Actuarial Review Return to Main Page
Flashlight on the Globe

Focus on International Research

ASTIN in Bergen

by Kris DeFrain, Member, CAS International Research Committee

This June, the International Actuarial Association and the Norwegian Actuarial Society hosted the 2004 ASTIN Colloquium in Bergen, Norway. Keynote addresses focused on climatic changes and insurance, insurance fraud from an actuarial perspective, genetics and insurance, and future challenges to actuarial science.

Climatic Changes and Insurance
Hans von Storch from the Institute for Coastal Research in Germany noted that perceived changes in storm climate do not appear factual. Insurance damages appear worse over time, but not if one adjusts for insured values. He cautioned against reliance on historical data, such as a long series of wind observations. Observation methods differ significantly over time, citing a change from visual observation to scientific measurement or a change from measuring wind by the ocean to measuring at an airport as examples.

David Anderson from Guy Carpenter in Sweden validated there is no clear-cut trend in natural peril losses in Scandinavia. Economic losses have doubled globally every ten years since the 1980s, but there have been significant increases in insured values.

Gerhard Berz of Munich Re in Germany stated that the worldwide-insured costs of great natural disasters have risen by a factor of almost 14 (after adjusting for inflation). Events and losses have been dominated by weather disasters. Reasons for the increases in frequency and severity include an increase in population, better standards of living, a higher concentration of people and value in large conurbations, an increase in insurance density, settlement in exposed areas, and a change in environmental conditions.

Insurance Fraud: Actuarial Perspective
Tommy Short, past president of the International Association of Special Investigation Units, U.S., described how costly insurance fraud is to both the industry and the insurance consumer. Numerous factors have combined to contribute to rising fraud costs including the Internet, lowered trade barriers, and the ease of moving across borders. Short described numerous examples of fraud. People acting alone (with multiple identities and cars), people acting in a group (with doctors, lawyers, and claim handlers involved), and organized crime can be players in fraud.

Montserrat Guillen, from the University of Barcelona in Spain, described the analysis of Spanish automobile insurance data to create a modeling system to score potential fraud. The models use a scoring system based on information about the insured, the vehicle, and the claim.

Irene Laegreid, of the SpareBank 1 in Norway, shared an insurance fraud statistical study based on Norwegian data. Factors such as age and sex were studied. Seasonal variation was found. Women were more likely to commit fraud in the early part of the year, right when Christmas bills are due. In auto, more problems occurred in February through April when new cars are being purchased. There were variations between urban and country areas and variations by line of business.

Richard Derrig, of OPAL Consulting and a visiting scholar at the Wharton School, University of Pennsylvania, noted that the differences between fraud and abuse should be considered instead of grouping everything under the “fraud” term. It is important for actuaries to get involved in this subject. Data is essential and patterns need to be researched.

Genetics and Insurance
D. Timothy Bishop, of Cancer Research U.K., University of Leeds in England, provided some scientific information about the identification and characterization of genes. Genes are often identified using family studies, and more recently through causation studies. Characterization of genes helps identify the pattern of mutation and calculates its population prevalence. Studies have shown, for example, that breast cancer risk increases significantly if the mother is affected and increases more when the relative has cancer early in life.

William Nowlan, of National Life Insurance in the U.S., discussed the use of genetic typing. He noted there were numerous findings of genes in the 1990s, but none recently. Even when a gene is discovered, it takes decades to determine treatment. DNA-based study might not be all that is needed. It’s possible that proteins, chemicals, and RNA might each play a role. He noted that family history is rarely used except in preferred underwriting. Most gene impairments are already evident from mammograms or x-rays by the time someone decides to purchase life insurance.

Piet de Jong, of Macquarie University in Australia, discussed the costs of adverse selection. Use of family history and other correlated information can limit losses and decrease required loadings. He asked whether the opportunity costs of not using some of the available information could be quantified.

Challenges to Actuarial Science in the 21st Century
Jean Lemaire, of the Wharton School in the U.S., provided a prediction of the needs for future actuarial science education. He advised actuaries to concentrate on in-depth knowledge of core information. A concentration on breadth of knowledge may result in loss of actuarial uniqueness compared to other financial analysts.

Working Sessions
Papers were presented in working sessions with focus on the following topics:

The papers from the working sessions are available on the colloquium Web site.

Click here to write a Letter to the Editors