Life insurers and annuity writers are now beginning to understand the impact of the COVID-19 pandemic on their lines of business, as mortality data for the year 2020 is reported and analyzed. While the pandemic has affected different carriers in different ways, future mortality rates are a key assumption for annuity writers.
With Milliman’s acquisition of Ruark Consulting in December 2021, the industry’s leading variable annuity mortality study has been rebranded as the Milliman Variable Annuity Mortality Study. The study is based on data from 2008 through 2020, totaling $674 billion in account value as of the end of the study period, with over 1 million deaths across 19 companies.
Request information about ordering the full report
The 2021 study reveals that mortality increased by 11% as a result of the onset of COVID-19, and that this had and may continue to have a material impact on the variable annuity industry.
Other study highlights include:
- We estimate 11% relative excess mortality for variable annuity contracts in 2020, as a result of the COVID-19 pandemic, similar to estimates found in population studies and in studies of life insurance mortality.
- Across the 2008-2019 period, variable annuity mortality improvement averaged 1.1% annually, consistent with the G2 projection scale. Mortality improvement was similar across benefit types. Mortality improvement was below average for smaller contracts and above average for larger contracts, suggesting a “wealth effect” in mortality improvement.
- Most standard industry tables and bases systematically overstate or understate mortality for various cohorts, even when they closely approximate aggregate mortality. Although the VM-21 basis modifies the 2012 IAM Basic table to reflect variable annuity experience by living benefit type and age-gender cohort, it still does not fully capture differences by age-gender cohort, nor does it capture durational anti-selection by living benefit type. To help our clients mitigate this issue, we used the data in this study to develop separate Milliman Variable Annuity Mortality tables for contracts with and without living benefits, with adjustment factors to reflect selection by contract duration.
- The lowest mortality is found on contracts with lifetime income guarantees, such as guaranteed lifetime withdrawal benefit (GLWB) and guaranteed minimum income benefit (GMIB) riders. The different rider types exhibit divergent patterns by contract duration as well, with low initial mortality for contracts with living benefit riders and high initial mortality for death-benefit-only contracts. Both the hierarchy of living benefit types, and the distinct durational patterns, suggest the presence of anti-selection in the absence of individual underwriting.
- Mortality exhibits anti-selection by contract size, with lower mortality rates on larger living benefit contracts and higher mortality rates on larger death-benefit-only contracts. More affluent policyholders are more likely to be more financially savvy and have greater access to expert financial advice, which translates to more savvy purchase and persistency behavior.
- Tax-qualified contracts exhibit systematically lower mortality rates than non-qualified, across all living benefit types, a magnitude similar to that reported in past studies. Qualified contracts are already benefiting from preferential tax treatment, so it could be that qualified policyholders who invest in a product with longevity benefits are those who expect to benefit most from them.
- Living benefit riders that offer lifetime income guarantees exhibit meaningful differences in mortality depending on the contract’s withdrawal history. Contracts that have taken no withdrawals exhibit the lowest mortality; those that have taken regular income withdrawals exhibit mortality close to average; and those that have taken excess withdrawals exhibit the highest mortality. These differences may reflect policyholders’ different incomes and liquidity needs in response to conditions that correlate with higher mortality.
- Other variables that were shown to have systematic differences in mortality included death benefit type and distribution channel.
Detailed study results, including company-level analytics, benchmarking, and customized assumption models calibrated to the study data, are available for purchase by contacting Timothy Paris at (312) 873-9719 or using this contact form.