Description: The two-parameter Pareto distribution is a commonly used model in reliability and risk modeling. Minimum variance unbiased estimates of the parameters of Pareto distribution are not known. In this paper, we propose and investigate a method of estimation of the parameters of the Pareto distribution. This method is based on minimization of the Kolmogorov-Smirnov distance between the empirical cumulative distribution function [cdf] and the cdf of the Pareto distribution.

Description: An excerpt from a presentation at the 2014 Living to 100 Symposium, this article describes an open-source model for use by actuaries attempting to estimate future health benefit costs, discusses its strengths and limitations, and projects that spending for those age over 65 will take more than $13 trillion by 2055, about 50 percent of total medical expenditures.

Description: Presenter offers ways in which actuaries might set credibility levels to be used by underwriters to set group health insurance expected benefits. From the Record of the Society of Actuaries, Vol. 24, No. 2.

Description: This research was funded by the Natural Sciences and Engineering Research Council of Canada [NSERC] Discovery Grant 36860–06. Loss reserving is one of the most challenging tasks facing actuaries. Numerous approaches have been developed to give reasonable estimates. Generalized linear models [GLMs] are becoming quickly popular statistical analysis methods to estimate loss reserves. However, most of these models are aggregate reserving methods based on loss development triangles, without using information with regard to the actual claims processes. This paper establishes a more sophisticated structural reserving method incorporating more detailed information, such as the premium exposure emergence pattern, the loss emergence pattern and the loss development pattern, within the framework of GLMs.

Description: This study note focuses on actuarial applications and loss distributions for insurance products.

Description: This is the abstract of the paper Minimum Quadratic Distance Estimators for the Zeta Parametric Family. In this paper, we propose a new estimator, based on quadratic distance, asymptotically fully efficient for parameter values greater than 2 and highly efficient for smaller values, but computationally more appealing than the maximum likelihood estimator we compare its asymptotic variance with that of the moment estimator and an estimator proposed by Seal, based on the ratio of the observed frequencies of the first two classes. We also discuss a goodness-of-fit test and a test to discriminate between the zeta distribution and the Poisson distribution truncated at 0.

Description: Panelists discuss the process and challenges presented by mergers or acquisitions. From the Record of the Society of Actuaries, Vol. 24, No. 2.

Description: In this paper, we study a survival regression model known as Cox's proportional Hazards model. We assume that the data available are already grouped and develop a minimum distance estimation procedure that is also applicable when the data are also left or right censored. Asymptotic properties of the estimators are established and the numerical implementation of the method based on an iterated reweighted least squares algorithm is discussed.

Description: By necessity, attributions of financial processes require approximations. Tools used for approximations include Aumann-Shapley, time step, step-through, average step-through, partial derivative, and average partial derivative. Significantly non-linear paths can cause problems during attribution. Time (theta) and market factors must be handled concurrently. Typically contractual factors can be handled separately. Average Step-Through, Aumann-Shapley and Aumann-Shapley-3 points perform the best given the same time period. Attribution should be done with fairly small time steps as opposed to using a lot of points in the Aumann-Shapley calculation, but using Aumann-Shapley-3 points as a last step has some advantages.

Description: The conventional way to analyze grouped observations on a single variable is to construct histograms or, sometimes, fit a normal distribution. We extend this latter approach by allowing for transformations to near normality.