Essays on forecast updating in a two-stage supply chain

Many companies, in such industries as high technology and fashion, face highly uncertain demand for their products. With uncertain demand and the need to make ordering and production decisions before demand is realized, companies can end up with not enough product to satisfy demand or too much inventory. Companies often respond to this problem by generating forecasts of demand that are periodically updated based on demand information that has become available to the company in order to create more accurate forecasts of demand.; As companies engage in forecast updating, they generally believe that generating more accurate forecasts of demand is a good thing, not only for the company who directly faces customer demand, but for the supply chain as a whole. In this research, we challenge these commonly held beliefs. We study two-stage supply chains. The main results include the following. (1) We find that better end product demand information does not always increase the expected profits for the downstream stage and the supply chain as a whole. We propose a contract so that the expected profit for the supply chain as a whole increases with better demand information. (2) We find that forecast updating can increase the volatility of production at the upstream stage and we propose a policy of using old forecasts at the downstream stage in order to reduce the volatility of production. We find that our policy can reduce the volatility, of production and surprisingly, produce very small increases in the expected system wide inventory and shortage costs compared to a system wide optimal policy. (3) We find that the upstream stage may not be motivated to help the downstream stage obtain the most accurate unbiased forecast of end product demand. We show that a standard buy back contract can align the incentives for both stages and achieve system wide optimal expected profit.; Our results suggest that it may not be wise for the downstream stage to focus on improving forecast accuracy without considering its impact on supply decisions, production volatility and supplier incentives.