Health Spending, Retiree Income, and How to Combine Pension and Health Affordability in Retirement

March 11, 2026

The burden of out-of-pocket (OOP) health spending relative to income is a key measure of retirement well-being, according to a new Issue Brief from the Center for Retirement Security at Boston College (CRR). However, even with Medicare coverage — and ignoring long-term care expenses — retirees still face sizable costs for premiums, copays, and uncovered services which consume a substantial portion of total retirement income (including Social Security). In short, manageable healthcare OOP spending is critical to meaningful retirement security — but after subtracting OOP spending, the typical retiree has only 71 percent of Social Security and 88 percent of total income left, CRR stresses. More recently, Congressional testimony offered by the National Institute for Public Employee Health Care Policy focused on drivers of those costs in the prescription drug supply chain and framed the issue through the lens of public sector purchasers attempting to keep coverage affordable and sustainable for employees and retirees. In short, CRR quantifies how much retiree health care — especially premiums and OOP — eats into retirees’ income, while the Institute explains why those costs are so hard to control in the drug supply chain and federal programs. The bottom line? Health costs shape retirement adequacy, and retirement systems should not ignore health‑care economics, even if they do not administer a retiree health plan.

As CRR notes at the outset, it is true that over the last decade, much has changed for Medicare beneficiaries in ways that could affect their OOP costs, with some changes tending to push down cost growth, such as the increase in Medicare Advantage coverage – with more than half of beneficiaries now opting for such, with an increasing share of these plans charging no premiums. But premiums in Medicare Part B have risen rapidly, “driven in part by coverage for expensive prescription drugs,” CRR points out.

The new CRR brief, which updates earlier research, examines the extent to which OOP medical expenses affect retirees’ finances. CRR explains that the primary question in its analysis is how OOP spending affects the share of Social Security benefits and total income available for non-medical expenditures, for older Americans overall and by subgroups.

CRR does so by using the 2018-2022 “waves” of the Health and Retirement Study (HRS) — a University of Michigan panel study that surveys a representative sample of more than 20,000 Americans over the age of 50 every two years, supported by the National Institute on Aging (NIA) and the Social Security Administration. The study explores the changes in labor force participation and the health transitions that individuals undergo toward the end of their work lives and in the years that follow by collecting information about income, work, assets, pension plans, health insurance, disability, physical health and functioning, cognitive functioning, and health care expenditures. This allows the calculation of the share of Social Security benefits and total income available for non-medical spending and explores how this measure – the post-OOP ratio – differs by age, health status, and income and, most importantly, supplemental insurance coverage. [“Waves” of the HRS refer to these biennial (every two years) survey interviews conducted since 1992.]

CRR explains that “the general public and most policy analysts tend to evaluate the adequacy of retirement income, and Social Security benefits in particular, based on the level of retirees’ total income.” But it points out that their income net of OOP medical costs, which are often considered nondiscretionary (i.e., related to chronic health conditions such as cancer, lung disease, stroke, heart problem, diabetes, and high blood pressure) — is more relevant to their purchasing power.

Furthermore, even though retirees aged 65 and older have Medicare, they still face considerable costs. CRR cites, for example, in the case of Medicare Part A — which covers inpatient hospital care and is financed primarily by payroll taxes — beneficiaries face cost sharing. Medicare Part B, which covers physician and outpatient hospital services, and Part D, which covers prescription drugs, are partly financed by premiums and include further cost sharing.

Finally, because Medicare’s OOP costs are often substantial, many enrollees buy supplemental coverage, which may include additional premiums. Retirees who do not do so face the full cost of the many services not covered by Medicare, such as dental, vision, and hearing. (NB: CRR stresses that spending on long-term care, which can be substantial, is excluded from this analysis “in order to focus on the impact of OOP spending in a typical year.”)

CRR also underscores that even though “recent evidence” indicates that most Medicare Advantage plans are now offered at no additional cost (beyond the Part B premium), these plans use more restrictive provider networks that may limit enrollees’ access to their preferred doctors and hospitals, with the corresponding “costs” this can impose on retirees.

CRR emphasizes that accounting for OOP cost burdens is important, “because it is crucial to know how much retirees who rely exclusively on Social Security have remaining for non-medical spending.” In addition, understanding how benefit adequacy varies by subgroups “helps identify those who may be particularly at risk.” Finally, with the growing importance of supplemental insurance, “participants need to understand what types of coverage are likely to leave them in the best position.”

In explaining its analysis, CRR makes the following key statements:

The sample is limited to respondents who are ages 65 and over and are receiving both Social Security and Medicare, and it excludes those who are working or report receiving health insurance from a current employer or spouse’s employer. In other words, the sample is limited to retirees fully detached from the labor force and reliant on Medicare.
The three key components of the study are Social Security benefits, total personal income, and OOP medical expenditures (excluding long-term care). These are derived from self-reported information in the HRS.
In terms of OOP expenditures, the HRS captures prescription drugs, special facilities, surgery, and medical visits to doctors, hospitals, and dentists. It also includes self-reported measures for premiums paid for Medicare Part D, Medicare Advantage, and private supplemental plans. Medicare Part B premiums are imputed from reported income. These components are combined to calculate the share of income remaining after out-of-pocket spending for each beneficiary in each year.

The CRR analysis examines the extent to which outcomes differ by age, health status, and household income.

For the full sample:

The median retiree spent $5,444 on medical costs in 2022 in “nominal dollars,” which represent the face value of money, spending, or income in a given year without adjusting for inflation or changes in purchasing power. (Remember, also, that the median is the exact middle point of a group – 50 percent above, 50 percent below — whereas the mean, or average, is the numerical sum divided by the count. The median is not affected by extremely high or low numbers. The average (mean) is highly sensitive to outliers, which can skew the result.) Premiums comprise the bulk of OOP costs.
For the median retiree, in 2022 (the latest data available), only 71 percent of the Social Security benefit remains after paying premiums and other OOP costs. CRR stresses that “OOP spending is much more burdensome at lower post-OOP income levels, pointing out that five percent of retirees have essentially none of their SS benefit left after medical OOP costs. At the 10th percentile, retirees spend all but one-quarter of their benefit on medical OOP costs. CRR says that these results “demonstrate that, for a large number of retirees, OOP costs comprise a sizable share of Social Security income.” [It should also be noted that CRR points out, in a footnote, that on top of medical spending – which most analyses treat as outside of the individual’s discretion – retirees face a substantial amount of other non-discretionary costs, such as housing expenses, taxes, and non-housing debt – which “consume about 30 percent of retirees’ household income, leaving even less for surprise expenses and any other desired spending.”]
When looking at total income, the median retiree has 88 percent of his total income left over, but five percent of the sample is left with as little as 40 percent of total retirement income after medical spending.

When CRR looked at the impact of OOP expenses from the perspective of age, it found that the shares of both Social Security benefits and total income remaining after OOP changes very little, showing just a slight decline as people grow older.

As far as the impact of health status, the share of Social Security benefits or total income available for non-medical spending is “surprisingly similar for retirees with and without health concerns,” CRR finds.

Finally, in terms of differences across the income distribution, CRR calls the pattern “predictable,” with the share of total income remaining after accounting for OOP costs rising with income. The highest quintile has 94 percent of total income remaining, even after accounting for income-related premium surcharges for Part B. The lowest quintile has 82 percent, which CRR explains incorporates the beneficial impact of Medicaid. Excluding those who report Medicaid coverage (about half of the lowest quintile’s sample), the post-OOP ratio falls to only 76 percent.

CRR also points out that one of “the more important distinctions with respect to OOP spending is the type of supplemental insurance retirees have.” For example, Medicaid enrollees have the highest share of income – both Social Security and total – remaining after OOP spending, which CRR says “is to be expected given that Medicaid often has no premiums and minimal cost sharing.”

Among the other groups, CRR finds it helpful to look by source of income separately. For example, with respect to Social Security, “surprisingly, those with just traditional Medicare appear to do the best, at least for the median retiree, followed by those with Medicare Advantage and those with retiree health insurance. CRR explains that these differences “are due entirely to premiums, despite the growth in zero-premium Medicare Advantage plans.” All three groups have similar Social Security income and spend a similar amount on cost sharing and uncovered services, but those with no supplemental insurance pay the least in premiums.

In conclusion, CRR found that, at the median, OOP medical costs – including premiums, cost sharing, and uncovered services (excluding long-term care) – leave only 71 percent of Social Security benefits available for spending on other items. Premiums for Medicare Parts B and D, Medicare Advantage, and supplemental plans (including retirement health insurance) “make up the lion’s share of medical spending for most retirees, except those with the highest spending.” Perhaps most disturbing is the fact that the share of income remaining after OOP spending is lower for those in poor health and low-income households. In short, with such a substantial portion of income going to medical costs, “retirees’ finances are more precarious than Social Security benefit levels alone might suggest,” CRR warns.

So, what can be done to address the situation, particularly as it impacts public sector employees, retirees, and their families? Recently, the National Institute for Public Employee Health Care Policy submitted testimony to the House Committee on Energy and Commerce Subcommittee on Health in response to its February 11, 2026, hearing, Lowering Health Care Costs for All Americans: An Examination of the Prescription Drug Supply Chain. [The Institute was formed in 2023 as a joint effort of the Public Sector HealthCare Roundtable and the Pharmaceutical Care Management Association (PCMA). It is a non-profit (501)(c)(3), nonpartisan, national policy institute focused on public policy areas impacting health care plans available to public sector employees, retirees, and their beneficiaries. The Institute conducts research, develops authored papers and issue briefs, and hosts events to raise awareness and educate the Washington, D.C. policy community.]

Prescription drug affordability is one of the Institute’s highest priorities, it said in its testimony. The Institute reported that public sector purchasers who responded to its annual Specialty Drug Survey in 2024 purchased over $38 billion worth of health care services including medical, behavioral, and pharmaceutical services. Of this, public sector purchasers spent over $13 billion supplying employees, retirees, and their dependents with pharmaceutical products, almost $6.5 billion of which were spent on specialty drugs.

Indeed, from 2022 to 2023, pharmaceutical spending increased by nearly 20 percent, largely driven by specialty drugs, the Institute emphasized, with approximately half of total commercial pharmacy costs attributable to specialty medications. “These findings underscore a clear trend: specialty drug spending is one of the most significant and fastest-growing cost drivers for public employers,” the institute underscored.

Their testimony essentially focused on drivers of those costs “upstream,” especially in the prescription drug supply chain, and framed the issue through the lens of public sector purchasers trying to keep coverage affordable and sustainable for employees and retirees. This compares to the CRR brief, which documents how OOP health spending erodes retirees’ income, even with Medicare and supplemental coverage.

As Microsoft Copilot characterizes it, CRR looks at the “the household‑level burden,” while the Institute focuses on “the system‑level mechanics that make that burden so hard to control.” In short, affordability is the central metric, with CRR treating affordability as the key measure of retirement well‑being — how much income is left after health spending – while the Institute defines its mission as protecting the affordability and sustainability of public sector health benefits for employees and retirees. Copilot says that they both implicitly share the same message: coverage is not enough — net, after‑health‑cost income is the real outcome that matters.

The Institute’s testimony supports efforts by the U.S. Food and Drug Administration to accelerate biosimilar development — a process designed to create a therapeutic biological product that is highly similar to an existing FDA-approved reference product, with no clinically meaningful differences in safety, purity, or potency — and reduce regulatory barriers. It also encourages Congress to advance policies that strengthen competition, pointing to particularly anti-competitive patent practices which can delay generic and biosimilar entry, increasing costs for public purchasers.

The Institute testimony also notes that demand for GLP-1 medications is creating unprecedented financial challenges for public sector health plans, which it says are “increasingly confronted with difficult tradeoffs: raising out-of-pocket costs for employees and retirees; reducing covered benefits; or increasing taxpayer burdens. “As high-cost therapies, including GLP-1s, cell therapies, and gene therapies, expand, fiscal pressures will intensify,” the Institute warns, urging sustainable pharmaceutical pricing which it says is “essential to maintaining comprehensive and affordable benefits for public employees and retirees.”

Stepping back, one important message from both the CRR brief and the Institute testimony is that OOP expenses – particularly those associated with prescription drugs – are a very real pressure point in retiree budgets. For example, changes in Medicare Advantage and Part D can materially affect retirees’ OOP burden, especially for prescription drugs, while the prescription drug supply chain—rebates, pharmacy benefit managers (PBMs), manufacturer pricing, and the dynamics associated with them — affect what public plans and their members actually pay. Once again, Copilot describes the situation as one in which the same drug‑pricing architecture that public plans that administer healthcare are trying to manage upstream shows up downstream as retirees’ OOP costs and income erosion.

When Microsoft Copilot was asked to identify lessons for state and local governments as a result of this intersection of the CRR brief with the institute’s testimony, it identified one notable recommendation: integrate retirement and health policy in governance. Specifically, it noted that pension boards, OPEB trustees, and health plan boards can often operate in silos, but the CRR brief makes clear that “health spending is a retirement‑income problem, not just a health‑plan problem.”

Also, Microsoft Copilot made the following observations when asked if there are lessons for public pension plans that are not directly involved with the provision of public employee healthcare that should be taken away from the CRR brief and the institute testimony. Understanding that AI is not always accurate — and its broad-brush generalizations may deserve a certain degree of skepticism — nevertheless, here are some thought-provoking observations:

Health costs are a retirement‑security variable, not a health‑plan variable. The CRR brief makes this unavoidable: retirees lose a large share of their income to premiums, cost sharing, and uncovered services. That erosion affects net replacement rates, member financial stress, longevity risk and spending patterns, and political pressure on the plan. “Even if a retirement system does not run a health plan, it is still responsible for delivering retirement security. Health spending is now one of the biggest determinants of whether that mission succeeds,” Copilot observes, and recommends that pension systems should consider integrating health‑cost assumptions into their retirement‑adequacy modeling, member education, and board reporting.
Drug‑supply‑chain dysfunction affects retirees even when the pension system is not the purchaser. The Institute’s testimony shows how opaque pricing, PBM spread, and manufacturer practices inflate costs for plans and members. For retirees, those distortions show up as: higher premiums in the health plans they enroll in; higher OOP costs for drugs; and greater volatility in annual health spending. “Even if a pension system is not the purchaser, its members are exposed to the downstream effects,” Copilot notes.
Coordination with the health‑benefits side should no longer be viewed as optional. Copilot points out that most states have fragmented governance: one entity runs pensions, another runs active‑employee health, another runs retiree health, and Medicare Advantage vendors operate in yet another silo. The CRR brief shows that retiree OOP burden is a retirement‑security issue. The Institute’s testimony shows that drug‑supply‑chain policy is a cost‑driver issue. “Together, they imply that pension systems should not treat health benefits as ‘someone else’s problem,’ and should consider establishing structured coordination with the health‑benefits authority to share data, align assumptions, and jointly monitor retiree affordability,” Copilot recommends.
Member education is part of fiduciary duty. Even if a retirement system does not administer health benefits, “it is often the only entity retirees trust for long‑term planning guidance,” Copilot observes. CRR’s findings show retirees systematically underestimate future premium growth; drug‑related OOP exposure; and the share of income consumed by health spending. The Institute’s testimony shows why these costs are structurally hard to control. Pension systems should consider incorporating health‑cost education into retirement counseling, seminars, and online tools — “not to give medical advice, but to help members plan realistically,” Copilot underscores.
Health‑cost risk is a driver of political and fiscal pressure on pension systems. When retirees face high health costs, Copilot points out that they often delay retirement; seek ad hoc COLAs; lobby for benefit enhancements; and increase political pressure on the system. “Even if the pension system does not run a health plan, it absorbs the political consequences of health‑cost inflation,” Copilot believes, and says monitoring retiree health‑cost trends can be a form of risk management for the pension system itself.
Actuarial assumptions should reflect health‑cost realities. Copilot finds that most pension systems model wage growth, inflation, COLA patterns, and longevity – but they “rarely model retiree health‑cost burdens, even though it affects retirement timing, member behavior, demand for COLAs, and long‑term adequacy of benefits.” Copilot therefore recommends that pension systems “should incorporate health‑cost sensitivity analysis into actuarial and policy modeling, even if they do not administer health benefits.”
Federal policy changes in Medicare and Part D affect pension outcomes. The CRR brief highlights how Medicare Advantage and Part D reforms change retiree OOP exposure, while the Institute’s testimony highlights how federal drug‑pricing reforms change plan and member costs. Copilot believes these federal levers influence retiree financial well‑being, demand for COLAs, retirement timing and member satisfaction with the system. Therefore, it recommends that pension systems should track federal health‑policy changes as part of their strategic environment, not as an externality.

The bottom line, Copilot stresses, is that “[e]ven when a public retirement system does not administer health benefits, health costs shape retirement adequacy; drug‑supply‑chain dysfunction drives retiree OOP exposure; federal reforms alter retiree financial security; and member behavior and political pressure flow back to the pension system.”

Copilot concludes that the CRR brief and the Institute’s testimony together “make one thing clear: retirement systems cannot ignore health‑care economics, even if they do not run a health plan.”

What do you think?