Financial Effects of an Out-of-Pocket Cap in Traditional Medicare

This economic evaluation assesses the financial outcomes of an out-of-pocket spending cap in traditional Medicare for Parts A and B.

Key Points

Question

What are the projected financial impacts of introducing an out-of-pocket (OOP) spending cap in traditional Medicare (TM) for Parts A and B?

Findings

In this economic evaluation using principles of microsimulation analysis of projected Medicare enrollment and spending from 2026 to 2034, an OOP cap of $3000 annually was estimated to reduce beneficiary spending by $1568 per year, a cap of $5000 annually by $1138 per year, and a cap of $10 000 annually by $707 per year, while increasing annual Centers for Medicare & Medicaid Services expenditures by $96 billion, $70 billion, and $39 billion.

Meaning

These findings suggest that implementing an OOP cap in TM would offer substantial financial protection for beneficiaries but impose substantial federal costs; a $10 000 cap may best balance affordability with fiscal sustainability.

Abstract

Importance

Traditional Medicare (TM) exposes beneficiaries to substantial financial risk because, unlike Medicare Advantage (MA), there is no out-of-pocket (OOP) maximum for medical spending. Policymakers have proposed introducing an OOP cap, but its fiscal and behavioral effects are uncertain.

Objective

To estimate the financial impacts of introducing OOP spending caps of $3000, $5000, or $10 000 annually in TM on beneficiary costs, enrollment patterns, and Centers for Medicare & Medicaid Services expenditures between 2026 and 2034.

Design, Setting, and Participants

This economic evaluation using principles of microsimulation analysis projected enrollment, spending, and health care use outcomes under baseline (no OOP cap) and 3 alternative cap scenarios using publicly available national Medicare and insurance data from 2020 to 2024 with projections through 2034. The model incorporated elasticity-based behavioral responses, including beneficiary switching between TM and MA, changes in service use, and increases to MA benchmarks and spending. Sensitivity analyses varied model assumptions probabilistically.

Exposure

Introduction of annual OOP caps for Parts A and B in TM at $3000, $5000, or $10 000.

Main Outcomes and Measures

Projected changes in beneficiary enrollment by coverage type, mean annual beneficiary OOP and premium spending, and annual Medicare expenditures.

Results

Between 2026 and 2034, a $3000 OOP cap was projected to reduce mean annual beneficiary expenditures by $1568, compared with $1138 under a $5000 cap and $707 under a $10 000 cap. Corresponding increases in annual Medicare spending were estimated at $96 billion (90% simulation uncertainty interval [SUI], $80-$114 billion), $70 billion (90% SUI, $55-$85 billion), and $39 billion (90% SUI, $29-$50 billion), respectively. OOP caps, especially at lower thresholds, would substantially shift beneficiaries from MA to TM.

Conclusions and Relevance

Results of this study suggest that implementing an OOP spending cap in TM would provide meaningful financial protection for beneficiaries but impose substantial fiscal costs on Medicare.

Introduction

Over half of Medicare beneficiaries are now enrolled in Medicare Advantage (MA).¹ Beneficiaries are increasingly choosing MA plans because they offer supplemental benefits (including prescription drugs, vision, dental, and hearing) and cost sharing protections, all at premiums that are substantially lower than in traditional Medicare (TM).² Beneficiaries’ relative ease of one stop shopping for MA—choosing a single plan that covers all their benefits as opposed to piecing together Medigap and stand-alone prescription drug plans in TM – has also likely contributed to the shift toward MA. In parallel, employer groups have moved away from offering retiree coverage that supplements TM and shifted toward employer-sponsored MA plans.^3,4

Despite the growing popularity of MA, it has important limitations that affect access to care and the quality of coverage. For beneficiaries, these include network restrictions, frequent denials and prior authorization requirements, and supplemental benefits that often fall short of marketing claims.^5,6 From a fiscal perspective, MA also places a growing burden on the federal budget: Centers for Medicare & Medicaid Services (CMS) pays MA plans 20% more to care for a similar beneficiary in TM due to overpayments from excessive risk coding, favorable selection, generous quality bonus payments, and other payment distortions.^2,7,8,9,10

In this context, beneficiary advocates, policymakers, and policy analysts have increasingly discussed ways to even the playing field between TM and MA. One policy under consideration is to offer an out-of-pocket (OOP) spending cap in TM for Parts A and B.^11,12 OOP spending caps are nearly universal health insurance benefit features in both public and private coverage, serving as a critical financial protection against catastrophic medical costs.¹³ However, OOP caps have been noticeably absent from the TM program. While the Inflation Reduction Act recently introduced an OOP cap for prescription drug spending under Part D, there is no such protection in Parts A and B. The absence of an OOP cap in TM exposes beneficiaries, especially those without supplemental coverage, to substantial financial risk and represents a critical policy gap in Medicare’s structure. Conversely, MA plans are required to offer such an OOP cap on medical spending, creating a substantial advantage that draws beneficiaries away from TM. An OOP cap in TM can be seen as an effort to improve benefit comparability and preserve beneficiary choice, while highlighting concerns about the fiscal impacts of continued beneficiary migration to MA.

Although an OOP cap for Parts A and B has conceptual appeal, its financial impacts for beneficiaries and CMS are unknown. Under an OOP spending cap, Medicare would assume financial responsibility for all beneficiary spending related to Part A and B above the cap. This would lead to transfer of financial risk from beneficiaries to CMS, with potentially substantial implications for Medicare spending. At the same time, an OOP spending cap in TM could lead to a complex set of behavioral interactions among beneficiaries and health plans. This includes beneficiary switching between MA and TM and changes in cost sharing, health care use, and spending. We used a microsimulation model to estimate the impacts of an OOP cap for Parts A and B in TM in financial outcomes. We assess both the distributional impacts for beneficiaries and the aggregate fiscal implications for CMS.

Methods

This study used publicly available data and did not include any direct interaction with or intervention involving human participants. As a result, the research qualified for exemption from institutional review board (IRB) review. Because this study used the principles of microsimulation, no established reporting guideline was directly applicable.

We developed a microsimulation model to project the financial impacts of 3 potential annual OOP Part A and Part B spending caps ($3000, $5000, $10 000) for TM beneficiaries between 2026 and 2034. Our model simulated enrollment segments within the Medicare population over time and updated enrollment, spending, and use outcomes annually in response to policy-induced changes in premiums and cost sharing. Behavioral responses were incorporated through elasticity-based switching between coverage types and price responsiveness in health care use. The model therefore tracked how the initiation of OOP caps propagates through the simulated population and generated projected outcomes over time, which is consistent with standard microsimulation approaches used in health policy analysis.¹⁴

We modeled the impacts of OOP caps at varying levels of generosity relative to the current OOP cap in MA (around $5000 for in-network care and $9000 total).¹⁵ Estimates of spending and enrollment with no OOP cap (the baseline scenario) were based on projections between 2026 and 2034 (expressed in 2024 dollars) from the 2025 Medicare Trustees Report.¹⁶

We modeled enrollment and spending separately across key insurance segments within TM (no supplemental insurance, employer-sponsored supplemental insurance, Medigap-individually purchased supplemental insurance, and TM-dual eligibility with Medicaid) and MA (individual MA, employer-sponsored MA, and MA enrollees with dual eligibility) using data on the share of beneficiaries in each segment.⁴ Relative spending differences, cost sharing, and spending above various OOP caps across segments were based on the 2021 Medicare Current Beneficiary Survey and recent literature.¹⁷ Medigap premiums and loss ratios were calculated using data from 2020 to 2024 from the National Association of Insurance Commissioners, accessed through the Mark Farrah portal.¹⁸ These data were combined with Trustee estimates of Medicare spending growth to estimate future Medigap premiums.

Microsimulation Model Design

Our microsimulation model was constructed at the segment-year level (eg, Medigap, no supplemental insurance, individually purchased MA, in 2026). The model was implemented in several steps. First, the baseline scenario was established based on CMS actuary reports and Congressional Budget Office models. Second, an OOP cap in TM was assigned at 1 of 3 levels: $3000, $5000, and $10 000. The cap decreased Medigap premiums because Medigap plans would no longer be responsible for OOP spending above the cap (which would be covered by Medicare). Based on new Medigap premiums and beneficiaries’ cross price elasticity of demand, beneficiaries (and their employers) decided whether to switch from no supplemental coverage to TM plus Medigap; switch from MA to TM plus Medigap; or remain in their current coverage. No beneficiaries switched to MA nor did they drop Medigap and remain in TM. Consistent with evidence related to plan choice in other settings, the switching phased in over time as beneficiaries became aware of the OOP caps.^19,20,21 Differences in effective cost sharing resulting both from the OOP caps and from plan switching were translated into use changes through assumptions about the price elasticity of demand for health care. Changes to MA spending, resulting from higher benchmarks from increases in TM spending, were also incorporated (eMethods in Supplement 1). Outputs from the simulation were then captured, including paid Medigap premiums, OOP spending, Part B premiums paid by beneficiaries, Medicare expenditures, and Part B premiums received by CMS (Box).

Box. Conceptual Flow of Simulation Architecture and Implementation.

1. Baseline Medicare Environment

• Enrollment across coverage types

• Spending and use

• Cost sharing

• Medigap premiums

2. Policy Intervention: OOP Cap in Traditional Medicare

• Annual cap on Parts A and B OOP spending

• Alternate cap levels ($3000/$5000/$10 000)

3. Policy Assumptions

• Centers for Medicare & Medicaid assumes liability for spending above the OOP cap, reducing Medigap actuarial liability

• Medigap premiums recalculated

4. Beneficiary and Employer Coverage Choice

• Beneficiaries and employers respond to changes in Medigap premiums

• Choose between traditional Medicare and Medigap or Medicare Advantage

• Switching across coverage types phases in over time

5. Changes in Cost Sharing and Utilization Response

• Cost sharing changes due to traditional Medicare OOP cap and switching across coverage types

• Lower effective cost sharing increases Part A and Part B service use and spending

6. Financial Outcomes

• Beneficiaries: OOP spending; Medigap premiums paid; Part B premiums paid

• Medicare: OOP spending above the cap; Part A and B spending; Part B premiums received

To characterize uncertainty in model estimates related to Medicare expenditures and beneficiary savings, we conducted a probabilistic sensitivity analysis in which key behavioral parameters were varied simultaneously across plausible ranges (eTable 1 in Supplement 1). We repeated the simulation 10 000 times and calculated 90% simulation uncertainty intervals using the 5th and 95th percentiles of the simulated outcome distribution. We then used regression analysis to assess the association of specific parameters and Medicare expenditures. This analysis varied the price elasticity of demand for medical care, the own price elasticity of demand for Medigap, the cross-price elasticity of demand for MA with respect to Medigap premiums, the cost differential between TM and MA, and the extent to which higher TM spending would increase MA spending. More detailed information on the simulation and assumptions is described in the eMethods, eTable 1, eTable 2, eFigure 1, and eFigure 2 in Supplement 1. Simulation model and statistical analyses were performed using Stata version 18 (StataCorp LLC).

Results

Baseline Scenario

Under the baseline scenario with no OOP cap, the TM population was projected to remain virtually constant between 2026 and 2034 while the MA population was expected to grow by 9.4 million beneficiaries (Table).^4,15 During this time, mean annual medical spending was expected to increase from $17 034 to $27 833 for TM beneficiaries and from $16 701 to $27 399 for MA beneficiaries (eFigure 1 in Supplement 1).

Table. Descriptive Statistics.

Variable	Baseline (counterfactual) scenario, No.a
	2026	2034
TM beneficiaries	33 967 000	33 966 000
MA beneficiaries	37 134 000	46 554 000
Beneficiaries by segment
TM Medigap	14 266 140	14 265 720
TM ESI	10 529 770	10 529 460
TM dual	5 434 720	5 434 560
TM No supplemental	3 736 370	3 736 260
MA individual	21 909 060	27 466 860
MA ESI	6 684 120	8 379 720
MA dual	8 540 820	10 707 420
Annual Part A and B spending per TM beneficiary, mean, $	17 034	27 833
Annual Part A and B spending per MA beneficiary, mean, $	16 701	27 399
Medigap premiums, mean, $	3829	6257

Abbreviations: ESI, employer-sponsored insurance; MA, Medicare Advantage; TM, traditional Medicare.

^a Own price elasticity of demand for Medigap, -1.4; Cross-price elasticity of demand for Medigap and MA, 0.41; Own price elasticity of demand for medical care, -0.1. Estimates drawn from 2025 Medicare Trustees Report and relative segment enrollment.^4,15

Enrollment Effects of OOP Caps

Figure 1 shows the projected impacts for 3 alternative OOPs for TM and MA enrollment between 2026 and 2034. It shows that the caps would substantially affect beneficiary enrollment decisions, shifting beneficiaries from MA to TM. Under the $3000 cap, TM enrollment was projected to be greater than MA enrollment until 2034 when enrollment would be equal. Under the $5000 cap, MA and TM enrollment growth were projected to be similar until 2030, after which MA enrollment growth would exceed that of TM. Under the $10 000 cap, MA enrollment was projected to be greater than TM enrollment through 2034 (eFigure 3 in Supplement 1). Variation in TM and MA enrollment across the cap levels was driven by beneficiaries switching from having no supplemental coverage or MA to Medigap because of lower Medigap premiums (eFigure 2 in Supplement 1).

Figure 1. Line Graphs of Projected Enrollment in Traditional Medicare (TM) and Medicare Advantage (MA) Under Alternative Out-of-Pocket Caps, 2026 to 2034.

Impacts of OOP Caps in Beneficiary Expenditures

Figure 2 shows the estimated impacts of OOP caps for beneficiary expenditures. The $3000, $5000, and $10 000 OOP caps would reduce annual beneficiary expenditures by a mean of $1568, $1138, and $707, respectively, between 2026 and 2034. OOP spending reductions accounted for approximately 68% of gross beneficiary savings with Medigap premium reductions accounting for the remaining 32%. Beneficiary expenditure reductions in TM would increase substantially over time in line with Medicare spending increases. We estimated that the actuarial values of TM with the $3000, $5000, and $10 000 OOPs are 87.9%, 86.8%, and 85.6%, respectively, compared with an actuarial value of 84% of the standard TM benefit.²² Across beneficiaries in TM and MA, the $3000, $5000, and $10 000 OOP caps would increase Part B premiums by $164, $113, and $61, respectively, per member per year.

Figure 2. Line Graphs of OOP Caps on Annual Per-Beneficiary Spending, 2026 to 2034.

OOP indicates out-of-pocket; total is less than the sum of reductions in OOP spending and Medigap premiums because it incorporates projected increases in beneficiaries’ Part B premiums from increased Medicare spending.

Implications of OOP Caps for Federal Spending

The $3000, $5000, and $10 000 OOP caps were projected to increase CMS annual expenditures by $96 billion (90% simulation uncertainty interval [SUI], $80 to $114 billion), $70 billion (90% SUI, $55 to $85 billion), and $39 billion (90% SUI: $29 to $50 billion) between 2026 and 2034 (Figure 3). Approximately 72% of the increase in CMS expenditures from an OOP cap occurred in TM, while the remaining 28% occurred in MA (eFigure 4 in Supplement 1). Also, of the additional CMS expenditures for the $3000, $5000, and $10 000 OOP caps, 9.9%, 8.0%, and 3.6% occurred below the caps because of beneficiaries switching into supplemental coverage with lower cost sharing.

Figure 3. Bar Graphs of Estimates of Mean Annual Changes in Expenditures for CMS and Beneficiaries Under Alternative Out-of-Pocket Caps.

Simulation intervals constructed by randomly varying model parameters over 10 000 simulation iterations. Whiskers indicate 90% uncertainty intervals. CMS indicates Centers for Medicare & Medicaid.

Sensitivity Analysis

Results were most sensitive to assumptions regarding the own-price elasticity of demand for medical care and the extent to which TM spending increases were passed through to MA (eTable 3 in Supplement 1). Varying medical care elasticity across its plausible range changed projected annual CMS expenditures by $44 billion under a $3000 cap, $38 billion under a $5000 cap, and $27 billion under a $10 000 cap. Assumptions regarding the share of TM spending increases that would be passed through to higher MA spending had a large effect on estimates, changing projected annual CMS expenditures by $27 billion under a $3000 cap, $20 billion under a $5000 cap, and $12 billion under a $10 000 cap. In addition, varying assumptions about the cross-price elasticity of MA enrollment with respect to Medigap premiums and the TM to MA ratio changed projected annual CMS expenditures by $9 billion under a $3000 cap, $14 billion under a $5000 cap, and $17 billion under a $10 000 cap. In contrast, varying the own-price elasticity of Medigap demand changed projected CMS expenditures by $1 billion or less annually across all cap levels.

Discussion

Our microsimulation model on the projected impacts of OOP spending caps in TM between 2016-2034 yielded 3 key findings. First, OOP caps, particularly the more generous $3000 and $5000 caps, were projected to substantially shift beneficiaries from MA to TM. Second, beneficiary savings from TM OOP caps were projected to be considerable, ranging from a mean annual savings of $1568 for the $3000 OOP cap (3.6% of median per capita income among Medicare beneficiaries) to $707 for the $10 000 OOP cap (1.6% of median per capita income).²³ Savings would accrue to all TM beneficiaries (and their employers) who purchase supplemental coverage and also to those whose OOP spending exceeded the cap. For instance, approximately 11% of TM beneficiaries would be expected to hit a $5000 OOP cap in a given year. Third, OOP caps were projected to be expensive, leading to mean annual increases in CMS expenditures under the base scenario of $96 billion for the $3000 OOP cap and $39 billion for the $10 000 OOP cap. The projected implications of OOP spending caps on CMS expenditures were most sensitive to assumptions about elasticity of demand for medical care.

Our research updates and extends other estimates of the impact of OOP caps in Medicare. An Urban Institute report analyzed the impact of introducing a $5000 cap on Parts A, B, and D spending,¹² finding that it would increase CMS expenditures by $39 billion in 2023. Although the Urban Institute analysis differs from ours because it includes a Part D spending cap and does not model enrollment shifts or increases in MA spending, the direction and magnitude of the findings are consistent. In both the Urban Institute and our analysis, OOP caps increase federal spending while reducing beneficiary liability and exerting downward pressure on Medigap premiums. The larger Medicare expenditures observed in our analysis likely reflect the later projection period and continued growth in Medicare spending.

Our findings are also consistent with Congressional Budget Office estimates from other Medicare policy settings, including the Medicare Part D redesign enacted in the Inflation Reduction Act, which projected higher federal spending as beneficiary financial liability was reduced through the introduction of an annual OOP cap.²⁴ While a separate analysis by the Congressional Budget Office examined the impact of an OOP cap for Part A and B as part of broader Medicare reforms, their approach did not identify the independent effect of an OOP cap.²⁵ To put the magnitude to our estimates of the impacts of an OOP cap in CMS expenditures, they are relatively close to the costs of adding dental, vision, and hearing care to TM, which the Congressional Budget Office estimated would average $35.8 billion per year from 2020 to 2029.²⁶

Policy Implications

Our findings suggest that OOP caps in TM would have substantial impacts for enrollment patterns, CMS expenditures, and beneficiary savings. Nonetheless, the caps are quite costly and would face an uphill battle legislatively, as seen through other recently proposed benefit enhancements in Medicare.²⁶ Under statutory pay-as-you-go rules, increases in mandatory Medicare spending are generally required to be fully offset over both 5- and 10-year budget windows. As a result, OOP caps would likely require substantial offsets or accompanying benefit or financing reforms to satisfy budget neutrality. Spending offsets to support an OOP payment cap could come through site neutral payment reform, targeted reductions in overpayments to MA plans, estimated to be over $84 billion a year, or other strategies.^27,28

In addition to supporting Medicare beneficiaries, OOP caps would also substantially benefit health care clinicians and hospitals. A large share of high OOP costs is never paid to clinicians and hospitals and turn into uncollected liability.²⁹ While CMS pays for a portion of this bad debt, OOP caps would reduce clinicians and hospitals’ administrative burden when seeking payment from beneficiaries and reduce an important part of payment uncertainty.

Limitations

Our study has several limitations. First, our model considered only the financial benefits to beneficiaries and did not incorporate important benefits for beneficiaries outside of financial savings. Evidence of beneficiary switching implies utility gains to beneficiaries even when beneficiary spending may increase (eg, if a beneficiary switches from no supplemental insurance to Medigap and then pays Medigap premiums). Beneficiary health may also improve through increased use of care. A more complete welfare analysis would incorporate utility gains from an OOP cap, including those resulting from greater financial security, improved mental health,³⁰ and avoiding the hassle and restrained choice from network restrictions and prior authorization in MA.³¹

Second, our model did not incorporate a comprehensive demand function for health insurance, which includes many relevant features related to supplemental benefits, network preferences, cost sharing, and price along with beneficiaries’ risk preferences. Instead, assuming all else being equal, beneficiary preferences and plan features were held constant, and that the OOP caps affected demand only through changing Medigap premiums. As a result, we were unable to assess certain outcomes (like beneficiaries choosing to drop Medigap coverage altogether if an OOP were implemented) and policy choices (eg, banning Medigap after implementing an OOP cap). However, this modeling limitation attenuates the costs of OOP caps to Medicare, as decreasing Medigap enrollment would increase cost sharing and decrease use. As a result, our modeling approach may overstate the costs of OOP caps to Medicare. Third, our analysis relied on several assumptions about beneficiary behavior, Medigap premiums, and the relative expenses of TM and MA. While some elasticity estimates were quite consistent in the literature (eg, own price elasticity for Medigap and medical care) other estimates were less certain (eg, cross price elasticity between Medigap premiums and MA). Our sensitivity analysis suggested that projected CMS expenditures were most sensitive to assumptions about the elasticity of demand for medical care, followed by assumptions about MA benchmark pass-through and other MA behavioral responses, and least sensitive to assumptions about the own-price elasticity of Medigap demand. In addition, baseline projections of Medicare enrollment and spending were derived from the Medicare Trustees Report. Deviations from these projections could affect estimated model impacts. We also did not consider the role of age and experience rating in Medigap, and how this may affect beneficiary choice and switching. While a more comprehensive analysis would account for state variations in Medigap regulations, sensitivity analysis found that our results were not overly sensitive to assumptions about beneficiary switching.

Fourth, our analysis did not consider heterogeneity across individuals and insurance markets. For instance, the effects of OOP caps are likely to vary across beneficiary income and health status and across different state regulatory regimes with respect to Medigap. Lower-income individuals and those with greater health care needs may benefit disproportionately from OOP caps, although this was not evaluated in our study. Relatedly, beneficiaries who switch supplemental insurance regimes likely differ in unobservable ways from those who do not switch,³² which we did not incorporate into our analysis. We also did not consider the potential dynamic response from MA and Medigap plans to attract beneficiaries in light of OOP caps in TM. Fifth, we did not model the impacts of uncollected liability, instead assuming that beneficiaries would be responsible for all OOP costs. In reality, beneficiaries have the ability to negotiate for lower bills or declare bankruptcy rather than paying large OOP bills. Finally, we did not consider the role of stand-alone Part D plans or the potential dropping of supplemental coverage altogether to limit the complexity of our model.

Conclusions

Results of this study suggest that introducing an OOP cap in traditional Medicare could substantially reduce beneficiary financial burden but would increase federal Medicare spending. These findings highlight the trade-off between improving financial protection for beneficiaries and the substantial federal costs of extending an OOP limit to traditional Medicare.

Supplement 1.

eMethods

eTable 1. Key microsimulation assumptions and parameter values

eFigure 1. Estimated effect of out-of-pocket caps on Medigap premiums

eTable 2. Estimates of cross-price elasticity of demand

eFigure 2. Phase-in function for beneficiary switching

eTable 3. Sensitivity of results to varying key parameter values

eFigure 3. Change in Traditional Medicare and Medicare Advantage enrollment from OOP caps

eFigure 4. Effects of out-of-pocket caps on components of CMS expenditures

Supplement 2.

Data Sharing Statement