---

COMPETITION TRIBUNAL

REPUBLIC OF SOUTH AFRICA

Case No.: 89/LM/Oct00

In the large merger between

Trident Steel (Proprietary) Limited (“Trident Steel”)

and

Dorbyl Limited (“Dorbyl”) for the acquisition of three operations of Baldwins Steel, a division of Dorbyl Limited

_______________________________________________________________________

Reasons for the Competition Tribunal’s Decision

________________________________________________________________________

Approval

The Competition Tribunal issued a Merger Clearance Certificate on 6 December 2000 approving the merger between Trident and three of Dorbyl’s subsidiaries, namely the Baldwins steel processing plants in Rosslyn, Durban and Port Elizabeth without conditions. The reasons for our decision are set out below.

The Merger Transaction

1. Trident Steel (Pty) Ltd, the acquiring firm is a subsidiary of Tristel Holdings (Pty) Limited, which is ultimately controlled by Aveng Limited, a large conglomerate with interests in the construction and engineering sector having substantial steel interests.

2. Trident Steel (Pty) Ltd is acquiring three plants from Dorbyl Limited, forming part of its Baldwins Steel division, as a going concern. More specifically, Trident Steel will purchase Baldwins’ three flat steel decoiling and cut-to-length service centres situate at Rosslyn, Durban and Port Elizabeth.

PART I: DOES THE MERGER LESSEN OR PREVENT COMPETITION?

The relevant products/services market

3. The two companies are steel merchants engaged in the processing of and supply of steel products to the automotive and non-automotive industries. The overlapping products are cut flat steel products which can be of varying quality, described, inter alia, as Improved Surface Finish (ISF) or non-ISF.

4. The two relevant markets are ISF and non-ISF steel products. Non-ISF products are utilised in both the automotive and non-automotive industries. Non-ISF steel products are of a normal specified quality finish and are often used within the automotive industry for the inner panels, as well as for the unexposed parts of motor vehicles. ISF quality steel is a high quality steel specific to the automotive industry. Motor car manufacturers require this type of steel specifically for the outer body panels of motor vehicles.

5. The parties’ business activities in respect of flat steel products involve the purchase of raw steel coils from South African and non-South African steel suppliers and the decoiling and cutting of the steel for purposes of producing flat steel products comprising steel sheets and blanks. These products are then sold to either car manufacturers or press shops.¹

6. Customer demand for superior high-grade surface quality steel material for the outer panels of motor vehicles (“outer blanks”), as opposed to normal quality steel, differentiate ISF outer blanks from the non-ISF category. Although non-ISF products can be produced on the same production line as ISF products the converse is not true. ISF products require the use of sophisticated and specialized machinery.

7. The Competition Commission and the parties are both in agreement that the relevant product markets can be differentiated as the markets for ISF and non-ISF products. We agree with this categorization of the relevant product markets. We will now proceed to examine each of these markets in turn.

IMPACT OF COMPETITION IN THE NON-ISF MARKET

8. Baldwin’s Steel is selling only its Rosslyn, Durban and Port Elizabeth plants. However, the remaining Baldwins plants, at Isando and Vanderbijlpark will continue to supply non-ISF materials to the motor manufacturing industry.

9. Baldwin’s currently has substantial spare capacity available at its Isando and Vanderbijlpark plants with respect to the processing of non-ISF steel products. Baldwins’ estimates that these two operations are utilizing only 40% of their capacity. Baldwins Steel plans to use this spare capacity in respect of these operations more efficiently post-merger. Intensifying the productive capacity of these remaining plants in respect of the non-ISF market will undoubtedly reduce its production costs.

10. The parties have accordingly argued that their ability to supply non-ISF products to the market will not be affected or reduced in any way, since these plants were previously not utilized to optimum capacity even though Trident is acquiring almost 50% of Baldwin’s capacity for cutting non-ISF blanks. Therefore, they contend that Baldwins remains in the same position to supply the market with non-ISF product as it was pre-merger.

11. The two remaining Baldwins plants will therefore continue to compete with Trident and other, smaller competitors in the non-ISF product category, as well as Macsteel, whose market share in the non-ISF steel blank sector is currently 15%. Although the parties suggest that Macsteel’s market share is likely to increase post-merger there is no evidence of why this should be so, but nevertheless it is likely that Macsteel will at least retain its 15% market share.

12. We are therefore satisfied that the proposed transaction does not alter the existing competitive situation in the non-ISF market.

IMPACT ON COMPETITION IN THE ISF MARKET

13. Baldwins intends to exit the market for ISF blanks by selling its Rosslyn plant to Trident, drastically reducing the level of competition in this market.

14. The Commission submitted that Baldwins and Trident each currently enjoy a 35% market share of the ISF market. Therefore, post-merger, the combined market share of the merged firm for producing ISF outer blanks for the domestic market will be 70%.

15. There are no other domestic competitors in this market currently, since the costs to establish a new operation with the sophisticated technology required to process ISF blanks, are prohibitive. The market for ISF products is small and already over-supplied, and one requiring huge capital investments, therefore unattractive to potential new entrants. Macsteel, could potentially enter the market for ISF products with lower set-up costs than a completely new entrant. However the Commission established in interviews with Macsteel, that they have never entered the ISF market and are unlikely to, despite the fact that they have the plant necessary to do so. ²

16. The Commission argues that the combined entity’s only other competitive threat comes from imports, which approximate 30% of the South African market (according to sales figures acquired from the National Association of Automobile Manufacturers of South Africa, NAAMSA) on the parties’ conservative estimate. The parties were not able to furnish direct statistics evidencing the market shares of overseas suppliers.

Will imports constrain the pricing behavior of the merged firm?

17. Once we have established that the only competitive restraint on the merged firm is from import competition we have to establish to what extent imports can credibly restrain it from exercising market power.

18. The Commission and the parties contended that this is an international market insofar as the customers, the car manufacturers, can rely on imports, whether from their parent companies abroad or other international suppliers.

19. However we have reason to doubt whether this argument is, in fact, valid for the following reasons-

1. The competing import is not a classic substitute for the ISF product.

2. Customer preference indicates that pricing issues are not determinative in their choice between domestic and foreign supply.

3. There is considerable skepticism about the potential for foreign competition to constrain domestic producers.

4. The barriers to entry created by tariff and incentive schemes undermine the competitive ability of foreign competition.

The nature of the competing product

20. The parties told us that imports come in three forms: fully-assembled or complete-built up (CBU), semi-knock down (SKD) or complete-knock down (CKD) form. While SKD form refers to the importation of only certain body panels of a vehicle, for example, the doors and the bonnet, CKD products come in packs, comprising all those components required for the assembly of a motor vehicle.

21. There is not enough evidence to support the contention that these products are directly substitutable for ISF steel products. The import market share information is ambiguous, at best. The parties’ ISF import market share estimate of 30% comprised fully assembled (or CBU) and complete (CKD) imports.³ Furthermore this figure is not composed purely of ISF product. According to the parties, only 60% of this figure comprised ISF product. They stressed that this was an estimate as better data was not available to them.

Customer Preferences

22. Though in this case there may be limited argument for some partial substitutability or interchangeability between ISF blanks and SKD or CKD packs, factors of inconvenience expressed by the customers, suggest that while these imports are an alternative, they do not necessarily regard them as an effective, competitive substitute for ISF outer panels. CBU products must immediately be disregarded since many of the car manufacturers agreed that car bodies imported in fully assembled form were not substitutable for the ISF steel panels (or blanks) produced by the parties. Many of the customers interviewed by the Commission reported that their decision to turn to imports might be motivated by reasons other than local price increases. Some customers have expressed the view that engineering, or unavailability of local supply, have frequently determined their decision to import from overseas. Delta stated that:

“ The local sheet metal component used in the third vehicle line has often been due to local engineering requirements and not necessarily favourable economics.”

Similarly, Nissan and Toyota reported that steel prices were not a significant factor in their decisions to import.

23. Some manufacturers and press shops expressed the view that products in this form could not be sourced as readily, speedily or as reliably, as local products could. Nissan expressed the view that it would “anticipate difficulty in obtaining commitment for continuity of supply from overseas sourcing, due to low call off requirements”. Baldwins itself admitted that on previous occasions, it has had to air freight material from Europe at great cost to meet delivery times on material committed.

24. Furthermore, South African export incentives create a preference on behalf of some manufacturers for sourcing local steel directly from Iscor. In other words, some manufacturers will lose out on an export credit incentive by virtue of the fact that they favour imports over utilising local Iscor steel. VW SA expressed the view that excessive price increases would impact their export rebates on exported steel parts. “VW SA could import their total requirements but this would have a major impact on future and existing export business”.⁴ Cost is a concern. Daimler Chrysler SA regards the import duties, logistics costs and cost of setting up infrastructure to support imports as additional costs. BMW’s press shop expressed the view that although total requirements could be sourced overseas, this would be costly and could result in the closure of local plants as it would in turn have to pass this cost onto BMW. Furthermore, they felt that “unrealistic” price increases by Trident post-merger would force press shops out of business.

25. On balance, the customers’ collective testimony suggests that they would not switch to importing fully assembled or CKD component packs as readily as the parties would have us believe. Cost is a factor, however other factors such as export rebates, engineering requirements, additional logistics costs associated with transport and warehousing and interruption of supply are also significant. In general, customer responses do not indicate that imports are worthy of much weight as an inhibitor of post-merger market power. They would not readily turn to such imports as an alternate source of supply in the event of a price increase.

26. We accordingly do not regard the overseas imported products (CBU products, SKD and CKD packs), though they may be an indirect form of substitution, as direct substitutes. Substitution in one of these forms will not necessarily constrain Trident’s market power or prevent Trident from raising prices on ISF flat steel products up to the import parity ceiling.

The potential for foreign competition to constrain

27. The mere presence of imports does not necessarily indicate conclusively that a market is international. Some writers and indeed other competition authorities take the approach that a market may be a national one punctuated with sporadic sources of supplies from overseas. This however would not necessarily warrant delineation as an international market.

28. When we talk of the relevant geographic market we refer to the area to which customers can “reasonably turn for sources of supply”.⁵ How do we know if a market is a national one with import competition or an international one?

29. Areeda has explained the distinction in the following way:

“When only actual imports are to be counted, courts say that the market is nationwide and includes all sales there. When the total output of foreign firms is to be counted, the market is said to be worldwide, or, alternatively, that it covers the US plus one more foreign region shipping to the US…”⁶

30. The approach to defining markets taken in the Australian Merger Guidelines is to define a market narrowly, but investigate the competitive role of imports with circumspection. One of the factors they consider is the extent to which imports are closely substitutable for the products of the merging firms from the perspective of their customers, without the need for supply substitution by the overseas producers.⁷

31. There is also some scepticism about whether import figures are a reliable indicator of competitive force. Competition academics have cautioned against relying too heavily on import market share data. Some writers have dispelled the relevance of import competition as being a reliable restraint on the market power of domestic firms. In the opinion of Porter and Sakabira who have researched the issue:

“…, our results are strongly suggestive of a view of competition as a dynamic process in which rivalry among locally based producers drives firms to constantly improve, in a way not substituted for by the presence of imports”. ⁸

These writers found that measures of import pressure do not necessarily account for local competition or domestic rivalry, by finding an insignificant correlation between such measures of import pressure and market share instability.⁹

32. It is recognized that trade barriers cannot be ignored in evaluating the domestic market. This approach has been enunciated in academic writing:

“ These factors often limit the ability of imports to restrain the exercise of power by a domestic firm. Thus, although an analysis of the reasons imports are entering the domestic market may lead to the conclusion that they will continue to restrain market power effectively, the existence of trade barriers and other trade-related costs may negate this effect.”¹⁰

It would therefore seem that where there are restrictions on entry of imports into a market, either by tariffs, quotas or anti-dumping laws, with the practical effect of reducing accessibility of imports into that market, the potential for foreign output to come in is unequivocally restrained, and the market must be defined more narrowly.¹¹ We now consider some of those restrictions as they affect this market.

Import Tariffs & Logistics Costs

33. The tariff on imports currently ranges from 32.5% - 43% depending on their form.¹² The parties’ understated these figures, suggesting an average of 5%. Tariffs are lower for CKD or SKD products, and gradually increase as the imports become more built-up. There are also transport and warehousing costs associated with importing overseas products. Though the differentials between the different type of imports has not been canvassed by the parties, again, one would presume that transport and warehousing costs would increase depending on the extent to which the imports are built-up.¹³

Export Incentive Scheme

34. The Export Incentive Scheme referred to earlier will also act as a disincentive to import for those manufacturers who qualify.

Currency Fluctuations

35. Exchange rate fluctuations can also influence customer demand for imports as substitutes. Even were the Tribunal to accept that the imports were an effective substitute for ISF steel blanks, in order to effectively curtail competition, imports must be competitively priced. ¹⁴

36. To illustrate the competitiveness of imports, the parties quote the per unit cost of the side frame for the BMW E46 (new 3 series) obtainable from Iscor as being substantially less than importing the equivalent product from Germany. However, these figures were calculated using the euro rate in existence in mid-October (i.e. 6.1 Rand per Euro). If one conducts the same calculation on current euro rates of 6.79 Rand per Euro,(the rate at the date of hearing in December 2000) imports become more expensive, indicating that exchange rate volatility can influence the import figure, making it variable, at best, and reducing the likelihood of their acting as a competitive restraint on domestic prices. ¹⁵

The following example is based on the parties’ cost figures quoted per unit for a BMW E 46 side frame imported from Germany and the cost of a local side frame sourced locally from Iscor:

IMPORT PRICE

(converted from Euros to Rands at current exchange rates of 6.79)

Present Imported Price (29 euros@6.79) R196.91

Import Duty @ 5% R 9.85

Total R206.75

DOMESTIC PRICE

Local Price¹⁶ R178.36

Iscor’s increases as of

Jan 2001 @ 10% R 17.84

Total R 196.20

This example indicates that imports are clearly not competitively priced on present exchange rates.¹⁷

37. Accordingly with respect to the ISF market under consideration, there is clearly a case for limitation of imports to actual imports. The parties have not documented that the imports are a regular, steady supply. Instead the ambivalent responses indicate that imports merely serve to overcome shortages in domestic supply on an ad hoc basis from the manufacturers’ parent companies overseas. While we must not exclude from our consideration the possibility of future potential sources of overseas supply as the demand for superior quality steel increases, we are compelled to “take the market as we find it.” ¹⁸ We could not possibly include in our assessment total foreign output simply because tariffs, logistics costs, customer preferences, product differentiation and exchange rate fluctuations all militate against this being regarded as a truly international market, thereby effectively excluding these other imports from the analysis.

38. We accordingly find that the market for processed flat steel products (outer blanks) is a national one, subject to some import competition.

Countervailing Power

39. The parties and the Commission made much of the countervailing power of the customers (the car manufacturers) and the suppliers (notably Iscor). They argue that since the customers are large multinational entities, with extensive overseas networks and resources, they have the ability to constrain any attempt by the merged entity to elevate prices. Similarly, Iscor’s power to impose supply side constraints would provide significant restraints on exercise of a monopoly power by the parties. Notwithstanding these arguments, the fact remains that countervailing power in a post-merger market-place where there will only be one domestic supplier of ISF outer blanks to the automotive industry, is unpersuasive. In reality, post-merger there will be no other domestic steel processor that will constrain the merged entity from pricing up to at least import parity.

40. Accordingly, considering that Trident will be the only domestic steel processor of outer steel blanks for the automotive industry post-merger, as well as the unreliability of imports as a likely competitive force, we find that the merger will result in a substantial lessening of competition in the ISF market.

PART 2: EFFICIENCY GAINS

41. Having found that the merger is likely to substantially prevent or lessen competition, Section 16(1)(a)(i) of the Act requires that we must next determine whether the merger is likely to result in any technological, efficiency or other pro-competitive gain¹⁹ which

“will be greater than, and offset, the effects of any prevention or lessening of competition, that may result or is likely to result from the merger and that would not likely be obtained if the merger is prevented...”

THE LEGAL ISSUES

42. The section was based on section 96 of the Canadian Act.²⁰ The Canadian Act itself seems to have been inspired by a trend in economic literature since the late 1960’s that recognized that a merger can both lessen competition and create efficiencies and that a proper enforcement policy should seek to maximize overall efficiency in the economy. This approach owes its origins to a series of articles written by the distinguished US economist Oliver Williamson who developed a hypothesis known in the literature as the “Williamson trade off”. The Williamson analysis is only relevant when the merger creates both market power and economies. Williamson argued that cost efficiencies would be far greater than social losses resulting from increased economic power. He demonstrated that a relatively small cost reduction would offset a relatively large price increase thereby making society indifferent to the merger.²¹

43. The Williamson model was attractive to many economists for its elegance and simplicity. The problem for antitrust enforcers was how to translate its framework into policy. In practice getting the data to satisfy the theoretical model is far more daunting . Even critics from the Chicago school such as Posner have proved skeptical:

“Not only is the measurement of efficiency … an intractable subject for litigation; but an estimate of a challenged merger’s cost savings could not be utilized in determining the total economic effect of the merger unless an estimate was also made of the monopoly costs of the merger – and we simply do not know enough about the effect of marginal increases in the concentration ratio…to predict the price effects.”²²

Similarly Fisher and Lande argue that:

“As a result of the complexities of a generalized Williamson tradeoff the ideal of a case by case balancing of efficiencies and market power effects becomes too unmanageable to be of any practical value , despite its initial appeal as a theoretical paradigm.”²³

44. Even Canadian commentators are skeptical about whether their section 96 has worked in practice. According to McFetridge section 96 has had little effect on merger enforcement in Canada.²⁴

45. American writers unlike their courts seem for the most part to recognize that efficiencies should be treated as a defence²⁵ although they differ on the extent to which they feel the defence should be recognised.

46. In Europe an efficiency defence has not been recognized in any decision to date nor is it clear that the Merger Regulations allow for it.²⁶ In some cases where efficiency issues are considered it would appear that the Commission views efficiencies as an offence rather than a defence as efficiencies might strengthen on this analysis a dominant position.²⁷

47. Neven et al in criticizing the Commission’s approach concede that in those situations there is a trade off but argue that this should not mean that these types of mergers should be condemned. They state that:

“Carried to its logical conclusion, such an argument would imply that, if only privately profitable mergers are proposed none should be allowed.” ²⁸

48. The Canadians incorporated “trade off” analysis into their statute through section 96, boldly treading where other authorities were still too reticent to go.²⁹ Our section 16(1)(a)(i) formulation as we indicated above, followed the Canadian lead. It remains for us then not to debate the desirability of such a test in the statute but how it should be interpreted.

49. The application of this provision of the Act raises several issues.(1) On who does the onus of establishing the efficiency gain rest? (2) What type of gains are acceptable (3) How is the offset or trade-off between the competitive loss and the efficiency gain calibrated? (4) Does the gain need to be passed on to the consumer? (5) Would the efficiency be obtained without the merger or put in another way is the efficiency merger-specific?

50. We will proceed to examine each one of these in turn.

The onus

51. We have previously held that the onus of establishing the efficiency defence rests on the merging parties³⁰. This approach is consistent with the approach taken in Canada and the United States.³¹ The significance of the Canadian authority is that as we pointed out earlier our section is closely modeled on theirs. As the OECD³² has explained the rationale for this approach is the fact that in a pre-merger notification system mergers must be evaluated before they can be implemented. It goes without saying that the task of identifying and quantifying claimed post- merger efficiencies at the pre-merger stage is difficult. Due to asymmetries in information it is the parties to the merger and not the competition authorities that are best placed to provide this information.

What types of efficiency gains are acceptable?

52. Every merger brings about some form of efficiency gain even if it is trivial. Did the legislature intend that each claimed cent in cost savings be factored into the trade –off of lost competition? We would suggest not and that what the legislature contemplated was either something more significant or enduring. In the United States courts have historically been extremely sceptical about efficiency claims. In United States v Philadelphia National Bank³³ the Supreme Court held:

“We are clear… that a merger the effect of which may be substantially to lessen competition is not saved because on some ultimate reckoning of social or economic debits or credits, it may be deemed beneficial”

53. In FTC v Proctor And Gamble³⁴ Justice Harlan in a concurring judgment wrote:

“Economies cannot be premised solely on dollar figures, lest accounting controversies dominate proceedings. Economies employed in defence of a merger must be shown in what economists label ‘real’ terms.”

54. Fisher and Lande’s interpretation of this is that the Court is saying Congress did not signal an intention to ignore economic values rather we ( the Court ) recognize our own limited ability to balance market power and efficiency effects.

55. It seems that the types of efficiencies that will be recognized are by no means clear-cut. Part of this is due to the paradox created between the desirability and the measurability of a claimed efficiency. As Robert Pitofsky has so trenchantly observed claims of efficiency are “easy to assert and sometimes difficult to disprove.”³⁵ The most beneficial efficiencies are those associated with innovation³⁶ or as they are otherwise known, “dynamic efficiencies”, because these are efficiencies to product or service quality - precisely those benefits competition seeks to induce.³⁷ Kattan argues that innovation has the quality of a public good in that its use by one party does not exclude others from using it simultaneously. He argues that despite protection afforded by intellectual property many innovations are imitated within a short time of their introduction.³⁸ Yet these efficiencies are also the hardest to quantify in practice.³⁹ At the other end of the scale are so called pecuniary efficiencies e.g. tax savings or lower input costs resulting from improved bargaining power with suppliers. These may be the easiest to “put a number” to, but are not considered real savings in resources and are less favored.⁴⁰ Production efficiencies are somewhere along the continuum between innovation and pecuniary efficiencies. Production efficiencies are those efficiencies that permit firms to produce more output or better quality output from the same amount of input.⁴¹

56. Production efficiencies can themselves be further classified into various types including plant level economies, distribution, procurement and capital cost economies, research and development. Not all merit equal recognition as part of an efficiency defence. Areeda treats plant size and plant specialization economies as those most worthy of recognition but is more sceptical about claims for others frequently raised which he describes as “ordinary efficiencies” e.g. distribution, procurement and overhead economies.⁴²

57. In Canada in the Hillsdown , the first case to deal with the efficiency defence in any detail, the Tribunal did not make any finding as to the types of efficiency that it would consider acceptable.⁴³

58. In its most recent judgment and indeed the only case thus far where the Tribunal has accepted an efficiency defence under section 96 the Tribunal in Commissioner of Competition v Superior Propane Inc and ICG Propane Inc (“Superior”⁴⁴) was far more solicitous about accepting efficiency claims but did not establish a set of criteria for determining which are worthy of recognition and which were not. This is perhaps because the Tribunal accepted a total surplus standard to its analysis of which we say more below.

59. The U.S. Merger Guidelines⁴⁵ provide some indications on their preference:

“The Agency has found that certain types of efficiencies are more likely to be cognizable and substantial than others. For example, efficiencies resulting from shifting production among facilities formerly owned separately, which enable the merging firms to reduce the marginal cost of production, are more likely to be susceptible to verification, merger-specific and substantial, and are less likely to result from anticompetitive reductions in output. Other efficiencies, such as those relating to research and development are potentially substantial but are generally less susceptible to verification and may be the result of anticompetitive output reductions. Yet others, such as those relating to procurement, management or capital costs are less likely to be merger-specific or substantial, or may not be cognizable for other reasons.”

60. There is also debate about whether only reductions in marginal cost should be included as efficiencies as fixed costs reductions have no effect on current price. Others argue that in the long run the cost of replacement is a marginal cost and should be recognised. ⁴⁶

61. In the Tongaat- Hulett case we gave the following examples of efficiency gains contemplated by the Act:

“One that for example evidences new products or processes that will flow from the merger of the two companies, or that identifies new markets that will be penetrated in consequence of the merger, markets that neither firm on their own would have been capable of entering, or that significantly enhances the intensity with which productive capacity is utilised. ”

62. We pointed out in Tongaat Hulett that these were not intended to be an exhaustive list . We would similarly be reluctant to propose a list in this decision although we come to a more tentative conclusion below in our conclusion.

Measuring the Trade off

63. Section 16, as we indicated earlier, requires that the efficiency gains must be “greater than” and “offset” the anticompetitive effects. This presupposes a weighing process, which suggests that the efficiencies must be capable of measurement, as opposed to broad speculative assertions. To give meaning to the efficiency assessment we need a way to verify the efficiency gains asserted and then establish how they “trade-off” against the loss to competition. Verification itself is conceptually difficult. First one must assess efficiencies quantitatively -then the likelihood they will occur. This is the approach taken in the US merger guidelines:

“Therefore the merging firms must substantiate efficiency claims so that the Agency can verify by reasonable means the likelihood and magnitude of each asserted efficiency, how and when each would be achieved, (and any costs of doing so), how each would enhance the merged firm’s ability and incentive to compete, and why each would be merger specific. Efficiency claims will not be considered if they are vague and speculative or otherwise cannot be verified by reasonable means”.⁴⁷

64. The Canadians take a similar approach:

“In general, parties should provide a reasonable and objectively verifiable explanation of why efficiencies that are available would not likely be sought by alternative means if the order were made.”⁴⁸.

65. Verification is not the only hurdle one has to cross in offset analysis. The assessment of the trade off is even more formidable. The case law and the literature suggest that two approaches can be followed; a formulaic approach such as that favored in the Superior case and a discretionary approach such as the US Merger Guidelines. The formulaic leads one to approach the problem as an economist would do in a classroom demonstrating Williamson’s trade off. Efficiencies claimed and deadweight losses are calculated in terms of a formula and then compared.⁴⁹ If the efficiency as calculated exceeds the deadweight loss the trade off requirement has been satisfied. One can see immediately why some find this approach attractive. Once the numbers have been verified the outcome is definitive. The problem with the formulaic approach is that the losses and gains are not always susceptible to measurement by the same units and on the same scale. The one may be quantitative and measurable in units such as rands, the other may be qualitative and defy easy calibration. How does one balance a loss associated with a possible 15% price increase with the gains associated with an innovation in product performance? Another problem with adopting measuring only deadweight loss is that market power effects may lead to price increases by other firms and thus the deadweight loss may be understated.⁵⁰

66. When adopting the flexible approach the competition adjudicator relies on its discretion rather than an equation. But the adjudicator can’t begin exercising its discretion unless it has formulated a policy approach to guide it in its evaluation. The danger with this approach is that it can lead to uncertainty – how will parties know in advance whether claims of efficiency will be accepted? Nevertheless we would not see these two approaches as mutually exclusive and a flexible approach that recognizes and weighs the evidence of a formulaic result has merit.

67. Sanderson is reassuring on this point:

“Indeed it is important not to view the tradeoff analysis as an exact science, even where quantitative estimates are available. Discretion has been exercised at various points in time, particularly when assigning probability weights to cost savings and when quantifying anticompetitive effects. The aim of the exercise is to compare two orders of magnitude – efficiencies versus anticompetitive effects – and not to make a decision based on the fact that n+1 n. Furthermore, comparing orders of magnitude generally is feasible”⁵¹

Must the gain be passed on to the consumer?

68. Perhaps the most controversial issue of all is who should benefit from the efficiency claimed? Put in another way, if efficiency defences are to be recognized, is it a requirement that they lead to lower prices for consumers i.e. consumer welfare, or is it sufficient that producers benefit, which means since we have already accepted that the merger is anticompetitive, there will be a wealth transfer from consumers to producers.

69. These are sometimes referred to as the choice between a consumer welfare and a total welfare standard.⁵² Under a consumer welfare standard efficiencies must be passed through to consumers in some proportion. Under a total welfare standard welfare transfers from consumers to producers are regarded as socially neutral – all that is required is that the transaction leads to an increase in the sum of consumer and producer surplus. On this approach the question of whose pockets should benefit is not considered to be of any economic significance since the wealth is not lost to society whether it transforms itself into lower prices for consumers or a greater dividend for the shareholders of producers. This answer is not a settled one in competition law. Neither in the United States nor Canada have Courts definitively answered this question and to the extent they show an inclination so do so they have come to opposing conclusions. In brief, in Canada the Merger Guidelines have adopted a total welfare approach but this approach was questioned by the Tribunal in the Hillsdown case where Justice Reed seemed to opt for a consumer welfare standard based on the legislative history of the Act.⁵³ A consumer welfare approach would ordinarily require a much greater magnitude of efficiencies than the total surplus standard.⁵⁴ This approach was not followed in the Canadian Tribunal’s most recent decision in Superior where the Tribunal held that the total surplus is the correct standard.⁵⁵ Here the Tribunal traded off the efficiency gain against the deadweight loss and coming to the conclusion that the former was the greater, found the efficiency defence had succeeded. The decision was not unanimous however.

70. The United States has a less complicated approach. In part this is due to the fact that the efficiency defence is a common law creation and not written into statute. As such, it is interpreted as a discretionary tool and does not require a trade off analysis bedeviled by statutory interpretation. Those who have addressed the issue refer to this as the “passing on” requirement. Areeda explains this as:

“whether all or at least most of the efficiencies will be reflected in lower customer prices rather than higher owner profits.”

71. In FTC v University Health Inc the Court held that :

“ ..a defendant who seeks to overcome a presumption that a proposed transaction would substantially lessen competition must demonstrate that the intended acquisition would result in significant economies and that these economies would ultimately benefit competition and, hence, consumers.”⁵⁶

72. In FTC v Staples⁵⁷ the Court applied a pass through rate analysis in rejecting an efficiency defence raised by the merging parties. The Court did not consider whether the efficiencies had to be passed through in order to be accepted and appears to have accepted this requirement as a given.

73. The Merger Guidelines as revised in 1997 state:

“To make the requisite determination, the Agency considers whether cognizable efficiencies likely would be sufficient to reverse the merger’s potential to harm consumers in the relevant market, e.g. by preventing price increases in that market”⁵⁸

74. One writer has cynically captured the approach of the United States Courts to efficiencies by observing that:

“Courts have tended to reject efficiency claims on evidentiary grounds in cases in which they found mergers to be anticompetitive and to credit claimed efficiencies when sustaining transactions on competitive grounds.”⁵⁹

75. A further problem is the credibility of claims that efficiency gains will be passed on. What if post-merger it is not implemented? Does one unscramble the merger on those grounds? What if other factors intervened preventing parties even in good faith from effecting the pass through? Requiring a pass on as a prerequisite for establishing the efficiency defence would be subject to the same criticisms that other price control remedies are viz. that it is not appropriate for the regulator to become a price setter.

Are the efficiencies merger-specific?

76. The final requirement of section 16(1)(a) is that it must be shown that the efficiencies “would not likely be obtained if the merger is prevented”. Expressed differently this is a requirement that the efficiencies must be “merger-specific” to be cognizable. If the efficiencies could come about through some other legal arrangement or organizational form that is not a merger, or if one of the firms could achieve a claimed efficiency on its own, the efficiency defence fails.

77. The Canadian Merger Guidelines in its categorization of efficiency gains excludes those claimed efficiency gains that:

“would be likely to be attained if the order that would be required to remedy the anticompetitive effect of the merger were made.”⁶⁰.

Textual analysis

78. Our statute differs from its Canadian counterpart in some important respects. Firstly our concept of efficiency is used in section 16 in combination with the words “technological or other pro-competitive gain”. Adopting an eiusdem generis approach and trying to discern a common meaning between these three words, this would suggest that in this context, efficiencies that equate to “technological gains” i.e. dynamic efficiencies or “pro-competitive gains” i.e. those that constitute real economies, not mere pecuniary gains, are to be favoured.

79. Secondly in the “purpose” clause, which we find in section 2(a) of the Act, efficiency is conceptually linked to notions of a dynamic nature:

“ to promote the efficiency, adaptability and development of the economy” [our emphasis]

This choice of language is, once again, suggestive of notions of dynamic and productive efficiencies.

80. Thirdly the use of employment as a public interest concern in section 16(3)(ii) which must be taken into account in assessing the desirability of the merger suggests that employment reduction should not be recognized as an efficiency in terms of section 16 (1)(a)(i). The legislature can hardly be seen to be giving a defence in one section (16(1)(a)(i)) and taking it away in another (section 16(3)(ii)).

Conclusion

81. This lengthy digression into comparative jurisprudence illustrates the Pandora’s box that the efficiency defence opens; for it admits of no simple solutions, small wonder why some have sought to keep well away from it. Nevertheless we believe that trawling through the literature and the case law despite the eddying currents of controversy that rage through them, some recurring principles emerge which suggest an approach to these issues that is both consistent with our statute and best practice. We propose the following test – where efficiencies constitute “real” efficiencies and there is evidence to verify them of a quantitative or qualitative nature, evidence that the efficiencies will benefit consumers, is less compelling. On the other hand, where efficiencies demonstrate less compelling economies, evidence of a pass through to consumers should be demonstrated and although no threshold for this is suggested, they need to be more than trivial, but neither is it necessary that they are wholly passed on. The test is thus one where real economies and benefit to consumers exist in an inverse relationship. The more compelling the former the less compelling need be the latter. When we talk of real economies we would, without proposing an exhaustive list, include dynamic efficiencies, production efficiencies ranging from plant economies of scope and scale to research and development efficiencies that might not be achieved short of merger. Pecuniary efficiencies would not constitute real economies nor would those that result in a mere redistribution of income from the customers, suppliers or employees to the merged entity. Without categorically rejecting them we would be more sceptical than the Canadian courts in accepting certain efficiencies such as administrative efficiencies since these can be established in most mergers. As our discussion of the textual features of our Act has shown, it could not have been the intention of the legislature that a merger that is anticompetitive could be immunized by a demonstration of savings on clips and clerks.

82. Whilst this approach may be criticized for giving the competition authority too much discretion at the expense of business certainty, the alternative which is to interpret this section as a mathematical comparison of two areas on a Williamson diagram, permits an approach so clinical and rigid that it would reduce the proper exercise of a discretion to a matter of calculus.

THE FACTUAL ISSUES

83. We turn now to applying this analysis to the current facts. The merging parties have identified three efficiencies that they associate with the merger. These are:

1. Plant scale efficiencies and plant use efficiencies

2. Supply production efficiencies

3. Volume discounts

We examine each one in turn.

Plant efficiencies

84. Baldwins currently manufactures ISF material at its Rosslyn plant. Motorcar manufacturers are continually setting higher specifications for the finishes to their vehicles, BMW being a prime example. This has placed pressure on Baldwins and Trident who must deliver ISF product that meets these more exacting standards. Unfortunately the steel supplied by ISCOR is not of the required quality and hence it requires better quality plant to clean it up so it meets the standard. Lack of sophisticated, state-of-the-art equipment has meant Baldwins has been forced to utilize its press feed line to process the outer blank ISF products, tying up capacity and decreasing efficiency. The press feed line’s designated purpose is actually for punching holes and dropouts in formed flat products, for example, pressing out windows to produce a window frame. Steel is then removed from the centre of the blank to produce the window frame. Trident does not experience the same problem with inferior quality steel. Its state-of-the-art equipment eliminates problem of inferior, dirty and unevenly oiled steel material obtained from ISCOR because they have washing and re-oiling capabilities that clean off surface defects. Additionally their processing lines are presently being under-utilised.

85. The parties estimate that the current capital expenditure required for the processing of outer steel blanks by Baldwins on its press feed line approximates R3,000 per hour. Should processing such outer blank products be effected on Trident’s processing lines, the cost will be reduced to R1,500 per hour. This amounts to a substantial cost saving of 50%. Consolidation of the firms’ manufacturing processes would reduce the amount of “scrap” generated by each firm individually. Baldwins scrap rate is estimated by the parties to currently be in the region of between 7%-8%, whereas Trident’s is lower, at 2%-3%. Once the manufacturing operations are integrated, the average will approximate 3%, on the parties’ submissions, generating cost savings that would add to the efficient use of Trident’s plant. It is accordingly clear that by acquiring Baldwins’ press feed line and utilising its own plant facilities and cut-to-length line to optimal purpose, Trident will process ISF blanks more efficiently and cost-effectively than before, ensuring the merged entity becomes a competitive, low-cost processor of steel blanks.

86. The merger will also allow for plant level re-organisation, achieving significant real economies. Trident’s excess capacity will be used to perform Baldwins’ existing cutting capacity. Baldwins’ press feed line would then be free to be used for its optimal function i.e. the pressing of blanks for windows and doors, a function Trident does not currently provide as it does not have the requisite machinery. Clearly the ability to provide this additional service will make Trident more competitive. By way of example, Trident say that this new capacity will enable them to compete for the processing of the outer steel blanks for the new Mercedes-Benz C-Class contracts, a contract that they would otherwise not be sourced locally. This contract is valued at R50 million per annum. None of the economies we have outlined above would have been achieved without the merger.

Supply efficiencies

87. The merger would also lead to production efficiency gains for the supplier. Iscor provides a standard list of products it supplies which incorporates a finite product range. The exact products that make it onto the standard list are determined by the amount of tonnage of the product merchants order annually. Iscor demarcates a minimum amount of tonnage merchants must order annually to make it worth their while to manufacture it, and therefore put it on their list. The combined entity would order more tonnage annually, thereby inducing Iscor to place that particular product on the standard list. If the product is not on the list, merchants sourcing from Iscor have to incur the cost of buying alternative products which subsequently have to be cut down to size by them. The balance is then disposed of as scrap. Therefore getting Iscor to supply as many products as possible on this standard list is crucial to minimize wastage in the steel merchants’ plants. For instance if Iscor’s standard list product is in a 1100mm form, but the parties only require 850 mm they would have to cut it down to size. The remaining material would have to be utilized as scrap or in some other lower value form. Pre-merger the parties’ individual ordering levels are too low, whereas post merger the combined quantities of both entities would ensure the order reaches the requisite level, to be placed on the standard list.

.

88. Although the parties were not able to precisely quantify the efficiencies that would result from such reductions of wastage, it is obvious that having such a standardized system would allow Iscor to make available correctly sized products, encouraging increased production and output levels, by getting the right product to market faster and reducing wastage. By optimizing their own processes, suppliers improve the efficiency of the industry as a whole.⁶¹

89. This same principle was referred to in Hillsdown⁶², where the Court quoted with approval the following speech by the former Director of Investigation and Research on 15 October 1988:

“ However cost savings resulting from larger volume orders, which enables the purchaser to attain economies of scale or incur lower transaction costs, may reflect real efficiency gains and consequently may be accepted for consideration. If the placement of larger volume orders also enables the supplier to reduce costs, part of which are transferred to the purchaser in the form of lower prices, then that part may also qualify as real efficiency gains.”

Volume discounts

90. The parties also claim a further efficiency gain because they will become entitled to volume discounts from Iscor. We treat this claim with much greater scepticism than the others. Volume discounts on their own do not, in the absence of other categories of efficiency gains, necessarily constitute the standard of efficiency contemplated in section by 16(a)(i)). These are not gains brought about by a saving of resources. As the Canadian Guidelines suggest:

“ this is contrasted against gains that are anticipated to arise as a result of increased bargaining leverage that enables the merged entity to extract wage concessions or volume discounts from suppliers that are not cost justified, representing a mere distribution of income to the merged entity from employees or the supplier, as the case may be. Such gains are not brought about by a saving in resource.”⁶³

Accordingly we have not taken volume discounts into account in weighing up the efficiencies.

Conclusion

91. The efficiencies the parties have claimed are in our view sufficient to be “greater than and to offset” any anticompetitive effect. Although we have insufficient evidence to quantify this in the form of calculations⁶⁴, the efficiencies claimed are so overwhelming, especially in relation to the plant re-organisation that is entailed and the reduction of the scrap rate that they suggest, that they will dwarf the anticompetitive effects. We must bear in mind that the merging firms ability to increase price is only up to the import parity price. Any move on their part to price above this will lead to customers sourcing overseas. Since this import parity price is not likely to be much higher than the current market price, the anticompetitive effects whilst real, are constrained.⁶⁵ Had this not been the case, we may have either found the trade off had not been sufficiently established or we might have considered approving the merger, but subject to appropriate behavioral conditions.

92. The efficiencies contemplated could not have been achieved without the merger. Baldwins produced evidence to demonstrate that its Rosslyn plant had been run at a loss for more than two years. The firm was not committed to expending any more on the plant and no other buyers could be found for it. Extracts from Director’s minutes dated 5 August 1999 show that the company was concerned about its Rosslyn plant’s profitability for some time and was investigating various options, prior to its ultimate decision to sell.⁶⁶ The supply efficiencies from Iscor required a single firms’ order and could not be achieved by the firms individually. Although there is no evidence that the efficiencies will be passed through to consumers in the form of lower prices, the nature of the efficiencies is such that this need not be shown in the context of this merger if we apply the proportionality test we have adopted above.

Public Interest Issues

93. If the merger proceeds, the parties estimate the number of retrenchments following the implementation of the merger will not exceed 10 and this will affect only management staff (general managers, sales managers, debtors clerks and inventory controllers). Thereafter, they estimate a further 40 employees will leave Trident’s employ at a normal industry rate of attrition. In contrast if the merger is prevented Baldwins would be forced to close down some of its plants and scale back at others leading to a greater loss of employment.⁶⁷

Conclusion

In light of the above the Tribunal is satisfied that although the merger does substantially prevent or lessen competition in the ISF market, the parties have successfully discharged the onus of proving that such anti-competitive effects are convincingly offset by the efficiency gains the merged entity, as well as the industry, are liable to experience as a result of the merger. For this reason the merger is approved.

30 January 2001

N.M. Manoim Date

Concurring: S. Zilwa and P.E Maponya

---