TY - UNPB

T1 - The impact of forecasting on the bullwhip effect

AU - Stamatopoulos, I

AU - Teunter, R H

AU - Fildes, R A

PY - 2006

Y1 - 2006

N2 - There has been strong empirical evidence that demand variability increases as one moves up the supply chain (from the retailer to the raw materials supplier), a phenomenon called bullwhip effect. This paper examines the bullwhip effect and in particular one of its main causes, demand forecasting. Key observations for the studies that deal with the impact of forecasting on the bullwhip effect are that: (1) all allow negative demands as well as negative orders for analytical tractability and (2) none considers the best exponential smoothing forecast without a prefixed smoothing constant. This paper validates the main findings in the literature when negative demands and negative orders are not allowed, using simulation. The main contribution is the inclusion of 'best' exponential smoothing as a forecasting method. This method is shown to explain some structural differences in bullwhip effect that have been observed in comparisons between naïve exponential smoothing and optimal forecasting. Therefore, it provides an important alternative to naïve smoothing for use in practice, especially as it is included in some of the more modern Demand Planning Systems.

AB - There has been strong empirical evidence that demand variability increases as one moves up the supply chain (from the retailer to the raw materials supplier), a phenomenon called bullwhip effect. This paper examines the bullwhip effect and in particular one of its main causes, demand forecasting. Key observations for the studies that deal with the impact of forecasting on the bullwhip effect are that: (1) all allow negative demands as well as negative orders for analytical tractability and (2) none considers the best exponential smoothing forecast without a prefixed smoothing constant. This paper validates the main findings in the literature when negative demands and negative orders are not allowed, using simulation. The main contribution is the inclusion of 'best' exponential smoothing as a forecasting method. This method is shown to explain some structural differences in bullwhip effect that have been observed in comparisons between naïve exponential smoothing and optimal forecasting. Therefore, it provides an important alternative to naïve smoothing for use in practice, especially as it is included in some of the more modern Demand Planning Systems.

KW - bullwhip

KW - supply chain

KW - forecast mis-specification

M3 - Working paper

T3 - Management Science Working Paper Series

BT - The impact of forecasting on the bullwhip effect

PB - The Department of Management Science

CY - Lancaster University

ER -