A measure of rationality and welfare

Evidence showing that individual behavior often deviates from the classical principle of preference maximization has raised at least two important questions: (i) How serious are the deviations? and (ii) What is the best way to analyse choice behavior in order to extract information for the purpose of welfare analysis? This paper addresses these questions by proposing a new way to identify the preference relation which is closest, in terms of welfare loss, to the revealed choice.