Pension Lump Sum or Annuity: How to Compare Two Offers That Are Not Comparable

A note on method: the internal-rate-of-return approach we use below is most closely associated with Michael Kitces, whose writing at kitces.com developed the hurdle-rate framing for this decision. We would recommend reading his work on it in full. The extensions to real returns, plan funding risk, and the survivor tax layer are our own.

Where the Two Numbers Come From

The monthly benefit is the plan speaking its native language: a formula output built from your years of service and your salary history, promised for life. The lump sum is that promise converted into a present value, and the conversion is where most people meet their first surprise.

Private plans generally price lump sums under Section 417(e) of the tax code, using prescribed interest rate and mortality assumptions. The one thing to internalize is the direction of the math: your future payments get discounted using corporate bond yield segments, so rising rates shrink the present value of the stream, and the lump sum offer shrinks with them. Falling rates do the opposite. Two coworkers holding identical pensions who retire eighteen months apart can walk away with meaningfully different lump sums for no reason other than where rates stood on each plan’s measurement date.

And that measurement date is quirkier than people expect. Most plans lock their rates annually using a lookback month written into the plan document, so the environment pricing your offer can already be months stale when the packet lands, and crossing a calendar year boundary can change the number. Nobody should time a retirement around bond yields. Everybody should know that the lump sum is a dated snapshot, not a permanent feature of their pension, and know when their plan takes the picture.

One more mechanic before the analysis starts: an underfunded plan may not be permitted to hand you the full lump sum at all. That rule surprises almost everyone it touches, and we cover it below, because most treatments of this decision never mention it.

The Math Everyone Runs First

Instinct says payback. Lump sum divided by annual payments equals years to catch up.

One example carries through this entire article: a 65-year-old offered $625,000, or $3,850 per month for life, which is $46,200 per year. That is a 7.39% payout rate on the lump sum, in the neighborhood of what many plans offer. The payback math says $625,000 divided by $46,200 is 13.5 years, so the annuity “catches up” around age 78 and a half. Outlive that age and the annuity was the right call. Fall short and it wasn’t.

We took apart this style of thinking in our piece on when to claim Social Security, and every part of the critique applies here. Payback is counting, not analysis. It turns a permanent financial election into a coin flip on your date of death. It assumes the lump sum earns exactly nothing for thirteen and a half years, which nobody would assume about any other dollar they own. It never mentions inflation. And it is silent about the scenario that ought to drive everything, which is not the likely outcome but the unaffordable one.

The question worth asking first is different: what return is this annuity actually paying me, and what would I need to earn on the lump sum to match it?

The Return the Annuity Is Really Paying

Set the lump sum against the payment stream and solve for the rate that equates them over a given lifespan. That rate is the hurdle an invested lump sum has to clear. Its defining feature is that it is not one number. It rises with every year you survive.

For our example:

• ●Death at 75: the annuity’s return was roughly negative 5.6% per year. You collected $462,000 against a $625,000 price and the pool kept the rest.
• ●Death at 79: roughly 0.5%. Look hard at that one. It is the payback point, and it exposes what payback actually measures: at the moment the raw dollars “catch up,” the annuity has paid you nearly nothing as a return.
• ●Living to 85: roughly 4.3% per year. A conservative portfolio can plausibly clear that, but not effortlessly.
• ●Living to 90: roughly 5.7%. Clearing that takes real equity exposure, held through whatever markets do, during the very decades when a bad early stretch is hardest to survive.
• ●Living to 95: roughly 6.4% per year, guaranteed, immune to markets. Almost no retiree can prudently target 6.4% on the same dollars they are living on.

That progression is the honest shape of the product. An annuity is not a fixed return; it is a return that keeps improving as long as you keep living, funded by the participants who stopped. Mortality pooling is the machine inside every annuity, and no portfolio can rebuild it, because a portfolio cannot earn anything from other people dying early. So the question was never whether you will beat age 78 and a half. The question is: in the long-life scenario, the exact scenario where money has to last, do you want a guaranteed 6%, or the obligation to go earn one?

Then Inflation Walks In

This is where our treatment departs from most of what gets written on this decision, and where we lean differently than much of the profession.

All the hurdle rates above are nominal, and the typical private pension annuity has no cost-of-living adjustment. It pays $3,850 in the first month and $3,850 in the three hundred sixtieth. Run 3% inflation against that fixed number:

• ●At 75, the payment buys what $2,865 buys today, roughly 74% of its starting power.
• ●At 85, what $2,132 buys, roughly 55%.
• ●At 95, what $1,586 buys, roughly 41%.

Live thirty years and about 59% of the check’s purchasing power quietly leaves, and the guarantee never lifts a finger, because the guarantee only ever covered dollars, not what dollars buy.

Now restate the hurdle rates in real terms at that same 3%:

• ●Death at 75: roughly negative 8.3% real.
• ●Death at 79: roughly negative 2.4% real.
• ●Living to 85, which is life expectancy territory for a healthy 65-year-old: roughly 1.2% real.
• ●Living to 90: roughly 2.6% real.
• ●Living to 95, an outcome any family would celebrate: roughly 3.3% real.

The age-85 line is the headline of this article. At roughly life expectancy, this annuity delivers about one percent per year in real terms. That is not a scandal. It is buying something no portfolio sells at any price: the complete removal of longevity and sequence risk. But it is nowhere near what the nominal numbers imply, and the space between the nominal picture and the real one is not a detail. It is most of the decision.

We lean toward weighting inflation risk more heavily than the industry default does. That is a house view, stated as one, and reasonable people weight it differently. Our reasoning: a 65-year-old is planning across a horizon that can plausibly run thirty years, recent memory has settled the question of whether inflation can still show up, and a fixed nominal payment is the most inflation-exposed instrument in a typical retirement. When one of two irrevocable options carries zero inflation defense, we would rather put that exposure on the table in numbers than nod at it in a closing caveat.

And the honest counterweights: if inflation averages under 3%, every real figure above improves and the annuity strengthens. Meanwhile the lump sum offers no inflation guarantee of its own, only inflation responsiveness, which is a weaker promise, and it arrives carrying the sequence risk the annuity eliminated. The lump sum is not a hedge. It is a claim on whatever future actually happens, which is different from a fixed payment, and also different from safe.

The Assumption Underneath Everything

Every number so far assumes the check arrives for life. That assumption deserves a look rather than a shrug, and this is the second place we lean differently than most.

The good news first, honestly stated. Corporate pension funding is strong right now. Wilshire put the aggregate funded ratio for S&P 500 plans at 108.7% at the end of June 2026, a genuine surplus, after improvement through most of the trailing two years. If your mental model of corporate pensions was formed watching 2008 through 2012, it is out of date.

Two caveats keep the subject open. First, an aggregate is not your plan. A 108.7% average blends plans well above and well below the line, and you are a participant in exactly one of them. Second, a large share of the improvement came from rates, not contributions. Higher discount rates shrink reported liabilities and flatter funded ratios without a new dollar entering the trust, and the same rate move that trimmed your lump sum offer polished your plan’s funded ratio. Rates falling would unwind both at once.

Behind private single-employer plans stands a real but capped federal backstop. For plans terminating in 2026, the Pension Benefit Guaranty Corporation’s maximum guarantee at age 65 is $93,477 per year for a straight-life annuity, which is $7,789.77 per month, and $7,010.79 per month for a joint-and-50% survivor form. The cap drops for benefits starting before 65 and rises for later starts. Most benefits fit under those ceilings with room to spare, and historically the large majority of participants in PBGC-trusteed single-employer plans have collected their full benefit, but an executive with a large pension should price the cap in dollars, not assume past it. Multiemployer plans, the union-industry kind, sit in a separate PBGC program with different and generally thinner guarantees, and those levels are not indexed.

Public pensions are a different conversation entirely, and it is where our concern actually lives. No PBGC stands behind a state or municipal plan. Their reported funding also rests on a valuation convention economists have been contesting for two decades: public plans discount their liabilities at the return they assume their assets will earn, commonly 6% to 7%, instead of at a rate that reflects what the liabilities are, which is close to guaranteed debt. Novy-Marx and Rauh made the case in the Journal of Economic Perspectives in 2009: already-promised state pension liabilities of $5.17 trillion against $1.94 trillion of assets, a $3.23 trillion hole invisible in headline state debt. Rauh and Giesecke’s follow-on work covering 2014 through 2022 found the hole deeper, not shallower: roughly $1.6 trillion unfunded on the plans’ own books, about 75% funded, versus roughly $5.1 trillion and under 50% funded at market discount rates, with unfunded liabilities growing by half through a decade of strong equity markets and rising contributions.

None of that predicts a missed check. Public pension promises usually carry state constitutional or statutory protection, and the political force behind honoring them is enormous. What it does say: the promise is optimistically valued, the shortfall is a draft on future taxpayers, and stressed plans sometimes make moves their participants feel immediately.

Texas wrote the case study. The Dallas Police and Fire Pension System went from 72% funded in January 2011 to 45% by January 2016, driven by an extraordinary portfolio, roughly 68% in alternatives and real estate against a national norm near 22%, sitting on top of chronic underfunding. When benefit cuts entered public discussion in 2016, participants in the plan’s Deferred Retirement Option Program started pulling balances, and the run took the funded ratio to about 35% by December, with projections then showing the fund exhausted around 2027. The legislature ultimately stepped in, and the Center for Retirement Research read Dallas as a story of exceptional decisions rather than an omen for public plans generally. The transferable lesson is smaller and sharper: the participants who treated “guaranteed” as a full sentence found out funded status could become their problem in months, and the ones who acted first were the ones already reading the reports.

So for anyone in a Texas public system, municipal, TRS, or otherwise, or any state’s plan: pull the funded ratio, note the assumed return discounting the liabilities, and read the actuarial valuation’s own projections. All public documents. They do not have to change your answer. They belong inside it.

Underfunding also carries a mechanical bite most people never see coming. Under Section 436 of the tax code, when a single-employer plan’s adjusted funding target attainment percentage falls below 80%, the plan may pay a lump sum only up to the lesser of half its value or the value of the PBGC-guaranteed benefit. Below 60%, or below 100% with the sponsor in bankruptcy, lump sums stop entirely. Read that as what it is: the self-directed exit gets rationed at exactly the moment the plan’s condition makes you want it. If your plan’s funding is soft, the lump sum window may be narrower than the packet implies, and the current AFTAP is worth confirming before assuming the offer survives into next year.

We lean toward treating funded status as a live input in this election, not scenery. Not because we expect defaults, and not as a claim about any particular plan or employer, but because the annuity’s entire case hangs on the word guaranteed, and that word is not equally strong everywhere it appears.

Two Insurance Policies Against Different Disasters

Take the arithmetic away and that is what these are.

The annuity insures against outliving money and against bad market order. It cannot run out and it cannot have a bad decade. As we wrote in our sequence-of-returns piece, the retirees who get hurt worst are not the ones who meet poor markets but the ones who meet them early while withdrawing, and the annuity simply cannot fail that way.

It is worth making that concrete, because the lump sum’s failure mode deserves the same daylight we gave inflation. Take our $625,000 rollover, withdraw the same $46,200 the annuity would have paid, and hand it ten years of returns: two losing years of 15% and 10%, and eight years of 7% gains. If the losing years come last, the account holds roughly $373,000 after the decade. If the identical returns arrive with the losing years first, it holds roughly $197,000, and continuing the same withdrawals puts it on a path to run dry around age 81, right about when the annuity’s hurdle rate was just getting interesting. Same returns, same withdrawals, different order, nearly half the money. That is the risk the annuity retires on day one, and it is why the guarantee is worth real money even at a modest real return.

The annuity also insures against a risk no packet names: you, later. Managing a portfolio at 65 and managing one at 87 are different jobs, and the check keeps arriving through cognitive decline, lost interest, and the persuasive relative with a sure thing.

The lump sum insures against inflation, early death, plan trouble, and rigidity. Rolled to an IRA it can grow, spend unevenly the way real retirements actually spend, absorb a roof or a wedding or a health event without permission, ignore whatever becomes of your former employer, and pass its remainder to your heirs. It also keeps tax doors open that the annuity nails shut, which the next sections cover.

Neither column wins on paper. They answer different versions of “what can you least afford,” and households honestly differ on that. If Social Security and other income already cover the essentials, the pension can reasonably be treated as investable wealth. If this pension is the difference between a funded retirement and an unfunded one, the floor can reasonably outrank everything else. Most families sit between, which is precisely why this is a modeling exercise and not a slogan contest.

The Survivor Decision Inside the Decision

Married retirees rarely face a binary. The menu reads: lump sum, single life, joint and 50% survivor, joint and 75%, joint and 100%, each step trading monthly income for protection. Federal law makes the joint and survivor form the default for married participants, and waiving it requires the spouse’s written consent, witnessed by a notary or a plan representative. The formality exists because the failure mode is severe: a single-life annuity ends at the first death, and pension income vanishing at the exact moment a household must reorganize has left a long trail of impoverished widows.

The survivor election also runs straight into a topic we’ve covered on its own: the widow’s tax penalty. Within a year or two of a spouse’s death, the survivor typically files single, with narrower brackets and a smaller standard deduction, while most of the household income keeps coming. A 100% survivor annuity delivers that income into the harsher structure. A lump sum leaves the survivor assets instead of income, more maneuverable against the squeeze, but it also hands full portfolio responsibility to a grieving spouse on the worst day for it.

There is a named middle path: take the larger single-life payment and spend part of the difference on life insurance for the surviving spouse, usually called pension maximization. It genuinely works sometimes. It fails badly when the coverage is underpriced in the illustration, underfunded in practice, lapsed at 82, or unavailable at reasonable cost because of health. It is a strategy to model in dollars against the joint-life election, lapse scenario included, never one to buy on the concept.

The Tax Fork

The paths separate at the first step, which is where the expensive errors happen.

The annuity: ordinary income, every year, for life. Simple, and permanently rigid. It sets a floor under taxable income that later stacks against Social Security taxation, the Medicare surcharges from our IRMAA piece, and required distributions from everything else.

The lump sum: tax-neutral only when moved correctly. A direct rollover to an IRA is a nonevent for current taxes and keeps full deferral. A check made out to you personally triggers mandatory 20% withholding and starts the 60-day clock, and completing the rollover means sourcing the withheld piece from other money. Cashing out entirely is almost never the plan: the whole amount stacks into one year’s ordinary income, running up through brackets and IRMAA tiers simultaneously. Practically, taking the lump sum means direct rollover to an IRA, where it lives like any other pre-tax balance: taxed on withdrawal, subject to required minimum distributions at 73, or 75 for anyone born in 1960 or later, and workable in the meantime.

Workable is the word payback math never finds. A rolled-over pension is raw material for the multi-year positioning we laid out in our gap-years piece: Roth conversions filling low brackets, income steered around subsidy and surcharge cliffs, the timing of income chosen instead of received. Electing the annuity surrenders that control permanently as part of the price of the guarantee. Some households should happily pay it. For others, the surrendered optionality is the single most expensive line in the deal, and it appears on no worksheet the plan will ever send.

How We Run the Decision

When we model this election for a family, it looks less like comparing two figures and more like stress-testing two entire retirements.

The starting point is the spending plan, not the pension. What do essentials cost, what already covers them, how much guaranteed lifetime income exists through Social Security? An annuity is at its best completing an income floor and at its worst duplicating one. The decisions also interlock here: delaying Social Security purchases a large, inflation-adjusted lifetime annuity on terms private markets struggle to match, and it is the one guaranteed income source in most plans that actually keeps pace with prices. A household planning to delay may simply need less fixed income from the pension.

From there we price the actual offer: hurdle rates across realistic lifespans, nominal and real, for every form on the menu, tested against the household’s real portfolio and real risk capacity instead of a stock assumption. We check the plan’s funded status and, where it matters, its Section 436 posture. We run the taxes both ways across a multi-year window with survivor scenarios included, because the version of this decision that ends up mattering is rarely the joint-life version. And we give honest weight to what numbers cannot hold: the value one spouse places on never watching a market again, the other’s realistic capacity to manage a seven-figure rollover at 85, and what the family wants left behind.

We lean away from both of the blanket answers this decision collects. “Take the lump sum, you’ll beat it in the market” values mortality pooling at zero and assumes a return sequence nobody is promised. “Take the annuity, guaranteed income wins” values inflation erosion, survivor tax dynamics, funding risk, and decades of forfeited tax flexibility at zero. Both slogans are free, and priced accordingly. This is a one-shot, irrevocable election over what is often one of the largest assets a household will ever hold, and it is exactly the kind of decision to run on your own numbers before anyone in the room, including you, falls in love with an answer.

Common Questions

How do I find out whether my pension plan is well funded?

Private single-employer plans must send an annual funding notice stating the funded percentage, and the plan’s Form 5500 is public. The number that controls lump sum availability is the adjusted funding target attainment percentage, or AFTAP, available from the plan administrator. Public plans publish an annual actuarial valuation showing the funded ratio and the assumed rate of return used to discount liabilities. Read that second number too: a lower assumption generally marks a more conservative and more believable valuation.

Is there deadline pressure I should know about?

Frequently. Lump sum windows tied to a retirement date, and one-time buyout offers to former employees, come with election deadlines. Because plans reset calculation rates annually, an offer straddling a year boundary can change size when the new rates arrive. And deteriorating funding can trigger Section 436 restrictions that shrink or remove the lump sum option going forward. None of that should rush a decision of this size. All of it says start the analysis when the packet arrives, not the week the form is due.

I am already receiving pension payments. Can I still be offered a lump sum?

Possibly. The IRS effectively ended these retiree buyout windows in Notice 2015-49, then reversed itself in Notice 2019-18, announcing it no longer intends to amend the regulations to prohibit them and will not treat them as violating the required minimum distribution rules while it keeps studying the issue. So sponsors may offer lump sum windows to retirees in pay status, and some do, to shed liabilities and PBGC premiums. If one reaches you, this article’s framework applies with one change: your hurdle rate runs from your current age rather than 65, which generally pushes it higher.

What happens to my choice if my former employer goes bankrupt or hands the pension to an insurance company?

If you took the lump sum: nothing. The money left the plan, and the plan’s fate stopped being yours. If you took the annuity, the payments continue but the backstop can change. A failed single-employer plan is trusteed by the PBGC under the 2026 caps above. A pension moved to an insurer through a group annuity purchase is thereafter backed by that insurer and, behind it, your state’s guaranty association limits rather than the PBGC. Those backstops hold comfortably for most benefit sizes. For an unusually large pension, check the limits in dollars before you elect.

Can I take part lump sum and part annuity?

Some plans allow a split, and where offered it deserves a hard look. The annuity portion can build the income floor while the rolled-over portion keeps growth, inflation responsiveness, legacy, and tax optionality alive. Availability and proportions are entirely plan-specific: check the summary plan description and ask directly, because election packets do not always advertise the option.

How does this interact with when I claim Social Security?

They are two halves of one income-floor question, best answered together. Delayed Social Security is additional guaranteed lifetime income that is also inflation-adjusted, which makes it a better instrument for floor-building than a fixed pension annuity. A household planning to delay may need less fixed income from the pension, strengthening the lump sum case; a household claiming early may value the pension floor more. Modeling the two jointly, across lifespans and survivor scenarios, routinely lands somewhere neither decision reaches alone.