A generalized single equation error correction model and its application to quarterly data

Abstract

In this paper, we specify a class of single equation 'error correction' models on the basis of a general autoregressive-distributed lag regression equation with one regressor and a white noise disturbance. This relationship is interpreted in terms of long run trends in the regressor and regressand and short run deviations from these trends. A parametrization which is useful for quarterly seasonally unadjusted data is proposed. The model is estimated by means of a non-linear least squares algorithm. Empirical results based on Norwegian quarterly national accounts data , illustrating the relationship between (i) household consumption and income, (ii) production and demand in manufacturing, and (iii) capital accumulation and production in manufacturing - are presented. Some experiences from forecasting exercises are also reported.