Dealing with Endogenous Shocks in Dynamic Friendship Network

Different types of shocks, or the treatment of one of the players in a specific network, may influence not only the future performance of themselves but also affect their network connections. It is crucial to explore the behaviour of the whole network in response to such an event. This paper focuses on the cases of endogenously formed shock. The logic used in the peer effect literature is adopted to develop the dynamic model and accounts for the endogeneity of the shock. The model allows us to predict the endogenous part of the shock and use the remaining unexpected component to estimate the effect of the shock on the changes in the performance of network connections. The identification conditions for effect are derived, and the consistent estimation procedure is proposed.