Abstract: Driven by the government's mechanism of reward-penalty, many manufacturers are engaging in collecting and remanufacturing operations. For the members of the closed-loop supply chain with the manufacturer collecting, they can forecast the uncertain market demand through using information technology. Generally, the forecast information is the private information of the retailer, which results in demand forecast information asymmetry in the closed-loop supply chain with the manufacturer collecting. In order to solve this problem, the demand forecast sharing decisions of the retailer is studied. Specifically, the following questions are mainly investigated:When both the manufacturer and the retailer can forecast, whether the retailer shares the forecast information voluntarily? If impossible, how to design a mechanism to induce it to share information? How does the government's mechanism of reward-penalty affect the equilibrium of information sharing? And how does the forecast accuracy affect the supply chain members' profits and the value of information sharing? A Stackelberg model between the manufacturer and the retailer is constructed. The manufacturer and the retailer first negotiate on information sharing agreement through a bargaining mechanism. And then the manufacturer acts as a Stackelberg leader and determines the wholesale price and the collection rate. As a follower, the retailer determines the retail price. This paper solves the multistage game by using a standard backward induction technique. In detail, the equilibrium decisions of the manufacturer and the retailer are first solved under the case of information sharing and non-information sharing, and then the firms' ex ante profits of both cases are computed, thereby obtaining the value of information sharing. Based on these results, this paper analyses the impact of demand forecast accuracy on the supply chain members' profits and the value of information sharing, and the values of information sharing and equilibrium information sharing decisions are investigated. The results show that in most cases, both the supply chain members can benefit from the precise demand forecasts. Besides, when the manufacturer's collection efficiency is high, voluntary information sharing is a unique equilibrium. When the efficiency is low, non-information sharing is a unique equilibrium. When efficiency is medium, the equilibrium of information sharing can be realized through a bargaining mechanism. Furthermore, the government's level of reward-penalty affects the equilibrium information sharing decisions. The research provides useful managerial insights to the government and supply chain managers in terms of sharing information, and collection and remanufacturing operations.

Key words: closed-loop supply chain, mechanism of reward-penalty, demand forecast, information sharing