Effects of monetary shocks on the dynamics of stumpage price and timber harvest

The traditional objective of timber forest management is the maximum sustained yield, which is consistent with economic efficiency only when the interest rate is zero and the timber price grows at an average speed of interest. It is known as the Hotelling rule. We test whether the Hotelling rule holds by using benchmark model and time trend model. The benchmark model shows that the changes in the stumpage price associate negatively with interest rate. The Hotelling rule requires a positive association. The time trend model confirms that in the long run timber price evolves at a speed similar to the interest rate.; Further study was focused on the stumpage prices and timber harvests of the Pacific Northwest Region. This region is divided into two sub-regions, the west side and east side. The west side is dominated by the Douglas fir species, while the east side is dominated by the Ponderosa pine species. Both sub-regions employ competitive auctions for the timber sales.; Structural Vector Auto Regression analysis reveals the following results. Shocks in the macroeconomic variables, timber demand, and timber supply affect the dynamics of the timber harvests and stumpage cut prices. Shocks in the aggregate price and aggregate output increase timber harvests, while shocks in the money demand decrease timber harvests. Response of the timber harvest to the monetary policy shocks is mixed; rises for the west side, but falls for the east side. Stumpage cut price reacts negatively to a positive shock in the aggregate price, monetary policy, and money demand. The reaction to the aggregate output is mixed for both the west and east sides. Shocks in the timber demand result in an increase in the timber harvest and stumpage cut price. Shocks in the timber supply cause the timber harvests to fall, but the stumpage cut price to rise. Under complicated responses of the timber harvest and stumpage cut price to a shock, a forecast using Bayesian Vector Auto Regression outperforms the one using Unrestricted Vector Auto Regression.