Causal inference on Nation-Scale graphs

Specialeforsvar: Frederik Kølby Christensen

Titel: Causal inference on Nation-Scale graphs

Abstract: This thesis approaches the problem of crime spreading from one person to another. By focusing on a particularly disadvantaged subgroup, young fathers, two methods of extracting causal inference, Instrumental Variables and G-computation, are compared and reviewed. G-computation is preferred. Using this, we find a possibility of an economically significant increase in the probability of crime on a male peer when his schoolmate commits crime. We do not find evidence of spreading when considering peers who are only related through living close to each other. This implies that crime may spread first and foremost
through realized social interactions, rather than through acquaintances. 

Vejledere: Niels Richard Hansen
                  Sune Lehmann, DTU
Censor:     Lasse Petersen, Mærsk Data