Switching Costs, Investment, and Collusion: Essays in Empirical Industrial Organization

Abstract

This dissertation consists of three essays in Empirical Industrial Organization. In each essay I develop and estimate an econometric model to quantify unobserved structural parameters of the underlying theoretical model. The first study develops a dynamic model of consumer behavior in the market for paid television services. A representative consumer supply side model is then used to recover the cost structure of cable providers. Simulation of counterfactual experiments allows me to measure the effect of satellite entry on optimal cable prices. Using data on cable and satellite systems across local U.S. television markets over the period 1997-2006, I estimate parameters of the structural models. Estimation results suggest presence of significant consumer switching costs, amounting to approximately $109 and $186 (in 1997 dollars) for cable and satellite systems respectively. In the absence of satellite competitor cable prices would be on average 19 percent higher than observed ones. The second essay evaluates the degree of collusion among sugar refiners in Ukraine. I show that government regulation may facilitate collusion between the members of the association of producers. I extend the existing empirical literature on collusion by developing an empirical model of collusive behavior when the objective function of a cartel is not known. Under the assumption that a cartel implements a proportional reduction technique of collusion parameter that measures the degree of monopolization has structural interpretation. Using data from the Ukrainian manufacturing enterprise register between 1993 and 2000, I estimate the degree of monopolization in the Ukrainian sugar refining industry. Estimation results suggest moderate degree of collusion ranging between 89 and 95 percent of the Cournot equilibrium quantity. The third essay proposes an empirical model of firm level fixed capital investment when financial markets are imperfect. In this case, firms operating in a "constrained regime" cannot afford the optimal level of investment predicted by their intertemporal maximization problem. I derive investment functions for constrained and unconstrained firms from a structural dynamic model of investment. I then derive a specification for a regime switching regression with unknown sample separation and without an explicit selection equation. To estimate the structural parameters, I use Ukrainian firm-level data between 1993 and 1998. Estimation results indicate that the probability of facing financial constraints in the Ukrainian economy was very high and ranged from 0.39 in the ferrous metallurgy to 0.79 in the power industry.