W e study a manufacturer's optimal multiple-sourcing strategies when some but not all suppliers face risks of complete supply disruptions. Using an approximate model, we show that the optimal unreliable orders are ranked by a simple and intuitive criterion, and are invariant of minor market size changes. Furthermore, when ordering from one reliable and one unreliable supplier, we show that the total order quantity and its allocation between the two suppliers are independent decisions. We then test and confirm the robustness of the insights without the approximation, as well as when we relax various assumptions.

We consider sealed- and open-bid total-cost procurement auctions where two attributes are used for contract award decisions: price, which is bid by the supplier, and a fixed cost adjustment, which is included by the buyer to capture nonprice factors such as logistics costs. Suppliers know only their own true production cost and their own cost adjustment, and the buyer does not know the suppliers' true production costs but does know all suppliers' cost adjustments, which she herself sets in order to make an informed total-cost decision. The buyer, who seeks to minimize her total (price and cost adjustment) procurement cost, can choose to run a first-price sealed-bid auction, where suppliers' bids are affected by their beliefs about each other's total costs, or a descending open-bid auction, where only the actual realizations of suppliers' total costs drive the auction outcome. We characterize the buyer's choice between the two formats as a threshold decision over suppliers' cost adjustments and analyze the effect of supplier beliefs on her decision. We also study the impact of additional suppliers on the buyer's decision, the effect of correlation between suppliers' production costs and their cost adjustments, and additive as well as multiplicative total-cost functions. The results suggest that procurement managers can use their evaluations of suppliers' cost adjustments to make better auction format decisions.procurement auctions, informed buyer, information disclosure

We study a supply chain consisting of one supplier and one OEM (original equipment manufacturer). The OEM faces stochastic demand for a final product that requires assembly of two major components, one of which is procured exclusively from the supplier. In the absence of competition, the supplier is able to make a take-it-or-leave-it offer to the OEM in the form of a menu of price-quantity contracts. The OEM possesses private information across two dimensions: (1) demand forecasts about the final product, and (2) production cost of the in-house component. Both pieces of information are relevant to the total supply chain profit, thus affecting the supplier's optimal offer. By initially assuming an exogenous information structure, we characterize the supplier's optimal contract menu for a simple case and demonstrate that more dimensions of asymmetric information are not always preferable for the OEM but could be beneficial for the supply chain. We subsequently examine whether this preference for one less dimension of private information implies disclosure of private information to the supplier when the information structure is endogenized. Our results indicate that if OEMs that are indifferent between disclosing and keeping information private choose to disclose it, disclosure of any verifiable information from all OEMs is always an equilibrium, whereas nondisclosure might fail to be an equilibrium. We also consider the possibility of the OEM and the supplier contracting at the ex-ante stage, i.e., before the OEM observes his private information. When both dimensions of the OEM's private information are verifiable and the cost of disclosing information is small enough, an ex-ante agreement on information disclosure is always possible; otherwise its feasibility depends on the problem parameters.

We study the pricing and capacity allocation problem of a service provider who serves two distinct customer classes. Customers within each class are inherently heterogeneous in their willingness to pay for service, but their utilities are also affected by the presence of other customers in the system. Specifically, customer utilities depend on how many customers are in the system at the time of service as well as who these other customers are. We find that if the service provider can price discriminate between customer classes, pricing out a class, i.e., operating an exclusive system, can sometimes be optimal and that depends only on classes' perceptions of each other. If the provider must charge a single price, an exclusive system is even more likely. We extend our analysis to a service provider who can prevent class interaction by allocating separate capacity segments to the two customer classes. Under price discrimination, allocating capacity is optimal if the "net appreciation" between classes, as defined in the paper, is negative. However, under a single-price policy, allocating capacity can be optimal even if this net appreciation is positive. We describe in detail how the nature of asymmetry in classes' perception of each other determines the optimal strategy.

No abstract