LIMRA's second-quarter survey puts total U.S. annuity sales at a record $123.9 billion, registered index-linked annuities at a record $23.3 billion, and single premium immediate annuities at a record $4.0 billion, up 12% year over year (LIMRA, July 2026). The RILA record is a crediting-rate story. The SPIA record is a mortality and interest-rate story, and it is the one that resets payout-rate and longevity assumptions for the record accumulation book still waiting to convert.
The Product Mix Behind the Record Quarter
Every major annuity product category grew or set a record in the second quarter, which is itself unusual. LIMRA's release attributes the breadth to "a combination of global tensions, market volatility and rising interest rates" that lifted demand across products simultaneously rather than rotating share from one category to another (InsuranceNewsNet, July 2026). Fixed-rate deferred annuities remained the largest single category at $44.7 billion, down a modest 2% year over year, while fixed indexed annuities fell 7% to $30.7 billion. Traditional variable annuities climbed 25% to $17.9 billion. The two payout categories, smaller in dollar terms but structurally distinct from everything else on the list, both set records: SPIA at $4.0 billion and deferred income annuities at $1.3 billion, up 32% from the first quarter.
| Product | Q2 2026 sales | YoY change | QoQ change |
|---|---|---|---|
| Total annuity sales | $123.9B | +4% | — |
| Registered index-linked annuity | $23.3B | +22% | +11% |
| Fixed-rate deferred annuity | $44.7B | -2% | +26% |
| Fixed indexed annuity | $30.7B | -7% | +14% |
| Traditional variable annuity | $17.9B | +25% | +4% |
| Single premium immediate annuity | $4.0B | +12% | +9% |
| Deferred income annuity | $1.3B | +5% | +32% |
"Total annuity sales set an all-time quarterly record, the 11th consecutive quarter above $100 billion" (Bryan Hodgens, LIMRA, July 2026). Year to date, sales have reached $231.3 billion, 2% above the first half of 2025 and a new first-half record on top of the $464.1 billion full-year record LIMRA reported for 2025 (LIMRA, 2026). RILA's own record drew a separate quote from LIMRA's annuity research lead: "RILA set another quarterly sales record as carriers continued to pivot toward these products and broaden their distribution" (Keith Golembiewski, LIMRA, July 2026).
Why the Income-Annuity Record Outweighs the Accumulation Headline
RILA, fixed indexed, and fixed-rate deferred products share a common actuarial structure: the insurer sets a crediting mechanism, a cap, a buffer, or a guaranteed rate at issue, and the liability tracks an index or a declared rate over a defined deferral period. The mortality assumption barely enters the pricing at all; the product is a wealth-accumulation vehicle with an insurance wrapper, and the carrier's exposure runs through hedge cost and crediting-rate risk, not longevity. That is why a 22% RILA surge and a 25% variable annuity jump are, actuarially, extensions of an options-pricing and asset-liability-matching problem the industry has run at scale for years.
SPIA and DIA sales are a different animal. The moment a buyer annuitizes, the insurer prices a lifetime income stream off two inputs set at that instant: the current yield curve and a mortality assumption selected once and then lived with for the life of the contract, often decades. A record SPIA quarter is a record quarter of underwriters committing to brand-new longevity assumptions at scale, at a point in the rate cycle when the 10-year Treasury yield sat at roughly 4.63% (Federal Reserve Bank of St. Louis, July 2026), materially above the sub-2% environment that shaped payout-rate assumptions through most of the 2010s. That combination, a higher discount rate meeting a growing volume of new payout business, is exactly the environment in which a stale or under-adjusted mortality basis costs the most money, in either direction.
Payout-Rate Mechanics: Yield and Longevity, Not a Hedge Budget
A SPIA's quoted payout rate is, mechanically, the reciprocal of the net single premium per dollar of annual income: äx = Σ vk · kpx, where v is the discount factor derived from current rates and kpx is the probability the annuitant survives to year k. Every basis point of yield the insurer can lock in on the backing bond portfolio lowers äx and raises the quoted payout rate for the same premium, which is the direct channel through which a 4.6% Treasury yield, versus a 2% yield, makes today's SPIA quotes look far more attractive than they did five years ago. That channel is also why LIMRA's release ties rising rates directly to demand: buyers are locking in income streams while the discount side of the pricing equation is doing them a favor.
The mortality side moves the other direction and dampens the effect. This publication's coverage of the SOA Research Institute and LIMRA's 2020-2024 Individual Payout Annuity Mortality Experience Study, spanning 3.1 million contract-years across 26 companies, found a survivor-selection effect in the 2023-2024 tail: annuitants who lived through the pandemic years now appear to run below the 2012 IAM Table's mortality projection, meaning they are living longer than the table assumes (SOA Research Institute, 2026). Longer expected survival raises every kpx term in the same summation that the yield curve is trying to shrink, which raises äx back up and works against the payout rate improvement a rate-driven buyer expects to see. A carrier quoting SPIA rates off the 2012 table with no post-COVID adjustment is either overpaying buyers relative to its true expected liability or, if it has already adjusted, quoting a rate the buyer perceives as less generous than the headline Treasury yield would suggest, a gap that shows up as a competitiveness problem rather than a solvency one, but a real pricing decision either way.
The two forces net against each other in a way a buyer never sees in a single quote. Move the discount rate alone from a 2% environment to today's roughly 4.6% and, holding mortality fixed, a $500,000 premium supports a materially higher monthly income for a life-only SPIA on a 70-year-old, since a higher v compounds favorably across every year of the annuity-due summation. Layer in a 2-to-3 percentage point downward revision to the mortality assumption, the range the SOA-LIMRA survivor-selection finding suggests for post-2022 issues, and a meaningful share of that rate-driven improvement is given back through a longer expected payment stream. The buyer experiences this as a payout rate that rose with Treasury yields but by less than a naive back-of-envelope calculation would predict, and the pricing actuary's job this cycle is to make sure the gap between naive and actual is the mortality basis working as intended, not a stale table masking margin erosion.
A Deferred Book Waiting to Become a Payout Book
The $44.7 billion of fixed-rate deferred and $30.7 billion of fixed indexed sales booked this quarter are accumulation liabilities today, but a portion of that book will eventually annuitize into SPIA-like payout streams, whether through an in-contract income rider or a full annuitization election at the end of the deferral period. That conversion path is precisely the seam where the reserving framework changed on January 1, 2026: VM-22 replaced the prescribed Commissioners Annuity Reserve Valuation Method with a principle-based, company-specific stochastic framework for non-variable annuities, running in parallel with the old regime through a transition period before mandatory application to all new business by January 1, 2029 (NAIC Valuation Manual, 2026 Edition). VM-22's scope explicitly spans both sides of the annuity lifecycle the industry just recorded records in: accumulation products such as fixed indexed and multi-year guarantee annuities, and payout products including SPIAs, pension risk transfers, and structured settlements.
Under CARVM, a company's payout-conversion reserve was largely a factor lookup, insensitive to whether the market environment at conversion looked like a 2% or a 4.6% yield world. Under VM-22, reserve levels move with market conditions and with the company's own best-estimate mortality and lapse assumptions, which is a more accurate reflection of economic risk but also a source of new earnings and capital volatility that did not exist under the old formula (Milliman, 2026). A record accumulation book converting into income streams during a period when rates move, even modestly, will generate reserve swings that a CARVM-era finance team never had to model. The record size of the 2025 and 2026 accumulation cohorts, on top of the mortality-basis question the SOA-LIMRA study already raised for the smaller payout book written today, means the reserve adequacy question is not confined to this quarter's $4.0 billion of new SPIA business; it extends to however much of the $464.1 billion 2025 vintage eventually annuitizes.
The timing compounds the modeling burden rather than simplifying it. Most fixed-rate deferred and fixed indexed contracts sold this quarter carry surrender charge periods running five to ten years, which means the bulk of the $75.4 billion in combined FRD and FIA sales this quarter will reach its first annuitization decision point somewhere between 2031 and 2036, a horizon over which nobody can assume today's 4.6% yield or today's post-COVID mortality basis will still describe the market. VM-22's stochastic reserving is explicitly designed to carry that scenario range forward rather than freeze it at issue, which shifts real work onto the assumption-setting process now: a company's projected annuitization rate assumption, the share of a deferred contract's policyholders expected to elect income rather than lump-sum withdrawal or death benefit, has to be revisited every valuation cycle against a demand signal that this quarter's SPIA and DIA records suggest is strengthening, not the flat, low single-digit annuitization rate many legacy models still assume by default.
The Peak 65 Demographic Wave Behind the Shift
More than 11,000 Americans turn 65 each day, over 4 million a year, and fewer of them carry the pension income that shaped prior retiree cohorts' guaranteed-income baseline (LIMRA, 2024). LIMRA's tracking shows the gap that creates: 72% of current retirees report their households receive enough lifetime-guaranteed income to cover basic living expenses, but only 47% of working adults ages 50 to 75 believe they will be able to say the same, an 11-percentage-point decline in confidence from 2017. That confidence gap is the demand-side mechanism underneath a $4.0 billion SPIA quarter and a $1.3 billion DIA quarter that grew 32% sequentially: buyers approaching retirement without a pension are converting savings into a substitute for the guaranteed income stream their parents got for free. LIMRA's own 2026 outlook frames the shift as durable rather than cyclical, projecting continued strength through 2028 on the combination of the Peak 65 wave, still-elevated rates, and a widening share of retirees with no employer-sponsored income floor (LIMRA, 2026).
The Assumption-Setting Shift From Crediting Rates to Longevity
The actuarial skill set that dominated the last four years of record annuity sales was built for accumulation products: option-adjusted crediting-rate design, cap and buffer calibration, hedge-cost budgeting against index volatility. None of that machinery prices a SPIA. A decumulation tilt, even a modest one measured against a $123.9 billion accumulation-heavy quarter, shifts pricing work back toward mortality table selection, improvement scale assumptions, and the interaction between a moving yield curve and a longevity basis that most pricing teams have not had to defend at this volume since before the 2010s low-rate era compressed payout-annuity new business into a niche line. Life actuaries setting 2027 payout-rate assumptions are now doing that work against a live VM-22 reserve framework, a post-pandemic mortality study still being digested, and a Treasury curve that has moved enough to change the competitive calculus on every quote. The RILA record will keep making headlines. The SPIA and DIA records are the ones that determine whether the next several years of decumulation gets priced right.
Further Reading
- SOA-LIMRA Payout Annuity Study Resets SPIA Mortality Pricing
- VM-22 Goes Live: Annuity Reserves Enter the Stochastic Era
- LIMRA Q1 2026: The RILA-FIA Product Shift and What It Means for Hedging and Pricing
- RILA Sales Hit $21.2 Billion as Total Annuity Sales Clear $100 Billion for a Tenth Quarter
- AM Best's Q1 2026 Data Shows an 18% Life/Annuity Income Decline That Traces to One Carrier
Sources
- LIMRA, “U.S. Annuity Sales Notch Tenth Consecutive $100 Billion+ Quarter,” LIMRA.com, July 2026
- InsuranceNewsNet, “LIMRA: Annuity sales set new quarterly record with $123.9B in Q2,” InsuranceNewsNet.com, July 2026
- LIMRA, “Final U.S. Retail Annuity Sales Set New Sales High, Totaling $464.1 Billion in 2025,” LIMRA.com, 2026
- LIMRA, “Future Retirees Face a Different Reality and Increasingly Look to Annuities for Financial Security,” LIMRA.com, 2024
- LIMRA, “The 2026 Annuity Sales Outlook Remains Strong,” LIMRA.com, 2026
- NAIC, Valuation Manual, January 1, 2026 Edition, VM-22 Non-Variable Annuity Reserve Requirements, content.naic.org, 2026
- Milliman, “VM-22 Readiness: Key Areas for Consideration,” Milliman.com, 2026
- SOA Research Institute, “2020-2024 Individual Payout Annuity Mortality Experience Study,” SOA.org, 2026
- Federal Reserve Bank of St. Louis, Market Yield on U.S. Treasury Securities at 10-Year Constant Maturity, FRED.stlouisfed.org, July 2026