Controlling likelihood beliefs even if these are unknown

ERIM Members Prof. Baillon and Prof. Wakker from Erasmus School of Economics published, together with Zhenxing Huang (Shanghai University) and Asli Selim (eBay), an article in Econometrica. The paper introduces two indexes of ambiguity attitudes (for aversion and insensitivity/perception), for which subjective beliefs can be controlled for even if those are unknown.

Problem of unobservable subjective beliefs

Most uncertainties in our everyday life, business decisions, and economic decisions, come with no probabilities specified (“ambiguity”). A problem for measurements of attitudes towards such ambiguity has, as yet, been that subjective beliefs confound them. A person may rather bet on a flip of a fair coin than on rain tomorrow, not because she is averse to the unknown (“ambiguous”) probability of rain, but because of a subjective belief that there will be no rain. It was unknown how to correct for such, unobservable, subjective beliefs. Measurements therefore were confined to artificial “Ellsberg urns,” where subjective beliefs could be controlled for, but this greatly hamperied validity and applicability.

Possibility to measure ambiguity

If a person prefers betting on heads up rather than on rain, and also on tails up rather than on no rain, then the person must have a special aversion to ambiguity. There cannot be a strong belief against rain and at the same time against no rain. In this and other clever combinations of several bets, the unknown beliefs drop from the equations and ambiguity attitudes still come out. Only now it becomes possible to measure ambiguity attitudes in situations that are relevant to our decisions, concerning investments, health treatments, policy decisions, everyday life, actions of others, and so on. The new method has already been applied to stock performances, written language, trust, and strategic moves of opponents (game theory). Several others are in preparation by teams in several parts of the world.

Changing views

This paper gives some surprising new insights. People have no particular aversion to uncertainties but instead, and more basically, they do not understand them. This could not show up with the artificial Ellsberg urns, but now fundamentally changes our views on ambiguity. The classical paradox of the co-existence of gambling and insurance can now be readily explained, as can the common underevaluations of precautionary actions such as safety belts. Nudging techniques should primarily go against cognitive misunderstandings, rather than just against aversion as thought hitherto.

Prof. Aurélien Baillon

Prof. Baillon (1980) joined Erasmus School of Economics in 2007, after obtaining a research master at the University of Toulouse and a Ph.D. at Arts et Métiers ParisTech. In the past years he was awarded a Veni and a Vidi grant from the Netherlands Organization for Scientific Research (N.W.O.) and a Starting Grant from the European Research Council. His research has been published in leading journals, including the American Economic Review and PNAS.

Prof. Peter Wakker

Prof. Wakker (1956) is professor of decisions under uncertainty at the Econometric Institute of Erasmus School of Economics (ESE). He works in behavioral economics.  He has joint papers with three Nobel prize winners, published in top journals in economics, psychology, health, and mathematics, was best Dutch economist four times, and received many awards.

Information about the article

Measuring Ambiguity Attitudes for All (Natural) Events

Aurélien Baillon, Zhenxing Huang, Asli Selim, Peter P. Wakker

Measurements of ambiguity attitudes have so far focused on artificial events, where (subjective) beliefs can be derived from symmetry of events and can be then controlled for. For natural events as relevant in applications, such a symmetry and corresponding control are usually absent, precluding traditional measurement methods. This paper introduces two indexes of ambiguity attitudes, one for aversion and the other for insensitivity/perception, for which we can control for likelihood beliefs even if these are unknown. Hence, we can now measure ambiguity attitudes for natural events. Our indexes are valid under many ambiguity theories, do not require expected utility for risk, and are easy to elicit in practice. We use our indexes to investigate time pressure under ambiguity. People do not become more ambiguity averse under time pressure but become more insensitive (perceive more ambiguity). These findings are plausible and, hence, support the validity of our indexes.