We compare the performance of top–down (TD) and bottom–up (BU) strategies for forecasting the demand of an item that belongs to a product family. The TD strategy forecasts the sum of the item demands and distributes it to the individual item based upon the historical demand proportion of each item in the family. The BU strategy forecasts each item demand individually using the historical demand data for the particular item. All the item demands, which may be correlated with each other, are assumed to follow a first-order univariate moving average process. As is common in a production-planning environment, the forecasting under both strategies is carried out using the exponential smoothing technique. We show that the performance of the two forecasting strategies is nearly identical, regardless of the coefficient of correlation between the item demands, the items' proportion in the family and the coefficient of the serial correlation term of the demand process. We further investigate the relative performance of the two strategies when a fixed (rather than the optimal) smoothing constant is used for forecasting the demand under both strategies.