Estimating Regional Softwood Lumber Supply in the United States Using Seemingly Unrelated Regression

In this article, we present estimates of regional softwood lumber supply functions in the United States using annual time series data for 1959 to 2009. Seemingly unrelated regression is used in a profit maximization framework to model softwood lumber supply as a function of lumber and stumpage prices, lagged supply, wage rate, and interest rate for the eastern and western United States. The effects of listing the northern spotted owl (Strix occidentalis caurina) as a threatened species and the US–Canada softwood lumber trade dispute are controlled for in empirical estimation. Results show that regional lumber supply is quite inelastic to lumber price and that stumpage price and bank prime rate negatively influence regional lumber supply. Results also suggest that present market supplies of softwood lumber have potential expansionary influence on future supplies, that listing of the northern spotted owl in 1990 reduced the lumber supply in the western region during subsequent years, that the US–Canada softwood lumber trade dispute/agreements favored regional lumber production in the United States during the period from 1996 to 2005, and that supply has declined during the recent period of economic recession. One of the most interesting issues in forest economics is concerned with the nature and extent of the market for timber (Uri and Boyd 1990). The forest product market in the United States is dominated by lumber. The estimated total lumber consumption and production in the United States in 2005 were 11.2 billion cubic feet and 7.9 billion cubic feet, respectively (Howard 2007). In that year, lumber consumption represented about 52 percent of the total forest products consumed, and lumber production represented about 47 percent of the total forest products produced. Softwood lumber has the largest share in the US lumber market. In 2005, 73 percent of the total lumber produced and 83 percent of the total lumber consumed was softwood lumber (Howard 2007). In the same year, softwood lumber imports to the United States were 3.6 billion cubic feet, or about 97 percent of the total lumber imports. Softwood lumber is mainly used in the housing construction sector in the United States. Examining the softwood lumber market is thus of significant importance to economic studies in forestry. The market for timber and softwood lumber has been widely investigated. For example, Hamilton (1970) estimated aggregate supply of National Forest timber. Supply elasticity specific to the California pine region was estimated by Frazier (1967). Robinson (1974) studied the markets for Douglas-fir (Pseudotsuga menziesii) and southern pine (such as Pinus taeda, Pinus echinata, and Pinus palustris) lumber. Luppold (1984) examined the US hardwood lumber market. Adams and Haynes (1980) were the first to estimate regional softwood lumber supply and demand functions under the Timber Assessment Market Model (TAMM). Regional estimates of aggregate lumber supply and demand functions, however, remain rare. Given the possible economic linkages between different regions in the country, simultaneous estimation of regional supply and demand using updated information is also lacking in the literature. We examine the supply side of the softwood lumber market in the United States at regional levels. Supply functions for softwood lumber are estimated for two regions of the United States: the eastern (USE) and the western (USW) United States. The USW includes Alaska, Arizona, The authors are, respectively, PhD Student, Professor, and Associate Professor, School of Forestry and Wildlife Sci., Auburn Univ., Auburn, Alabama ([email protected], zhangdw@auburn. edu, [email protected]). This paper was received for publication in December 2010. Article no. 10-00071. Forest Products Society 2010. Forest Prod. J. 60(7/8):709–714. FOREST PRODUCTS JOURNAL Vol. 60, No. 7/8 709 California, Colorado, Hawaii, Idaho, Montana, Nevada, New Mexico, Oregon, South Dakota, Utah, Washington, and Wyoming. The USE includes the remaining states in both the southern and northern United States. The regional divisions are based on the volume of softwood lumber supply. The USW has traditionally been the largest softwood lumber producing region (Howard 2007). Although the share of the southern United States has steadily increased since listing of the northern spotted owl (Strix occidentalis caurina) as a threatened species in 1990, the average softwood lumber production in the USW was about 1.5 times higher than the average production in the southern United States between 1965 and 2005. The average production in the northern United States was less than one-tenth of that in the USW during the same time period. Thus, instead of treating the northern United States as a separate softwood lumber producing region, the southern and northern United States are combined to form the USE. The softwood lumber market in both the USE and the USW are likely to be subject to spillovers from economywide or worldwide shocks contemporaneously. Thus, the supply equations of the two regions may be linked by the fact that their disturbances are correlated. For this reason, we use Zellner’s (1962) seemingly unrelated regression (SUR) model to estimate the supply coefficients of the two regions. Regional supply functions are estimated using a profit maximization framework and time series data for 51 years, from 1959 to 2009. Price elasticities of supply are estimated with respect to both output and input prices. We also examine the effects of listing the northern spotted owl as a threatened species and the US–Canada softwood lumber trade disputes/agreements on regional supply of softwood lumber in the United States.