Description: Discount rate curve construction has been a topic of renewed interest following recent changes in financial reporting standards. Interpolation is a key component of curve construction and can be conducted via an assortment of methods. This article explores six popular interpolation techniques and assesses how they perform with regards to desirable features of such methods: continuity of forwards, stability and locality, simplicity, and market consistency.

Description: Actuaries can leverage cluster modeling to reduce model runtime and cloud compute costs, which can be a real concern as companies transition from an on-premise grid to a scalable cloud solution. We consider practical considerations associated with clustering, through both US statutory and US GAAP lenses. Lastly, the results from a variable annuity liability clustering case study are presented.

Description: Risk management techniques do not include system science. System science requires that the uncertainty of the world is incorporated into the results. The articles explains the difference between risk and uncertainty. Learn how risk practices create an overoptimistic view of future outcomes.

Description: Milliman was engaged by the SOA to conduct two Accelerated Underwriting (AU) surveys, one of direct companies and another of reinsurers. The surveys explored actual results from 2019–2021 and expected results from 2022.

Description: This Q&A article discusses the use of machine learning to analyze unstructured data. Supriya Ramachandra, a data scientist at SIKKA, explains how her team uses natural language processing techniques, technical aspects of neural networks, LSTM modeling, and natural language processing. Machine learning allows for the classification of large amounts of unorganized data, providing actionable information for actuaries.

Description: Article about Prompt Engineering Tips.

Description: Underwriting specialty risks in life insurance can be challenging, but due diligence is at the core of what allows us to make the best and most informed decision.

Description: This is an update, as of February 2025, on the Reinsurance Section’s sponsored research in process and studies recently completed.

Description: One common barrier to employing GLM modeling is the lack of sufficient data. This limitation often comes from the combination of the considerations in ASOP No. 25 “Credibility Procedures” and the fact that GLMs assign full credibility to the data. Explore the possibilities of credibility-based regression by expanding the intended purpose of modeling analysis from producing a single model as output to producing data insights, credible differences, and a range of reasonable models.