1. The European Commission and the National Competition Authorities commonly use sampling methods. For instance, the recent Commission pharmaceutical sector inquiry is based on the analysis of a sample of substances used in prescription medicines. Competition authorities also exploit samples to perform substantives analyses of proposed mergers and to assess the anticompetitive nature of a given conduct. There are good reasons for this frequent use of sampling techniques. First and foremost, gathering all the information needed to construct a comprehensive dataset in a given competition case is often costly and time-consuming, if not impossible. Sampling methods allow one to infer some knowledge about an entire “population” (e.g., a series of prices) from a selection of individual
|
LAW & ECONOMICS: SAMPLINGS - SAMPLES - QUANTITATIVE ANALYSES - DATASETS - INFORMATION - POPULATION - PRICES - STATISTICAL TESTS - EMPIRICAL RESULTS - EXTRAPOLATION - REPRESENTATIVE
The use of sampling by the EU Commission and the NCAs: “You don’t have to eat the whole ox to know that it is tough”
The European Commission and the National Competition Authorities commonly based their quantitative analyses on samples, rather than comprehensive datasets. To collect all the information regarding a given “population” (e.g., a series of prices) is costly and time consuming, if not impossible. Statistical tests allow one to extrapolate empirical results computed on the basis of the sample to the entire population under scrutiny. However, such an extrapolation comes at a cost: it is not perfectly accurate. It is lying within a confidence interval and it depends on a certain margin of error. Competition authorities therefore need to ensure that the quality of the sample analysed is good enough to infer correct general conclusions. A sample must be large enough to ensure that the estimate is sufficiently accurate. Moreover, a sample must be representative of the entire population it is based on. If a sample is biased, it is impossible to easily draw any conclusion from it: in such a case, sample-based results are erroneous.
Access to this article is restricted to subscribers
Already Subscribed? Sign-in
Access to this article is restricted to subscribers.
Read one article for free
Sign-up to read this article for free and discover our services.
