A0498
Title: Forecasting with heterogeneous spatiotemporal models
Authors: Cynthia Yang - Florida State University (United States) [presenting]
Abstract: The purpose is to investigate the forecasting performance of dynamic spatial panel data models that incorporate heterogeneous coefficients and latent common factors. A comprehensive comparison of the predictive accuracy of univariate versus multivariate models is conducted, with and without accounting for spatial interdependence, across various forecasting horizons. Monte Carlo simulations are employed to assess the finite sample properties of competing models and estimators. An empirical application focused on house price forecasting demonstrates the practical advantages of the spatiotemporal models.
