Abstract
Micro data from a dental insurance natural experiment is used to analyze why agents opt out of insurance. The purpose is to relate the dropout decision to new information on risk, acquired by the policy holder and the insurer. The results show that agents tend to leave the insurance when reclassified into higher premium classes, or when experiencing unexpectedly low dental care costs within the insurance. They are more responsive to higher premiums than to lower expected costs. The results show updating on dental risk to be asymmetric, giving agents and insurer partly different information sets. Agents do not take the insurer's information fully into account, even though it is public, and higher premiums are viewed as higher prices on insurance, rather than a fair risk reassessment. The decision is also based on old information, indicating that the cognitive process of updating information is slow.