We propose to relate the social structure of towns and cities to the presence of corruption in their local governments using data from Hungary. We measure corruption risk using data on public contract awards and social structure using data from iWiW, a defunct online social network. We find that towns with a fragmented social structure have significantly more corruption risk in their contracts, while towns made up of people with more diverse networks have significantly less corruption risk. These network measures provide substantial additional explanatory power for corruption risk above a baseline model controlling for covariates including education, income, and employment.
We suggest that the online social network can be used to measure the social organization or ``social capital'' of a town. We find that excess fragmentation in the social network enables in-group favoritism and corruption via social homogeneity. We measure fragmentation by applying a scaled version of modularity . This scaling enables us to compare fragmentation of settlements of different sizes. On the other hand, we find that diversity in the social network moderates corruption by facilitating impartiality in public life. We measure diversity at the ego level by measuring the extent to which its alters form overlapping communities. We average this measure over the users of a town to measure its collective network diversity.
Figure 1: The fragmentation and diversity of online social networks of residents of Hungarian cities predict corruption in the city’s public contracts. Fragmented cities consisting of densely connected groups with few outside links are significantly more corrupt. Cities in which the typical resident has a more diverse social network are significantly less corrupt.