“I approached the problem of utility measurement in 1923 during a stay in Paris. There were three objects I had in view :
:(I) To point out the choice axioms that are implied when we think of utility as a quantity, and to define utility in a rigorous way by starting from a set of such axioms;
:(II) To develop a method of measuring utility statistically;
:(III) To apply the method to actual data.
The results of my study along these lines are contained in a paper “Sur un Problème d’Économic Pure”, published in the Series Norsk Matematisk Forenings Skrifter, Serie I, Nr 16, 1926. In this paper, the axiomatics are worked out so far as the static utility concept is concerned. The method of measurement developed is the method of isoquants, which is also outlined in Section 4 below. The statistical data to which the method was applied were sales and price statistics collected by the “Union des Coopérateurs Parisien”. From these data I constructed what I believe can be considered the marginal utility curve of money for the “average” member of the group of people forming the customers of the union. To my knowledge, this is the first marginal utility curve of money ever published.”

Frisch (1932) New Methods of Measuring Marginal Utility. Mohr, Tübingen. p. 2-3: Quoted in: Dagsvik, John K., Steinar Strøm, and Zhiyang Jia. &quot; A stochastic model for the utility of income http://www.ssb.no/a/publikasjoner/pdf/DP/dp358.pdf.&quot; (2003). <br class="br">1930s