Overview of Market Risk Benefits

Market Risk Benefits (MRBs) became relevant with the introduction of LDTI over recent years. Defined by the ASU, an MRB is:

• “A contract or contract feature in a long-duration contract issued by an insurance entity that both protects the contract holder from other-than-nominal capital market risk and exposes the insurance entity to other-than-nominal capital market risk.”

• “A benefit is presumed to have other-than-nominal capital market risk if the net amount at risk would vary significantly in response to capital market volatility”

Said simply, if a product feature causes the company to take on some capital market risk, then that feature likely needs to be valued as an MRB. MRBs cover benefits paid to the policyholder in addition to the account value upon the occurrence of a specific event.

Market risk benefits require a full retrospective calculation (back to the issue date) and require a fair value calculation based on stochastic projected cash flows using risk-neutral scenarios and market participant assumptions.

How do I know if I have an MRB?

Prior to the adoption of US GAAP Long Duration Targeted Improvements (LDTI), guaranteed benefits (GMDB, GMWB, GMIBs, etc.) were typically valued on a GAAP basis using SOP 03-1. In some cases, they may have been valued as an embedded derivative. Under LDTI, these benefits are valued as MRBs.

MRB GAAP Reporting Impacts

There are two ways in which the change in MRBs are recorded in income:

1. The change in Fair Value of the MRB (ignoring changes in own-credit-risk) is applied to Net Income.

2. The change in Fair Value due to a change in own-credit-risk is applied to Other Comprehensive Income.

Own credit risk refers to a company’s risk that they will be unable to fulfill future obligations. In the context of MRBs, an own-credit-risk spread is developed to account for this risk and added to the discount rates used in the MRB valuation. Any changes in the MRB balance caused by a change in this spread, due to some event like a credit rating downgrade, would be reported in Other Comprehensive Income.

MRB Valuation Approaches

There are two widely accepted approaches to calculating an MRB: a non-option approach (i.e., Attributable Fee) and an option approach. Both approaches require the same stochastic projection output using risk-neutral scenarios.

The decision as to which of these methods to use is highly dependent on product design.

Non-Option Approach

• Commonly referred to as “Attributable Fee” method and requires the calculation of an attributed fee percent (AF%)

• Most common for VAs since there is no host to adjust

• Commonly used for non-VAs when rider fees are sufficient because it is simple

• Typically results in MRB = 0 at issue

As long as fees are sufficient at issue (i.e. PV of Fees > PV of Benefits), the MRB liability will start at 0 and emerge over time using the locked in AF%. If fees are insufficient, the MRB will be non-zero and result in a loss at issue. In the case of insufficient fees, an option approach is typically explored.

Option Approach

Some form of an option approach is common for Fixed Indexed Annuities as GMxB benefits are often implicitly priced through investment spreads instead of explicit rider fees. Additionally, an FIA product has an existing FAS 133 host that can be adjusted to avoid a loss at issue from a non-zero MRB balance.

This approach fits in with the FAS 133 approach for FIAs, and instead of bifurcating the premium between the host and embedded derivative (VED), the premium is trifurcated between the host, embedded derivative, and MRB.

Even if fees are insufficient and a non-zero MRB at issue exists, the total GAAP reserve is not impacted as the initial host is reduced to offset the non-zero MRB.

Looking Forward

Undoubtedly as time goes on, more methodologies will emerge as staples in the industry. Not all contracts fit nicely into the non-option and option approaches laid out above, and the introduction of new product types in the market will challenge the “norm.” Companies will continue to long for an approved simplified deterministic approach for MRBs, removing the need for complex stochastic modeling entirely.