A central theme in the international debate on genetic testing concerns the extent to which insurance companies should be allowed to use genetic information in their design of insurance contracts. We analyze this issue within a model with the following important feature: A person’s well-being depends on the perceived probability of becoming ill in the future in a way that varies among individuals. We show that both tested high-risks and untested individuals are equally well off whether or not test results can be used by insurers. Individuals who test for being low-risks, on the other hand, are made worse off by not being able to verify this to insurers. This implies that verifiability dominates nonverifiability in an ex-ante sense.