Subjective Survival Probabilities and Life Tables: Evidence from Europe

Understanding the variability of survival probabilities, both between and within cohorts, is important in order to understand life-cycle decisions under uncertainty. In this paper we analyze the subjective probabilities of survival to specific target ages provided by respondents to the Survey of Health, Ageing and Retirement in Europe (SHARE). To evaluate how these probabilities compare with objective data from life tables, and avoid the problems associated with a naive use of period life tables, we construct cohort life tables from the sequence of period life tables available in the Human Mortality Database and use them to compute actuarial probabilities of survival to the same target ages. We find that male subjective survival probabilities are close to the probabilities computed from our cohort life tables, whereas female subjective probabilities are always lower. We also find that subjective survival probabilities are on average higher for more educated people, those with higher household income, and those with better health. This evidence suggests that both income and health matter for own assessments of subjective survival.