| This dissertation consists of two essays. The connecting theme for this body of work is social interaction. I define social interaction, for this dissertation, as non-market based face-to-face interaction between individuals where each is interested in a response. Non-market based means, for example, the individual is not trying to negotiate a price of a good or a labor contract, but is merely interested in the other's input, company, or friendship. Within the dissertation, I examine the relationship between social interaction and local area population density and the effect of social interaction on the urban consumer's location and housing consumption decision.;The first contribution investigates the relationship between social interaction and neighborhood population density. Understanding the causal impact on social interactions from a change in neighborhood density is a challenging empirical question; the choice of location for the individual may include a desire for population density or social interaction, or, most likely, both. I address endogeneity or simultaneity using an instrumental variables technique. I use the exogenous variation in a set of geological instruments to predict the neighborhood population density. The instruments include seismic hazard, landslide hazard, and the presence of sedimentary rock beneath the soil. These instruments effect population density through the types of structures that are placed on top of the land. The results suggest that an increase in neighborhood population density increases some types of social interaction. Friends interaction increases as population density rises, but neighbor interaction falls with increases in population density. Group involvement appears to have no consistent significant effect from a change in population density. This implies a need for theoretical models involving social interaction to be flexible to differing types of social interaction.;The second contribution brings social interaction into a standard urban consumer theory model, the Alonso-Muth model. I start by outlining the canonical model and providing intuition for the inclusion of social interaction. Social interaction is allowed to take multiple forms in the model; interaction can be exogenously or endogenously determined across an area, increase or decrease with population density, or be a substitute or complement to housing consumption. This flexibility provides multiple cases and sub-cases of how social interaction could appear in reality. I capture the effects of changes in social interaction on housing demand and location choice with a number of comparative statics. Lastly, I examine the predictions of the endogenous social interaction model with empirical data. The empirical results buttress the theoretical model's predictions. |
