45% at risk of running short in retirement, study finds

Rather than offering single headline numbers, the center examines the granular tradeoffs policymakers and plan sponsors face when considering changes to how Americans save for retirement.

The model is built on top of the Survey of Consumer Finances, a representative sample of the U.S. population. It tracks people's wages, savings, and job changes throughout their career—and reflects actual behavior, such as whether they cash out retirement plans when changing jobs or make pre-retirement withdrawals.

Once people enter retirement, the model incorporates Social Security claiming patterns, longevity, and long-term care costs. This granular approach captures what most retirement calculators miss: the dynamic financial decisions people make as circumstances change, not the static assumptions of fixed savings rates or preset withdrawal schedules.

The focus is on retirement income adequacy and the likelihood of running short of money, not on wealth balance or replacement rates alone. That shift in lens reveals which households are truly at risk.

It also incorporates Morningstar's capital market assumptions and a detailed database of household behavior during both the accumulation stage and retirement.

The richness of that data allows the model to project what kind of expenses individuals will have in retirement—from basic consumption to long-term care costs—with a level of detail that generic rules of thumb cannot match.

But the crisis is concentrated: lower-income workers, single women, Black and Hispanic households, and those who did not consistently save or participate in employer plans show markedly higher rates of retirement income inadequacy.

Those who have consistently saved and regularly participated in a plan tend to be in pretty good shape. Gen Z workers in the status quo analysis actually look slightly better than older age cohorts, many of whom experienced disruption during the shift from defined-benefit to defined-contribution plans.

The disparity reflects real structural differences: lower-income workers have less access to employer-sponsored plans, less ability to accumulate savings, and fewer resources to cushion against shocks or health crises.

When they are included, it climbs back to 40-45%. Long-term care has a very large impact on retirement adequacy numbers—and most retirement planning underestimates or ignores it entirely.

The uncertainty around long-term care costs is substantial: timing, type of care needed, whether it is institutional or home-based, family support options, and asset spend-down all vary by individual. That complexity, combined with high costs, makes long-term care one of the largest retirement risks most Americans do not plan for.

Participation in employer-sponsored plans is lower for lower-income workers than for middle- and higher-income workers. That participation gap directly feeds into retirement income shortfall estimates. Research shows that lower-income workers have a pretty good chance they could run short of money in retirement.

Plan access matters because without an employer plan, workers have fewer mechanisms to build retirement savings—and they are less likely to have access to advisers who can help them navigate other options. Policy initiatives targeting universal market participation, such as state auto-IRA programs, aim to cover workers without adviser access so they can start building equity.

The income gap in plan sponsorship is a structural problem that concentrates risk in the population that can least afford it.

Simulations at Morningstar show that only a very small percentage of the 401(k) participant base could save enough additional money to immunize themselves from a potential 22% reduction in Social Security benefits. The window to do so is narrow: workers already over age 50 cannot save enough in their 401(k) without contributing ridiculously large amounts—25% of compensation per year or more.

Younger workers have longer planning horizons and may have more options, but they must still stress test their plans for both payroll tax increases and benefit reductions. Most workers cannot simply save their way out of a significant Social Security cut through the qualified plan system alone.

Alternative approaches are needed: deferring retirement age, making additional contributions beyond qualified plans, or strategically using 401(k) savings in combination with other income sources. A single lever—401(k) contributions—will not solve the problem for most households.

Automatic enrollment is valuable, but simulations show that access alone is simply not going to be the answer for retirement income adequacy for many individuals. Enrollment must be paired with high default deferral rates and automatic escalation to move the needle on retirement outcomes.

Easy access to withdraw money—especially in Roth IRA situations—undermines savings in simulations. Locking up money until age 62 or a similar age restriction, combined with autoportability features that move money between jobs rather than allowing it to be cashed out, addresses the leakage problem that drains retirement accounts during career transitions.

One-year snapshots of plan participation numbers are insufficient for measuring whether changes actually improve retirement income adequacy. The test must be whether participants have enough at retirement, not whether they were enrolled.

The Savers Match offers a 50% match on the first $2,000 of workplace or IRA contributions per individual, for a maximum match of $1,000. Single filers get the full credit with modified adjusted gross income under $20,500 and see the credit phase out completely at $35,500.

For many lower-income workers, saving $2,000 per year may be out of reach despite the match incentive. Morningstar research estimates a 10-12% wealth increase for Gen Z and millennial households under the current program—a meaningful but modest gain that does not close the retirement income gap on its own.

Researchers created 43 plan prototypes to test the impact of voluntary versus automatic enrollment, escalations, default deferrals, and different match rates. By moving from one match rate to another, plan sponsors can make a huge difference over the decades that a young participant still has until retirement.

Changes near retirement have minimal impact because there is no time left for compounding. The impact also differs dramatically across different participant cohorts within the same plan, which means plan sponsors should ask not for average effects but which specific at-risk populations will respond, by how much, whether the impact persists over time, and what compounding horizon each group has.

The takeaway is clear: small tweaks to plan design can unlock large retirement outcome differences for younger workers, but only if sponsors focus on the at-risk population—those unlikely to have sufficient retirement income under the current plan design.

The Retirement Savings for Americans Act would introduce a federal thrift savings plan into the market for workers not covered by an employer-sponsored plan, automatically defaulting them into the program. Similar auto-IRA proposals offer a theme-based approach.

Multiple states now have state-managed plans that automatically enroll workers without employer coverage. The goal is to extend coverage to gig workers and others unlikely to have employer plans while reducing the burden on small employers to set up and maintain their own retirement plans.

Economies of scale in providing these services allow smaller companies to access the kinds of insights and optimization that were once available only to large plans.

Small employers can now access research from companies that have tested different plan designs to find what works best. That democratization of plan expertise helps smaller firms compete on retirement benefits without the massive cost burden that used to come with customization.

Long-term care needs vary by person, gender, income level, and wealth. There is no single number that advisers should plug in and use. Sensitivity testing should explore different scenarios: nursing home versus home healthcare, the role of family support, and the eventual spend-down to Medicaid eligibility.

Many retirees spend down their assets and eventually rely on Medicaid to cover long-term care. Advisers should discuss this approach with clients and model it as a realistic outcome, rather than assuming all long-term care costs will be paid out of pocket. Morningstar is studying long-term care insurance products, including combinations with whole life, universal life, or annuity products, to understand how those vehicles fit into comprehensive retirement income planning.

Existing data sources systematically undercount long-term care costs because they miss unpaid care provided by spouses or children. When families rely on family members to provide care, those costs do not appear in financial data—so any model built on financial records alone will underestimate the true burden.

Spencer Look and Jack VanDerhei's research framework simulates long-term care likelihood using data from the Health and Retirement Study, then models costs using Genworth assumptions. Their model includes Medicaid functionality based on asset and income limits, capturing how people actually navigate eligibility and benefits in the real world.

VanDerhei disputes the assumption that spending down assets to Medicaid eligibility is an acceptable retirement strategy most clients would embrace. If an adviser told a client "You don't have to worry about long-term care because we're going to spend you down to the point where Medicaid takes care of you," that strategy may not align with what retirees actually want to do.

Different treatment of long-term care across models is a major reason retirement crisis statistics vary so widely. Long-term care assumptions are fundamental drivers of the differences between Morningstar's findings and other researchers' conclusions—which means the assumptions themselves are part of the policy debate, not settled fact.

Simple life expectancy tools use only age and gender to project longevity. Advanced tools incorporate health status, income, and marital status—factors that studies show can definitely affect life expectancy and retirement security.

Spencer Look recommends sensitivity testing for high-income earners in good health, testing assumptions well above the age-90 baseline to account for the possibility of living into the mid-90s or beyond. No single life expectancy number should be relied upon; planning must account for uncertainty around the actual age of death.

That uncertainty compounds the long-term care problem: longer life increases the odds of needing extended care, and the compounding of that risk over a potentially 30+ year retirement is large enough to reshape retirement adequacy estimates.

Morningstar's upcoming publication will focus on access to retirement plans and increasing participation rates among workers currently excluded from employer plans. Reducing that participation gap is a direct lever for reducing income shortfalls.

The center continues to study long-term care insurance and its role in retirement income adequacy—recognizing that traditional insurance products, as well as new hybrid designs, need to be understood in the context of retirement planning.

A paper in progress examines modern tontine-style designs where participants pool money and share returns without individual guarantees. These pooled lifetime income products represent a different approach to managing longevity and long-term care risk—one that spreads costs across a group rather than concentrating them on individuals or relying solely on Medicaid.