By Maryanne Angumuthoo Friday, April 09, 2010

Lesson 1: In the context of a horizontal merger, a substantial lessening of competition cannot be inferred from the fact that a market is highly concentrated or the fact that a significant increase in concentration will result post-merger, where that market is substantially differentiated; having regard to differences in market characteristics including the product range and mix, customer profiles, and overall prices achieved by the merger parties.
Lesson 2: Economic analysis is only one step in a multi-evaluation process and does not obviate the need for a contextualised fact-based analysis of the particular market.
Facts and Findings
The facts of this transaction are that Massmart Holdings Limited (active in the retail and wholesale of grocery products, liquor and general merchandise in South Africa) sought to acquire a 75% interest in Finro Cash & Carry (a single family-owned independent wholesaler of groceries and general merchandise in Port Elizabeth). Massmart owns, amongst others, the Makro wholesale chain and Jumbo Cash & Carry. In the Port Elizabeth area specifically, Massmart owned two grocery wholesalers (Weirs Cash & Carry and Makro). 
The Competition Tribunal’s decision is noteworthy for three reasons:

It establishes the principle that where a market is substantially differentiated, a substantial lessening of competition cannot be inferred from the fact that the market is highly concentrated or the fact that a significant increase in concentration will result post-merger.
It lays a good foundation for the consideration of survey evidence, statistical analysis and the use of economic modeling tools to predict likely post-merger unilateral price effects - as an indicator of an incentive to engage in a substantial lessening of competition.
It illustrates the importance of good old-fashioned merger analysis which typically involves questioning relevant witnesses (customers and competitors of the merging parties) and obtaining evidence (about the market generally and the perceived effects of the merger) based on the experience of parties active in the relevant market.

The Competition Commission recommended prohibition of the merger on the grounds that the merger parties were close competitors and the merged firm would be able to significantly raise prices post-merger.
The Competition Tribunal came to a different conclusion and found that Finro and Weirs were significantly differentiated and as such could not be said to be close competitors. The Commission’s evidence illustrating the merged firm’s post-merger incentive to raise prices was therefore weak.
The Tribunal defined the product market as the market for the wholesale of grocery products (food and non-food) and general merchandise and the geographic market as Port Elizabeth and its regional surroundings. The horizontal overlap lay in the wholesale of grocery products. (The vertical overlap raised no concerns and need not be considered further).
The Tribunal’s decision focused on the theory of harm concerning anticompetitive unilateral effects. Unilateral effects occur when a merger enhances the ability of the merged firm to exercise market power independently. 
In deciding that the merger parties were not close competitors as a result of the market for the wholesale of grocery products (food and non-food) and general merchandise being characterised by a considerable degree of differentiation, the Tribunal took into account market characteristics which indicated differences between Finro and Weirs and Makro in relation to product range and mix, customer profiles, overall price, margins, location of the stores, delivery services and credit terms offered.
After considering market characteristics, the Tribunal considered market shares and market concentration levels.
Pre-merger, Massmart had a20% to 30% market share and Finro 10% to 20%, while post-merger the merged firm would have approximately 30% to 40% market share. This led to the Tribunal concluding that the relevant market would be highly concentrated post-merger and that the increase in concentration as a result of the merger would be significant.
However, the Tribunal noted that one cannot simply infer a substantial lessening of competition from a highly concentrated market or a significant increase in concentration where the market is significantly differentiated. In such markets, market shares and/or concentration levels are less informative about the degree of rivalry between firms and the merged entity’s potential market power post-merger.
More detailed analysis is required to infer a substantial lessening of competition, considering, inter alia, the characteristics of the relevant market and specifically the degree of market differentiation, or, in other words, the closeness of competition.
Economic analysis
The Commission in its economic analysis had applied diversion ratio analysis (DRA), an economic tool used to measure the degree to which merger parties are competitors; to measure the closeness of competition between the merger parties.
The Tribunal noted that it was standard practice in differentiated markets to determine diversion ratios as a quantitative measure of the closeness of competition between merger parties and then combine it with pre-merger gross margins and use economic modeling to predict potential price raising consequences of the merger.
Simply put, DRA asks: if Firm A raised prices what fraction of customers would switch to rival Firm B and vice versa?
Economic theory suggests that where the competitive restraint is removed (in this case by the merging of two competitors who previously acted as restraints upon each other), the merged firm would be incentivised to increase prices, since it would recoup the sales that would previously have been lost by customers switching.
The Commission conducted a customer survey to determine revenue diversion ratios. Using economic simulation, the Commission combined the ratios with the gross margins of the parties’ wholesale outlets to predict the likely post-merger price effects of the proposed deal.
As described, the Commission determined the diversion ratios in order to use the ratios as an input in an economic model that would predict the post-merger profit-maximising incentives of the merger parties, or, in other words, to predict the level of price increases by the merged firm post-merger.
The economic model compared a scenario where firms competed with each other in a scenario where the merged firm acts as a single profit-maximising entity.   Interestingly, the economic model predicted post-merger increases for Weirs at 0,6% (considered insignificant by the Tribunal) and Finro at 2%.
The Tribunal was extremely critical of the Commission’s lack of analytical robustness in the customer survey. It observed that sufficient statistical grounds existed for it to be sceptical of the Commission’s estimates:

the customer survey sample size was too small resulting in extreme uncertainty;
the Commission failed to evaluate the quality of the quantitative survey against industry, commercial and economic facts, including the views of market participants, company documents and other qualitative information sources;
the integrity of Nielsen data generated by the survey was questionable; and
there were concerns about the questions in the context of the relevant market characteristics – for example, the survey question considered price increases of 25%, which were absurd in the context of the wholesale grocery market. 

The Tribunal was equally critical of the Commission’s economic modelling and concluded that the Commission’s simulation:

showed a weak ability of the merged entity to increase prices;
did not allow for external supply side factors (reactions from rivals through new entry, expansion or product offerings or synergies that lower marginal costs and reduce the predicted price increase); and
did not take into account a number of constraints that mitigated the merged entity’s ability to raise prices (new entry by competing wholesalers, rivals responses, buying groups that served independent retailers and wholesalers and direct supply by manufacturers to retailers which bypassed wholesalers).

On this basis, the Tribunal held that the merger was unlikely to substantially prevent or lessen competition from a horizontal or vertical perspective.
Lessons and Guidelines
Market shares and concentration levels are not effective proxies for the likely extent of sales diversion between merger parties in a differentiated market as they are not a reliable indicator of market power in a differentiated goods market.
Even where a target firm is an effective competitor to the acquiring firm (as was the case with Finro in relation to Weirs and Makro), the effectiveness of the target and a high post-merger concentration must be assessed within the context of other qualitative and quantitative evidence taking into account:

substantial market differentiation;
the existence of significant competitors post-merger (in this case there were several significant competitors including three large wholesalers with market shares greater than 10%);
the ability of competitors to increase market share (at least one competitor had increased its market share in the previous two years and successfully repositioned its product offering); and
competitors that would become significant players in the near future (one such players existed in this case).  

This is the first contested matter before the Tribunal that involved extensive economic modeling and customer survey and statistical data analysis. As such, the comments and recommendations made by the Tribunal are of relevance to the approach that it will adopt in future cases.
As a general principle, the Tribunal indicated that it was highly supportive of the increased use of economic analysis in merger cases supported by economic evidence. 
The following guidelines are extracted from the decision:

Statistical data and analysis are case-specific and should therefore be assessed case-by-case;
Customer surveys are useful additions to evidence provided they are well designed. (The Tribunal accepts uncertainty regarding the reliability of surveys and this does not detract from their useful contribution as evidence);
Questions in surveys that are likely to give rise to bias must be avoided and the integrity of a survey must be maintained; 
To provide empirically robust measurements, a survey should be designed with its intended purpose as the key objective (which was not done in this case);
Surveys should be custom-designed to the relevant market and take into account and be tested against commercial realities before being utilised;
Economic modeling is only one step in a multi-evaluation process and does not obviate the need for contextualised qualitative analysis focused on a particular theory of harm (here unilateral anticompetitive effects and adverse public interest effects on small business);
Supplementary evidence in the form of large scale quantitative surveys are rigid and do not provide in-depth or detailed responses. (In the Tribunal’s view, the Commission should have obtained these detailed responses from grocery manufacturers and buying groups); and  
Merger simulation models like the ones in this case are static and focus exclusively on customers’ demand-side responses (gross margins and revenue diversions ratios) and do not consider off-model external supply-side responses. The failure to account for these overstates likely post-merger anti-competitive effects.

The Commission sought to prove that the merged firm would be able to significantly raise prices post-merger, but its own painstaking economic analysis revealed the exact opposite.
Why did it go to such trouble on the economic evidence but ignore the facts at its disposal – facts that pointed it to approving the merger? Was the Commission determined to prohibit the merger and sought to rely on economic theory to arrive at the foregone conclusion?
Whatever the reasons, this case provides a useful lesson about the pivotal role that should be played by the traditional methods of obtaining factual evidence to support a case. Perhaps this is why the Tribunal’s finding that the market was substantially differentiated did not signal the end of its analysis.
The concern about a horizontal overlap in merger analysis is that the merged entity will acquire market power in the market in which the overlap exists that may allow it to act to the detriment of customers and competitors. If the market is significantly differentiated, then the extent of the identified horizontal overlap is limited and consequently, the ability to acquire market power. In such circumstances a substantial lessening of competition post-merger is unlikely and the Tribunal need not have gone further.
However, having easily determined, by a simple factual analysis of the relevant market, that the merger parties were not close competitors, the Tribunal then engaged in a lengthy consideration of the economic analysis used by the Commission. Perhaps it did so in order to prove that this analysis supported its conclusion on the facts (that the merger parties are not close competitors) and, importantly, to prove, in careful detail, that the Commission had reached the wrong conclusion on the same facts, as a result of ignoring traditional methods of merger analysis.
While the Commission should be given credit for its engagement in economic analysis, its eagerness to find economic support may have caused it to lose sight of the factual case that needed to be made – one that took the commercial realities of the wholesale grocery market into account, realities that would have been readily apparent if it had obtained detailed responses from grocery manufacturers and buying groups.