The National Association of Insurance Commissioners (NAIC) is evaluating a new economic scenario generator (ESG) for U.S. principle-based reserving (PBR). This paper presents an illustrative framework to quantify the potential impact of the scenario reform on a prototypical in-force block of variable annuities (VAs) using key economic scenario sets from the NAIC’s first field testing. The analysis focuses on VM-21 statutory reserve and C-3 Phase II capital impacts, capital market sensitivities, and measures of scenario dispersion in the tail.
Scenarios Reviewed:
The paper compares three different real-world ESGs:
- AAA Interest Rate Generator (AIRG)
- Conning GEMS with Generalized Fractional Floor (GEMS GFF)
- Conning GEMS with Alternative Shadow Floor (GEMS ASF)
Key Observations on Scenarios:
- GEMS sets exhibit higher interest rates, wider rate dispersion, and more curve inversion compared to the AIRG.
- GEMS models have higher fixed income returns due to higher credit spreads.
- GEMS GFF has the widest range of rate dispersion, while GEMS ASF dampens the frequency and severity of negative rates.
Model Description:
The valuation model uses 1,000 scenarios, actuarial assumptions consistent with VM-21 Standard Projection plus a 10% margin for adverse deviation, and assumes all assets earned the net asset earned rate. The liability modeling assumes a representative in-force liability and uses the prescribed withdrawal delay cohort method for GLWB riders.
Impact Analysis:
- TAR and Capital Requirements: GEMS scenario sets produce materially higher TAR than the AIRG, but hedging significantly mitigates this impact. For a $20 billion account value block, TAR increases by 2.4% (GEMS GFF, unhedged) and 1.0% (GEMS GFF, hedged) compared to the AIRG.
- Scenario Distribution: GEMS GFF exhibits a wider range of GPVADs than the AIRG or GEMS ASF sets. Hedging tightens the distribution of TAR across scenario sets.
- Market Sensitivities: GEMS sets exhibit more equity sensitivity than the AIRG, particularly GEMS ASF. Rate sensitivity is also higher in GEMS sets, driven by their risk premia model for equity returns. Hedging converges the level of delta across scenario sets.
- New vs. Existing Business: New business has a lower TAR due to future fee income recognition, while existing business has higher TAR due to withdrawal activity. GEMS scenarios have a lower TAR impact on new business compared to the AIRG.
- Rider Mix: GLWB riders require a larger TAR than the base case rider mix across all scenario sets, while GMDB riders have a lower TAR in GEMS scenarios. The GLWB vs. GMDB mix relationship is more substantial in GEMS scenarios.
- Fund Mix: Higher equity allocation increases TAR, with GEMS scenarios being more sensitive to this change. Hedging significantly reduces sensitivity to the fund mix in GEMS scenarios.
Conclusion:
The ranking of TAR levels across scenario sets is consistently AIRG, GEMS ASF, and GEMS GFF. The GEMS scenarios have higher interest rates, fixed income returns, and volatility. The impact on individual companies varies significantly depending on rider mix, moneyness level, fund mix, block age, and other factors. Robust dynamic hedging is the most effective way to insulate business from potential scenario reform impacts. Rate sensitivity is likely to increase, potentially bridging a statutory balance sheet closer to a more "economic" sensitivity profile. Equity sensitivity may be more complex but could also increase.
