In October 2014, the Commission published the result of its long-awaited food retail study. The study provides an extensive empirical analysis of the impact of bargaining power and private label on choice and innovation. However, the study suffers from a crucial shortcoming: none of the countries with high levels of retail concentration have been included in the empirical analysis. As such, the study does not really deliver on its promise to investigate the impact of retail concentration and is unlikely to quell the debate on this issue.
Part of a longer term review of retail market dynamics, the study was launched following complaints raised by various stakeholders that retailer buyer power and the increasing importance of private label may negatively impact on choice and innovation. Similar concerns have been expressed in a number of contributions in the economic literature (example here). The key aim of the study was to examine these concerns empirically.
As part of the study, an extensive dataset was compiled in order to be able to undertake an econometric study of the driving factors behind the development of choice and innovation over time. Of particular interest was how each of choice and innovation are influenced by (i) retailer concentration; (ii) supplier concentration; (iii) the relative bargaining power of retailers and suppliers; and (iv) the share of private label. An econometric analysis is capable of “isolating” the impact of each of these factors, controlling for the impact of other factors that may also have an impact (for example, GDP or unemployment).
The study was undertaken at local level and included a total of 300 shops situated in seven countries. (Some qualitative case studies were also undertaken for a number of categories, half of which relating to fresh products, covering a small number of additional countries.)
One of the key findings of the study is that in moderately concentrated retail markets, an increase in retailer bargaining power does not seem to result in less choice and innovation in food products. Up to a point, the same conclusion was reached for the share of private label. Once the share of private label increases beyond a certain point, private label may become detrimental for choice and innovation.
A crucial limitation of the study is however that the empirical analysis is based on data from just a handful of countries. Moreover, as it happens, none of these countries include countries with highly concentrated retail markets. (The limitation is due to the data at local level, required given the study’s set-up, not being available for many countries.)
To illustrate, using the key measure of concentration relied on in the study, the Herfindahl-Hirschman index (HHI) for the seven countries included in the study ranged from 1200 to 2000 with an average of 1600. Conversely, the HHI for the 18 countries not included in the study ranges from 1700 to 3900 with an average of 2400. (All figures are for 2012 and have been rounded. A market may be considered moderately concentrated if the HHI exceeds 1500; a HHI value in excess of 2500 typically is taken to signal a concentrated market.)
This being the case, the study does not really deliver on its promise to investigate the impact on retail concentration on choice and innovation. It is not possible to investigate this impact if none of the countries with high levels of retail concentration are actually included in the study. In short: the study does not provide enough food for thought.
________________________
To make sure you do not miss out on regular updates from the Kluwer Competition Law Blog, please subscribe here.
We welcome this feedback on the modern retail study from Mr.
van der Veer. He highlights an important point about the lack of comprehensive relevant data for the highly concentrated member states. In an ideal world the study would have included all member states so as to cover the full range of situations in the EU. However the inclusion of local level data over the whole period investigated by the study was essential in order to bring data-focused analysis to the debate on how retailer practices upstream impact on the product offer available to consumers downstream. This unfortunately meant that those member states for which there
was no data at the shop level over the whole period had to be excluded from the study.
Nevertheless, this limitation in data availability is not as detrimental to the usefulness of the results of the study as Mr. van der Veer has concluded. Whilst the study does not represent the situation in member states with retail markets which are highly concentrated at national level, the study still captures the situation in the majority
of member states in the EU.
Mr. van der Veer notes that “A market may be considered moderately concentrated if the HHI exceeds 1500; a HHI value in excess of 2500 typically is taken to signal a concentrated market”. Within the sample of 26 member states for which HHI data at the national level is available, only 7 countries have an HHI above 2500 (based on 2012 figures), which represents just over one quarter of the sample. Of those, only one member state (Finland) has an HHI of above 3500, and two (Latvia and Sweden) have an HHI of more than 3000. The highly concentrated member states are therefore in a minority.
This means that 19 member states (nearly three-quarters of the sample) have a low-to-moderately concentrated retail market (HHI of below 2500), which coincides with the low-to-moderately concentrated coverage of the study.
As mentioned, we welcome feedback on the study, and the Commission invites stakeholders to submit their responses to the study in the public consultation which is open until 30 January 2015. (Please see http://ec.europa.eu/competition/sectors/agriculture/overview_en.html for more information). We are currently considering how to follow up the retail study with further analysis, and we strongly encourage stakeholders to put forward any suggestions.