Události

Ambuj Dewan, Ph.D.

USC Dornsife Institute for New Economic Thinking,
University of Southern California, USA

________________________________________

Author: Ambuj Dewan

Abstract: Many of the objects about which people have to make economic choices have multiple attributes or dimensions (e.g. the specifications of a computer, the qualifications of a job candidate), and a decision-maker (DM) must learn about these dimensions before coming to a decision. Moreover, these dimensions may covary with each other, thereby complicating the information acquisition process. This paper presents and characterizes the solution to a multidimensional information acquisition problem with Gaussian prior beliefs. The solution's most important characteristic is that more attention is paid to the “orthogonalized” dimensions that the DM believes vary more ex ante. The model is extended to non-Gaussian priors and incorrect prior beliefs. Applications to consumer choice and discrimination in hiring are also presented, providing information-theoretic explanations for choice overload, focusing illusions in the valuation of multiattribute goods, and the effects of anti-discrimination initiatives.
________________________________________
Full Text: “Costly Multidimensional Information”