Backcasting can be described as forecasting or extrapolating backwards in time. This is typically done by applying forecasting methods to an existing time series in reverse chronological order, starting from the end of the time period of interest and going back to its beginning.
In the context of statistical classifications, such as the NACE, an important criterion when selecting a backcasting method is to ensure consistency between:
- the time series that is already published, which uses an old version of the classification, and
- the backcasted time series, which uses the new version of the classification.
Backcasting for NACE
For many European statistics, there is a need for providing historical time series using the same version of the NACE. Therefore, backcasted statistics for past time periods broken down by NACE Rev. 2.1, instead of NACE Rev. 2 will be published for many statistical domains.
For certain European business statistics, backcasting and double reporting requirements are set out in Commission Implementing Regulation 2024/1840 (amending 3 implementing regulations as regards references to the statistical classification of economic activities NACE Rev. 2).
Guidance for statistics producers
Eurostat has developed a backcasting manual which provides an overview of the methods applied to the backcasting of time series for data using categories of economic activity.
In addition, the manual describes several methods for reclassification of statistical units in a statistical business register. This is needed because the applied backcasting methods require that statistical business registers are double coded according to the old and the new classification for at least one reference year.
The manual also discusses how to mitigate the statistical disclosure risks that arise due to the parallel dissemination of the same time series according to 2 versions of the same classification.
