Identification and estimation of latent variables and their effect on social and economic outcomes

Abstract

Recently, there has been strong interest among economists in the impact of social and cultural factors on economic outcomes. For instance, concepts like culture, social capital or social attitudes have been used to explain several individual and group outcomes such as labor supply, health, financial development or economic growth. In this spirit, in this thesis, I explore differences in individuals’ attitudes, their determinants, and their potential to explain individual behavior. The following are some of the findings. Personal and demographic characteristics, especially education, influence attitudes towards the peace process in the Palestinian-Israeli conflict (Chapter 1). Trust influences the type of child care that mothers use, and this has an effect on female labor supply. Since trust differs across European countries, it may explain differences in female labor supply (Chapter 2). Trust also influences individuals’ investment decisions; individuals with less trust tend to invest more in housing and less in financial assets (Chapter 3). Trust and attitudes towards reciprocity affect individuals’ civic engagement differently. People with more trust participate more through existing formal institutions. People with high levels of reciprocity also tend to participate more, but if their levels of trust are not so high, they may choose a more informal (less traditional) way of doing it (Chapter 4). Good measurement of the latent variables (like trust or attitudes towards reconciliation and concessions) is crucial for understanding the effects of individual unobservable traits such as attitudes on observable outcomes, or the effects of observable personal and demographic characteristics on the formation of those attitudes. It also helps overcome the critique sometimes directed at the applied behavioral economics literature that some researchers make claims that go beyond what the statistical results justify. Therefore, one of the goals of this thesis is to use a rigorous measure of these latent variables. To achieve this, I estimate attitudes and the effects of the individuals’ latent traits on specific outcomes using a hierarchical item response model.