Labor demand and factor substitution in the Western Washington sawmill industry

The sawmill industry in Western Washington has gone through a long period of declining employment per unit output. During the early 1980's the employment pattern showed that the loss of jobs was not being shared equally by both skilled and unskilled workers. The objective of this study is to examine the phenomenon of the "hoarding" of skilled labor during troughs in production cycles in the sawmill industry; to explain the drop in unskilled labor relative to skilled labor during the study period. Dynamic factor demand for heterogeneous labor is modeled with output as a choice variable of the firm. The approach estimates the system of the profit function, in normalized quadratic form, three variable input demand equations and two quasi-fixed input investment equations over the historical period, 1980-1988. Behavior of the model is analyzed for the period 1990-1996 by solving the system given hypothetical price scenarios.;The model employs firm and county level pooled, cross-sectional time series data. Unskilled and skilled labor (both production and non-production workers) are modeled separately as variable and quasi-fixed inputs, respectively, and the adjustment costs of skilled labor and capital are explicitly modeled. Endogeneity of stumpage price and quasi-fixed factor adjustments is tested and simultaneous equation (three stage least squares) estimation is used.;The analysis of demand elasticities and adjustment rates of quasi-fixed factors suggests that skilled and unskilled labor react differently to changes in relative prices. Skilled labor adjusts slowly to optimal levels (fifty percent at the end of two years). Decomposition of elasticities confirms the restricted cost study results that labor is a long run substitute for wood and energy while the unconstrained analysis highlights the importance of output effects. The results imply that policies that affect the price of outputs have the greatest impact on both skilled and unskilled labor, while labor price changes have stronger effects on unskilled labor than on skilled labor. Policies that influence sawlog prices promote moderate changes in labor demand, while changes in capital price have the smallest influence on labor. The analysis confirms the existence of technological change in the industry in the past, but no statistically significant direct linkage between technological change and unskilled labor use was found.