Equilibrium-correction vs. differencing in macroeconometric forecasting |
| |
Authors: |
yvind Eitrheim Tore Anders Huseb Ragnar Nymoen |
| |
Institution: | a Research Department, C-51, Norges Bank, Central Bank of Norway, Box 1179 Sentrum, N-0107 Oslo, Norway;b University of Oslo, Department of Economics, P.O. Box 1095, Blindern, N-0317 Oslo, Norway |
| |
Abstract: | Recent work by Clements and Hendry elucidate why forecasting systems that are in terms of differences, dVARs, can be more accurate than econometric models that include levels variables, EqCMs. For example, dVAR forecasts are in some cases insulated from parameter non-constancies in the long run mean of the cointegration relationships. In this paper, the practical relevance of these issues are investigated for RIMINI, the quarterly macroeconometric model used in Norges Bank (Central Bank of Norway), an example of an EqCM forecasting model. We develop two dVAR versions of the full RIMINI model and compare EqCM and dVAR forecasts for the period 1992.1–1994.4. We also include forecasts from univariate dVAR type models. The results seem to confirm the relevance of the theoretical results. First, dVAR forecasts appear to provide some immunity against parameter non-constancies that could seriously bias the EqCM forecasts. Second, the misspecification resulting from omitting levels information generates substantial biases in the dVAR forecasts 8 and 12 quarters ahead. |
| |
Keywords: | Forecasting Macroeconomics Equilibrium-correction Differenced VAR models |
本文献已被 ScienceDirect 等数据库收录! |
|