Loss Models: From Data to Decisions, Second Edition

Rating:
Summary: great introduction to models needed in insurance
Review: When I took a job to model prediction of loss reserves for workers compensation insurance, I began to realize that the traditional statistical methods that I generally relied n would not help me (without modification). The required modification would be either to transform variables or to model long-tailed probability distributions. This is because in the insurance business you have to reserve for those big catastrophies. The cost data for workers compensation data generally show a high frequency of low to moderate costs (say in the range of $1000 to $50,000). However occasionally there are a few cases of severe injury causing permanent disability which could run over 1 million dollars. Even though the probability of occurrence is small the cost is so high that it cannot be ignored. Such claims will surely be found when large insurance company cover millions of employees over many years.

The problem occurs when insuring for floods, earthquakes, fires and other disasters. Stuart Klugman and Bob Hogg in 1984 wrote the first introductory text to acquaint statisticians with such probability models that are important in the insurance business. Other books covering the subject were covered in books on risk theory designed for actuaries. This book covers all the topics and assumes mathematical and staistical knowledge at the level of the book by Hogg and Craig (so some calculus is required).

Rating:
Summary: great introduction to models needed in insurance
Review: When I took a job to model prediction of loss reserves for workers compensation insurance, I began to realize that the traditional statistical methods that I generally relied n would not help me (without modification). The required modification would be either to transform variables or to model long-tailed probability distributions. This is because in the insurance business you have to reserve for those big catastrophies. The cost data for workers compensation data generally show a high frequency of low to moderate costs... . However occasionally there are a few cases of sever injury causing permanent disability which could run over 1 million dollars. Even though the probability of occurrence is small the cost is so high that it cannot be ignored. Such claims will surely be found when large insurance company cover millions of employees over many years.

The problem occurs when insuring for floods, earthquakes, fires and other disasters. Stuart Klugman and Bob Hogg in 1984 wrote the first introductory text to acquaint statisticians with such probability models that are important in the insurance business. Other books covering the subject were covered in books on risk theory designed for actuaries. This book covers all the topics and assumes mathematical and staistical knowledge at the level of the book by Hogg and Craig (so some calculus is required).