Essays on decision-making under risk and ambiguity in theory and laboratory experiments
Open access
Autor(in)
Datum
2017Typ
- Doctoral Thesis
ETH Bibliographie
yes
Altmetrics
Abstract
This dissertation discusses selected aspects of choice behavior under risk and ambiguity in theory and laboratory experiments. The first chapter briefly introduces the Ellsberg paradoxes and the economic literature on choice under ambiguity that has emerged in response to them. In particular, the chapter shows how Ellsberg's thought experiments challenge the subjective expected utility hypothesis and the preference axioms underlying it. It then describes how alternative preference models can accommodate the Ellsberg paradoxes and more generally ambiguity averse behavior. Chapter 2 focusses on the smooth ambiguity model by Klibanoff, Marinacci, and Mukerji (2005). The chapter first suggests a characterization of increases in risk aversion within the smooth ambiguity model. It then shows that an increase in risk aversion is qualitatively different from that under expected utility, due to the incomplete separation between risk and ambiguity attitude. The chapter notably clarifies how ambiguity perception and attitude depend on risk aversion. A portfolio choice problem illustrates that comparative statics results on risk aversion in the smooth ambiguity model can differ from those obtained under the standard expected utility framework. In chapter 3, I present a method to systematically test independence axioms and compare decisions under risk and under ambiguity in laboratory experiments. Using an Anscombe-Aumann setting, I show that acts and mixture operations between acts can be represented in an intuitive and straightforward way using Ellsberg urns. Consequently, decision makers' choices between bets on Ellsberg urns directly reveal preferences over Anscombe-Aumann acts and possibly violations of independence. Importantly, the same urns can be used to observe choices under risk and under ambiguity: Under risk, the composition of the urns is revealed to the decision maker, while under ambiguity it is not. The design is thus suited to compare violations of independence under risk to violations of independence under ambiguity. In chapter 4, I apply the method developed in chapter 3 to an incentivized laboratory experiment. The experiment is designed to compare Allais-type violations of independence under risk and under ambiguity. I find systematic violation of independence due to Allais-type common ratio effects under ambiguity but not under risk. At the same time, I do not observe strong evidence for systematic violations of independence due to Allais-type common consequence effects either under risk or under ambiguity. The results suggest that ambiguity models which are compatible with common ratio effects have a descriptive advantage over models which are not. Moreover, the results indicate that systematic violations of independence - and hence deviation from expected utility - are more prominent under ambiguity than under risk. Mehr anzeigen
Persistenter Link
https://doi.org/10.3929/ethz-b-000179562Publikationsstatus
publishedExterne Links
Printexemplar via ETH-Bibliothek suchen
Verlag
ETH ZurichThema
Microeconomics; Decision theory; Experimental economics; AmbiguityOrganisationseinheit
03877 - Bommier, Antoine / Bommier, Antoine
Zugehörige Publikationen und Daten
ETH Bibliographie
yes
Altmetrics