AbstractIn the forestry literature, stochastic extensions of the classic Faustmann model have become the predominant models of optimal harvesting on even‐aged timber stands. A recent attempt to estimate a stochastic version of the Faustmann model left several puzzling results unexplained (Provencher, 1995). This paper reports on new estimation and analyses undertaken to resolve these puzzles. Of particular note is the result that the value of using correct price forecasts in the harvest decision is much lower than indicated by recent simulation studies, perhaps because variability in convenience yield dampens the expected gain from exploiting stochastic variation in timber price