Background Image

Statistical Modeling

Robert Bear's statistical modeling work has included the following accomplishments:

Modeling Risk

Consulting assignments modeling the risk that insurance loss reserves will develop adversely, especially for long-tailed lines of business.

Risk Underwriting

Worked with PXRE Group's underwriting, actuarial, and systems professionals to measure the potential impact of each risk underwritten on economic capital requirements and profitability levels. This required extensive work with alternative catastrophe models.

Fitting Loss Severity Distributions

Extensive experience fitting loss severity distributions and loss emergence patterns to insurance claims data.


Pricing Reinsurance Treaties

Extensive experience pricing reinsurance treaties required the development of complex loss simulation models. This work has included numerous risk transfer analyses for finite reinsurance contracts.

Research

Research performed as a consultant in developing and testing econometric loss reserving models.

Results

Econometric research at Insurance Services Office leading to improved forecasting techniques in ratemaking (received outstanding performance award for Inflation Adjusted Trend Procedure).

Method

In his statistical modeling work, he applies the Industrial Version of the Decisions Tools Suite of Excel add-ons developed by Palisade Corporation which includes the widely used @Risk simulation tool, StatTools for statistical analyses, Evolver for optimizing non-linear models, and neural tools for fitting neural networks which are predictive models.

As Chairperson of the Casualty Actuarial Society's Loss Simulation Model Working Party (LSMWP), he led the development of the Loss Simulation Model that can be used to test loss reserving methods and models. He contributed to the LSMWP paper on "Modeling Loss Emergence and Settlement Processes" that was published in the 2011 Winter CAS E-Forum.

Robert Bear also authored an article on "An Introduction to Extreme Value Theory," that was published in the May 2011 edition of the CPCU International Insurance Interest Group periodical. This topic has attracted much interest after recent natural and financial catastrophes.

Background Image

Our Process

Discuss Goals

Discuss goals of risk model with client so that important risks are modelled without making the model unnecessarily complex.

Research

Determine availability and sources of appropriate company specific data and search for appropriate industry data to supplement company specific data.

Develop & Test Model

Develop model and apply tests of model accuracy and reasonableness. Summarize modeling results for client and determine if further refinements are needed.