Network Games With Triadic Interactions

ABSTRACT We propose a model for network‐mediated interactions incorporating higher‐order relationships such as triadic interactions. The marginal utility of an individual's action depends explicitly on endogenous actions of multiple units. Conditions ensuring a unique Nash equilibrium are derived using a variational inequality approach. We estimate the model parameters via maximum likelihood and derive their large sample properties. Applying our model to data from a primary school teamwork experiment in China, we find that the peer effects associated with externally assigned teammates vary depending on preexisting friendship networks. The teamwork intervention enhances peer influence on attitudes while diminishing academic competition among students.