For the second case on sharing information for collaborative forecasting, known as collaborative planning, forecasting and replenishment (CPRF), usually involves two parties, the manufacturer and the retailer. The collaborating parties would jointly generate a forecast and plan for that forecast. The desired effect would be to make the supply chain more efficient since the forecast is coordinated and carried more information. Yossi [54] studied a two-stage supply chain involving a supplier and a retailer. He created two models, (1) a decentralized structure where each member performs local forecasting and integrates adjusted forecasts into his replenishment process and (2) a centralized structure where the two members jointly forecast and update, and compared the two models with a benchmark model where forecasts are not integrated with the replenishment process. In the following year, Yossi [55] studied the case of auto-correlated demand on the same two-stage supply chain. He created three models, (1) retailer and supplier coordinate their policy parameters but do not share observations, (2) supplier manages the supply chain’s inventory without information of retailer’s observations, and (3) full sharing of observations with collaborative forecasting. The insight derived was, VMI and CPRF becomes more important as the demand process is more correlated across time, and as company’s ability to explain the demand uncertainty through early demand information improves.