Pension Lump Sum vs. Annuity: Which One Is Better?

Neither is better on its own. A lump sum is meant to be the present value of the monthly annuity you were already promised — that’s the legal floor, though most plans pay exactly the floor and no more. What actually decides it is your spouse, whether the payments adjust for inflation, and the interest rates in force the month your offer is calculated.

Everything below is how each of those works. I am not going to hand you a rule, because the honest answer changes with facts a rule can’t see.

How is a pension lump sum actually calculated?

The plan takes the monthly benefit it owes you, projects it across a life expectancy, and discounts it back to today’s dollars. Two inputs do the work: a mortality table and an interest rate.

Both are prescribed. Federal law sets a floor — the present value of your benefit “shall not be less than the present value calculated by using the applicable mortality table and the applicable interest rate” (26 U.S.C. §417(e)(3)). The interest rate is not one rate but three segment rates, published monthly by the IRS.

A plan can be more generous than that floor. Most are not. So in practice, your lump sum is a discounted-cash-flow calculation with someone else’s assumptions, and the only variables that move it are your age, your monthly benefit, and the rates.

Why did my lump sum offer drop when interest rates went up?

Because a discount rate and a present value move in opposite directions. The higher the rate used to discount your future payments, the less those payments are worth today. Nothing about your pension changed. The arithmetic did.

This matters more than it sounds, because plans lock rates in advance. A plan has to specify a lookback month, which may be any one of the first five full calendar months before the stability period — the stretch during which the rate stays fixed (Treas. Reg. §1.417(e)-1). A stability period can be as short as a month or as long as a year.

So a lump sum quoted in January and one quoted in July can be calculated off rates from very different months. Two facts determine what any particular quote means: which month’s rates it was built from, and the date that calculation expires. Neither is obscure — the plan document specifies both.

The monthly annuity, meanwhile, does not move with rates at all. It is the fixed side of the trade.

What is the break-even point, and what does it leave out?

Divide the lump sum by the annual pension. That gives you the number of years of payments it takes to get your money back with no growth, no inflation and no tax.

Illustrative only. These figures are made up to show the arithmetic, not drawn from any plan.

• Lump sum offered: $900,000
• Monthly pension: $4,500, which is $54,000 a year
• $900,000 ÷ $54,000 = 16.67 years, or 16 years and 8 months

That is the entire calculation. It is the only number in this decision computable in ten seconds, and it is also the crudest.

What it ignores: what the lump sum would earn, what inflation does to a fixed check, the tax on either path, and — the largest omission — what a surviving spouse receives. I am not going to fill those gaps with a projected rate of return, because a return I made up would drive the whole answer. So break-even locates the offer in a neighborhood; the sections below are what it cannot see.

There is a second number that takes about as long to get and says considerably more. Ask an insurer what it would charge you today for an immediate annuity paying the same monthly amount, for the same lives, in the same form. That quote is what the open market prices your plan’s offer at. If the lump sum is comfortably larger than the quote, the plan is offering you more than the income is worth to buy; if it is smaller, the plan’s income is the better-priced side. It turns an unanswerable “which is better” into two figures you can set beside each other, and it uses somebody else’s mortality and interest assumptions rather than one I invented. It is not the whole decision — the sections below are still what break-even cannot see — but it is the closest thing here to a price.

Why can a lifetime check out-pay the same money invested?

There is a reason the annuity side can pay out at a rate a portfolio cannot safely match, and it is not investment skill. It is pooling.

A pension or an insurer is not funding your payments alone. It is funding payments for a whole group of people the same age, out of one pot, and it knows with reasonable accuracy how many of them will still be alive in each future year. Every person in the pool who dies earlier than average leaves behind money that was reserved for payments never made, and that money funds the checks of the people who live longer than average. Actuaries call the difference this makes mortality credits — the extra return, above interest and investment growth, that a survivor earns simply by having survived.

A portfolio you manage yourself has no pool. It is funding one life, so it has to be sized for the long tail of that single life rather than the average of many. That is the structural reason a self-funded withdrawal rate is set well below what an income contract for the same money will quote: you are carrying alone the risk the pool spreads across thousands.

Two things about that mechanism are worth holding onto. Mortality credits are small early and grow steeply with age, because they only exist to the extent the pool has thinned — which is why the advantage of pooled income is mostly an advantage in your eighties and nineties, not your sixties. And they are paid for: the money that funds them is the money you do not leave to anyone. Pooling and bequest are the same dollars viewed from opposite ends, so the value of a mortality credit depends entirely on which of the two you were going to want.

I am describing the mechanism, not pointing at an answer. Whether pooled income is worth its price in your case turns on the survivor terms, the inflation treatment, the tax path and what else already covers your fixed spending — all of which are below.

What does my spouse actually get, either way?

This is usually the real decision, and it is the one most often described badly. You will see it framed as security versus flexibility. That framing is wrong, and it is wrong in a way that costs money.

Choosing a lump sum over a joint-and-survivor annuity is not choosing flexibility. It is a transfer of risk onto the surviving spouse. Two specific risks: the risk of outliving the money, and the risk that whoever manages the account after you are gone is not you.

Here is the mechanism. A qualified plan’s default form for a married participant is a qualified joint and survivor annuity — a survivor benefit “not less than 50 percent of (and is not greater than 100 percent of)” the amount payable while you are both alive (26 U.S.C. §417(b)(1)). Electing anything else, including a lump sum, requires your spouse’s written consent, acknowledging the effect of the election and witnessed by a plan representative or a notary public. The election window is the 180 days ending on the annuity starting date (26 U.S.C. §417(a)).

That signature requirement exists because Congress decided this was a decision one spouse should not make alone. The consent form is the document on which a lifetime income guarantee for the survivor is relinquished, which is what it is designed to make visible.

None of which settles the question in the annuity’s favor. A rolled-over lump sum is inheritable and a pension generally is not — the check stops, or steps down to the survivor and then stops. Where leaving something behind is a goal, the whole job of legacy sits with the portfolio. The size of the trade is knowable in advance: plans quote a monthly amount under each survivor percentage they offer, and the gap between the single-life figure and the joint-and-survivor figure is the price of the guarantee in dollars per month.

If it helps to have the numbers in one place, I keep them all in the Ultimate Retirement Guide.

Does the pension have a cost-of-living adjustment?

Whether it does changes the comparison more than almost anything else on this page, and the plan document is where the answer is written.

A fixed pension pays the same nominal dollars in year one and year thirty. What it buys in year thirty is a different question, and the answer is uncomfortable. Most corporate plans have no cost-of-living adjustment. Public-sector and union plans sometimes do, sometimes capped, sometimes discretionary rather than promised.

If the payments do adjust for inflation, the annuity is far harder to replicate and break-even understates it. If they don’t, the comparison is a nominally growing pile against a check losing purchasing power each year. A plan that says it “may” grant an increase has not promised one.

Here is the framing I keep coming back to: a pension is a bond you never had to buy. It does the job bonds do in a portfolio — steady, predictable income that holds up when markets are down — without occupying a dollar of space in it. Which means the pension’s presence changes what the rest of the portfolio is being asked to do. A fixed pension resembles a bond with no inflation protection. A pension with a contractual cost-of-living adjustment resembles an inflation-linked one. Those are not the same instrument, and the difference sits at the center of this comparison.

What happens if my employer fails?

Single-employer defined benefit plans are insured by the Pension Benefit Guaranty Corporation, a federal agency. If the plan terminates without enough money, PBGC steps in as trustee and pays benefits — up to a maximum.

The maximum is set at age 65 in the year the plan terminates — reduced if payments start earlier, and increased if you wait past 65 to start them, so this is not a one-directional adjustment. It is also reduced if the form includes a survivor benefit. Two more limits are worth knowing. PBGC does not pay cost-of-living increases on the benefits it takes over. And a benefit increase that has been in effect for less than 5 years when the plan terminates is only partly guaranteed — the guarantee phases in over those years rather than applying in full from the start (29 CFR §4022.25, “Five-year phase-in of benefit guarantee”).

Multiemployer plans are covered under an entirely different and far lower formula, based on years of credited service rather than a dollar maximum. If yours is a union plan, the single-employer numbers do not describe your coverage.

For most people the guarantee comfortably covers the whole benefit, and employer credit risk does not enter the comparison at all. Where a benefit is large enough to run past the maximum and the sponsor’s finances are in question, part of the promised income sits outside the guarantee — which is one of the few places where the two options differ on something other than tax and timing.

How are the two options taxed?

Differently in timing, not in character.

Annuity payments are ordinary income as received. If you made after-tax contributions, part of each payment is a tax-free recovery of that cost — the IRS Simplified Method spreads it over your expected payments. Once the cost is recovered, everything is taxable.

A lump sum rolled directly to an IRA is not taxable at the time of the rollover. Nothing is triggered, nothing is reported as income. But note directly: if the plan pays the money to you instead of the IRA, it must withhold a share for income tax, and you have to make that withheld amount up out of pocket to complete the rollover.

The rollover doesn’t erase the tax. It relocates it. That IRA becomes subject to required minimum distributions starting at 73 if you were born between 1951 and 1958, or at 75 if you were born in 1960 or later (Treas. Reg. §1.401(a)(9)-2(b)(2)(iv)–(vi)). Anyone born in 1959 sits in a gap the regulation has not resolved. If you were born in 1950 or earlier, this is moot — you are already past the start age and RMDs on that IRA have already begun.

One more asymmetry, and it is a state one. Everything above is federal. States do not all treat pension income and IRA withdrawals the same way, and several exempt some or all of a pension while taxing an IRA distribution in full. Where that is true, rolling the pension into an IRA can permanently give up a state exclusion the pension carried and the IRA does not. It is worth reading your own state’s rule before the election, not after, because the rollover is the step that cannot be undone.

And a creditor one. A benefit sitting in an employer plan carries ERISA’s own protection: “Each pension plan shall provide that benefits provided under the plan may not be assigned or alienated” (29 U.S.C. §1056(d)(1)). An IRA’s protection is a different and more conditional thing. In bankruptcy, rollover money and its earnings are left out of the dollar cap Congress put on IRA exemptions, so that route is well covered (11 U.S.C. §522(n)). Outside bankruptcy there is no federal rule at all, and what a creditor can reach comes down to your state’s exemption statute. For most households this never comes up. For someone in a profession with real liability exposure, it is a live difference and it changes at the moment of the rollover.

There is a symmetry people miss here. Annuity payments are the required distribution from that plan — a life annuity satisfies the distribution rules on its own (Treas. Reg. §1.401(a)(9)-6). Roll to an IRA and the required withdrawal becomes a separate event on a separate schedule, sized by a balance that grows until it starts.

Which path is cheaper in tax turns on whether the low-income years fall before or after the pension would have started. A pension fills the lowest brackets before a dollar of personal savings is touched. In the years before it starts, those brackets are as empty as they ever get.

What makes one household’s answer differ from another’s?

The facts that matter are not evenly distributed, and they pull in opposite directions.

Facts that increase what the annuity is doing: fixed spending that no other income source covers, a spouse with little income of their own, a contractual cost-of-living adjustment attached to the benefit, a preference against managing a large portfolio deep into old age, and health and family longevity that lengthen the payment horizon. Each of these raises the number of checks the annuity is expected to write and the cost of replicating them.

Facts that reduce what the annuity is doing: a baseline already covered by Social Security and other income, no survivor to protect, a benefit large enough to run past the PBGC guarantee where the sponsor’s finances are in question, a health condition that shortens the horizon, and a legacy goal that a stream of payments cannot serve. Each of these shrinks the job the guaranteed income was there to do.

Most households have some of both, which is why the arithmetic often lands close. And a two-pension household is making two of these comparisons, not one.

What do people get wrong about this?

• Running break-even and stopping. It ignores inflation, tax, and the survivor entirely.
• Calling the lump sum “flexibility.” It is flexibility plus a risk transfer, and only one of those gets mentioned.
• Assuming the offer is a fixed quantity. It is a calculation off a specific month’s rates, with an expiration date.
• Comparing a single-life pension to the lump sum. If you are married, the joint-and-survivor figure is the real comparison.
• Ignoring the cost-of-living adjustment question. It is the difference between two different products.
• Treating the pension as a reason to invest conservatively. It is already doing the job bonds do, so it changes what the rest of the portfolio is being asked to do rather than adding a constraint on top of it.
• Deciding in the ten minutes HR gives you. The election period is measured in months, not minutes — and you can revoke an election and make a new one at any point during that window, so the first form you sign is not necessarily the last word.

When this does not apply to you

• If your plan offers no lump sum. Plenty don’t. The decision is which annuity form, and the survivor section is the part that matters.
• If the pension is small relative to your assets. At a certain ratio it stops driving anything and the effort belongs elsewhere.
• If you are single with no beneficiary you care about. Most of the survivor analysis drops out and this becomes closer to a pure break-even and longevity question.
• If your benefit is in a multiemployer plan. The insurance limits above describe a different system than yours.
• If your spending is already covered without the pension either way. Then this is an estate question wearing an income question’s clothes.
• If the annuity has a genuine, contractual cost-of-living adjustment. Replicating an inflation-adjusted lifetime income stream out of a portfolio is a materially harder problem than replicating a fixed one, and the two sides of the comparison rarely stay close.
• If the rollover would not in fact be managed. The lump sum’s case rests on the money being invested and drawn down deliberately; where that is not what would happen, the comparison is not the one described here.

The pension decision is really a decision about what your baseline income looks like for the rest of your life, and everything else gets priced off it. If you want the layer above this one — which accounts to spend from once the income floor is set — that is withdrawal strategy.