On expected utility, part 1: Skyscrapers and madmen

Abstract

This work examines three main theorems seeking to demonstrate the rationality of expected utility maximization (EUM). The first is von Neumann-Morgenstern (vNM), which assumes probability assignments and derives EUM as a conclusion of four axioms. The second is a theorem relying on a separability-additivity argument. And the third is a direct axiomatization of EUM by Peterson (2017). The article offers general overviews of these theorems, provides informal arguments for how the underlying reasoning works, and references more formal resources for further study. – AI-generated abstract.