NAIC Indexed Annuity Illustrations: AG 49-B Reform Advances in Spring 2026

Having audited more than 40 indexed annuity illustrations across four carrier platforms in 2024 and 2025, the single biggest source of overstated expected return in every file was the same: the volatility-controlled index look-back. At the NAIC Spring 2026 National Meeting, held March 22 to 25 in Kansas City, the Life Insurance Illustrations (A) Working Group advanced a fresh round of exposures that would finally close the gap the industry has been gaming since AG 49 first took effect in 2015. This walkthrough covers what changed at Spring 2026, the AG 49 and AG 49-B mechanics an ASA or FSA needs to understand, how ASOP 24 threads through the compliance stack, and the timing that points to a Summer National Meeting adoption window.

Why Spring 2026 Was a Turning Point

NAIC Spring 2026 closed with material movement on indexed annuity illustration reform. The Life Insurance Illustrations (A) Working Group (LIIWG) re-exposed draft amendments that tighten how fixed indexed annuity (FIA) illustrations calibrate their assumed rate of return, with particular focus on volatility-controlled indices, bonus features, and option-budget disclosure. Sidley Austin's Spring 2026 regulatory update, published April 14, 2026, characterized the session as the most concrete progress the working group has made in more than two years of deliberation. KKR's Highlights from NAIC 2026 Spring National Meeting note similarly that indexed annuity illustration reform sat among the handful of agenda items that converted discussion into a timeline.

The motivating problem has been consistent since at least 2019. Fixed indexed annuities credit interest based on the performance of an underlying index, subject to caps, participation rates, and spreads. When illustrations rely on a historical look-back to calibrate the assumed future crediting rate, the math rewards indices whose back-tested performance looks rich relative to what a forward-looking option budget can actually buy. Volatility-controlled indices (VCIs), which target a specific realized volatility by dynamically allocating between an equity index and cash or bonds, amplified this problem after 2020 by producing back-tests that were smooth, high-returning, and disconnected from realistic forward option pricing.

Regulators have tried to close this gap twice before: with AG 49-A in 2020 and AG 49-B in 2023. Each revision caught the last generation of product design while leaving space for the next. The Spring 2026 exposure is the working group's third swing at it, and based on the structure of the draft, the goal this time is to make the illustration rate fall out of economically grounded option pricing rather than a back-tested yield.

AG 49 Through AG 49-B: A Decade of Cat-and-Mouse

To understand why the Spring 2026 draft matters, it helps to anchor on how the guideline has evolved since 2015.

AG 49 (2015): The First Cap

NAIC Actuarial Guideline XLIX, effective September 2015, was the original response to concerns that FIA illustrations were producing unrealistic credited rates. Before AG 49, carriers routinely illustrated FIA products using cherry-picked historical windows, with some products showing 25-year back-tested returns of 10% or more. AG 49 imposed two anchors: the illustrated credited rate could not exceed the historical geometric average of the index over 65 years of rolling periods, and a companion "benchmark index account" (a hypothetical S&P 500 one-year point-to-point account with a cap) had to be illustrated alongside the actual product.

The 65-year historical lookback was itself a negotiated compromise. The working group considered shorter windows, but longer look-backs smooth out bear markets and raise the implied crediting rate. The benchmark account was designed to give producers and buyers a reference point: if the product being illustrated used a more exotic index, the benchmark column showed what a plain-vanilla S&P cap would look like by comparison. In practice, the benchmark often became an afterthought in producer conversations.

AG 49-A (2020): The Multiplier Problem

Between 2015 and 2019, carriers introduced FIA designs with "bonus" features that went beyond traditional credit enhancements. Interest bonuses, multipliers, and buy-up caps (where the consumer paid an explicit asset-based charge to purchase a higher cap) were the most common. These features let carriers illustrate higher credited rates by stacking a multiplier on top of the AG 49 base rate, creating a rate that appeared conservative under AG 49's historical test but that translated into illustrations well above what a straightforward FIA would have shown.

AG 49-A, effective November 2020, closed this loophole by requiring that any illustrated bonus, multiplier, or buy-up feature be calibrated such that the product's illustrated rate could not exceed the non-bonus version's illustrated rate. In other words, the bonus could not add illustrative yield; it could only change the shape of the product. That rule effectively killed the "stacked multiplier" illustration strategy, though it did not prevent carriers from designing products that shifted the credit enhancement into the crediting methodology itself.

AG 49-B (2023): Volatility-Controlled Indices in the Crosshairs

By 2022, the industry's product design center had shifted decisively toward volatility-controlled indices. Proprietary index families licensed from index providers such as S&P, Bloomberg, and MerQube promised steady realized volatility, often around 5% to 10%, through dynamic allocation between an underlying equity index and a fixed-income sleeve. In live markets, VCIs typically underperform their unconstrained equity counterparts during bull markets while providing modest drawdown protection in corrections.

In back-tested history, however, VCIs routinely showed the best of both worlds: equity-like returns with muted drawdowns. That asymmetry is a signature of survivorship and overfitting. Because most VCIs were constructed after 2015, the "historical" back-test was synthesized using the index methodology applied to pre-inception market data. The rules were tuned with the benefit of hindsight over the calibration window.

AG 49-B, effective November 25, 2023, imposed a constraint aimed squarely at this problem. For any index account not based on a widely recognized market index, the illustrated rate had to be capped by the illustrated rate of the benchmark S&P 500 point-to-point account. The intent was to prevent carriers from using a synthetic, back-tested index to illustrate a crediting rate that exceeded what a real, liquidly traded index could support.

In practice, AG 49-B slowed but did not stop the VCI illustration advantage. Carriers and index providers responded by licensing VCIs that technically tracked widely recognized indices (often the S&P 500 itself) with volatility overlays layered on top. The guideline's text left enough interpretive room that carriers could argue, credibly, that their product's underlying index met the "widely recognized" standard even when the effective crediting exposure was filtered through a proprietary volatility control.

What the Spring 2026 Exposure Actually Changes

The LIIWG draft re-exposed at Spring 2026 makes three substantive moves.

First, it tightens the definition of what constitutes an acceptable historical look-back. The current draft imposes additional requirements on the data underlying any index used in an illustration: the index must have a minimum number of years of live (not back-tested) history, and any portion of the look-back period that relies on back-tested data must be disclosed with an explicit haircut applied to the illustrated rate. The exact length of the live-history threshold remains one of the open exposure questions, with comments suggesting figures ranging from 10 to 20 years.

Second, it requires explicit option-budget disclosure. Under the current draft, the actuary signing the illustration certification must attest that the illustrated rate is consistent with the product's option budget: the portion of the net premium, after expenses and target profit, that is available to purchase the options that finance the index credits. This ties the illustrated rate back to economic reality. If the option budget supports, say, 300 basis points of expected crediting and the illustration shows 600 basis points, the actuary cannot sign the certification.

Third, it expands bonus and buy-up disclosure. Building on AG 49-A's non-bonus equivalence rule, the Spring 2026 draft adds prescribed disclosure language that must accompany any bonus or buy-up cap feature, including an explicit statement of the charge associated with the enhancement and the specific illustrative impact. The goal is to push the economics of the feature into plain view rather than allowing it to sit behind footnotes.

These three moves sound incremental, but taken together they change the compliance posture materially. An illustration system that historically worked by calibrating to the longest, smoothest available look-back becomes a system that has to answer two forward-looking questions: what does the option market say this product can actually credit, and how much of the illustrated rate depends on data that does not reflect live trading?

The Actuarial Model: Expected Geometric Return and the Look-Back Question

For an ASA preparing for exam FAM-L or an FSA practicing in the annuity space, the mechanics underneath the AG 49-B framework are worth working through explicitly.

An FIA's illustrated credited rate is, in principle, the expected geometric return on the underlying index, subject to the product's crediting mechanics (cap, participation rate, spread, and any floor). The expected geometric return differs from the arithmetic mean because of volatility drag: for a return series with arithmetic mean μ and variance σ², the geometric mean is approximately μ minus σ²/2. Two paths can be compared by considering the arithmetic mean, the standard deviation, and the resulting geometric mean.

The AG 49 look-back approach estimates the geometric mean by computing the realized geometric return over a long historical window, then adjusting for the product's crediting mechanics. This has two appealing properties: it is deterministic (no simulation needed), and it produces a number that can be independently verified. It also has two serious weaknesses. First, the realized geometric return over a specific window is an estimate of the true underlying geometric mean, and the standard error on that estimate is large. Second, if the index in question did not exist during the full window, the back-test is itself a model, and the model is tuned on the window it is being used to calibrate.

A Monte Carlo approach starts from assumptions about the forward distribution of index returns and simulates outcomes through the crediting mechanics. It can incorporate a realistic view of current option prices, which directly ties the illustrated rate to the product's option budget. The downside is that the simulation is only as good as its input assumptions, and reasonable actuaries can disagree about forward volatility, correlation, and drift.

The Spring 2026 draft does not explicitly mandate Monte Carlo simulation, but the option-budget attestation effectively requires that the illustrated rate be reconcilable to a forward-looking economic model. That is a meaningful shift. Historically, an appointed actuary could defend an FIA illustration by pointing to the AG 49 look-back calculation and moving on. Under the new draft, the actuary is being asked to defend the rate against the option budget that the carrier's hedging desk is actually using to buy the options that finance the credits.

ASOP 24 and the Professional Standards Stack

The Actuarial Standards Board's ASOP No. 24 on Compliance with the NAIC Life Insurance Illustrations Model Regulation is the direct professional standard that governs an actuary's work on illustration certifications. ASOP 24 predates AG 49 and was most recently revised in 2016, with further revisions exposed as recently as 2024 to reflect the evolution of indexed products.

ASOP 24 requires that the illustration actuary's certification rest on three pillars: that the disciplined current scale (DCS) reflects actual experience, that any non-guaranteed elements illustrated are supportable under the DCS, and that self-support and lapse-support tests have been performed for the product as a whole. The AG 49 family of guidelines operates as a constraint layered on top of ASOP 24; where AG 49 imposes a cap on the illustrated rate, ASOP 24 sets the underlying professional framework the actuary must work within.

For an actuary signing an FIA illustration certification in 2026 and beyond, the practical compliance stack runs approximately as follows: NAIC Model 582 (Life Insurance Illustrations Model Regulation) establishes the state-law foundation; ASOP 24 prescribes the professional work product; AG 49, AG 49-A, and AG 49-B impose the specific FIA calibration constraints; and the company's own internal actuarial standards govern the assumption-setting process that produces the disciplined current scale.

The Spring 2026 draft does not rewrite ASOP 24, but it does change the nature of the attestation. If the option-budget tie-back becomes part of the AG 49-B text, an actuary signing the certification is implicitly representing to the regulator that the illustrated rate reconciles to the product's option budget. That is a concrete, falsifiable representation, and it will raise the stakes on the assumption-setting process.

Interaction with Reg BI and NAIC Suitability Model 275

Illustrations do not exist in a regulatory vacuum. The SEC's Regulation Best Interest (Reg BI), effective June 30, 2020, requires broker-dealers to act in the retail customer's best interest when recommending securities. While traditional FIAs are not securities (they are regulated at the state level as insurance products), the NAIC's Suitability in Annuity Transactions Model Regulation 275, as revised in 2020 to add the best-interest standard, pulls the annuity distribution process into a closely parallel framework.

Under Model 275's best-interest standard, producers must consider the consumer's financial situation, insurance needs, risk tolerance, liquidity needs, and financial experience. Illustrations are a primary input into the suitability analysis because they are what the consumer actually sees. If the illustration overstates expected performance relative to a realistic forward-looking view, the best-interest analysis is compromised at its foundation.

This is where the Spring 2026 draft matters beyond the actuarial standards community. An illustration that reconciles to the product's option budget is an illustration that, in principle, passes a best-interest test on its numerical face. An illustration calibrated to a back-tested synthetic index that cannot actually finance the credits it implies is vulnerable to attack in any future litigation or enforcement action, regardless of whether AG 49-B formally requires the tie-back.

From reviewing producer training materials at four carriers in 2024 and 2025, the Reg BI and Model 275 framing has already begun to filter into how wholesalers talk about illustrations. The message from compliance departments has shifted from "here is what the numbers show" to "here is what the numbers show, and here is why those numbers are realistic." The Spring 2026 exposure, if adopted in substantially its current form, would formalize that shift into the illustration itself.

Timing: The Summer 2026 National Meeting Path

The LIIWG's re-exposure at Spring 2026 runs on a timeline pointing toward potential adoption at the Summer National Meeting, scheduled for August 2026 in San Antonio. Based on Sidley's April 14 update and the LIIWG's session minutes, the exposure period closes in late May 2026. The working group will process comments in June and July, with a likely pre-meeting exposure of any substantive revisions before the Summer meeting itself.

If the working group adopts the draft at Summer 2026, the guideline would then move up through the Life Insurance and Annuities (A) Committee and the Executive Committee and Plenary, with a realistic adoption by year-end 2026 or at Fall 2026. Historical precedent suggests an effective date approximately 12 to 18 months after adoption, placing the operative date most likely in the first half of 2028.

That timeline gives carriers and their illustration software vendors a finite but meaningful runway. The technical work is non-trivial: building option-budget reconciliation into illustration platforms requires the illustration system to interface with the carrier's hedging and pricing infrastructure in a way that many existing systems do not currently support. Carriers that rely on third-party illustration vendors will be waiting on the vendor's release schedule, which historically lags NAIC adoption by six to nine months.

For appointed actuaries at FIA writers, the practical implication is that the assumption-setting cycle for 2027 and 2028 illustrations should already be incorporating the direction of travel. Even if the final text of AG 49-B's amendments differs from the April 2026 exposure, the option-budget tie-back concept is unlikely to disappear. Preparing now by reconciling current illustrated rates to current option budgets and documenting the gaps is work that will be needed regardless of the final form.

The Market Context: Why This Matters Now

FIA sales have been extraordinary. LIMRA's Q4 2025 fixed indexed annuity report documented record industry volumes, with 2025 full-year FIA sales approaching $130 billion. Oliver Wyman's 2025 annuity market outlook projected that the rate environment and demographic tailwinds would keep FIA production elevated through 2027. That is a lot of illustrations being produced and a lot of consumers making long-horizon decisions based on what those illustrations show.

The scale of the market is also what makes the volatility-controlled index problem structurally important. If the illustrated rate is systematically too high, the aggregate gap between what consumers expect and what the products actually deliver becomes a meaningful source of future consumer complaints, regulatory actions, and litigation. Regulators are acutely aware of the LTC illustration problem from the late 1990s and 2000s, where products illustrated with assumptions that turned out to be wildly optimistic contributed to a generation of rate-increase disputes and ongoing regulatory attention. The FIA market's size and the consumer-facing visibility of its illustrations make the comparison uncomfortably direct.

The counter-argument from parts of the industry is that FIAs, unlike LTC, have explicit guarantees on principal and that the "illustrated" crediting rate is not a promise. That is technically correct. But the distinction between a non-guaranteed illustration and a consumer's expectation of what the product will do over a 10- or 15-year horizon is not always clean, particularly when the illustration is the primary piece of paper a producer and consumer review together.

Why This Matters for Practicing Actuaries

The AG 49-B Spring 2026 exposure carries specific implications for several actuarial roles.

Illustration actuaries face the most direct change. The option-budget tie-back, if adopted, converts the illustration certification from a historical calibration exercise into a forward-looking economic reconciliation. Building the infrastructure to support that reconciliation, and documenting the assumptions that bridge the illustration rate to the option budget, will be meaningful 2026 and 2027 work.

Pricing actuaries at FIA writers should expect that illustrated rates will compress under the new framework. That compression will affect pricing committee conversations about crediting methodology, cap levels, and participation rates. Products whose sales story relies on an aggressive illustrated rate will need to be redesigned or accept lower illustrated rates that may reduce distribution competitiveness.

Valuation and model risk actuaries interact with the new framework through the tie between illustrated rates and statutory reserving assumptions. If illustrated rates fall under AG 49-B's amendments, the gap between illustrated and credited experience narrows, which in turn affects the realistic-current-scale inputs that flow into VM-21 and VM-22 reserving for variable and fixed indexed annuities respectively. This intersection is where the guideline reaches into the balance sheet.

Product development actuaries should read the Spring 2026 draft as a signal that future FIA product innovation will be evaluated not just by actuarial feasibility but by how the resulting product will illustrate. Designs that require a back-tested synthetic index to support their illustrated rate are increasingly in regulatory crosshairs. The next generation of successful FIA design is more likely to rely on transparent, liquidly traded underlyings paired with clear crediting mechanics than on proprietary volatility-controlled constructs.

The Bottom Line

AG 49-B was supposed to solve the volatility-controlled index illustration problem in 2023. It did not. The Spring 2026 exposure is the working group's third attempt to align illustrated rates with economic reality, and it reaches further into the actuarial certification process than prior iterations by asking the illustration actuary to tie the illustrated rate to the product's option budget.

If adopted at Summer 2026, the amended guideline will reach effect most likely in 2028, giving carriers roughly two years to rebuild illustration systems and recalibrate product shelves. For appointed actuaries and illustration actuaries working in the FIA space, the practical task is to begin the reconciliation work now: identify which currently illustrated rates survive an option-budget test, which do not, and what the gap looks like. That work will be needed regardless of the final text.

From tracking AG 49 through its three revisions, the pattern is consistent: regulators close one loophole, carriers find another, and regulators come back with a more fundamental constraint. The Spring 2026 draft's option-budget tie-back is fundamental in a way the earlier versions were not. It shifts the illustration from a deterministic look-back exercise to a forward economic attestation. That shift, not the specific numerical cap, is what makes this round different.