Pension Risk Transfer 2026: Bid Economics on Standard vs. Complex Populations

Standard US pension buyouts now price at 100.1% of accounting liability, against a 103.4% all-insurer average, so a structured competitive process is worth 3.3% (Milliman, April 2026). On clean retiree-only populations that compression is essentially finished. On plans carrying disabled lives above 15%, non-standard benefit forms or unresolved data gaps, valid bids come back 200 to 400 basis points apart, and the 3.3% is not available at all.

Key Takeaways

• 100.1% of ABO on competitive bids against 103.4% across all insurers. The 3.3-point gap is $16.5 million on a $500 million plan, and it exists only where the population is standard enough for insurers to price the same risk.
• 50 to 100 basis points is the bid spread on a clean retiree population; above 20% disabled lives it runs 200 to 400, because the assumption set stops being credible on plan-specific data.
• Duration-7 annuity purchase rates hit 4.94% and duration-15 5.02% in May 2026, the highest since June 2025, with the short-duration spread at -0.18%: pricing marginally below the accounting discount rate.
• More than 20 life insurers quote US business, roughly double a decade ago, but a plan with mixed actives, over 20% disabled lives and amended benefit histories may draw four to six responses.
• Benefits above the PBGC maximum guarantee stay an employer obligation whether or not a group annuity is purchased, so part of the liability does not transfer at any price.

What the 3.3% Is and Where It Comes From

The Milliman index carries two numbers. Competitive-bid retiree buyout cost fell to 100.1% of the accumulated benefit obligation, while the average across all insurers stood at 103.4%. The gap is the quantified value of running a structured process rather than a single-carrier negotiation: on a $500 million plan, $16.5 million.

It does not arrive passively. It reflects inviting five or more qualified insurers, standardizing data and benefit documentation so bids are comparable, and allowing each pricing team enough calendar time. Mercer's global index puts the same discipline at 3% to 4%.

Rates are cooperating. October Three put duration-7 annuity purchase rates at 4.94% and duration-15 at 5.02% in May 2026, each the highest since June 2025, with the short-duration spread turning slightly negative at -0.18%. Annuity pricing below the accounting discount rate is an unusually favorable signal for a standard retiree lift-out.

Volume confirms the depth: at or above $45 billion for four consecutive years, with $48.8 billion in 2025 splitting $31.3 billion of buyouts and $17.5 billion of buy-ins. The insurer field doubling over a decade is what compressed standard pricing to near book.

Credibility Is Why the Saving Disappears

On a clean retiree population, every insurer is pricing materially similar longevity exposure off industry pensioner tables and its own improvement scale. The inputs are well characterized, the data is abundant, and valid bids land within 50 to 100 basis points. That is how the market reaches 100.1% of ABO.

Disabled-life pricing is a different actuarial problem. Each insurer forms its own working assumption for the mortality and comorbidity profile of that specific population, from case data that almost never reaches full credibility. Expected mortality relative to standard pensioner tables, comorbidity-adjusted life expectancy by disability type, and claim continuation rates in the first years after transfer all vary across pricing models.

Where the plan-specific data cannot credibilize the assumption, judgment carries the weight, and judgment diverges across twenty underwriting teams. That is the whole mechanism behind a 200 to 400 basis point spread on a plan where disabled lives exceed 15% to 20%. It is not a failure of competition; it is competition pricing an assumption rather than a table.

Population Characteristic	Typical Valid Bid Count	Approximate Bid Spread
Standard retirees, clean data, flat-life annuities	15–20	50–100 bps
Disabled lives 15%–20% of population	8–12	150–250 bps
Disabled lives above 20% of population	5–8	200–400 bps
SS level income options or early retirement subsidies	8–14	100–200 bps
Significant data quality deficiencies (missing beneficiary, term dates)	4–8	150–300 bps
Combined complexity (disabled lives + non-standard forms + data gaps)	2–5	300–500 bps

Data quality moves the same lever without changing the risk at all. A plan going to market with incomplete termination dates, missing beneficiary designations or unreconciled benefit calculation histories draws fewer valid bids and wider spreads, because each insurer prices its own uncertainty about a population it cannot characterize. Remediation on moderately deficient populations typically recovers 50 to 150 basis points, which on a $200 million transaction is $1 million to $3 million against an advisory cost that rarely approaches it.

Non-standard forms narrow the field rather than widening the spread. Social security level income options need pricing infrastructure some insurers have built and others have not, so a process sent to twenty insurers on a plan where 30% of participants hold that option may return valid pricing from 8 to 12.

The Insurer Count Describes a Plan Most Complex Cases Are Not

Aon's count of more than 20 active US insurers is accurate for clean populations, full data packages, conventional flat-life annuity forms and standard benefit histories. It does not describe the field for anything else.

A plan terminating 2,500 retirees on standard forms with complete history may draw 15 or more valid bids. The same premium quantum on 2,500 participants with a mixed active and retiree population, disabled lives above 20%, early retirement subsidies from a closing facility, and a benefit history spanning three formula amendments across two prior acquisitions may draw four to six, not all of them competitive across every benefit form. The competitive dynamics at ten bidders are materially different from those at twenty, and the sponsor who priced its expectations off the headline count meets that difference in the first round.

What survives the transaction is the sharper constraint. Benefits payable above the PBGC maximum guarantee ceiling remain an employer obligation outside the insurance framework regardless of whether a group annuity is purchased. A bid package that includes those excess benefits without explicit population segmentation supports two incompatible readings: the insurer has priced the full benefit, or the sponsor intends to retain the excess. That gap does not resolve itself at close, and it surfaces at participant communication, where individual notices have to state what is insured and what is not.

Against that, the PBGC premium clock is what a completed transaction actually stops. The 2026 flat-rate premium is $111 per participant, up from $106 in 2025, $86 in 2021 and $35 in 2012, with no statutory ceiling; the variable rate adds $52 per $1,000 of unfunded vested benefits, against $9 per $1,000 in 2012. On 1,000 participants that is $111,000 a year of pure cost, and a plan 95% funded on a vested basis with $5 million unfunded pays another $260,000. The premium stream is recurring and rising. The excess-benefit obligation the buyout leaves behind is neither, which is precisely why it goes unnoticed.