Does Urban Density Promote Social Interaction? Evidence from Instrumental Variable Estimation

To test the common assumption that increased local area population density implies high levels of social interaction, I use instrumental variable estimation to examine a causal link between social interaction and urban population density. I instrument for local population density using earthquake and landslide hazard rates in addition to the presence of hard rock beneath the soil. I find the social interaction type matters when determining the relationship with population density. Social interaction with friends is positively related with population density, whereas neighbor interaction is negatively related. Group involvement does not seem to be systematically related to population density.