Smarter products and policies can connect wary investors to lifetime income
When it comes to later-life worries, two-thirds of Americans say they are more scared of running out of money than death itself. Given that mindset, it makes perfect sense that an even higher percentage of people in or near retirement say they like the idea of guaranteed income: a steady payment that keeps coming for as long as they live.
Yet just 12% of Americans at least 50 years old who have at least $100,000 in savings have bought the product designed to provide exactly that: a traditional life annuity, which turns an upfront payment into guaranteed income for life.
This conundrum, known as the “annuity puzzle,” is not new. A seminal 1965 economics paper by Menahem Yaari made the case that a rational person would choose annuities as a form of later-life financial protection.Twenty years later, economist Franco Modigliani used his Nobel lecture to note that households rarely bought annuities — and the reason why remained poorly understood.
In the 40-plus years since, a great deal of research has made the puzzle far less puzzling. In a National Bureau of Economic Research working paper, UCLA Anderson’s Hal Hershfield, Cornell’s Suzanne Shu, University of Illinois’s Jeffrey Brown, Hebrew University’s Abigail Hurwitz, York University’s Moshe Arye Milevsky, University of Pennsylvania’s Olivia Mitchell and retirement-income consultant Tamiko Toland draw on that accumulated knowledge to explain what drives the disconnect and to explore potential solutions.
The DIY Retirement Plan
They bring plenty of expertise to the conversation as all have been central contributors to the research. In “The Annuity Puzzle Revisited: Barriers, Behavior, and Policy Paths to Lifetime Income,” they lay out a case that now that we understand saver wariness — and it is not irrational — better product design and important regulatory changes can help more people embrace what they intuitively want: more financial security in retirement.
The urgency could hardly be greater. Today’s private-sector retirees are the first generation to have largely DIY’d their retirement. The rise of the 401(k) that began in the early ’80s allowed corporate America to shift the burden of retirement saving from employer to employee. Today, far fewer private sector firms offer traditional pensions.
When Modigliani delivered his Nobel lecture, 401(k)s were still in their infancy, with total assets below $150 billion. Today there is about $10 trillion in 401(k)s and another $19 trillion in individual retirement accounts. At retirement, savers must switch gears entirely — from accumulating money to figuring out how to live off it without running out. Which explains both the outliving-one’s-money terrors and the potential for annuities to reduce that fear by adding a guaranteed lifetime income stream alongside Social Security.
Deconstructing the Puzzle: A Three-Part Diagnosis
What old-school economic theory failed to incorporate is that people are human. The advent of behavioral economics, and the need to understand individual motivations in a retirement ecosystem that has flipped from employer pensions to employee-driven retirement planning, has identified very logical reasons for the annuity disconnect. This working paper is essentially a review of the key advances over the past few decades.
The collaborators sort the drivers of the puzzle into three main avenues.
Rational Economic Trade-Offs
Consumers who express interest in guaranteed lifetime income and then refuse to buy an annuity aren’t being abjectly irrational. There are very real trade-offs that come into play. Signing an irreversible annuity contract requires forking over a big sum of money that effectively reduces cash on hand. If a retiree experiences a sudden medical emergency, faces long-term care costs that can easily run to thousands of dollars a month or harbors a deep desire to leave a financial legacy for children or grandchildren, locking up their nest egg is a genuine risk.
Co-author Mitchell, executive director of Wharton’s Pension Research Council and one of the most influential researchers on the structure of the U.S. retirement system and the challenges it presents for households, has repeatedly highlighted these competing realities: uncertain medical expenses, the desire to leave money to family, differing health profiles and varying levels of existing guaranteed income. Those factors create legitimate reasons to hold on to savings, even for people who value the protection lifetime income provides. But that does not argue against annuities so much as against all-or-nothing annuitization. The more practical goal is helping retirees consider whether some additional guaranteed income, layered on top of Social Security, can reduce the risk of running out of money while still leaving savings available for other needs.
A Confusing and Fragmented Product Market
The annuity market is complex, lumping together simple guaranteed lifetime income products with other types of annuities that often combine an investment component tied to stocks, and other features — protection from market volatility, payments to beneficiaries, etc. — that can be compelling but also come at a cost of a reduced annuity payment.
A retiree who is open to the idea of using some savings for guaranteed income and has done plenty of saving inside a closed and regulated 401(k) is faced with venturing out into the retail world of financial services and sorting through the menu of products labeled as an annuity that have very different features and tradeoffs.
Annuities are insurance products and the California Department of Insurance’s webpage, “Annuities What Seniors Need to Know,” for instance, runs to more than 3,500 words and raises as many concerns as it puts to rest.
Is the insurer underwriting the annuity financially sound today — will it be 30 years from now? What are the terms of your state’s guaranty fund in case of insurer insolvency, or is it the state where the insurer’s based? Little wonder that savers often turn to financial advisors to sort through the complexity. Advisors, though, may have a strong incentive, in the form of higher fees, to sell a complex annuity, rather than a simple one.
Indeed, the more complex annuities dwarf the simple income ones. The authors note that single premium immediate annuities — the specific tool designed to solve the problem of outliving your money — account for roughly $15 billion in annual sales, compared with $500 billion for annuities built around investment growth and market protection.
Shu’s research with UCLA Anderson’s Robert Zeithammer and Duke’s John Payne adds an important point here: Consumers are not judging annuities only by expected payouts. Showing cumulative payouts across different life expectancies, or illustrating how annuity payments map to routine monthly expenses, can make the product easier to understand and more appealing — without changing the annuity itself. People don’t need a better deal. They need a clearer picture.
Annuities Are a Behavioral Quagmire
Even if a retiree gets past the rational trade-offs and the confusing product market, buying an annuity requires managing emotional resistance. For starters, after decades working so hard to save for retirement, handing over a chunk of that money in exchange for income that will arrive slowly, over an unknowable number of years, is clearly not an easy self-ask.
Brown’s foundational work on how an annuity is framed can melt some resistance. In research published in the American Economic Review in 2008, Brown and his co-authors found that when study participants considered an annuity as an investment — focused on returns and account balances — just 21% preferred it. When the same product was framed as insurance that would provide lifetime income, more than 70% preferred it. That finding exposed a major disconnect: The financial industry was often trying to sell a peace-of-mind product in the return-heavy language of investing.
And earlier research had already shown that the insurance frame was not just clever marketing. In a 1999 American Economic Review paper, Brown, Mitchell, MIT’s James Poterba and TIAA-CREF’s Mark Warshawsky estimated how much lifetime income buyers could expect from individual annuities.
A typical 65-year-old from the broader population could expect to receive 80 to 85 cents back in payments for every premium dollar paid. For actual annuity buyers, a subset of the population, who tend to live longer, the estimated average payout was more than 90 cents. We routinely pay home and auto insurance premiums hoping to never collect a penny back through a claim. An annuity is unusual insurance: The buyer is paying for protection against the risk of living a very long time yet, on average, can still expect to receive much of the premium back as income. Live a long time and the payouts will exceed the premium cost.
Shu’s fairness research adds another dimension. In separate work with Zeithammer and Payne, she found that many consumers view traditional annuities as unfair: If the buyer dies sooner than expected, the insurer keeps the money. That sense of unfairness turns out to be a powerful predictor of annuity avoidance — and it responds not to better payouts but to better explanation of how the product actually works. When people understand that buyers who die early effectively help fund the income of those who live longest, and that no one is betting against the house alone, resistance softens.
Hershfield’s future-self research helps explain another piece of the resistance. An annuity is designed to protect an older version of oneself — perhaps an 85- or 90-year-old still very much alive and still needing income. But people often feel psychologically distant from that future self. The visible loss of cash today can feel more real than the benefit to someone who is hard to imagine.
Hurwitz’s research on longevity awareness adds yet another layer. A collaboration with Mitchell and Hebrew University’s Orly Sade established that people often misjudge the odds of living to a very old age, which means they may undervalue the protection an annuity provides. Her research has shown that giving people objective information about longevity risk can increase interest in guaranteed income. In other research with Mitchell, she has also found that retirees who passed on purchasing an annuity earlier in their life often wish they had made a different choice.
Milevsky brings a mathematician’s rigor to deconstructing how annuities can be best structured (and the timing of purchase) to help individuals weigh guaranteed income against other needs. He has been a driver of the message that rather than asking retirees to make a single, irrevocable decision at retirement, approaches that break the problem into smaller, more manageable pieces over time can be more palatable. And partial annuitization offers a middle ground: Retirees can create an additional stream of income they cannot outlive while still retaining substantial savings for health care expenses, family needs and unexpected emergencies.
And the timing of the decision matters. Research cited in the paper — including work by Shu — suggests that workers in their 40s are often more receptive to guaranteed lifetime income than people approaching retirement. It seems that decades of saving can create a fierce sense of ownership that makes using a portion of it to buy an annuity feel less like purchasing insurance and more like surrendering a hard-earned asset.
Getting the System to Step Up with Solutions
The collaborators synthesize the decades of research into recommendations to revamp the guaranteed-income annuity landscape. The answer, they argue, is not to keep expecting individuals to solve decumulation (the rendering of savings into income) on their own. Retirement plan sponsors, regulators and product designers need to make plain-vanilla guaranteed income easier to find, understand and choose. This echoes research Shu collaborated on that addressed the lack of decumulation nudges in the retirement product landscape in which accumulation nudges abound (automatic enrollment, automatic contribution rate escalation etc.)
That is where Toland’s work comes to the table. A former head of annuity research at CANNEX, an annuity data and analytics firm, and senior executive at TIAA, a leading annuity provider and retirement plan provider, she has worked deep inside annuity product design, pricing and plan-level implementation. Her research with Michael Finke, professor of wealth management at The American College of Financial Services, has advanced the case for 401(k)s to build in automatic annuity defaults of a modest portion of savings into lifetime guaranteed income. This dovetails with Mitchell’s research that retirees have legitimate reasons not to annuitize everything, while reflecting the behavioral evidence that many people who could benefit from guaranteed income find it hard to buy in.
The authors point to the December 2025 announcement from Vanguard, a major 401(k) provider, and TIAA, as an example of where the industry can step up to help savers more easily get the extra guaranteed income they consistently say is wanted. For 401(k) participants approaching retirement who own a Vanguard target date fund inside their retirement account, Vanguard will now include TIAA lifetime-income annuities as a replacement for part of the fixed-income portion of the TDF.
Target-date funds are already the default landing place for many 401(k) savers, with total assets near $5 trillion. The paper notes that three-quarters of participants mistakenly believe that target-date funds currently provide guaranteed income at retirement. Adding lifetime income to a target-date fund would provide a feature savers already think they have.
Big shifts such as this are only now possible because of a shift in the regulatory landscape. A big roadblock has been that the federal regulations mandating how retirement plans can operate has not explicitly given plans a “safe harbor” for offering annuities. Without that protection, litigation-fears have kept plans from delivering the guaranteed income they know many of their participants want (and need).
Recent changes now make it more practical for plans to finally deliver in-plan annuities. The SECURE Act of 2019 created a fiduciary safe harbor for employers selecting annuity providers, and SECURE 2.0, enacted in 2022, continued efforts to make lifetime income a more practical component of workplace retirement plans. The Department of Labor’s 2025 Advisory Opinion pushed further, clarifying that income benefits can be incorporated into default investment portfolios and giving plan sponsors additional confidence that they can offer these features without running afoul of their fiduciary obligations.
That policy progress helps explain why workplace retirement plans are increasingly central to the annuity-puzzle solution. With liability concerns now off the table, it’s up to plan sponsors and plan providers to step up and offer this feature that has repeatedly been flagged as a high priority by plan participants. Rather than asking retirees to navigate a fragmented retail market on their own, lifetime income can be built into the same plan architecture that already helps workers save.
Mitchell and colleagues have modeled one version in which retirement plans would automatically direct up to 20% of a retiree’s 401(k) balance above a threshold into a lifetime income stream at age 67, while leaving the remaining 80% available for other needs. Defaults are not mandates — savers can opt out. But the default moves the decision and process to add lifetime income from the intimidating retail marketplace into the retirement plan itself, where savers are already accustomed to being guided through complex decisions.
The annuity puzzle was largely a theoretical conundrum for economists in the days before DIY retirement planning became the norm. Today it is a real-world challenge for millions of savers who are scared of running out of money and drawn to the idea of a paycheck that lasts for life. The research now exists to help policymakers, plan sponsors and product designers make that a realistic option for retirees tasked with making sure they don’t run out of money.
Featured Faculty
-
Hal Hershfield
Professor of Marketing and Behavioral Decision Making
-
Suzanne Shu
Professor Emeritus of Marketing
About the Research
Hershfield, H. E., Shu, S., Brown, J. R., Hurwitz, A., Milevsky, M. A., Mitchell, O. S., & Toland, T. (2026). The Annuity Puzzle Revisited: Barriers, Behavior, and Policy Paths to Lifetime Income (No. w35145). National Bureau of Economic Research.
