This paper discusses the stability of a multi-sector inventory cycle model incorporating inventories of inputs and, to some extent, of outputs. It emphasizes the effect on stability of adopting new assumptions about sales forecasting behavior; these new assumptions are embodied in a "customers' demand" forecasting model, presented in Section 1. It is argued that stability conclusions are somewhat sensitive to the forecasting assumptions adopted; so conclusions should be drawn with caution, given our lack of knowledge concerning actual forecasting techniques used by business.