Description: Explores the open vs. closed systems debate with regards to actuarial modeling and exposes the over-simplification. Valuable for anyone who uses models.

Description: This paper presents an approach to analyzing continuing care retirement community [CCRC] data, and demonstrates the methods using data from a large CCRC. It is assumed that residents make transitions among a number of states that represent the levels of care required by residents. There is randomness associated with both the transition times and the states entered at these times. The model is conveniently characterized in terms of transition intensity functions, which represent the instantaneous rates of transition between pairs of states.

Description: Although rate structures where low-cost cohorts subsidize high cost cohorts may be justified from a actuarial, marketing, or legal standpoint, the subsidization creates risks which must be understood and monitored. This article presents the concept of subsidization signature to measure and interpret the effects of rate subsidization, and to assess the sustainability of rating structures and the sensitivity of business mix assumptions.

Description: This article goes through some basic data exploration techniques to show how different approaches look.

Description: Panel discussion about ways can insurance carriers and health care coalitions work together, what data are coalitions requesting from health carriers, how health carriers respond to those requests, and implications for health carriers.

Description: A review of the Bridging the Gap Series: Application of Predictive Modeling in VA/FIA Risk Management at the Equity Based Guarantees Conference in Chicago.

Description: Using Healthcare Claims Data to Create Actuarial Models.

Description: Research report sponsored by SOA to model and compare projected costs of different proposals for a Medicare prescription drug coverage plan. 3 types of costs are studied: net benefit cost estimates per member per month, continuance table data, and projected annual costs to the government over 10 years for two selected plan designs.

Description: Even the best algorithms are only effective if they are trained on appropriate data. This article discusses common pitfalls in creating a training data set.