Evaluating Forecasts from State-Dependent Autoregressive Models for US GDP Growth Rate. Comparison with Alternative Approaches

Abstract The aim of the paper is to compare forecasting performance a class statedependent autoregressive (SDAR) models for univariate time series with two alternative families nonlinear models, such as SETAR and GARCH models. study conducted on US GDP growth rate using quarterly data. Two methods forecast comparison are employed. first method consists in evaluation average by measures root mean square error (RMSE) absolute (MAE) over different horizons, while second make use one most used statistical test accuracy Diebold-Mariano test. JEL classification numbers: C22, E37, F47. Keywords: Nonlinear series, rate, Forecasting accuracy.