GKI Economic Research Co. compared its key annual macroeconomic forecasts published between 1995 and 2011 with the actual data. Differences between the predictions and the actual data are not only the results of forecast errors. The quality of the forecasts and their reliability may be affected by several other factors as well. They include, for example, external economic effects, government intervention in the economy and the continuous adjustment of statistical data.
GKI issues forecasts six times a year. Their accuracy is improving over time as more and more data are available. GKI provided the most accurate forecasts for the consumer price index. The mean absolute difference between GKI forecasts and actual inflation figures in the given period was 0.7 percentage points, which is extremely low. The average absolute difference between the forecasts and actual data of the volume index of GDP, which is the most common economic indicator, was 1 percentage point in the period between 1995 and 2011, which is also a favourable value. This is proved by the fact that the mean difference of preliminary statistical figures (usually published in March of the following year) and final official ones of GDP released by the Central Statistical Office amounts to 0.6 percentage points, which is only slightly better than the forecasts of GKI made many months or even a year before. The reliability of GDP and CPI forecasts issued by GKI in the period between 2006 and 2011 was comparable with the average reliability of expectations of 25-30 professional institutions published by Reuters.
The authors of forecasts always assume the possibility of erring. It is no coincidence that the introductions of economic analyses and forecasts often contain a disclaimer. Of course, GKI assumes no responsibility for business decisions based on its forecasts. However, GKI assumes responsibility for the forecasts themselves; this is why it regularly publishes such comparisons.