Ambiguity in Markets and Social Interactions: Theory and Applications – AMSI

In this project, we take seriously the concern that information from data, while objective, might not be perfectly suitable for the prediction to be made. Data sets differ according to their degree of precision and relevance to the prediction at hand. We allow the decision maker to express preferences for the type and quantity of information he would like to receive. We can use these preferences to elicit the uncertainty he associates with different sources of information, but also his attitude to uncertainty. In turn, these two characteristics will determine whether and how much he will be willing to pay in order to obtain additional information. This question is highly relevant in models, in which information is endogenously generated by the actions of the market participants. For example, in a model of technological adoption, ample amounts of data will be available about traditional technologies, whereas data on innovations will be scarce. Innovation thus presents a dilemma for pessimistic agents: while they are the ones who would most benefit from experimenting with the new technology and generating additional data about its performance, they are also those who are most cautious in choosing uncertain alternatives. A choice of a market, in which to trade presents another example: market institutions differ w.r.t. the type and quality of information they provide to market participants. At the same time, it is the agents’ behavior in the market that determines prices and allocations and thus generates the information available in the market. Our theoretical model links beliefs (and thus, behavior) to available data, and allows us to analyze the dynamics of such complex systems.

The project will generate the following five milestones:
Milestone 1: A model of decisions informed by data capturing ambiguity and preferences for information. We will identify the decision maker’s attitude towards ambiguity and relate it to his willingness to pay for additional information.
Milestone 2: Study of the evolution of attitudes towards ambiguity and their impact on learning and social outcomes. In this model, we are interested in the evolution of optimistic and pessimistic sentiments and their effect on prices and allocations, as well as on the information provided by the markets.
Milestone 3: A model of technology adoption in response to climate change: analysis and policy implications. In this part, we examine the conditions under which a society will adopt a new technology, in order to adapt to changing climate conditions. We study the dynamics of the adoption process and evaluate policies which lead to fast adoption.
Milestone 4: A study of the long-run impact of heterogeneity in ambiguity attitudes on prices and allocations in financial market. In this model, we ask whether investors who are neutral towards ambiguity will be more successful in the long-run than those who are averse to ambiguity. This is a test of whether ambiguity and ambiguity aversion have persistent effect on asset prices.
Milestone 5: A study of the impact of ambiguity on market participation and market structure. We will analyze the dynamics of prices and allocation and provide guidance for institutional design and market regulation.

We envision that the methodological contributions of this project will be applicable to a wide range of important economic phenomena in the areas of macroeconomics, finance and industrial economics, besides those indicated in this project. The results of all stages of the project allow us to generate empirically testable predictions. In particular, the results of Milestone 1 can be tested using an experimental design. The results of the other tasks can be taken to data, tested in an empirical model and used to generate predictions. Furthermore, the results obtained in Milestones 3, 4 and 5 generate important policy implications and provide guidance for institutional design. Hence, our theoretical results will have a broader impact on the economic community than indicated by the scope of the project itself and can be put in the service of the global community.

Since the start of the project, Juergen Eichberger and Ani Guerdjikova have co-written two papers: “Ambiguity, Data and Preferences for Information: A Case-Based Approach” and “Technology Adoption and Adaptation to Climate Change”. The first paper provides the basic theoretical model for decision making when information is given by data. The second studies technology adoption in view of adaptation to climate change. These papers have been presented at several international conferences, notably RUD 2012 and SAET 2012 and will be presented at the European Meetings in Malaga in August 2012. A one-day workshop on “New Developments in Decision Theory under Uncertainty” has been held in May 2012 at the University of Cergy-Pontoise. The workshop presented recent work in decision theory and provided researchers with a forum for exchange and discussion on the topic.

Most economic environments are characterized by uncertainty. Economic agents thus have to make decisions, the consequences of which cannot be known ex ante. Decision theory studies this process of decision making and tries to identify how the description of uncertainty influences beliefs, and through them, the evaluation of uncertain acts. This, in turn, allows us to make testable predictions about behavior and study the implications of uncertainty on economic outcomes.
Ambiguity, i.e., the fact that decision makers face uncertainty about the probability distribution governing the outcomes, has been found to have a significant impact on human behavior. In this project, we study decision making under ambiguity when information is provided in the form of data. While data carry objective information about the stochastic process of outcomes, this information might be insufficient to uniquely identify the probability distribution of outcomes. We examine the relation between objective information provided by data and decision maker's beliefs about uncertain outcomes. We analyze how the type and frequency of observations combine with the subjective characteristics of the decision maker to determine his beliefs and, in turn his behavior in view of this ambiguity.
An important question in decision theory concerns the value of additional information. While it is known that for Bayesian agents, additional information is always beneficial, this need not be the case for agents who are not neutral towards ambiguity. We thus study the value of acquiring additional data as a function of the characteristics of the data and the individual characteristics of the agent, such as his degrees of optimism and pessimism.
In the second part of the project, we apply the framework developed above to different economic problems characterized by ambiguity. We study the behavior of ambiguity-averse investors in a financial market and identify the long-run impact of ambiguity and ambiguity-aversion on equilibrium prices and allocations. We analyze the process of technology adoption as an important mechanism of adaptation to climate change. In particular, we examine the impact of optimism and pessimism on learning and on optimal technology choice. We also evaluate different policies designed to stimulate early adoption. Finally, we study the decision of market participation, when markets provide different type and different quantity of information. We show that agents self-select into markets depending on their preferences for information, which, in turn determines the market structure in a given economy. We use these insights to derive implications for the optimal design and regulation of markets.