Description: What is a "super" model? How do we develop, maintain, and visualize them? What is the difference between models and assumptions? Most importantly, how can we use all of this to quantifiably improve risk management?

Description: This article discusses considerations for statistical biases that may arise in LTC blocks due to historical benefit reductions. The article explores these biases in the context of experience studies and experience projections.

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: Deep learning neural network applied to experience analysis for post-level term lapse and mortality.

Description: In recent years, vast amount of data from emerging experiences has helped insurers gain a deeper understanding of their policyholders and the factors driving their behaviors. However, due to limitations in traditional experience study methods, this understanding typically remains at the cohort level. This series of articles proposes a novel machine learning-based model that analyzes and predicts policyholder behaviors at the individual policy level. Policyholder embedding representations are learned using a Recurrent Neural Network (RNN) that takes into account the sequence of historical transactions associated with a policy. Additionally, a Neural Network with a multi-head architecture is trained to predict the policyholder's next action based on the policyholder embedding and other policy attributes. The resulting model offers several advantages over traditional experience studies and assumption-setting methods, including the ability to set policy-level assumptions that improve reserve accuracy, predict multiple behaviors simultaneously while capturing the interrelationships between them, and generate policyholder embedding representations that are useful for segmentation.

Description: In this article, we introduce a novel machine learning-based model that analyzes and predicts policyholder behaviors at the individual policy level. Policyholder embedding representations are learned using a Recurrent Neural Network (RNN) that takes into account the sequence of historical transactions associated with a policy. Additionally, a Neural Network with a multi-head architecture is trained to predict the policyholder's next action based on the policyholder embedding and other policy attributes. The resulting model offers several advantages over traditional experience studies and assumption-setting methods, including the ability to set policy-level assumptions that improve reserve accuracy, predict multiple behaviors simultaneously while capturing the interrelationships between them, and generate policyholder embedding representations that are useful for segmentation.

Description: A short article looking at the results from an annual assumption survey of VA writers with a focus on changes companies observed and planned to make due to COVID-19. In the article we discuss what companies should be considering as they approach the 2020 year end assumption setting and audit season.