Testing for asymmetric information and learning: evidence from the Korean auto insurance market

Abstract

This article draws on a set of data from the Korean automobile insurance industry to investigate the presence of asymmetric information and learning. I find that there exists a conditional correlation between the choice of Collision and Comprehensive coverage and the probability of accidents. This finding implies that the better insured actually cause more claims, although the difference in the number of claims seems less significant in economic contexts. I confirm that the degree of information asymmetry is more significant in contracts with new policyholders, young drivers with less than 3 years of driving experience, and senior drivers with more than 7 years of experience. Tests using premiums provide results consistent with tests considering variables related to the characteristics of policyholders, characteristics of their vehicles, and the contract choices they make.