Rounding patterns of survival probabilistic expectations and their consequences on inference.

The focus of this thesis is on subjective expectations, how they are formed and their consequences in the financial market participation. Data on subjective expectations are taken from the SHARE survey: the Survey of Health, Aging and Retirement in Europe, which will also allow to make comparisons of results across countries. The thesis analyze the measurement error issue when using subjective expectations, due in particular to the fact that people tend to "round up" their responses by often using easily communicable values. I estimate the magnitude of this measurement error through an algorithm that transforms individuals' point responses into interval responses based on individual response patterns. In the final part of the thesis, I present an empirical application in which I analyze the relationship between investment in the financial market and the probability with which the individual believes he will survive beyond a certain target age. I compare model estimates using point responses and interval responses in turns. The results suggest that measurement error in this specific empirical application has no significant impact on the conclusions.