Evaluation of Norges Bank's projections for 1999.(central bank of Norway)(Statistical Data Included)

In order to provide the central bank with an optimal basis for the conduct of monetary policy, the central bank must evaluate its projections for economic developments. Norges Bank has previously published analyses of its projections for 1996, 1997 and 1998. This article evaluates the Bank's projections for 1999.

The article starts with an analysis of the projections in the December 1997 Inflation Report. The baseline scenario in the report implied a cyclical turnaround in 2000, but the turnaround occurred as early as mid-1998 and was followed by slow growth in 1999. The analysis of main developments was broadly in line with actual developments, but events in 1998 led to an earlier-than-projected turnaround. The projections for 1999 were therefore not particularly accurate. However, the report also included alternative projections where important assumptions were changed, for example a substantial depreciation of the krone and a rise in interest rates. Developments in the latter half of 1998 and into 1999 were closely in line with this alternative. The projections for 1999 published in the December 1998 Inflation Report were closer to the mark.

Norges Bank's projections are compared with those of Statistics Norway and the Ministry of Finance. Norges Bank's forecast for wage growth can be said to be at least as good as that of the two other institutions. There are no systematic differences with regard to the accuracy of price inflation projections.

Norges Bank's analyses of developments in the Norwegian economy are published in our quarterly Inflation Report. The projections for developments in the Norwegian and world economy provide an important background for the formulation of monetary policy. In addition, the analyses serve as a basis for advice on the orientation of economic policy in general. Since 1994 Norges Bank's macroeconomic model RIMINI, developed in the Research Department of the Bank, has been the main tool for our analyses. The RIMINI model is an econometric model with almost 370 equations. Around 70 of them are estimated on the basis of historical data, while the remainder are definitional relationships. A revised RIMINI model, based entirely on the Revised National Accounts, is now being tested. The new model will be operational in the course of the winter. With the introduction of the new model, it will also be possible to analyse the model's and the model users' contribution to forecast errors. Alternative modelling tools are also being developed to supplement the RIMINI model on some points.

Norges Bank aims to produce the best possible projections for the Norwegian economy. It is important to reveal errors in order to improve the analyses. Evaluations of Norges Bank's projections in relation to those of other institution also contribute to this.

Norges Bank places emphasis on transparency and the availability of its forecast work. This work also includes analyses of earlier projections. The projections are based on a model that is publicly known, and the Bank's use of the model is published. The purpose is to provide others with the basis for evaluating how we arrived at our projections and their accuracy. Systematic evaluation also places greater demands on consistency and documentation of the projections in the Inflation Report, which in itself will improve the quality of the analyses.

Analyses of forecast errors in Norges Bank's model-based projections are presented here for the fourth time. Previously, we have evaluated and attempted to find the sources of the forecast errors in the projections for the years 1996 to 1998. As in the article in Economic Bulletin 1999/2 (Jore 1999), this article also looks at projections two years ahead.

For further discussion of forecast errors in general, see previously published articles (Madsen 1996, Jore 1998, Jore 1999).

Forecasts for 1999 published in December 1997

The projections in the December 1997 Inflation Report pointed to higher growth rates over the next two years. A cyclical turnaround with stagnating employment and rising unemployment was expected from 2000. The cyclical turnaround in the baseline scenario was predicted in the light of the deterioration in competitiveness as a result of high wage growth relative to our trading partners, in addition to an expected decline in petroleum investment.

Developments in key parameters through 1998 and 1999 turned out to be very different from that assumed towards the end of 1997. First, there were the contagion effects of the Asian crisis. Financial market turbulence spread to our part of the world and contributed, in conjunction with the weakening of competitiveness as a result of high wage growth, to a depreciation of the krone and higher money market rates. Growth in the world economy slowed, which led to a decline in international producer prices. The fall in oil prices over 1998 contributed to a general fall in prices. On the other hand, petroleum investment was substantially higher than we had assumed both in 1998 and 1999.

In several areas, developments were different from our assumptions. A number of factors indicated that price inflation would be higher than expected:

- As a technical assumption, the krone was assumed to remain unchanged at the previous month's level to the end of the projection period. This implied a virtually unchanged exchange rate between 1997 and 1999. However, the import-weighted exchange rate index (I-44) depreciated by 2.2 per cent between 1997 and 1998, followed by an appreciation of 1.2 per cent between 1998 and 1999.

- Public consumption was estimated to increase by 2 per cent in 1998 and 2 1/2 per cent in 1999. Growth was almost twice that in 1998 at 3.8 per cent, while the estimate for 1999 proved to be accurate.

- We assumed a 17 per cent fall in petroleum investment between 1997 and 1999. However, petroleum investment expanded by a good 5 per cent during the period. …

To read the full text of this article and others like it, subscribe today!