Abstract
If a group of horizontal rivals gets together to agree on a way to structure efficient production, are they violating competition law? The issue could arise where a group of producers of agricultural products gets together to form a cooperative or even where professionals in the same field get together in a partnership. On the face of it, each supply agreement between the producer and the cooperative or partnership is vertical, but the design of the collective rules, which govern for all, involves horizontal coordination. This article takes as the starting point the example of dairy cooperatives as they emerged in the later part of the nineteenth century as a solution to a challenge offered by new technology. We use the landmark contract law case of McEllistrim v. Ballymacelligott Cooperative to illustrate the ways in which competition law could be engaged when cooperatives are formed. Comparisons of Ireland and Denmark in the period leading up to the decision suggest that not only might the restraint be ancillary, but if not, it reduced costs, increased quality, and was welcomed by consumers (though these were in England rather than in Denmark or Ireland). The restraint also appears essential in some form, suggesting that either ancillarity or the application of Art. 101(3) Treaty on the Functioning of the European Union would have allowed the restraint to be used.
Keywords
[I]f every unilateral action that restrained trade were subject to antitrust scrutiny, the courts would be forced to judge almost every internal business decision. Overregulation of unilateral conduct might threaten the competitive enthusiasm that the antitrust laws seek to promote.
1
I. Introduction
Competition law rarely engages with what goes on within the firm, treating it as a black box. 2 The same is arguably the case when it comes to the economic theories commonly seen as relevant to competition law. 3 The exceptions are the theories and laws surrounding vertical relationships including vertical mergers, where the efficiency effects of structuring the ownership and the contracts governing the vertical value chain are actively considered. These theories are only a small part of the economic theories dealing with what goes on within the firm. An important focus for these theories is how incentive problems can best be solved, either through internal command structures and incentives or through arms-length market transactions based on more or less relational contracts. 4 An important issue for competition policy is whether the solution to incentive problems arises from a desire to soften competition or to reduce costs.
While the issues related to the internal structure of the firm are mostly related to the vertical structures where block exemptions may control any interference by Art. 101 of the Treaty on the Functioning of the European Union (TFEU) into the optimal cost-minimizing structure of a firm or a supply chain, there are two concerns that have a horizontal element. One is whether incentive provisions within the firm increase the incentives for collusive behavior across horizontal rivals and how compliance programs can be constructed to deal with these. 5 The other, which we will pursue in this article, is whether collective agreements about incentive provisions can or should be exempt from competition law scrutiny, either directly because the term is an ancillary restraint or because it meets the exemption through Art. 101(3) TFEU.
The inspiration for this article is a 100-year-old case, McEllistrim v. Ballimacelligott Cooperative. 6 The suspect contract term, which was ultimately voided by the House of Lords, was a restraint of trade forcing members of a cooperative to supply all their milk to the cooperative. 7 The concern over the term was heightened by what was considered a penalty clause to apply where the term was breached. The term was included in what appeared, at first glance, to be a vertical agreement between a supplier (member) and a processor (the cooperative). Indirectly and in substance, however, it involved a horizontal agreement between the farmers who had set up the (vertical) rules governing the cooperative in the first place. 8 In the House of Lords decision, the majority interpreted the agreement between a supplier/member of a cooperative creamery and the creamery itself as a single freestanding vertical contract which included an illegal restraint of trade. The minority took a broader view, seeing the vertical contract in the context of a cooperative involving coordination between horizontal competitors, which itself could be viewed as a nexus of vertical contracts. In this view, the restraint of trade is essential to the success of the cooperative. Importantly, the restraint of trade is a part of a contract that is standard for the relationship between all suppliers to the cooperative and the cooperative itself. The wording of that standard vertical contract is agreed among the suppliers who set up the cooperative in the first place. These suppliers are horizontal rivals, and hence, the construction of the cooperative may be a horizontal agreement, capable of violating Art. 101 TFEU. The creation of the cooperative could possibly be interpreted as a horizontal agreement with vertical effects. 9 This then questions whether there are cases where the formation of a collective solution (such as a cooperative or a partnership) to a set of incentive problems solved through various vertical restraints can fall foul of competition law. In such an inquiry, both the vertical and horizontal elements play an important role in the basic question: Why is it important for members (who are horizontally related) to agree on a common (standard) set of vertical contracts to be used in the relationship between these same members and the cooperative as a separate legal entity?
We will use the McEllistrim case to shed light on this. An advantage of these early cases is that the technology, and hence, incentive problems are relatively simple, affording a clearer view of the principles behind any intervention. However, the case also serves to illustrate the implications of having a greater focus on incentive problems within firms and supply chains. While the focus will be on a single organizational form, the cooperative, we believe that there will also be implications for partnerships and associations.
This article is organized as follows. Section II describes the industry, the associated incentive problems, and ways in which organizations and procedures can be fashioned to solve these problems. Section III outlines the key facts in the case and explains why the minority decision shows a better appreciation of the incentive problem. Section IV considers how the issues raised by McEllistrim might be dealt with by competition law. The horizontally agreed vertical restraint may be outside the scope of competition law if the restraint of trade is part of a nexus of contracts and hence the restraint is ancillary. If not outside the scope, the agreements may be rescued by the exception created by Art. 101(3). Finally, the contracts may be “internal” to the firm. The case of McEllistrim serves as an interesting test of whether the definition of the boundary of undertakings is helpful once we acknowledge that firms and contracts are there to solve incentive problems and that solutions to incentive problems are usually associated with some form of risk allocation. Section V concludes this article.
II. The Context: A Challenge for Butter Producers
For farmers and smallholders with small herd sizes who wanted to produce butter, the challenge prior to 1880 was to produce enough butter of sufficient quality that it could be traded other than locally. Combining milk from several suppliers was not an efficient option because the transportation of milk on poor road surfaces essentially homogenized the milk, thereby reducing the amount of cream which could subsequently be extracted significantly. A technical innovation around 1880, the continuous cream separator, offered a potential solution. The new technology was efficient at extracting cream from transported milk. This in theory made possible commercial creameries that, by collecting milk from a number of farms, could produce larger more homogeneous 10 batches of fresh butter, which could fetch much higher prices at market.
In practice, the successful establishment of such dairies required significant investments on two fronts. Firstly, the building of the actual dairy, a processing plant with significant specific assets and a minimum efficient scale which in almost all cases exceeded the capacity of even the largest milk producers. This raised the potential danger of opportunistic holdup by the suppliers of raw milk. 11 The consequence of holdup is particularly significant when it involved the larger farms, leading to potential tensions between suppliers with different herd sizes and a reluctance of those with smaller herds to participate.
Secondly, investment was required to facilitate a change in farming practices. The milk yield of a cow is basically related to how much a cow eats. 12 With food (grass) more plentiful in the summer, it makes sense to engage in summer dairying, ensuring that the cows were in calf over the winter. To enable continuous use of the processing plant across the year required a switch to year-round dairying. 13 To enable this (some), farmers would have to change the period of the year during which a cow is not producing milk, at least for a part of their herd. They would also have to increase winter feeding and possibly even have to stable the cows. 14 If the dairy was a failure, the cost of changing the timing of calving, as well as purchasing and/or storing additional fodder, would be significant, especially for farmers with small herds. 15 Hence, the necessary change is risky for the individual supplier and requires considerable trust in the behavior of other suppliers, the success of the dairy, and in the terms of trade with the dairy. To ensure that suppliers do not find that they were the only one who has actually implemented the change, there needs to be some element of coordination.
These two collective action problems arose from significant investments which either had a high degree of asset specificity or took time to reverse were dynamic and needed to be solved for a significant period of time. Meeting these challenges requires trust, by the dairy in the suppliers and by the suppliers in the dairy. Trust in agreements could arise because these were self-enforcing due to social sanctions or because legal instruments, such as contracts, could and would be used. 16 In some cases, a combination of social and legal sanctions may be needed or more efficient. 17
To appreciate the background to the McEllistrim case and in particular the response of the Irish dairying industry to these challenges, it is useful to look to Denmark. Irish butter exports to England had fallen behind Denmark’s from the mid-1890s and the contractual terms which were core to the McEllistrim case were arguably part of an attempt to catch up with a major rival. A leading scholar, Kevin O’Rourke, states that the use of Denmark as a benchmark by which to judge Irish performance “was the universal practice among Irish agricultural reformers at that time.” 18
Denmark embraced cooperation as a means of exploiting the new technology of the continuous cream separator very soon after its invention, with the first cooperative formed in 1882. The number of cooperatives grew almost exponentially in the period leading up to the turn of the century. The Danish cooperative movement was very much a grassroot movement designing their own rules with little input from any central body. Denmark did not have a specific cooperative law and arguably did not get one until 1999. 19 Instead, cooperatives were based on the freedom to form associations and to fashion enforceable rules which would bind members. The rules of a cooperative are equivalent to terms in a contract and their enforcement depends on whether they would be enforceable by contract law. Generally, the Danish courts enforced the rules of associations so long as they were sufficiently clearly drafted that members knew what they signed up for. 20 That included terms that were restraints of trade and terms which dictated remedies for breach which were clearly penalty clauses.
The typical Danish cooperative was characterized by a commitment to supply all milk to the dairy for an explicitly stated finite period, typically 10 years. Some cooperatives had rules specifying the conditions under which exit could take place, but many explicitly barred exit during the period or made this at the discretion of the board. 21 This restraint of trade went some way to solve both holdup concerns. The dairy was viable because it would have sufficient ongoing supply of quality raw milk and for that reason investment in winter dairying was less risky. Cohesion of the supply was also secured through very significant damages in case of breach of the terms. 22 The penalties for refusal to supply were severe, typically a fine often combined with exclusion. Exclusion meant the loss of any claims on the assets of the cooperative, any share in the surplus of the current accounting year, and a requirement of paying the relevant share of any outstanding debt. 23 At least part of the costs of exit and in particular the payment of the debt and the forfeit of the assets are almost certainly a penalty rather than a measure of the damages caused by the exit. 24 Such penalties were enforceable, unless extreme.
Cooperatives very quickly became the dominant organizational structure to make use of the opportunities offered by the continuous cream separator. While private dairies had been formed before the first cooperative, private dairies generally struggled, and by 1900, there were five times as many cooperatives as private dairies. 25 While the reason for the dominance of cooperatives has never been conclusively explained, it could be argued that it relates to a residual level of trust required when milk was paid for according to quality. Suspicions about the veracity or fairness of the quality assessment may be difficult to allay solely through contracts. Cooperatives had members take turns in being present in the weighing in and testing of the milk delivered. 26
While the dominant organizational form in Denmark was the cooperative, in Ireland, the picture was much more mixed. The Irish cooperative movement started much later and the driving force was a private organization, the Irish Agricultural Organization Society (IAOS), which was supporting the cooperatives with help and advice, including on what the rules of the association should be. Ireland had a substantial private dairy sector throughout the relevant period. 27 Moreover, while the Danes focused almost exclusively on butter production, sometimes combined with cheese production, in Ireland the largest (private) dairy, the Condensed Milk Company of Irelands, also known as Cleeve’s, concentrated on the production of condensed milk. 28 This created additional competition for the raw milk supplied by farmers.
Legally, in Ireland, a cooperative is a society registered under the Industrial and Provident Societies Act (IPSA) 1893.
29
This provides the right to form an association for an economic purpose, but it also places some restrictions on the rules of the association. What rules for members would be enforceable also depended on what contract law would allow.
30
The early set of rules, designed by the IOAS, included a three-year commitment to supply all milk to the dairy.
31
We see this in the initial set of rules for Ballymacelligott, rule 20, which reads: Any member, who shall supply milk to any creamery other than that owned by the society, for the space of three years from the date of his admission to membership, without the consent in writing of the committee, shall forfeit his shares, together with all money credited thereon.
32
After the society shall have started a creamery for its members, no individual member of the society, so long as he continues a member thereof, who shall have milk to sell, the produce of a cow or cows kept or grazed on lands within the area defined in rule 5, shall, without the written consent of the committee first obtained, sell any such milk to any creamery other than a creamery of the society, or to any company, society, person or persons who sells milk, or manufactures butter for sale. Any member of the society committing a breach of this rule shall pay to the society, as and for liquidated damages, and not by way of penalty, the sum of one shilling per cow per day for every cow’s milk sold contrary hereto.
36
Rule 16 provides that a member whose shares have been transferred or cancelled thereupon ceases to be a member, but a member is not otherwise entitled to withdraw from the society. Rule 21 requires the consent of the committee to any transfer of shares but the committee are not bound to assign any reason for refusal.
37
The restraint of trade and associated penalties designed by the IOAS did not enable a successful switch to winter dairying.
39
“The chief difficulty…is the falling off of the milk supply during the months from October to April. During a large part of this period, the whole of the creameries are working at no more than half-time, and many of them are close for altogether for a month or two.”
40
In more recent research, Henriksen and O’Rourke argue that, [v]iewed from the outside, one of the most noticeable features of the Danish dairy industry was its year-round exports of butter. This contrasted sharply with the very seasonal production patterns common elsewhere. For example, in Ireland in 1909, creameries produced just 22 per cent of their annual output in the six months November–April; 45 per cent of annual output came in the three months June, July, and August.
41
A. Comparison of Outcomes
To understand the importance of the case to the cooperative movement in Ireland, it is important to understand that the impact of the continuous cream separator and the various adopted organizational forms of the dairies was very different in Denmark and Ireland: While the Irish dairy industry was larger than the Danish one as late as the 1870s, Danish butter exports exceeded their Irish counterparts in the late 1880s, and were almost three times higher by 1914…Both countries’ butter exporters were overwhelmingly reliant on the British market; Ireland’s share of that market dropped from over 50 per cent in 1860 to just 12 per cent in 1910, while Denmark’s increased from 0.6 per cent in 1860 to 37 per cent in 1914…Price evidence shows clearly that this loss of market share was The Danish dairying societies, recognizing that a regular supply of milk is a sine qua non of successful co-operative dairying, bind their members to send all their milk to the societies to which they belong under penalty of the payment of a heavy forfeit. As such agreements have been declared non-enforceable at law in Ireland, the dairy associations are called upon to take stringent measures to protect the great majority of the members from injury at the hands of the few.
45
B. But What Are the Competition Concerns?
Before looking more closely at the McEllistrim decision, it is worth spelling out some of the potential competition issues seen with today’s eyes. Firstly, as noted above, the majority in the House of Lords treated the arrangement simply as a vertical contract between supplier and purchaser imposing a restraint of trade, potentially foreclosing a rival dairy. 47 The extent to which this is a serious problem would, among other, depend on the size of the geographical market and how many dairies could be accommodated at a minimum efficient scale. With butter production and without extensive refrigeration, two factors in particular determined the geographical extent of a market for inputs. One was how early the milk had to be collected to ensure the same-day processing, and the other was how late the skimmed milk left over was returned to the members to be fed to pigs. 48 Similarly to modern cases regarding ready-mixed concrete, a lot would have depended on drive time. The importance of a guaranteed supply is clearly the greater, the fewer potential suppliers are located nearby.
The focus in this article, conversely, is on possible horizontal effects that arise from this ostensibly vertical arrangement. These effects arise because the terms of the vertical agreement to supply milk are based on a collectively reached agreement about the terms, decided among the farmer members of the cooperative, which are offered to individual members on a take it or leave it basis. In that regard, Denmark and Ireland offer two different but related concerns. In Ireland, the terms are drawn up by a central organization, the IAOS, which some might argue acts as the hub in coordinating farmer behavior. Others might argue that the IAOS acts as a trade association and engaging in legitimate standard setting. In Denmark, where rules were designed locally by the individual cooperative, the suppliers played the dual role as the designer of the restraint of trade and the potential victim thereof. In a parallel to the US Apple e-book case, where the publishers and Apple got together to restructure how books were transacted, the dairy and the suppliers got together and designed the collectively best contract, which in this case was the same for all.
In our understanding of whether there could have been a modern competition law infringement and if so what lessons it might have for us, it becomes important to appreciate whether the focus is on the individual restraint or the collectively reached restraint: that is, whether the problem is essentially vertical or horizontal in its focus. If the restraint is viewed as part of a broader horizontal agreement, it is possible that the restraint is ancillary to the creation of the nexus of contracts which make up the cooperative and in that case outside the reach of EU competition law. 49 As we shall see, the court did not agree on this. Secondly, even if the agreement breaches Article 101(1) TFEU (or the UK domestic equivalent contained in the Competition Act 1998), the term might nevertheless survive following an Art. 101(3) TFEU analysis.
III. Relevant Details from the Case
As a starting point, it is worth pointing out that McEllistrim was the culmination of a series of challenges to the rules of Irish cooperatives. The distaste for the vertical restraint in Irish contract law already existed as highlighted by the quote from McCabe above. 50 It was also, as pointed out by McLaughlin and Sharp, an intensely political case. 51
Ballymacelligott Cooperative Agricultural and Dairy Society Limited was a cooperative society formed in 1903, “for the purpose of developing and improving the industry of dairy farming in Ireland by carrying on the manufacture and sale of butter and cheese and other milk products from milk supplied by such of its members as had milk to sell.”
52
In 1915, they decide to expand by building a new dairy in Gortatlea for which they take up a sizable loan. Following suggestions by the IAOS, to which they were affiliated, they also put a new set of rules regarding the phrasing of the restraint of trade, discussed above, to a general meeting. The new rules were adopted by a vote of 987–146, representing 87% for and 13% against,
53
well in excess of the two-thirds majority needed by their old rules.
54
McEllistrim was 1 of the 14 who had voted against the change in rules. The claim by the appellant was that: the effect of the new rules was to make it impossible for members ever to withdraw from the society and to compel them to sell all their produce to the society at a price to be fixed by the society itself; and that such restrictions were ruinous to members and were unnecessary for the protection of the society’s interests.
55
The House of Lords judgment is not unanimous but is split four to one, with Lord Birkenhead, Lord Finlay, Lord Atkinson, and Lord Shaw of Dunfermline finding for McEllistrim, and Lord Parmoor dissenting.
Lord Birkenhead laid out the principle that “[a] contract which is in restraint of trade cannot be enforced unless (a) it is reasonable as between the parties; (b) it is consistent with the interests of the public.” 56 The question is whether the combined protection of the cooperative provided by rules 6(2), 16 and 21 is indispensable. Lord Parmoor put it slightly differently: “a contract in restraint of trade is not invalid so long as the restriction is no more than an adequate protection to the interests of the party in whose favor it is imposed.” 57
Arguably, the big difference between the analysis of Lord Atkinson and Lord Parmoor is what they view as the unit of assessment. Lord Atkinson looks at the individual contract between supplier and dairy, as evidenced, for example, by his focus on only one half of rule 6, discussed further in Section III(B). Thus, he approached the arrangement as a purely vertical relationship. Lord Parmoor explicitly looks at the nexus of contracts which makes up the dairy—that is, as a manifestation of the underlying horizontal coordination—and argues that it is legitimate for the cooperative to have rules which help it becoming a success as a cooperative. The rules hence need to be judged against the problem they are designed to solve. This difference in perspective may explain their different conclusions.
A. The Holdup Problem
In his testimony, Mr. Byrne, the secretary and manager of Ballymacelligott, was clearly aware of the holdup problem posed by the asset specific investment they had decided to make. In cross-examination, he said, “that the society would not take the risk of a new creamery without some guarantee that it would be supplied with milk.” 58 Subsequently, “[i]n re-examination he said that the necessary capital was borrowed from the bank on the guarantee of some of the members, and unless some such rule were adopted the members who had not guaranteed could leave the creamery without milk.” 59 In other words, some of the members had a much more significant level of liability and they would reasonably expect to have some sort of protection from a “run from the cooperative.” 60 Lord Finlay also appreciated the potential holdup problem arising from undersupply, as “no one could object to a provision, say, that before withdrawing his supply from the society a member should be bound to give the society reasonable notice.” 61 Generally, there is an appreciation of the need for some degree of commitment to supply in order to deal with the holdup problem.
Lord Atkinson pursues the tactics used by neighboring private dairies of offering what we would now refer to as teaser rates. “By the words ‘equal terms’ he [Mr. Byrne] referred to the fact that they found special prices had been offered to some members to take them away, and that by these words he meant ‘competition by the offering of special prices.’” 62 Mr. Byrne offers the rules of the cooperative as a protection against “the competition of those who may pay a special price for the purpose of killing the new branch in its infancy.” 63 Mr. Byrne is essentially referring to what we would now refer to as a strategy by other enterprises of “raising rivals costs.” This concern is lend credibility by the rather asymmetric structure of the processing market where one firm, Cleeve’s (the Condensed Milk Company of Ireland Ltd.), was clearly dominant as “the value of the share capital in Cleeve’s alone, at £350,000 in 1923, dwarfed that of the entire cooperative sector.” 64 Cleeve’s had dairies located in various locality which were in competition for the supply of milk with local cooperatives. Rather than understanding how the cost structure of the dairy creates problems when combined with predatory behavior by others, Lord Atkinson concludes that what the cooperative “desired to effect by these new rules was to protect themselves from the competition of proprietary creameries.” One might wonder whether today, the argument might be that the protection offered by competition law from an abuse of dominance would reduce the fear of opportunistic behavior by others.
B. The Tie-Up
The issue is the legality of rule 6, which specifies the restraint of trade, together with the rules regarding exit, rules 16 and 21. While rule 6 specifies both the duty of the cooperative and the suppliers, the majority focused solely on the latter, found in rule 6(2). Only Lord Parmoor considered rule 6 in its entirety.
65
Lord Finlay provided the most expansive assessment about the combined effects of rules 6(23), 16, and 21: But the combined operation of the rules to which I have referred goes far beyond what is reasonably necessary for the protection of the society. How can it be necessary that without any power of withdrawal a member should be restrained from supplying milk to other dealers unless and until he can get the consent of the committee to a transfer of his shares? Such a provision is in my opinion unreasonable and might be oppressive in the highest degree. It far exceeds anything required for the reasonable protection of the society’s business, and on the principles which have been repeatedly affirmed by this House and by the Judicial Committee of the Privy Council it is not enforceable.
66
Lord Parmoor questions whether the backstop is de facto infinite because there is no evidence of anyone trying to block exit. Apparently, McEllistrim had not sought to obtain the consent of the committee of management before the commencement of this expensive litigation. There is no reason that it might not have been given in the case of the appellant, and there is evidence that whenever a member has applied for a transfer it has been granted.
68
Lord Parmoor is fairly explicit in his view that the others have not understood the purpose and nature of cooperatives.
69
Showing a deeper engagement with the purpose of the cooperative, he argues that the relationship between a co-operative society and its members is similar not to the relationship between a producer and an ordinary wholesale purchaser, but to the relationship of a partner to his partnership firm, and that it is a fallacy to regard the impugned restrictions as though they had been imposed upon the vendor at the instance of an outside wholesale merchant, purchasing solely for his own benefit.
70
The Law Lords all argue that some level of commitment to supply is desirable and it is arguable that they would have been prepared to countenance a restraint of trade for a fixed period. This follows from the Tipperary case. 71 But whether this was for 6–12 months or more is not clear from the decision. From the Tipperary case, it would appear that there is no clear guidance from the court as to what would constitute a convincing justification for a particular length of the restraint.
C. Key Points from the Case
The first point to make from the case is that the contracts in question are, on their face, vertical contracts containing a vertical restraint of trade. In the judgments, therefore, the only contentious issue is whether or not the contracts should be analyzed individually as purely vertical restrictions, as done by the majority, or should be analyzed collectively as an expression of the underlying horizontal coordination, as done by the minority.
The second point is that the purpose of the restraint of trade (and for that matter the penalty clauses) is to solve holdup problems which otherwise would either preclude the establishment of dairies or reduce their efficiency significantly. The choice of contractual terms appears to be motivated by efficiency concerns and concerns about unfair competition from others that may exacerbate any holdup problem.
The third point is that the content and wording of these contracts may have been the result of a “conspiracy,” in that the rules of the cooperative are set collectively and were the same for all. In the Irish context, to the extent that anything cartel like is going on, this must relate to the activity of the IAOS. It is the IAOS who designs the rules and more importantly, it appears as if the IAOS supports the cooperative dairy cases both with advice and financially. Are we looking at a concerted practice promoted by the IAOS? Alternatively, if the vertical contracts are standard form contracts, is this a case of (tacit) collusion among the local farmers?
The fourth point is that in the Irish context, we need to keep our mind on the market for raw milk, which is concentrated and which produces other outputs than butter (and cheese). There is potentially significant buyer power among the large private condensed milk producers. However, while some processors of milk may have had monopsony power, the market for the final product was very competitive. The change to commercial butter making in Denmark through the creation of cooperatives employing the continuous cream separator increased both the amount and quality of butter produced, and the branded Danish butter in addition fetched higher prices on the English market than had previously been the case. The effect of the restraints on price, quantity and quality, must have increased consumer surplus. Had Irish farmers been able to emulate the Danes, the expansion in the quantity of high-quality butter would have been even greater with a likely dampening effect on final butter prices. Thus if, with modern eyes, the agreement violates Art. 101(1) TFEU, it might also be seen to have beneficial effects on efficiency and consumer surplus.
The final point is that the questions asked by the court in McEllistrim are similar to the questions we would ask both in assessing whether the restraint is ancillary and whether it would satisfy the criteria in Art. 101(3) TFEU. Is the restraint essential? Is the restraint in the public interest (i.e., does it increase efficiency and will consumers get a share of the resulting surplus)? Below, we shall argue that the answer to the first question rests on whether or not the restraint was de facto indefinite and if not, whether it was for and unnecessarily long period. The answer to the second question is a clear yes.
IV. Giving McEllistrim a Modern Competition Treatment
Cooperatives continue to exist and are in some EU member states still the dominant organizational structure in some sectors, in particular in the processing of agricultural products. Today, such sectors are not just subject to contract law which has traditionally had a view on restraints of trade, but also to competition law. What then can we learn from the McEllistrim decision?
In this section, we focus solely on the restraint of trade binding the supplier to supply all their milk to the cooperative so long as they remain a member. An agreement in restraint of trade on the face of it would look like an obvious candidate to infringe Art. 101(1) TFEU. However, it is still necessary to assess whether Art. 101(1) is in fact engaged or whether instead the restraint is ancillary. First, consider the connection with recent cases looking at a mix of vertical and horizontal effects which have been considered by competition authorities.
A. Horizontal Agreement with Vertical Effects
The abstract version of McEllistrim is one of N farmers who get together, possibly with the aid of a trade association, to agree on the wording of a common vertical contract with a dairy, which they will all sign. The dairy is a collective of the same farmers, who have a shared right to the surplus of the dairy according to cooperative principles, that is, relative to their transaction with the dairy. Among the term of the vertical contract is a commitment exclusively to trade with the processor for a specified number of years.
The abstract case contains both horizontal and vertical elements. The horizontal element is the agreement among suppliers to use the same standard contract which has, as one of its effects, a restraint of trade which would deprive any outsider of supplies from the farmers who have entered the vertical agreement. The agreements commit the collective not to supply others. The vertical agreement is the agreement to which each member signed up. This is a purely vertical agreement, which on its own, unless the particular supplier has a very substantial share of the total locally available supply, will have very limited effects on other users of the product supplied to the dairy.
There are some logical problems with analyzing the case. For example, the cooperative does not exist until the members who were party to designing the vertical contract have actually signed up in sufficient numbers. Thus, while the cooperative dairy is a separate legal entity (both in Denmark and in Ireland), it could not meaningfully have been part of any conspiracy. The group who initially signed the contracts may not include all those involved in the “horizontal conspiracy” to write the contract. Identifying who might have violated competition law may hence not be a simple matter. For those who were not actively involved in the design of the contract but who joined the cooperative later, they clearly were not part of any horizontal agreement but merely signed a vertical contract with potentially minimal exclusionary effect. 72 Imagine, for example, that whether or not a farmer who does not sell to a creamery joins a new creamery, which may then be the difference between whether or not this alternative dairy is viable. Not allowing the farmer to join the established cooperative would seem an extreme intervention in their freedom to trade. Another issue relates to any subsequent revisions to the rules and by implication the vertical contracts. If passed by the required majority of members, such a revision would legally bind any minority who had voted against. Would the action of having voted against be sufficient to absolve the minority in the participation in a concerted practice? Would the ability to provide evidence turn on whether or not the vote was by secret ballot? 73
The type of case above is arguably about a horizontal agreement between suppliers on a contract which involves vertical restraints with horizontal effects. Similar issues can arise in vertical agreements that combine agency pricing with platform parity or retail Most Favored Nation (MFN) clauses. Such agreements have recently been challenged by various competition authorities in markets such as online hotel booking, 74 e-books 75 , and insurance. 76 The impetus for the vertical agreement can in theory come from the upstream or downstream market. An interesting comparison arises where at least the upstream firms are involved in designing the vertical agreement. As an abstract example of this type of cases, consider L publishers getting together to take control of price setting. They do so by agreeing to insist on an agency arrangement with the retailers, in return for which the vertical contract will also include a retail MFN to ensure that the retailers are not used as vehicles for price discrimination. This example is close to the U.S. E-book case where the focus was on the conspiracy element—the publishers and Apple getting together to agree on standard terms. In the platform cases, including e-books, the isolated effect of the constructed vertical contract may have anticompetitive effects. The motivation for the contract is more difficult to be sure of. It has been argued that the key reasons for obtaining control over retail pricing related to externalities arising from retail platforms using e-books as loss leaders. Two separate effects from that could have been a legitimate concern. One is that a greater pool of locked-in consumers attracted to the platform due to the low prices could have an adverse effect on future bargaining between publishers and platforms. Another that the price of e-books could affect the reference price of hardcover books, a market with both greater volume and margin. So far, the potentially pro-competitive arguments have not found favor with competition authorities. 77 The possible difference between the E-book case and a modern version of McEllistrim is whether the key motivation is legitimate or not.
B. Ancillary Restraints—Is Art. 101(1) Even Engaged?
Recall, as per Lord Parmoor, that the purpose of the restraint is to solve a problem which would undermine, potentially fatally, a cooperative dairy, thereby potentially leaving the milk producers with no one to supply or facing poorer terms of trade and depriving the economy of income from exports. If we take a holistic view of the restraints and look at their combined effect, it may be possible to argue that the vertical agreements were ancillary restraints required to realize the objectives of the underlying horizontal coordination.
The obvious case to consider relates to a Danish cooperative, Gøttrup-Klim Grovvareforening v. Dansk Landbrugs Grovvareselskab.
78
The exclusivity arrangements, which the cooperative Dansk Landbrugs Grovvareselskab had with its member, were challenged. It was argued, in response, that the restraints did not have the purpose of restricting competition but rather of enabling the members, through the membership, to exercise countervailing buyer power against large multinational producers of fertilizers and other products. The Court of Justice of the European Union agreed, holding that: [A] provision in the statutes of a cooperative purchasing association, forbidding its members to participate in other forms of organized cooperation which are in direct competition with it, is not caught by the prohibition in Article [101(1) TFEU], so long as the abovementioned provision is restricted to what is necessary to ensure that the cooperative functions properly and maintains.
79
How we view rule 6 of McEllistrim depends on what we think the main transactions is. If we take the main transaction the organization of a successful cooperative butter making facility, which would be more in the spirit of Lord Parmoor, then arguably it is subsidiary and in some form also necessary. The only argument would be how long the restraint needed to be in place. An indefinite restraint without a unilateral exit possibility is unlikely to pass the second part of the test, but a finite restraint, or a restraint which binds as long as the supplier is a member would do so. On this interpretation of the main transaction, Art. 101(1) is not infringed.
If, however, the main transaction is the supply contract between a farmer and a dairy (i.e., if this really is simply a vertical arrangement), which arguably is how Lord Atkinson would see it, then it would be hard to argue that the restraint is ancillary. The same would be the case if the length of the restraint was deemed excessive. In that case, Art. 101(1) is infringed and we move on to ask whether Art. 101(3) would exempt the restraint of trade. 82
C. Is the Restraint of Trade Exempt by Art. 101(3) TFEU?
There are four criteria (efficiency gains, fair share for consumers, indispensability of the restrictions, and no elimination of competition) which have to be met in order for an agreement to be rescued by Art. 101(3) TFEU. In this section, we look at each of these in turn, using the Danish experience as the counterfactual. It is worth noting that just as in the case of an ancillary restraint and in the analysis by the House of Lords, the key question remains: is the restraint essential?
1. Efficiency Gains
The efficiency argument is probably the easiest to argue. Introduction of the continuous cream separator clearly was efficiency-enhancing since more, better quality, butter could be manufactured. It was well established that in Denmark, minimum efficient scale required milk from approximately 400 cows. The number of cows required would not have been smaller in Ireland. Given the substantial fixed costs involved, to make efficient use of the creamery, the raw milk would have to be available year-round.
This does not imply that cooperatives were necessarily the best solution. One of the challenging questions for Denmark has always been where the source of domination of cooperatives over private firms came from. We know that private dairies also wrote long-term contracts with terms not dissimilar from the rules of the cooperatives. The following story about a private creamery at Snogbæk, set up by Jørgen Jensen in 1879, three years before the first cooperative, illustrates this: Jørgen Jensen created a 10 year contract with the farmers in Øster Snogbæk that they would supply their milk to him. The agreement was presented in 11 paragraphs and took into consideration both the interests of Jørgen Jensen and the suppliers, as was the case in similar agreements from other creameries at this time. However, there are parts of special interest - for example in §5: the supplier commits himself to organize the calving such that at least 1/3 of the herd calves before the end of November. He similarly commits himself not to feed the cows with other root crop than mangel-wurzel and carrots. The daily amount must not exceed one “skæppe” (about 17.39 liters)…. Suppliers have to submit to fairly rigorous conditions from Jørgen Jensen. But they accepted these and more and more joined up as suppliers.
83
[T]he producer would always distrust the exactitude of the control apparatus, and, as this can only give an approximate result, when it is below the truth he suffers irreparable loss. Under the co-operative system, on the contrary, the producer recovers in the annual dividends compensation for this temporary loss. Moreover, this loss itself is distributed equally amongst the producers, who are the only shareholders; for if the apparatus does not indicate exactly the quantity of cream contained in the sample of milk, it does indicate without any error the proportionate richness in cream of the different samples tested at the same time.
85
2. Fair Share for Consumers
Because this is a market where the product is produced in one country and sold in another, it matters whose consumer surplus we are looking at. The Danish consumers may not necessarily have gained as the best quality was exported to England. The quality would have improved but the price would also have risen. However, if we focus on the biggest market for the butter, England, then the effect is unambiguously positive. The success of the Danish cooperatives introduced a completely new player to the market. Commonly, we would expect more intense competition to result in a better offer to the consumer. Much of this may have been in terms of better quality, but not exclusively so, since the Danes would have initially flooded the market with high-quality butter and we would expect prices as a result to have fallen. Given the effect of the Danish success, a similar boost to output and quality in Ireland is most likely to have led to higher quality and lower prices and hence increased consumer surplus.
3. Indispensability of the Restrictions
As per the European Commission’s Guidelines on the application of Art. 101, two conditions need to be met for an agreement to be indispensable:
87
The agreement must be reasonably necessary in order to achieve the efficiencies, and the specific restriction must also be reasonably necessary for the attainment of the efficiencies. As argued above, both private and cooperative dairies in Denmark used long-term contracts with restraints of trade, typically for a 10-year period with often no, little, or expensive opportunity for early termination of the contract. That large suppliers could throw their weight about is illustrated by a case where the threat of exit was used proactively in order to influence the decision of the creamery. A letter of September 1890 from a large dairy farmer to the board of a newly started cooperative is illustrative of the latter: By checking the fat content of my milk and by comparing it to that of other cooperative members I have become aware of the fact that the fat content of the milk supplied from many of the members is so low that it can only be explained by adulteration or by other kind of inadmissible conduct. I assume that the esteemed board will act vigorously upon this information now that I have called its attention to it so that it will not be necessary for me to carry the matter any further.
88
It naturally all depends on what we mean by essential. That a cooperative with a 10-year loan and unlimited liability needs a 10-year restraint of trade may not be controversial, but there are other ways to organize the making of butter. Many predicted that in Denmark after the first loan period, cooperatives would restructure as private enterprise. The capital equipment was in place as was a set of competent dairy men to manufacture the butter. The reverse seems to have happened. While most Danish cooperatives formed for another loan period (and did so over and over again until a much later consolidation), many share companies converted to cooperatives at the earliest convenient moment. At renewal, the Danish cooperatives typically took up new loans, passing the capital on to the members, necessitating a repeat of the restraint of trade. The popularity of the restraint can possibly be explained by the way in which exit was typically handled. 90 The cost of imposing the restraint seems slight compared with the potential cost of the demise of the cooperatives.
However, it is hard to argue that the restraint would have to be longer than the danger imposed by strategic holdup, that is, the natural life of the project. 91 A challenge for Ballymacelligott in particular and the IAOS rules more generally was the indefinite nature of the tie. No one seemed able to offer a clear justification by the IAOS or its members. The IAOS and in particular its leading lights, Robert Andrew Anderson and Horace Plunkett, were well aware of the Danish cooperatives rules and could easily have copied these. The limited liability, which characterize the Irish cooperatives, may have harmed them since this weakens the downside risk for the members. However, the apparent hostility to cooperative ventures from people such as Lord Atkinson would not have helped.
4. No Elimination of Competition
When it comes to this part of the test, the restraints would potentially be more in trouble in Denmark than Ireland, since in Denmark, the cooperatives outperformed the private dairies, while in Ireland, private dairies thrived. If the issue is the survival of alternative forms of production, then if anything, it is the Danes who would be more likely to fail the test. At the same time, rather than the number of commercial butter producers falling in Denmark, the number of dairies rose very rapidly during the heyday of the restraint. If we interpret this part of the test as relating to as-efficient competitors, neither the Danish nor the Irish cooperatives managed to remove as-efficient dairies.
D. Broader Issues
Modern firms bring together a number of actors to source the necessary capital, labor, and raw material. Some of these actors are inevitable horizontal rivals. Hence, a more fundamental issue, which chimes with the initial quote from American Needle, is whether those involved in providing the resources involved in producing something can do so freely without the interference of Art. 101(1), or they have to rely on Art. 101(3) and the benefits from the production being sufficient to overrule any concern about potential horizontal cooperation.
Alison Jones offers a clear exposition of what constitutes a firm for the purpose of EU competition law
92
but also acknowledges that “it may not always be easy to identify an undertaking and its boundaries.”
93
The key competition law question is whether an “agreement is “within” (only Art. 102 TFEU may be engaged) or “outside” (Art. 101 TFEU is engaged) a firm.
94
An important element of any definition appears to be independence: The CJ has also held that, in the context of parent-subsidiary relationships, entities will constitute a single economic unit if a subsidiary “enjoys no economic independence” or if the entities “form an economic unit within which the subsidiary has no real freedom to determine its course of action in the market” but carry out the instructions issued by the parent company controlling them.
95
Another key indicator relates to risk. Hence, “the agent who does not accept the risk of the principal’s business seems to be viewed not as an entity engaged in economic activity but as incorporated within the undertaking, providing services for and at the direction of the principal.” 97 The discussion of risk appears to use the principal-agent model, but without the formality imposed by economists. Consider the simplest standard economics P-A model. 98 The principle has a project, which she cannot work on. The project is offered to the agent, who can. The outcome of the project is uncertain but affected by the efforts put in by the agent. Ex post, the principal can observe the outcome but not the effort of the agent. A particular outcome may hence be the consequence of good or bad luck on the one hand and high or low effort on the other. The principal does not have the time or ability to monitor the agent. Clearly, if the agent is offered a flat fee for carrying out the task, the agent has no incentive to work hard. An obvious alternative is what amounts to selling the project to the agent, who will keep the whole surplus in return for a flat fee payable to the principal. This alternative puts the entire risk of the project on the shoulders of the agent. If the agent is more risk-averse than the principal, this allocation of risk is not efficient. Hence, the typical principal agent contract is based on a trade-off between incentive provision and risk-sharing. To help the discussion, let α be the agent’s share in the output. Clearly, the higher the α is, the more risk the agent assumes. Now, consider the standard competition law approach to deciding on whether or not the agent is a separate economic entity. Clearly, where the agent assumes all the risk, α = 1, the agent is an undertaking. If we are taking the case law seriously, then equally clearly the agent who takes on no risk, α = 0, is an employee and hence not a separate economic entity. The implication is that there must be some value of α where the agent switches from being an employee to being a separate undertaking. The difficulty in explaining how the value of α is to be determined should either lead to a rejection of risk allocation as a way to resolve the question of what is an undertaking, or to a very careful outlining of what risks are referred to.
Jones proposes “a narrower theory of an economic unit, confining it to parent/subsidiary relationships where the former has sole control over the latter and the power to control positively the subsidiary’s strategic behavior in the market.” 99 While that sounds very sensible, the issue of how to operationalize this still seems challenging.
In the context of the cooperatives, both the suppliers of raw milk, the dairy and even the manager of the dairy took on risk when they set up and ran a cooperative. The suppliers were subject to a large degree of command and control from the cooperative, while the manager had considerable freedom to act and faced associated risks through sharing in the success (or failure) or the dairy. The current case law understanding of what constitutes a firm (or an undertaking) does not fit terribly well with the sort of cooperatives created particularly in Denmark in the late nineteenth century. This seems an area that deserves further thought and research.
V. Conclusion
The McEllistrim case is about a vertical restraint of trade, which is essential for the success of an effective introduction of new technology. The vertical restraint is part of a series of standard contracts, which together makes up the nexus of contracts which we refer to as the dairy. The question mark hanging over the restraint is whether it is in reality finite. What does appear to be true is that it is a case of poor drafting, not by the cooperative itself but by the IAOS.
The term is on paper a violation of competition law and it is hence interesting to speculate on how the case would run as a modern competition law case. One could argue, and this would follow the arguments of one of the Law Lords fairly closely, that the restraint is ancillary to the underlying horizontal cooperation and hence competition law is not violated. Alternatively, one could take the approach of the majority of the Law Lords and focus narrowly on the vertical contract, in which case Art. 101(1) would be violated and it then all comes down to Art. 101(3). We have argued that the only possible stumbling block is whether or not the restraint is of a sensible finite length, long enough but no longer that necessary to serve its purpose. Here again, it depends on which Law Lord one is prepared to follow. Since the majority showed little sign of understanding the cooperative principles, it depends on how important one views the cooperatives in the wider economy. The shift in dominance on the English butter market from the Irish producers to the Danish producers might tip the balance in favor of the minority, Lord Parmoor. The case was always about the legality of a term in a vertical contract necessary to create an efficient organization for the purpose of production. As such, the case has more to tell us about the theory of the firm than competition. It does serve to remind us that there is a potential conflict between company law and competition law and that Art. 101(3) has potentially a lot of lifting to do to ensure that we do not throw efficiency enhancing innovations out.
The case also reminds us that in many situations, the case can be argued either as a competition case or a contract case. We saw above that in reality, there is very little difference between the outcomes from the two approaches. However, given the greater familiarity of the courts with contract law, it might be worthwhile to focus on bringing the cases as such.
Footnotes
Author’s Note
All errors and shortcomings are the author’s.
Acknowledgments
The author would like to thank Eoin McLaughlin, Sebastian Peyer, Niamh Dunne, and in particular Elias Deutscher for helpful comments and discussions, as well as participants at the workshop “The ‘Acceptable’ Cartel? Horizontal Agreements under Competition Law and Beyond” (London School of Economics, Mar. 22, 2019) for helpful comments.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.