Academian comments on Harsanyi’s Social Aggregation Theorem and what it means for CEV

[Academian]

Non-VNM agents satisfying only axiom 1 have coherent preferences… they just don’t mix well with probabilities.