Anat Bracha

Applied Microeconomic Theory
Decision Making under Uncertainty
Psychology and Economics

Microeconomic Theory
Decision Theory
Psychology and Economics

May 2001 (Oral) International Economics, Game Theory
May 2000 (Written) Microeconomic and Macroeconomic Theory

Multiple Selves and Endogenous Beliefs

Economics
   Professor Donald Brown
   Professor Stephen Morris
   Professor Leeat Yariv
Psychology
   Professor David Armor

May 2005

Ph.D., Economics, Yale University, expected May 2005
M. Phil., Economics, Yale University, December 2002
M.A., Economics, Yale University, May 2001
B.A., Economics, Bar-Ilan University, Israel, Summa Cum Laude, 1997

Whitebox Advisors Doctoral Fellowship, 2004-05
Cowles Foundation Prize, Yale University, 2003
The Georg Walter Leitner Grant for political economics, 2002
John Perry Miller Fund Fellowship, 2002 (declined)
Yale Fellowship, Sept. 1999 – 2003
Rector prize for the top student in economics (graduate level), Bar-Ilan University (Ramat-Gan, Israel), 1998
Dean honor for top 3 percent students in economics (undergraduate level), Bar-Ilan University (Ramat-Gan,
     Israel), 1994-1997

Instructor
     Introductory Microeconomics, Yale University, Summer 2003
Head Teaching Assistant
     Introductory Microeconomics, Yale University. Coordinated and managed teaching assistants, Fall 2003
Teaching Assistant
     Intermediate Microeconomics, Yale University, Fall 2001
     International Trade (Graduate level), Yale University, Sprincg 2002
     Game Theory (Undergraduate level, Math Econ), Yale University, Spring 2003
     Introductory Microeconomics, Fall 2003, Spring 2004
     Microeconomic theory (Undergraduate level, Math Econ), Yale University, Fall 2004
     Introductory Microeconomics, Bar-Ilan University, Israel, 1998-1999
     Introductory Macroeconomics, Bar-Ilan University, Israel, 1998-1999
     Introductory Microeconomics, Netanya Academic College Law School, Israel. Emphasis on Law and
         Economics, 1998-1999

Research Assistant, Bar-Ilan University, Israel
     Analyzed ethnic discrimination using SAS; performed cost-benefit analysis of education loans, 1998-1999

Consultant, Deloitte and Touche Management consulting, Tel-Aviv
     Served as management consultant for industrial firms. In particular, devised cost-saving,
     productivity-enhancing reorganization programs. Provided expert opinion in disputes over regulatory
     policies of the Israeli government, 1997-1998

Sergeant, Israel Defense Forces, 1992-1994
     Served as a radio frequency interference technician. Conducted experiments on communication devices
     for the Israeli Army, as part of a team of technicians and engineers

13th European Workshop on General Equilibrium Theory, Venice, June 2004
19th Annual Congress of the European Economic Association, Madrid, August 2004

The Russell Sage Summer Institute, Trento, July 2004

Information is at the core of economic analysis, as it determines the economic agent’s action and strategies. In many circumstances the relevant information is the individual’s type; e.g., her tastes, abilities and risk level. A standard assumption in the literature is that the individual perfectly knows her type. This assumption has an intuitive appeal. However, in reality it is frequently violated. People find it difficult to determine their value of a product or to calculate their risk of being injured in a car accident. Instead, agents use perceptions which they form in a dynamic progression. In forming such perceptions, psychologists have recorded systematic biases and the influence of affective considerations. Affective considerations stand for the role of emotional motivation in determining one’s belief. These observations are at the heart of my dissertation, as described below.

The first paper, Affective Decision Making in Insurance Markets, is a theoretical decision-making model incorporating affective considerations into the expected utility paradigm. This model is illustrated for the insurance context, although it can be extended to other markets as well. More specifically, the decision maker is modeled as an agent with two inner processes labeled the "rational account" and the "mental account". The rational account coincides with the process specified by the traditional economic model. Namely, it maximizes expected utility by choosing action (insurance) for a given risk perception. The mental account is a process where risk perception is formed. Risk perception is determined to maximize expected utility net of mental costs. In other words, the mental account chooses the risk perception which optimally justifies her insurance decision. Affective motivation – the desire of feeling good about her insurance decision – is captured by the expected utility term. This also captures the dependency of risk perception on outcome, as supported in numerous psychological studies. Moreover, the distinction between two thinking processes is supported by research in psychology and is known as the dual processes theory.  In order to make a decision, the two accounts interact; I model this interaction as a simultaneous-move intrapersonal game, where choice is a pure strategy Nash equilibrium of the game. Pure strategy Nash equilibrium reflects consistency of the two accounts once a choice is made. However, one can think of an adjustment process which converges to such a choice, i.e., a process of reaching an agreement. During this process the two accounts are constantly in disagreement, that is, cognitive dissonance. Note that in this model both risk perception and insurance level are determined by choice. This gives rise to, generally, a multiplicity of equilibria and shows that almost surely the agent’s decision departs from the traditional economic model. It also shows that one can not conclude the absolute risk aversion property of the utility function from the data and allows for a negative correlation between objective risk and insurance level. I am currently exploring the design of insurance contracts for affective agents. I conjecture that with affective agents, insurance companies will offer some packages not for the purpose of selling them, but for changing the agent’s risk perception and hence leading her to buy more insurance.

The multiplicity of equilibria in affective decision making (ADM), as described above, reflects several risk perceptions that the agent can hold. Since risk perception is not fixed in the ADM model, it is conceivable that this model can rationalize every finite data set. This conjecture is addressed and proven false in the second paper, Consistency and Refutability of Affective Choice. More specifically, this paper formulates the ADM model in the state preference framework. Every insurance level and insurance premium can be translated into consumption and price vectors. In the case of two states of the world, multiplicity of equilibria is reflected as points on the budget line. Each such point can be achieved by maximizing the agent’s utility function, subject to a budget line; the multiplicity of equilibria is due to a change of the utility function with risk perceptions. This formulation, using a revealed preference argument, allows one to prove that the ADM is testable in insurance markets and complete financial markets. Testability implies that there exists a finite data set which ADM can rationalize, as well as a data set which refutes it. This result extends to equilibrium models of mutual insurance and a class of Knightian uncertainty models. Moreover, testability of the ADM model leads to the conclusion that there exists an axiomatization of affective choice. To see this note that a finite family of Afriat’s polynomial inequalities characterizes the solution of the ADM model. Applying the Tarski-Seidenberg theorem to this finite family of polynomial inequalities, I show that there exists an equivalent set of polynomial inequalities on the observables that characterize choice. Characterizing affective choice by a set of polynomial inequalities on the observables constitutes an implicit axiomatization of affective choice. That is, we prove the existence of the axiomatization, but the existence proof is not constructive. In addition, we extend the notion of Pareto optimality to ADM models of mutual insurance and prove a version of Arrow’s theorems on optimal risk sharing.

The third paper, A Perceived Screening Model with Bayesian Agents in Monopolistic Markets (PSM), considers the interaction among players when agents are Bayesian and have imperfect information about their own type. To be more precise, each agent observes a noisy signal about her type and then computes her type assessment (posterior belief) using the conditional probability of the signal and the known distribution of types in the population. The monopolistic firm is assumed to control the conditional probability of the agent’s private signal. This can be viewed as advertising by the firm in an attempt to influence agents’ information. Therefore, the paper examines the incentives a monopolistic firm has, in a market with Bayesian agents, to distort (or not) the agents’ information about their type. Our results show that there are incentives for the firm to distort the agents’ information, even if agents use Bayesian updating. Two cases are considered: price discrimination and insurance markets. In the price discrimination model, the firm prefers perfectly informed or completely uninformed agents, but in the insurance model the company always prefers completely uninformed agents.