2004 journal article
Forest forecasts: does individual heterogeneity matter for market and landscape outcomes?
FOREST POLICY AND ECONOMICS, 6(3-4), 243–260.
author keywords: timber supply; amenity demand; wildlife habitat; forest inventory projection; southern US forests
UN Sustainable Development Goal Categories
15. Life on Land
(Web of Science; OpenAlex)
Source: Web Of Science
Added: August 6, 2018
Recent econometric analyses have shown that timber supply choices reflect heterogeneous preferences for amenities and management of forests in the US South. However, this evidence is insufficient to determine whether timber market models that rely on conventional timber supply specifications will suffer from significant forecasting biases. The goal of this paper is to evaluate the nature and extent of such bias by (a) modifying the Sub-Regional Timber Supply (SRTS) model to reflect landowner heterogeneity; and (b) using estimated parameters to tie timber markets to heterogeneous individual supply choices. We find that conventional models will underestimate the ending period inventory volume in the younger age classes of all forest management types, except planted pines. These aggregate results mask interesting sub-regional patterns, as exemplified by mixed-pine forests of Virginia mountains, Florida panhandle, and North Carolina mountains, and natural pine forests of North Carolina piedmont. Compared to empirically valid models, conventional models will also estimate (a) lower timber prices, higher harvests and substantially higher inventory for softwood species; and (b) higher prices, lower harvests, and higher inventory for hardwood species. A case study from North Carolina also indicates significant differences in habitat forecasts for 61 species of birds, amphibians, and reptiles. We conclude with a synthesis of the key underlying forces that supplement or mitigate the heterogeneity impact, and a discussion of the bias-vs.-efficiency tradeoffs confronting policy makers and policy analysts who rely on forest sector projection models.