Timber trade in the United States of America 1870 to 2017. A socio-metabolic analysis

Abstract

The importance of international timber trade in The United States forest transition is poorly understood. Here, we synthesize a variety of historical sources to establish a consistent socio-metabolic dataset for U.S. physical trade in timber and timber products from 1870 to 2017, distinguishing three product categories (raw, primary and final products) and major trading partners. The United States increasingly relied on net imports of primary and final timber products, mainly from Canada, but also from Asia, and Europe, while emerging as a net-exporter of raw timber. The growing physical import dependence coincided with domestic forest recovery, indicating that imports contributed to reducing pressures on domestic forests. The structure of timber trade suggests that the United States also outsourced labor for wood processing.

Keywords:

• Timber trade
• timber products
• forest transition
• United States
• ecologically unequal exchange

Introduction

The United States have emerged as one of the key players in global biomass trade during the past 200 years, with wood products being at times of particular importance (Krausmann and Langthaler 2019). Biomass trade has important implications for ecosystems, land and resource use in both exporting and importing countries (Kastner et al. 2021). In particular, trade in biomass can affect forest dynamics due to a spatial displacement of deforestation (Pendrill et al. 2019). In addition, imports of wood may allow for a forest transition in the importing country while increasing deforestation in the exporting country, thus outsourcing pressures on forests from one country to another (Gingrich et al. 2019; Meyfroidt and Lambin 2011).

In the United States, forests started recovering in the early 20th century after centuries of deforestation (MacCleery 1993; Magerl et al. 2019), a phenomenon described as forest transition (Mather 1992). The forest transition in the United States has been linked to domestic intensification of agriculture, as well as to changes in energy use (Magerl et al. 2022). However, the effect that wood trade had on domestic forest recovery in the United States has, to our knowledge, never been investigated empirically.

Our study addresses this research frontier and inquires to what extent trade in forestry products (“timber trade”) played a role in forest dynamics in the United States from 1870 to 2017. We use a socio-metabolic approach to quantify indicators from the framework of economy-wide material flow accounting, including physical imports, physical exports, and physical trade balance, discerning three types of timber products according to their degree of processing, and identify their main countries of origin and destination. Linking these data to previously-published data on dynamics in forest area, domestic wood harvest, and forest biomass stocks (Magerl et al. 2019, 2022) enables us to explore to which extent wood extraction processes abroad have contributed to the U.S. forest transition.

Previous work showed that the United States was a net exporter of biomass until WWI (Manthy and Potter 1978; Potter and Christy 1962; Vanek 1959). Thereafter, the United States turned into a net importer of biomass, driven by an “insatiable appetite” for tropical wood imports as well as crisis events such as the Dust Bowl years (Tucker 2000). However, to our knowledge, none of these studies have investigated the long-term links between wood trade and recovering forests in the United States, or its implications for ecologically unequal exchange. Research on the history of timber trade has been conducted for other countries, usually covering shorter time periods (Clifford and Castonguay 2022; Daheur 2022; Gingrich 2011; Iriarte-Goñi 2013), as well as globally for recent decades (Kastner, Erb, and Nonhebel 2011). Noting the lack of studies on the impact of biomass trade prior to 1950, Brolin and Kander (2022) suggest that studies of a particular country should be extended over a long period of time. On trade’s crossing of territorial boundaries, the British case has been much debated (Pomeranz 2021; Theodoridis, Warde, and Kander 2018), and some of its features, such as overcoming land constraints through horizontal expansion from imports, can serve as inspiration for this study. The nature of the traded goods has important implications for the use of the associated land to produce them, whether for direct exports or for indirect needs, such as the use of water. Since land use is a pervasive driver of climate change (IPCC 2021; Meyfroidt and Lambin 2011), international trade and associated impacts on land and resource use are of particular importance and have gained increased attention in recent decades (Kastner et al. 2021). We conceptualize the externalization of environmental burdens exerted by wood harvest through trade as “ecologically unequal exchange” (Eisenmenger and Giljum 2007; Hickel et al. 2022; Meyfroidt et al. 2022). Research has shown that globally, higher-income countries tend to import natural resources from lower-income countries, thus externalizing environmental pressures including land- and resource use abroad, as well as outsourcing labor (Dorninger et al. 2021). To our knowledge, trade in U.S. timber products in the context of ecologically unequal exchange has not been thoroughly investigated yet over a long-term basis. In this study, we collected data on the main timber trade partners of the United States, investigating to what extent the country imported which timber products from which countries.

A material environmental history approach, as defined by McNeill (2003, p. 6) “concerns itself with changes in biological and physical environments, and how those changes affect human societies. It stresses the economic and technological sides of human affairs”. We demonstrate in this study how social metabolism approaches can be integrated into historians’ tools for analyzing the macro-social processes of 19th and 20th century history. The work at hand studies trade in timber products in the United States from 1870 to 2017, adopting a socio-metabolic perspective. We aim at advancing the field of quantitative history by proposing an innovative approach, combining methods from social ecology and material environmental history. In particular, we present a comprehensive dataset for all timber products traded by the United States, relying on the consistent compilation of data derived from diverse historical statistics, and applying the methodological framework of economy-wide material flow accounting (Fischer-Kowalski et al. 2011). The innovations of this database are (1) its completeness over a very long period of almost a century and a half, and (2) its precision through a categorization between different types of timber products. The conversion of all product categories into uniform physical units i.e., metric tons, allows for a consistent analysis over time. As the products traded are diverse in nature, a typology according to their degree of processing makes it possible to specify the nature of the trade that the United States has established, and the temporal trajectories of the products traded. The collection presented here constitutes a historiographical innovation since research to date has, to our knowledge, not been sufficiently comprehensive to establish links between timber trade and the forest transition in the United States. Our work thus allows us to develop a new narrative on the timber trade of the United States.

This paper proceeds as follows: The next section presents the conceptual approach and describes the methodological work of collecting, compiling, and processing the data used to develop the novel dataset on long-term U.S. timber trade. The third section presents the results of the established dataset. Finally, the last section discusses the results in the context of the political-economic history, as well as their implications for the U.S. forest transition and for ecologically unequal exchange.

Methods and data

Concepts

The socio-metabolic indicators to investigate trade dynamics in timber products used here stem from the methodological framework of economy-wide material flow analysis, or accounting (“MFA”) (Krausmann et al., 2008; Haberl et al. 2004). It allows for a consistent compilation of annual material throughput of a national economy, including trade of materials between countries. It establishes categories of resource flows of different materials by converting all data into biophysical units (i.e., metric tons) and consistently classifying them into material categories aggregated into biomass, fossils, metallic and nonmetallic minerals. MFA has been successfully used to describe long-term dynamics in national resource use, based on historical and contemporary statistical records (Krausmann, Gingrich, and Nourbakhch-Sabet 2011).

MFA approaches have also been applied to study the historical dynamics of international trade in resources, particularly in agricultural and biomass products (Krausmann and Langthaler 2019). Quantifying physical trade flows allows to describe the “teleconnections”, the social, ecological and economic impacts activities of consumers in one place have on producer in another place, that can manifest over large distances (Friis et al. 2016; Schaffartzik et al. 2015). Relevant indicators from MFA include physical exports (in metric tons per year), physical imports (in metric tons per year) and the physical trade balance (PTB), i.e., imports minus exports. The concept of PTB makes it possible to question the dependence of a national economy on the consumption of products from foreign lands (Gingrich 2011; Pendrill et al. 2019). All trade indicators derived from economy-wide MFA refer to the traded material crossing a national boundary within one year, neglecting e.g., the amount of raw material processed upstream to generate this trade flow, or the environmental impact caused by the respective material extraction. A multitude of approaches coexist that quantify the ecological impacts of biomass trade in different ways for more recent time periods, based on a variety of data sources and modeling assumptions (Bhan et al. 2021; Schaffartzik et al. 2015), which were beyond the scope of this study. Despite its limitations, the physical trade balance has many advantages for environmental history: it is a simple indicator based on international trade statistics which are more or less readily available for longer time-periods, it enables determining the dependence of a given territory on other geographical spaces, and thus delineates the importance of extractivism, or on the contrary, of ecological imperialism.

Data sources

We compiled a consistent database on physical imports and exports of timber products of the United States for the period 1870–2017, based on published data covering overlapping timespans (1870–1905, 1900–1970, 1965–2017). We chose to use a limited amount of data sources to allow for investigating this long time series while minimizing breaks and inconsistencies (for a discussion of data processing, including commensurability and correction procedures performed, see below). We develop time series of the quantity of timber trade flows in metric tons, as well as their structure in terms of primary vs. processed products. In addition, we investigate the main trade partners of the United States over time. Our analysis focuses on timber trade, i.e., industrial wood including all products commonly cut from round sections of trees, excluding firewood. According to Gierlinger and Krausmann (2012), firewood quantities traded were only of minor importance. We therefore consider them as negligible, along with other specific forest products such as Christmas trees, Christmas greens, marine stores, and other non-timber forest products (e.g., raisin, turpentine).

The data used for the period of the 1870s to 1905 stem from a book by forestry economist James E. Defebaugh (1907), who was associated with the U.S. Forest Service. It provides data on imports and exports of timber, shingles, sawn lumber, board planks, joists and scantlings, which are based on tables of exports and imports from statistics prepared by the U.S. Treasury Department and the Department of Commerce and Labor. The studied period corresponds to a centralization of statistical production around the Bureau of Statistics attached to the Department of Commerce and Labor. This new institution guaranteed a better quality in the estimates of imports and exports, in particular with regard to Canada, which had been underestimated until then. Nevertheless, Defebaugh (1907) stated that even the establishment of the new statistical organization couldn’t hide the fact that many of the collected data were “faulty” and “largely understated” (ibid., p. 535). We therefore point out that data for the first period are, at best, an approximation.

For the period 1900–1970, we use data compiled by the National Bureau of the Census of the US (Historical Statistics of the United States, Colonial times to 1970), including statistics on forest products traded, e.g., lumber, logs, plywood, and newsprint (U.S. Bureau of the Census 1975). These data were originally obtained from several reports on wood products and forest statistics published by a number of different entities (U.S. Forest Service, Hair and Ulrich 1967 The Demand and Price Situation for Forest Products; Treasury Department 1903 Foreign Commerce and Navigation of the U.S.; Department of Commerce 1966 Imports of Merchandises for Consumption). Since the Bureau of the Census reports, to our knowledge, the longest consistent timeline for U.S. trade of forestry products, we chose it as our main data source for this period. However, some of the data were aggregated and some data gaps exist (see also next chapter). We supplement missing data with figures from the Statistical Abstract series (U.S. Department of Commerce and Labor 1908, 1921, 1930), since this series is consistent with the Historical Statistics of the United States, Colonial times to 1970 and the Foreign Commerce and Navigation series, reporting the same or very similar numbers.

Finally, for the period 1965–2017, we use comprehensive data published by the U.S. Forest Service Statistics Section (U.S. timber production, trade, consumption, and price statistics 1965 to 2017) covering international trade of lumber, plywood and veneer, pulp-wood based products, logs, and pulpwood chips (Howard and Liang 2019).

Data on trade partners were obtained from Defebaugh (1907) for the first, and from Howard and Liang (2019) for the last period. For the second period, however, no consistent data on countries of origin and destination of timber products were available.

The dataset presented and analyzed here aims at displaying the potentials for a social metabolism approach in investigating long-term trade dynamics from a material perspective. While necessarily remaining coarse in terms of topical, material and geographical resolution, the dataset serves as an entry point for further research on United States’ wood use, trade, or processing by helping to situate major periods, products, or trading partners within a long-term and macro-economic context.

Data processing

Harmonizing product categories

Since data on timber products are reported in different units of volume or mass, we used a range of conversion factors to translate them to metric tons. Table 1 shows all the compiled product categories, associated with their sources, and the factors used to convert them. See Appendix A for more information and original sources for these factors.

Table 1. Processing steps and factors to convert different reported units of the products traded into metric tons and the sources reporting the original figures for the United States between 1870 and 2017.

Sources of data and definitions	Units used in the original sources	Conversion factors to metric tons (see Appendix A for sources)	Product definitions and conversion steps
Defebaugh (1907)	Mahogany in board foot	*0.00348*0.576	Timber of tropical hardwood species. We first converted board foot to cubic meters (factor 0.00348), after which we applied a standard volume-to-weight conversion factor of 0.576 for hardwoods.
	Board and other sawn lumber in board foot	*0.00348*0.491	Board is timber that has been cut into varying rectangular widths and lengths: often, the timber is treated (with lamination), before the sale. Timber boards can be used in a wide range of construction projects. Sawn lumber: logs that are cut lengthwise to make long pieces and crosswise to length depending on standard or customs. We first board foot to cubic meters (factor 0.00348), after which we applied an average standard volume-to-weight conversion factor of 0.491.
	Shingles (units)	*1000*0.000973*0.491	Wood shingles are thin, tapered pieces of wood, used to cover roofs and walls of buildings to protect them from the weather. Historically, their dimensions are 10 to 19 mm height, 76 to 203 mm width and 360 to 919 mm length. In the United States the most common shake – a basic wooden shingle made from split logs, is about 460 mm long. Sources: U.S. Department of the Interior, National Park Service (Wood shingle). This information helped us to get a sense of the volume of a shingle, using average lengths: 1 shingle = 15*141*460 = 972,900 mm^3 = 0.000973 m^3 Next, we converted from volume to weight by applying a standard volume-to-weight conversion factor of 0.491
	Logs in thousands of board feet	*1000*0.00348*0.405 / 0.576	A log is a part of the trunk or a large branch of a tree that has fallen or been cut off. Logging – the process of cutting, processing and moving to a location for transport-, is the beginning of a supply chain that provided raw material for many sectors (housing, construction, energy, paper). We first multiplied by 1000 then by 0.00348 to get cubic meters, after which a standard volume-to-weight factors was applied, in the case of softwoods, 0.405, for hardwood 0.576.
	Timber hewn and sawn in cubic feet	*0.02832*0.491	Timber hewn is a log that has been converted from its rounded natural form into lumber with more or less flat surfaces with an ax. It is an ancient method, and is now used only occasionally. Sawn timber is timber that is cut from logs into different shapes and sizes. Timber that has been sawn to size but has not been finished, or is rough. First, we converted from cubic feet to cubic meters by multiplying by 0.02832, then to metric tons with an average standard density conversion factor of 0.491.
	Board planks, joists, scantlings and staves in thousand feet	*1000*0.00348*0.491	A timber board is an intermediate product that has been cut into varying rectangular widths and lengths. It is often treated before sale. A board plank is a timber that is flat elongated and rectangular, with parallel surfaces that are higher and longer than wide. Used primarily in carpentry, but also in ships, house and bridges. A joist, also used in construction, is a structural member that spans horizontally between the foundation of a building, or between walls. It is a sort of beam (that is a structural element that resists loads applied laterally to their axis), usually made of solid timber. Data were converted from thousand feet, to feet (*1000), then to cubic meters (0.00348), and then to metric tons (0.491).
U.S. Bureau of the Census (1975)	Lumber total in trillion board feet	*1,000,000,000*0.00348* 0.491	We first multiplied by 1,000,000,000, then by 0.00348 to convert from board feet to cubic meters, then from volume to weight (0.491).
	Lumber (softwoods and hardwoods) in trillion board feet	*1,000,000,000*0.00348* 0.405 (softwoods); *1,000,000,000*0.00348* 0.576 (hardwoods)	Same procedure as for total lumber, except for using a different standard volume-to-weight conversion factors for softwoods (0.405) and hardwoods (0.576)
	Logs (softwoods and hardwoods) in million board feet	*1,000,000*0.00348*0.405 (softwood); *1,000,000*0.00348*0.576 (hardwood);	We first multiplied by 1,000,000, then by 0.00348 to convert from board feet to cubic meters. Next we converted volume to weight by using different standard conversion factors for softwoods (0.405) and hardwoods (0.576)
	Plywood from mm sq. ft. (softwood)	1000*0.295*0.405	We first multiplied by 1000, then by 0.295 to convert from square feet to cubic meters. Next, we converted volume to weight by using the standard conversion factor for softwoods (0.405)
	Pulp products in short tons	*0.907	To convert quantities given in short tons into metric tons, we multiplied by 0.907
Howard and Liang (2019)	Timber products (softwoods and hardwoods) in million cubic feet	*1,000,000*0.02832*0.405 (softwood) *1,000,000*0.02832*0.576 (hardwood)	We first multiplied by 1,000,000, to convert from cubic feet to cubic meters by 0.02832, then used different standard conversion factors for softwoods (0.405) and hardwoods (0.576)

Note: See text and Appendix A for details about the conversion factors.

Source: Defebaugh (1907), U.S. Bureau of the Census (1975), and Howard and Liang (2019).

The categories of timber products traded differ from period to period, as do the names used to designate them. For example, for imports, Defebaugh (1907) lists the following categories: “Timber hewn and sawn” (1898–1905), “Boards and other sawed lumber” (1871–1905), “Mahogany” (1899–1904), “Shingles” (1872–1905), “Wood pulp” (1890–1904), “Logs and round timber” (1895–1905). On the other hand, the U.S. Bureau of the Census (1975) uses the following categories: “Lumber” (1900–1970), “Plywood and veneer” (1938–1970), “Pulp products” (1900–1970), “Logs” (1939–1970). To develop a consistent time series, we established categories of forestry products that could be applied for the entire period (Table 2). While the products traded are very diverse in nature, a typology according to their degree of transformation makes it possible to categorize them.

Table 2. Classification of aggregated wood product categories used in this analysis and corresponding products reported in original sources, and respective FAO categories and products.

Aggregated product categorization used in this analysis	Raw materials	Primary products	Final products
Products reported in original sources	Logs and round timber	Sawed lumber, wood pulp, sawn and hewn lumber, lumber (softwood & hardwood) and pulp wood chip	Shingles, boards, planks, joists and scantlings, plywood, pulpwood-based products, panel products, plywood and veneer
Corresponding products classified by the FAO	Wood in the rough	Wood chips and particles, residues and recoverable wood products and Sawn wood.	Veneer Sheets, and Wood-based panels

Table 3. Average U.S. imports and exports by main trade partners (countries and world regions) as fractions of total timber products traded for three periods, 1871–1905, 1965–1996, 1997–2017.

	Physical imports			Physical exports
	1871–1905	1965–1996	1997–2017	1871–1905	1965–1996	1997–2017
Canada	87%	45%	66%	3%	9%	20%
Mexico, Central, South America	<1%	<1%	2%	21%	<1%	14%
UK	<1%	<1%	<1%	5%	<1%	1%
Europe, other	<1%	<1%	4%	5%	4%	8%
China	<1%	1%	3%	<1%	2%	7%
South East Asia	<1%	2%	2%	<1%	<1%	1%
South Korea	<1%	2%	<1%	<1%	2%	2%
Japan	<1%	<1%	<1%	<1%	27%	8%
Other countries/regions	12%	49%	20%	66%	55%	39%

Sources: Defebaugh (1907), Howard & Liang (2019).

Notes: Data missing between 1906 and 1964; between 1965 and 1996, pulpwoodbased products were not reported. All products (RM, PP, FP) covered between 1997 and 2017. Yearly average shares of total.

Three degrees of processing were harmonized with the recent classification made by the Food and Agricultural Organization of the United Nations in 2022 (FAO 2022). These categories, originating in 1956 from the FAO’s aim to enhance economic integration on global timber markets, serve us as very rough proxies for the degree of processing in the timber products traded (Bemmann 2016). The first category is called “raw materials”: it does not correspond to trees that have just been felled, as these are not marketed abroad, but to products that have only undergone a relatively simple transformation in order to be transported (e.g., removal of bark, branches and foliage). In the historical records, log and round timber correspond to this category, which the FAO classifies as “Wood in the rough”. The second category, “primary products” (also referred to as “semi-final products” in statistics) includes products that have undergone more than one transformation for a specific use, without being directly usable as such. These include the following products: sawed lumber, lumber (softwood & hardwood), timber (hewn and sawn), wood pulp (sawn and hewn), and pulp wood chips. This category corresponds to the FAO (2022) classification “Wood chips and particles, residues and recoverable wood products” and “Sawn wood”. Finally, the third category, “final products” includes the different timber products which, after having undergone several transformations, are ready to be used for a specific purpose. We have grouped the following products under this terminology: shingles, boards, planks, joists and scantlings, plywood, pulpwood-based products, panel products, plywood and veneer. These categories correspond to the following sections determined by the FAO (2022) classification: “Veneer Sheets,” and “Wood-based panels.”

Data gaps were found in the Bureau of the Census records. In the period 1900–1970, for aggregated totals as well as for some individual products, only net imports (imports minus exports) instead of imports were reported. To obtain import figures, we simply summed the net imports with exports if both flows were reported. However, for some individual products, data were not available for the entire time-period for all trade flows: Import numbers for the final products category (in particular plywood and veneer) were only available from 1927. Logs (raw material category) were only available from 1950. Only for lumber and pulp products, data were available for the entire period 1900–1970. Moreover, the methodological notes of the Bureau of the Census specify that certain final products, such as paper and cardboard, were aggregated under pulp products. Using additional data from the Statistical Abstract series, we could reconstruct most of the missing products for the period 1900–1938. For the period 1939–1950, we subtracted the sum of the three product categories from the total aggregated imports and divided the remainder among the three product categories, using constant 1938 shares of each category of products in the total for this period.

Data for exports split by product categories 1900–1970 were more consistent. Similar to imports, we used data from the Statistical Abstract series to overcome data gaps. Only for exports of raw materials, no data were available between 1938 and 1950. However, this category before and after this period represented only a small share of total exports (5–7%). Therefore, figures for exports for these 12 years may be slightly underestimated.

Our approach allowed us to establish a good approximation of the total trade volumes. Periods where the different sources overlap show good consistency in trends and quantities. The curves reported by Defebaugh (1907) and the U.S. Bureau of the Census (1975) intersect in the early 1900s. Around the end of the 1960s, Howard and Liang (2019) and the U.S. Bureau of the Census (1975) data show good agreement, despite slightly lower quantities for the raw material imports and higher numbers for exports reported by the latter. In total, the different sources (Defebaugh 1907; U.S. Bureau of the Census 1975; Statistical Abstract 1908, 1921, 1930) diverge on average by 3% in imports and by 6% in exports in 1900–1905, and by 2% in imports and by 15% in exports in 1965–1970 (U.S. Bureau of the Census 1975; Howard and Liang 2019).

Trading partners

Beyond distinguishing timber trade by product categories, we attempted to empirically investigate the importance of different foreign trading partners. However, this analysis was limited by data availability. Although for the beginning and the end of the analyzed period data were available, no consistent time series on countries of origin and destination of forestry products traded by the United States could be established for the entire timeframe.

The data of the Bureau of the Census extending from 1900 to 1970 contained no indication of trading partners. Moreover, the tables by Howard and Liang (2019) showing trading partners for the period 1965–2017 did not cover all products for the whole period and did not specify all trading partners. Due to these data limitations, we were only able to approximate timber trade with major partners in the periods 1871–1905 and 1965–2017.

Data for imports by trade partners for the period 1871–1905 were available for board planks, wood pulp, shingles, and mahogany. Exports in this period were only reported for boards, planks, deals, joists and scantlings, as well as shingles. For 1965–1996, imports and exports of paper products were not reported by trade partners. From 1997 onwards we could use supplemental FAO data, adding the missing portion of paper products.

Since only some of the products traded were reported by trading partner (especially for the first period), we aggregated imports and exports by trade partners and subtracted them from the total imports and exports (Figure 1). We summed the remainders thereof under the category “Other countries/regions”. Therefore, the respective shares for trade partners in total imports and exports in the first two periods can only be regarded as approximations. While not encompassing in temporal and structural coverage, this analysis enables us to investigate major trading partners of large timber product groups, and discuss the effect of the U.S. timber trade on forest dynamics in particular countries at the very beginning and at the very end of our period of analysis.

Figure 1. Physical imports and exports of timber products of the United States, 1870–2017. Combined timelines. For sources, products considered, and conversion methodology see Table 1 and text. Data in million metric tons per year. Physical exports are displayed as negative flows.

Results

Trends in imports, exports, physical trade balance

Figure 1 shows total aggregated physical imports and exports of timber products, as well as the physical trade balance based on the sources described in Table 1. For the rest of this work, we will aggregate the imports and exports of the different sources. Between 1870 and 1915, the physical trade balance was negative, meaning that the United States exported more than it imported. In comparison to later periods, quantities were limited and did not exceed 10 Mt/yr (Mt/yr = Megatons/year = Million tons per year). From 1915 onwards, the physical trade balance became positive until the end of the period studied (except for 1930), meaning that the United States turned into a net-importer of timber products already in the early 20th century. Imports increased until the 1960s to reach more than 20 Mt/yr. Despite five relative declines corresponding to periods of crisis, the physical trade balance remained positive: During the crisis of the 1930s, in the aftermath of the oil price shocks in the early 1970s and early 1980s, and again after the early 1990s recession, which had resulted in a slump of office buildings (Gardner, Hipple, and Nardone 1994; Walsh 1993). The biggest drop - of the order of 30 Mt imported - occurred in the aftermath of the 2007 financial crisis. However, the overall trend until the 2007 crisis was a strong expansion of imports, to a peak of 67 Mt in 2005, resulting in an increase of net imports to the United States across the entire time series.

The United States also exported timber throughout the whole period. Before 1900, exports of timber amounted to no more than 4 Mt/yr. Until the crisis of the 1930s, exports had increased to almost 10 Mt/yr. However, after 1929, exports declined sharply, remaining below 2 Mt per year in the late 1930s and during the second World War. Exports did not return to the 1929 level until the mid-1960s (in 1965, 7 Mt were exported). They experienced strong growth between 1965 and 1990, reaching 30 Mt/yr. After a slight decline in exports until 2007 (18 Mt were exported in 2006), exports again rose to almost 30 Mt in 2017.

Physical trade balance of forestry products

The physical trade balances (PTB) of the three product categories distinguished in our analysis show diverging patterns (Figure 2). Before 1915, the United States exported more primary and final products than raw materials. Between 1915 and 1960, final and primary products constituted the bulk of imported products, hence developments in the total physical trade balance largely reflect developments in these categories. The PTB of raw materials was close to zero caused by low volumes of both imports and exports in raw materials, almost canceling each other out. From 1965 to the 2000s, primary products constituted the majority of net imports. There was a complete reversal of the trends compared to those between 1870 and 1915: Indeed, in addition to the fact that the total PTB turned largely positive, the composition of the products traded also changed remarkably.

Figure 2. Physical trade balance (PTB) of the main timber product categories traded in the United States, 1870–2017. From 1905 to 1950, data on raw materials and finished products are not accessible or aggregated in the primary products (e.g., paper and cardboard are listed with wood-pulp products, i.e., in the Pp category). Negative values indicate net exports, positive values net imports. Data in million metric tons per year.

The PTB of raw materials was close to zero before 1961, with net imports below 700,000 tons/yr. This indicates that the raw material wood resources of the United States until the middle of the 20th century were essentially consumed domestically. In the second half of the 20th century (from 1960 onwards) the PTB of raw timber materials was progressively negative. In fact, for the rest of the observed period, raw materials were the only product category of which exports exceeded imports. From 1970 onwards, net exports of raw materials increased to −5 Mt/yr and peaked at −9 Mt in 1990, after which they decreased but remained negative at around −5 Mt/yr. The PTB in primary products was negative before 1917, at around −2.5 Mt/yr, reflecting a relative specialization of the United States in the export of these products at that time. After a slight negative trend until 1929, the PTB in primary products became positive, and grew strongly to 10–20 Mt/yr between the 1960s and 1980s. During this period, it fluctuated between decreases (10 Mt in 1960 and 13 Mt in 1980; 10 Mt in 1990 and 8 Mt in 2010), and peaks (20 Mt in 1978, 23 Mt in 1985 and 37 Mt in 2005). Similar to primary products, the PTB of final products also reversed during the period under review. Before 1905, their PTB was mostly negative at around −1.5 Mt/yr. Especially after 1912, it turned positive, remaining between 5 and 10 Mt/yr, despite a peak in 2003 of around 15 Mt. After 2007, the PTB of final products dropped sharply to between 1 and 3 Mt/yr, reflecting a relative change from im- to exports of these products. In the second half of the 20th century, the United States turned into a net importer of primary and final timber products, while continuously acting as a net exporter of (a lesser amount of) raw timber materials.

Trade partners

Due to the limited data availability, we show average shares of imports and exports for main trade partners for the three distinct periods where a reasonable quantification was possible, discerning those trading partners presented in the sources used.

1871–1905

From 1871 to 1905, the main trade partners of the United States were Canada, the UK and other European countries, as well as Mexico, Central and South America (Table 3). Wood imports in this period came almost entirely from Canada (87% of total imports).

The composition of countries the United States exported timber to was different. Mexico, Central and South America were the most important destinations for U.S. timber exports (21%). On the contrary, only 3% were exported to Canada, 5% to the UK and another 5% to other European countries. The largest share of total U.S. timber exports however, on average 66%, was sent to other countries and world regions.

In terms of product groups, according to Defebaugh (1907), the majority of primary products exports (logs and sawed timber) and raw materials were sent to the UK (nearly 85% of the total logs or 19,391 tons). Final products (boards, planks, joists and scantlings, and shingles) by contrast were exported to Mexico, Central and South American countries, which in total absolute quantities were the largest trade partner of the United States for exports until the last decades of the 19th century. Shingles were exported to the West Indies (mostly to Antilles, Cuba, Jamaica, Puerto Rico, Martinique and Guadeloupe). The UK was for a time the best customer of U.S. primary forestry products (21% of total lumber exported in 1900), alongside Canada (16% of total lumber in 1900). The Central and South American countries, particularly Cuba (10% of total lumber in 1900), Mexico (13% of total lumber), and Argentina (14% of total lumber), also imported large quantities of wood from the United States.

1965–1996

In the second period, from 1965 to 1996, the composition of U.S. trade partners diversified, with China, South East Asian countries and in particular Japan becoming important, especially in terms of U.S. timber exports (Table 3). In comparison to the period 1871–1905, the share of wood exported to Mexico, Central and South American countries in the period 1965–1996 had dramatically declined to an average of below 1% of total exports. Japan emerged as the largest single market for U.S. timber, accounting for almost one third of total U.S. wood exports.

The trading partners from which the United States imported timber products changed much less dramatically: Canada remained the most important partner in the late 20th century. However, the average contribution of Canada in total timber U.S. imports declined strongly, from 87% to around 45% in this period, while on average almost half of wood products imported stemmed from other countries and world regions.

In terms of products, in the 1960s, the United States started to import hardwood (veneer and plywood) from Asia and Southeast Asia. On average, over the period 1965–1996, 50% of U.S. hardwood plywood imports came from Japan, the Philippines, China and South Korea. According to Darr (1975), the United States purchased 89% of the total hardwood plywood exported from South Korea, and respectively 75% from Taiwan, and 83% from the Philippines in this period. For these three countries, long-term political and economic relations with the United States have contributed to this development.

1997–2017

In the last period observed, the composition of U.S. trade partners had changed further, although it must be noted that only for this period, data for all wood products traded by destination and origin became available, since paper products by trade partners were only reported in this period by the FAO. These products represented a considerable portion, accounting on average for 40% of total exports and 28% of total imports during this period.

Between 1997 and 2017 Canada still was the most important country from which the United States imported wood (on average 66% Table 3). However, the mix of countries from which the United States imported wood had further diversified – the average contribution of European countries had risen to 4%, China accounted for an average of 3% of total imports, while another 2% came from Mexico, Central and South American countries.

Compared to the other two periods, exports to Canada from 1997 to 2017 had significantly increased, accounting on average for 20% of total wood exports. Mexico, Central and South America where the second most important export destination for the United States in this period, accounting for an average of 14% of total wood exports. Japan, China, and European countries each accounted for around 8% of all wood exports during this period.

Discussion

Reconstructing almost 150 years of trade in timber products based on socio-metabolic indicators, the database presented here provides new insights into the dynamics of physical timber trade in the United States in terms of traded products and trading partners. We interpret these data (1) in view of their connection to the U.S. forest transition, inquiring whether and to which extent domestic forest recovery was enabled by net imports of timber, and (2) with regards to their implications for ecologically unequal exchange, investigating which (types of) products were traded with which trading partners across time, and implicitly on which resources (wood or labor) the United States gained access across the time period.

The shift of the United States from a net exporter to an importer of large amounts of timber products resonates with previous work on forest transitions: Across the time period, domestic forest biomass stocks increased by 47%, starting from the early 20th century onwards (Table 4; see also Birdsey, Pregitzer, and Lucier 2006; Magerl et al. 2019). This coincided with the period when the physical trade balance of U.S. forestry products became positive (even if gross exports continued to increase), suggesting that trade may, overall, have contributed to facilitating the forest transition in the United States by means of externalizing wood production abroad (Gingrich et al. 2022; Magerl et al. 2022). According to Fedkiw (1989), imports significantly contributed to increasing American consumption during the 20th century, suggesting that without trade, the same nature of extraction of wood, all else being equal, would have contributed to a severe degradation of American forests. In the early decades of the 20th century, public concerns arose about an imminent “timber famine”, should harvest levels continue (MacCleery 1993). A report by the USDA Forest Service (1920) underlined the strong reliance of the United States on timber imports, confirming the important role of timber trade for forest change. The supply of timber was still a national concern in the 1960s and 1970s, when softwood lumber and plywood prices rose significantly, driving the increase in imports from Canada (Fedkiw 1989; Lower, Carrothers, and Saunders 1938). Our results confirm that particularly Canadian forests were instrumental in providing timber products to U.S. processing industries and consumers, while, surprisingly, the United States increasingly exported raw wood in the late 20th century. We identify three distinct phases in this development.

Table 4. Forest transition indicators, based on Magerl et al. (2019), Magerl et al. (2022) and their connection to international trade of wood in the United States – area (mha = Mega hectar = million hectar), biomass stocks (GtC = gigatons carbon = one billion tons carbon), harvest, and trade (Mt/yr = megatons per year) for the three periods discussed (1870–1914, 1915–1960, 1961–2017) in absolute numbers (left panel) for selected time points and compound average annual growth rates for the three periods (right panel).

			1870	1914	1960	2017	1870–1914	1915–1960	1961–2017
Forest area
			Mha				%Δ
			303.0	257.6 ^#	255.3	257.1	−0.4%	0.1% **	<0.1%
Forest biomass
			GtC				%Δ
	Stocks		NA	28.6^+	29.1	42.1	NA	0.2% *	0.6%
			Mt/yr				%Δ
	Harvest	Total	61.8	276.4 ^#	263.7	282.9	3.9%	0.3% **	0.6%
	Imports	Total	1.2	2.9	18.7	42.6	2.2%	2.1%	1.2%
		RM	NA	0.3	0.5	0.3	NA	3.5%	0.8%
		PP	1.2	1.2	8.5	27.0	< −0.1%	3.4%	1.6%
		FP	0.05	1.4	9.8	15.2	8.9%	1.4%	0.5%
	Exports	Total	0.6	6.4	3.2	30.0	6.2%	−1.1%	1.8%
		RM	NA	0.3	0.4	4.9	NA	2.7%	−0.1%
		PP	0.4	3.9	1.4	11.7	6.3%	−1.5%	3.2%
		FP	0.3	2.4	1.4	13.4	6.0%	−1.2%	2.0%
			Mt/yr
	PTB	Total	0.6	−3.5	15.6	12.6
		RM	<0.1	<0.1	0.1	−4.5
		PP	0.9	−2.7	7.1	15.3
		FP	−0.2	−1.1	8.4	1.8

Notes: Absolute exports displayed as positive numbers. Forest area: productive protected and used forests, including shrub- and woodlands. Stocks: total above- and belowground living and dead biomass, including litter. Harvest: industrial wood and fuelwood, including forest biomass grazed. Blue hatchings indicate % growth, red hatchings indicate % decrease. NA: no data available. Δ indicates % change.

⁺ = 1907, ^# = 1910, * = 1907–1960, ** = 1910–1960.

In the period before the First World War (1870–1914), U.S. forests were under severe pressure, expressed by the net export of wood products, the large increase in wood harvest quantities, and the substantial decline of domestic forest areas (Table 4). The agricultural frontier expansion had led to large-scale forest clearing during European settlement since the 17th century, contributing to the observed declines in forest area (MacCleery 1993). Timber products consumed domestically, providing for the expansion of railroads, telegraphs, ties, utility poles and fences, stemmed mainly from domestic forests. Together with other pressures like forest fires, this caused forest degradation and deforestation (Williams 1988, 1992). However, trade in timber products compared to the later periods was limited in quantities, explained by the high trading tariff barriers surrounding wood and its products (Defebaugh 1907). The first protectionist measure of the period, the Morill Tariff, was enacted in 1861, in the midst of the American Civil War, and remained in place until 1915. In the context of the conquest of the American West and the extension of the settlement frontier, European settlers saw no need to bring massive amounts of wood from other continents. Defebaugh (1907) describes the shift in this frontier between 1860 and 1900 (p. 521): The New York area and Pennsylvania dominated the market at first, before giving way to Michigan and the Great Lakes region. When the supply of the upper Midwest dwindled, the American wood industries flourished in the Pacific Northwest, in states like Idaho. Our data display that total export increases in this period were driven by primary and final timber products. This indicates that the United States acted as a provider of timber products to global consumers (mostly in Latin America) - The domestic economy was able to generate value added in domestic sawmills.

In the second period between 1915 and 1960 (Figure 2, Table 4), the physical trade balance was characterized by the increase of net imports of primary and final products to meet the needs of an emerging consumer society in a context of economic crisis and war. The increase in imports in the first half of the 20th century, and in particular in the 1930s, is largely attributed to the increase in pulp imports. This is explained by the strong increase in paper consumption, whose value in millions of dollars doubled between 1930 and 1950, driven by the emergence of a paper-, and in particular newsprint industry. The development of new printing techniques (offset technique) and the rising demand for products such as cardboard boxes, books, tickets for fines, and single-use sanitary paper products contributed to increases in pulp imports (USDA Forest Service 1958). While net imports increased in this period, domestic forest biomass started recovering, despite a continuous slight decline of forest areas (Table 4). During the Great Depression (1930–1935), domestic production of industrial wood sharply dropped (Magerl et al. 2022). Many timber companies were forced to shut down during this period (Duncan 1941). The subsequent increase in the use of paper and pulp-based products drove the imports of processed and final products, stemming largely from Canada. The role of the United States thus changed not only in terms of its role from supplier to recipient of wood in global markets. Additionally, a larger share of (net) imported products were already (semi-)processed abroad.

The Forest Reserve Act, which set national forests out of use, may have played an additional role in changing the values of wood products and thus influenced the structure of wood trade. In 1891, the Forest Reserve Act was established, and large amounts of national forests were categorized as “federal lands” (Steen 2004), limiting logging and other uses of forests. It also changed public land policy in favor of conservation with the proclamation of lands as national reserves (Walton and Rockoff 2013). Indeed, the period of limited logging of national forests coincided with the increase in imports of timber products. Similarly, the entry of the United States into the second world war in 1941 highlighted the strategic role of forests, and the constitution of forest reserves, since during the war, wood became an essential construction material. It was used for barracks, billets, ships and docks, but also for war factories, housing, gun stocks, explosives, aircraft, boxes and crates for war supplies (USDA Forest Service 1948).

Finally, in the third period 1960–2017, the United States started to increasingly export raw materials of wood, while in sum still remaining a net importer of timber products, due to the continuously high imports of processed and finished products. Similar to the previous period, forest biomass stocks recovered during this time (Table 4). Despite the increase in national wood harvest during this period, demand accelerated further, driven in particular by the massive increase in housing construction, thus enlarging imports (Howard and Liang 2019).

However, softwood supply became a problem in the early 1960s, due to price pressure: during this period, lumber and plywood prices rose well above the overall rate of inflation, in response to rising housing demand and the scarcity of softwood lumber on the supply side. The shortage contributed to rising housing costs. The Johnson (1963–1969), Nixon (1969–1974), Ford (1974–1977) and Carter (1977–1981) administrations took on the issue of softwood supply, notably through improved national monitoring, such as the large national forest inventories. The strategy was to reduce timber and plywood prices to counteract the effects of inflation through consistent monitoring of the timber reserves. However, in favor of forest conservation and recreational use, the expansion of logging operations in national forests were limited. Hence, timber harvest and sales from national forests during the 1970s remained below those achieved in the 1960s, despite a law passed in 1974 for national forest use and management planning (Fedkiw 1989).

In addition, the United States negotiated international treaties eliminating almost all tariff barriers on timber. With the Uruguay round and the North American Free Trade Agreement (NAFTA), most tariffs on forest products were on track for total elimination by the late 1990s. As a result of these treaties, American consumption of wood products surged, provided importantly from international markets, with Canada continuing to act as the major supplier of processed and finished wood products (Table 3). Lumber consumption peaked in 2005, which was a record high that even exceeded levels in the early 1900s when lumber was the most important raw material used in the United States for construction, manufactured products, and shipping (MacCleery 1993). The 2007 financial crisis was a revealing moment of the powerful links between the real estate bubble, wood imports and the role of trade globalization. The impact of the subprime crisis contributed to a drop in softwood lumber imports of over 64% (Howard and Liang 2019). The last period thus continued some of the trends from the mid-20th centuries (increasing net imports of processed materials, visible impacts of economic crises), but added the feature that simultaneously, a consistently high amount of raw material was exported, mainly to Japan, Canada and China (Table 3).

In the context of ecologically unequal exchange, our findings suggest that U.S. timber trade does not seem to reflect the predominant global patterns of asymmetrical net trade flows of biophysical resources from poorer to richer countries (Dorninger et al. 2021): the United States imported mainly from Canada, mirroring the political ties of the United States with its Northern neighbor.

Nevertheless, the United States became a major importer of primary and final products over the period observed. This shows how the country increasingly gained access not only to foreign timber, but also to labor for wood processing and manufacturing, in Canada but also increasingly in China, Southeast Asia and Europe. These findings also stress that the U.S. wood industry became optimized for producing raw materials rather than manufactured products.

Our study opens up a range of opportunities for future research. The U.S. case and its specific characteristics could be further compared to other case studies on wood trade and forest change e.g., to that of Germany (Daheur 2016). Studying the specific types and species of wood traded could further enhance our understanding of what drives trade dynamics. For example, this could explain why certain types of wood were preferred over others and what their origin was. A more detailed analysis of long-term U.S. trade data in the context of ecologically unequal exchange would require a more consistent data set of trade partners, ideally combined with monetary and energy flows. This could further increase our understanding of the teleconnections between domestic forest transitions and international wood trade.

Conclusion

This study provides a new, quantitative and comprehensive approach for analyzing the connection between the history of U.S. wood trade and its domestic forests. We synthesized data from a variety of sources over a long time-period to present a novel dataset, covering almost 150 years of timber trade history. Using this dataset, we analyzed changes in the structure, magnitude and trajectories of U.S. trade of timber products, as well as major trading partners. We contextualized our results in terms of forest transitions and ecologically unequal exchange.

Our results confirm previous research highlighting the importance of net timber imports in the context of forest transitions for the specific case of the United States. From a net exporter of high-quality timber products, the United States has shifted to a net importer of processed timber products. This shift began in the mid-20th century, mainly driven by the rising demand for manufactured timber products, but also due to a ban of political trade tariffs. At the same time, the United States became a supplier of raw wood to global markets. The main trade partner for wood imports and exports of the United States was Canada, making it an unusual case in terms of ecologically unequal exchange. However, European and Asian countries emerged as important suppliers of primary and final timber products at the end of the observed period, reflecting how the United States outsourced labor for wood manufacturing abroad.

The methods presented here can be applied in future research to more systematically investigate the biophysical dimensions of international trade of other countries. The dataset, beyond serving to contextualize future work on wood trade, processing and use in the United States, can additionally inspire further research on the ecological impacts and long-term legacies of trade in exporting countries, or the increasing international connectedness of biomass production and use during the 19th and 20th centuries. Our study thus presents a valuable new perspective for historical research in general and material environmental history in particular, proposing a biophysical approach to analyze long-term trade dynamics connected to forest transitions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (ERC StG ‘HEFT’, Grant Agreement No. 757995). The Open access publication was supported by BOKU Vienna Open Access Publishing Funding.

Appendix A

Table A1. Conversion from imperial to metric units by original source.

Abbreviation	Unit	Conversion to	Factor	Source
ft³	Cubic feet	Cubic meters	0.02832	FAO (2020)
FBM	Board feet	Cubic meters	0.00348	FAO (2020)
CD	Cord	Cubic meters	2.36000	FAO (2020)
rndwd eq ft³	Cubic feet roundwood equivalent	Board feet	6.95200	McRoberts et al. (2007)

Table A2. Conversion of wood products from volume to weight by wood type.

Wood type ^a	Density [tons at 15% moisture content/m³]
Coniferous/softwood	0.405	Krausmann et al. (2013)
Non-coniferous/hardwood	0.576
Average	0.491

^aUsed for all industrial timber products and fuelwood.