Hierarchical screening for capacity allocation in distribution systems

We consider the capacity allocation in a decentralized distribution system with a continuum of retailers, a number of distributors, and a supplier. Retailers privately observe their local demands, and distributors are equipped with information technology that facilitates finer prediction of retailers’ demands. The supplier builds capacity, but observes neither local demands nor the precision of distributors’ technology. We show that a distributor’s profit weakly increases in capacity and precision, but she cannot capitalize on her technology with sufficiently small capacity. High production cost may induce the supplier to accept retailers’ orders directly. With low production cost, the supplier delegates more capacity to distributors who know local demands better, and no distributor is excluded. We then numerically investigate scenarios with two types of distributors and uniform distribution of retailers’ type. With moderate production costs, the supplier should offer a wholesale price contract, even though some distributors do utilize technology to screen retailers. Quantity discount contracts are offered to distributors when production cost is low, and the supplier may extract full surplus from all distributors when technology is relatively common. Retailers need not receive more capacity when distributors have better technology, and quantity allocation among retailers may be non-monotonic in local demand.