Working Paper 9

Jobs, Revenues, and Taxes from Timber Harvest;
An Examination of the Forest Industry Contribution to the Washington State Economy

September 2007

C. Larry Mason
and
Bruce R. Lippke

Rural Technology Initiative
College of Forest Resources
University of Washington
Box 352100
Seattle , WA 98195-2100
www.ruraltech.org

Acknowledgements

This report represents a synthesis of information and analysis provided from many sources. A model of economic multipliers linked to annual timber harvest activity has been developed. The work of the research team included review of the literature, surveys of industry sectors, interviews with individuals, companies, and other organizations, collection of census employment data, development of estimated economic impacts by forest industry sector and indirect employment associated with timber harvests by state timbersheds using the Washington Input/Output Model, and the synthesis, analysis, and report of all collected information and modeled outputs. Members of the research team included Larry Mason, Project Coordinator for The Rural Technology Initiative (RTI) and Bruce Lippke, Economics Professor at the University of Washington College of Forest Resources and Director of RTI. Administrative support was provided by Clara Burnett.

Critically important to the completion of this project were the countless conversations with loggers, mill managers, timber merchandisers, consultants, scientists, other forestry professionals, and agency representatives that provided inside information and professional insights in regards to current harvesting, manufacturing, and employment developments within the region.

A list of the most prominent information contributors presented in alphabetical order includes: Phil Aust (Washington Department of Natural Resources), William Beyers (Professor, Geography, UW), Brian Boyle (Northwest Environmental Forum), Rose Braden (Research Consultant, Center for International Trade in Forest Products, UW), John Calhoun (Director, Olympic Natural Resource Center, UW), Richard Conway (Dick Conway & Associates), Ivan Eastin (Director, Center for International Trade in Forest Products, UW), Steve Fredericks (Washington Department of Revenue), Indroneil Ganguly (Research Associate, Center for International Trade in Forest Products, UW), Peter Heide (Washington Forest Protection Association), Rick Lockhart (Washington Employment Security Department), Cassandra Moseley (Ecosystem Workforce Program, Institute for a Sustainable Environment), John Perez-Garcia (Professor, Forest Economics, UW), James Restivo (U.S. Dept of Census), Bill Pickell (Director, Washington Contract Loggers Association), David Richards (Washington Department of Natural Resources), Alex Roubinchtein (Washington Employment Security Department), Jon Tweedale (Washington Department of Natural Resources), Debra Warren (Economist, USDA PNW Research Station), Chris Westwood (Washington Department of Revenue), Tom Wolford (Director, Washington Pulp and Paper Foundation).

Special thanks go to the many company representatives that took time from their busy schedules to respond to the project surveys and interviews.

This work was made possible by a research grant through the USDA Pacific Northwest Research Station and administered by the Olympic Natural Resource Center of the University of Washington . Additional support was provided through the Future of Washington Forests Project funded by the Washington State Legislature and administered by the Washington Department of Natural Resources.

Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the funding agencies or project cooperators.

Abstract

Shifting timber harvest levels, locations, management treatments and log qualities influence both the level and location of employment for forestry industry companies which, in turn, precipitate fluctuations in downstream economic activity. Adjustments to industrial infrastructure bring implications, important to local and regional policy makers; for rural as well as urban economies. Economic models constructed to estimate such impacts for Washington State are known to be dated and may not accurately capture important job, revenue, and tax ramifications associated with changing forest management treatments, mill locations, workforce availability, harvest volume fluctuations, and other factors. This project creates updated multiplier estimators with which to calculate direct and indirect economic impacts from changes in forest industry activity in Washington State . Surveys and interviews of logging contractors, sawmills, log truckers, paper companies, and state agencies were undertaken and the resulting information was combined with North American Industry Classification System (NAICS) employment data from the United States Census Bureau to create a forest sector transaction matrix of purchases and employment based upon the Washington Input/Output Model. Management alternatives, such as expanded use of thinning, and processing changes, have been examined and described so that parties interested in specific regional impacts can customize the results for local application. The unique contributions to employment and state revenues that result from Washington Department of Natural Resources timber sale programs are highlighted.

Keywords: Forest Industry, Economic Multipliers, Washington Input/Output Model, Washington Department of Natural Resources

Executive Summary

Shifting timber harvest levels and activity locations influence the distribution of employment within the forest sector, which, in turn, precipitates fluctuations in downstream economic activity. Adjustments to industrial infrastructure bring implications, important to local and regional policy makers, for rural as well as urban economies. Economic models constructed to estimate such impacts for Washington State are known to be dated and may not accurately capture important job, revenue, and tax ramifications associated with changing forest management treatments, mill locations, workforce availability, harvest volume fluctuations, and other factors. This project creates updated multiplier estimators with which to calculate direct and indirect economic impacts from changes in forest industry activity in Washington State . Surveys and interviews of logging contractors, sawmills, log truckers, paper companies, and state agencies were undertaken and the resulting information was combined with North American Industry Classification System (NAICS) employment data from the United States Census Bureau to create a forest sector transaction matrix of purchases and employment based upon the Washington Input/Output Model. Management alternatives, such as expanded use of thinning, and processing changes, have been examined and described so that parties interested in specific regional impacts can customize the results for local application. The unique contributions to employment and state revenues that result from Washington Department of Natural Resources timber sale programs are highlighted.

Employment, revenue, and tax multipliers

Basic forest industry employment classifications that have appeared in prior studies are Primary Wood, Secondary Wood, and Primary Paper. Primary Wood includes logging and primary processing such as lumber and veneer manufacturing. Secondary Wood includes door, molding, furniture, and other downstream wood product manufacturers that purchase intermediate products from Primary Wood processors. Primary Paper includes pulp and paper manufacturers whose products are generated from raw materials such as pulp logs and chips that are purchased as intermediate products from Primary Wood processors. These three groupings are the aggregated sector classifications that have been used for this analysis.

Employment data (U.S. Census) for these sectors of the Washington forest industry for the years 1998-2004 were averaged to create mean annual (seven-year average) employment figures by sector by county and then aggregated to report by timbershed, state east and west regions, and for the entire state. Total timber harvest volume data (all owner and by county for 1998-2004) were averaged to create mean annual harvest volumes and aggregated to report by timbershed, state region, and for the entire state. Average annual jobs by sector and location were then divided by the average annual harvest volume for that area of the state to generate estimated average industry employment by sector per million board feet of timber harvest per year (jobs/MMBF/YR) for the associated geographical area.

Estimated average industry jobs/MMBF/YR of timber harvest (1998-2004) were introduced as input variables to the Washington Input-Output Model to create estimates of total downstream economic activities/MMBF/YR of timber harvest in 2004 dollars. An iterative routine was employed to segregate direct and indirect forest industry jobs from total forest jobs (the known input data derived from census information). Once direct forest industry jobs were determined, indirect non-forest jobs estimates can be calculated by the model along with labor income, gross state product, and state and local taxes (Table ES-1).

Table ES-1. Employment, Economic Output, and Tax impacts generated by one million board feet of annual timber production in Washington State (employment in person-years, output in millions of 2004 dollars calibrated to the Washington State Input-Output Model NAICS).

Table ES-2, presented below, summarizes the average total annual employment and revenue impacts from timber harvest in Washington from 1998 through 2004. Total jobs and economic benefits for the logging sector have been calculated based upon an average annual harvest of about 3.9 billion board feet. Processing jobs are based upon the total harvest net of log exports which averaged approximately 660 million board feet per year. Paper employment, as a consequence of timber harvest, is considered to be fifty percent of total state paper employment since approximately half of the state paper production is from chip supplies imported from out-of-state suppliers or from recycled material. Total direct and indirect employment was found to be 109,253 jobs with total labor income of $3.7 billion and an average annual wage of $34,000. Forest industry annual wages average $39,000 with a low for the furniture sector of $30,000 and a high for the paper industry of $60,000. The average annual Gross State Product (GSP), which includes labor income, business taxes, and capital income, resulting from timber harvest, was found to be $6.3 billion with State and Local Taxes equal to around $713 million per year. The average annual Gross State Product for Washington from 1998-2004 was $226.1 billion. The forest industry contributed about 3% of GSP during that period.

Table ES-2. Total Jobs and Outputs/Total Harvest MMBF/YR 1998-2004 Ave (employment in person-years, output in millions of 2004 dollars calibrated to the Washington Input-Output Model NAICS).

Table ES-3 shows a summarized distribution taken from this analysis of forest industry jobs by sector as a percentage of state totals for the Westside verses the Eastside of Washington. Comparisons of percentages by timbershed for harvest volumes and forest industry jobs are shown in Table ES-4. The west contributes 79% of the total state timber harvest volume but benefits from 83% of total forest industry employment. Note that five out of the seven timbersheds contribute a larger percentage share of state timber volume than comes back in relative employment benefit. The counties within these timbersheds are generally considered to be the more rural areas of the state. The more urban Puget Sound area, which has historically been the distribution and transportation center for Washington , accounts for only 27% of the timber harvest but benefits from 47% of the jobs. While forest industry employment and revenue streams may leverage higher as a contributor to less diverse rural economies, this analysis shows that almost half of the actual economic benefit occurs in more urban areas.

Table ES-3. Forest Industry Employment and Economic Impact West and East as a Percentage of State Totals by Sector

Table ES-4. Forest Industry Employment by Sector by Timbershed and Harvest Volumes by Timbershed as a Percentage of State Totals

Impacts of treatment alternatives

Estimates of alternative treatment employment impacts were developed from relative harvest cost comparisons provided by the Washington Department of Natural Resources and a survey of Washington Contract Loggers Association membership. For this demonstration, the average regeneration harvest cost per thousand board feet (MBF) of logs removed is assumed as a baseline and given the relative value of 1. Average selection harvest costs were found to be 1.89 times more costly than the baseline regeneration harvests. Average commercial thinning costs were found to be 2.32 times more costly than the baseline regeneration harvests. Long rotation harvests are assumed to have the same cost as the baseline. Assuming that operation cost increases are accompanied by proportionate worker-hour increases, then the estimated cost increases can be used to adjust the economic multipliers. Adjustments were also made to reflect effects on secondary and primary paper processors.

Table ES-5. Adjusted employment impacts for different treatments for Western Washington.

Table ES-6. Adjusted employment impacts for different treatments for Eastern Washington.

Returns from state timber sales to trust beneficiaries

Stumpage revenue generated from DNR timber sales has a uniquely powerful impact on state wealth as it is reinvested within the state; approximately seventy-five percent of stumpage revenues are reinvested for the public good (as returns to trust beneficiaries) in government projects and services in Washington. An estimate of the added benefits provided by DNR stumpage revenues was calculated based upon the assumption that DNR revenues provide public services that otherwise would require tax increases therefore these revenues function as additional state labor income. The average annual timber volume harvested from state forests for the years 1998 through 2004 was 562 million board feet. The average stumpage return was $329/MBF. Total average annual revenue exceeded $185 million with seventy-five percent returned to the trust beneficiaries and the state economy equal to $134 million per year. To model the impact of the addition to the state economy from returns to trust beneficiaries, we added an additional labor income benefit of $246,874 per million board feet of harvested timber to state baseline multipliers and calculated the cumulative impact based upon an average annual DNR timber removal from 1998-2004 of 562 MMBF (Table ES-7,8,&9; employment in person-years, output in millions of 2004 dollars calibrated to the Washington Input-Output Model NAICS).

Table ES-7. Annual economic outputs from DNR timber sales without re-investment of trust revenues.

Table ES-8. Annual economic outputs from DNR timber sales with re-investment of trust revenues.

Table ES- 9. Difference in annual economic outputs from DNR timber sales with re-investment of trust revenues.

The economic multipliers used in the analysis are best current estimates grounded by surveys, census data, and engineering estimates linked to harvest volume inputs and developed with the Washington Input-Output Model.

Table of Contents

List of Figures

List of Tables

Introduction

A Changing Forest Economy

The development of the forest industry in Washington spans a history of rapid transition from naturally endowed and commonly held forest resources to managed forests in a complex mixture of public and private ownerships. One hundred years ago, good timber sold for $4 per thousand board feet (MBF) and $30,000 was enough build a sawmill (Andrews 1957). In 1910, Washington mills produced four billion board feet of lumber; most of which was shipped via rail and water to other states and countries. At that time Washington was the largest lumber-producing state in the United States (U.S.). The largest saw mill in the world was located at Port Blakely. Lumber production provided 61% of all manufacturing jobs and was an important determinant in the region's economic growth. Export of Washington 's rich timber resources induced extensive investment in regional distribution and transportation facilities that in turn quickened the pace of in-migration. New residents found forest products employment to be readily available with wages significantly higher than the national average. The mechanized nature of forest products processing led to growth of repair and maintenance industries that evolved into specialized machinery producers. Prosperity created a growing demand for locally produced goods and services; encouraging investment in domestic infrastructure including education. Transportation centers evolved quickly into cities. The Seattle population was 237,000 in 1910; equivalent to 21% of the state total of 1,142,000 just twenty years after statehood (Tattersol 1960).

One hundred years later, Washington annual lumber production has increased to approximately six billion board feet per year (BBF/YR) making the state the second largest producer in the U.S. behind Oregon . The largest sawmill in the world is now located in British Columbia . As it was in 1910, more than 80% of Washington-produced lumber is sold out-of-state. But, gone now is the expansionist philosophy of property ownership and natural resource exploitation that facilitated the rapid development of the Washington economy in the early twentieth century. In more recent years, concerns about species habitats and old growth protection have resulted in court injunctions and policy shifts that have caused disruptions to timber harvests and processing activities. In 1988, the timber harvest from National Forests in Washington State was 1.5 BBF/YR. In 2002, the federal harvest in Washington was 72 million board feet per year (MMBF/YR); less than 5% of the 1988 level. Timber harvest levels were reduced on Washington Department of Natural Resources (DNR) forestlands from 826 million board feet in 1988 to 457 million in 2002; a 45% decline.

As a result of dramatic reductions in public timber harvests, during the period from 1988 to the present, many sawmills, pulp mills, and veneer mills in Washington , unable to secure sufficient log supplies or to respond to rapidly shifting markets, closed operations. Forest industry employment dropped by more than 10,000 ( Warren 2000, 2005) creating downstream economic impacts that rippled through the state economy. However, the consequences of changes in timber supply were more profound than closed businesses and lost jobs; a total restructuring of the forest products industry occurred which produced some seemingly incongruent outcomes such as increases in annual state lumber production coincident with mill closures (Bergvall and Gedney 1970, Bergvall et al. 1979, Larsen 1992 & 2003, Perez-Garcia 2005). In 2006, forest industries provided 11% of Washington 's manufacturing jobs (Eastin et al. 2007). Seattle population has increased to 583,000; equivalent to only 9% of Washington total population of 6,396,000, but the Seattle metropolitan area, which didn't exist in 1910, has grown to support a population of 3,264,000; equivalent to 51% of the state total (U.S. Census).

Input-Output Analysis

Attempts to give precise form to the effects of expansion and contraction of regional economies, such as those described by the Washington examples above, have been made by scholarly economists through the years. Perhaps the most prominent approach is that of "inter-industry" or "input-output" analysis (Leontief 1936, Isard 1960, Leontief 1966, Yan 1969, Hewings 1985). Harvard University Professor, Wassily Leontief, was awarded the Nobel Prize in Economics in 1973 for his work, beginning in the 1930s and lasting through his career, in pioneering development of input-output economic analysis. This method of economic analysis is based on an understanding that a relatively stable flow of goods and services among elements of the economy can be used to create a detailed statistical picture of an economic system. Put simply, to engage in production, each industry or sector of an economy must purchase inputs . Inputs include raw materials, semi-finished goods and capital equipment bought from other industries (sectors). In addition, labor must be hired and business taxes must be paid. The outputs produced by each sector are then sold to consumers (final users) or to other sectors which then use these intermediate products as inputs. A table summarizing the origin of various sector inputs and outputs that, when totaled, estimate final demand in an economy is called an input-output table.

Leontief took this concept a step further and created an input-output model by converting inter-industry transactions into "direct purchase coefficients". This is accomplished by dividing each inter-industry transaction by the respective industry total input. The result is known as the "Leontif Inverse Matrix" which has been widely used to forecast interactive multi-sector response to economic variables. In the Leontief input-output model, the economic system is assumed to have n industries with two types of demands on each industry: external demand (from outside the system) and internal demand (demand placed on one industry by another in the same system). It is assumed that there is no over-production so that the sum of the internal demands plus the external demand equals the total demand for each industry. Examples of applications include: measuring the economic interdependence of a region's industrial structure, providing a set of precise and sensitive disaggregated (sector-specific) multipliers, estimating inter-industry "ripple effects" and earnings-induced changes in personal consumption, calculating regional impacts from changes in final demand, and many other areas of economic investigation. Special attention is given to understanding the implications of export, import, and internal demands on the economy. Employment, labor income, and personal consumption expenditures can also be estimated.

A State Input-Output Study for Washington was first published in 1967 (1963 data) with subsequent state input-output tables developed in later years for 1967, 1972, 1982, and 1987. The input-output table was constructed by surveying all of the industrial sectors in the state to determine their total sales and purchases. That information is used to create a matrix, or table of numbers, that relates the outputs of all industrial sectors (sales to consumers, to industry, and exports) to the inputs to those industries (all the goods, services, labor, and etc. that the industries must purchase in order to generate the outputs). In a balanced system, the total outputs must equal the total inputs. The latest version, released in 2004, updated the State Input-Output Table to 1997 (Washington Input-Output Analysis 2004). This is the model that was selected to aid development of the analysis in this report. The Washington Input-Output Model and supporting interpretive materials are available from the Washington Office of Financial Management web site, http://www.ofm.wa.gov/economy .

Methods

Selection of the Modeling Framework

This project was undertaken to create updated and detailed economic estimators in order to calculate direct and indirect impacts from changes in forest industry activity for each of the seven timbersheds in Washington State . The Washington Projection and Simulation Model (WPSM): A Regional Interindustry Econometric Model (Conway 1990 and 1994) has been used successfully in prior studies to analyze both direct and indirect downstream economic impacts of changes to the forest industry (Lippke et al 1996, Lippke and Conway 1994). However, this model has not been updated since 1994 as the classification system for statewide economic activity data collection, upon which it was originally based, has since been changed from collection of data by establishment to collection of data by job function. As a result, the WPSM was not considered suitable for use in this project.

The Standard Industrial Classification (SIC) was originally developed in the 1930's to classify establishments by the type of activity in which they are primarily engaged and to promote the comparability of establishment data describing various facets of the U.S. economy. However, over time, new industries evolve and old industries decline. In recent years, rapid changes in both the U.S. and world economies brought the SIC under increasing criticism. In 1992, the U.S. Office of Management and Budget (OMB) joined with the Bureau of Economic Analysis, the Bureau of the Census, and the Bureau of Labor Statistics to undertake development of a new industrial classification system. The result is the North American Industrial Classification System (NAICS). NAICS has a five-digit structure for industry classifications that has been organized in a hierarchal manner to facilitate sector examination at different levels of aggregation. Seventy-nine new industries were recognized (previously not captured by SIC) and another 186 were revised. The updated 1997 Washington Input-Output Model offers the first State NAICS analysis model that includes 39 sector classifications. This model also provides for inflation adjustment of analysis results to 2004 dollars. Tables displaying NAICS code and attributes are included in the Appendix.

The Washington Input-Output Model is a static model representing a snapshot of inter-industry transactional relationships for the Washington State economy while the WPSM is a dynamic econometric model designed to more fully capture ripple effects. Both systems, however, use a similar inter-industry activity matrix developed to characterize the relationship between the forest sector and the rest of the state economy (Bourque et al. 1977, Bourque and Conway 1977). The WPSM was designed to combine input-output tables with export base theory, income-expenditure models, and time projection capabilities (estimated rates of change such as population) to more fully capture future downstream indirect effects of re-investment, government services, and other activities usually resulting in higher impact estimates. For this reason the analytical results developed with the State Input-Output Model tend to understate indirect impacts (as compared with WPSM) and therefore should be regarded as conservative economic estimates (Conway 1994, Lippke and Conway 1994). To aid comparability of results, the economic multipliers developed for this project have been published in similar tabular presentation as those in prior studies.

U.S. Census Employment Data

Basic forest industry employment classifications that have appeared in prior studies (Lippke et al 1996, Lippke and Conway 1994) are Primary Wood, Secondary Wood, and Primary Paper. Primary Wood includes logging and primary processing such as lumber and veneer manufacturing. Secondary Wood includes door, molding, furniture, and other downstream wood product manufacturers that purchase intermediate products from Primary Wood processors. Primary Paper includes pulp and paper manufacturers whose products are generated from raw materials such as pulp logs and chips that are purchased as intermediate products from Primary Wood processors. These three groupings are the aggregated sector classifications that will be used for this analysis.

NAICS is organized around 20 main sector classifications that contain a total of 2328 industries (see Appendix). The U.S. Census Bureau provides NAICS employment data (available online) by year, by sector, by county, and by state for years 1998 through 2005. Data for Washington counties for all sectors of the forest products industry were subset from the total NAICS database and downloaded for review. Data were selected for the years 1998-2004, the baseline years for this investigation. All forest industry classifications were subset further for best fit to the input-output model and to reflect all industries engaged in timber processing as described above and in prior investigations that are likely to be sensitive to fluctuations in timber harvest volumes (see Appendix). For example, secondary paper products manufacture was excluded as was wood construction as these are consumer-driven activities that are largely independent of the origin of supply. Other classifications that might logically be considered as forest industry such as some forestry support services and log trucking were also excluded (in this case by the model) as these industries are picked up as indirect services to the forest sector by the agriculture and transportation sectors. Our focus is on the impact of supply changes within Washington State i.e. we do not want to infer changes in employment activity that are not likely to be dependent upon state timber harvesting activities.

Survey Data

Mail surveys were conducted for the logging industry and the sawmill industry with analysis of response data augmented by interviews with forestry professionals and agency representatives. Pertinent reports from the literature were reviewed. Survey respondents provided valuable data of interest to this investigation that was used to enrich and confirm census information. Examples include average worker age and relative scarcity of skilled employees. Collection of survey data also provided important opportunity to compare, "ground truth", census data analysis against actual reporting. For example, a major disagreement between survey and analysis results with regards to average wages or jobs per unit harvest would likely highlight analytical error. Survey responses combined with historical data from the literature provide valuable insights to aid data interpretation, suggest possible cause and effect relationships, and develop trend analysis. For example, estimates of sawmill overrun (lumber output per log input) gleaned from analysis of survey data and combined with historic records can help us to understand how the number of sawmills could decline while the volume of lumber production increased and employment remained stable. The census data, however, provided the needed detail, not available from survey data, with which to link to the Input-Output Model to develop county-by-county employment information (NAICS) which could then be aggregated to create timbershed and state comparisons.

Harvest, Export, and Other Data

Historical record of timber harvest in Washington by owner type is available from the DNR web site (DNR 2007). Data for Washington annual log export volumes were provided by Debra Warren at the USDA Forest Service PNW Research Station. Estimate of the percentage of chip supply to the Washington pulp and paper industry that originates instate verses out-of-state was provided by Tom Wolford, Director of Washington Pulp and Paper Foundation. Historic data on Gross State Product (GSP), Labor Income, Personal Income, and State and Local Taxes were supplied by the Washington Department of Revenue, U.S. Census Department, Washington Office of Financial Management, and the U.S. Bureau of Economic Analysis. Volumes and values of harvests on DNR (state) forestlands were provided by DNR.

Timbersheds

There are seven timbersheds in Washington as displayed in Table 1 and in the Appendix. Analysis has been conducted and results are reported by timbershed, for the Eastside, for the Westside, (there are two timbersheds in the east and there are five timbersheds in the west), and for the total State.

Employment Analysis

Employment data (Census) for sectors of the Washington forest industry (as mentioned above) for the years 1998-2004 were averaged to create mean annual (seven-year average) employment figures by sector by county and then aggregated to report by timbershed, state region, and for the entire state. Employment data were averaged over the seven-year period to smooth variability and error associated with county level data availability and proprietary limitations on local reporting. Total timber harvest volume data (all owner and by county for 1998-2004) were averaged to create mean annual harvest volumes and aggregated to report by timbershed, state region, and for the entire state. Average annual jobs by sector and location were then divided by the average annual harvest volume for that area of the state to generate estimated average industry employment by sector per million board feet of timber harvest per year (jobs/MMBF/YR) for the associated geographical area.

Direct and Indirect Economic Impacts

Employment multipliers are standardized measurements of economic impacts for purposes of comparison. The employment multiplier is the ratio of the total employment impact to the direct employment impact. In this report, we estimate employment multipliers by individual sector as well as for the aggregate forest products industry. Employment multipliers per input variable of interest are reported as jobs/MMBF/YR of harvested timber volume. Average employment estimates captured by surveys of the forest sector reflect both inter-industry and intra-industry transactions. Intra-industry transactions occur when some parts of the forest sector outsource to other parts of the sector resulting in within-sector transactions. For example, one logging company may purchase standing timber but contract with another company to undertake the harvest. Indirect forest industry employment figures are estimated as a second of jobs that are created when one company in an industrial sector purchases services and/or products from another company in the same sector. The sums of direct and indirect forest industry activities are the forest sector totals as influenced by instate harvest activities. This approach allows a direct comparison to the sector by sector jobs picked up in our survey and in past employment publications. For our purposes, both the direct employment and other within-sector employment sum to comprise a Total Forest Industry Employment estimated as job-years/MMBF/YR for logging and other forest industry sectors. In addition to forest industry economic activity, many other jobs and economic impacts are created as a result of downstream economic activities generated by timber harvest. Non-forestry indirect economic activities include all jobs, revenues, and taxes created by the economic activities of the forest products industry in response to external and internal demands linked to annual timber harvest activities. In addition to jobs, other measures of economic activity, important to Washington State , are reported including Labor Income, Gross State Product, and State and Local Taxes. Indirect economic impacts are reported as millions of dollars/MMBF/YR of timber harvest and jobs/MMBF/YR of timber harvest.

Results - Jobs-to-harvest ratios

Employment data and the Washington Input-Output Model

Total annual forest industry employment, from Census data, is shown below in Figure 1 with total timber harvest volume for Washington by year for 1998-2004. Figure 2 shows annual harvest volumes presented with the ratios of forest industry jobs by year to each million board feet of timber harvested. The average ratios for forest industry jobs per MMBF/YR (1998-2004) are presented below in Tables 2A & 2B for each timbershed, the West, the East, and the whole State. Average annual total state timber harvest for 1998-2004 equaled around 3.9 billion board feet per year (BBF/YR). For each MMBF/YR of timber harvested an average of 11.28 forest industry jobs were generated.

Note that the distribution of forest industry jobs is dominated by Puget Sound manufacturing with comparatively little secondary economic benefit shared with more rural timbersheds. As a result, in Figures 3 and 4, it can be seen that the North and South Puget Sound timbersheds have greater employment benefits per unit of harvested timber than other Washington areas.

Figure 1. Total Washington timber harvest and total forest industry jobs/year 1998-2004.

Figure 2. Total Washington timber harvest and forest industry jobs/MMBF/year 1998-2004.

Figure 3. Average forest industry Jobs/MMBF/YR timber harvest by timbershed 1998-2004.

Figure 4. Average forest industry Jobs/MMBF/YR timber harvest by sector and by timbershed (1998-2004).

Estimated average industry jobs/MMBF/YR of timber harvest (1998-2004) were introduced as input variables to the Washington Input-Output Model to create estimates of total downstream economic activities/MMBF/YR of timber harvest in 2004 dollars. An iterative routine was employed to segregate direct and indirect forest industry jobs from total forest jobs (the known input data derived from census information). Once direct forest industry jobs were determined, indirect non-forest jobs estimates can be calculated by the model along with labor income, gross state product, and state and local taxes.

Economic impacts as previously estimated

The basic sector outputs from the model used for economic impact assessments in the prior studies are provided in Table 3. The model provides direct and total employment, direct output, gross product and state & local taxes for logging and sawnwood, as two components of primary wood; secondary wood, primary paper, and a total forest sector impact. Employment and economic outputs are presented as estimates of job-years and millions of 1992 dollars for one million board feet of annual timber harvest in Washington State (MMBF/YR).

Table 3. Employment, Economic Output, and Tax impacts generated by one million board feet of annual timber production in Washington State (employment in-person years, output in millions of 1992 dollars adapted from WPSM 1994).

Update employment impacts

For more current estimates of economic activity, we update the employment data to 2004 based upon current employment averages used with the state input-output model as described above. Results are shown in Table 4 below. Prior and current results are presented for comparison; however, the WSPM data (1992) originated with SIC employment classifications while the updated data (1998-2004) is from NAICS. Data source differences create some sector inconsistencies for direct comparisons. It should also be noted that, in the prior study, indirect forest industry employment was reported as part of total employment, while, in the updated analysis, indirect forest sector employment has been isolated for addition to direct forest jobs to generate an estimate of total forest industry employment comparable to employment census publications.

Logging sector direct employment shows slightly higher labor intensity in 2004 than in 1992 which is consistent with harvest of more difficult locations and smaller diameter timber. The NAICS logging classification also captures some forestry activities such as timber tract operations, reforestation, and gathering of non-timber forest products. Sawnwood, however, shows a decline in employment per unit input indicating productivity gains by fewer but much larger and efficient sawmills than in prior studies. Secondary wood processing is split out into two categories, other manufacturing (such as molding, millwork, doors, and windows) and wooden furniture manufacture, picking up some value-add manufacturing activity either not present in 1992 or not included in earlier models. Primary paper direct employment per unit input is very close to the prior assessment indicating that, while gross levels of activity in this sector may have declined and per-unit-outputs may have increased, the per-unit-input productivity has remained relatively constant. The 2004 table is expanded to show the impact detail of the subset of indirect jobs within the forest sector i.e. some parts of the forest sector outsource to other parts of the sector and appear as within-sector transactions as mentioned above. This approach allows a direct comparison to the sector by sector jobs picked up in our surveys, which include both the direct employment and other within-sector employment comprising a Total Forest Industry Employment estimated as job-years/MMBF/YR for logging and other forest industry sectors. This presentation of direct and indirect forest sector detail is intended to resolve confusion that, in the past, arose when direct job estimates (from input-output analysis) were compared against data from business-reported employment roles and were found not in agreement.

The Total Washington Direct Forest Industry employment appears higher than in prior studies with some gains explained by increased productivity as evidenced by record lumber production. Increased jobs/MMBF/YR of timber harvest can also be explained by inclusion of two secondary manufacturing sectors as mentioned above and detailed in the Appendix. The differences in sector results are also, in part, a relic of data inconsistencies between SIC and NAICS classifications, however, differences in sector classifications should largely disappear in summary comparisons of total industry employment. Labor Income is computed from employment using average labor rates within the Input-Output Model (I-O) adjusted to reflect 2004 dollars. Total Employment is estimated in the model including inter-industry transactions and indirect purchases and services for a direct plus indirect job estimate in support of the Forest Sector economic activity (multiplier). Gross State Product (GSP) and State and Local Taxes are computed based upon ratios to Labor Income for 2004. State & Local taxes are estimated directly from the statewide economic activity and do not reflect extra taxes leveled on the industry such as the timber excise tax or additional revenue streams to counties such as Payments in Lieu of Taxes (PILTS) from the federal government. Comparison of summary results from 1992 and 2004 shows increased forest industry employment per unit input attributable to reasons suggested above. Decreases in the total direct and indirect employment multiplier result in part from the previously-discussed differences between the dynamic (WPSM) verses static (I-O) models. However, the decline in total employment levels is more directly related to the declining harvest with its diminished influence on the total state economy.

Table 4. Employment, Economic Output, and Tax impacts generated by one million board feet of annual timber production in Washington State (employment in person-years, output in millions of 2004 dollars calibrated to the Washington State Input-Output Model NAICS).

It is important to note that not all activities generally perceived to be within the Forest Sector are actually defined as part of the Total Forest Industry in the economic models. For example, trucking is considered by the employment model as an indirect service provided by the Transportation Sector while some forestry support services, such as forestry consulting, firefighting, and pest control, are classified as Agriculture Sector activities. The employment benefits of these sectors are not unrecognized however, as many of these activities are captured as indirect employment in the Total Employment estimates generated by the Forest Sector activities.

Employment data adjustments were made to reflect estimated corrections for influences of log exports and chip imports on processed volumes but not foreign purchased port activities. For the five Westside timbersheds, while logging jobs were considered as necessarily associated with all harvested log volumes, processing sector estimates of jobs/MMBF/YR (Sawnwood, Secondary Manufacturing, and Paper Processing) were adjusted to reflect reductions, as created by log exports, to total harvest volume available for instate process. While the trucking jobs connected to log exports are likely captured by logging indirects, previous studies suggest that the magnitude of other export preparation activities, such as Stevedore jobs that are purchased by foreign consumers, are not captured as indirect activity in the Input-Output Model. It is assumed that there are no exported logs from the two Eastside timbersheds.

Employment from pulp and paper manufacture, as a consequence of state timber harvest volumes, is considered to be only fifty percent of total state paper employment since approximately one half of the state pulp and paper production is made from chip supplies imported from out-of-state suppliers or from recycled materials (Wolford pers. com.).

While some logs harvested in Washington are likely to be processed in neighboring states, the reverse is also true. Some volumes of logs from other states enter Washington for processing. The net impact from interstate log traffic is believed to be small and beyond the scope of this study. It is assumed, therefore, that for our economic estimates, log traffic for processing into- and out-of-state (excluding international log exports) has a neutral effect on employment. The distributional impacts associated with primary and intermediate product shipments between timbersheds, however, are captured by the actual census employment data by sector.

Total state economic impacts from annual timber harvest

Table 5, presented below, summarizes the average annual employment and revenue impacts from timber harvest in Washington from 1998 through 2004. Economic results are presented in millions of 2004 dollars. Total jobs and economic benefits for the logging sector have been calculated based upon an average annual harvest of 3.9 billion board feet. Processing jobs are based upon the total harvest net of log exports which averaged 660 million board feet per year. While the trucking jobs connected to log exports are likely captured by logging indirects, international export preparation acivities such as Stevedore jobs are not. Paper employment, as a consequence of timber harvest, is considered to be fifty percent of total state paper employment since approximately half of the state paper production is from chip supplies imported from out-of-state suppliers or from recycled material. Total direct and indirect employment was found to be 109,253 jobs with total labor income of $3.7 billion and an average annual wage of $34,000. Forest industry annual wages average $39,000 with a low for the furniture sector of $30,000 and a high for the paper industry of $60,000 (wage detail is provided in the Appendix). The average annual Gross State Product (GSP), which includes labor income, business taxes, and capital income, resulting from timber harvest was found to be $6.3 billion with State and Local Taxes associated with this activity level equal to around $713 million per year. The average annual Gross State Product for Washington from 1998-2004 was $226.1 billion. The activities introduced by the timber harvest contributed about 3% of GSP during that period.

Table 5. Total Jobs and Outputs/Ave Total Harvest MMBF/YR 1998-2004 (see also Table 9)

Close examination of the total jobs and outputs/average total harvest with adjustments to manufacturing sectors to reflect the impacts of log exports (Table 5) shows a decrease in multipliers/MMBF of total harvest volume for the manufacturing sectors as compared to Table 4. These changes (presented in Table 6 below) are specific for the average total harvest verses exported timber volumes for the years 1998-2004. The authors recommend that use of the updated multipliers developed in this report should begin with figures as shown in Table 4. Totals are then calculated based upon the harvested and export volumes of interest as demonstrated in Table 5. Depending upon different total harvest to exported volume ratios, the multipliers for manufacturing sectors will self-adjust.

Table 6. Adjusted manufacturing employment multipliers as a result of 1998-2004 average log exports.

Table 7 shows a summarized distribution taken from this analysis of forest industry jobs by sector as a percentage of state totals for the Westside verses the Eastside of Washington. Comparisons of percentages by timbershed for harvest volumes and forest industry jobs are shown in Table 8. The west contributes 79% of the total state timber harvest volume but benefits from 83% of total forest industry employment. Note that five out of the seven timbersheds contribute a larger percentage share of state timber volume than comes back in relative employment benefit. The counties within these timbersheds are generally considered to be the more rural areas of the state. The more urban Puget Sound area, which has historically been the distribution and transportation center for Washington , accounts for only 27% of the timber harvest but benefits from 47% of the jobs. While forest industry employment and revenue streams may leverage higher as a contributor to less diverse rural economies, this analysis shows that almost half of the actual economic benefit occurs in more urban areas.

Table 7. Forest Industry Employment and Economic Impact West and East as a Percentage of State Totals by Sector

Table 8. Forest Industry Employment by Sector by Timbershed and Harvest Volumes by Timbershed as a Percentage of State Totals

Detailed tables for the state, state regions, and by timbershed are given below. These analyses provide updated multipliers for employment, labor income, GSP, and state and taxes derived from one million board feet of timber harvest per year. Multipliers have been developed for use in analysis of timbersheds, state regions, and total state activities linked to timber harvest. An important application for forest industry multipliers is in investigation of comparative harvest scenario outcomes such that the marginal benefits of different business and policy choices can be better understood. The state, the regions and the timbersheds each have unique multipliers that are dependent upon sector employment and harvest. Table 9 shows state multipliers from which analysis in this report, unless otherwise specified as by region or timbershed, is based.

Table 9. Total State Jobs and Outputs/Ave MMBF/YR and Total Jobs and Outputs/Ave Total Harvest MMBF/YR 1998-2004.

Table 10. Total Jobs and Outputs/Ave MMBF/YR and Total Jobs and Outputs/Ave Total Harvest MMBF/YR 1998-2004 for Western Washington and Eastern Washington.

Table 11. Total Jobs and Outputs/Ave MMBF/YR and Total Jobs and Outputs/Ave Total Harvest MMBF/YR 1998-2004 for West Timbersheds: North Coast and South Coast.

Table 12. Total Jobs and Outputs/Ave MMBF/YR and Total Jobs and Outputs/Ave Total Harvest MMBF/YR 1998-2004 for West Timbersheds: North Puget Sound and South Puget Sound.

Table 13. Total Jobs and Outputs/Ave MMBF/YR and Total Jobs and Outputs/Ave Total Harvest MMBF/YR 1998-2004 for West Timbershed : Southwest.

Table 14. Total Jobs and Outputs/Ave MMBF/YR and Total Jobs and Outputs/Ave Total Harvest MMBF/YR 1998-2004 for East Timbersheds: East Cascades and Inland Empire.

Results - Impacts of Treatment Differences

While Table 4 presents updated estimates for current levels of timber industry productivity and state impact, further refinements can be developed to reflect the economic impacts of different management treatment alternatives that in turn can be linked to a timber and forest structure supply model. As noted in Lippke et al (1996) a first thinning treatment is much more labor intensive than a final regeneration harvest as a consequence of lower volume and smaller pieces, while a second and third treatment are just slightly more labor intensive then regeneration harvests. Similarly, thinnings with much smaller diameter trees are weighted more heavily to chips for pulp and paper and less heavily to secondary wood uses whereas longer rotations produce higher quality and increases in secondary manufacture.

Estimates of alternative treatment employment impacts were developed from relative harvest cost comparisons provided by the Washington Department of Natural Resources and a survey of Washington Contract Loggers Association membership. For this demonstration, the average regeneration harvest cost per thousand board feet (MBF) of logs removed (Regen - 50/50 as an average with WCLA and DNR reporting) is assumed as a baseline and given the relative value of 1 (Table 15). Average selection harvest costs (Select - 50/50 with only WCLA reporting) were found to be 1.89 times more costly than the baseline regeneration harvests. Average commercial thinning costs (Thin - 50/50 as an average of with WCLA and DNR reporting) were found to be 2.32 times more costly than the baseline regeneration harvests. Long rotation harvests are assumed to have the same cost as the baseline (Regen - 50/50). Assuming that operational cost increases are accompanied by proportionate worker-hour increases, then the estimated cost increases can be used to adjust the economic multipliers in Table 4 to estimate comparative impacts of forest management alternatives on employment, revenues, and taxes.

Table 15. Washington comparative harvest costs by treatment type with cost multipliers developed with regeneration harvest cost as the baseline.

Impact of Treatment Differences

Different treatments produce different log outputs that serve different end uses. By noting how these treatment differences affect the sector level data in the model, a first-order estimate of the impact of different treatments can be obtained. Listed below are sample treatment alternatives typical of management options that vary by landowner type and objectives. For example, large private forestland owners, obligated to investment responsibilities, maximize economic returns through short commercial rotations whereas public forestland managers may favor longer rotations and thinning harvests to integrate financial return with habitat enhancements. Table 16 summarizes the sector impact responses for different westside treatments, as described below, relative to the average base case and Table 17 scales the treatment response to our normalized treatment size of one million bd. ft. harvest. Table 18 and 19 provide similar response tables for the eastside.

Treatment Alternatives

Base (Regen - 50/50): Regeneration harvest at 50 years after full plantation stocking followed by a pre-commercial thin (PCT) for density and quality control (the dominant historic treatment for Western Washington ).

First Thin (Thin - 50/50): Commercial thin at around age 30, roughly half by ground processor equipment (low slopes) and half by skyline (steep slopes) with costs almost twice the baseline.

Second Thin (Select - 50/50): Commercial thin at about age 50 with fewer small logs flowing to chips and more large logs for increased secondary manufacture.

Third Thin (Select 50/50): Commercial thin at about age 70 with more uniform and larger logs.

Long Rotation Harvest (Regen - 50/50): Regeneration final harvest of large mostly uniform trees producing more quality wood and secondary manufacture but less chips.

Sector Impacts

Logging (Prim W.): More costly thinnings with more jobs based on lower volume operations and handling of smaller diameter material.

Sawnwood (Prim W.): Sawnwood plus plywood with more quality volume from longer rotations.

Other MFG and Furniture (Sec W.): Half sensitive to higher quality wood (doors, windows, cut stock, mill work) but including pallets (low quality), trusses (above average structural quality), and furniture. Excludes building construction.

Primary Paper: Paper mills receive chip volumes primarily from mill processing residuals and smaller diameter logs.

Table 16. Key employment impact changes to characterize treatment differences based on comparative processing costs and changes to manufacturing activities relative to the base for Western Washington.

Table 17. Adjusted employment impacts for different treatments for Western Washington .

Table 18. Key employment impact changes to characterize treatment differences based on comparative processing costs and changes to manufacturing activities relative to the base for Eastern Washington .

Table 19. Adjusted employment impacts for different treatments for Eastern Washington.

Employment and Other Economic Outputs per unit of volume for different treatments

We have chosen not to include future productivity gains in the labor projections since it is easier to evaluate impacts on a current productivity or constant labor productivity basis. We prefer to display the impact of different treatments directly for demonstration rather than to confound the comparative treatment outcomes with unreliable estimates of productivity in the future. Labor income is generally the more stable measure of employment over a projection period as productivity gains are generally passed on to wage gains so the labor wages are not likely to be compromised by productivity changes over time. Comparative impact tables are provided below by treatment type for the Westside and Eastside (Tables 20 & 21).

Table 20. Total employment, economic output, and tax impacts for treatment alternatives as generated by one million board feet of annual timber production for Western Washington (employment in person-years, output in millions of 2004 dollars calibrated to the Washington State Input-Output Model NAICS).

Table 21. Total Employment, Economic Output, and Tax impacts for treatment alternatives as generated by one million board feet of annual timber production for Eastern Washington (employment in person-years, output in millions of 2004 dollars calibrated to the Washington State Input-Output Model NAICS).

Economic impact estimators are provided above to characterize the comparative effects of different management treatments for both the Westside and Eastside that reflect regional infrastructure differences in harvesting and processing. In Western Washington , current industrial forestland management favors short rotation regeneration harvests for maximum economic return (base case) which provide 12.59 direct jobs/MMBF/YR while other private and public forestland owners are likely to include more thinning treatments that increase jobs benefits. Determining a base case scenario for the Eastside is somewhat more complicated as multiple cohort and repeated thinning harvests are more common than the single cohort management regimes of the Westside (Oliver et al. 1994). However, the significance of the lower job multipliers on the Eastside should not go unnoticed as compared to the Westside. Lower multipliers combined with much smaller harvest volumes on the Eastside result in Westside timber activities dominating the State Forest Sector contribution to the state economy. However, emerging policy recognition of the need to reduce stocking densities within Inland West forests for dual benefits of forest fire hazard mitigation and utilization of biomass for renewable energy may generate future increases in forest management activities and employment for Eastern Washington (DNR 2004a, 2004b, The White House 2003).

While the long rotation treatment alternative increases the jobs per unit of removals by as much as 45% it should be noted that jobs benefits occur very far in the future and may be subject to unforeseeable changes including productivity gains and policy changes prior to the activity. Thinning treatments, such as the first thin, increase the job impact a modest 10% since added logging labor is largely offset by secondary (value-add manufacturing) labor declines, however, unlike the long rotation, the impact is near-term and can increase economic activity early in the rotation period. Since different management treatments are practiced in different timbersheds with different infrastructure and employment bases, we can logically expect somewhat different job impacts in each local region. Tables presented above for West verses East have been developed to demonstrate such differences. Future research to investigate marginal trade-offs in forest landowner net present value returns from alternative management approaches as compared to potential increases in state employment and tax revenues could provide useful cost/benefit analysis to inform forest policy in regards to the public benefits that could arise from management incentive programs coupled with markets for environmental services such as carbon sequestration or habitat.