Testing life-cycle models and other economic models of saving and consumption at micro level requires knowledge of individuals' subjective believes of their mortality risk. Previous studies have shown that individual responses on subjective survival probabilities are generally consistent with life tables. However, survey responses suffer serious problems caused by focal responses of zero and one. This paper suggests using a Bayesian update model that accounts for the problems encountered in focal responses. We also propose models that help us to identify how much each individual deviates from life table in her subjective belief. The resulting individual subjective survival curves have considerable variations and are readily applicable in testing economic models that require individual subjective life expectancies.
