An 'ordinal utility' theory of choice under uncertainty

Using an ordinal approach to utility, in the spirit of Hicks (1962, 1967a), it is possible to generalize the theory of choice under uncertainty. The basic assumption is to summarize any probability distribution into its moments, so that preferences over distributions can be mapped into preferences over vectors of moments. This implies that ‘uncertain prospects’ (assets, portfolios, lotteries etc.), like Lancaster’s (1966) consumption goods, are bundles of characteristics, and can be directly priced, at the margin, in terms of their moments. The ‘independece axiom’ and expected utility are not required and St. Petersburg’s, Allais’s and other paradoxes may easily be solved.