Measuring Comovement by a Smooth Transition Simultaneous Equation Model

Mardi | 2018-09-26
Salle des thèses 16h – 17h20


This article develops a flexible econometric framework to investigate the comovement between two endogenous variables within a nonlinear simultaneous equation model. The model controls also for indirect dependence which intervenes through common observed and unobserved factors. The comovement is modeled as a smooth and potentially asymmetric function of the magnitude of the endogenous variable. The threshold at which a shock is transmitted is estimated with the other parameters of the model. We investigate the properties of an accurate estimation method which takes into account endogeneity, and a testing procedure for simultaneity in the presence of nuisance parameters under the null hypothesis. In a two-equation setup, we study the conditions on the parameters which ensure the uniqueness of the implicit reduced form of the model. Multiple equilibria have important drawbacks such as the unsuitability of the model for forecasting. We illustrate the methodology by studying the impact of the recent crises on the comovement between the sovereign and banking sectors for eight developed countries.