Welcome

Emil P. Iantchev

iantchev@syr.edu
110D Eggers Hall, Syracuse University
Syracuse, NY 13244
(315) 443-4079

Department Seminar

Department Site

Syracuse University

Research

Asset-Pricing Implications of Biologically Based Non-Expected Utility

Adaptive models in mathematical biology suggest that evolutionary successful species must be intrinsically more averse to aggregate than idiosyncratic shocks because the former can lead to extinction. To test the relevance of this idea to economics, I embed the non-expected utility specification implied by such models into an economy with partial risk sharing due to limited commitment. For relative risk aversion towards idiosyncratic bets around 5, the non-expected utility version is consistent with both the equity premium and the degree of consumption smoothing observed in the data, unlike the expected utility model which, given the degree of idiosyncratic risk sharing observed, cannot generate realistic equity premia for such low values of risk aversion.

Evolutionary Selection of Modular Decision Architectures (with A. Bisin)

We study the evolutionary properties of decision processes. In particular, we show that in the presence of harmful mutations, a population of decision makers who possess an architecture consisting of hierarchically organized decision modules will have a strictly higher asymptotic growth factor than a population of decision makers with a unitary, fully connected, decision architecture. In addition, we show that under imperfect precision of the architecture and small cyclical fluctuations in the environment, conflict among the reference policies of decision modules arises as an evolutionary equilibrium. Finally, we show that economic models of multiple decision processes can be represented as examples of the type of modular hierarchy we investigate.

Risk_or_Loss Aversion? Evidence from Personnel Records
I use data on individual pay and output from Safelite Corp. to estimate a principal-agent model with moral hazard where the agent has preferences approximating various specifications of reference dependence, diminishing sensitivity, and loss aversion, as well as the case of CARA utility. Estimating the preference parameters via GMM shows that the specification with loss aversion and diminishing sensitivity fits the data better that any of the other nested specifications. Furthermore, goodness-of-fit simulations suggest that the model with loss aversion and diminishing sensitivity has root mean squared error that is 18 percent smaller than the CARA model.

An Evolutionary Approach towards Time Preferences (with A. Robson and B. Szentes)
We consider how choice behavior would be genetically determined when there are intertemporal and intergenerational tradeoffs. A particular gene is the result of evolution if it is not possible for a rare mutant gene to grow at a faster rate. Our goal is to represent the choice behavior of such a resulting gene by a preference relation. We show that if choices affect the number of offspring but not the descendants’ reproductive ability, such a representation is given by discounted expected utility, where the discount factor is the population growth factor, and the utility function is the product of the fertility function at each age and the survival functions to that age.
We then consider the case where newborn offspring are heterogeneous due to receiving transfers from parents. Now the preference representation is more subtle. The discount factor is still the population growth factor, but the utility function is the sum of the expected discounted reproductive values of the individuals whom the parent’s choices directly affect.

Moral Hazard and Bertrand Equilibria under Reference Dependence and Loss Aversion

This paper studies frictionless markets with moral hazard where risk neutral principals compete a la Bertrand for the services of agents whose preferences exhibit reference dependence, loss aversion, and diminishing sensitivity. The equilibrium contract consists of three regions. If output falls below a critical value, the agent incurs the biggest possible penalty. When output is above this critical value but below an upper threshold, the agent receives a constant transfer equal to his reference (expected) payoff. And if output exceeds the upper threshold, there is a region with increasing additional rewards.