Should systems like LINK be used for long-range forecasts and simulations?

Abstract

Systems like LINK should not generally be used for long-range forecasts and simulations because they eschew many significant long-term issues. World-wide fluctuations in the division of labour should be endogenous, and similarly the various momentous, but partly induced, international 'shocks' like the energy crisis and the breakdown of world monetary order. The same applies for several aspects of the national macroeconometric models making up systems like LINK. Consequently, if such systems are used mechanically to churn out long-range forecasts and simulations, the results will generally be misleading and the inflationary impact of the policy simulations is likely to be underestimated. © 1985.