Accuracy in forecasting macroeconomic variables in Iceland

Abstract

This paper discusses accuracy in forecasting of macroeconomic time series in Iceland. Until recently only the National Economic Institute (NEI) did macroeconomic forecasting in Iceland. Extensive analysis of forecasting can therefore only be done for the forecasts made by this institution during 1974-2002.

The paper analysis macroeconomic forecasts published by the Central Bank of Iceland (CBI). It also analysis the accuracy of the first realeases of data from Statistics Iceland as “forecasts” of final (or the most recent) data during recent years. Forecasts made by international institutions like OECD and IMF are not included.

The paper finds that errors in forecasting of GDP and private consumption have declined and that the performance of the forecasting for these variables has improved on some measures. But the volatility in the series has also decreased so when the forecast errors are compared to measures of the shocks that hit the economy the forecasting of changes in GDP do not seem to have improved. For some of the main components of GDP like export, imports and investments, the forecast errors have not decreased.