|
Activity Number:
|
318
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Risk Analysis
|
| Abstract - #300706 |
|
Title:
|
Endogeneity of an Ordered Explanatory Variable in a Bivariate Model: Accidents and Deductible Choice for Young Drivers
|
|
Author(s):
|
Michel Grun-Rehomme*+ and Olga Vasechko and Noureddine Benlagha
|
|
Companies:
|
Université Paris2, ERMES and Research Institute of Statistics and Université Paris2, ERMES
|
|
Address:
|
12 place du panthéon, paris, 75005, France
|
|
Keywords:
|
Bivariate ordinal model ; automobile insurance ; adverse selection ; accident probability
|
|
Abstract:
|
Automobile insurers, using different observable explanatory factors to estimate the likelihood of accidents, have sought to better align premiums with accident probabilities. This raises the question of whether the insured drivers have changed their risk behavior. We analyze the coverage in the French automobile insurance market for information about the risk-aversion of drivers. This study looks at conditions of the insurance market. We consider four types of deductible and four frequencies of accidents (0, 1, 2 and more than 2). Autonomous estimates of the number of accidents may include an endogeneity bias. Introducing an accident variable with four modalities into the bivariate modeling enables us to test the significance of engine power in the equation of the coverage choice.
|
