An Evaluation of the Methods to Forecast Tax Revenues
In June of each year, the Federal Public Service Finance (FPS Finance) draws up estimates of the expected tax revenues for the following year. Based on these forecasts, the initial budget is drawn up in September and voted on by Parliament before the end of December. In February, FPS Finance also estimates the tax revenues for the current year. These estimates form the basis for any adjustments to the budget, the adjusted budget. This study addresses both forecasts of tax revenues by the FPS Finance. These include not only the revenues destined for the federal level, but also the revenues that the federal government, after collection, transfers to the social security system, the regions, and the communities.
In 2015, doubts arose as to the quality of these forecasts. They were often thoroughly revised during the adjustment of the budget. Therefore, on 22 October 2016, a task force was set up to analyse the forecasts and the forecasting methods. As part of its task within that task force, FPS Finance decided to launch an evaluation on the forecasting of tax revenues. This text summarises the main findings of this evaluation.
The evaluation can be divided into five parts. The first part was an international comparison of the forecasting methods used. The disaggregated method was then described and discussed in detail. This method is used to forecast approximately 85% of the tax revenues, in particular for personal income tax, corporation tax, value-added tax (VAT), and excise duties. The third part was the quantitative evaluation of the forecasts that are made on the basis of this disaggregated method. The forecasts outside the disaggregated method were then evaluated analogously, both in terms of method and quantitative performance. The fifth and final part of the report summarised the findings and identified several possible areas for improvement.