In time series models, the number of parameters increases quickly with the number of variables, so that usually only small-scale multivariate models are considered. Factor models can cope with many variables without running into scarce degrees of freedom problems. Hence, in this paper we construct a large macroeconomic data-set for China, with about 41 variables, model it using a dynamic factor model, and compare the resulting forecasts with ARMA models. Finally, the factor-based forecasts are shown to improve upon standard benchmarks for GDP at virtually no additional modelling or computational costs.
展开▼
