A0276
Title: A score-driven filter for time-varying regression models with endogenous regressors
Authors: Noah Stegehuis - Vrije Universiteit Amsterdam (Netherlands) [presenting]
Francisco Blasques - VU University Amsterdam (Netherlands)
Abstract: A score-driven model is proposed for filtering time-varying causal parameters using instrumental variables. The causal parameter is updated at each time point by the score of the likelihood function. In the presence of suitable instruments, it is shown that it can uncover dynamic causal relations between variables, even in the presence of regressor endogeneity which may arise due to simultaneity, omitted variables, or measurement errors. Due to the observation-driven nature of score models, the method is simple and practical to implement. The asymptotic properties of the maximum likelihood estimator are established and show that the instrumental variable score-driven filter converges to the unique unknown causal path of the true parameter, whereas the existing score-driven procedure does not. Further, the finite sample properties of the filtered causal parameter in a comprehensive Monte Carlo exercise are analysed. Finally, the empirical relevance of this method in an application is revealed to aggregate consumption in macroeconomic data.