Probabilistic Survey Questions and Incorrect Answers: Retirement Income Replacement Rates

Abstract

We use probabilistic survey questions to analyze the determinants of individuals’ expected value of and the uncertainty in the retirement income replacement rate. We find that the expected replacement rate is U-shaped in age and is substantially lower for highereducated individuals. Uncertainty in pension income decreases with age, but increases with education. An important aim of this paper is to investigate if the use of probabilistic survey questions yields an endogenous sample selection by removing individuals that provide answers incompatible with probability theory. We find that not accounting for endogenous sample selection biases the results toward more pessimistic expectations and excess uncertainty in the replacement rate. These biases are most prevalent for less-educated individuals.