The role for competition policy in economic development: the South African experience

Abstract

Competition policy is part of the new international orthodoxy in economic policy and, at the same time, was viewed in South Africa as a crucial element of economic transformation. This article reviews the role of competition policy in economic development and the experiences of developing countries such as Brazil and South Korea. It then assesses the effects of competition policy in South Africa after 1994, with the main focus being on the performance of the new competition institutions established in 1999. The case of the steel industry is used to assess the approach and impact of the institutions in a concentrated sector that has simultaneously undergone processes of liberalisation and domestic consolidation.

1 INTRODUCTION

Competition policy is part of the emerging orthodoxy in economic policy. It is promoted by major donors and international organisations as part of the government's role in establishing rules for the business environment in the context of reduced state intervention and trade liberalisation (see, for example, Unctad, 1997; Stiglitz, 1998; World Bank, 1999). South Africa has followed this international pattern with a new Competition Act coming into force in 1999. The legislation, however, also reflects the African National Congress (ANC) government's concerns with the concentration of ownership and control in the South African economy.

The new South African competition law forms an important part of reforms designed to both address the historical economic structure and encourage broad‐based economic growth. The government has recently developed the ‘Microeconomic Reform Strategy’ in which the role of competition policy is identified as central to the efficient outcomes of markets. Competition policy is seen as important in increasing competitive market pressures, leading to firms becoming more efficient and internationally competitive. It is also viewed as important for the improved participation of black‐owned companies in the economy.

There is little doubt that corporate ownership and control in the South African economy is highly concentrated. In the latest available manufacturing census (StatsSA, 1996), for 46 per cent of the 57 main product groupings the largest four firms account for more than half the output, while in a further 35 per cent of groupings the four firms' concentration ratio is between 0,25 and 0,50. Concentration is even greater if measures of firm size are based on control, which is often exerted through minority stakes and holding companies.

Historically, four main conglomerate groupings – Anglo American Corporation, Sanlam, Liberty Life and Rembrandt/Remgro – have controlled the majority of economic activity (Table 1). While the control exerted by the four largest groups declined to 1998, this pattern has reversed in recent years.2

Summary of control of JSE market capitalisation (% of total)¹

	1985	1990	1991	1992	1993	1994	1995	1996	1997	1998	1999	2002^4
Anglo American	53,6	44,2	42,4	33,7	38,2	43,3	37,1	27,5	22,6	17,4	22,3	25,4
Corporation
Sanlam	12,2	13,2	13,2	15,6	12,0	10,5	12,4	11,0	10,6	9,9	12,5	4,2
Liberty Life	2,0	2,6	3,7	4,7	6,2	7,2	7,3	11,1	11,9	9,5	7,1	3,2
Rembrandt/Remgro	3,8	13,6	15,2	14,6	15,5	13,0	7,8	10,6	9,9	9,0	10,4	13,7
SA Mutual/Old	10,6	10,2	10,4	14,2	10,7	9,7	11,2	10,2	11,4	8,8	10,6	13,3
Mutual
Anglovaal^2	2,1	2,5	2,9	2,9	3,4	3,6	2,9	3,0	1,5	0,8	0,7
Black‐owned groups^3	–	–	–	–	–	–	–	6,3	9,3	8,9	6,8
Top four groups	80,2	81,3	82,0	79,9	79,2	80,1	79,4	69,7	64,9	53,9	62,2	59,8
collectively
Notes: 1. Control is assessed by McGregors, taking into account the various cross-holdings of shares that exist, and may be associated with a relatively small direct shareholding in any given company. 2. In 1998, the Anglovaal shareholding was split equally, giving the Hersov and Menell families each control over 0,4 per cent of the Johannesburg Stock Exchange (JSE) capitalisation. 3. The black‐owned groups are identified as such by McGregors on the basis of ownership. McGregors did not include such an identification for 2002. 4. Shares for 2002 were calculated from the identification of control and the market capitalisation for November 2002 given in McGregors (2003). The holdings of the Rupert and Hertzog families in Richemont and Venfin meant they were included under Remgro, although the two companies are held separately. Sources: McGregors (1999, 2000, 2003).

The opening‐up of the economy through trade liberalisation has also seen increased concentration in many sectors. This is a result of consolidation, with inefficient firms closing down or being taken over, and of closer focus by companies on their core activities. Economies‐of‐scale arguments have also been used in several sectors to support mergers and acquisitions.

This article focuses on three key groups of issues. First, it briefly discusses the role of competition policy in economic development and in developing countries. Secondly, it examines the ways in which competition measures have been applied in practice in South Africa and the development of the institutions under the 1998 Competition Act (South Africa, 1998) compared with the old Competition Board. Thirdly, it analyses recent cases and current competition issues in the steel sector to explore the issues of industrial development and competition policy in more detail.

2 ECONOMIC DEVELOPMENT AND THE OBJECTIVES AND PRACTICE OF COMPETITION POLICY

Competition is commonly discussed by microeconomists in terms of different market structures, against the benchmark of perfect competition where firms are all price takers. However, competition itself relates to behaviour, in particular the degree and nature of inter‐firm rivalry (Vickers, 1998). In orthodox economics, firms that are dominant are able to make profits by constraining supply and raising prices above marginal (and average) cost. Where there are a small number of firms, they can collude rather than compete so as to collectively generate monopoly profits.

Leaving aside the per se type prohibited agreements, the challenge of competition law is to address those practices that have a negative effect on economic efficiency through raising prices and barriers to competition (Hay, 1993). However, assessing this may be very difficult, especially in the presence of economies of scale, shocks to production costs and multiproduct firms. These and other concerns underpin what Phlips (1998) has termed the ‘indistinguishability theorem’. That is, it may be very difficult to tell whether a particular outcome is due to deliberate uncompetitive behaviour or is the result of non‐collusive or price‐taking behaviour, given the conditions of the industry. At the very least, competition assessments are very information intensive, especially given the asymmetries that exist between the firm and the authorities. Unless the competition authority is to become a regulator and determine acceptable price levels directly, then it must be able to distinguish conscious anticompetitive actions from the ‘innocent’ interplay of other factors.

If dynamic factors and externality effects are taken into account, then it would be more appropriate to apply a concept of optimal rather than maximum competition (Singh & Dhumale, 1999; Singh, 2002). In this case, competition policy should be seen as part of industrial policy, following the examples of Japan and South Korea (Amsden & Singh, 1994). The active rivalry of firms and the impact on their investment and production decisions are encouraged as an integral part of working towards identified industrial development goals (Sakakibara & Porter, 2001).

2.1 Competition and different business models

Management and economic history address the important role of competition within economic development from somewhat different starting points to textbook microeconomics. A firm's strategy is fundamentally about its competitive position vis‐à‐vis its rivals. This position is due partly to the environment within which it operates in terms of access to inputs, its cost structure and so on, and is partly determined by its decisions. For example, vertical integration may be an important part of ensuring better access to key inputs than rival firms. Similarly, advertising to build a brand identity is part of maintaining leadership, distinguishing one's product and deterring entrants – in each way reducing the negative immediate effect of competitive pressures on product price and revenue streams.

The orientation of large businesses and the ways in which they interact or compete are a central part of a country's development trajectory (Chandler, 1990; Chandler et al., 1997, 1998). The internal organisational capabilities of firms are essential to the ways in which they adopt and exploit new technologies, and realise economies of scale and scope (Best, 2001). Through cooperation and interrelationships, firms within large corporate groupings in Japan and South Korea have been able to build dynamic competitive capabilities (Amsden, 1989, 1997). However, competition between groups has been a very important disciplining and motivating factor in ensuring that interfirm cooperation is part of constructive processes and not aimed at collusion and rent extraction. The dynamic rivalry that constitutes competition is thus a very important element of Chandler's characterisations of different capitalisms.

The behaviour of firms and their interrelationships cuts across the adoption of different production systems, the different business models governing firms, and factors such as innovation and skills development (Best, 2001). South Africa's new industrial policy identifies the utilisation of knowledge in production and effective employment of information and communications technologies as fundamental to building production capabilities. The nature and extent of competition will influence the pressures to deploy such technologies.

2.2 The practice of competition policy in developing countries

Different developing country experiences highlight the importance of moving beyond the formal regulatory framework in considering countries' competition policy choices, and they reveal a diversity of approaches in practice. In general, it is the criteria that are applied, the implementation approach and the overall level of commitment that are most influential (Hoekman, 1998).

In contrast with South Africa, and along with industrialised countries such as Japan and Norway, several developing countries emphasise measures for addressing the behaviour of dominant firms. This is consistent with the significance of economies of scale, dynamic effects related to technology and the importance of production linkages in processes of industrialisation. In countries with small domestic markets, the assumptions for perfect competition to apply and be first best are even less likely to hold, and the nature of interfirm rivalry and dynamic efficiency effects requires more attention in competition policy.

For example, the objectives of the South Korean Fair Trade Commission (KFTC) are to encourage free and fair competition, prevent the concentration of economic power and thereby promote ‘balanced development’ (Wise, 2000). Most importantly, free and fair competition is defined broadly in the sense of a competitive industrial structure and the control of potential abuses and imbalances in the bargaining power between parties, specifically in subcontracting arrangements. The KFTC examines ‘unreasonable’ practices and ‘unjustifiable’ restrictions on competition. By comparison, a recent survey of competition policy in Brazil finds that too much attention has been paid to mergers, which are essentially to do with competition in terms of the number and size of firms, and that more attention needs to be given to preventing damaging collusion by firms and abuse by dominant firms of their position (Clarke, 2000). With reference to Brazil, it has also been argued that it is important to avoid an overly legalistic approach in dealing with dominance due to issues of transition and industrial development under liberalisation (Stevens, 1995).

Despite Japan and South Korea having competition laws strongly influenced by the United States (US), the competition institutions and implementation of competition measures have been closely linked to the government's industrial policy (Amsden & Singh, 1994). In South Korea, while the share of the top 30 conglomerates or chaebols in manufacturing and mining production has remained at about one‐third since the 1980s (Wise, 2000), dynamic competition processes have played a very important disciplining role (Khemani, 1994; Amsden & Singh, 1994). Ensuring rivalrous behaviour rather than addressing structure has been the more important concern for the competition authority. The KFTC identified 332 cartels between 1981 and 2001 (OECD, 2002) and in 2002 alone, 43 cartels were detected and fines imposed (Lee, 2002). The KFTC also dealt with 3 130 cases related to unfair subcontracting in 2001 (OECD, 2002). In addition, the KFTC has been able to ‘designate’ the largest conglomerates, meaning that these groupings were subject to close monitoring of their performance and possible abuses of their dominant position.

The importance of appropriate implementation mechanisms is further highlighted by the many examples of countries with competition laws but no effective mechanisms for their realisation. The Philippines has had antimonopoly laws since 1935, and the 1987 national constitution declares that ‘the State shall regulate or prohibit monopolies when the public interest so requires. No combinations in restraint of trade or unfair competition shall be allowed’. There is, however, no central agency to oversee the legislation and no administrative mechanisms to give effect to the provisions (CUTS, 2000). Countries with legislation that was not being implemented until the last decade include Argentina, Brazil and Mexico, highlighting the importance of political will and the appropriate institutional framework (Stevens, 1995). There are also countries that have achieved rapid growth, such as Malaysia, which have not prioritised competition policy. Alternative policy instruments have been used to impact on the behaviour of large companies.

Appropriate and effective competition regimes must also take account of countries' institutional capabilities. United States‐style competition policy, drawing on large numbers of lawyers and economists, is not an optimal use of scarce expertise and will favour well‐organised interest groups and large corporations (Stevens, 1995; Laffont, 1998). By comparison, the interpretation of the legal measures in South Korea prohibiting ‘unreasonable’ practices by dominant companies and ‘unjustifiable’ restrictions on competition allows the KFTC substantial discretionary power. As important, perhaps, for the ability to exercise this power is the close links of the KFTC with the powerful Economic Planning Board, of which it was a part until 1994 (Sanekata & Wilks, 1996; Wise, 2000). By comparison, in countries such as Peru and Venezuela the competition authorities may be independent but their effectiveness is hampered by a lack of resources and political economy factors within the countries (Rodriguez & Williams, 1998).

The experience of developing countries therefore demonstrates a diversity of approaches rather than a focus on one model, especially when the institutional set‐up and practice are examined.

3 COMPETITION POLICY IN SOUTH AFRICA

South Africa's Competition Board, which existed until 1999, operated under the provisions of the Maintenance and Promotion of Competition Act, No. 96 of 1979. This legislation provided for a review of mergers and acquisitions, restrictive practices and monopoly situations. The Board was, however, an administrative body without executive authority, which was housed within the Department of Trade and Industry and made recommendations to the Minister. The Minister then decided whether to accept or reject the recommendations, and the action to be taken.

The Board had powers to review acquisitions, which had the effect of restricting competition. This was taken as limiting the Board to considering horizontal transactions. It had wide‐ranging powers to investigate transactions but, with no pre‐merger notification required of firms, the Board relied on complaints from other parties or voluntary notification. In practice, many of the investigations proceeded on an informal or consultative basis. Only a relatively small number of mergers attracted the attention of the Board and very few were assessed in any depth (Table 2).

Acquisition assessments by the Competition Board

	1993	1994	1995	1996	1997
Preliminary assessments	41 (40)	42 (40)	34 (30)	24 (20)	35 (35)
Consultations	10 (10)	–	2 (2)	– ^2	– ^2
Investigations	3 (2)	3 (2)	3 (1)	6 (4)	2 (1)
Notes: 1. Numbers refer to total cases, i.e. completed and uncompleted. Numbers in parentheses refer to cases completed in a particular year. This means that the same case may be recorded in two or more years if the investigation was carried over. 2. In 1996 and 1997, preliminary assessments and consultations are grouped together. Preliminary assessments refer to assessments of whether transactions could constitute acquisitions as defined under the Act. Consultations refer to formal processes of engagement in connection with an acquisition. Sources: Competition Board (1996) and various other years.

The Board also had wide‐ranging powers to investigate restrictive practices, defined as ‘any agreement, understanding, act or omission, and situation which restricts competition by having or likely to have the effect of, inter alia, restricting the production or distribution of any commodity, enhancing or maintaining prices, preventing or retarding the introduction of new technology, and preventing or restricting new entrants into any market’ (Competition Board, 1996: 3). In addition, there were also provisions to address ‘monopoly situations’ defined as ‘where any person, or two or more persons with a substantial economic connection, control in the Republic or any part thereof, wholly or to a large extent, the class of business in which he or they are engaged in respect of any commodity’ (Competition Board, 1996: 30). These provisions enabled the Board to address both structure and behaviour, with possible remedies (to be ruled by the Minister) that included divestment. As with mergers and acquisitions, the Board's actions were relatively limited, although it received many complaints and became increasingly active towards the end of its life.

Rather than the strength of the legal provisions, the most important issues for understanding the operation of the Competition Board are its link with the government and its institutional capacity. The impact of the Board essentially depended on whether the government wished to use it as a tool for pursuing policy goals. With the apartheid government's close relations with big business, this was not the case before 1994. Members of the Board changed under the post‐1994 democratic government, and there were several important decisions such as the prohibition of major elements of the proposed SASOL–AECI merger in 1998 (Competition Board, 1998).

The Board's capacity also restricted its scope. With few staff and resources it could not take on large cases. It had also developed a culture of informal negotiation with firms, relying on them for information rather than seeking independent sources. This makes it difficult to assess the approach adopted and evaluation conducted. The Board also continued to be complaints driven instead of assessing practices in highly concentrated industries. Again, however, the change in government had gradually brought about a change in approach. There were several investigations of behaviour in concentrated industries, such as pharmaceuticals.

3.1 The Competition Act of 1998

The major criticism by the ANC‐led government of the competition legislation it had inherited was that it was weak in addressing the extent of concentration of ownership and market share (DTI, 1997). It was widely accepted that the Competition Board was relatively ineffective (Fourie et al., 1995). After being on the policy agenda since 1994, negotiations on a new law were undertaken in earnest in 1998, with the new Competition Act coming into force on 1 September 1999. The Act made provisions to establish a Competition Commission with primary responsibility for determining and investigating cases under the Act, and a Competition Tribunal to rule on most cases. A Competition Appeal Court was also to be established.

The Act deals with two main areas: prohibited practices (covered in Chapter 2 of the Act) and mergers (Chapter 3). The prohibited practices are further separated into restrictive practices – either horizontal or vertical – and abuse of a dominant position. The objectives of the Act are broad and take into account a range of concerns that will not necessarily be consistent with each other in the actual evaluation of cases. They are stated in section 2 of the Act as follows:

The purpose of this Act is to promote and maintain competition in the Republic in order:

1.	To promote the efficiency, adaptability and development of the economy;
2.	To provide consumers with competitive prices and product choices;
3.	To promote employment and advance the social and economic welfare of South Africans;
4.	To expand opportunities for South African participation in world markets and to recognise the role of foreign competition in the Republic;
5.	To ensure that small and medium‐sized enterprises have an equitable opportunity to participate in the economy; and
6.	To promote a greater spread of ownership, in particular to increase the ownership stakes of historically disadvantaged people.

There has been much discussion and criticism of the broad scope of the objectives (Reekie, 1999). To an extent the objectives reflect the differing pressures on policy makers, and their prioritisation will depend on the development of precedents from cases. When the criteria set out in the Act for the evaluation of cases are examined, it is evident that economic efficiency is the overriding principle. This reflects the fact that the South African Act drew heavily on the Canadian legislation for provisions governing mergers and on Australian legislation for prohibited practices.

The major change in the legal provisions from the previous Act was the introduction of compulsory pre‐merger notification for all deals above specified thresholds. An even greater change in the institutional framework was to make the new institutions independent of the government. Concerns over limiting the discretionary power of the competition authority led to a separation of powers with three different institutions. There is also an emphasis on transparency, and the ability of interested parties to participate in hearings of the Tribunal. Tight timeframes and procedures have to be followed by the authorities in investigating cases.

Apart from a role in appointments of the Commissioner and Tribunal members, the Minister only has a right to make representations on public interest grounds as a party to merger proceedings (section 18 of the Act), or may make representations at hearings of the Tribunal where there is a material interest on the part of the government (section 53).

3.2 An overview of cases

The effectiveness of the institutions can be assessed in a number of different ways. These include the processing of cases, the selection of complaints, the enforcement of rulings, and advocacy. Awareness of the measures of the Competition Act itself will affect the behaviour of firms, especially if they believe there is a credible threat of detection and remedy for actions that breach the provisions of the Act.

3.2.1 Mergers

The most striking feature of the institutions' activities is the amount of effort expended on merger evaluation in contrast with the concern that had been expressed previously on the need to address concentration. During the period from September 1999 to March 2002, a total of 958 mergers were notified to the Commission, of which 36 were large mergers on which the Commission was required to make a recommendation for judgement by the Tribunal. The rigorous standards required for merger evaluation, compulsory notification and the increase in mergers and acquisitions all contribute to explaining why so much of the authorities' energies have been expended in this area.

While only very few mergers have been prohibited or approved subject to conditions (Table 3), important precedents have been set. Merger decisions also give an insight into how different concerns of competition, economies of scale, international competitiveness and public interest factors such as employment are being taken into account.

Mergers prohibited, or approved subject to conditions, 1999–2003

Merging parties	Ruling	Reason/conditions
Bromor foods–Game	Approved subject to conditions	Brands to be maintained as competitors
Joshua Doore–Ellerines	Prohibited	Horizontal merger in concentrated market
Nasionale Pers–Educor	Approved subject to conditions	Subject to maintaining competing operations
Tongaat Hulett–Transvaal Suiker Beperk	Prohibited	Horizontal merger in concentrated market
Nampak–Malbak	Approved subject to divestment	Horizontal overlap in one market
Nedcor–Standard Bank	Prohibited by the Minister of Finance	Horizontal merger in concentrated market
Iscor–Saldanha Steel	Approved subject to conditions	Failing firm justified anticompetition merger, with conditions that compete in the local market
Nestle–Heinz/Tiger Foods (Pet foods)	Approved subject to conditions	Horizontal merger, subject to divestment
Mondi–Kohler Cores & Tubes	Prohibited	Vertical merger, but possibility of foreclosure and collusion
Astral–National Chick	Approved subject to conditions	Horizontal and vertical merger, with conditions on conduct and divestment
Distell–SFW	Approved subject to conditions	Horizontal overlap in one market
Joshua Doore–Profurn	Approved subject to conditions	The failing firm justified the merger, on condition that the firm should continue to buy from independent suppliers
Coleus (SAB)–Rheem	Approved subject to conditions	Vertical merger, subject to future divestment
Chemserve–Senmin	Approved subject to conditions	Continue to supply competitors of vertically integrated subsidiary
Note: In the Unilever–Bestfoods merger, divestment of some operations was agreed on by the parties before the Tribunal hearing.

Section 16 of the Act sets out the statutory standard for merger evaluation. This includes assessing competition in the identified market, taking into account the actual and potential level of import competition, ease of entry, countervailing power, as well as the removal of an effective competitor. Technological and efficiency gains that could offset any potentially anticompetitive effects resulting from the merger are also considered. Public interest issues that may be taken into account include employment; the ability of small businesses, or firms controlled or owned by historically disadvantaged persons, to become competitive; and the ability of national industries to compete in international markets. These are, however, secondary concerns that may be set against competition implications if they are deemed to be very significant.

Several mergers that have increased concentration have been approved with conditions. The power of the Tribunal has been exercised in these cases by its analysis of factors determining economic efficiency over time. International competitiveness and economies of scale were taken into account in allowing the mergers of Trident Steel and Baldwins Steel, and Iscor and Saldanha. The need for consolidation in the face of increased international competition and/or a failing firm underpinned qualified approval in many other cases in Table 3, such as those of Nampak–Malbak, and Coleus–Rheem. In the ruling on the Trident Steel and Baldwins Steel merger, the Tribunal discussed the various efficiency considerations along with international precedent. It approved the merger without conditions, despite a clear finding that ‘the merger does substantially prevent or lessen competition’ in one of the key markets (Competition Tribunal, 2001). While it acknowledged the difficulty in quantifying claimed efficiencies (especially dynamic ones), it set a precedent for judging the ‘order of magnitude’ in balancing the differing effects. The weight placed on such efficiencies is consistent with the concern over encouraging investment and taking into account employment, which has social efficiency effects where there is a high of unemployment.

The second notable pattern in merger rulings has been scepticism of vertical mergers that do not, by definition, directly result in increased market concentration. The Tribunal has instead emphasised the following concerns (taken from Riordan & Salop, 1995), which highlight a willingness to take into account behavioural issues, even though in the case of mergers this is a forward‐looking exercise:

1.	The potential to raise rivals' costs by means of foreclosure
2.	Where a merger increases the ability of firms to coordinate conduct, such as through the exchange of competition‐sensitive information
3.	Where a merged firm can evade price regulation

Based on the rulings on mergers, a balance is being struck between competition, a broad definition of efficiency and public interest concerns such as employment. While the construction of efficiency appeared in earlier evaluations to be biased towards a structural understanding of competition rather than the behaviour of firms, this has been less evident in more recent cases.

3.2.2 Prohibited practices

Between September 1999 and March 2002, the Commission received 329 complaints of prohibited practices, of which the majority referred to multiple provisions, or to section 8 of the Act dealing with abuse of dominance. Very few, however, have been referred to the Competition Tribunal for adjudication. While this is partly due to complaints being misguided or frivolous, given the high levels of concentration in the economy and indications of monopolistic or collusive behaviour, the limited progress in this area is of concern. There have been several consent orders in cases where a per se provision of the Act had been transgressed. Aside from these, the Tribunal has only ruled on two substantive restrictive practices:

1.	Abuse of dominance by Patensie Citrus packing and marketing operation
2.	Minimum resale price maintenance by Federal Mogul

Apart from situations where arrangements are illegal per se under the Act, the evidence required is quite onerous. For example, to prove that a dominant firm is abusing its market power by charging excessive prices requires very detailed economic analysis. In such cases there is also a problem of an appropriate remedy, given that the authorities have expressly stated that they do not see their role as being regulators of price.

The Commission's first major investigation into horizontal restrictive practices of alleged collusive behaviour in the cement industry was subject to a court challenge following a ‘search and seizure’ operation on 2 August 2000. Such operations are allowed under the Act where there are reasonable grounds to believe that a prohibited practice has taken place and where people have information on the premises relevant to the practice. The search also only took place after the Commission had repeatedly requested information. The court challenge was successful on the grounds that the Commission had wrongly invited the press to the operation. While this may appear to be a special case linked to learning by new institutions, firms have been highly successful in delaying cases by taking a succession of legal points.

The crucial role of addressing the anticompetitive practices of large firms is thus very demanding of the institutions. In this area, much of the action by the authorities ultimately rests on judgement and economic evaluation, although the provisions in the South African law are framed in clear‐cut legal terms that allow little discretion. This also sets very high standards for such cases in terms of both analysis and information, and allows considerable scope for parties to delay cases by taking legal points. While the weightiness of the matters suggests the need for high standards, and checks and balances on the institutions, the institutions must be able to take decisive action if they are to have an impact in this important area of economic policy. It is also notable that participation by small and medium firms, which would be subject to the negative effects of anticompetitive practices, has been poor and that such firms have been unwilling to lay complaints against dominant firms on which they will continue to depend.

In two important product groupings, food and pharmaceuticals, the government has now established separate committees to monitor pricing and make recommendations. In the case of food prices, this accompanied a Competition Commission investigation at the behest of the government. In the case of pharmaceutical prices, the formation of the committee did not wait for the resolution of a Commission investigation. In any event, it appears as if the competition authorities and legislative provisions do not in themselves provide sufficient safeguards against the exercise of market power. The question of firm behaviour and appropriate policy tools is now addressed in the case of the steel industry.

4 COMPETITION POLICY AND EFFICIENCY IN CONCENTRATED MARKETS – THE CASE OF THE STEEL INDUSTRY3

The South African steel industry is well developed, with a low cost base due to abundant local inputs such as iron ore, and strong production capabilities reflected in a major export orientation. As in many countries, the state played a major role in the industry's growth but has now largely withdrawn. This withdrawal of support included the liberalisation of protection, with import tariffs being reduced to just 5 per cent by 1996. Partly as a response to this liberalisation there has been a major programme of restructuring and consolidation of South African production, consistent with the large economies of scale in the sector and the lumpy capital‐intensive nature of investment. The main producer, Iscor, accounts for more than 70 per cent of the local market for carbon steel. Restructuring has also brought efficiencies and foreign ownership. The world's second largest steel producer, LNM, had acquired 47 per cent of Iscor by 2003.

Steel is also an important input product for many downstream industries. For example, it accounts for 38 per cent of inputs to fabricated metal products (StatsSA, 1998). It is thus an interesting case in which to examine the government's industrial policy emphasis on further beneficiation, given the high levels of exports of basic iron and steel compared with the poorer trade performance of downstream, more labour‐intensive sectors.

As will be discussed, changes in the structure of the sector were explicitly approved by the competition authorities because of their efficiency‐enhancing effects. One would, however, expect a profit‐maximising firm reaping such efficiencies (and with its market position strengthened by the efficiencies) also to exert its market power where possible. The provisions of the Competition Act addressing prohibited practices guard against this. The sector therefore provides an ideal case in which to evaluate the impact of competition policy and the institutions charged with implementing the legislative provisions.

4.1 The changing structure of the South African steel industry

While the nature of the steel industry meant that it has always been highly concentrated, this has increased in the past decade in production, distribution and basic processing. This includes the acquisition of Saldanha by Iscor approved in 2002, the acquisition of the operations of Baldwins Steel by Trident Steel in 2000, and the merging of the third and fourth largest steel merchants, Baldwins/Kulungile and Abkins Steel. It also includes the formation in the mid‐1990s of MacSteel International BV as a joint venture by Iscor and MacSteel, South Africa's largest steel merchant. All four of these transactions were investigated and ruled on by the competition authorities.

The acquisition of the operations of Baldwins by Trident Steel was approved by the Competition Tribunal despite it leading to one firm having a 70 per cent share of the local market for improved surface‐finish outer body panels for the auto sector, with the remaining 30 per cent accounted for by imports. As mentioned, the approval was granted on the basis of the order of magnitude of the dynamic efficiencies to be reaped from the combined operations. It clearly entrenched economic efficiency as the primary criterion employed in analyses of cases rather than the promotion or protection of competition for its own sake.

The acquisition of Saldanha Steel by Iscor was also approved despite increasing concentration in an already concentrated market, based on the efficiencies to be reaped. In addition, in this case the Tribunal took cognisance of the fact that Saldanha was a ‘failing firm’ with a history of major losses since its inception. However, the Tribunal required the restriction that had been placed on Duferco (which rolls the steel coil from Saldanha) from selling in the local market to be removed.

In both of these cases the Tribunal appears to have been vindicated by subsequent developments. Trident Steel has invested in expanded operations in the production of auto steel, while Iscor quickly turned Saldanha into a profitable business. The merger of the third and fourth largest steel merchants was approved on the grounds that it would ensure a more effective competitor in the South African market, given economies of scale and scope. This is a very recent deal and too early to evaluate.

The joint venture between Iscor and MacSteel to provide for the exclusive right to export Iscor steel was approved by the Competition Board in 1995 for an initial two‐year period and subject to conditions preventing discrimination between customers, except as might be justified by different volumes purchased. As will be discussed below, the excess of local steel production over local demand means that separating exports from steel supplied to local buyers is crucial to being able to charge a price in the local market that is above the export price. The exclusive arrangement has ensured that separation for Iscor.

The analysis of the steel industry contained in the rulings by the competition authorities indicates that they believed that they had an important role to play in guarding against anticompetitive behaviour in the steel industry, given its concentrated nature. This is consistent with literature reviewed earlier, which emphasises the importance of competition policy in liberalised economies.

4.2 Pricing and behaviour in the steel industry

Iscor has clearly stated that it practises import parity pricing in the local market for flat steel products, although it now terms this ‘international price parity’. This means pricing up to the equivalent cost to a local buyer of importing despite South African production being far in excess of local demand. The price is calculated by taking the price in international markets and adding on the costs of shipping, import duties and related port charges (IDC, 2000). Currency changes necessitate being able to adjust the price on a month‐by‐month basis. Thus, an ‘import parity price discount’ is used to lower the list price, which is changed every three to four months, to meet the import parity price. There are also significant non‐price deterrents to importing in the form of the time lag from order to delivery and potential problems with recourse if there are quality or delivery problems. These deterrents, in fact, allow an ‘import parity plus’ basis for pricing where only the very largest buyers, which are readily able to import, pay the actual import parity level.

As a result, the local price is a significant margin above the international price – typically around 30 per cent. For the approximately 40 per cent of Iscor production that is exported, the price received is the international price less the shipping and related costs. This implies that the price paid by local purchasers is an even greater margin above the price received for exported steel. Clearly there is scope for arbitrage as an export customer of Iscor's (e.g. in Europe) paying the net export price would have a strong incentive to resell the steel into the South African market if it was able to do so without incurring the costs of shipping the product to Europe and back to South Africa. The arrangement with MacSteel International BV effectively ensures that this is not possible, as exported steel can be segmented from the local market.

The practice of import parity pricing might be defended as necessary for maximising returns from the local market in order to be able to support the operations as a whole, given the much lower prices in the export market. It is also interesting that import parity price discounts are not used in the long products where there are more domestic producers.

While consistent with profit maximisation, import parity pricing is inefficient in terms of both allocative efficiency and dynamic efficiency considerations. It may also not be in the long‐term interests of the local steel industry. The mark‐up effectively means that there is no benefit passed on to local downstream producers from low steel production costs in South Africa. (South Africa has among the lowest steel production costs in the world.) This suggests underconsumption of steel and, as an intermediate product, the higher local steel price effectively means that downstream industries are underdeveloped.

There is recent concrete evidence of pricing negatively affecting downstream competitiveness. According to metal product firms, prices to local buyers of flat steel products increased by more than 40 per cent in 2002, following the depreciation of the rand. These price increases were not therefore related to changes in costs of production, most of which are rand based. This would also have been the case if prices were based on export parity levels, but the latter levels would reflect the opportunity cost of supplying the local market, which is the revenue lost from reduced exports.

The sharp price increases have led to downstream firms being uncompetitive against imports and in export markets. Despite export rebates (paid for steel embodied in exported end products), major exporting firms in products such as mining equipment have seen export orders lost to international competitors in countries (e.g. Germany) with much higher production costs but cheaper steel. The pattern of exporting unbeneficiated and capital‐intensive steel while being uncompetitive in beneficiated and more labour‐intensive products is thus being reinforced. This is despite the relative factor costs in South Africa favouring downstream production.

Following the recovery of the exchange rate, prices have been slow to come down, demonstrating considerable ‘stickiness’. Price reductions in effect require increases in the import parity price discount, which appear to have occurred only when there have been actual switches to imports. For example, the import parity price discount on galvanised steel used in roofing was reportedly increased to R2 190 per tonne in July 2003. This is a substantial discount from a price of approximately R7 500 per tonne for 0,35 mm galvanised steel sheet.

In a dynamic context, the demand in the ‘high return’ local market depends in turn on the performance of the steel‐using industries, which in part depends on their competitiveness – both against imports and in the export market. An imported metal product means a loss of South African manufacturing production and lower local demand for South African steel. Pricing for short‐term profit maximisation therefore conflicts with conditions for long‐term growth of the industry.

The ‘Williamson trade‐off’ suggests the weighing‐up of cost efficiencies achieved by Iscor, encouraged by profit seeking, against the deadweight losses from the exertion of market power. The point made here is, however, that the behaviour of a dominant firm has extensive knock‐on effects on downstream firms beyond static deadweight losses. Indeed, recent patterns could even be characterised as deindustrialisation.

Given the greater labour intensity of downstream manufacturing, the question is a much broader one, namely whether competition policy is able to address the behaviour of large firms. Can it weigh up the different considerations necessary, given the issues flagged by the authorities themselves in the judgements on the structural changes? To date, this is doubtful and there are even indications that the behavioural conditions imposed by the authorities are not being heeded, for example, in the case of the Saldanha merger. Experiences in the steel industry show that South Africa needs to address the conduct of large firms, and that the overwhelming focus on merger analysis should be of concern.

5 CONCLUSIONS

High levels of concentration mean that there are great analytical demands on competition institutions in relatively small developing countries such as South Africa, even though their capabilities in terms of information gathering and analysis are weaker than those of their industrialised counterparts.

Previous legislation and the operations of the Competition Board were criticised mainly for their weakness in addressing the behaviour of large firms. However, as noted in a recent peer review of South Africa's competition policy regime, ‘to a surprising extent, competition policy in South Africa is merger policy’ (Wise, 2003). The review further notes that this is understandable in the context of South Africa's relatively high level of development and that the ‘focus on mergers carries forward the theme of correcting excessive concentration of economic power.’ This appraisal fails to recognise that the existing high levels of concentration in the South African economy will not be affected by merger assessment.

Economic theory and the experience of developing countries suggest that the focus should be on the behaviour of firms and, in particular, the degree of rivalry. This means that competition authorities must be able to undertake economic evaluations and have the ability to make and enforce rulings based on the analysis of different effects that cannot be perfectly quantified. The South African regime, however, emphasises checks on discretionary action by the authorities within a rules‐based regime. This reflects a preoccupation with international ‘best practice’ that tends to ignore the experiences of industrialising countries and the diversity in competition regimes in developed economies.

Developments in the South African steel sector illustrate the orientation towards merger review. The approval of mergers due to efficiency gains has not been accompanied by effective measures on conduct. Conditions placed on mergers are also not necessarily being complied with, and price setting by the dominant steel firm utilising its market power appears to have significant negative effects on buyers and downstream industries.

The answer to the relative lack of progress in South Africa in addressing the behaviour of firms with market power may be to continue to develop the institutional capacity of the competition authorities. Weaknesses, however, also reflect intrinsic features such as the asymmetry of information between the authorities and firms. In addition, for firms under investigation it is worth their while to pay highly to ensure they acquire the best expertise, at rates that public institutions cannot match. The more successful the institutions, the greater are these effects as more is at stake. Instead, it may be worth considering provisions that increase the institutions' powers in addressing firm behaviour, as in some other industrialising countries. For example, by ‘designating’ firms that are overwhelmingly dominant or where industries are particularly sensitive, as in South Korea, information on firms' behaviour could be required in order to enable the institutions to better monitor for possible abuses. The establishment by the South African government of committees to monitor prices of food and pharmaceuticals is an ad hoc response to such a need and is not necessarily the most effective one.

Finally, a focus on improving economic evaluation of firm behaviour by the competition authorities, as well as on the different measures that could be employed to impact on the nature of interfirm rivalry and behaviour, suggests the need to develop links between the competition authorities and other public institutions that have strong information‐gathering and analytical capabilities. There may have been an overemphasis on the separation of the competition institutions for fear of their independence being compromised. However, without improved abilities to collect and interpret information, and develop applicable remedies, they will remain relatively ineffective in addressing the effects of existing concentrations.

Additional information

Notes on contributors

Simon Roberts

Notes

Note, by using market capitalisation we are including the value of non‐South African operations. This overstates the significance in the South African economy of corporations with extensive international operations.

The author has been engaged to work on several competition cases, including a case still under way in the steel sector.