On the deadweight cost of production requirements for geographically differentiated agricultural products.

This article investigates the welfare effects of alternate collusion schemes in a context where collusion is authorized in order to cover fixed costs. We find analytical evidence that whenever producers as a group are allowed to control quota levels, an input quota policy entails a smaller absolute deadweight loss than an output quota policy. This means that in the presence of fixed costs, collusion schemes where producers collude on input use (e.g., land) may be socially preferable to arrangements where producers can directly control output. Such schemes are found in some European geographical indicator (GI) markets, where certification costs have been argued to be high (Marette, Crespi, and Schiavina 1999). This result contrasts with a conjecture by Lence et al. (2007), previously published in this Journal.

Recent literature on collective marketing schemes for agricultural products has identified instances where a benevolent planner may find it optimal to allow collusion among many producers in order to provide sufficient economic incentives to create new, high-valued differentiated products (Marette, Crespi, and Schiavina 1999; Marette and Crespi 2003; Lence et al. 2007). (1) The surplus transfer is deemed necessary to compensate for nonconvexities in the production set arising from the presence of large fixed costs, typically certification costs that may be shared among producers (Marette, Crespi, and Schiavina 1999; Marette and Crespi 2003), or an investment cost incurred in the first stage of marketing the differentiated product (Lence et al. 2007). (2) Although such transfers entail efficiency losses, the benefits arising from consumption of higher-quality products may outweigh those losses and result in higher social welfare. (3)

That transfers may be desirable to provide break-even production incentives in the presence of nonconvexities is a well-known fact in public economics. Without necessarily invoking economic justifications for such transfers, the traditional agricultural economics literature has investigated the relative efficiency of various redistribution schemes, such as government-managed subsidies and quotas (see, for instance, Wallace 1962; Floyd 1965; Gardner 1983; Gardner 1987; Gisser 1993; Bullock and Salhofer 2003). Although inspired by the same fundamental question (how to transfer economic surplus to producers efficiently), the emerging literature on collective marketing arrangements for agricultural products is novel in that it disregards traditional transfer schemes to investigate the relative efficiency of self-administered policies for the production sector. More specifically, policy makers are now assumed to design antitrust policies that allow for different types of collusions between producers, typically quotas on output or on an input (in particular, land), and producer groups are responsible for choosing the quota level and allocating rights among individual producers.

In a sense, this newer literature has set a bridge between traditional agricultural policy and the market power literature by considering policy instruments where producers, as a group, are allowed to influence the extent of the distortion from the competitive equilibrium. The policy question that arises is therefore more complex and can be framed within the following optimization program:

(1) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where DWL denotes the social cost or deadweight loss, P represents the policy choice (typically, the type of collusion that should be exempt from antitrust scrutiny, say, a quota on land), [lambda] represents the level or intensity of collusion among producers (e.g., the number of acres in the quota), T represents the realized transfer to producers, and [tau] is a policy objective (the minimum transfer required to achieve the desired incentive). This program can be contrasted to the traditional farm policy paradigm, where the policy maker would choose both the type (P) and the intensity ([lambda]) of the policy to achieve any given level of transfer [tau]

(2) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII].

The recent interest in analyzing transfer policies within the framework implied by program (1) likely comes from the observation of collective marketing arrangements for so-called "geographically differentiated agricultural products" (GDAPs), also known as protected designations of origin (PDOs) or protected geographical indications (PGIs), mainly present in Europe, as well as marketing arrangements within certain U.S. marketing orders or producer cooperatives that amount to controlling supply, directly or indirectly (Carman and Alston 2005). (4) The literature has analyzed such marketing schemes from both a positive and a normative standpoint, trying to justify observed policies that may appear inefficient or providing arguments in favor of implementing them more systematically.

The Economic and Policy Significance of Lence et al.'s Conjecture

One of the most important pieces of literature in the field to date is the article by Lence et al. (2007). The authors use a general model of aggregate supply to compare the profit incentives of alternate collusion schemes in the context of geographically differentiated agricultural products and argue that allowing producers to collude on output, to restrict acreage, or to adopt cost-enhancing production requirements may have positive social welfare effects, whenever those schemes, through a rise in output price, enable producers to cover the fixed cost of developing the differentiated product.

An incidental claim in Lence et al.'s (2007) article (referred to as L's conjecture) is that, conditional on the GDAP having been developed, stronger forms of collusions (like collusion on output) may be preferable to weaker forms of collusions (like the implementation of cost-enhancing production requirements), even from a social welfare standpoint, the reason being that "legislation aimed at curtailing the market power of producer organizations may induce large technological distortions" (Lence et al. 2007, p. 962). The idea behind L's conjecture is that by preventing producer organizations from colluding on output, stringent antitrust rules only leave those organizations with indirect tools to control output, leading to inefficient production. The authors then suggest that because GDAP legislation in the United States provides producer organizations with fewer instruments to control supply than in the European Union (EU), the incentive to distort production practices to limit supply is probably larger in the United States, leading to larger overall deadweight losses. Taken literally, this conclusion could mean that unless antitrust regulations effectively prevent producer organizations from distorting production practices to restrict output, society would be better off by allowing producer organizations to directly control output levels.

Although Lence et al.'s (2007) observation that indirect ways to control output, such as input restrictions, entail additional technology-driven deadweight costs is correct, they do not explicitly recognize that this very inefficiency also affects the ability of producers to distort output price, therefore leading to a smaller consumer loss than that under output collusion. How these two opposing effects affect the overall deadweight cost comparison has not been resolved in the literature to date.

The present article investigates this question in a partial equilibrium context. Our findings cast serious doubt on the validity of L's conjecture by showing that, if producers are only allowed to resort to a production-distorting instrument to limit output (specifically, an input quota), they will, in fact, make production choices that are less costly for society as a whole than if they had been allowed to directly control the output level. Therefore, whenever both policies generate sufficient rents to cover fixed costs, an input quota is socially more desirable than an output quota. (5) The result is derived using an equilibrium displacement model and by assuming locally linear demand and input supply schedules. It is robust to whether the quasi-rent to suppliers of the unrestricted input is included in the profit maximization problem facing the colluding industry. (6)

The policy implications of our findings are potentially important. The input quota scenario that is the focus of our theoretical analysis is not limited to acreage restrictions. Rather, it covers many types of cost-enhancing production requirements such as capacity constraints or input tax schemes with redistribution of tax revenues, all of which can be recast in terms of a quota on some well-defined input. We would argue that these types of production requirements are the norm, rather than the exception, in many GDAP markets. Indeed, the adoption of production constraints by producer groups is essential to the recognition of GIs in the EU, because the stated rationale behind the PDO/PGI protection is that the registered products differ significantly from their generic counterparts due to specific geographical factors and specific production methods (Council of the European Union 2006). Therefore, unless antitrust policy effectively prevents them from doing so, there is little doubt that, in defining production standards, producer groups will try to maximize joint profits.

One recent illustrative example of the strategic adoption of cost-enhancing production requirements can be found in the French Comte cheese market. Comte is France's top-selling PDO hard cheese. In May 2007, the French government adopted a new regulation (proposed by the Comte consortium) that limits the number of basins to be used in cheese factories, the capacity of those basins, and the frequency of their use (Republique Francaise 2007). (7)

Print Get the Mag Weekly Updates