Accounting for uncertainty in a forest sector model using Monte Carlo simulation

Large scale forest sector models tend to have thousands of parameters, of which many have values that are unknown or not known exactly. Despite the wide use of such models in the policy analyses, the impact of the overall uncertainty on the results and the distribution of the resulting variable values have not been considered. it would be important for the users of the results to know the robustness of the model outputs and conclusions made to variations in model inputs. This study attempts to systematically account for the uncertainties in forest sector model analysis. Monte Carlo simulation is applied to a spatial partial equilibrium model for the Finnish forest sector in order to examine how the uncertainty over the parametric data and output price developments affects the model projections. The focus is on the sawlog market and sawnwood production. As a policy example, the impacts of forest conservation set-asides are examined in the presence of these uncertainties. The uncertainty in the basic parametric data seems to have rather moderate impact on the results. Instead, the unpredictability of the world market prices for the forest industry products is an important source of variation among the model projections. Hence, any single model projection should be perceived to be strongly conditional to the assumptions made concerning the developments of world markets. The model projections for the relative impacts of forest conservation seem to be rather robust despite the uncertainties. (C) 2009 Elsevier B.V. All rights reserved.