Network Games with Local Correlation and Clustering

In many contexts an individual's beliefs and behavior are affected by the choices of his or her social or geographic neighbors. This influence results in local correlation in people's actions, which in turn affects how information and behaviors spread. Previously developed frameworks capture local social influence using network games, but discard local correlation in players' strategies. This paper develops a network games framework that allows for local correlation in players' strategies by incorporating a richer partial information structure than previous models. Using this framework we also examine the dependence of equilibrium outcomes on network clustering — the probability that two individuals with a mutual neighbor are connected to each other. We find that clustering increases the provision of public goods and makes it easier for a beneficial new technology to break into a market dominated by a less preferred standard