In late January, the Commission finally published the non-confidential version of its decision in Unilever/Sara Lee Body Care (adopted in November 2010). The decision, reached after a Phase II investigation, is notable because it marks the first time in many years where the Commission objects to a proposed merger on the basis of, among other pieces of evidence, a merger simulation.
The key area of overlap between the Parties was in deodorants. After having decided that separate markets existed for male and non-male deodorants, the Commission found that the Parties were close competitors with regard to certain brands and that Sara Lee contributed to important innovations in the market. But in addition to these “traditional” pieces of evidence, the Commission appears to have put significant weight on the fact that the merger simulation showed price increases (as also shown by the fact that the decision devotes an annex comprising more than 70 pages to the simulation exercise). In order to remedy the Commission’s concerns, Unilever ultimately offered to divest Sara Lee’s Sanex brand.
Merger simulations are based on a formal economic model as described in the economic literature. Once this has been specified for the industry in question, both as far as the demand side and the supply side are concerned, it can be used to produce estimates of price increases that a merger may generate. Merger simulations require extensive datasets, which in fast moving consumer goods (FMCG) industries are typically readily available in the form of retail scanner data.
Following a long time in which the Commission did not rely on this technique at all, the first recent case in which the Commission referred to a merger simulation was Kraft/Cadbury, another FMCG case decided in 2010. In that case, the Commission carried out its own economic analysis, not using merger simulation, to show that the proposed merger did not raise competition concerns in the market for chocolate tablets in the UK. But in addition, it referred to a merger simulation undertaken by the parties which, as the Commission claimed, provided further evidence that the proposed transaction was unlikely to give rise to significant price increases.
In Unilever/Sara Lee, the Commission has gone much further by undertaking a simulation itself in order to support a finding against the proposed merger. The Commission carried out a simulation for eight markets (male and non-male deodorants in four countries), finding estimated price increases between 1 and 6%. Competition concerns were ultimately raised for five of these markets, all with estimated price increases of 2% or higher. In the three remaining markets for which no concerns were raised, estimated price increases were all around 1%. Even though these estimated price increases naturally have to be assessed in conjunction with the other evidence on which the Commission relied, the fact that the Commission is prepared to find competition concerns if estimated price increases are as low as 2% suggests a very low threshold for intervention.
This applies in particular given the uncertainty that is inherent in this type of analysis. While the simulation that the Commission undertook in Unilever/Sara Lee is more detailed than the one it referred to in Kraft/Cadbury, the Commission’s approach also relies on many assumptions to which the results of the analysis can be sensitive. Questions also remain regarding the extent to which the analysis undertaken by the Commission adequately takes account of, to mention just a few issues, the fact that the demand for FMCG goods can be heavily driven by promotions (during which many consumers tend to purchase significant quantities which they then store and consume later); the act that while the merger occurred at the upstream level between manufacturers, the simulation was undertaken using data on prices against which their products are sold to final consumers via retailers; or retailers’ buyer power. More generally, merger simulation models, which are entirely static in nature, tend to always predict a price increase in the absence of efficiencies. By contrast, real-life markets may be characterized by entry, innovation, product repositioning, etc., all of which may defeat a hypothetical price increase but none of which are captured by merger simulation models.
Competition authorities are increasingly using techniques aimed at directly estimating the likely price effects of mergers. For example, the UK authorities now regularly employ techniques such as Upward Pricing Pressure (UPP) or Illustrative Price Rise (IPR) tests, which are essentially highly simplified merger simulation approaches. While the Commission does not (yet) appear to have made use of such techniques, it is now showing a renewed willingness to challenge mergers based on full simulation approaches. While the Commission accepts that this technique has shortcomings (as do the simpler approaches), it insists that the exercise increases the overall reliability of its assessment. Against this background, it will be interesting to see how the Commission’s use of these approaches will evolve going forward.