Tests for Long-Run Relationships in Hardwood Lumber Prices

Abstract

Abstract Hardwood lumber prices are unique because of the large number of marketable species and variability of prices across species. Previous research showed that long-run fashion decisions regarding species selection may be influenced by price, so the interaction between fashion and species price may act to keep prices (hence, demand) of different hardwood species together in the long run. To test this hypothesis, we examined the joint lumber price behavior of six major hardwood species representing different appearance characteristics in the Appalachian hardwood region. Bivariate and multivariate price cointegration tests within lumber grades of these mainly nonstationary price series, conducted using a consistent vector error-correction rank and lag-order model selection procedure, revealed no stable long-run statistical relationships, rejecting the principal null hypothesis. Current relative price levels therefore cannot be used to infer future relative levels. Supplementary vector autoregressions of mostly differenced series, however, indicate that some interspecies price relationships exist. Such relationships, however, were mostly confined within appearance groups and only rarely across groups.