Exclusive versus common dealership.

• Author: Gabrielsen, Tommy Staahl

1. Introduction

   How should a producer organize its distribution system? This is one of the main issues in the vertical restraints literature. One of the questions posed in this literature is whether producers of imperfect substitutes should use a distribution system with separate exclusive dealers or use a common dealer.

   Different distribution systems have benefits and drawbacks. A common distribution system, for instance, will give rise to collusive retail pricing (Bernheim and Whinston 1998; O'Brien and Shaffer 1997). However, because the dominant dealer may not accept your product, only a limited amount of this profit can be extracted by each producer. The authors show that producers selling to one common retailer can only extract each product's incremental contribution to the industry profit. In contrast, when producers use separate exclusive dealers (i) the joint collusive retail prices are no longer achievable and (ii) the producers can extract all retail profits (Bonanno and Vickers 1988). Therefore, there is a tradeoff between the retail prices that can be achieved in different structures and the amount of profits that you as a producer can extract from your dealers. Comparing the above-mentioned structures, O'Brien and Shaffer (1993) found that the exclusive distribution system will be preferred by the producers.(1)

   In the existing literature there is an asymmetry between the two distribution systems. Each producer can extract all potential retailer profit in the exclusive distribution system, but not in the common distribution system. We argue that full extraction of retail profit is unlikely to be an equilibrium outcome in an exclusive distribution system because, if the dealers earn no profit, each of them could threaten to defect to a rival producer and thereby place the original supplier out of the market (foreclosure).(2) Given that such a defection is possible also in the exclusive distribution system, we show that there is a symmetry between the producers' possibilities for profit extraction in the two systems. In both systems, the producer can only extract its product's incremental contribution to the industry profit.(3)

   We show that a system with a common dealer generates higher aggregate industry profit. In addition, we show that the profit that can be retained by a common dealer is also less than what can be jointly retained by two exclusive dealers. Therefore, there is no tradeoff for the producers when moving from an exclusive to a common distribution system for the following reasons. First, adding a second product in a system with two exclusive retailers generates more rivalry between the products than when a second product is added to the sales of a single common retailer. The obvious reason for this is that a common retailer internalizes the competitive externality between the products, which cannot be achieved when the products are sold by separate exclusive dealers. Consequently, the aggregate industry profit is higher under common than under exclusive dealing.

   Second, the producers can only extract their incremental contribution to the industry profits in each system. This contribution is the difference between the industry profit when both products are sold and the maximal profit when only one product is sold (i.e., the monopoly profit of that product). Because incremental contribution is increasing in the aggregate profit when both products are sold, each producers' profit also increases when industry profit goes up. For the same reason, the retail sector's rent decreases when industry profit increases. Because aggregate profits are larger under common dealership, the rent retained by the common dealer is less than the aggregate rent that can be retained by the exclusive dealers. It is the attractiveness of the common dealer, higher industry profits, that limits the rents that this dealer can retain. Consequently, the producers prefer a system with a single common retailer.

   The rest of the paper is organized as follows. In section 2 we review in more detail contributions to the literature that are most relevant to our analysis and we also relate our results to the existing literature. The model and the rules of the game are presented in section 3, and the equilibrium outcomes are analyzed in section 4. In section 5, we offer concluding remarks, and we discuss the implications for antitrust policy.

2. Related Literature and Results

   For each of the two distribution systems we consider, there are several theoretical studies. In the following, we refer briefly to the results from those studies.

   First, we consider the exclusive distribution system. Rey and Stiglitz (1988, 1995) study a setting where two producers may choose between having one and many exclusive retailers each. They find that each producer chooses to have one exclusive dealer each. The reason is that such a distribution system would dampen competition. By assumption, there is no intrabrand competition. This will encourage each producer to behave less aggressively concerning price setting and thereby dampen the interbrand competition. Bonanno and Vickers (1988) study a game in which each producer of a differentiated product can either integrate with its exclusive dealer or delegate sales to him through observable two-part tariffs. It is shown that each producer has an incentive to delegate (separate), and in equilibrium, wholesale prices above producer costs are offered.(4) The intuition is that when the game between the dealers is in prices, a high wholesale price will induce your dealer to set a high retail price. Because retail prices are strategic complements, this will induce the dealer of the rival product to raise his price as well. This strategic effect is beneficial to each producer as long as the increased profit at the retail level can be appropriated by the producers with fixed fees in the wholesale contracts.

   Second, we consider the common distribution system. In this distribution system, the issue of foreclosure through exclusive dealing clauses is discussed in the literature. Bernheim and Whinston (1998) and O'Brien and Shaffer (1997) analyze this question in a model with two producers of differentiated products and one single retailer. They show that foreclosure is never an equilibrium outcome.(5) The reason is that a rejected producer will lower the fixed fee he charges from the retailer until the net profits for the dealer from carrying his brand is nonnegative. This means that a producer only can extract his product's net contribution to the profit earned by the dealer. Because the incremental contribution of a product equals the difference between the industry profit when both products are sold to consumers and the industry profit when only the rival brand is sold, each producer has an incentive to offer wholesale contracts that maximize the industry profit. In equilibrium, the producers offer wholesale contracts, where wholesale prices are set equal to producer marginal costs, and the dealer sets retail prices as a multiproduct monopolist facing true production costs. Therefore, the joint collusive profit Of the industry is realized and each producer collects its product's incremental contribution to this profit. Because the products are imperfect substitutes in demand, the retailer will earn some profits. The profit earned by the retailer increases when products become closer substitutes (because then each product's incremental contribution becomes smaller).

   As noted above, these two outcomes from different distribution systems are compared in an article by O'Brien and Shaffer (1993). In a model with constant returns to scale and linear demands, the authors show that an exclusive distribution system will always be better for the producers because all profit can be extracted from exclusive dealers. In the specified game, the producers choose whether to have one or two retailers the first stage. If one retailer is chosen, we are back in the setup of Bernheim and Whinston (1998) and O'Brien and Shaffer (1997). If two retailers are chosen, we assume the producers only offer contracts to separate dealers, which is the setup of Rey and Stiglitz (1988, 1995) and Bonanno and Vickers (1988). We see that foreclosure is not an issue in the exclusive subgame of their model, therefore, the producers can achieve full franchise extraction of retail profit. We argue that, in many circumstances, this is not a realistic assumption. If the number of dealers is limited as in O'Brien and Shaffer (1993) (there are only two), dealers should be able to earn some rent. It seems natural to us that if a dealer receives a contract proposal that generates no rent to him, he could threaten to defect to the rival product.

   Generally, there are a number of issues to which a prospective dealer should pay attention when having received a contract proposal from a specific producer. Perhaps the most natural question to ask is whether there are more dealers of the same product in your area. If there are more dealers, retail competition will be tough, but if you are the sole distributor, you will enjoy some market power over consumers. In this way, a prospective dealer should try to estimate his expected profit from accepting the contract. If the expected dealer profit from a contract is low, a dealer may wish to defect to rival producers. We show that if this is a viable option for the dealers, it cannot be an equilibrium when dealers earn zero profit in an exclusive distribution system. Producers must now ensure that their dealers are left with sufficient profit so that they will not want to change the product they handle. The implications for the producers are that the profitability of an exclusive distribution system is reduced. In fact, we show that it is reduced to the extent that the producers now choose a common distribution system.

   In our model, the potential for foreclosure implies that there is a symmetry...