A cause of preference is not an object of preference

Abstract

Welfare economists sometimes treat a cause of preference as an object of preference. This paper explains that this is an error. It examines two examples where the error has occurred. One is from the theory of endogenous preferences. The other is from the theory of extended preferences. People have erroneously been led to believe that everyone must have the same extended preferences, and this has led them to think that extended preferences can be the basis of interpersonal comparisons of wellbeing. But actually the basis of interpersonal comparisons must come from elsewhere.