and transparency is essential to maintaining those standards. As an organization, the Urban Institute does not take positions on issues, but it does empower and support its experts in sharing their own evidence-based views and policy recommendations that have been shaped by scholarship. Funders do not determine our research findings or the insights and recommendations of our experts. Urban scholars and experts are expected to be objective and follow the evidence wherever it may lead. 500 L’Enfant Plaza SW Washington, DC 20024 www.urban.org