IMPORTANT LEGAL NOTICE - The information on this site is subject to a disclaimer and a copyright notice.
The Economics of Verticals
Luc Peeperkorn, DG IV-A-1
Competition Policy Newsletter 1998 - number 2 - June
As recognised in the Green Paper on Vertical Restraints in EC Competition Policy (Green Paper), the economic analysis of vertical agreements has in the past been the subject of heated debate between economists. (1) By the early 1980s the position had swung from regarding them as suspect for competition, to a generalised perception that they were innocuous for competition. Nowadays there is a new emerging consensus and economists are becoming more cautious in their assessment of vertical agreements and less willing to make sweeping generalisations.
A first and central element of this new consensus is the importance of market structure in determining the impact of vertical agreements on competition. Economics tells us that in the field of vertical restraints competition concerns can only arise if there is insufficient inter-brand competition, i.e. if there exists a certain degree of market power. On the one hand, the fiercer inter-brand competition, the more likely it is that vertical restraints have no negative effect or at least a net positive effect. On the other hand, the weaker inter-brand competition, the more likely it is that vertical restraints have a net negative effect. This means that the same vertical restraint can have different effects depending on the market structure and on the market power of the company applying the vertical restraint.
This of course raises the question of what is meant by market power? In economics market power is usually defined as the power to raise price above the competitive level (in the short run marginal cost, in the long run average total cost). In other words that a firm by changing its output has perceptible influence on the price at which it can sell and that by charging a price above the competitive level it is able, at least in the short term, to obtain supra-normal profits. Most economists would agree that there already exists market power below the level of dominance as defined by the Court of Justice. This view was also expressed in the Green Paper, to indicate that vertical restraints can harm competition below the level of dominance and therefore that Article 86 and merger control will not suffice and Article 85 needs to be applied to vertical restraints. (2)
A second element of this new consensus is that it is generally recognised that vertical restraints are on average less harmful than horizontal competition restraints like price fixing or market sharing. The main reason for treating a vertical restraint more
leniently than a horizontal restraint lies in the fact that the latter may concern an agreement between competitors producing substitute goods/services while the former concerns an agreement between a supplier and a buyer of a particular product/service. In horizontal situations the exercise of market power by one firm (higher price of its product) will benefit its competitors. This may provide an incentive to competitors to induce each other to behave anti-competitively. In vertical situations the product of the one is the input for the other. This means that the exercise of market power by either the up-stream or down-stream company would normally hurt the demand for the product of the other. The companies involved in the agreement may therefore have an incentive to prevent the exercise of market power by the other (so called self policing character of vertical restraints).
However, this self-restraining character should not be over-estimated. When a company has no market power it can only try to increase its profits by optimising its manufacturing and distribution processes, with or without the help of vertical agreements. However, when it does have market power it can also try to increase its profits at the expense of its direct competitors by raising their costs and at the expense of its buyers/consumers by trying to appropriate some of their surplus. This can happen when the up-stream and down-stream company share the extra profits or when one of the two imposes the vertical restraint and thereby appropriates all the extra profits.
In this article I will not go into the question when is inter-brand competition weak and when is market power present. This can only be assessed on a case by case basis. One could only observe that, as most markets are rather competitive, in many instances vertical restraints are unlikely to have significant negative effects. A large majority of actual vertical agreements are therefore unlikely to be of interest from a competition policy point of view.
The success of any future policy will depend on it being able to provide a sufficiently wide and well defined safe harbour to exclude this large majority of agreements from competition policy scrutiny. In addition its success will depend on the predictability and clarity of policy towards agreements that are outside the defined safe harbour. It is in this light that in this article an economic classification of vertical restraints is presented. This classification is based on the main negative effects that may result from the different vertical restraints. It also describes the positive effects linked to vertical restraints. Finally it draws some general conclusions.
The final formulation of policy will of course not only depend on the economics of vertical restraints. It will also and most importantly be determined by the choice of policy objectives, the effects on enforcement costs, the effects on legal certainty for industry etc. For example, the assessment of a resale restriction may be very different when not only the protection of competition but also the goal of market integration is considered. However, these issues are also not covered in this article as it concentrates on the economic classification of vertical restraints.
2. The negative effects
2.1. Individual vertical restraints
To analyse the possible negative effects it is appropriate to divide vertical agreements into four groups: an exclusive distribution group, a single branding group, a resale price maintenance group and a market partitioning group. The vertical restraints within each group (as opposed to between these groups) seem to have similar negative effects on competition.
Before describing the four groups a number of general points need to be made. Firstly, the analysis applies to both goods and services, although certain restraints are mainly used in the distribution of goods. This is why throughout this text the term good(s) means both good(s) and service(s) unless otherwise stated. Secondly, vertical agreements can be concluded for intermediate and final goods and services. Unless otherwise stated the analysis and arguments in the text apply to all levels of trade and the neutral terms supplier and buyer are used. When only a specific level is implicated this is indicated. Thirdly, the termino-logy used may confuse some as it at times differs from the current legal definitions. The classification is based upon what could be described as the basic components of vertical restraints. In practice many vertical agreements make use of more than one of these components. To give an example, exclusive distribution is usually limiting the number of buyers the supplier can sell to and at the same time limiting the area where the buyers can be active. The first component may lead to foreclosure of other buyers while the second component may lead to price discrimination.
Exclusive distribution group
Under the heading of exclusive distribution come those agreements/components that have as their main element that the manufacturer is selling only to one or a limited number of buyers. This may be to restrict the number of buyers for a particular territory or group of customers, or to restrict the kind of buyers. The group comprises exclusive distribution and exclusive customer allocation as the supplier is limiting its sales to only one buyer for a certain territory or class of customers. It also comprises exclusive supply and quantity forcing on the supplier, where an obligation or incentive scheme agreed between the supplier and the buyer makes the former to sell on a particular market only or mainly to one buyer. For example, when a manufacturer pays a shelf allowance to a retailer it will be in its interest to concentrate its sales with this retailer so as to spread the cost of the allowance. Lastly, this group comprises selective distribution, where the condi-tions imposed on or agreed with the selected dealers may limit their number.
There are two main effects on competition: (1) certain buyers within that market can no longer buy from this particular supplier, i.e. it leads to foreclosure of certain buyers, and (2) as far as the distribution of final goods is concerned, since less distributors will offer this good it will also lead to reduced intra-brand competition. In the case of wide exclusive territories or customer allocation the result may be total elimination of intra-brand competition. When the exclusive distribution type of agreement is used rather selectively, that is not many stores can carry the product, it also leads to less in-store competition and reduced inter-brand competition.
Single branding group
Under the heading of single branding come those agreements/components that have as their main element that the buyer is induced to concentrate his orders for a particular type of good with one supplier. The group comprises non-compete and quantity forcing on the buyer, where an obligation or incentive scheme agreed between the supplier and the buyer makes the latter purchase its requirements for a particular good or service and its substitutes only or mainly from one supplier.
There are two main effects on competition: (1) other suppliers in that market cannot sell to the particular buyers, i.e. foreclosure of certain suppliers, and (2) as far as the distribution of final goods is concerned, the particular retailers will only sell one brand, therefore there will be no in-store competition in their shops. Both effects may lead to a reduction in inter-brand competition.
The reduction in inter-brand competition may be mitigated by stronger ex-ante competition between suppliers to obtain the single branding contracts, but the longer the duration the more likely it will be that this effect will not be strong enough to fully compensate for the lack of inter-brand competition.
Resale price maintenance group
Under the heading of resale price maintenance come those agreements/components that have as their main element that the buyer is obliged or induced to resell not below a certain price, at a certain price or not above a certain price. This group comprises minimum, fixed,
maximum and recommended resale prices. Maximum and recommended resale prices, although in theory unlikely to have negative effects, may work as fixed RPM. As RPM relates to the resale price it is mainly relevant for the distribution of final goods.
There are two main effects of minimum and fixed RPM on competition: (1) the distributors can no longer compete on price for that brand, leading to a total elimination of intra-brand price competition, and (2) there is increased transparency on price and responsibility for price changes, making horizontal collusion between manufacturers easier, at least in concentrated markets. The reduction in intra-brand competition may, as it leads to less downward pressure on the price for the particular good, have as an indirect effect a reduced level of inter-brand competition.
Market partitioning group
Under the heading of market partitioning come what may appear at first sight a miscellaneous group of agreements/components but that have as their main element that the buyer is restricted in where it either sources or resells a particular good. This group comprises exclusive purchasing, territorial sales restrictions, customer sales restrictions, after-market sales restrictions, prohibitions of resale and tying.
The main effect on competition is a reduction of intra-brand competition that may help the supplier or the buyer (in case of after-market sales restrictions) to partition the market and thus hinder market integration. This may facilitate price discrimination. Tying is slightly the odd one out. Its main effect is that the buyers may pay a higher price for the tied good than they would otherwise do but it may also lead to foreclosure of other suppliers and reduced inter-brand competition in the market of the tied good.
2.2 Combinations of vertical restraints
The next question to be considered is whether a combination of different vertical restraints increases the negative effects. In the Green Paper a rather prominent place is given to the argument that certain combinations of vertical restraints are better for competition than their use in isolation from each other (3). Although this may occasionally be the case, it does not appear to be the general rule. In general the opposite seems true, a combination usually aggravates the possible negative effects.
For example, a combination of one of the restraints of the single branding group with one of the exclusive distribution group combines a reduction of inter-brand competition with a reduction of intra-brand competition. In the case of final goods a market is created with local brand monopolists without in-store competition. Also, to foreclosure at manufacturer level is added foreclosure at the retail level. This means that not only may it be difficult for a manufacturer to sell a new brand as stores are tied, but also that new entrants to the retail market may have difficulty obtaining some of the leading brands. This results in a situation where it may be both difficult to find outlets and unprofitable to set up new outlets.
Another example is the combination of one of the restraints of the exclusive distribution group with one of the RPM group. To the reduction of intra-brand competition of the first is added the elimination of intra-brand price competition of the second. This quickly leads to a total elimination of intra-brand competition. This elimination of intra-brand competition may also help to sustain collusive tendencies between manufacturers facilitated by RPM. In general, this combination does also not make sense from an efficiency point of view as both protect the margin of the retailer. One of these restraints would normally suffice to overcome, for example, a free rider problem between retailers.
Lastly, a combination of one of the restraints of the single branding group with one of the RPM group may combine a reduction of inter-brand competition resulting from a lack of in-store competition with a facilitation of collusive behaviour between the manufacturers induced by RPM. Collusive behaviour may become easier as the lack of in-store competition takes away some of the competitive pressure. In addition the reduction of inter-brand competition is combined with a loss of intra-brand price competition resulting from RPM.
A number of combinations may however be viewed more positively, where it can be argued that one of the vertical restraints limits the possible negative effects of the other. In the combination of exclusive distribution with maximum RPM the latter restraint may help the supplier to limit possible price increases the buyer may want to implement under the protection of the territorial exclusivity obtained. The same reasoning can be applied to the combination of selective distribution and maximum RPM. Also the combination of exclusive distribution with quantity forcing on the buyer may work in the same way as the latter may prevent the distributor from raising his prices.
There are three combinations that are particularly negative from a market integration perspective: (1) territorial sales restriction combined with selective distribution at the same level of distribution, (2) exclusive distribution combined with exclusive purchasing, and (3) selective distribution combined with exclusive purchasing. These combinations help to make a distribution system more watertight by making arbitrage, either by final customers or by distributors, more difficult if not impossible.
3. The positive effects
It is said that in a number of situations the usual arms length dealings between manufacturer and retailer, detailing only price and quantity of a certain transaction, lead to a sub-optimal level of investments and sales. The following generalisations can be made about this:
1) The first and main reason why this (i.e. a sub-optimal level of investments and sales) is supposed to happen is the existence of some form of free rider problem. The person who makes an effort may not be able to appropriate all the benefits his or her effort engenders and may therefore be inclined to invest sub-optimally. This may be the result of free riding by one retailer on the promotion efforts of another retailer. Exclusive distribution or similar restrictions or RPM may be helpful in avoiding such free riding. Free riding can also occur between manufacturers where one invests in promotion in the shops for its brand, thereby also attracting customers for its competitors. Non-compete type restraints can help to overcome this.
For there to be a problem there needs to be a real free rider issue, something that is not always so obvious. Free riding between retailers can only occur on pre-sales services and not on after-sales services. The good needs to be relatively new or technically complex as the customer otherwise may very well know what he wants from past purchases. And the good must be of a reasonably high value as it is otherwise not attractive for a customer to go to one shop for information and to another to buy. On top of this, when all these conditions are fulfilled it must not be practical for the manufacturer to agree with the retailers effective service requirements concerning the pre-sales services. (4)
Free riding between manufacturers is also limited by rather strict conditions. It can be the case that a manufacturer who invests in promotion of his own product is also increasing demand for his competitors products. A vertical restraint may however not be helpful in addressing such a free rider problem. When for example advertising in the (national) media leads in general to extra demand with all outlets a vertical restraint will not help. Only in case the promotion leads to extra
demand via certain retail outlets, for example because these outlets are carrying the promotion, a non-compete type of agreement may help capture the full benefits. In addition, such free riding can only occur on pre-sales service, it must not be possible to make the promotion brand specific and it is only likely for relatively new and complex products as customers may otherwise know very well what they want already. (5)
2) A second general point that needs to be made concerns the possible divergence between what is privately efficient and efficient from a total welfare/consumer point of view. What is privately efficient is not always good for total welfare. To go back to the free riding between retailers or between manufacturers. Lets suppose a real free rider problem exists and sales can be expanded by inducing more pre-sales services although this would also lead to higher prices. When these extra services are valued equally by the majority of consumers this may very well lead to higher total welfare. But when the infra-marginal consumers (that is those who are already buying at the current price/service level) know what they want and do not appreciate the extra service, they only suffer from the higher price, especially if there is insufficient inter-brand competition. It may be privately efficient to increase the service level to attract more marginal consumers and thereby increase sales, but total welfare may nonetheless suffer.
3) A special form of free riding is the certification free rider argument. The hypothesis is that certain retailers perform a valuable service by identifying "good" products. The fact that these retailers sell a certain product signifies to the consumer that it is a good buy. This hypothesis may sometimes be useful for explaining the introduction of new products. New and complex products are first stocked by high quality, high margin stores where they are bought by avant-garde consumers. Gradually its reputation becomes established and demand grows enough for it to be sold through low price chains. If the manufacturer can not initially limit its sales to the premium stores, it runs the risk of being delisted and the product introduction may fail. If this is true, a problem analogous to invention patent protection exists. It may be necessary to provide temporary protection against price discounters to help the introduction of the product. However, a period of protection which is too long may only delay the product moving into the mature, price competitive stages of its life cycle, to the disadvantage of consumers. This means, at best, that there may be a reason to allow for a limited duration a restriction of the exclusive distribution or RPM kind; - enough to guarantee introduction, but not so long as to unduly delay large scale dissemination.
4) Yet another special form of free riding is the so-called 'hold-up' problem. Sometimes there are specific investments to be made by either the supplier or the buyer, such as in special equipment or training. In such a case, after the investments have been made the investor becomes to a certain extent prisoner to the other side. The balance of power will shift. In fear of this the necessary investments may not be made, unless ex-ante supply arrangements can be fixed. The investor fears that the other side will free ride on its investment. However, as in the free riding example between retailers, there are a number of conditions which have to be met before such a risk is real. Firstly, the investment must be sunk and specific to deal with that other party only. Secondly, it must be a long-term investment which is not recouped in the short run. And thirdly, the investment must be asymmetric; i.e. one invests more than the other. Only when these conditions are met can there be a real reason to have a vertical restraint for a limited duration, of the non-compete type when the investment is made by the supplier and of the exclusive distribution or exclusive supply type when the investment is made by the buyer.
5) The last reason for sub-optimal sales, also discussed in the Green Paper, that should be mentioned, is the problem of 'double marginalisation'. In case both the manufacturer and the retailer have market power each will set its price above marginal cost. They both add their margin that exceeds the one that would exist under competition. This may result in a final price that even exceeds the monopoly price an integrated company would charge, to the detriment of their collective profits and consumers. In this, arguably rather hypothetical case, quantity forcing on the buyer or maximum RPM could help the manufacturer bring the price down to joint profit maximising level.
6) In the economic literature it is explained that there is a large measure of substitutability between the different vertical restraints. This means that the same inefficiency problem can be solved by different vertical restraints. For example, as explained above, the problem of free riding between retailers or the certification free rider problem can be solved by means of exclusive distribution or fixed or minimum RPM. This is of importance as the negative effects on competition may differ between the various vertical restraints. This plays a role when indispensability is discussed under Article 85(3). As RPM is generally considered to be less acceptable from a competition point of view this may be a reason only to allow exclusive distribution or other less serious restraints and not RPM.
4. Differentiation between vertical restraints
A last question to be considered is whether some restraints are more or less harmful than others? To answer this question both negative and positive effects need to be considered. It may be a problem that some vertical restraints that may be most effective to solve a certain free rider problem may also have the most serious negative effects. In addition the answer will depend on the goals being pursued by competition policy. It is therefore not possible to answer this question solely on the basis of the economics involved. I will therefore just draw the attention to some general rules that can be formulated.
As a first general rule, it can be said that exclusive agreements are generally worse for competition than non-exclusive agreements. Exclusive agree-ments make, by the express language of the contracts or their practical effects, one party fulfil all or practically all its requirements from another party. For example, a non-compete obligation makes that the buyer purchases only one brand, while quantity forcing may leave the buyer scope to purchase competing goods. The degree of foreclosure is therefore different, while often the efficiencies are remarkably similar.
A second general rule, applicable to all four groups, can be formulated: restraints agreed for intermediate goods are in general less harmful than restraints affecting the distribution of final goods. At an intermediate level both the supplier and the buyer are usually professional and knowledgeable. This makes a possible loss of intra-brand competition less important because it stimulates specialisation which leads to comparative advantages.
A third general rule, also applicable to all four groups, can be formulated: the possible negative effects of vertical restraints are reinforced when not just one supplier with its buyers practices a certain vertical restraint but when also other suppliers and their buyers organise their trade in a similar way. These so called cumulative effects can be a problem in a number of sectors. To make a valid assessment of the effects of such a cumulation of vertical agreements requires a sector wide investigation and overview.
Within the RPM group, fixed and minimum price maintenance are evidently the serious restraints. Maximum and recommended prices, when really maximum or recommended, are clearly much less and possibly not at all restrictive.
Within the market partitioning group, which assumes a market integration objective, restriction of resale and after-market sales restrictions seem the worst as they allow market partitioning without clear possible efficiencies. Tying is in general considered a somewhat less serious restriction. It concerns the possible extension of market power from one market into another. Possible efficiency arguments ("need to assure the buyer uses the right sort of input for the fragile machine we sold him as breakdowns may hurt our products image" or "joint delivery is cost saving") may be limited. Exclusive purchasing is the least serious restriction within this group.
On the one hand economics tells us that it is only when inter-brand competition is reduced that it may be necessary for a competition authority to intervene against vertical restrictions. It is only in such conditions that it may have to reduce the restrictions on inter- or intra-brand competition which may result from vertical agreements.
On the other hand economics also tells us that vertical agreements are often necessary to realise efficiencies and may help firms to enter new markets.
At the same time rather strict conditions have to be met before one can actually speak of a real free rider problem that justifies the imposition of vertical restrictions. The case is strongest for vertical restrictions of a limited duration which help the introduction of new complex products.
This leads to the conclusion that in situations where the parties have (considerable) market power a case by case approach is warranted for vertical restrictions
(1) Green Paper on Vertical Restraints in EC Competition Policy, COM (96) 721 final, of 22.01.1997, point 54.
(2) Green Paper, point 303.
(3) Green Paper, point 67.
(4) The standard argument against contractability of service requirements is that the costs of monitoring and the contract costs may be prohibitive for the manufacturer in case of a large number of small retailers.
(5) Promotion that only creates image without there being real extra quality has doubtful welfare effects and does not need to be protected.
[ SPEECH ] - [ Previous ]