Essays in Microeconomics

This dissertation analyzes problems related to the the economics of incomplete information and to the theory of matching markets. Chapter 1 defines a family of functions that measure the distance between opinions; Chapter 2 investigates how to measure the cost of an experiment; and Chapter 3 studies a model of two-sided matching with countably many agents. Chapter 1 introduces six axioms that a measure of disagreement should satisfy, and characterizes all the functions that satisfy them. The disagreement measures characterized generalize the Renyi divergences, and include the Kullback-Leibler divergence and the Bhattacharyya distance. Two applications are then studied. The first application provides a necessary and sufficient condition under which public information reduces expected disagreement between Bayesian agents. The second application shows that the measures of disagreement here defined are useful to understand trading under heterogeneous beliefs. Trade volume and gains from trade are increasing in some of the measures of disagreement. Chapter 2 introduces seven postulates for a cost of information function. The main result of this chapter is the proof that there exists a unique function that satisfies these postulates. Differently from the cost functions commonly used, the function found in Chapter 2 is independent of the experimenter’s beliefs, and it is additive in independent experiments. Similarly to other cost functions, it is increasing in the informativeness of the experiment, and it is separable in the signal realizations. Chapter 3 analyzes two-sided one-to-one matching with countably infinite agents. It shows that the set of stable matching is non-empty if and only if agents’ preferences admit a maximum on all subsets. This requires generalizing the Deferred Acceptance algorithm, which also allows to find the man-optimal and woman-optimal stable matchings. It is then shown that, like in the finite model, the set of stable matchings is a complete lattice under the preferences induced by men (or women). Unlike in finite models, the set of matched agents may vary across stable matchings and some implications for dynamic matching markets are discussed