Printer Friendly



A practical method for tracking the effect of curtailment announcements on lumber supply is described and tested. Combining announcements of closures and curtailments with mill capacities enables the creation of accurate forward-looking assessments of lumber supply 1 to 2 months into the future. For three American and Canadian lumber-producing regions, the method produced projections of supply that were within 0.5 percent accuracy for a 7-month period encompassing June to December 2000. Announcements of production intentions are an important piece of market intelligence that can help guide assessments of demand-supply in a volatile market environment.

In the ebb and flow of commodity markets, demand and supply are often out of balance and producers and consumers have to continually adjust to uphold equilibrium. Usually the adjustments take the form of changes in price, but when an imbalance is prolonged, physical changes in production also take place.

Such was the case in North America in the year 2000 when rising interest rates dampened demand while past profitability led to a buildup of softwood lumber capacity. Prices began their yearlong decline in January, when the price for benchmark spruce-pine-fir (nominal 2- by 4-inch [standard 38- by 89-mm]) peaked at $350. By May, prices had fallen to $260, a level approaching the break-even point for many mills. At that time, the industry, which had been in full production mode, began to announce limited output curtailments. As prices continued to weaken into the summer and fall, the pace of announcements picked up, and by December more than 240 mills of an approximate universe of 1,250 had announced some form of action to curtail supply. In a volatile market, the item of greatest interest is price and its likely direction. Production curtailment is a response to price weakness that from the producers' viewpoint has reached unfavorable levels. Knowledge of future production (supply) is valuable as a guidepost to likely future price direction, and the question addressed here is how accurate were these announcements as predictors of near-term supply?


As a production curtailment was announced, the number of affected working days was noted. The volume of lumber output foregone was calculated by multiplying the last known annual capacity [1] by the number of closed working days and dividing this by the number of working days in a year (i.e., 250 days minus holidays). These data were aggregated by region and month. Corresponding to each region and month, a baseline aggregate production was calculated by taking the production rate of the previous year (1999), dividing it by the number of working days in 1999, and multiplying that by the number of working days in the current year (2000). The projection for 2000 results from subtracting the calculated foregone output from this baseline. These projections were compared to the production figures as reported a couple of months later by various trade associations and statistical agencies.


Table 1 contains comparative production data between 1999 and 2000, as reported for three key producing regions, and 2000 production figures as projected by this method. Staffing with the West region of the United States, we see that the projections correctly forecasted declines in production compared to 1999, but actual reductions turned out to be even greater. The overestimation of output was on the order of 3 percent for the 7-month period.

For the South, the results were more accurate. A large underestimation in September was offset by a similar overestimation in October, which could have been due to a mill-reporting error to the data-gathering agency's statistician. The overall underestimation error was about 1 percent.

For British Columbia, production was also underestimated. From June through November, the underestimation error was a mere 1 percent, but a large 19 percent discrepancy in December brought the average to a more substantial 4 percent. A review of this with industry observers led to the conclusion that much of the announced downtime at year-end was actually normal holiday-related curtailment rather than a unique measure prompted by prevailing market conditions.


The amount of planned future production is a critical market variable. Data on lumber production are gathered weekly from a subset of mills in the United States and reported with a lag of about 2 weeks. Monthly production figures that include estimates for nonmember mills are reported with about a 2-month delay. Canadian production data are compiled by Statistics Canada and reported with about a 3-month lag. In certain times when markets are weak, announced cutbacks and curtailments may be used to augment these data to indicate prospective market supply. As gathered and compiled during the course of the second half of 2000, these announcements pointed to a 6.9 percent reduction in production in three major regions for the period June to December 2000 as compared to the same period in 1999. When the official data were compiled and reported, they showed a nearly identical 6.5 percent decline.

Curtailment announcements are not definitive indicators of future supply in and of themselves. First, mill capacities are often increased through improvements and de-bottlenecking projects as well as through new construction. New mills are generally known and can be accounted for, but capacity creep or even the change in the number of scheduled hours worked is beyond the radar screens of most observers. To the extent that mills not curtailing production have increased their capacities as compared to that of the prior year, this method overestimates the amount of reduction. Second, some of the announced curtailments may not be completely carried out. If the change in plans is not announced, curtailments will also be overestimated. Third, not all mills report their curtailment intention. In that case, this method leads to underestimation of the reduced output. Additional events that influence supply can include weather-related curtailments, strikes, and chip-disposal difficulties (for integrated mills). All of these can be incorporated in calculations of future supply to the extent the affected mills make known their situation, but over the study period such factors were not reported to have played a role.

The regional evidence from this cycle indicates a mixed bag of effects from these factors. In the aggregate, however, the errors generally cancelled each other out. In the 7 months during which the comparisons were made, the direction and magnitude of changes were closely predicted, with a less than 0.5 percent overall error rate toward overestimating curtailments or underestimating production. With additional follow-up to ensure that announced curtailments are market related rather than normal seasonal events, even better results might have been achieved.

In general, it is in the interest of producers to announce their production intentions. This can help market participants better evaluate the demand-supply balances at a given time and perhaps dampen pricing volatility. Several business and trade news organizations disseminate such announcements efficiently, providing both producers and purchasers with a piece of information that, if detailed and complete, can be used to accurately project supply several months in advance. Especially in periods of market weakness, significant curtailment announcements can have a beneficial psychological effect on buyers and prices, but if abused, such as announcing a routine holiday or maintenance closure as a major market-related measure, that impact of the announcement could be diluted.

The author is an Economist, USDA Forest Serv., Forest Prod. Lab., One Gifford Pinchot Dr., Madison, WI 53705. This paper was received for publication in January 2001. Reprint No. 9238.

[C]Forest Products Society 2001.

Forest Prod. J. 51(7/8):44-46.

(1.) Spelter, H. and T. McKeever. 1999. Profile 1999: Softwood sawmills in the United States and Canada. Res. Pap. FPL-RP-579. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 76 pp.
TABLE 1.--Projected and
actual softwood lumber
production in three regions.
Region and actual/ Monthly production
projected production June July August September
 (MMBF) a
 Actual 1999 1,577 1,466 1,511 1,464
 Actual 2000 1,494 1,226 1,371 1,333
 Projected 2000 1,539 1,329 1,463 1,325
 Percentage of error (%) 3 8 7 -1
 Actual 1999 1,561 1,453 1,361 1,379
 Actual 2000 1,507 1,343 1,424 1,338
 Projected 2000 1,523 1,345 1,356 1,241
 Percentage of error (%) 1 0 -5 -7
British Columbia
 Actual 1999 1,234 1,136 1,096 1,244
 Actual 2000 1,189 1,038 971 1,069
 Projected 2000 1,210 978 933 1,080
 Percentage of error (%) 2 -6 -4 1
 1999 4,372 4,055 3,968 4,087
 2000 4,190 3,607 3,766 3,740
 Projected 4,272 3,652 3,752 3,646
 Percentage of error (%) 2 1 -0 -3
Region and actual/
projected production October November December
 Actual 1999 1,476 1,438 1,407
 Actual 2000 1,480 1,329 1,169
 Projected 2000 1,491 1,347 1,209
 Percentage of error (%) 1 1 3
 Actual 1999 1,463 1,356 1,293
 Actual 2000 1,390 1,270 1,112
 Projected 2000 1,469 1,226 1,104
 Percentage of error (%) 6 -3 -1
British Columbia
 Actual 1999 1,211 1,104 962
 Actual 2000 1,195 1,132 967
 Projected 2000 1,227 1,072 784
 Percentage of error (%) 3 -5 -19
 1999 4,150 3,898 3,662
 2000 4,065 3,731 3,248
 Projected 4,187 3,645 3,097
 Percentage of error (%) 3 -2 -5
(a) MMBF = million
board feet.
COPYRIGHT 2001 Forest Products Society
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2001 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Publication:Forest Products Journal
Article Type:Statistical Data Included
Geographic Code:1USA
Date:Jul 1, 2001

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters