Nonlinear factor models for network and panel data

Factor structures or interactive effects are convenient devices to incorporate latent variables in panel data models. We consider fixed effect estimation of nonlinear single-index models with factor the unobservables, which include logit, probit, ordered probit and Poisson specifications. establish that estimators model parameters average partial have normal distributions when two dimensions grow large, but might suffer from incidental parameter bias. also show how can be applied capture important features network such as reciprocity, degree heterogeneity, homophily variables, clustering. illustrate this applicability an empirical example a gravity equation international trade between countries using multiple factors.