What were the causes of the statistically large price shocks in the seventeen series? We believe that the changes experienced by the price series can be at be partially explained by three alternative factors. First, demand for timber might have spiked because of the large jumps in U.S. GDP and housing starts at the beginning of 1992. Second, supply of timber might have become restricted due to particularly bad weather for harvesting in the first quarter of 1992, relative to the last quarter of 1991.
Third, there might have been some reporting changes at Timber Mart-South that were independent of the region shifts. If this last factor is responsible for the significant dummy coefficients, then there is reason to be cautious when applying the weighting factors that we have developed in constructing new TMS series from the old series prices. In that case, we would recommend avoiding using the weighting scheme for those seventeen series with identified statistically significant coefficient dummies.
