Forecasting economic time series using flexible versus fixed specification and linear versus nonlinear econometric models英文电子书.pdfVIP

ELSEVIER International Journal of Forecasting 13 (1997) 439-461 Forecasting economic time series using flexible versus fixed specification and linear versus nonlinear econometric models a.:~ N o r m a n R. S w a n s o n , H a l b e r t W h i t e b ~The Pennsylvania State University. 521 Kern Graduate Bldg.. Department of Economics. University Park. PA 16802. USA ~Research GroupfiJr Econometric Analysis. Department of Economics. University of California. Sun Diego. La Jolla. CA 92093. USA A b s t r a c t Nine macroeconomic variables are forecast in a real-time scenario using a variety of flexible specification, fixed specilication, linear, and nonlinear econometric models. All me,dels are allowed to evolve through time. and our analysis focuses on nltKlel selection and performance. In the context of real-time forecasts, flexible specilication models (including linear autoregressive models with exogenous variables and nonlinear artificial neural networks) appear to offer a useful and viable ahernative to less flexible fixed specification linear tm~dels for a subset of the economic variables which we examine, particularly at forecast h~rizons greater than I-step ahead. We speculate that one reason for this result is that the economy is evolving (rather slowly) over time. This feature cannot easily be captured by lixed specilication linear models, however, attd manifests itself in the form of ewdving