Search

The article analyzes the possibility of using a seven−year development trend model to forecast the average price of wood sales (CGUS). Price models and forecasts were made on the basis of three different time series covering 7−year periods. The forecast of timber prices in individual years was developed on the basis of extrapolation going beyond the set of information adopted for research used to build the development trend model. We found that the size of the indicators of determination of the analyzed functions ranged from 0.5 to 0.6, and the convergence rates from 0.5 to 0.4. This means that the linear functions were characterized by poor adjustment to real prices. Despite this, the timber price forecasts differed from the empirical prices from PLN 1.81 to 13.91. In relative categories, the forecast error was in the range from 0.92 to 7.28%. Therefore, in our opinion, the research process presented can be a reference point as a comparative element to verify forecast results for other (non−linear) price forecasting models. The process of modeling timber raw material prices should be extended to other predictors related to the forest marketing chain. The predicted results from these models cannot be worse than those obtained by forecasting using linear functions., Sylwan 164 (3): 206-215