Friday, March 12th, 2004 by Juan Briones

The ongoing process of modernization of the EU Competition policy has already had significant effects in most areas from both a procedural and substantive point of view.

The importance of economic analysis that the Commission necessarily had to develop to handle mergers has now permeated entirely the other areas of competition policy, and most notably vertical restraints.

Entirely? Well, not entirely! One small group of indomitable lawyers still holds out against economic analysis from the muddy trenches of Article 82.

Judgements of the CFI.

Last autumn, the CFI adopted two judgements [ 1 ] (“Michelin” and “BA”) relating to abusive loyalty discount schemes. Both judgements refer to a consistent case law that can be traced back to 1973 (Suiker Unie v. Commission) but that in fact had not been applied to my knowledge since 1993 (BPB Industries Plc and British Gypsum Ltd v Commission).

The Michelin and BA judgements are mainly concerned with the loyalty schemes themselves (and consequent discrimination aspects). Unlike previous cases, there are no issues of exclusive purchasing or selective foreclosure of market access to rival firms.

A subject of major concern is that in both cases, loyalty schemes are treated as a prohibition per se if firms are found dominant. The same idea underlies the initial decisions of the Commission condemning Michelin and BA. Both the Commission and the Court make abstraction of any meaningful market analysis.

Both companies were loosing market share, but it is considered that they would have lost even more in the absence of the loyalty schemes. Complainants are one distributor (among all the distributors that were being allegedly discriminated) in the Michelin case and one competitor in BA, not any discriminated travel agent. The competitors’ market share was by the way on a raising trend.

In both cases, the Court opts for a per se rule and does not consider necessary to assess the impact of the loyalty rebates on the market or on competition. The Commission does consider briefly possible foreclosure effects in its decisions, but stating clearly that those effects are not required neither to reach a negative decision nor to assess the gravity of the infringement. .

At worst, the effects on competition of loyalty schemes with retail distributors can be assimilated to exclusivity clauses in vertical contracts. It appears somewhat contradictory to maintain the traditional policy and an absolute per se prohibition with regard to loyalty discounts now that the Commission has acknowledged the complexity of vertical relations between firms.

Is dominance sufficient to justify such a divergent approach? It depends. A necessary condition is of course that the relevant market and the dominant position are established taking into account the complexity of vertical relations, and above all, the way competition works between companies with different degrees of vertical integration.

Vertical integration and firm boundaries.

Vertical relations between two firms are much more complex than the relation between a firm and end consumers. In the first case, there are two agents taking decisions on prices and sharing of profits, allocation of risks, sharing the burden of promotional effort, providing the incentives to provide an optimal quality of service, and deciding on the adoption of new technologies.

On the second case, only one agent decides. Moreover, retailers compete with each other, whereas consumers do not.

These decisions are taken in the context of interdependence of manufacturers’ and retailer’s profits, potentially incomplete contracts and potentially significant transaction costs, often in a situation of asymmetrical information. Market imperfections in the form of externalities [ 2 ] between the manufacturer and the retailer are almost always present, and they are at the heart of the defendants that vertical restraints enhance overall efficiency more often than not.

Considerations about all these variables, as well as considerations internal to the firm will determine whether a firm decides to integrate downstream or not. Firms will typically balance the pros and cons of coordinating internally a number of functions (internal planning), or letting the market carry out that function (market prices will base decisions). The decision depends on balancing whether the firm finds more profitable internal coordination than purchasing in the market a distribution service.

Since not all firms have the same management, or the same information or the same perception of risk, it is not unusual to find competitors with different degrees of vertical integration, or competitors combining a mixture of vertical integration and vertical contracts with third parties. This is the situation in the Michelin case.

Technological change may substantially alter the economics of vertical integration and vertical arrangements. The rapid development of Internet connections has had an impact on the marketing of air transport services, and air carriers have reacted by selling increasing amounts of tickets through Internet platforms. The remuneration and the margins of travel agents have become as a consequence under pressure. This trend was starting in the BA case.

Tyres and air tickets.

The competitive process in these two markets appears to be substantially different, and the incentive schemes supplied to distributors to grow sales volume and market share have different implications for competition.

For one thing, distributors of tyres can build inventories when prices appear attractive, whereas the transport service cannot be stored. These inventories will then compete with the sales of the manufacturer through its own integrated distribution network. The pricing of tyres at wholesale level and any discount scheme for tyres has to accommodate this fact

A second, more important difference between the distribution of truck tyres in France and air tickets in the UK is that the economics of downstream integration appear to be quite different.

In BA, the competing carriers in the UK do not have any significant downstream integration into travel agencies. In the absence of any exclusivity, all have access to air travel agents, and will compete to develop sales either though travel agencies or through Internet. Although BA granted fidelity rebates, it was at the same time reducing the overall remuneration to travel agencies.

In Michelin, like in BA, there is no question about exclusivity or resale price maintenance. But the degree of vertical integration and the complexity of vertical arrangements is substantially higher.

In the French market for truck tyres, about one quarter of sales consist of direct sales from manufacturers to large fleet owners, truck garages and dealers. The remaining is sold through specialised dealers. Of 2200 sales outlets, about 500 belong to integrated manufacturer’s networks. All the significant manufacturers are vertically integrated into distribution to varying degrees. Integrated networks, at least the Michelin network, do market competing brands. We can assume that inter-brand competition existed also in the competitors’ network.

Increasingly, manufacturers are engaged in extending their respective degree of downstream integration in a race for acquiring points of sale. This constitutes an important aspect of the competitive process.

Finally, independent dealers are consolidating into larger networks, looser franchise networks or buying associations (which happen to account for one third of the market).

In other words, there are at least five different routes to the market: own sales network, independent networks, buying groups, competitors’ network and direct sales.

The main tyre manufacturers in France are all large groups that compete not only in quality, prices, advertising and service, but also compete to achieve an efficient distribution. However, there are a number of different options and all manufacturers use a combination of competing distribution channels.

Presumably, all manufacturers offer price discounts and promotional budgets to independent distributors to increase the level of service and advertising for their products at the points of sale. And presumably, all provide incentives to distributors to grow their respective sales volumes and market share.

Unfortunately, neither the Commission nor the Court seems to have given importance to this issue. Since other firms cannot be dominant, from a strict legal point of view, there is no reason to examine in detail what the competitors’ commercial practices are.

But from an economic point of view, it is of utmost importance to establish if the incumbent is behaving differently from its competitors to assess the degree of dominance in the first place.

A mere computation of market shares and the identification of some “comparative advantages” (technical lead as a result of higher investment in R+D, better products with longer life time, larger sales force) should not suffice to establish dominance anymore [ 3 ].

Consequences of per se prohibition of certain discount schemes.

What can an incumbent with a 50% plus market share do to react to losses of sales volume and market share? Does it just have to sit and wait until its share falls below the presumption of dominance, and then reintroduce so-called loyalty schemes, to be in the starting situation again?

Managers are under an obligation to grow the sales of their companies and will provide incentives to distributors to sell more of its products if the market and competition so requires. There are different options and so-called loyalty schemes are probably the least anticompetitive option among the various options offered by vertical arrangements.

But if this option is closed, other ways will be used. In my view, the per se prohibition of loyalty schemes will lead the market leaders, such as Michelin, to increase their degree of downstream vertical integration.

When retailers are so crucial to sale volumes and the margin of manoeuvre of the market leader is severely limited by competition policy, the balancing decision as to the optimal boundary of the firm will be severely affected.

Also, the competition between different channels will be distorted, because of an additional incentive to vertically integrate. The market leader, who is exposed to the sanction under Article 82, would start the process, and if it decides to eliminate inter-brand competition in its network, competitors will no doubt follow suit.

What are the effects in the medium and long run? Firstly, competition authorities will have shaped the structure of a market in full abstraction of consideration of efficiency and competition. Secondly, independent distributors will be progressively marginalised, clearly an undesired effect for a competition authority.

Favouring the development of vertical integration against other distribution channels will thus ultimately raise barriers to entry in the market for truck tyres.

In BA, the case for this result appears to be less clear, but this only because of differences in the competitive process in the market for air travel. The case for airlines integrating downstream into travel agency appears much more dubious, as reflected by the lack of vertical integration of competitors in the UK.

In all probability, the capillarity required to retail travel services and the fact that travel agents typically market a wider range of services to their clients, discourage air carriers to integrate downstream. The costs of coordinating these activities within the firm may overcompensate any advantage arising from a full control of the point of sale.

If this may be true in air transport services, it does not necessarily hold in other markets. Only economic analysis and relevant economic evidence about the competitive process and the role therein of distributors can tell the difference.

The policy question that is raised by comparing the BA and Michelin cases is that similar rebate schemes may have very different impact on different markets, and that forbidding certain schemes across the board may have very different consequences on competition, and even have anticompetitive effects.

There is one thing that strikes the attention when reading the decisions of the Commission and the judgements of the CFI in these cases: there is very little reference and almost no analysis of the competitive process or of the characteristics and capacities of competitors.

The analysis focuses on the characteristics of the rebate scheme in itself, and if it fits in the model developed in the constant jurisprudence, in that they create strong pressure at the margin to increase the sales of the market leader, the scheme constitutes an abuse.

The analysis is entirely legal and ignores the effective impact of the scheme and the effective impact of its prohibition.

Entirely? Well, for all practical purposes, yes.

This entry was posted on Friday, March 12th, 2004 at 10:35 pm and is filed under e-konomica articles. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.

Previous: Price dispersion and competition policy | Next: Commission develops new policy for predatory pricing in network industries.