Optimizing loblolly pine management with stochastic dynamic programming
MetadataShow full item record
Two approaches are taken to compare site preparation intensities: a quasideterministic approach, where expected cash flows are discounted with risk-adjusted discount rates, and a stochastic approach, where probability functions of cash flows are used to maximize expected utility from net present values. The stochastic approach is further divided into non-adaptive scenarios and adaptive scenarios, where the investor can gather additional price information during the life of a stand to optimize the harvest decision. The adaptive management problem is solved with stochastic dynamic programming. For each possible harvest age, an optimal reservation price below which the forest landowner should not sell the stumpage is calculated.
The study shows that the use of a single risk-adjusted discount rate is generally inadequate to compare different management intensities. The stochastic approaches reveal that the optimal management intensity depends on the degree of risk aversion, with increasing risk aversion leading to a lower intensity level. Given the possibility of catastrophic losses, the adoption of a feedback harvesting policy strengthens the already dominant influence of risk aversion and does not generally lead to an increase in management intensity.
The study's results suggest that even if the landowner is managing the forest solely for financial reasons, some of the reluctance to invest in intensive forestry may not indicate a lack of interest or information but simply an economic reaction to risk, especially in regions with a high potential danger of catastrophic losses.
- Doctoral Dissertations