Evaluation of U.S. southern pine stumpage market informational efficiency
The literature on informational efficiency of southern timber markets conflicts. Part of this conflict is because of differences in how efficiency was tested. In this paper, price behavior tests are based on deflated ("real") southern pine (Pinus spp.) sawtimber stumpage prices, using some of the same data and tests used in previous research and some new data and tests. Here, different results are found in many cases regarding price behavior, as compared with the existing literature. Using a valid and consistent data-based model selection procedure, augmented Dickey-Fuller tests cannot reject a ' of a unit root for most deflated monthly and all quarterly southern pine timber price series evaluated. Regressions of long-term deflated timber price ratios on their own lags lead to results similar to those offered by other authors when not corrected for bias but produce fewer similarities when bias is addressed. The results of those regressions support a contention that most of the monthly series contain nonstationary as well as stationary components and that quarterly prices tested in this framework using data through 2001 are closer to pure nonstationary processes. These results have implications for harvest timing approaches that depend on serial dependence of timber prices, provide support for certain kinds of policy and catastrophic shocks modeling procedures, and address the valid- ity of statistical approaches best suited to evaluating interconnections among timber markets.