Classifying Changes in the Consolidation Perimeter in the German Group Accounts Statistics Using Statistical Learning

Abstract

The Deutsche Bundesbank publishes aggregated rates of change for the revenue, the earnings before interest, taxes, depreciation and amortisation, and the earnings before interest and taxes of the German non-financial groups listed in the Prime Standard of the Frankfurt Stock Exchange. Where there are changes in the consolidation perimeter, it is crucial to ad-just these growth rates to obtain an unbiased impression of economic developments. However, an adjustment is only possible if it is known which groups have undergone a significant change in group structure. To find evidence of changes in perimeter, different statistical learning approaches are tested and combined. This is the first paper to classify changes in the consolidation perimeter for data quality maintenance. A combination of a random forest and a weighted nearest neighbours approach, which are both oversampled, provides the best results. The approach avoids overlooking changes in the consolidation perimeter in an efficient and affordable way.