The aim of this study is to analyze the profitability of thinnings at the forest holding level. Analysis is carried out with MELA-software of the following thinning treatments: 1) thinning models of Forestry Center Tapio, 2) no thinnings, 3) postponement of the first thinning to the mean height of 13m (following thinnings according to Tapio's thinning models), 4) thinning models with interest rates of three % and 5) five % (Pesonen & Hirvelae 1992a). The data were collected from the forestry board districts of Satakunta, Pohjois-Karjala and Keski-Pohjanmaa. Basic data consisted of all forest holdings over 5 hectares with forestry plan for 1987-1989, summing up to 5266 forest holdings. Systematic stratified sample weighted by the acreage of forest holdings was taken from the basic data. The final data contained 115 forest holdings from each selected district, altogether 345 forest holdings. Two different strategies represent the objectives of forest owner. A forest owner aiming at the greatest allowable cut based on sustained yield is described by maximization of net present value at the beginning of the planning period with sustainability constraint. A forest owner who is saving cutting possibilites is described by maximization of net present vlue at the end of the planning period with minimum income requirement. The 3% thinning models proved to be the best thinning programme for maximizing net present value both at the beginning and at the end of the planning period. The unthinned alternative resulted in a loss of the net present value in all comparisons. That was due to the high mortality with subsequently reduced potential of the future revenues. Tapio's models maximized saw-timber yield. Early first thinning, small harvest yields and high interest rate decreased revenues as compared to the 3% models and the postponement of first thinning. The size of forest holdings and district had no impact on profitability of the thinning alternatives. Results at the forest level were parallel to those at the stand level (Pesonen & Hirvelae 1992a). The difference between the 3% model and the no thinning -alternative was smaller than what it was at stand level, because the latter involved the whole rotation. At the forest level, sustainability and other contraints evened out the differences between the programmes. Also the timing and areal concentration of thinnings compensate for the differences of programmes at the forest level. Thinnings are essential from the point of financial return. Postponement of the first thinning at least to the mean height of 13m improves the financial return considerably, even if the subsequent thinnings are based on Tapio's models. The 3% model gives the best financial return with greater harvest yields and a shorter rotation than Tapio's model.
