Introduction: How VM-22 PBR is poised to change annuity pricing

Principle-based reserving (PBR) has arrived for a range of products reserved under VM-22.¹ Despite ample discussion among regulators, field testing, and industry commentary in the years preceding formal adoption of the new reserving guidance, annuity writers are still determining how to incorporate these changes into pricing practices for affected products. This comes at a time when the annuity industry has seen sustained sales growth across multiple product lines, including fixed-index annuities (FIAs) and multi-year guaranteed annuities (MYGAs), due to both the interest rate environment and demographic surge in retirement age individuals. With new entrants still joining an already competitive market, annuity writers’ pricing practices must maintain the level of sophistication necessary to succeed. VM-22 PBR both represents a significant increase in complexity relative to the existing commissioners’ annuity reserve valuation method (CARVM) and serves as a new measure by which to evaluate product profitability and risk. This paper discusses several important considerations for insurers to keep in mind as they adapt their pricing models and philosophy to a VM-22 PBR world.

Historically, pricing of MYGAs and FIAs in the United States was driven by company experience supplemented by industry experience when needed, competitive positioning, and compliance with minimum statutory requirements. Key profitability metrics included present value measures such as internal rate of return (IRR), profit margins, and statutory earnings projections. Pricing typically relied on deterministic scenarios for asset yields and policyholder behavior, commonly with conservative assumptions for lapse rates and credited interest rates to support profitability targets.

Common shortcuts included simplified projection models based on limited scenarios and aggregate assumptions for policyholder behavior. Product features and crediting strategies were often benchmarked against those of similar products, leading to the use of “rule of thumb” margins or quick stress tests in place of comprehensive stochastic analysis. While these practices supported efficient product development and speed to market, they lacked the rigor and risk sensitivity that frameworks like VM-22 are intended to provide.

Although the same profitability metrics are expected to remain in focus, the decision process around them should dramatically change. Here, we outline seven factors annuity writers should consider as they roll out pricing models following VM-22 PBR adoption.

1. Assumption setting must evolve under VM-22 PBR

With the shift to PBR, insurers’ approaches to setting assumptions must evolve. They can now reflect best-estimate assumptions—subject to provisions for adverse deviation (PADs)—which represent greater flexibility than assumption setting under CARVM. The PADs are generally not prescribed, but VM-22 does provide guidance for setting them and also requires an annual disclosure using a prescribed set of assumptions.² Additionally, the assumption-setting process extends to both assets and liabilities in the PBR framework, expanding the inventory of assumptions insurers must evaluate on a regular basis. Whereas valuation departments responsible for reserving in-force business will rely on guidance notes within the VM-22 language, pricing departments may embed margins based on separate actuarial judgments and could rely on other assumption-setting processes.

For example, for a FIA with a guaranteed lifetime withdrawal benefit (GLWB), when policyholders begin taking withdrawals (i.e., utilizing their rider benefit) is an important actuarial assumption. Pricing teams may simplify the assumption across a few model cells with a certain level of nonutilization embedded based on actuarial judgment. However, the valuation team may set the assumption based on a more robust withdrawal cohorting method derived from guaranteed actuarial present value curves as documented in the valuation manual, along with a PAD for conservatism around rates of benefit nonutilization. As sales occur, if the reserve strain is higher for the block than expected in pricing, then realized profitability will be lower than expected.

The potential inconsistency between valuation and pricing assumptions, on both best-estimate and prudent bases, is not new to annuity writers. However, under VM-22, where valuation departments have more discretion in assumption and PAD setting than before, the risk of inconsistency is heightened and requires increased coordination between insurers’ pricing and valuation functions.

2. Reserve projection under VM-22

With most annuity products in scope under VM-22 requiring a stochastic reserve (SR) calculation, pricing actuaries will need to evaluate the complexity and value trade-off of incorporating a nested stochastic calculation in earnings projections. A full recalculation of the projected reserve will introduce significant complexity to pricing models, as it will require future scenario sets to be generated, along with additional projection model configurations and run time. Further, the assets and liabilities of the block must be incorporated in any future SR calculations, reinforcing the importance of robust pricing assumptions. For example, since VM-22 PBR is intended to penalize poor asset-liability management (ALM) within the reserve calculation, oversimplifying the asset portfolio composition and/or future reinvestment activity could lead to an ALM mismatch that emerges in future SR calculations, leading to unrealistic reserve strain building up.

Alternatively, pricing actuaries could look for other proxies for the full SR recalculation. Some example approaches could include deterministic scenario, reduced scenario techniques, proxying reserve proportional to another projected risk measure, or even simplifying the reserve strain down to an expected spread on the assets. These types of simplification could substantially reduce complexity and improve efficiency of pricing models but sacrifice accuracy. The lack of accuracy could take the form of legitimate misstatement of reserve levels or of the timing of the change in reserve along pricing projections, both of which could materially impact expected profitability using measures like IRR.

One middle ground could be to perform the robust recalculation of the SR along pricing projections periodically. These results could be used as a reasonability check against more agile simplifications of the projected reserves or potentially even used directly to calibrate an alternative. Like many other pricing assumptions, though, the chosen reserve methodology should be sensitivity tested to establish appropriate expectations.

3. Repricing cycle frequency: The impact of VM-22 PBR

Building on the consideration for reserve projection methodology, it is important to highlight how VM-22 PBR could influence insurers’ approaches to the pricing cycle. It is common for companies to fully reprice core products once or twice a year and establish a number of levers that can be pulled for swift rate-setting adjustments as the economic or competitive environment moves around. However, insurers could use the added complexity and sensitivity of VM-22 PBR as a rationale to increase the frequency of repricing products. For example, the SR calculation relies on the recently adopted GOES scenarios.³ If the GOES scenarios undergo an unexpected model change that impacts the distribution of results underpinning the SR, insurers will want to be quick to understand and be able to incorporate pricing changes to accommodate. Similarly, with largely principles-based assumption setting, the VM-22 reserves for some products may be more sensitive to emerging policyholder behavior trends than the prescribed assumptions embedded in the prior reserving framework.

4. Incremental versus standalone pricing presents opportunities under VM-22 PBR

VM-22 PBR introduces the opportunity for insurers to account for the natural diversification benefit across policies and products when pricing in expected reserve levels for new business written. This question—whether to price business on a standalone or incremental basis—is similar in nature to how insurers have historically grappled with expense allocation. Pricing on a standalone basis has the advantage of allowing insurers to better assess the profitability of individual products or cohorts of policies. However, pricing on an incremental basis captures the legitimate economic value of new business, which likely will be an important factor that improves IRRs and serves as a justification for some annuity writers to offer more competitive products.

The diversification benefit present when pricing on an incremental basis is not clear-cut, though. Pricing teams could consider aggregation of varying degrees, either among policies within the same product line, across different annuity products that are reserved together under the new VM-22, or potentially with legacy in-force business, depending on where regulators land with retrospective application of VM-22.⁴ Pricing actuaries should be cautious when incorporating a reasonable benefit of aggregation, though, since pricing cells are hypothetical in nature and the actual new business mix could experience different levels of diversification than what was priced in. Additionally, the aggregation benefit will likely vary based on market conditions, which layers another dynamic element into sensitivity analysis for pricing metrics.

5. Asset granularity and optimization: VM-22 PBR requires thoughtful asset assumptions

VM-22 PBR necessitates more thoughtful asset assumptions than the prior framework. Similar to how liabilities are often represented by a limited number of pricing cells, insurers’ strategic asset allocation (SAA) is typically represented by a relatively small number of hypothetical assets as opposed to the reality in which portfolios often have hundreds or thousands of individual assets. This limitation raises a question around how much ALM analysis can be performed as part of the pricing function, where a small number of hypothetical assets may lead to unrealistic levels of cash flow mismatches with the liability, even if other initial measures like yield and duration are reasonable. In addition to the cash-flow-matching element of ALM risk, optionality of certain assets (e.g., callability on some bonds, prepayment risk on mortgage-backed securities) would likely manifest in a robust VM-22 SR calculation but could be difficult to approximate in a pricing model. Asset risks like this could detract from an insurers’ ability to meet their target SAA, which is particularly important given the asset-intensive nature of products in scope under VM-22.

Default assumptions prescribed by VM-22 could also affect what insurers deem the optimal SAA. Previously, insurers may have priced based on best-estimate default assumptions, with embedded reserving methodology being disconnected from the underlying investments. However, now, with the PBR framework, the reserve must reflect the assets backing reserves and use default costs prescribed by the National Association of Insurance Commissioners (NAIC) for those assets. Differences naturally arise between company best-estimate and NAIC default assumptions, by both asset class and credit rating. Because of this, it is possible that a hypothetical portfolio with the “optimal” yield net of defaults along the best-estimate pricing path is suboptimal within the VM-22 reserve. Likewise, the prescribed reinvestment guardrail in VM-22 is commonly more conservative than insurers’ reinvestment strategy, which could produce a similar disconnect between what is “optimal” under VM-22 versus a best-estimate path.

6. Reserve structuring shake-up introduced by VM-22 PBR

It is conventional wisdom that as regulation changes, the industry adapts as market participants seek corporate structures that are the most capital efficient subject to their business constraints. This has led to insurers embracing a menu of reserving options for annuity business sold in recent years, as evidenced in particular by the annuity reinsurance flows into non-U.S. jurisdictions. Evaluating where and how to hold annuity reserves is ingrained in the pricing strategy of many annuity writers, whether the possibilities are internal risk transfer options, reserving financing, or external reinsurance. There are a number of implications to each of these options, but VM-22 PBR begins a new era in which insurers have another reserving methodology to contemplate in their pricing strategy. As shown by the field-testing model office results, certain products experienced an increase in reserves relative to CARVM, whereas others saw a decrease.⁵ This could be one influencing factor in the rate of adoption of VM-22 PBR and whether insurers shift their approaches to reserving and/or the use of reinsurance on the affected annuity products.

7. Annuity product design: VM-22 PBR may not have an impact

A fair question to ask is whether this regime change in reserving framework could lead to new product innovations, as it is possible that certain product features can interact favorably with regulatory guidelines in a manner that improves profitability. However, existing insurers have already been fairly creative in recent history with the use of reserve structuring solutions. These solutions often trend toward evaluation of products under a more economic reserve definition compared to CARVM, which is closer in nature to the objectives of VM-22 PBR. Through this experience, many have already identified the sensitivities and risks of certain product features that may exhibit similar effects under VM-22. For example, market value adjustment (MVA) features may be particularly impactful within VM-22 due to how the PBR approach rewards strong ALM, but most companies have already introduced MVA features on annuity products where possible.

Certainly, significant product innovation is still expected over the next few years as competition in the annuity industry is sustained. However, the product design changes are not expected to be driven purely by the adoption of VM-22. Still, the PBR framework may allow insurers more flexibility to reflect first-principles cash flow modeling of new products and features.

Conclusion: A new approach to annuity pricing under VM-22 PBR

Pricing models come with a range of complexities and span the spectrum between art and science, but VM-22 PBR reform represents an opportunity to revamp pricing models at a new level. Pricing actuaries should be proactive in understanding the requirements of VM-22 PBR and work with valuation, investment, and risk functions to coordinate assumptions and methodologies, ensuring new business written meets both the risk and return standards expected. The seven considerations highlighted above are not an exhaustive list but illustrate the sweeping effects VM-22 could have on insurers with FIAs, MYGAs, or other non-variable annuities as a core line of business. Despite what has already been adopted, there are still potential changes on the horizon as regulators evaluate retrospective application of VM-22, C-3 Phase 1 reform, and other aspects of annuity valuation that should be incorporated into pricing strategies. In the meantime, insurers should be educated on and prepared for the pricing opportunity that VM-22 PBR has created.

¹ Specifically, the VM-22 statutory framework will apply prospectively to new business issued starting January 1, 2026. However, insurers will be given a three-year transition period with a mandatory application for all business issued starting January 1, 2029.

² Insurers must calculate the standard projection amount (SPA) and indicate whether an additional SPA exists above the company’s stochastic reserve. At the time of publication, regulators are currently reviewing APF 2025-12, which contains language to strengthen the guidance around the SPA.

³ GOES is “Generator of Economic Scenarios”, which represents another significant reform effort by the NAIC to replace the prior economic scenario generator with a new model maintained by Conning.

⁴ At the time of publication, regulators have been actively reviewing the potential and scope of retroactive application of the new VM-22. Currently aggregation across accumulation and payout product lines is allowable under certain conditions for business issued after VM-22 PBR adoption, but potential for business aggregation could expand under retrospective application.

⁵ For example, FIAs without a GLWB saw a modest increase in reserves relative to CARVM, whereas FIAs with a GLWB saw a notable decrease in reserves. See American Academy of Actuaries. (March 12, 2025). VM-22 field test: Summary of participant results— addendum for ASPA analysis [Slide deck]. Retrieved February 10, 2026, from https://content.naic.org/sites/default/files/call_materials/20250312F%20SPA%20Addendum%20Participant%20Results%20for%20LATF%20%281%29.pdf.