Modeling local timber harvests using a set of error-related economic relations

by Stone, Trevor M.

Abstract (Summary)
Recent changes in public timber supplies in the Pacific Northwest have increased the

importance of the role private timber plays in the forest products industry and local

communities. Most economic models of timber supply, however, have emphasized

national or regional markets where data are adequate and statistical testing methodologies

relatively well documented. Little attention has been paid to modeling timber harvests at

the local market level. This study attempts to develop an economic model to explain

timber harvests at the county level where previous efforts, which have emphasized a

simultaneous equations approach, have met with poor results.

A set of economic timber harvest relations was tested for eight counties in Northwest

Oregon using the seemingly unrelated regressions (SUR) technique. For industrial

landowners, a present net worth maximization model was used where harvest is a function

of stumpage price, discount rate, and level of growing stock inventory. For non-industrial

private landowners a utility maximization model was used where harvest is a function of

stumpage price, personal income, and level of growing stock inventory. Parameter

coefficient estimates developed using SUR were compared with those developed using

ordinary least squares (OLS) to evaluate the adequacy of the error-related approach.

Results of the study showed significant contemporaneous correlation between harvests in

the counties of the study region for both industrial and non-industrial landowners.

Therefore parameter coefficient estimates obtained using SEJR are more efficient than

those obtained with OLS. The greatest improvements in modeling efficiency were

observed for non-industrial owners. Furthermore, the present net worth maximization

model used for industrial landowners appears to reasonably represent the harvest

motivations of those landowners. However, the high standard errors and poor

explanatory power observed in the non-industrial landowner estimations suggest that the

utility maximization model used for those landowners needs to be re-evaluated.

For both landowners, the level of growing stock inventory plays a large role in

determining timber harvests. Policy makers and analysts interested in predicting countylevel

timber harvests should advocate the collection of more extensive county-level

inventory information than is currently collected. The recording of county-specific price

information would also prove valuable for future analyses by eliminating the need for a

proxy variable.

Bibliographical Information:

Advisor:Johnson, K. Norman; Johnson, Becky; Rettig, Bruce

School:Oregon State University

School Location:USA - Oregon

Source Type:Master's Thesis

Keywords:logging econometric models


Date of Publication:11/15/1995

© 2009 All Rights Reserved.