Economic forecasts are a fundamental and critical building block of fiscal policymaking. Their accuracy shapes budgetary plans and influences outcomes. Since the future is uncertain, in any given year forecasts will almost always be off the mark by varying degrees. What matters over the medium and long term is whether they are biased. For instance, an optimistic bias on economic growth can lead to unsustainable expenditure plans, which in turn may weigh on the sustainability of public finances more generally. Recognising the importance of budgetary forecasts, the EFB Secretariat has launched a forecast tracker. Apart from offering all interested parties the possibility to check the forecast performance in a user-friendly manner, the tracker is also a one-stop-window for the academic community who can download the underlying data here for further research. The current dataset covers the period 2000 to 2025, on which basis one can calculate the forecasts errors from 2001 to 2024.Mean forecast errorsThe forecast error is defined as the difference between (i) the outturn of a given variable for year T as recorded in the spring of year T+1 and (ii) the forecast of the same variable for year T produced in year T-1. The analysis shows 3 different statistics to assess forecast accuracy:The mean error (ME) represents the average forecast error for each country over a specified period of time. A positive (negative) value means that on average outturns exceed (fall short of) forecasts. For instance, in the case of economic growth a positive value means that forecasts tend to be cautious (optimistic). Positive and negative forecast errors may offset each other over time even if very large. The mean absolute error (MAE) is the average of the absolute forecast errors for each country over a specified period time. In contrast to the ME, positive and negative forecast error do not offset each other. As a result, the MAE is a measure of cumulated forecast errors in a specified period of time.The root mean squared error (RMSE) is the square root of the average squared forecast errors for each country over a given period of time. The RMSE provides an estimate of the standard deviation of the forecast errors. By squaring the errors, larger errors are given more weight, making the RMSE especially useful when large deviations are particularly problematic.
