Supply chain management of differentiated agricultural products under imperfect information

Participants in a supply chain of agricultural value added products face two significant challenges. First, many of the costly distinctive traits being desired by consumers are difficult (if not impossible) to observe even after consumption. In order for markets for these classes of goods to develop, firms touting the quality of the product need to be trusted. Hence, maintaining an excellent reputation is essential for firms to keep their customers' good will. Secondly, production is conducted in an environment of yield uncertainty, making it impossible for producers and processors to predict with certainty either the quantity or the quality of the input that will be available in any given season. In short, production, processing, and marketing of some value added products require tighter coordination mechanisms than those afforded by open market transactions.;Chapter 2 studies the merit and feasibility of co-existence of spot and contract markets for a value added product. Co-existence of contract and spot markets arise as an equilibrium for a wide range of distinct parameterizations of the model. The fundamental economic factors influencing the prevalence of each market are identified. In order to obtain co-existence, both yield uncertainty and a vigorous competition whenever a spot market arises are needed. Chapter 3 analyzes the optimal choice of a quality assurance system (QAS) a processing firm should require from its suppliers in the presence of imperfect information. We find that firms will require more stringent QAS when quality is easier to discover, and reputations are firm specific (as opposed to industry-wide). Also, monopolists will implement more stringent QAS than duopolists, indicating that market concentration is not necessarily welfare reducing in this environment. Policy implications are drawn. Chapter 4 proposes a flexible framework that can be used empirically by a group of producers to sort their product into quality classes based on the results of potentially imperfect tests. The models are designed to obtain thresholds for certification, and provide insights on why finding the optimal thresholds to certify tenderness has been so elusive. An illustration of how to put the framework to work is provided.