401(k) participants are largely on their own when it comes to solving the “nastiest, hardest problem in finance” of converting their account balances to retirement income. The 2019 SECURE Act endorsed one solution: incorporating annuity products — like those currently sold by insurance companies to individuals — into 401(k) investment menus.
But there’s another approach that would generate higher lifetime income: uninsured longevity pools in the form of “modern tontines.” Longevity pooling is the actuarial underpinning of the lifetime income provided by insured income annuities and traditional pension plans. Tontines unbundle the pooling, avoiding some of the baggage (fees, opacity, etc.) associated with insured annuities.
Tontines can provide any payment pattern available with insured products, such as income that lasts for life or until the later death of the pool member and her spouse. The income amount is based on the investment amount, the payment form selected, expected investment returns, and expected probabilities of death at each age.
The two key defining features are: (1) remaining account balances at the pool member’s death (or a participating spouse’s, if later) are reallocated to surviving members, and (2) there is no guarantor. So, while income lasts for life by construction, benefit amounts are not guaranteed but are adjusted periodically so that the pool remains fully funded, i.e., the actuarial present value of the pool’s projected payout stream is kept equal to the value of assets.
Income amounts do not have to be especially volatile, however. They can be engineered to remain relatively steady by investing in safe bonds and derivatives that match the projected payout stream, analogous to how traditional pension plan sponsors hedge liabilities. Alternatively, risky investments like equities can allow income to possibly grow in exchange for taking on the risk that income will decrease or be more volatile. Further, for a decent-sized pool, favorable or adverse longevity experience, which also results in income adjustments, will generally be small and emerge slowly over time.
Benefit adjustments by source can be calculated and straightforwardly disclosed. Modern tontines can be indefinite “open pools,” with new retirees entering continuously.
Longevity pooling works because, while it is impossible to know how long any given person will live, it is possible to know within a narrow range how many people out of a large group will be alive at any point in the future. Because the money left over from those who die is available to those still alive, the income for everyone in the pool can be higher during their lifetimes.
To illustrate the impact of pooling, consider a 65-year-old male with a life expectancy of 22 years (to age 87) who wants to use $1 million to generate safe income by investing in low-default-risk bonds (like Treasuries) expected to yield 4.25% per year. How much annual income can this generate?
If invested to pay out 21.8 years (life expectancy), $69,677
If invested to pay out 25 years (to age 90) $64,245
Pooling — income lasts for life, $73,138
With pooling, income for life is nearly 5% higher than the income paid over a fixed period equal to life expectancy, and almost 14% higher than a 25-year fixed payment period (to age 90), perhaps chosen for conservatism.
Features can be included to provide a legacy in the event of early death, but that would mean lower lifetime income. This is also true for commercial annuity products and traditional pensions.
The most straightforward approach to generating lifetime income with insured products would be for employees to purchase income annuities at retirement from an employer-curated list of insurers using 401(k) money rolled over into an IRA at retirement.
One advantage of this approach relative to tontines is that insurers will guarantee the amount of income. Disadvantages include the risk that the insurer may default on its promise, and, because the insurer must hold reserves against the various risks it is taking on and make a profit for its shareholders (except for mutual insurance companies), lifetime income is likely to be lower.
Most insured products marketed to 401(k) plans are not this simple, however, but are adapted from complex retail products predominantly used as tax-advantaged savings vehicles, not to provide retirement income. In practice, only 5% to 10% of these retail policies are “annuitized.”
When such product features are used “in-plan,” e.g., integrated with a target date-fund (a “qualified default investment alternative”), a portion of the employee’s account balance is typically moved to an insured annuity investment each year starting, for example, at age 55. By age 65, the accumulated annuity investment will comprise maybe 30% of the participant’s total balance and can be used to buy lifetime income guarantees.
But there are often additional complications. Probably the most popular income delivery approach marketed to 401(k) plans — the guaranteed lifetime withdrawal benefit (GLWB) — has been described as providing “the income certainty of a guaranteed solution with the growth potential, liquidity, and flexibility of a non-guaranteed solution.” There are guarantees against investment losses on portfolios of risky assets. Participants can take withdrawals from their accounts at any time.
This sounds great, but the amount available for withdrawal at retirement may be less than the amount available for conversion to lifetime income, the difference being effectively a withdrawal penalty. And during retirement, withdrawals greater than the guaranteed income level may cause a disproportionately large reduction in the income amount guaranteed.
It’s unlikely that all these nuances are fully understood by most 401(k) participants or administrators, or their investment consultants. And these features and guarantees typically cost participants about 1% of their insured account balance per year.
Also, the complex risks associated with these products are challenging for insurers to manage. Many that sold similar retail products suffered financial difficulties during the global financial crisis. Employers needn’t worry, however. The lobbying efforts of the insurance industry resulted in broad, albeit conditional, employer protections from fiduciary liability under the SECURE Act if an insurer defaults.
Modern tontines have zero default risk.
While modern tontines would be new in the 401(k) market and currently challenging to implement under ERISA, they have been successfully used in the U.S. in contexts where ERISA restrictions don’t apply, and elsewhere.
In the U.S. higher-education and non-profit market, CREF (TIAA affiliate) has been issuing “variable annuities” (different from the retail product), which are essentially open-pool modern tontines, since 1952. Some national umbrella church organizations utilize uninsured longevity pooling to provide lifetime retirement income funded by 403(b) plan account balances at retirement.
In Canada, the open-pool University of British Columbia Variable Payment Life Annuity (VPLA) was modeled on CREF. Canadian income tax rules were adopted in 2021 to facilitate the use of VPLAs in other contexts.
Legislation or regulatory guidance facilitating the use of modern tontines in 401(k) plans would greatly improve the retirement system. But tontines might already be usable in certain contexts. Could a national union like the UAW set one up for its members who no longer have traditional pension benefits, funded with 401(k) balances rolled over to IRAs? Could a corporation use a modern tontine to satisfy the annuity payout requirement in a money purchase plan (a DC plan alternative to a 401(k))? Could defined contribution plans with tontine payouts be an alternative to fiscally out-of-control public pension plans? These are questions for lawyers and others, but it’s time for us all to loosen the retirement income conceptual straitjacket.
Larry Pollack is a consultant and pension actuary. He is the principal at LIP Consulting LLC, based in Hoboken, N.J., and publishes the Substack newsletter Pension Questions. This content represents the views of the author. It was submitted and edited under Pensions & Investments guidelines but is not a product of P&I’s editorial team.