Abstract: | The great variability and inaccuracy of past energy supply and demand forecasts suggests that they are an unreliable basis for policy analysis and planning. This paper argues that unreliability cannot be eliminated from these analyses by increasing their sophistication, since many of the problems involved in forecasting inhere in the techniques themselves or in the institutional context within which they are used. This argument is illustrated by an examination of the natural gas forecasts of the Canadian National Energy Board. |