The Withdrawal Delay Cohort under VM-21/AG-43: The case for random sampling

  • Print
  • Connect
  • Email
  • Facebook
  • Twitter
  • LinkedIn
  • Google+
By Benjamin J.L. Buttin, Matthias Kullowatz, Zi Xiang Low, Zohair Motiwalla | 25 March 2019

In early December 2017, the National Association of Insurance Commissioners (NAIC) released proposed revisions to the existing U.S. variable annuity statutory framework. These revisions were promulgated as redline updates to the existing Actuarial Guideline 43 instructions, paving the way for VM-21, “Requirements for Principle-Based Reserves for Variable Annuities.” Most observers expect that the final version of VM-21 will be formally adopted at the NAIC Summer meeting in August 2019, for a January 1, 2020, effective date.

One of the more challenging components of the VM-21 Standard Projection is the Withdrawal Delay Cohort Method, which is a prescribed approach for determining the timing of policyholder election for GLWB and Hybrid GMIB policies. Adhering to the prescriptive process for the Withdrawal Delay Cohort Method outlined in VM-21 can result in significant run-time challenges for companies. This paper employs a random sampling approach to address this run-time challenge that we believe may be helpful to companies that are dealing with this issue.

If you have any questions or comments on this paper or would like to discuss this topic further, please do not hesitate to reach out to Zohair at