Title: Threshold regression with social interactions
Authors: Antri Konstantinidi - University of Cyprus (Cyprus)
Andros Kourtellos - University of Cyprus (Cyprus) [presenting]
Yiguo Sun - University of Guelph (Canada)
Abstract: A new class of social interaction models that generalize the spatial autoregressive model is developed in order to allow for threshold effects. Threshold-type nonlinearity is one of the most interesting forms of nonlinearity as it is often implied by economic theory. In particular, we study identification and develop a statistical theory for the threshold parameter as well as the regression coefficients. We exploit the non-linear structure of the threshold regression as well as the network structure to overcome identification issues that arise in the linear-in-means model and propose a GMM approach for the estimation. We assess the performance of our methods using a Monte Carlo simulation and provide an empirical illustration.