Genna Robb

In September 2015, the Competition Tribunal found that Media 24 had contravened the Competition Act by engaging in a predatory strategy in order to eliminate a small competitor in the community newspaper market in Welkom.[1] The Tribunal found that the prices Media 24 had charged prices for advertising for one of its two titles in Welkom were lower than its cost of producing the newspaper over a sustained period. As the newspapers were distributed for free to readers, advertising costs were the only available source of revenue, so Media 24 was operating the paper at a loss. The Tribunal moreover found that Media 24 had done so with the intention of eliminating its competitor, and that following the exit of the competitor it had raised prices in the market in order to recoup the earlier losses. This had harmed customers (both advertisers and readers) by reducing price competition and limiting choice.

In the context of South Africa’s concentrated economy, dominated by large entrenched firms, and the goals of the Competition Act which espouse the protection of small businesses, this is the type of offence which should be strongly sanctioned. However, due to the fact that the Tribunal made this finding under section 8(c) of the Act, a catch-all section for exclusionary abuse, rather than under the specific section dedicated to predatory conduct (section 8 (d) (iv)), Media24 escaped a financial penalty and instead will face only such remedies as the Tribunal finds necessary to prevent a repeat of the conduct. Such remedies are unlikely to address the anti-competitive harm which has been caused (not least because the competitor in question has already been driven out of the market) and are unlikely to prove a strong deterrent to firms considering similar conduct in future.

In this piece we unpack the approach to assessing predation taken in the Tribunal’s decision and ask whether the Act, or the Tribunal’s interpretation of it, tends to stack the odds against a small firm complaining of predatory conduct by a large multi-product incumbent.

What are the key issues raised in the decision?

A wide range of important economic issues are raised in the Tribunal’s lengthy decision. For the purposes of this piece we have focussed on those which directly affected the question of whether the conduct should be considered a contravention of section 8 (d) (iv) (and hence eligible for a fine on the first offence) or section 8 (c) (with no fine for a first offence). Thus the focus is on the price-cost tests used and how the relevant costs have been calculated. The discussions of intent, recoupment and effects, whilst important in supporting the theory of harm and establishing the harmfulness of the conduct, are not integral to the determination of relevant costs and hence we do not discuss them further here.

Predatory conduct by a dominant firm occurs when the firm charges prices which are below cost for a limited period of time with the intention of driving a smaller rival out of the market, based on the belief that it will subsequently be able to increase prices above the competitive level.[2] Given that pricing below costs will result in sacrificing profits or even incurring losses, predation would be irrational if the firm did not believe that it would be able to recoup the sacrificed profits at some point in the future. If the firm does not charge prices below costs, then predation does not occur, and the conduct is simply an example of aggressive competition.

Tests for predation therefore generally focus on the issue of profit sacrifice and involve comparing the revenues generated by the product with the costs involved in producing it in order to determine what, if any, sacrifice the firm has made. A key question for the enforcement of competition law is therefore which costs are relevant for such an assessment.

Economic theory suggests that a profit-maximising firm would not price a unit of output below the marginal cost of producing that unit, as this would result in it incurring losses, and it would be better off not producing the marginal unit. Therefore, if a firm is found to be charging a price which is below its marginal cost over a prolonged period, it can be inferred to be deliberately sacrificing profit, and therefore to have predatory intent.

In practice, marginal cost is difficult to measure and so a range of alternatives have been proposed in order to determine whether a firm is charging predatory prices. One such alternative, first proposed by Areeda and Turner (1975), suggests taking average variable cost (AVC) as a proxy for marginal cost[3]. Variable costs are those costs which vary in proportion to output. This approach was adopted by the drafters of the South African Competition Act, and as a result, section 8 reads:

“It is prohibited for a dominant firm to –

(a)…

(d) engage in any of the following exclusionary acts, unless the firm concerned can show technological, efficiency or other pro-competitive, gains which outweigh the anti-competitive effect of its act:

(i)……

(iv) selling goods or services below their marginal or average variable cost” [own emphasis]

More recently, however, there have been criticisms of the AVC benchmark. These relate to the necessity of deciding over what period costs should be considered variable, and the fact that short run variable costs are not the only determinants of firms’ prices.[4] Critics also point out that the AVC benchmark could be “gamed” by firms over-investing so as to have spare capacity but low variable costs, and that it is not very useful in network industries where variable costs are low but sunk costs are high.[5]

In light of this, the concept of average avoidable cost (AAC) has been introduced as an alternative cost benchmark. Avoidable cost involves comparing the incremental cost of remaining in the market with the avoidable cost of exiting it. If the firm would be better off discontinuing the line of business, then it is behaving irrationally in staying in the market.[6] Some consider AAC to be more appropriate than AVC for determining whether a firm has engaged in predatory conduct, as it measures both the variable and product-specific fixed costs which would be saved if the firm decided to exit rather than predate. Thus AAC may be a better measure of the dominant firm’s profit sacrifice.[7] As shown above, however, AAC does not appear explicitly in section 8 (d) (iv).

Three main issues arise in relation to the definition and measurement of relevant costs in the Tribunal’s decision. The first is whether or not AVC and AAC can be considered to be equivalent, both in this particular case and in terms of the South African Competition Act. The second is whether it is appropriate or desirable to include an estimate of the dominant firm’s opportunity cost in the calculation of its avoidable costs. The third is the question of how to treat shared costs for a multi-product firm. We deal with each of these three in turn.

Are AVC and AAC the same?

The first question the Tribunal asks in its decision is whether or not AVC and AAC can be considered to be equivalent for the purposes of the test prescribed in section 8 (d) (iv). The Tribunal is of the view that Section 8(d)(iv) does not compel one to use only AVC or MC, but “requires that the measure of costs chosen would yield a result that would still be below at least one of them”.[8] The Tribunal acknowledges that there is a difference between AVC and AAC in that AAC includes product-specific fixed costs[9]. However, in this instance it notes that AAC is appropriate as the theory of harm concerns a fighting brand. A fighting brand is a particular product which the incumbent firm uses to drive a new entrant out of the market. This strategy usually involves targeting the particular sub-segment of the market where the entrant’s products are situated, in order to lower the cost of predation and avoid having to price below cost across the whole market. In this instance, Media24 had two titles in the Welkom area, and used one of them in particular to target the entrant. As the Tribunal notes, it is therefore appropriate to consider whether the incremental revenues generated by the fighting brand exceed the costs which could have been avoided if the brand had exited.[10]

The Tribunal explains the relevance of the AAC test, saying: “There may be cases where one might be able to rely on another cost standard, say AAC, and demonstrate that it yields a result where the firm concerned is pricing below its AAC, but that the costs relied on for this exercise are not only avoidable but also variable.”[11] In other words, the Tribunal considers that AAC may be an appropriate test in some instances, but prices must also fall below AVC in that instance for it to be considered a contravention under section 8 (d) (iv).

The Tribunal cites the EC’s abuse of dominance guidelines in support of this approach: “In most cases the AVC and AAC will be the same, as often only variable costs can be avoided. However, in circumstances where AVC and AAC differ, the latter better reflects possible sacrifice; for example, if the dominant undertaking had to expand capacity in order to be able to predate, then the sunk costs of that extra capacity should be taken into account in looking at the dominant undertaking’s losses. Those costs would be reflected in the AAC, but not in the AVC.”[12] [Tribunal’s emphasis]

The Tribunal does not ultimately decide whether it would accept the AAC benchmark, as it finds the Commission has not shown that Media 24 was pricing below AAC in any event (this will be discussed further below). But it implies that it may have been inclined to do so because, in its view, AVC would have been equivalent to AAC in this case.

The second half of the EC quote therefore raises a potential conundrum for the Tribunal. It sets out a clear situation where there is a better economic justification for using AAC as a benchmark than AVC, but where, according to the Tribunal, it would not be possible to make a finding under 8 (d)(iv) since AAC would likely be higher than AVC. This begs the question of why use the AAC test at all if it will only identify the same instances of predation as the AVC test and continue to exclude situations where there are important product-specific fixed costs which should be taken into account?

Furthermore, in the current instance, we are not convinced that the two measures are equivalent. The fact that the fighting brand was kept in the market in order to exclude the smaller competitor means that its avoidable costs would have included all of the product-specific costs of operating the fighting band, some of which should have been classified as fixed unless an extremely long time-horizon was used for the definition of variable costs.

Ultimately, it seems to us that the Commission was correct in proposing AAC as the more relevant benchmark, whether or not pricing below AAC also amounts to pricing below AVC in this case. The Tribunal’s attempts to relate AAC to AVC confuse the issue somewhat by ignoring the important differences between the two measures. An alternative approach could have been to accept AAC as an appropriate and valid proxy for marginal cost (as is the case with AVC according to the Act); or to determine that AAC could not be used as a benchmark under the current formulation of section 8 (d) (iv), in which case there may be a case for the amendment of the Act.

Is opportunity cost a relevant part of AAC calculation?

One of the critical factors in the Tribunal’s finding that Media 24 did not price below AAC was its decision to exclude opportunity costs from the AAC calculation. The Commission had contended that this should rightly be included as part of the AAC calculation since some of the profits that Media 24 could have made through its main title, were sacrificed to the fighting brand, as some advertisers would have chosen its main title if the fighting brand was not an option.[13] In other words, if Media 24 had closed down the fighting brand title, it would not have lost all of the revenues associated with the fighting brand, as some of its revenues would have been diverted to Media 24’s other title. The Commission argued that this was relevant to the consideration of whether keeping the fighting brand in the market was a rational decision or not.

The Tribunal considers this argument in detail, but notes in its decision that whilst there is support for the idea in an EC guidance document on exclusionary abuse and one UK case (Cardiff Bus), it is not a widely used approach currently. In addition, in this particular instance, there was contention around how much of advertising revenue would have been diverted to Media24’s main title if the fighting brand had closed down (what the correct ‘diversion ratio’ was), and the Tribunal concluded that the evidence was unclear. It consequently adopted a “pragmatic approach” to exclude opportunity costs from the calculation.[14] Although some estimates for the possible diversion of revenue were available, the Tribunal argues that taking a conservative approach to the available evidence would not help, since: “Once the basis to the approach is accepted to be built on unreliable foundations it does not help to construct something less ambitious but still suspect as a means of calculation on top of it.”[15] Thus, opportunity costs were excluded altogether from the calculation of AAC.

In our view, there is a clear rationale for considering the diverted revenues. The difficulty of estimating them with precision does not divert from the heavy preponderance of evidence that they in fact existed, and thus were relevant to the case. The exercise proposed by the Commission was to compare the incremental returns to keeping the fighting brand in the market to the incremental returns if it were to exit the market. The returns to staying in the market amount to the revenue made by the fighting brand minus the costs of keeping the fighting brand in the market (all the costs of operating the fighting brand). The returns to exiting include the revenue diverted from the fighting brand to the main title minus the unavoidable costs of operating the fighting brand and the costs of exit. For it to be rational to stay in the market, the returns to staying in the market must be greater than the returns from exiting. This means:

Revenue FB – all costs > diverted rev – unavoidable costs – cost of exit

Which can also be written as:

Revenue FB > avoidable costs + diverted rev – cost of exit

If the revenue of the fighting brand is greater than the avoidable costs plus the diverted revenue from the fighting brand to the main title minus the cost of exit, then the decision to continue to operate the fighting brand is rational. If not, then the fighting brand can be assumed to be in the market only to engage in the predatory strategy.

Therefore, the diverted revenues are of critical importance to the calculation of whether the observed pricing is part of a competitive or predatory strategy. In making the decision to keep the fighting brand open, Media 24 knew that if it closed the fighting brand, it would not lose 100% of its revenue. This would have weighed in the decision to keep it open or not, and therefore it is important to take this into account in evaluating the rationality of its decision. Clearly the diverted revenue would be an amount greater than zero, given that the Tribunal had found that the titles were in the same market. It is unfortunate that better evidence on the likely diverted revenue was not available, but a better approach may indeed have been to make a conservative estimate of what that number could be. Particularly since, depending on the level of diversion ratio which was chosen, the diverted revenues could have made the difference in determining whether or not price was below AAC.

What to do about shared costs

A second issue which affected the assessment of whether Media 24 priced below AAC was the question of Media 24’s shared costs. As a large multi-product firm, there were a number of shared costs from the larger operation which were partially apportioned to the fighting brand. Media 24 argued that it would not be able to re-deploy any of these shared costs if it closed down the fighting brand, making them unavoidable. The Tribunal accepted this interpretation and ruled that all shared costs should be excluded from the AAC calculation.

The Tribunal recognised that such an approach has the potential to disadvantage small firms: “Of course this means that on an AAC standard, the large scale multiproduct firm has a greater ability to claim that costs are unavoidable because the alleged predator‘s costs become trivial fractions of the larger pie meaning that they make no sense to avoid when the unit is closed.”[16]A small, single-product competitor is not able to share costs across products or business units and must cover the full cost of producing the relevant product with the price charged for that product. Thus excluding shared cost from the calculation necessarily places the small firm at a disadvantage as the price may be below its avoidable costs but not the large firm’s avoidable costs. On the other hand, it may be unfair and undesirable to penalise large firms for their size and lower costs.

The Tribunal suggests as a solution that this favours the use of average total costs (ATC) as an appropriate cost standard in such cases. The less stringent approach to the price-cost test then implies a lesser finding under 8 (c) of the Act, rather than 8 (d) (iv) and consequently no penalty for a first offence. This trouble with such an approach is that in cases such as the current one, where there is clear evidence of predatory intent and significant harm to competition and consumers, the sanction ends up seeming disproportionately small in comparison with the deliberate anti-competitive behaviour. This will also reduce the extent of deterrence of such behaviour.

What chance does a small firm have of bringing a successful complaint under 8 (d) (iv) against a much larger multi-product firm?

The fore-going discussion has highlighted the difficulty which a small entrant faces when trying to bring a complaint of predatory conduct against a large multi-product firm. According to the Tribunal’s decision in this case, in order for it to find a contravention of section 8 (d) (iv), the complainant effectively needs to show that prices charged are below AVC for a sustained period with no shared costs or diverted revenue to be taken into account. It has been illustrated above that this may be an inappropriately high benchmark from an economic point of view, at least in some cases, and that the resulting calculation does not necessarily reflect the reality and gravity of the underlying situation.

This matters all the more because, as highlighted by the Tribunal in this case, one of the main economic theories of predation emphasises the incentive of the dominant firm to engage in predatory conduct in order to create a reputation, possibly in multiple markets, for an aggressive response to entry. Thus the effects of predation can go beyond the effect of eliminating or marginalising a competitor in one market. Without competition law that can effectively deter such conduct, entry may be discouraged, especially in some highly concentrated South African markets.

When drafted, the decision to include the phrase “marginal or average variable” cost in section 8 (d) (iv) was no doubt intended to improve clarity, by providing guidance on the expected evidentiary standard. However, by doing so the section has been made excessively rigid, unable to change in accordance with best practice in economic theory, or deal with the full range of cost conditions experienced by firms. A change to the Competition Act, or at least the interpretation of it, may be necessary in order to ensure we do not err on the side of protecting dominant firms at the expense of competition.