Kombination von FADN- und IFCN-Datensätzen in der Politikfolgenanalyse - untersucht am Beispiel der EU-Milchmarktpolitik

Abstract

An important task in scientific policy consulting is to estimate how agricultural policy reforms and changing framework conditions impact farms and to what extent the farms can adjust. Analyses at the single farm level carried out to date are largely based on bookkeeping results of the Information Network of Farm Bookkeeping (INLB, or rather FADN) and on data from typical model farms from the International Farm Comparison Network (IFCN). The FADN Data are well suited for short term estimates without significant changes in quantity. The FADN lack the depth required (above all the quantity framework and the process specific cost classification) for differentiated analyses with consideration of farm adjustment reactions. The strength of the typical IFCN farm data is the depth of documentation in the calculation of near-reality farm adjustment reactions. Due to the minimal number of typical farms, only limited conclusions can be derived on the regional or member country level. The goal of this project is thus to develop a methodical concept to link to the individual farm data centres FADN and IFCN for improved policy impact assessment. The comparative advantages of both data sources shall be used with linkages. In this project, the empirical analysis is done on the example of the EU Milk Market policy. The linkage approach classification consists of classifying typical IFCN farms selected by consultants in the total number of farms represented by FADN farms. In the first step, the IFCN and FADN data are compared on the basis of structural statistics and the FADN farms selected that are most similar to the IFCN farms (similarity index: Euclidean distance). The results show that if a similarity index is very narrow, only a relatively small portion of entire scope of farms and dairy production are represented. In the second step, it is tested how the structurally organized IFCN farms can be estimated with regard to their economic performance ability. Here it can be seen that only part of the typical farms can be classified as having an average performance level and thus the analysis results of these farms form conclusions for a relatively small farm group of the whole. Ultimately, the question arises of whether the FADN data is better for the selection of the typical farms, instead of later establishing how representative the chosen typical farms really are. That is why a supplementary process was developed with which the selection of typical IFCN farms can be conducted on the basis of statistical data (FADN data) in order to attain better representativity. German dairy farms from the FADN data were clustered on the basis of five classification characteristics according to the Ward Process. Overall, 36 clusters were developed out of which the structural givens for typical IFCN farms could be derived. It is recommended to, with the help of consultants, select farms that best characterize the region on the basis of the cluster with the highest production. In the linkage approach simulation, the objective is to make a larger number of FADN farms usable for differentiated simulation calculation in the TIPI CAL model. Since the FADN Data was primarily gathered for income analyses, those variables not found in FADN must be supplemented for calculations in the TIPI CAL model. Two policy scenarios and an adjustment strategy were calculated. In accordance with expectations, the FADN and IFCN farms had similar income effects and adjustment potentials. The transfer of developed linkages to a computer program is recommended so that a larger number of FADN farms can be analysed without a great deal of effort.