Vertical mergers that look horizontal— optical illusion or reality?

A case study transcript from presentations to antitrust practitioners, 2019

Ok, good afternoon and thanks so much for inviting us to share with you today. Thanks also for the great introduction. I have also invited my colleague Albertus to join in with me today. Ideally, we would like this to be a relaxed and informal discussion, so he will be adding in on some points of interest to him. We would also love to have comments and questions at any point, so please feel very welcome to raise discussion as we go.

As stated, we will share with you our recent experiences regarding the Timrite-Tufbag merger. This is an intermediate merger than was heard by the Competition Tribunal last year, with reasons published in May this year. Throughout this period, we advised the merging parties, also providing expert testimony at the Competition Tribunal hearing. So today we will share a little on that process, of advising and testifying. After this we will also share a little on the actual economics of the case. In our opinion this case provides many possible discussions and learnings, so hopefully we can elicit some of those.

To begin it is necessary to state a little about the companies concerned in the merger. Timrite is owned by Thebe investments, one of the largest investment houses in South Africa. Timrite, as it names implies, started in the timber industry, making and supplying timber support products to mines. In the last decade it however also saw a gap in the market for other forms of underground support products to mines, in particular mining bags, which I will explain in a moment. For this to work well Timrite signed an exclusive dealing arrangement with Tufbag whereby Tufbag would make certain mining bags, and Timrite would supply these to the mines. For this the companies shared joint IP on specific products in respect of the agreement, which was termed the ‘product supply agreement’ or ‘the PSA’. These mining bags are essentially large polypropylene bags that are pumped with substances such as cement and grout and used to form walls, roofs and air channels in underground mines. Many of these mines are more than a kilometer deep and extremely dangerous and so these bags are used to keep the mines from caving in, allowing for safe mining practices. The way in which the bags are designed and pumped is also highly technical, in order to avoid bursts, mining shaft collapses, and ultimately mining incidents involving workers. For these purposes, bags are produced; Tufbag makes and Timrite supplies in excess of 273,000 such bags per year. Within this context, Timrite and Tufbag wished to merge, given that Tufbag’s shareholders wished to exit the business, and Timrite saw many potential efficiencies from the merging, which was essentially a furthering of the vertical arrangement that they already had in place with Tufbag.

So, it is on this merger that we advised. We will return to the firms and their businesses in a moment, but first a few words on the Tribunal hearing process. This merger between Timrite and Tufbag was the first that followed what is commonly termed a ‘hot tub’ hearing process. This essentially means that evidence is provided by the experts concurrently rather than sequentially.

In this case the opposing side was the Commission itself, who had recommended to the Tribunal that the merger be prohibited (we were not involved at this stage). But regardless of who the opposing side is, as in any hearing, the Tribunal panel needs to elicit all of the relevant information from both sides. In any given week they may be presiding over many matters, so this also needs to be as efficient as possible.

As I am sure you then know, in this process economists often have an important role to play, in acting as expert witnesses and advising the Tribunal panel.

This is however a role that has been hotly debated in recent years, most recently in this country in the 2015 excessive pricing case against Sasol Chemical Industries. In this case Judge Dennis Davis set out the responsibilities of an expert witness. These responsibilities are also in line with international guidelines. As is often quoted at the start of South African Tribunal hearings, the Australian Federal Court states: “…”

Aligned with the responsibilities of an expert, the South African Competition Commission and Tribunal have in recent years often required pre-hearing ‘hot tub’ meetings. These are meetings held between opposing expert economists (or other experts) in a case, wherein they are requested to independently discuss and document commonalities or points of agreement in their approaches. This is intended to speed up the actual hearings by focusing the discussion on the contestable points and thereby increasing the efficiency with which the experts are able to assist the Tribunal.

Following the introduction of ‘hot tub’ pre-hearing meetings, the use of ‘hot tub’ hearings have now also been introduced, with the Timrite-Tufbag merger being the first in this respect and a number since then. In terms of procedure, the experts were allowed to open with a 15-minute presentation of their key points, with the remainder of the experts’ testimony taking the form of direct discussion between the two experts. To facilitate this process of concurrent evidence, a specific pre-hearing meeting was held in order to establish and agree on the main points of discussion. These items were then raised by the Tribunal in the hearing, as an opening to each discussion point. Each point was then debated between the two experts, with each requested to respond to the points raised by the other, and with allowance for direct questions, both between the experts, and also from the Tribunal members. Following this, examination by Counsel was allowed, but the duration was significantly lessened due to the fact that the key points and views were already rapidly and explicitly covered through concurrent evidence. An additional advantage of the proceedings in this instance was that the experts were allowed to be assisted with documentation by their team, further increasing the efficiency with which key evidence could be raised and discussed.

Beyond the general efficiency gains in this instance, this case was well suited to the use of concurrent evidence, given that the frameworks of analysis by the experts differed fundamentally. Hence the traditional proceedings may have left the Tribunal with two vastly different testimonies to assess. In this context, the proceedings mandated both experts to deal with all relevant issues, with each responding to the other on all points; again, eliciting the most relevant facts for the Tribunal’s assessment. In addition to this and in our opinion, the experts were forced to crystalize the main points of their arguments clearly and concisely. Since each expert knows that the other expert would be able to expose an inappropriate answer immediately, or reinforce an appropriate one, the evidence generally proceeds directly to the critical points of difference. This can aid the Tribunal (or court) in understanding a point more clearly before moving on to another point. Secondly, it allows the economists to directly test one another’s assumptions on a more technical level than would generally be possible through traditional cross-examination. This aids in one economist being able to correct any “misinformation” provided by the opposing expert economist that may not be picked up by the cross-examiner, particularly where it relates to technical, yet important, details. Concurrent evidence also allows this to be done in “real time” before the Tribunal (or court). Third, all of the evidence concerning each issue is dealt with in a logical progression and can be found in one place in the transcript.

In summary then of the process and our impressions as economists, it appears that this the use of hot tub hearings may be useful, but this should be assessed on a case by case basis. Since then there have been a number of hot tub hearings, as I am sure many of you have participated in.

Now having shared a bit on the process, we can move on to the economics of the case. Thanks to the pre-hearing hot tub in this case we were able to enter into the hearing with many points of agreement between the opposing economists. The cornerstone of any good economic analysis is often the market definition. In this case we reached agreement on many points, but not on exactly how broad the market was and whether it included older designs of mining bags. Accordingly, we could not agree on how many players were in the markets in question. We also analyzed this as a vertical merger, whilst the Commission stated that it was a horizontal merger, as I will show in the next slide. Therefore, we were technically not in agreement or disagreement here, given that the Commission had no vertical analysis.

The main points of contention were therefore highlighted under the Commission’s horizontal assessment of the merger. This was also the main reason for the Commission’s initial prohibition of the merger. Essentially, they stated that, but for the product supply agreement between the two firms (that is the exclusive arrangement in respect of specific IP-related products that we discussed earlier), the merging parties would be competing horizontally. I.e. if it weren’t for this agreement, Timrite would be manufacturing and Tufbag would be supplying to the mines. The Commission therefore put forward theories of horizontal and potential competition, which we discuss in a moment. Essentially, they argued that if the merger would to be prohibited and also in the absence of the agreement to date, Timrite would manufacture, adding competition at that level of the value chain. Their conclusion therefore followed that there may be a range of counterfactuals that would be more competitive than the merger.

In contrast to this and through much research and analysis, we found that that the merging parties were in a vertical relationship and that they were not horizontal competitors. This followed on from the fact that the PSA was in respect of specific IP-related products only and therefore it did not preclude horizontal competition between the merging firms outside of these products. We also found that it was unlikely that, absent the transaction, the merging firms would begin competing horizontally. We therefore found that the facts did not agree with the Commission’s theories of harm, specifically relating to horizontal and potential competition. Accordingly, we found that the merger counterfactual, against which the merger should be assessed, was the status quo, and that therefore no competitor or competitive constraint would be removed by the merger. Upon further investigation we also found that there were many pro-competitive effects deriving from the vertical agreement between the parties to date, that would be furthered by the merger.

Put differently, the main points included that the CC saw this as a horizontal merger, whilst we assessed it as a vertical merger. The vertical aspects were not contentious as the CC did not cover these as a result. So the only contentious points were on the horizontal aspects. Here the CC advanced two theories of harm. Firstly, one of market allocation, i.e. that absent the PSA, the merging parties would be competing. We negated this by demonstrating that this was in respect only in specific products, and that it was terminable, and therefore that the PSA did not allocate markets or prevent competition. Secondly and following from this the CC advanced the theory that the merger removes potential competitors. We negated this by first contextualizing this as a theory with a very high bar, that in this case, was not met, and also showing that there was no indication that even if one were to enter into the others market absent the merger, the CC had not in any way demonstrated that that would provide for more competition.

Arriving at these conclusions ahead of the hearing involved a great deal of work, but we will explain in brief how we got there.

Starting on market definition, as mentioned we could not agree on this. We found that there was a market that was broader than engineered mining bags — the market which the CC agreed Timrite in any event created. We rather argued for a broader market that included other types of mining support bags. This was supported by many facts, perhaps most strongly the fact that mines procure overall solutions and the precise choice of bag is dependent on the mining engineer and the exact use case. The relevance of this broader market was that there were many more players in the markets concerned — both at the upstream and downstream level. In any event, the Tribunal decided not to conclude on this and our case did not turn strongly on this.

Market definition aside, the main concerns raised by the Commission had to do with their perception that absent the PSA, the merging parties would be competing horizontally. The PSA however as we showed in the proceedings was for the joint development of specific products covered by IP, in order to protect their investment. This can also be viewed as an exclusive dealing arrangement in respect of very specific products. But importantly it did not in any way preclude the merging parties from competing in respect of any other products within the relevant market but outside of that jointly developed IP.

Further to this and in considering the CC’s concerns, we noted that the European Commission’s Guidelines of Vertical Restraints, which were particularly helpful here. More specifically this says that:

“Two companies are treated as actual competitors if they are active on the same relevant market. A company is treated as a potential competitor of another company if, absent the agreement, in case of a small but permanent increase in relative prices it is likely that this first company, within a short period of time normally not longer than one year, would undertake the necessary additional investments or other necessary switching costs to enter the relevant market on which the other company is active. This assessment has to be based on realistic grounds; the mere theoretical possibility of entering a market is not sufficient.”

Against this context, we reasoned that entry into a market must be based on sound evidence and not theory; and that in this case there was no evidence of constrained competition to date, and no compelling evidence of competition absent the merger going forward. It therefore followed that no potential competitor would be removed as a result of the merger.

On this point the CC however disagreed, focusing on Timrite’s potential entry into Tufbag’s market, which they found would have happened absent the agreement, and hence also absent the merger, which is really an extension of the agreement in many ways. Looking to the literature we however find that one would need to in such a case assess the probability of entry and then also crucially the competitive effects of that entry.

The wide literature on this topic also highlights that the bars for the doctrine of potential competition are high. In the US we found that there has not been a federal court of appeals merger decision applying this doctrine since the 1980’s. The recent Dow Dupont merger also indicated that proof of significant intentions to enter are required. In this case and as you know this was relevant in respect of maize products and the intention to enter the South African market was well documented and specific, hence requiring third party licensing in this case.

Again, on against these bars, we do not find that the merger in question removed any potential competitor.

This was also the finding of the Tribunal, where they acknowledged that Timrite has considered entering the upstream market but that crucially this had never occurred as it was clear that every time it was considered, it was found that Timrite did not have the necessary operational experience at this level of the value chain. Second to this and perhaps even more compelling is the fact that every consideration by Timrite to enter the upstream level of the value chain was never intended to absorb all Tufbag’s production, but rather as a backup plan for the very real times where there were supply disruptions and business had to somehow continue. This was also considered to therefore be at a very limited level and never for full entry into that upstream level. And so, it was on this basis that the Tribunal dismissed the CC’s theory of potential competition and removed by the merger of a potential competitor.

Following on from our dismissal of the CC’s main concerns — which the Tribunal later agreed with us on — we turned to an assessment of the current vertical relationship between the merging parties. Here the theory that postulates that there may be many pro-competitive aspects of vertical restraint agreements — in helping to realize efficiencies and the development of new markets, as in this case. More specifically here we found that the PSA promoted product development, that might not have otherwise been seen. The PSA also provided quality assurance to mines in that there was no uncertainty as to who the product came from and who would be liable in the case of any product defects. The agreement setup further promoted the avoidance of double marginalization and hence also the possibility to reach scale. Following on from this we found that the merger would further all of these pro-competitive aspects, and crucially inefficiencies would be improved, for more continuous supply to mines. Pre-merger this was really a problem, with many supply disruptions due to labour protests and hence problematic stopping and starting of supply. Not to say that this would not occur post-merger, but Timrite would likely have more control over this, being fully integrated.

On the points of input and customer foreclosure we did not find any issue and no issues were raised in the CC’s two reports prior to the hearing. In the hearing however, this was then raised as an issue, not per se related directly to the merger, as we will show in a moment, but for which the merging parties rapidly put forward remedies that alleviated all concerns. I will speak to this with the counterfactuals.

Finally, we turned to the market position of the merging parties with the thinking that an exclusive dealing arrangement may be deemed to be anti-competitive if there has been an increase in prices or a decrease in output that cannot be explained by other market factors. On this point and in analyzing Timrite’s pricing we noted that Timrite agrees on pricing up front with clients and signs an agreement in this way. Timrite cannot then increase prices outside of this agreed pricing mechanism, unless there is written agreement. As the buyers are large mines that are price sensitive this really constrains Timrite’s pricing to them. A few clients we noted also did not have contracts and the same buyer power applied here. Pricing was also often on stringent tender bases. This was significantly compelling to indicate that any market power in any way attained by the merger could not be exercised, any in any event we found there to be a significant degree of competition in the relevant markets and reasonable entry barriers such that most clients did not even require products that were covered under IP and evidence from them further supported that they shop around after contracts expire.

Following on from this assessment, we considered the correct counterfactual. As I am sure you know, this is generally the ultimate point from which to assess the merger, i.e. considering whether or not the alternative world, i.e. absent the merger, would be more or less competitive. In essence our discussion has shown that the counterfactual that we determined was the status quo, with the possible exit of Tufbag’s shareholders over time. We hence determined that the merger would hence not remove competitors but would further the pro-competitive effects of the vertical agreement. Crucially though, we also had to assess and address the Commissions’ counterfactuals. Usually one would put forward one counterfactual from which to assess a merger, but in this case the CC put forward a range. These are listed on the left-hand side here.

Whilst they may seem complicated at first, these are quite simple to understand. In the first instance the CC hypothesis that, absent the merger, each merging party enters into the others market. Then in addition to this, the other suppliers with which Timrite has agreements are freed up for supply to other downstream players, including the two of them, who are now both operational downstream. Counterfactual 3 then considers only the additional part of counterfactual 2, with freed up supply, but no potential entrants. Counterfactual 4 and 5 then consider the freed-up supply but with only one or the other entering the others market. Having already discussed our responses to the CC’s theories of harm, these are quite easy to address. In the case of entry by one or the other into the others market, we found that the CC did not provide practical evidence that was sufficient to prove this, both absent the agreement, as well as absent the merger. More importantly however, we found that the CC had provided no evidence to show that this would be more competitive, i.e. going beyond the simply assumption that a count of one more competitor is good. And then in the case of the additional freeing up of supply, we found that this was the first mention of any foreclosure concern by the Commission, in respect of Timrite’s agreements with other parties (beyond Tufbag). At the time of the hearing, Timrite however had no further exclusive agreements, and agreed that none would be made going forward. This concern was hence also negated, and any forward-looking concerns alleviated. In summary then, we could not find a credible counterfactual amongst those put forward by the CC that could lead to the argument that the merger would result in a less competitive world.

On this basis we concluded. In terms of market definition, we found that the markets are larger than those products covered by the PSA; that there are no foreclosure concerns. In terms of our vertical assessment we found that that the PSA was pro-competitive, and that the merger would likely further some of these effects We also found that the context in which the merger was taking place was characterized by a high degree of competition and buyer power, as well as reasonable barriers to entry. And finally, in terms of our horizontal assessment, we found that there was no evidence of constrained competition to date, no evidence of potential competition by the merging parties going forward; and no potential competitor removed as a result of the merger.

*I provided written and oral evidence in this matter, whilst employed by Econex and instructed by Hogan Lovells