In this paper, the bullwhip effect in a multistage supply chain has been analyzed using double moving average and double exponential smoothing forecasts. We find out that double exponential smoothing and double moving average forecasts can sometimes have a variance reducing effect on the supply chain. In particular, the bullwhip effect exists for positively correlated or independent and identically distributed (i.i.d) demands. However, for negatively correlated demands, bullwhip effect does not necessarily exist. In addition, our analysis shows that the order variance based on simple moving average forecast is always larger than the order variance based on double moving average forecast, and the order variance based on simple exponential smoothing forecast is always larger than the order variance based on double exponential smoothing forecast. For the independent and identically distributed (i.i.d) demands , it can be shown that the order quantity fluctuation range for double exponential smoothing is greater than that for double moving average.