Spatial demand forecasts are hardly discussed in the scientific literature. In the real estate literature a lot has been written about forecasting, but most of these studies are about prices and hardly about the amount of real estate that is needed. In this paper we analyse the demand forecasts for industrial land in the Netherlands. These forecasts are part of the Dutch planning process. For each new industrial estate or office park, forecasts have to be made which should provide evidence that there is sufficient demand for such new sites.

At first we compare demand forecasts with actual demand. It is concluded that most forecast overestimate demand and that this more often occurs on lower than on higher spatial scales. Secondly, we try to find an explanation for the inaccuracy. This is done by analysing various demand forecasts models on national, regional and local level. This analysis is completed by a review of various second-opinion reports on forecasts and reports by regional and local Courts of Audit about the development of industrial estates. We conclude that the main explanation for the inaccuracies is the assumed success of economic growth policies in the models (either by choosing high growth scenario’s or by assuming successful policy interventions). This implies strategic behaviour of forecasters, planners and politicians and shows resemblance with research on the accuracy of travel demand forecasts.