West Fraser offered a range of outlooks in its second quarter earnings report. It noted that over the medium term, improvements in housing affordability supported by U.S. government policy changes and the expected normalization of…
Forisk’s 2018 Lumber And Structural Panel Projections
Forisk projects U.S. softwood lumber consumption of 49.8 BBFT in 2018. This represents a 3.7% increase from our 2017 forecast of 48.1 BBFT and is 5.4% higher than 2016 actuals of 47.3 BBFT. Softwood lumber consumption increased every year since 2009, but remained 26.4% below the 2005 consumption high of 64.2 BBFT. U.S. softwood lumber production is forecasted to increase 5.0% to 35.6 billion board feet in 2018. The South drives this growth with production expected to rise 6.4% for the year, reaching 19.5 BBFT. This would be a new high for the region, surpassing the 19.0 BBFT of production in 2005. In the Base Case, the South’s share of national lumber production increases 0.8% to 55% in 2018.
We project U.S. softwood lumber self-sufficiency to reach 71.4% in 2018 as imports level off, and U.S. softwood production increases relative to consumption. We expect net imports to remain stable in 2018 due to physical constraints on Canadian lumber producers and their inability to grow exports to the U.S.: raging wildfires; stagnant softwood lumber capacity; increasing domestic softwood consumption; two new tariffs imposed on Canadian softwood lumber imports in 2017; and reductions in annual allowable cuts.
In the Forisk Base Case, total structural panel consumption increases from 31.2 billion square feet in 2017 to 36.6 billion square feet by 2022, increasing at a compound annual growth rate of 3.3%. OSB consumption is forecasted to rise 3.1% over 2017, while plywood consumption is forecast to increase 2.1%. For context, our October 2017 forecast was 3.5% lower than actuals, while our initial January 2017 forecast came in at 0.9% below actuals.
Each year we update a multi-phased study for estimating U.S. structural panel consumption. We test variables such as housing starts, GDP, population and others. Outputs from all approaches – regardless of the approach, variables, or form – tend to correlate highly with each other. Simpler models performed better; our final model had an R-squared of 0.97. Then, we evaluate results of our model relative to history and research by the USDA Forest Service and APA-The Engineered Wood Association prior to incorporating the projections into price forecasts for timber and delivered logs across the U.S.
From FORISK Consulting: https://forisk.com/blog/2018/03/15/forisk-research-quarterly-frq-excerpt-2018-lumber-structural-panel-projections/