A global vector autoregression model for softwood lumber trade.
Abstract A novel econometric approach is developed in this chapter, namely, the Global Vector Autoregressive (GVAR) model. It provides a comprehensive framework for analyzing the country-level impacts of various domestic, foreign, and/or global shocks on softwood lumber trade. The GVAR approach is applied to Canada-U.S. trade in softwood lumber and used to analyze the effect of external shocks on Canadian lumber prices. Findings indicate that Canada's export prices are positively correlated to U.S. housing starts and real GDP. Further, using impulse response functions, it is used to examine the effects on regional lumber export prices in Canada of: (1) a change in U.S. housing starts; (2) a reduction in U.S. GDP by one standard deviation; (3) a COVID-19 induced decline in U.S. GDP (of three standard deviations); (4) an increase in global oil prices; and, in the Appendix, (5) an increase in the long-term interest rate. Price impacts vary a great deal by Canadian region depending on the type of shock, with the propagation mechanism in Alberta significantly different from that in other regions. For example, with an oil price shock and because Alberta is a major exporter of oil, the lumber export price remains high even as the shock dissipates over time.