رهبری قیمت عمودی در بازارهای ذرت محلی در بنین

This paper considers vertical price relationships between wholesalers and retailers on five local maize markets in Benin. We show that the common stochastic trend and the long-run disequilibrium error must explicitly be considered to correctly interpret the restrictions on the error–correction structure in terms of economic power in the channel. Interesting differences between markets are found. In the two major towns, retailers play a more prominent role in the price formation process than generally assumed in the literature on development economics. In the two larger rural centers, however, wholesalers involved in arbitrage among urban markets do influence price formation.

In the literature on industrial organization, retail prices are often assumed to be determined by wholesale market conditions (see, for example, Tirole, 1988 and Martin, 1993). Likewise, in the marketing literature on the functioning of food markets in tropical countries, the vertical price leadership of wholesalers is often conjectured but not empirically tested. In contrast to the vertical case, much empirical research on spatial price linkages between agricultural commodity markets in developing countries is available (e.g., Ravallion, 1986, Baulch, 1997, Badiane and Shively, 1998, Kuiper et al., 1999 and Abdulai, 2000). Most studies on spatial price linkages focus on two issues: are markets integrated and are prices dominated by a central market? Most results confirm that markets are competitive and integrated, although less than perfect. Transaction costs hamper market integration and not all spot markets perform equally well. Moreover, as Baulch (1997; pp. 478–479) rightly points out, “market integration does not itself, however, imply that food markets are competitive. The spatial arbitrage conditions are also consistent with such oligopolistic pricing practices as basing point pricing (Faminow and Benson, 1990)”. Consequently, tests for market integration should also be accompanied by an investigation of the wholesale–retail price relationship in order to assess the power of wholesalers being involved in spatial arbitrage. Accordingly, we want to draw attention to the wholesale–retail price relationship. Can we find evidence for vertical price leadership of wholesalers or do they lack the market power to impose prices on local retailers? To put it differently, can we find some empirical support for the popular complaints regularly expressed by retailers and local market authorities about the market power of wholesalers? Looking at the Benin maize market, Lutz (1994) found that retail and wholesale price series in the same market place cohere, which implies that retail margins are stationary. This result suggests that retailers are indeed passive decision makers, following wholesale prices without taking local supply and demand conditions into account. However, other evidence provided by a survey among traders does not support this conclusion and shows that a large number of wholesalers supply the urban market from different surplus regions, while urban retailers actively search for wholesalers proposing the lowest price (Lutz, 1994). Moreover, in rural areas retailers can choose to buy either from wholesalers or at the farm gate. Buying directly from farmers may provide retailers some freedom to set prices. Consequently, it is not a clear matter whether wholesalers or retailers or both have some market power and are able to influence price formation. In an earlier study on price arbitrage in the wholesale segment of the maize market, we concluded that all wholesale markets played a role in the price formation process (Kuiper et al., 1999). None of the price series of any of the wholesale markets were found to be dominant: all price series were interdependent. The arbitrage process corresponded to a network with a number of interdependent wholesale markets; there were no autarkic markets and transportation costs did not show a stochastic trend. The study, however, did not incorporate the price series observed on the retail segments. In the present paper we focus on this omission, questioning the relationship between prices in wholesale and retail market segments in various markets for the same sample period as in Kuiper et al. (1999). The questions we set out to answer are: is there a difference in wholesale–retail price relationships in towns and rural centers, and is there any evidence for wholesale market dominance vis-à-vis the retailers? Most studies on vertical price relationships published to date in marketing and industrial organization (see, for example, Gerstner and Hess, 1991 and Lee and Staelin, 1997 and the references they cite) have used comparative statics to study channel behavior; the long-run relationships derived have not been empirically tested. Our study differs in that its main focus is on empirical analysis. In order to verify whether the price formation process is driven by retailers, wholesalers, or both, we can distinguish two segments in the market: the retail segment and the wholesale segment. We assume that actors in both segments try to maximize profits. To examine whether or not wholesalers are price leader vis-à-vis the retailers in the sense of Stackelberg leadership, one can consider the long-run equilibrium (i.e., cointegrating) relationship between the wholesale and retail prices and test whether or not wholesale prices and retail prices respond to deviations from the equilibrium price. The basic assumption we make is that the common stochastic trend observed in the cointegrated wholesale–retail price series is generated by local supply and demand conditions (seasonal price trend). Three models then become interesting for the study of price adjustment: Model 1 in which both retailers and wholesalers have some freedom to respond to deviations from the equilibrium price; Model 2 in which only retailers have the power to respond to deviations from the equilibrium price; and Model 3 in which only wholesalers have the freedom to respond to deviations from the equilibrium price. In Model 1, wholesalers have sufficient power vis-à-vis the retailers to behave as vertical price leaders, although retailers can still maximize their profits dependent on the wholesale price being set by the wholesalers. This model applies if both retail and wholesale traders exercise some market power, for example, if alternative market opportunities exist for both actors. In contrast, in Model 2 the retailers do not allow wholesalers to influence short-run retail price deviations and leave them with only the option of setting wholesale prices on the basis of the wholesalers' unit costs (i.e., farm gate price plus a margin to enable the wholesaler to survive), which represents the common stochastic trend that drives the two prices, the retail price and the wholesale price, in the long run. Market power for retailers may be the result of a temporarily abundant supply in the wholesale segment and a lack of alternative market opportunities for wholesalers. Lastly, in Model 3, only wholesalers are able to set their prices in the sense of Stackelberg leadership and to respond to price deviations from the equilibrium. The situation applying in this model is one in which large numbers of retailers buy from farmers and wholesalers to serve local consumers, whereas the local wholesalers are also involved in regional market arbitrage and ship to urban markets. Consequently, the retail price is getting stuck to the common stochastic trend from which the wholesale price can deviate in the short run by price arbitrage among the spatially dispersed wholesale markets. In deriving the testable implications of the hypotheses about economic power in marketing channels, we explicitly take the common stochastic trend and the deviations from the long-run vertical price equilibrium into account. This is the major contribution of our paper to the debate on vertical price leadership and we will show that if the common stochastic trend and disequilibrium error are not explicitly assigned to certain variables in the channel model, one can easily be wrong about how restrictions on the error–correction model must be interpreted in terms of vertical price leadership. Furthermore, since we wish to test between the three theoretical models outlined above, another important advantage of our empirical method is that it nests these tests in one procedure. The article analyzes the process of price formation for maize in five market places in the south of Benin: two towns (Bohicon and Cotonou) and three rural centers (Azové, Dassa and Kétou). Section 2 discusses the relevance of the three above-mentioned models. In Section 3, the method of analysis is presented. We formulate the long-run model and derive its testable implications on the short-run price system. Section 4 presents the empirical results and Section 5 the conclusions.

In this paper, we proposed a method for empirically testing whether or not wholesalers have some price setting power vis-à-vis the retailers. The method was applied to three models that were considered possible candidates for describing the vertical price relationships in the marketing channels of local maize markets in Benin. A salient feature of our method is that the common stochastic trend and the deviations from the long-run vertical price equilibrium must be assigned to the variables in each model being considered. Doing this for the application in this paper, we found that the exclusion restrictions on the error–correction structure led to testable implications discriminating between the three models. As far as our limited evidence goes, we conclude that retailers do not allow wholesalers to behave as vertical price leaders in the sense of Stackelberg leadership, unless the wholesalers are involved in market arbitrage. In fact, in the towns wholesalers do not have alternative market opportunities and retailers dominate the local market price formation process. In Kétou, the few retailers that exist seem to be able to exploit some opportunities for monopolistic competition. In the two larger rural centers considered in this study, Dassa and Azové, wholesalers dominate: retail prices are stuck to the stochastic trend, while wholesalers have alternative arbitrage opportunities, giving them some freedom to influence prices. Our empirical results indicate that relations between wholesalers and retailers vary between market places. In contrast to common assumption in development studies, retailers play a crucial role in the price formation process. Local market conditions are decisive in the distribution of market power among retailers and wholesalers. Consequently, the statement ‘the retail market segment is dominated by the wholesale segment’ needs to be tested, before it is imposed as an assumption on a model.