Project topics and materials
Get Complete Project Material File(s) Now! »
Chapter 2 Buyer Alliances in Vertically Related Markets
Introduction
Alliances of multiple buyers to deal with their suppliers is a widespread phenomenon in vertical markets. Examples include group purchasing organizations that negotiate tariffs with medical device manufacturers on behalf of hospitals; independent drug-stores who join buyer groups to negotiate wholesale contracts with drug manufacturers (e.g., Numark in the U.K., Giphar in France); advertisers in the online ads industry who delegate their bidding campaigns to specialized agencies in order to get advertisement space on search engines and social networks (Decarolis, Goldmanis and Penta, 2017); buyer alliances formed by food retailers to negotiate trading terms with their suppliers (e.g., Dobson et al., 1999).1
In practice, the competition concerns about buyer alliances (or buyer groups) have long been analyzed with a strong presumption of legality by antitrust agencies (Carstensen, 2010). At first glance, alliances may generate better trading terms for buyers resulting in cost savings which could then be passed on to final consumers without generating any potential market power effects as opposed to horizontal mergers. However, recent investigations conducted by competition authorities have noticed risks of adverse ef-fects such as collusive behavior between retailers due to exchanges of information, and have claimed for further control of such practices.2
The main contribution of this paper is to shed light on a new mechanism emerging from the inability of suppliers to price discriminate between the members of a buyer alliance. I show that, in the absence of such discrimination, theoretical predictions about the effects of buyer alliances on the bargaining power of firms and prices paid by final consumers are ambiguous. To gain further insights, I consider the issue from an empirical perspective using homescan data on bottled water purchases in France for the year 2013. The empirical framework builds on Bonnet, Bouamra-Mechemache and Molina (2017) who develop a structural model of demand and supply to recover the division of surplus in bilateral oligopolies with differentiated products (see Chap-ter 1). I first estimate consumer demand to measure the degree of product differenti-ation which drives the pricing behavior of firms in the French bottled water market. On the supply side, I consider a setting of vertical contracting between multiple man-ufacturers and retailers which allows for balanced bargaining power and takes into account the impact of (negotiated) wholesale tariffs on the downstream price com-petition between retailers. Bargaining power of firms are recovered based on new conditional moment restrictions which approximate Chamberlain (1987) optimal in-struments. Using estimated parameters of the structural model, I perform simulations to analyze the economic effects of buyer alliances. My focus is on three alliances that have been formed between competing retailers on the French food retail sector in 2014, namely: Carrefour (21:8%) and Cora (3:3%), Groupe Auchan (11:3%) and Système U (10:3%), ITM Entreprises (14:4%) and Groupe Casino (11:5%).3 Empirical results con-trast with the standard intuition that alliances generate more beneficial trading terms for retailers. The results show that buyer alliances weaken the bargaining power of retailers and allow upstream manufacturers to increase their price-cost margins by 2:80%. I find that the increase in wholesale prices is passed on to final consumers and reduces industry profit by 0:69%.
This article relates to the literature on buyer power in vertically related markets which, dating back to Galbraith (1952, 1954) and his concept of countervailing buyer power, analyzes the potential for large buyers to obtain lower trading terms and pass on the resulting benefit to final consumers. Whether consumers should welcome big retailers has been a controversial issue in the economic literature (e.g., Stigler, 1963; Hunter, 1958) and remains subject to ongoing research. Large buyers are often con-sidered being able to secure lower input prices from their upstream providers because they have better outside options — e.g., credible threats of vertical integration (Katz, 1987) — or they act as gatekeepers due to the absence of rivalry on the market. It has also been emphasized that whether big buyers obtain more favorable trading terms de-pends on the curvature of suppliers’ profit functions (Chipty and Snyder, 1999; Nor-mann, Ruffle and Snyder, 2007). Effects of retail concentration on both upstream and downstream markets (e.g., horizontal mergers, entry or exit of rival retailers) have also been extensively analyzed in the literature (Dobson and Waterson, 1997; Iozzi and Valletti, 2014; Gaudin, 2017). However, instead of considering consolidation on both sides of the market, my paper focuses on the case in which retailers form alliances on the upstream market but remain competitors at the downstream level.
In this respect, several articles have pointed out that such alliances can be used by buyers to coordinate their purchasing policy and stimulate upstream competition by reducing the number of suppliers to deal with (Inderst and Shaffer, 2007; Dana, 2012; Chen and Li, 2013).4 By contrast, my article considers that buyer alliances do not directly modify the buyer-seller network but enable downstream firms to affect threat points in negotiations by precipitating bargaining breakdowns with multiple retailers at the same time as in Caprice and Rey (2015). Prior to the empirical analy-sis, I present the main insights in a stylized model of vertical relationships with one upstream manufacturer and two downstream retailers. In this setting, firms operate under constant returns to scale, per-unit wholesale prices are determined through bi-lateral bargains, and retailers compete in prices on a downstream market. I show that a buyer alliance which aims at maximizing total profit of its members and securing nondiscriminatory trading terms via centralised negotiations generates two main ef-fects on the bargaining power of firms. On the one hand, joining forces deteriorates the outside option of the manufacturer which strengthens the bargaining position of retailers. On the other hand, this increase in bargaining power can be mitigated by the fact that the buyer alliance provides similar trading terms for both retailers. Initially pointed out by O’Brien (2014) in the context of banning price discrimination in inter-mediate good markets, I show that this effect lessens the ability of retailers to extract price concessions from the manufacturer and undermines their bargaining power.
On the empirical side, my framework is in line with the literature on structural models of vertical relationships and bilateral oligopolies. One strand of this literature has analyzed vertical contracting in settings where multiple upstream players make (simultaneous) take-it-or-leave-it offers to downstream players (Villas-Boas, 2007, 2009; Ho, 2009; Bonnet and Dubois, 2010; Bonnet et al., 2013; Goldberg and Hellerstein, 2013). Since powerful firms operate on both sides of the French bottled water mar-ket, my structural modeling approach follows a recent stream of articles which de-velop empirical models of bargaining to analyze buyer-seller interactions with con-tracting externalities (Draganska, Klapper and Villas-Boas, 2010; Crawford and Yu-rukoglu, 2012; Grennan, 2013; Gowrisankaran, Nevo and Town, 2015; Ho and Lee, 2017). Grennan (2013) is of particular interest since he simulates the formation of a group purchasing organization that negotiates with medical device manufacturers on behalf of hospitals in the U.S. coronary stent industry. His findings show that such an alliance may increase prices in favor of manufacturers, thereby reducing the bargain-ing power of hospitals. However, the fact that hospitals behave as local monopolists and do not strategically compete with one another on a downstream market implies that his framework differs from mine in many aspects.5
The rest of this article is organized as follows. In Section 2, I develop a stylized model of vertical relationships to shed light on the main economic forces at play when two retailers form a buyer alliance. Section 3 presents the data, the structural model of demand and supply, and the results. Simulations of buyer alliances are considered in Section 4, and Section 5 concludes.
Theoretical Insights
In this section, I design a stylized model of vertical relationships to provide insights on the main effects generated by a buyer alliance on the bargaining power of firms and final prices paid by consumers.
Setup. Consider an upstream manufacturer A which produces a brand and sells it to two symmetric retailers, R1 and R2, competing for consumers on a downstream mar-ket. The two retailers are supposed to be differentiated, reflecting differences in their sales services or location. There are two differentiated products on the market, indexed by j = 1;2 (a product is defined as a brand-retailer combination). For each product, the marginal cost of production incurred by the upstream manufacturer and the marginal cost of distribution for the retailers are assumed to be constant and normalized to zero for the sake of convenience. The product distributed by Rj is sold to consumers at price pj and its competitor at price p j . I suppose that consumer utility functions and budget constraints lead to the following demand function qj (pj ; p j ) for Rj ’s product Timing, solution concept and information. I consider a two-stage game in which the upstream manufacturer and the two downstream retailers interact as follows:
Stage 1: The upstream firm engages simultaneously in a bilateral bargaining with each retailer on the wholesale market. Contracts are secret and consist of a per-unit wholesale price paid by the retailers.
Stage 2: Retailers engage in a simultaneous price-setting competition on the downstream market.
This two-stage game is solved by using a refinement of the Perfect Bayesian equilib-rium concept. In the upstream market, I employ a semi-cooperative approach pio-neered by Horn and Wolinsky (1988) to determine the surplus division between the manufacturer and its retailers: the “Nash-in-Nash” bargaining solution. This bar-gaining protocol corresponds to a delegated agents model in which separate repre-sentatives are (simultaneously) sent by firms to every bilateral negotiations in order to bargain over trading terms on their behalf. Because each delegated agent partic-ipates only in one bilateral negotiation and cannot communicate with their counter-parts (even those coming from the same firm), it is assumed that firms’ delegates hold “passive-beliefs” over deals reached elsewhere (McAfee and Schwartz, 1994).6 There-fore, this bargaining model exhibits a setting in which contracts are binding, firms behave schizophrenically, and negotiators have “passive-beliefs”.7 In the downstream market, competition between retailers is modeled as a Nash pricing game with interim observability (Rey and Vergé, 2004), that is, wholesale prices negotiated with the up-stream firm are fully revealed to retailers before they set their prices.8
The purpose of this simple model is to illustrate the effects of a buyer alliance formed by R1 and R2 on the equilibrium retail and wholesale prices. First, I study a benchmark setting in which R1 and R2 negotiate separately their wholesale tariff with the upstream firm. Then, I investigate the case in which R1 and R2 negotiate together through a buyer group. Proceeding backwards, I consider beforehand the downstream price competition between the two retailers.
Table of contents :
Introduction
1 The Downstream Competition Effects in Bilateral Oligopolies: A Structural Bargaining Approach with Limited Data
1 Introduction
2 Data and Institutional Details
3 Consumers Demand for Soft Drinks
3.1 The Demand Model
3.2 Identification and Estimation of the Demand Model
4 The Supply Model
4.1 Stage 2: Downstream price competition
4.2 Stage 1: Bargaining between manufacturers and retailers
4.3 Identification and Estimation of Bargaining Stage
5 Empirical results
5.1 Demand Side
5.2 Supply Side
6 Counterfactual Experiments
7 Concluding remarks
Appendix A Price-cost margins of the manufacturers
Appendix B Proof of Lemma
Appendix C Estimation of the out-of-equilibrium retail prices
Appendix D Counterfactual algorithm
2 Buyer Alliances in Vertically Related Markets
1 Introduction
2 Theoretical Insights
2.1 Downstream Price Competition
2.2 Manufacturer-Retailer Bargaining
2.3 Manufacturer-Retailer Bargaining with a Buyer Group
3 Empirical Analysis
3.1 Data and Industry Background
3.2 Consumer Demand for Bottled Water
3.3 Downstream Competition and Manufacturer-Retailer Bargaining
3.4 Results
4 Simulations of Buyer Alliances Formed by Downstream Competitors
5 Concluding Remarks
Appendix A Theoretical Insights on Buyer Alliances
A.1 Conditions for Existence and Uniqueness
A.2 Retail Pass-through
A.3 Buyer Group Effects: Computational Details
Appendix B Empirical Bargaining Framework: Technical Issues
B.1 Computation of the Out-of-Equilibrium Retail Prices
B.2 Derivation of the Manufacturers’ Price-Cost Margins
B.3 Algorithm to Approximate the Optimal Instruments
Appendix C Demand Results: Tables and Figures
Appendix D Simulations of Buyer Alliances: Technical Issues
D.1 Ex-Post Upstream Margins
D.2 Ex-Post Retail Pass-through
D.3 Counterfactual Algorithm
3 Full-line Forcing Practices in Vertically Related Markets†
1 Introduction
2 The model
3 Pure component
3.1 D chooses the assortment fAh;Bl g
3.2 D chooses the assortment fAh;Al g
4 Full-line forcing
4.1 D chooses to sell the bundle fAh;Al g
4.2 D chooses to sell fBl g
5 Full-line forcing vs pure component selling strategy
6 Illustrative example
7 Concluding remarks
Appendix A Pure component selling strategy
Appendix B Full-line forcing strategy
