Dissertations

Thursday, 13 April, 2023

Pavel Ilinov: Essays on Endogenous Information Acquisition in Economics

Dissertation Committee:
Jakub Steiner (CERGE-EI and the University of Zurich , Chair)
Jan Zápal (CERGE-EI)
Ole Jann (CERGE-EI)

Defense Committee:
Krešimir Žigić (CERGE-EI, Chair)
Yiman Sun (CERGE-EI)
Marek Hudík (VŠE)

Referees:
Eugen Kováč (University of Duisburg-Essen)
Fabio Michelucci (Ca' Foscari University of Venice)

Abstract:

The first chapter studies the properties of the stochastic choice data in the problems with endogenous information acquisition. We consider two types of models: (i) a rational inattention problem and (ii) a conformity game, in which fully informed players find it costly to deviate from average behavior. We show that these problems are equivalent to each other, both from the perspective of the participant and the outside observer: each individual faces identical trade-offs in both situations, and an observer would not be able to distinguish the two models from the choice data they generate. We also establish when individual behavior in the conformity game maximizes welfare.

The second chapter focuses on the flexibility of the endogenous information that an agent can acquire. We show that the principal can strictly benefit from delegating a decision to an agent whose opinion differs from that of the principal. We consider a ``delegated expertise'' problem in which the agent has an advantage in information acquisition relative to the principal, rather than having preexisting private information. When the principal is ex ante predisposed towards some action, it is optimal for her to hire an agent who is predisposed towards the same action, but to a lesser extent, since such an agent would acquire more information, which outweighs the bias stemming from misalignment. We show that belief misalignment between an agent and a principal is a viable instrument in delegation, performing on a par with contracting and communication in a class of problems.

The third chapter investigates the intertemporal trade-off of endogenous information in dynamic decision-making problems. We extend the classical search framework by allowing the decision-maker to choose information endogenously and flexibly. We consider a stylized problem of an interview design in which the manager chooses the best among ex ante identical candidates. The manager learns the qualities of candidates sequentially and can choose the information structure by herself at a cost. Our findings vary depending on the number of available candidates. We show that with two candidates, the optimal learning strategy is not unique: the manager may learn different information about the candidates, and there is a continuum of optimal learning strategies. However, the manager, on average, chooses candidates uniformly across all optimal strategies. On the contrary, when three candidates are available, the optimal strategy is unique. The manager engages in ``cherry-picking'' behavior: the difficulty of an interview decreases according to the order of candidates. Nevertheless, the manager optimally balances between the difficulty and the informativeness of an interview. In the optimum the manager discriminates against the last candidate and chooses him least often. Additionally, we connect our findings with the serial-position effect from the psychological literature. With three alternatives the primacy effect is present: the average amount of information obtained decreases in the order of candidates. We also show that our results are robust to several extensions.

Full Text: "Essays on Endogenous Information Acquisition in Economics"