Three Essays in Microeconometrics

This dissertation examines three distinct issues using microeconometric techniques. The first two chapters fall in the realm of discrete choice models and try to make allowance for limited attention. The third chapter focuses on firm behavior and investigates the impact of ownership concentration on productivity. Chapter 1 predominantly builds on the consideration capacity model in Dardanoni, Manzini, Mariotti and Tyson (2019). In the attempt to behavioralize rational choice theory, their model identifies the distribution of cognitive characteristics in a population of agents who are observed choosing repeatedly from a single menu. By exploiting algebraic arguments, we first generalize the identification result. Then, we propose an Expectation-Maximization algorithm which is able to recover the distribution of individuals' cognitive characteristics in a non-parametric framework. The algorithm is applied to both simulated and real market data. The first application is meant to show that model primitives can be estimated with a high degree of accuracy. In the second one, instead, it is shown that a substantial fraction of individuals is either low or moderate attentive thereby contradicting full rationality which would require subjects to pay attention to all the alternatives of a given menu. Chapter 2 resorts to a parametric setup and investigates asset allocation choices in defined contribution plans of a sample of U.S. workers. We propose a multinomial logit model in which the choice of a given financial instrument is preceded by a probabilistic consideration set formation process. Our results show that failing to recognize the relevance of limited attention can induce misleading evaluation of the effects of demographic and job-specific characteristics on the process through which workers decide how to allocate their contributions. Chapter 3 analyzes the relationship between ownership structure and firm performance using total factor productivity (TFP) to measure firm value. Adopting a structural approach à la Olley and Pakes (1996), we proposes a semi-parametric model which controls for firms' unobserved heterogeneity and for the endogeneity of not only input factors but also of a relevant corporate governance variable, namely ownership concentration. The method is applied to Italian manufacturing data coming from the ORBIS dataset which are enriched with information on ownership provided by the Italian Securities Commission (CONSOB). Results highlight the presence of a non-monotonic relation between productivity and the degree of ownership concentration (an inverted U-shaped relationship). We argue that our findings depend on the interplay between the monitoring dimension and shareholder conflict dimension associated with ownership concentration.