Essays on Microeconomic Theory

Abstract

This collection of essays investigates issues related to information acquisition in the presence of permanent ambiguity, perception errors or computational constraints. After a brief introduction, Chapter 1 introduces a decision maker who faces a sequence of non-identical experiments with a finite numbers of possible outcomes. The law generating the observations changes from period to period and the decision maker does not know anything about how the laws evolve. She has probabilistic beliefs over sets of possible laws and makes inferences on the true set of laws. I consider updating mechanisms characterized by consequentialism and such that updating can only depend on current beliefs and the current observation. To stay as close as possible to the standard Bayesian framework, I assume that the relative frequencies of observations converges to some limit frequency and that this limit frequency is compatible with only one of the possible sets of laws. I then show that, when the number of possible outcomes is larger than three, the individual may never learn what the true set of laws is.

In Chapters 2 and 3 I consider a decision maker receiving signals. The signals specify a subset of the set <italic>S</italic> of states of the world in which the true state in included. However, the decision maker does not observe the signals correctly: each signal is perceived as a possibly different subset of <italic>S</italic> according to a function <italic>v</italic> mapping true signals into perceived ones. Chapter 2 introduces the concepts of underconfidence and overconfidence for this setting and analyzes the consequences of different behavioral sources of misperception on the set of fixed points of the mapping function <italic>v</italic> and on the relation between each true signal <italic>A</italic>. and the corresponding perceived signal <italic>v(A)</italic>. In Chapter 3 I add an ex-ante stage to the model. The decision maker is aware of her ability to manage a limited number of different signals, smaller than the total number of signals that she may receive. She therefore chooses an optimal function <italic>v</italic> mapping received signals into perceived ones, subject to the constraint on the number of the latter. The chapter studies the properties of this optimal mapping.