Milevsky Ch. 4: Longevity Risk Pooling — What an Annuity Actually Buys You

"The mortality credit is the only free lunch in retirement finance, and it is served with a very long list of side dishes you have to pay for." — the honest version of the pooling argument

Investment Background

There is exactly one thing a lifetime income contract does that a portfolio cannot do at any price, and understanding it precisely is the difference between using these products well and being sold them badly.

That thing is the mortality credit. When a group of 80-year-olds pools capital and agrees that the pool pays income to whoever is still alive, the capital of those who die early is redistributed to those who live long. Nobody's investment return improved. No manager was skilled. The extra income comes from a source unavailable to an individual: other people's deaths.

The size of that credit is what surprises people. At 65 the mortality credit embedded in a lifetime payout is small — worth perhaps 1% a year over a bond of similar duration. At 80 it is substantial, plausibly 4–5% a year. At 90 it dominates the payout entirely. This is the direct consequence of the compounding hazard rate from Chapter 2, and it is why the when of pooling matters at least as much as the whether.

Milevsky's framing is that this is genuinely a form of insurance, not an investment, and should be judged as insurance. You do not buy homeowners insurance expecting a good return. You buy it because the uninsured outcome is unacceptable. Lifetime income is insurance against the specific outcome of outliving your money, and it is priced accordingly — which means it will usually look like a bad deal on a spreadsheet that assumes you die on time.

The decision this chapter is about is therefore not "annuity: yes or no." It is: what portion of my essential spending, if any, is currently exposed to my own longevity, and is pooling the cheapest way to close that gap?

The Wall Street Translation

What the Pool Buys, and What It Costs

What you get	What you give up
Income that cannot be outlived	Liquidity — the capital is gone and irreversible
Mortality credits — extra yield from pooling	Bequest — most of the capital dies with you
Removal of sequence risk on that slice	Purchasing power, if the payout is nominal
Simplicity — no decisions required at 90	Counterparty exposure to a single insurer
Behavioral relief — spending guilt disappears	Fees embedded in the payout, often opaque

Both columns are real. The industry sells the left and buries the right; the anti-annuity camp does the reverse. Neither is being honest.

The Costs, Stated Plainly

Inflation erosion is the largest and most underappreciated. The overwhelming majority of income annuities sold in the U.S. are nominal — the payment never rises. At 3% inflation, a $2,000 monthly payment has $1,100 of today's purchasing power after 20 years, and about $830 after 30. A retiree who annuitizes their essential floor at 70 with a nominal contract may find that at 92 the contract covers barely half the floor it was bought for. Inflation-adjusted versions exist but start roughly 25–30% lower, and most buyers reject them for exactly that reason — which is choosing the version that fails slowly over the version that starts small.

Irreversibility is not a footnote. A single premium immediate annuity has no surrender value in any meaningful sense. If you are diagnosed with a serious illness at 74, three years after annuitizing, you cannot undo it. If you need $80,000 for a medical or family emergency, that capital is not available. This is the reason no household should annuitize an amount that leaves them without substantial liquid reserves.

Insurer credit risk is real, if usually modest. The payment stream is a general obligation of one insurance company for potentially thirty years. State guaranty associations provide backstops, but the limits are typically in the $250,000–$300,000 range of present value per insurer per state, and they vary. A retiree placing $600,000 with one carrier is holding uncompensated concentration risk.

Fees are embedded and hard to see. A plain immediate annuity's cost shows up as a lower payout than the pure actuarial value — commissions, expenses and profit margin, often equivalent to several percent of premium. That is tolerable for a simple contract. It becomes serious in variable and fixed-index annuities with lifetime withdrawal riders, where all-in costs of 2.5–3.5% a year are common and the "guarantee" applies to a benefit base rather than to money you can withdraw. Those products are a different animal and deserve far more skepticism than a plain payout contract.

Bequest is materially reduced. Capital committed to a lifetime pool largely does not pass to heirs. For some households that is irrelevant; for others it is the whole point of the estate. It must be a stated decision, not a discovered consequence.

Why This Is Not a Portfolio Replacement

The single most important framing in this chapter: pooled income is a hedge against one specific risk, applied to one specific slice of the balance sheet. It is not a substitute for the investment portfolio and it is not a strategy for growing wealth.

The reason is structural. A lifetime payout contract is backed by the insurer's general account, which is overwhelmingly investment-grade bonds. Its long-run real return, net of the mortality credit, is bond-like. That is entirely appropriate for the floor — money whose job is to arrive reliably, not to grow. It is entirely inappropriate for the growth engine, whose job over a thirty-year retirement is to outrun inflation, and which is done best and cheapest by a broad, low-cost, globally diversified index portfolio.

The right mental model is a barbell: a modest floor of pooled, market-independent lifetime income covering essential spending, and a large, cheap, indexed core carrying everything above it. Annuitizing the whole balance sheet does not make you safe; it makes you a bondholder with no inflation protection and no liquidity, which is a different way to fail.

Execution Rules

1. Only ever pool the essential-floor gap, and never more than 25–30% of investable assets. Compute essential spending, subtract Social Security and pension, and consider pooling only that remainder. If the gap is zero, you do not need an annuity regardless of how good the pitch sounds.

2. Exhaust Social Security delay first — it is the best annuity available. Delaying from 62 to 70 buys roughly 76% more inflation-indexed lifetime income, backed by the federal government, with zero commission and full CPI adjustment. No commercial product matches it on any dimension. Only after this is fully used should a commercial contract be considered.

3. Restrict yourself to plain vanilla contracts you can price-compare. Single premium immediate annuities and deferred income annuities have one number that matters — the payout per $100,000 — which you can quote across five carriers in an afternoon. Variable and fixed-index annuities with living-benefit riders cannot be compared this way, carry 2.5%+ all-in costs, and should be declined by default.

4. Split across at least two highly rated insurers, and stay near guaranty limits. Use carriers rated A or better, and size each contract so that its present value sits within your state's guaranty association coverage where practical. Concentration in one carrier for thirty years is an unpaid risk.

5. Price the inflation-adjusted version even if you reject it, and never annuitize your liquidity. Get quotes on both nominal and CPI-adjusted payouts so the erosion is a decision you made with numbers in front of you. And keep at least two years of spending plus a dedicated medical and long-term care reserve entirely outside any annuitized capital.

Retirement Application

Take a household with $1.4 million invested, $54,000 of combined Social Security, and $76,000 of essential annual spending. The uncovered essential gap is $22,000.

The complete answer is not a $400,000 annuity. It is a sequence. First, delay the higher earner's benefit to 70, which might raise combined Social Security to $66,000 and shrink the gap to $10,000 — closing more than half the exposure with the cheapest, fully inflation-indexed instrument available. Second, wait. At 66 the mortality credit is small, and the household's health and spending pattern are still uncertain. Third, revisit at 75–78, when the same income costs meaningfully less capital, and if the gap still matters, pool roughly $130,000–$160,000 across two carriers to close it.

The remaining $1.25 million stays exactly where it belongs: a broad, low-cost global index core funding discretionary spending, inflation protection and any bequest. Nothing about buying a floor changes the case for indexing the rest — if anything it strengthens it, because a household whose essentials are secured can hold equities through a bear market without being forced to sell.

Risk Management

• Do not annuitize into a rate trough. Payouts track interest rates. Annuitizing the entire floor in a single transaction locks a lifetime of one day's rates. Ladder purchases across two or three years, or across ages 75, 78 and 81, to diversify rate timing.
• Watch the joint-life decision carefully. A single-life contract on one spouse leaves the survivor exposed. Joint-and-survivor at 100% costs more per dollar of initial income but eliminates the widow's-cliff failure mode, which is one of the most common real-world retirement disasters.
• Beware the pitch that arrives with a free dinner. Commission on complex annuity products routinely runs 5–7% of premium, which is a powerful incentive to sell the complicated version. Any recommendation you did not seek out deserves a fee-only second opinion.
• Health matters and the pricing does not reflect yours. Standard annuity pricing assumes an above-average-health buyer. If your health is genuinely impaired, you are subsidizing the pool and should generally not annuitize; medically underwritten contracts exist but are uncommon in the U.S.

What This Chapter Cannot Do

This chapter cannot tell you whether to buy a lifetime income contract. That depends on your essential floor, your existing lifetime income, your health, your bequest intentions, your spouse's situation, your state's guaranty limits and current rates — and one of those changes every quarter.

It also cannot make the products cheap or transparent. Commercial annuities carry real costs, and the U.S. market is dominated by complex products whose pricing is deliberately hard to compare. Nothing here should be read as an endorsement of the industry's sales practices.

And it makes no case against indexing — the opposite. Pooling exists to protect the floor precisely so that the growth portfolio can stay invested, stay cheap, and stay indexed. Any presentation that uses longevity risk as an argument for moving the whole portfolio into insurance products has inverted the argument.

---

Key Takeaway: An annuity buys mortality credits — extra income funded by the pool's early deaths — and nothing else that a portfolio cannot provide more cheaply. Pool only the essential-floor gap, exhaust Social Security delay first, insist on plain contracts split across highly rated insurers, accept the inflation, liquidity and bequest costs explicitly, and keep the rest of the balance sheet in a low-cost index core.