Forecasting home appliances sales incorporating adoption and replacement

Empirical research on diffusion of consumer durable goods has mostly focused on adoption disregarding the importance of replacement sales. This article develops a demand model that incorporates both elements. Unlike previous studies, we allow adoption to depend on several economic variables (e.g., product price, disposable income, energy prices, and new private housing starts), and compute replacement sales from micro data rather than from an arbitrary survival function. The model was adjusted to U.S. data of electric heaters for 1946–1995, and shows that sales forecasting can be improved by allowing adoption to depend on various economic factors.