Published by EH.Net (August 2019)
Ivan Moscati, Measuring Utility: From the Marginal Revolution to Behavioral Economics. New York: Oxford University Press, 2019. ix + 326 pp. $35 (paperback), ISBN: 978-0-19-937277-5.
Reviewed for EH.Net by Harro Maas, Centre Walras-Pareto, University of Lausanne.
Measuring Utility tells the story of a fundamental problem in the history of economics that arose after the labor theory of value gave way to a theory that took marginal utility, in one or another form, as a determinant of prices in the market. Is utility measurable and if so, how to measure it? This question was first posed by William Stanley Jevons, who admitted there was no unit of measurement. Jevons added this did not mean no unit of measurement could be found in the future. Jevons’s question lost its urgency in the economic mainstream once economists realized they could derive their most important theoretical results without measuring utility at all, but in the process concepts of measurement and of utility themselves changed substantially.
Most economists will probably nod approvingly and say they know the history of this problem — which, of course, was solved when the concept of utility changed from cardinal to ordinal and became nothing more than a numerical index for the preferred notion of preferences. When pressed by the historian that questions about the measurability of utility returned after the Second World War with the rise of Expected Utility Theory (EUT) and then behavioral economics, they will probably start muttering that this historian should please let them get on with their job. But in the economists’ standard narrative, it remains an enigma how an ordinal and cardinal conception of utility could happily live side by side in the postwar period.
Ivan Moscati attacks this enigma head-on by reducing the history of the measurability of utility to its barebones. Moscati provides just sufficient information about the economists involved to know who they are, and where and with whom they interacted. He also discards larger questions about welfare and other theories that might complicate this history. The seemingly trifling addition of ratio-measurement to the standard distinction between cardinal and ordinal utility enables Moscati to open up a rich history that involved some of the greatest minds in economics, psychology and measurement theory. Moscati thus follows the “dream of Jevons” to measure utility, from the marginalist revolution in the last quarter of the nineteenth century, via the rise of Expected Utility Theory around mid-twentieth century, until the increasing skepticism about the validity of EUT with the rise of laboratory experimentation and behavioral economics in the mid-1970s.
Moscati’s distinction between ratio, cardinal and ordinal utility motivates the introduction of a second story line on measurement theory. Moscati follows this story line from the late nineteenth century until the theory of measurement that became a reference point after the Second World War: the theory of measurement of Patrick Suppes and his collaborators that axiomatized measurement as a mapping of objects into numbers. In the process, measurement changed from the nineteenth century notion of ratio-scale measurement, in which one was searching for a fundamental unit of measurement, to a conception of measurement as the ordering of objects on progressive criteria that end in the most constrained form of ordering: ratio-scale measurement, or equivalently, unit measurement. Thus equipped, Moscati explains how nineteenth century economists were searching for ratio-scale or unit measurement, something more restrictive than cardinal measurement (a concept they never used), while twentieth century economists loosened their utility concept to the even less restrictive ordinal scale. Cardinal utility could then re-enter the scene in the post-war period, because the axioms of EUT did not imply a return to ratio-scale measurement, and were not concerned with the shape of preferences, but with ordering conditions on decision-making under risk. When economists started to test predictions about these decisions in laboratory settings, the mood changed from the endorsement of EUT to increasing skepticism about its empirical merits, and a far broader range of scientists than merely economists reinvigorated efforts to find a mental or neurological unit of measurement.
The set-up of Moscati’s book follows this storyline. The book is divided in four parts which each start with a chapter on developments in measurement theory. This is followed with a detailed account of debates about the measurability of utility. Each part summarizes the results along five themes, the meaning of measurement, the scope of the utility concept, its status, what data are relevant for its measurement, and what utility theory aims to explain. The first part covers the early marginalists and early twentieth century economists still adhering to the idea of measuring utility as a unit, such as Eugen von Böhm-Bawerk or Irving Fisher. The second part covers debates about whether utility is ordinal or cardinal, starting with Vilfredo Pareto, moving to John Hicks and Roy Allen, and ending with Paul Samuelson’s Foundations of Economic Analysis. This part includes a concise discussion of Norman Campbell’s theory of measurement, which still hinged on ratio-scales as the gold standard of measurement, as well as succinct discussions of Ragnar Frisch and Léon Thurstone’s early attempts to measure utility or indifference curves respectively. Responses to both attempts were largely negative, in the case of Thurstone because of fundamental criticism of his hypothetical methodology that lacks proper incentives to elicit true preferences, canonized as the Wallis-Friedman critique.
Part three introduces the psychologist Stanley Smith Stevens’ instrumentalist theory of measurement and the debate between Milton Friedman, Leonard Savage, Jacob Marschak, Samuelson and William Baumol about the relation between Expected Utility Theory and standard ordinal utility theory. Moscati guides readers through a minefield of distinctions in which utility functions with regard to EUT and standard, riskless utility theory, became gradually cleared up, up to the point that all American economists involved were drawn to Friedman and Savage’s side and endorsed EUT as normatively compelling. Moscati then confronts this new American consensus with the French skepticism at the Paris conference in 1952 where Maurice Allais articulated his famous paradox and with the instrumentalist interpretation theory measurement received after Friedman’s equally famous essay on economic methodology of 1953. As a result, the relevance of the distinction between cardinal and ordinal utility vanished and the emphasis was placed on the theory’s predictive value. These chapters, on which Moscati has published earlier, form the nucleus of the book.
Moscati is to be lauded for guiding his audience with a clear voice through materials that are by no means easy to disentangle. This is a remarkable accomplishment. The book loses some of its force in the fourth part, which to me reads too much as a survey of experimental efforts to measure utility where I would have expected a sustained contrast between Stevens’ operationalism and the theory of measurement of Suppes et al. Moscati rightly questions whether Stevens can be considered to defend a representational theory of measurement, but too easily classifies Suppes et al. as representational. Representational of what? Stevens rightly questioned whether Suppes’ measurement theory can be considered measurement if the actual, empirical procedures of measurement are never specified. Stevens’ concerns could have led to a sustained reconsideration of the ontological commitments of economists and psychologists, conveniently distinguished in the preface as “mentalist” versus “instrumentalist.”
Does this bring us back to the nineteenth century and the search for a something to be measured –a unit of measurement? It seems so. In the epilogue, Moscati signals that all directions in recent research in utility measurement adopt a mentalist view, which effectively amounts to a search for a mental something to be measured on a ratio-scale. Instead of signaling this, this reviewer would have appreciated a sustained discussion of its implications where it seemed most fit: in the confrontation between the axiomatic and operationalist view on measurement. This remark does not diminish the value of Moscati’s book. For years to come, Measuring Utility will be the work of reference for those seeking to the understand the meaning and significance of debates about the measurability of this central concept in microeconomic theory.
Harro Maas is a professor in history and methodology of economics at the Centre Walras-Pareto for the history of economic and political thought at the University of Lausanne, Switzerland. He is the author of William Stanley Jevons and the Making of Modern Economics (Cambridge University Press, 2005), Economic Methodology: A Historical Introduction (Routledge, 2014), and co-editor (with Mary S. Morgan and Andrej Svorenčík respectively) of Observation in Economics: Historically Considered (Duke University Press, 2012) and Experimental Economics: Witness Seminar on the Emergence of a Field (Springer, 2016). He is currently working on a history of consumer governance from the early nineteenth century to the present.
Copyright (c) 2019 by EH.Net. All rights reserved. This work may be copied for non-profit educational uses if proper credit is given to the author and the list. For other permission, please contact the EH.Net Administrator (administrator@w3needs.com). Published by EH.Net (August 2019). All EH.Net reviews are archived at https://eh.net/BookReview.