Leonard N. Stern School of Business, New York University

Department of Information, Operations & Management Science

Operations Management Research Seminar

Topic: "Strategic Commitment for Optimal Capacity Decision Under Asymmetric Forecast Information*"

Speaker:   Ozalp Ozer, MS&E, Stanford University
Date: Friday, November 7, 2003
Time:   2:30 pm - 4:00 pm
Place: 5-80 KMC

Abstract. Due to long procurement lead-time, suppliers typically decide on capacity, such as the purchase of a critical component, before the final production at the manufacturer’s site. Under such a production scheme, the initial capacity decision has significant impact on the profits, in particular for short product life cycle industries, such as fashion and high-tech. To make the right capacity decision, it is crucial for the supplier to have accurate demand information. It is often the case, however, that manufacturer has more accurate forecast. To make the matters worse, she also has every incentive to inflate the forecast to secure more capacity if the supplier were to ask for her forecast information.

In this talk, we will present a model for a supply chain that faces the above problem. We will address how different contracts affect the optimal capacity decision and hence the profitability of the supplier and the manufacturer. In particular, we will consider four contracts -- the wholesale price contract, the pay back contract, the capacity reservation contract, and the advance purchase contract. The wholesale price contract is the most prevalent contract in practice because of its simplicity. But the simplicity comes with a price: it does not achieve credible information sharing, hence resulting in high inefficiency. Advance purchase contract enables the manufacturer to signal her forecast update (a signaling game) while the capacity reservation contract enables the supplier to “smoke out” this information from the manufacturer (a mechanism design). Both of them achieve credible information sharing hence improving the efficiency of the system. But which is better depends on the problem setting such as degree of forecast information asymmetry and the capacity acquisition cost.