Since 2016, so-called industry concentration has gradually received more attention among competition policy practitioners. In the US in particular, both academics and some policymakers have argued that American industries and the economy at large have become more concentrated. In South Africa, lawmakers have cited high and persistent levels of concentration as motivation for enacting changes to the competition law. In fact, the aim of South African lawmakers is now explicitly to “deconcentrate” the economy[i].

The concentrated nature of markets and the associated anti-competitive effects are not new ideas. In fact, these have generally been associated with the outdated ideas of the so-called Structure-Conduct-Performance school of thought. However, given recent pronouncements on increased levels of concentration, economists are slowly starting to respond to the idea that whole economies are concentrated or have become more concentrated. A recent piece by Greg Werden and Luke Froeb[ii] (both at the US Department of Justice) shows that concerns about high or increasing concentration in particular sectors or industries in the US is unwarranted. They explain how a misunderstanding of the concentration concept has fuelled the current debate.

In competition policy, concentration metrics are intended to measure the inequality of the shares of competing firms in a particular market. For example, when a few players dominate the market – i.e. enjoy large market shares – market concentration is considered high. On the other hand, when the market is characterized by a large number of players each commanding a small share of the market, market concentration is considered low. Werden and Froeb point out that the critical issue is that the concentration relates to a specific product and geographic market. The notion of a highly concentrated industry or sector, which typically consists of many (hundreds of) different markets, is therefore at best uninformative and, at worst, misleading.

There is a clear distinction between the concept of a market and the concept of an industry. In competition economics, a market relates to a particular set of substitutable products or services[iii]. An industry, in contrast, is a collective noun that consists of many hundreds of markets. We might talk of the ‘agricultural processing industry’, but this industry consists of a great many markets. One would have to consider the various products and services, at various levels of supply chains, in which agro-processing firms are involved. Relying, for example, on total turnover of a few large conglomerates to argue that the South African agricultural processing industry is concentrated is not informative for the purposes of understanding competition.

If the distinction between market and industry is interesting when it comes to the US antitrust debate, it is of critical importance in South Africa, where policy makers are actively seeking to reform competition policy on the basis of a ‘deconcentration’ argument. In our case, admittedly, policy makers have also relied on an empirical argument related to market, rather than industry, concentration. Even so, I note below that the empirical analysis is opaque and, given the information provided, not able to sustain the argument that concentration is high. Furthermore, lawmakers also erroneously extrapolate from particular market observations to larger industries and, indeed, the national economy.

In the background note accompanying the draft Competition Amendment Bill[iv] competition in South Africa, much is made of an empirical analysis of 2,150 merger decisions of the Competition Commission over the period 2009-2016. Without providing the public with an extensive analysis and the data, lawmakers merely assert that many of these mergers took place in highly concentrated markets and then implicitly argue that this is sufficient evidence that South African markets in general remain highly concentrated. This rudimentary analysis merits an extensive discussion[v], but it is sufficient to point out that the mergers are not dated and that markets are not identifiable in the dataset, which makes any inference about trends or indeed levels of concentration questionable. Market definition in merger cases is by definition specific to the set of circumstances being investigated. Aggregating different market definitions from different mergers into one industry is therefore a flawed approach which cannot be expected to provide objective evidence of concentration levels.

More important, though, is the assumption of lawmakers that the set of affected markets are representative of the South African economy. Given that a large number of these transactions may relate to particular industries, it is unlikely that this is the case. Even if one looked beyond this data to consider, for example, the extensive enforcement against collusion and other abusive practices in recent years, it is still not clear that the larger South African economy is uncompetitive or, at the very least, that most markets have remained uncompetitive. In the academic sphere in South Africa, much is made of evidence of persistently high mark-ups in the manufacturing industry, but there is little research on data of the past ten years[vi]. Competition in a great many other sectors and, given the previous comments, in specific markets within sectors have received far less attention.

The debate on competition policy in South Africa is marked by a certain one-sidedness, that seeks to accept as a given that large South African companies are not competing or are uncompetitive[vii]. The implication is that the competition authorities should do more.

While competition problems are perennial, the evidence is too limited to conclude that competition across many thousands of South African markets is weak or that “concentration” in the economy at large is high. At the very least, it is necessary to update and significantly expand the existing research on market concentration and competition in South Africa. This is one of the urgent tasks for South African economists, given the prominence that policymakers have assigned to competition problems as binding constraints on economic growth and investment. It seems particularly premature to recast policy in the absence of extensive research[viii].

Prof Willem Boshoff

Co-Director, CCLE



[i] Hyperlink to background document motivating the amendments 


[iii] This substitutability is determined by reference to various economic and legal tests, of which the Hypothetical Monopolist test is the most prominent.

[iv] Hyperlink to Competition Amendment Bill

[v] For some high-level views, see the CCLE’s comments on the proposed amendment bill at

[vi] With the notable exception of Fedderke, J.W., Obikili, N. and Viegi, N. 2017. Markups and Concentration in South African Manufacturing Sectors: An Analysis with Administrative Data. South African Journal of Economics 86(S1): 120-140. This analysis is based on 4-digit SIC manufacturing industry codes, which is subject to the same criticisms raised by Froeb and Werden.

[vii] For a notable exception, see Du Plessis, S.A., Katzke, N. Gilbert, E. and Hart, C. Mark-ups and competition: a comparison of the profitability of listed South African industrial companies. Stellenbosch Economic Working Paper 02/15.

[viii] More generally, there is a growing realization of the importance of countering fraught government-business relations in South Africa with better academic research, see: Keeton, G. 2017. “Post-truth”, “alternative facts” and “fakenomics”. South African Journal of Economics 86(1): 113-124.