Social interactions and sickness absence: Family, colleague and neighborhood effects

Abstract

Social interactions play an important role in economic decision making. In this thesis I examine the effects of social influence on sickness absence. While I study social interaction effects among i) family members, ii) colleagues, and iii) neighbors, the main focus are on neighbors. I use Norwegian data provided by Statistics Norway to analyse if the percentage of one’s family members, colleagues, and neighbors on sick leave affects an individual’s likelihood of being on sick leave. The results indicate presence of social interaction effects in all three groups. The estimated family effect suggests an increased sick leave probability of 0.07 percent if the percentage of one’s family members on sick leave increases with one percent. Individual’s faced with a one percent increase in the percentage of his or her colleagues on sick leave, show a 0.28 percent increased probability of receiving sick pay. Finally, the estimated neighborhood effect show that a one percent increases in the percentage of neighbors on sick leave show an increased sick leave probability of 0.39 percent. While I do find support for social interaction effects across all models, I acknowledge the methodological challenges related to estimating group effects using the above approach. To address one of these challenges, I estimate a neighborhood model that accounts for the so-called reflection problem. The result from this model is consistent with the simpler model, providing further evidence of social influence on individual sickness absence by neighbors. Lastly, I estimate a spatial model to assess the impact by nearby geographical neighborhoods on sickness absence in a focal neighborhood. The result indicates that the percentage of individuals on sick leave in a focal neighborhood is influenced by these neighboring neighborhoods.