Experiments in Economics

Decision Making and Markets

John D Hey

• Éditeur: WSPC
• Année: 2018
• ISBN:
  9789813235823
• Sujet: Economia
• Sous-sujet: Teoria economica

-->

This book provides the most important publications of John D Hey over his almost 50-year career in academia, concentrating primarily on his publications in the field of experimental economics. This is a field that has grown dramatically over the last 30 years, and John D Hey has contributed significantly to its growth and development. The papers included in this volume cover the whole range from individual decision making, both static and dynamic under risk and uncertainty, through games, bargaining and auctions, to markets. The author has contributed in all these fields, and has pioneered much new methodology.

--> Sample Chapter(s)
Introduction
My Experimental Meanderings
Chapter 1: Mixture Models of Choice Under Risk --> Contents:

• Preface
• About the Editor
• Acknowledgements
• Introduction
• My Experimental Meanderings
• Risk:
  • Mixture Models of Choice Under Risk
  • Does Repetition Improve Consistency?
  • Investigating Generalizations of Expected Utility Theory Using Experimental Data
  • Discriminating between Preference Functionals: A Preliminary Monte Carlo Study
• Ambiguity:
  • When and How to Satisfice: An Experimental Investigation
  • The Explanatory and Predictive Power of Non Two-Stage-Probability Theories of Decision Making Under Ambiguity
  • Assessing Multiple Prior Models of Behaviour Under Ambiguity
  • The Descriptive and Predictive Adequacy of Theories of Decision Making Under Uncertainty/Ambiguity
• Dynamics:
  • Strategies in Dynamic Decision Making — An Experimental Investigation of the Rationality of Decision Behaviour
  • Dynamic Decision Making: What Do People Do?
  • Naive, Resolute or Sophisticated? A Study of Dynamic Decision Making
  • How Far Ahead Do People Plan?
• Noise:
  • Why We Should Not Be Silent about Noise
  • Comparing Theories: What Are We Looking For?
  • Are Preference Reversals Errors? An Experimental Investigation
  • Which Error Story is Best?
  • Experimental Investigations of Errors in Decision Making Under Risk
• Multi-Agent Models:
  • Does Money Impede Convergence?
  • What Price Compromise?
  • Experimental Evidence on English Auctions: Oral Outcry Versus Clock
  • Do Markets Drive Out Lemmings — or Vice Versa?

-->
--> Readership: Graduate students and researchers in the field of experimental economics. -->
Experimental Economics;Risk;Ambiguity;Markets;Auctions;Bargaining;Econometrics;Methodology0 Key Features:

• New ways of thinking how to experimentally test economic theories
• New methodological approaches to conducting experiments
• New methods for analysing experimental data

Part 1

Risk

Chapter 1

Mixture models of choice under risk

Anna Conte^a,b, John D. Hey^c,d,∗, Peter G. Moffatt^e

^aStrategic Interaction Group, Max-Planck-Institut für Ökonomik, Jena, Germany

^bCentre for Employment Research, University of Westminster, London, UK

^cLUISS, Rome, Italy

^dUniversity of York, UK

^eSchool of Economics, University of East Anglia, Norwich, UK

A R T I C L E I N F O

Article history:
Available online 13 October 2009

JEL classification:
C15
C2
C51
C87
C91
D81

Keywords:
Expected utility theory
Maximum simulated likelihood
Mixture models
Rank dependent expected utility theory
Heterogeneity

A B S T R A C T

This paper is concerned with estimating preference functionals for choice under risk from the choice behaviour of individuals. We note that there is heterogeneity in behaviour between individuals and within individuals. By ‘heterogeneity between individuals’ we mean that people are different, in terms of both their preference functionals and their parameters for these functionals. By ‘heterogeneity within individuals’ we mean that the behaviour may be different even by the same individual for the same choice problem. We propose methods of taking into account all forms of heterogeneity, concentrating particularly on using a Mixture Model to capture the heterogeneity of preference functionals.

© 2009 Elsevier B.V. All rights reserved.

1.Introduction

As is clear from Starmer (2000), the past five decades have witnessed intensive theoretical and empirical research into finding a good descriptive theory of behaviour under risk. Since the general acceptance of the criticisms of Expected Utility made by Allais (for example, in Allais, 1953) and others, theorists have been active in developing new theories to explain the deficiencies of Expected Utility theory. Hey (1997) provides a list¹ of the major theories at that time: Allais’ 1952 theory, Anticipated Utility theory, Cumulative Prospect theory, Disappointment theory, Disappointment Aversion theory, Implicit Expected (or linear) Utility theory, Implicit Rank Linear Utility theory, Implicit Weighed Utility theory, Lottery Dependent Expected Utility theory, Machina’s Generalised Expected Utility theory, Perspective theory, Prospect theory, Prospective Reference theory, Quadratic Utility theory, Rank Dependent Expected (or Linear) Utility theory, Regret theory, SSB theory, Weighted Expected Utility theory, and Yaari’s Dual theory. All these theories were motivated by the inability of Expected Utility theory to explain all observed behaviour. This burst of theoretical activity took place in the last thirty years or so of the 20th century. Since then, activity has been concentrated more on discovering which of these theories are empirically most plausible and robust; see, for example, Hey and Orme (1994). This period of empirical work revealed clearly that there is considerable heterogeneity of behaviour both between individuals and within individuals. By ‘heterogeneity between individuals’ we mean that people are different, not only in terms of which type of preference functional that they have, but also in terms of their parameters for these functionals. By ‘heterogeneity within individuals’ we mean that the behaviour may be different even for the same choice problem. Econometric investigation has to take these heterogeneities into account.

Some of the empirical literature adopted the strategy of trying to find the best preference functional individual by individual; see, for example, Hey and Orme (1994) and Gonzales and Wu (1999). Another part of the literature attempted to find the best preference functional across a group of individuals, by, in some way, pooling or aggregating the data; see, for example, Harless and Camerer (1994). In fitting data subject by subject, the problem of heterogeneity within subjects becomes immediately apparent in two different ways. First, when confronted with the same decision problem on different occasions, people respond differently. Second, and perhaps more importantly, it was soon realised that none of the long list of preference functionals listed above fitted any (non-trivial) data exactly. Economists responded in their usual fashion — by declaring that individuals were noisy in their behaviour, or that they made errors of some kind when taking decisions. At this point, interest centred on ways of describing such noise and incorporating it into the econometric investigation. A number of solutions were proposed: the constant-probability-of-making-a-mistake model of Harless and Camerer (1994), the Fechner-error model adopted by Hey and Orme (1994), and the random-preference model of Loomes and Sugden (1998), implemented econometrically by Loomes et al. (2002). In the first of these, subjects in experiments are thought of as implementing their choices with a constant error; in the second, subjects were perceived as measuring the value of each option with some error; in the third, subjects were thought of as not having precisely defined preferences, but preferences drawn randomly from some probability distribution. The tremble model, analysed in Moffatt and Peters (2001), can be considered like the constant-probability model but perhaps appended to one of the other two types. A useful discussion of the relative merits of these different models can be found in Ballinger and Wilcox (1997), which concludes that the constant-probability model on its own is dominated by the other two approaches. Further results can be found in Buschena and Zilberman (2000).

Those economists who followed the measurement error story soon realised that the error might not be homoscedastic and could well depend on the nature of the choice problem (see, for example Hey, 1995). Indeed, Blavatskyy (2007) argues that, with the appropriate heteroscedastic error specification, Expected Utility theory can explain the data at least as well as any of the generalisations (after allowing for degrees of freedom). Not all would go as far as this, but the incorporation of some kind of error story has led to the demise of many of the theories noted in the list above. Two remain pre-eminent: Expected Utility – henceforth EU – theory; and Rank Dependent Expected Utility – henceforth RDEU – theory (Quiggin, 1982). Machina (1994) comments that the Rank Dependent model is “the most natural and useful modification of the classical expected utility formula”. In certain contexts, for example the Cumulative Prospect theory of Tversky and Kahneman (1992), the theory is enriched with a context dependent reference point. Nevertheless, the consensus seems to be that EU theory and RDEU theory remain the leading contenders for the description of behaviour under risk.

As we have already remarked, some of the investigations of the appropriate preference functional have taken each individual separately and have carried out econometric work individual by individual. There are problems here with degrees of freedom and with possible over-fitting. Other investigations have proceeded with pooled data — from a set of individuals. The problem with this latter approach, even though it saves on degrees of freedom, is that individuals are clearly different. They are different, not only in terms of which type of preference functional that they have, but also in terms of their parameters for these functionals. The latter can be taken care of by assuming a distribution of the relevant parameters over the individuals concerned and in estimating the parameters of this dis...

Table des matières

• Cover
• Halftitle
• Title
• Copyright
• Preface
• About the Editor
• Acknowledgements
• Contents
• Introduction
• My Experimental Meanderings
• Part 1: Risk
• Part 2: Ambiguity
• Part 3: Dynamics
• Part 4: Noise
• Part 5: Multi-Agent Models