A Network Structure of ROSCAs (Rotating Savings and Credit Associations) : ERGMs (Exponential Random Graph Models) Applied to a Leaders' Network in Rural Uzbekistan

This paper empirically analyzes a network structure created by the ROSCAs (Rotating Savings and Credit Associations) related to a leaders' network in rural Uzbekistan. The estimation methodology is based on the recent development of ERGMs (Exponential Random Graph Models) whose approximate maximum likelihood estimators are produced by MCMC (Markov Chain Monte Carlo) algorithms. The paper reveals the tendencies of the transitive triad structure of the network that can facilitate the tracking of defecting members.