Measuring Concentration in the South African Economy: Some Cautionary Remarks
June 30, 2022
Measuring Concentration in the South African Economy: Some Cautionary RemarksDownloadsDownload Article
The absolute and relative sizes of large firms in South Africa have attracted growing attention from the competition authorities in recent years.1
The most recent evidence of a policy concern with concentration is the Competition Commission’s concentration tracker report, in which the Commission seeks to measure concentration in the South African economy at large.2
The concentration tool of the Commission is relevant to corporate counsel and other practitioners, as the Commission implies that some of its future policy actions will be motivated by concentration concerns. Indeed, the recent tracker flags several industries and indicates the likelihood of future market inquiries on the basis of these concentration findings.3
Understanding the Approach
An understanding of the economics content of the Commission’s approach to measuring concentration is important to constructive policy dialogue in South Africa and potentially critical to competition law proceedings, including market inquiries.
Economics provides important guidance when evaluating the Commission’s concentration findings, both conceptually (in terms of appropriate measures of concentration) and empirically (in terms of acknowledging and dealing with data challenges underlying concentration calculations).
Indeed, economics gained its prominent role in competition policy exactly for its contributions in clarifying and measuring market-based concepts, including concentration.
Conceptually, it is worth making three remarks about the Commission’s approach to measuring concentration from an economic perspective.
First, concentration is traditionally a marketbased concept in competition policy. As practitioners know, competition policy views a market as a carefully defined collection of products that are substitutable to buyers.4 Concentration measures are based on revenue or volume market shares enjoyed by the various firms operating in the market.
The Commission acknowledges that its concentration measures are not market based, but it nevertheless infers market power on the basis of these sector measures.
For example, the Commission claims that “[its] study shows that once markets are concentrated, the trend is towards more concentration”(own emphasis).7
In some sectors, relatively similar players may be active across most of its constituent markets, so that sector-level concentration could be an indicator of market power across markets in a particular sector.
Second, in contrast, the Commission’s tool focuses on concentration as an industry- or sector-based concept. Any one industry or sector may consist of several and indeed hundreds to thousands of different markets, implying that the Commission’s approach effectively pools this diversity into aggregate numbers.6
Such conceptual ambiguity is also important for another, if related, goal of South African competition policy, namely promoting HDI9 ownership. Higher levels of HDI ownership need not translate into a broader ownership distribution, measured in terms of the size of shareholders.
This conceptual ambiguity is worsened further when referring to sectoral concentration to make claims about ownership concentration.
Beyond providing conceptual clarity in dealing with concentration claims, economics also provides guidance for empirical analysis, equally relevant to the concentration findings of the Commission.
Third, the Commission’s flexible approach to concentration goes even further. The Commission claims a necessary link and, at times, an equality, between market concentration and ownership spread, by which it means the distribution of corporate ownership. The promotion of a greater spread of ownership is a goal of South African competition law.8 But it is critical not to conflate the concept of ownership concentration with that of market-share-based concentration. Large oligopolists may well be owned by a variety of owners.
Concentration estimates are based on particular firm data, which are used to estimate relative firm sizes. The Commission will have to be more open to publicly sharing its firm data and justifying its choices for categorising firms into particular sectors.
More importantly, if the current concentration tool is indeed to be a ‘tracker’, future updates must be expected.
Yet future updates to the Commission’s concentration findings are likely to be based on data samples that vary in terms of the number and identity of firms included, and their categorisation, compared to the current sample.
The subsequent concentration findings may be comparable to previous findings, but only to the extent that statistical and economic differences in samples are properly accounted for.
For all these reasons, great care is required when relying on concentration analysis to orientate the direction of South African competition policy.
As always, a proper economic analysis of competition concerns entails far more than an analysis of structure.
Even if structure is considered important, its measurement entails key definitional and data decisions that must be both open to interrogation and consistent with the principles of competition economics.
1: A strong concern with so-called ‘de-concentration’ is clear from the amendments to the Competition Act, but also from preceding and subsequent policy documents and investigations, which identify firm size and distribution as a constraint on South African economic growth and dynamism.
3: See slide 15 of the presentation “Measurement of concentration and participation in the South African economy: levels and trends”, dated 15 March 2022.
4: See, for example: Boshoff, W.H. 2014. Market definition as a problem of statistical inference. Journal of Competition Law and Economics 10(4): 861-882.
5: The Herfindahl-Hirschman Index, a commonly relied upon concentration metric, is defined as the sum of the squared market shares (for an accessible introduction, see: https://www.justice.gov/atr/herfindahl-hirschman-index).
6: For an accessible piece by prominent US practitioners, Luke Froeb and Greg Werden, see: https://www.competitionpolicyinternational.com/dont-panic-a-guide-to-claims-of-increasing-concentration/
7: See slide 13 of the presentation “Measurement of concentration and participation in the South African economy: levels and trends”, dated 15 March 2022.
8: See, for example, the second paragraph of page 1 of the ‘Summary Report of Findings and Recommendations’ in relation to the concentration tracker.
9: Historically Disadvantaged Individual
June 30, 2022