Page of

A Project of Uneven Liberalization

A Project of Uneven Liberalization

(p.85) Chapter Three A Project of Uneven Liberalization
Global Rivalries
Amy A. Quark
University of Chicago Press

Abstract and Keywords

Chapter three demonstrates that projects to create new institutions are often trial-and-error, ad hoc efforts, as institutionalist scholars suggest, but they are also driven by competitive efforts to shape the terrain of market competition. Actors create new institutions to solve the problems they face given their historically and spatially specific position within patterns of capital accumulation. Moreover, the efficacy of these institutions is limited by the patterns of conflict that they generate. This chapter traces the origins of the contemporary struggle among a U.S.-led coalition of the state, transnational merchants, and cotton producers, its rivals in China, and more marginalized actors. It explores the rise of a US-led neoliberal project in the 1970s and the efforts of the US state and transnational merchants to recast quality standards and dispute settlement to privilege their preferences in this liberalizing environment. As these U.S. institutions came to be seen as de facto global institutions, however, they generated new patterns of conflict, which limited the enforceability of these rules.

Keywords:   Conflict, Market liberalism, Neoliberalism, United states, Merchants, Enforceability

When I visited Banikoara, a small village in the northern part of Benin, in 2006, the cotton harvest was almost over. The long journey from the country's capital, Cotonou, to Banikoara had involved a motorcycle taxi, a longdistance bus, and finally for the last few hours a taxi packed with people and supplies. With the windows rolled down, my face became caked with dust from the red soil as I observed the countryside. Farmworkers were gleaning the last bits of cotton lint from fields that had already been picked over. As trucks carried the season's yield toward the local cotton gins, tufts of cotton spilled out, lining the road in their wake.

With simple buildings and dirt roads, Banikoara seemed far away from other stops on my research circuit, which included plush office buildings and five-star conference hotels. But as I began to talk with farmers and local leaders, it became clear that the issues debated at international conferences and in corporate boardrooms were just as salient to the people of Banikoara as to those working for transnational cotton merchants or the USDA. Indeed, as the small tenant farmers in the post-Civil War US South and the cotton growers who marched alongside Gandhi in the khadi movement were concerned about their ability to get a fair price for the cotton they grew, so too were the residents of Banikoara.

A local farmer demonstrated this to me through the meticulous records he had been keeping since the 1970s. He had tracked his yields and the prices he had received each year. Things were tough in 2006. As the farmer explained, prices were low and the cost of inputs kept increasing. It did not help that the state-owned cotton company, SONAPRA, had been undergoing a process of privatization at the behest of the World Bank and International (p.86) Monetary Fund. Due to the disruptions caused by this restructuring process, many farmers had not yet been paid for their previous year's crop. I later visited the local market where farmers were delivering their cotton. I watched as the cotton was weighed and graded before it was sent off to the gin where the cotton fibers would be separated from the seeds, pressed into bales, and prepared for export. Other cotton farmers at the market expressed similar concerns. One farmer joked that the market was where they brought their cotton not to get paid, but to avoid getting paid!

Leaving the market, a local official steered me toward his office. He wanted to ensure that I understood the bigger picture that the farmers of Banikoara were facing, particularly given my position as a researcher from a US university. He dug in his desk drawer and pulled out a stack of papers. It was a petition signed by many in the community, calling for the elimination of US cotton subsidies.

From the cotton fields of Banikoara to the highest-level trade negotiations at the WTO, the issue of cotton subsidies framed cooperation on all other issues. Not only did cotton farmers in Banikoara see many of the challenges they faced as related to the subsidy issue, so too did many in the international community. Tensions over what was seen as a highly uneven US project of market liberalization spilled over into sector-specific negotiations over the harmonization of quality standards and dispute settlement arrangements. A representative from the ICAC explained that collective action on quality standards and other issues was difficult given resentment toward the United States, especially with the failure of the Doha round of WTO negotiations. “The psychological effect of the failure of Doha has been enormous,” he noted, “People are angry and are less likely to cooperate.” In 2003, the Cancun WTO meetings had collapsed. Further cooperation on trade liberalization was stalled, and, as the Organisation for Economic Cooperation and Development reported, “The failure to reach agreement on cotton was a key reason for this impasse” (as cited in Heinisch 2006:262).

In this chapter, I explore the origins of these contemporary axes of conflict by tracing how strategic actors pieced together new rules of the game for the cotton trade from the 1970s to the 1990s. In doing so, I demonstrate the utility of institutional analyses that characterize institution-building as a trial-and-error, often ad hoc, process, but insist that institution-building must also be understood as driven by competitive efforts to shape the terrain of (p.87) market competition. Actors craft new institutions to address the problems they face given their historically and spatially specific position within patterns of capital accumulation. Moreover, the efficacy of these institutions is limited by the patterns of conflict that they generate.

This chapter explores how US actors reconstituted the institutional arrangement created out of the struggle with Liverpool merchants (see chapter 2) as part of a broader US project of market liberalism. Much as Liverpool merchants forged their private authority over quality standards in an effort to navigate the uncertainties of a liberalizing economy in the 1870s, a century later US textile manufacturers, cotton producers, merchants, and the US government retooled the existing quality classification system and dispute settlement arrangement as they sought to address both the new opportunities and the new competitive pressures presented by the broader, US-led liberal market project. In doing so, they created a new de facto global quality classification system and dispute settlement arrangement during this period.

Moreover, just as Liverpool merchants' position as de facto transnational rule-makers generated particular patterns of social conflict and resistance, US producers, the USDA, and transnational merchants faced growing resistance to their new governance arrangements as they came to be seen as de facto global arrangements. I will demonstrate how tensions began to build as the new governance arrangements were interpreted as part of the United States' uneven liberalization project that served to stack the deck in favor of US players. These tensions would ultimately boil over into more coherent redirection and protection strategies on the part of emerging rivals and marginalized actors, respectively.

The contemporary axes of conflict are not simply carbon copies of the earlier struggle between actors in the United States and Britain. While the patterns of conflict generated by projects of market liberalism share similarities, they also take historically specific forms. The US neoliberal project, for example, was not an entirely coherent liberalization project. Rather, it unfolded in a highly uneven manner as it was influenced by domestic conflict within the United States over the trajectory of liberalization. The unevenness of this project shaped the nature of quality governance as well as the institutional strategies that challengers developed in response to it.

The Uneven Project of us Market Liberalism

At the turn of the twentieth century, the USDA had wrested control over two key quality governance functions—the definition of quality and the (p.88) creation of benchmark standards—from Liverpool merchants and maintained authority over these functions into the postwar period. From the 1970s to the 1990s, the USDA cast these functions in new form as part of a broader project of market liberalism. During this period, the US state, in cooperation with private firms and cotton producers, constructed an automated system of cotton classification that would replace the existing manual system of classification. This new system emerged out of and was shaped by the particular contours of the US liberal market project.

By the 1970s, US firms were facing a crisis of profitability. The 1950s and 1960s had brought prodigious expansion in world production and trade. Through this expansion, Japan and the Western European countries, especially Germany, successfully recovered from WWII and began to compete and even forge ahead of US firms in one sector after another, from textiles to steel (Silver and Arrighi 2003:342). The US state responded to these competitive pressures through a “revival of the liberal creed” (Silver and Arrighi 2003:341).

The form that this project of market liberalism would take, however, was shaped by domestic conflict. Distinct segments of the US economy experienced this crisis of profitability differently. Importing firms and transnationally competitive exporters saw the solution in transnational expansion and began to push the US government to more aggressively pursue a trade liberalization agenda. However, US sectors such as textiles, apparel, and agriculture, which would not be competitive in a liberalized environment, sought protectionist measures (Chorev 2007). Given trenchant domestic conflict, the US government carved out a liberal market project that was highly uneven across different economic sectors. Nowhere was this more apparent than in the textile, apparel, and agricultural sectors, which would become the crux of what some claimed to be the hypocritical nature of the US liberal market project. While pushing other countries to enact trade liberalizing policies, the United States would establish its powerful position in the cotton textile sector through protectionist policies.

First, the US state maintained its long-standing agricultural support programs during this period, which allowed the US producers to establish a dominant position, claiming over one-third of all exports during the 1970s and 1980s (ICAC 2010a) (see Figure 3.1). The US agricultural subsidy regime was designed to combat declining agricultural incomes resulting from overproduction and declining commodity prices. Although the problem of overproduction dated back at least as far as the Civil War (Wright 1978), the problem was not directly addressed until the New Deal when the federal (p.89)

A Project of Uneven Liberalization

Figure 3.1. Major Cotton Exporters, 1980–2009. Adapted from ICAC (2010a). Each year refers to a marketing year—that is, 2000 refers to the marketing year from August 2000 to July 2001.

government began to face growing unrest in the wake of the Great Depression. The Agricultural Adjustment Act (AAA) of 1932 (revised in 1938) offered price supports, or subsidies, to wheat, corn, and cotton producers as an incentive to cooperate with acreage controls (Friedmann and McMichael 1989; Winders 2009).1 This supply management policy did reduce cotton production and export market share temporarily in the 1930s. At the same time, however, it had unintended effects. As conflicts with sharecroppers and wage laborers, as well as competition on the export market, intensified, large cotton producers began to use government subsidies to finance the mechanization of cotton picking, rapidly displacing sharecroppers in the 1940s and 1950s. When combined with the introduction of synthetic fertilizers, insecticides, and herbicides following World War II, the transition to mechanized cotton production effectively undermined prevailing attempts to restrict the supply of cotton and allowed the United States to regain its competitive position on the export market (Mann 1987). The subsidy regime would remain in place into the twenty-first century, albeit with some changes.2 While supporting farm incomes in the United States, these policies indirectly served to artificially lower the international price of cotton, thus discouraging international production and competition. Rather than addressing the problem of overproduction, domestic cotton producers and merchants sought to offset declining prices with demand-side strategies intended to create new markets for US exports (Brown 2000).

(p.90) The United States also ensured the continued competitiveness of its textile and apparel industry through protectionism, despite a broader policy of trade liberalization. Asian textile and apparel manufacturers had grown quickly in the immediate postwar period through a combination of state management, US support, and Western market access. In hopes of containing Soviet expansion in the 1950s and 1960s, the US government had provided funds to help modernize textile industries in places such as Japan and Hong Kong, South Korea, Singapore, and Taiwan, the so-called Four Asian Tigers, and provided access to US markets (Rosen 2002). While US textile manufacturers balked at the practice, cotton producers and merchants directly benefited and lobbied in support of efforts to create new markets abroad. For example, cotton merchants lobbied in support of the Marshall Plan, which helped finance the sale of cotton throughout Europe, and Public Law 480, which allowed over 25 countries to import agricultural commodities including cotton with local currency (Brown 2000; Dunn 1992).

With this support, Asian textile manufacturers had begun to outcompete Western manufacturers in their own markets. Thus, while the US government began to use the General Agreement on Tariffs and Trade (GATT), bilateral negotiations, and structural adjustment programs through the IMF to evoke liberalization policies from its trade partners, the powerful textile and apparel lobby successfully persuaded the US government to negotiate the Multi-Fibre Arrangement (MFA) in 1974—and renew it in 1977, 1981, and 1986—to ensure protection from foreign imports. Through the MFA, textile and apparel exporters, and particularly the major exporters in Japan, Korea, Taiwan, and Hong Kong, agreed to a system of “voluntary” quantitative restrictions on textile and apparel exports to the United States (Rosen 2002).3

The stated goal of the MFA was to facilitate the “orderly” expansion of trade in textiles and apparel, such that it could benefit both countries in the global North and those in the global South. Through the MFA, states established individual quotas specifying the precise quantities of different textile and apparel products that could be exported from one country to another. In practice, MFA import quotas limited the expansion of textile production and exports from “developing” countries, and particularly the most competitive countries in East Asia (Rosen 2002). Western countries were to gradually open their markets to exports from the South, while maintaining safeguards to minimize “disruptions” in their domestic markets. In practice, however, quotas became more rather than less restrictive as the United (p.91) States and European Union sought to protect their domestic manufacturers and “play politics” with textile and apparel quotas (Dicken 2007).

Though not a party to the original MFA, China also signed bilateral agreements establishing MFA-style quotas. The Chinese government took full advantage of possibilities to expand their textile and apparel exports, manipulating the quota regime where possible (Rosen 2002). However, China's continued expansion, like that of MFA countries, was ultimately limited by its access to Western markets.

Together, the MFA and the bilateral agreements between the United States and China had critical implications for the logic of trade in the textile and apparel sector and thus for the geography of the global cotton trade. As the size of countries' textile and apparel sectors was limited by the size of their allotted quota, the quantity of cotton they imported and their import market share was also limited by the size of their quota, effectively decentralizing cotton imports. For example, from 1974 to 1992, US cotton producers claimed, on average, around a 35-percent market share of cotton exports while the largest cotton importer in this period, Japan, had only a 16-percent share of cotton imports. No other single importer accounted for more than 9 percent of total world imports (ICAC 2010a). The MFA thus indirectly reinforced the power of US producers vis-à-vis cotton importers by minimizing their dependence on any one importer.

Protection through the MFA allowed textile manufacturers to stave off their own decline within the United States' broader project of market liberalism. However, given that the ultimate goal of the MFA was to eventually liberalize the apparel and textile trade, another critical component of this strategy was cost cutting and technological upgrading to enhance their competitiveness vis-à-vis competitors in East Asia. In the 1960s and 1970s, textile and apparel manufacturers successfully lobbied the government for tax write-offs and other assistance to invest in new technologies that would increase economies of speed and deskill the labor process (Rosen 2002:90–91). To this end, US textile manufacturers adopted a new spinning technology: open-end (rotor) spinning. Open end spinning was attractive to US manufacturers for several reasons (Perkins et al. 1984:484–85). Openend spinning brought textile manufacturers one step closer to completely automated yarn manufacturing, allowing them to reduce their labor force. Moreover, it reduced the number of processes in yarn spinning, streamlining production. Finally, open-end spinning brought significant economies of speed; its production rate was three to five times faster than the ring spinning technology that had conventionally been used. Through these (p.92) technological upgrades, the textile industry became more consolidated as the competition to gain financing for new capital investments pushed out smaller firms and brought a wave of buyouts and mergers, as well as greater vertical and horizontal integration (Rosen 2002:92).

The push to establish a mechanized system for cotton classification emerged out of the intersection of textile industry efforts to sharpen their competitiveness and the growing threat of a new rival for cotton producers: synthetic fibers (USDA 1984:48). Open-end spinning brought significant economies of speed, but it created other problems for US textile manufacturers. Openend spinning produced yarn and thus fabrics that were not as strong as those spun using the traditional ring spinning technology, which was still used by the United States' competitors in East Asia. As Jacobson and Smith explain:

Open-end yarn—however efficiently it might be spun—was lower in strength, and its performance in weaving was in fact less efficient than that of the ringspun alternative. The fabric made from open-end yarns tended to lower tensile and tear properties, and this caused problems with garment manufacturers farther down the line. (2001:212)

As such, some of the main users of cotton fabrics—blue jeans-makers like Levi Strauss and Company—began to complain that their fabrics were ripping and wearing out more quickly. One Levi's representative expressed to its textile manufacturer suppliers, “We don't care which one you use [open-end or ring spinning], but you better turn out the same product” (as cited in Jacobson and Smith 2001:213). Levi's and other companies began to introduce new strength standards, creating greater pressure on US textile manufacturers to balance the economies of speed they could gain with open-end spinning with the quality demands of its buyers. One way to overcome this problem was to pay greater attention to fiber quality (Jacobson and Smith 2001:209). Manufacturers using open-end spinning required a particular quality of cotton—cotton with stronger, finer fibers—in order to withstand faster spinning speeds and maintain yarn and fabric quality (Perkins et al. 1984). As textile manufacturers had to buy cotton of higher quality along these characteristics, the accurate measurement of cotton fibers became more important.

For textile manufacturers looking to more carefully manage fiber quality, synthetic fibers like polyester and rayon had certain advantages over cotton. Synthetic fibers could be produced with constant quality characteristics, which made them ideal as processing inputs. Cotton, in contrast, varied significantly with environmental conditions and production and ginning (p.93) practices (Mann 1987; USDA 1984). With these advantages, industrial producers of synthetic fibers both at home and abroad began to eat away at cotton's fiber market share within the United States. In the 1920s, cotton held an 88-percent share of the US domestic apparel market. As petroleumbased synthetics such as acrylics, polyesters, and spandex were introduced in the 1930s and gained widespread use after WWII, cotton's market share eroded to 66 percent and continued to decline to its all-time low of 33 percent in the US domestic market in 1973 (Jacobson and Smith 2001). Cotton researchers and industry players recognized that textile manufacturers could use cotton fibers as efficiently as synthetic fibers only if instruments could be developed to more precisely measure cotton fiber characteristics (Jacobson and Smith 2001:206).

Thus given their particular position with the competitive dynamics of the United States' uneven liberalization project, in the 1960s and 1970s, the USDA, with the support of US cotton producers' organizations, began researching the possibilities of mechanized classification—of making quality standards that gave more precise processing-relevant information (see Table 3.1). Mechanized classification could sharpen the competitiveness

Table 3.1. Key Cotton Fiber Characteristics and Their Relationship to Processing Efficiency and End-Product Quality. Adapted from Cotton Incorporated (2010); Perkins et al. (1984); USDA (2001).



Fiber length

Affects yarn strength, yarn evenness, yarn fineness, and the efficiency of the spinning process.

Length uniformity

Affects yarn evenness and strength, as well as spinning efficiency. Related to short fiber content.

Fiber strength

Affects yarn strength, allows faster processing speeds.


Finer fibers (with low micronaire) may require slower processing speeds. Finer fibers can produce stronger yarns (more fibers per cross-section). The more mature the fibers, the better the absorbency and retention of dyes.


Affects the ability of fibers to absorb and hold dyes and finishes and can reduce processing efficiency.

Trash/Leaf Content

Must be removed at a cost.

Short Fiber Content

Ratio of short to long fibers, which affects yarn strength and fineness.


Tangled knots of fibers that cause undyed spots during dyeing and finishing of yarn and fabrics.

(p.94) both of US cotton producers vis-à-vis corporate synthetic producers and of US textile manufacturers vis-à-vis competitors in East Asia. The USDA had begun scientific research on fiber quality measurement in the 1910s and 1920s (see chapter 2). From those early years, the application of scientific principles to improving the efficiency of textile manufacturing was understood as part of a strategy of national competitiveness. As an early USDA scientist commented after hearing a presentation on new developments in US fiber science: “I feel like shouting from the housetops that work like this will determine national security and supremacy” (letter from Dr. R. Y. Winter, director of the North Carolina Experiment Station, 1935, cited in Palmer 1961). Through this earlier research, a number of measurement instruments had been developed to more accurately measure some quality characteristics of cotton. While these measurement instruments did come to be used in some commercial contracts, they were mostly used by individual merchants and textile manufacturers to manage their inventories (Palmer 1961). Cotton breeders also began to use these measurement instruments in their efforts to breed cotton to meet spinners' demands (USDA 1984). These instruments did not, however, come into widespread use for the commercial trade as they were manually operated and rather time-consuming to use, making it impractical to employ them in the classification of the entire US crop. The exception to this was the micronaire instrument, which measured the fineness of cotton fibers, for which a rapid and reproducible measurement instrument was developed and included in the USDA's Official Classification procedure in 1966 (Perkins et al. 1984:447).

The goal in the 1960s and 1970s was thus to make these measurement instruments faster through automation and to integrate them into a single system that could be used to classify cotton for the commercial trade. The USDA worked with several private instrument manufacturers to achieve this aim. A Swiss company, Uster Technologies, eventually came to hold the patent for this new measurement system known as the High Volume Instrument or HVI system. The first US classification office was equipped with mechanized classification technology in 1980. By 1991, the United States had switched entirely to mechanized classification, and merchant and textile manufacturer associations around the world approved the HVI system for inclusion in the Universal Standards Agreement in 1995 (Knowlton 2005).

This new classification system represented a significant shift from the traditional quality standards based on manual classification, by shifting the definition of, as well as the “ways of knowing” about, cotton quality. (p.95) In the traditional manual system, US cotton was differentiated on international markets according to national origin (e.g., United States), staple length (the length of fibers), and grade (based on color, trash content, and gin preparation). Traditional manual classification was considered more of an art than a science. Some compared the expertise required to manually grade cotton to the evaluation of wine (DAGRIS 2002, as cited in Bingen 2006:229). An expert classer drew on all of her senses to evaluate a cotton sample. A classer judged the “touch” or feel of the cotton by pressing it into the palm of her hand to observe how quickly it resumed its form afterwards. Next the classer held a small sample of cotton between the thumb and forefinger of each hand and, through “a process of lapping, pulling, and discarding,” made the fibers parallel to judge the length, as well as the uniformity of fiber length in the sample (Perkins et al. 1984:446). These processes also revealed to the classer any leaves or seed fragments, which were considered trash, and allowed the classer to identify tangled knots of fibers (neps). The most experienced classers even ran the fibers across their tongue to assess “stickiness,” which results from inadequate pest management, and pulled the cotton next to their ears to detect the brittleness of the sound when the fibers separated and broke (Bingen 2006:229–30). Through this process and through comparison with the official benchmark standards prepared by the USDA, a classer would determine the cotton's length, as well as its grade (e.g., Middling or Strict Low Middling). The grade represented a composite evaluation, based on a classer's “overall impression” of the sample's color, leaf/trash content, and gin preparation, which included factors such as neps and overall smoothness or roughness (Perkins et al. 1984:441).

With the introduction of mechanized classification, classers' “sensory science” was replaced by machine testing. Rather than offering more general evaluations of staple length and the various factors contributing to grade, the HVI system drew on advances in electronics, lasers, and spectral analyses to provide more accurate and detailed measurements of individual fiber characteristics (Perkins et al. 1984:449). The mechanized system integrated three measurement instruments, each feeding data into a central computer. The first instrument measured a fiber's resistance to a puff of air to evaluate its fineness. The second employed a colorimeter, a device that detects subtle variations in cotton color, ranging from gray to yellow to white. A video camera was also used to detect leaves, stems, and other trash. Finally, the third instrument used air to draw fibers out to measure their full length and then pulled the fibers apart to test their strength (Lee 1996).

(p.96) These mechanized measurements were more precise than manual evaluations. With the mechanized system, fiber length came to be measured not in 32nds of an inch but in 100ths of an inch. They were also much faster. Instead of being able to test about forty samples in an eight-hour day, the mechanized system could measure 800 (Lee 1996). And in place of an overall evaluation of color, trash content, and gin preparation, the mechanized system provided individual measurements of length uniformity, fineness/maturity (micronaire), strength, and color (see Table 3.2).4 Moreover, in the manual system, the legitimacy of the standards was ultimately rooted in the benchmark standards, or the physical representations of the length and grade standards that the USDA prepared. These physical representations were based on fiber science and prepared by classing experts but were ultimately subject to inspection and approval by the US domestic industry and signatories to the Universal Standards Agreement. The mechanized classification system was also based on the preparation of benchmark standards, or physical samples of cotton considered to have “known” values that were used as reference material to calibrate the classification instruments. These benchmark standards were prepared through a tedious and exacting process by highly specialized experts. While they were accepted through the Universal Standards Agreement in 1995 as the International Calibration Cotton Standards (ICCS), the US industry and most foreign signatories to the Universal Standards Agreement lacked the specialized expertise required to evaluate them. The automated classification system thus significantly deepened the basis for the legitimacy of standards in a highly specialized Western science.

The data offered by the new mechanized classification system allowed

Table 3.2. Manual Classification versus High Volume Instrument (HVI) Classification.

With Manual Classification, cotton is sold based on…

With HVI Classification cotton is sold based on…

National Origin




Grade (overall assessment)

Length Uniformity

Micronaire (fineness/maturity measured by instrument since 1963)

Micronaire (fineness/maturity)

Color (evaluated manually until 2000)

Trash (evaluated manually)

(p.97) textile manufacturers to use a scientific approach to managing its raw material inputs. In 1982, Cotton Incorporated, a US cotton producer-funded research organization, introduced a computer software system, the Engineered Fiber Selection (EFS) system, to help textile manufacturers make use of the measurements from the United States' new mechanized classification system to increase their competitiveness (Jacobson and Smith 2001). The EFS software linked detailed fiber measurements to the precise characteristics of the yarn that textile manufacturers wanted to produce. It used regression equations to estimate the specific quality of cotton to purchase, the precise way to mix these cotton bales, and the spinning machine settings to use in order to produce a desired quality of yarn. This allowed textile manufacturers to match machine speeds and settings to fiber properties in order to increase processing efficiency, to better control end product quality, and to better manage raw material purchasing (Smith and Zhu 1999).

Moreover, the EFS software controlled for variance not only within a lot of bales but also over long periods of time. This feature was critical to overcoming a key advantage of synthetic over natural fibers: the variation created by fluctuating weather conditions from year to year and by geographically dispersed producers with different climates, soil qualities, etc. “The EFS program, when fed HVI data, could assure that the mix of cotton a mill laid down today would be just like (have the same specifications as) the one it had laid down a month ago and would lay down a month from now” (Jacobson and Smith 2001:215). It would later be estimated that the EFS system could save textile manufacturers between two and six cents per pound of cotton they used (Cotton Gin and Oil Mill Press 2005). The system quickly gained wide adoption among US textile manufacturers in the 1980s and early 1990s.

In sum, the USDA constructed a new automated classification system to replace manual classification as the basis for the Universal Standards Agreement. The USDA retooled this institutional arrangement in response to the changing competitive positions of both US cotton producers and textile manufacturers within the broader US project of market liberalism. Both US cotton producers and textile manufacturers stood to lose to competitors in the global South under the broader trade liberalization agenda and successfully lobbied for protectionist measures to shield them from this outcome, ultimately generating a highly uneven liberal market project. The retooled classification system represented part of this effort to navigate a liberalizing environment as it presented a way to enhance the global competitiveness of both US cotton producers and textile manufacturers.

(p.98) The Us System as the de Facto Transnational System

As competitive pressures intensified within the United States' uneven project of market-led development, the United States' mechanized classification system became the de facto global system. The pressure for the globalization of this classification system was rooted in the unexpected outcomes of protectionism within the broader trade-liberalizing agenda. While the textile industry had successfully maintained protectionism, liberalization in other sectors—such as finance—allowed apparel manufacturers and later retailers to operate within the loopholes and inconsistencies of the MFA rules. Apparel manufacturers realized they could effectively outsource their apparel and textile manufacturing to countries in the global South with access to quota and dramatically reduce production costs. The MFA thus gave Western apparel firms and retailers incentives to build complex subcontracting networks by piecing together import quotas allotted to a broad range of countries (Collins 2003; Rosen 2002). Indeed, as Dicken notes (2007:261), “the entire clothing industry of some developing countries was, in effect, created by MFA quotas.” By the 1990s, if apparel firms and retailers had not shifted production overseas, they increasingly could not compete (Collins 2003).

This intensifying competition also spilled over into quality demands. In the contest for consumer dollars, retailers and the new “branded marketers” (apparel firms that handle design and marketing but outsource production) began to shift toward product differentiation as a basis for competition (Larsen 2003:11). As Jonathan Zeitlin and Peter Totterdill explain, the “struggle for competitive advantage has come to center increasingly on retailers' and manufacturers' efforts to target specific groups of customers defined in new ways, to seduce customers with attractive, fashionable garments; to respond rapidly to short-term trends in the sales of individual product lines” (1989:162). To this end, fashion cycles intensified as retailers began to switch out their inventory six or eight times a year rather than twice, while products and styles proliferated in a “dizzying diversity in fabric, design, and style” (Abernathy et al. 1999:9; Collins 2003).

The competition among retailers and branded marketers generated growing consolidation among firms through mergers and acquisitions and thus growing power to pass on these competitive pressures to weaker actors upstream. Indeed, Richard Appelbaum and Gary Gereffi reported that “the large merchandising firms can afford to squeeze hard … [turning] (p.99) up the pressure on their contractors to make clothes with more fashion seasons, faster turnaround times, lower profit margins, greater uncertainty about future orders and frequently worse conditions for workers” (1994:52, 60). These pressures were in turn passed on from apparel manufacturers to textile manufacturers in the form of more exacting quality specifications, demand for more sophisticated textiles, and, of course, lower price points (Gibbon and Ponte 2005:179; Larsen 2003:11). For textile manufacturers around the globe, these demands turned their attention toward the quality of their raw materials for the production of both high- and low-quality yarns and textiles. Careful attention to cotton quality was critical in order to meet higher quality specifications and to create new types of yarn and fabrics. As Colwick, Lalor, and Wilkes explain:

The increasing consumer demand for quality and performance translates into new demands on cotton as a fiber. Fine yarns are needed for sheer fabrics. Strong yarn is needed for easy-care fabrics because easy-care finishes weaken fabrics (like wrinkle-free). Mature, nep-free fiber is needed to manufacture velours, corduroys, and knitted fabrics that are to be dyed in shades where fabric quality depends on uniform dye uptake. (1984:392)

However, cotton quality was also important for products such as blue jeans, which could be made of fairly low quality cotton. Raw materials—cotton and other fibers—were the most important factor influencing yarn quality and represented about 50 percent of the cost of yarn (Estur 2004b). With narrowing margins, textile manufacturers sought to purchase the lowest quality of cotton possible to make a given end-product. Under these conditions, textile manufacturers around the world became more interested in the savings that could be gained by scientific management of processing based on mechanized cotton fiber measurements. This became a concern not only for textile manufacturers but also for major cotton exporters wanting to ensure their competitive position vis-à-vis US exporters.

As demand for mechanized classification grew, the US classification system and benchmark standards gained status as the de facto global standards by the 1990s. Most other cotton-producing countries still used manual classification, and no other comprehensive mechanized classification system existed. US cotton producers thus enjoyed a first-movers advantage. Some textile manufacturers in Europe and East Asia began to adopt mechanized systems and to license the EFS software by the late 1980s. To facilitate this adoption as a way to create markets for US cotton, the US producer-funded Cotton Incorporated began to license their EFS software to textile manufacturers (p.100) abroad, focusing particularly on Asian textile manufacturers. In 1989, Cotton Incorporated opened an office in Singapore to offer technical support to textile manufacturers in Thailand, Indonesia, the Philippines, and Malaysia who adopted the HVI system and licensed the EFS software (Jacobson and Smith 2001:296).

With an EFS license, textile manufacturers received two service visits a year from Cotton Incorporated and could attend seminars on how to use the EFS system to achieve “lean” manufacturing. In return, the textile manufacturer had to commit to increasing the percentage of US cotton that they used (Robinson 2006). Indeed, Cotton Incorporated saw the EFS system as a way to encourage foreign textile manufacturers to choose US cotton over both synthetics and cotton from other producing countries.

In short, as major retailers began imposing new quality demands on their suppliers around the globe, demand for more precise measurements of cotton fibers gave the United States' mechanized classification system status as the de facto global standard. As such, the new mechanized classification system helped to reinforce the position of US cotton producers as dominant exporters vis-à-vis their competitors both in synthetic fibers and in other cotton-producing countries—a position that was ultimately undergirded by the United States' ongoing agricultural subsidy regime. As we will see in the following chapter, this would become a source of international rancor and would spark both a protection strategy and a redirection strategy against the de facto global status of the US classification system and standards and the broader liberal market project in which they were embedded.

Shifting Quality Classifications in Comparative Context

The growing demand for more exacting quality differentiations in the cotton industry reflects two broader dynamics common to a range of other commodities in this period. The first is the dynamic interplay between the competition among capitalist firms to achieve ever-greater economies of scale and/or speed and the material characteristics of raw materials for production. Building on Bunker's earlier work (1984, 1985), Bunker and Ciccan-tell (2005) identify this interplay in relation to extractive industries. They argue that, under the competitive demands of capitalism, firms and states create technological innovations to achieve greater economies of scale and speed. However, as economies of scale and speed increase, so do the quantities and precise qualities of natural resources required by the system.

(p.101) For example, at the end of the twentieth century, Japanese engineers designed and built supertankers and enormous dry-bulk carriers that ultimately gave them a significant advantage over the United States in the competition to dominate world trade. To realize these technological innovations, however, they created new raw material quality demands; that is, they required steel of much higher tensile strength and flexion than currently existed and thus had to secure access to huge deposits of iron and coal of precisely specified purity, grade, carbon and sulfur content, hardness, and moisture content (Bunker and Ciccantell 2005:10–11). In short, technological change emerging out of the competitive dynamics of capitalism generates increasing demands on raw material quality and the information available about it. Busch et al. (2006) note parallel dynamics in the soybean industry, as the adoption of more capital-intensive technology to extract oil from soybeans created demand for new information about soybean quality. In a similar way, the competition to reduce labor costs and achieve greater economies of speed in the textile industry created new demands for stronger fibers that could withstand faster processing speeds, thus generating demand for new ways to evaluate fiber quality.

In addition to the interplay between competitive dynamics and material constraints, the greater importance of quality in the cotton trade in recent decades reflects a broader “turn to quality” in the economy more generally. In recent decades we have witnessed a shift from an economy of quantities, or an economy in which price competition is the major axis of struggle, to an economy of qualities, or an economy in which competition unfolds increasingly in relation to nonprice considerations (Callon et al. 2002). Just as the apparel industry shifted to product differentiation and intensified fashion cycles in the 1980s and 1990s, with implications for the cotton trade, so too have a range of other sectors come to place greater attention on quality differentiation as consumer demands shift and/or firms differentiate products to generate greater consumer demand. This shift has been well documented in food and agricultural commodities as competition in products as diverse as fresh fruit and vegetables, coffee, soybeans, and wheat have increasingly come to turn on quality attributes (see Daviron and Ponte 2005; Dolan and Humphrey 2000; Friedberg 2004; Magnan 2011). Hatanaka, Bain, and Busch argue that “quality is becoming a central component of economic competition in the global agrifood system…. Retailers are competing on other factors besides price, such as quality, convenience, and production practices” (2006:42).

(p.102) Transnational Merchants: Expanded Networks, de Facto Dispute Settlement Authority

While the USDA claimed de facto global authority over the definition of quality and the creation of benchmark standards from the 1970s to the 1990s, Western transnational merchants began to extend onto a global scale their private authority over the third critical quality governance task: dispute settlement. In the 1970s and 1980s, Western merchants began to expand their sourcing and distribution networks abroad, particularly as textile manufacturing shifted increasingly toward Asia. Many smaller and mid-sized merchants and producer cooperatives, operating around the world with domestic or regional foci, continued to prosper by trading domestically. However, at the same time, a handful of large merchants, some who had moved in and out of the cotton trade for over 100 years and some new to the cotton trade, began to again expand operations globally. Taking advantage of the shifting geography of textile production, these merchants competed through mergers and acquisitions to create broad-based sourcing and distribution networks worldwide.

These included European and US merchants with a history of transnational trade, US-based merchants looking to expand outside their domestic market, and highly capitalized grain merchants diversifying into new commodities. It was the merging of the strengths of these three types of firms that resulted in significant consolidation in cotton trading. For example, Cargill, a US-based grain merchant, acquired the Memphis-based cotton merchant Hohenberg Brothers in 1975. Cargill was attempting to diversify into other commodities within the US market, and Hohenberg Bros, was one of the largest domestic cotton merchants. In 1981, Cargill further acquired the Liverpool-based Ralli Brothers and Coney. This British cotton merchant had a long history of transnational cotton trading, allowing Cargill to transnationalize their newly acquired cotton business (Cargill 2005b). Giant multicommodity merchants like Cargill gained critical advantages over the older firms concentrated on cotton alone. They could shift funds from one commodity to another in response to price changes and profit opportunities, and it was generally only these firms that had the capital and expertise to speculate in the commodity futures markets (see also Talbot 2004).

By the late 1990s, trading cotton came to be characterized by increasingly high barriers to entry. Building global sourcing and distribution networks (p.103) required very high levels of working capital and access to credit, accumulated market knowledge (in all countries involved as well as in trade-related services like transport, insurance, and financial services), the ability to manage risk through diversification and/or price risk management, and intangibles such as reputation (Gibbon 2001; Larsen 2003). This went hand in hand with the increasing role of financial capital in the economy in this period.

Merchants assumed a critical coordinating role, procuring specific volumes and quality mixes from cotton-producing countries around the world in order to offer a wide selection of alternative cottons to geographically dispersed textile manufacturers. With the growing adoption of mechanized classification of cotton, the largest merchants began offering cotton to meet more exacting specifications within narrow tolerances (Ruh 2005:11). To be competitive, merchants thus had to minimize inventories while maintaining a high level of responsiveness to the precise demands of manufacturers (Gibbon 2001). As well, reliable delivery times became increasingly important. Merchants began to offer delivery in regular monthly, weekly, or even daily shipments as textile manufacturers began to operate on a just-in-time delivery basis (Çalis¸kan 2010:67; Ruh 2005:11). Finally, merchants increasingly offered brokerage services for buyers, including futures and options. Although some regional merchants and large-scale producer cooperatives (such as those in the United States) forged their own marketing relationships, most cotton was sold through the large transnational firms given the costs of performing these functions (Gibbon 2001).

Some scholars have seen this as the emergence of a “trader-driven” commodity chain in the cotton trade, given the “huge disparities in resources (financial and informational)” that emerged between merchants and both their suppliers and clients (Gibbon 2001:352). During this period, trader-driven chains began to emerge for a range of commodities, such as coffee, cocoa, and cashews. Many of these commodities were characterized not only by the growing power of transnational merchants but also by growing consolidation among downstream buyers (e.g., coffee roasters, chocolate manufacturers, and apparel retailers) (Gibbon 2001). In this respect, many trader-driven chains fed into buyer-driven commodity chains, which were becoming increasingly common in this period (Gereffi 1994).

As a new cadre of transnational merchants gained dominance in this period, they attempted to extend their rules for private dispute settlement on an increasingly global scale. Since its beginnings in the nineteenth century, the Liverpool Cotton Association (LCA) had continued to operate as a primary (p.104) trade association for European-based merchants, albeit with disruptions during the World Wars. For their part, US merchants had established their own national association, the American Cotton Shippers Association (ACSA), in 1924 (ACSA 2011). By the 1970s, however, as the transnational trade in cotton began to expand significantly, these two trade associations, which counted in their membership most transnational merchants and whose memberships were increasingly linked due to mergers and acquisitions, decided to join forces and harmonize their rules for cross-border trade. Given that ACSA's rules had become more domestically focused in the postwar period, members of ACSA agreed to use the LCA's rules and arbitral body for dispute settlement in the transnational trade.

Transnational merchants' ability to extend the private authority of the LCA to settle disputes was facilitated by, and took particular form in relation to, the United States' uneven market-led development project. In their relations with textile manufacturers, transnational merchants were empowered to largely impose their private authority over dispute settlement through the combined effects of the MFA and the structural adjustment programs of the IMF and World Bank. The MFA had the indirect effect of creating a geographically diffuse cotton trade. This created a new opening for transnational merchants who could supply cotton to these geographically diffuse and relatively small-scale buyers. Structural adjustment programs (SAPs), pushed by the IMF and World Bank, further reduced the bargaining power of textile manufacturers. SAPs were developed as a policy discourse by World Bank President Robert McNamara and cultivated by Western elites, “in keeping with the ascendant Reagan revolution” (Babb 2005:200). First introduced as a response to the debt crisis facing many countries in the 1980s, SAPs cast policies of liberalization and privatization as the path to “development” for countries in the global South, undergirded by the coercion of conditional lending (Babb 2005; McMichael 2004). A key target of these policies was state trading enterprises (STEs) that had bargained collectively on behalf of textile manufacturers (and/or cotton producers) to secure a better price vis-à-vis transnational merchants. Privatization thus gave transnational merchants greater leverage to impose their private authority as arbiters of quality disputes, as well as contract disputes more broadly.

The Indian case is illustrative here. In the nationally focused development model adopted by India in the post-independence period, an STE, the Cotton Corporation of India, held a monopoly over cotton imports and exports. In 1991, however, the Indian government ended this monopoly as (p.105) part of a broader adoption of liberalization and deregulation policies, and private merchants and spinners were allowed to import and export cotton (Landes et al. 2005). Given the historical development of the sector, Indian spinners, compared with those in other major textile-producing countries, were relatively small and not vertically integrated—that is, spinning, weaving, finishing, and apparel manufacturing processes were often done by separate firms (Landes et al. 2005; Tewari 2008). Without the STE to purchase cotton on their behalf, these small Indian yarn manufacturers and import firms, inexperienced in transnational transactions, purchased cotton directly from transnational merchants. A representative of the Indian Cotton Mills Federation explained the implications:

In cotton, it is often the seller who decides the conditions of contracts and in most transactions his writ ultimately runs. The fact that cotton exporters are few in number and huge in size, whereas importers are often thousands of small spinners only further compounds this position. (Jaipuria 2005)

Many of the firms importing cotton were small and had little bargaining power vis-à-vis merchants who insisted on using LCA contract rules and arbitration to settle disputes (Business Line 2004; Gurumurthy 2001). Indeed, it was reported that LCA contract and arbitration rules were used for all cotton imported to India by the early 2000s (Economic Times 2003). As these small players had difficulty challenging transnational merchants' position as de facto transnational rule-makers, these simmering tensions would eventually develop into a protection strategy.

In the 1980s and 1990s, transnational merchants' position as powerful middlemen was further strengthened by the implementation of structural adjustment programs (SAPs) in cotton-producing countries. First, SAPs encouraged countries to adopt an export-oriented model of agriculture. The West African bloc of cotton-producing countries, for example, claimed a position as a significant exporter from the 1980s to the 2000s by focusing on developing their export-oriented cotton sectors. Indeed, some analysts consider cotton “a rare economic success story in sub-Saharan Africa” (Tshirley et al. 2008:123). While sub-Saharan Africa's share of world agricultural trade fell by half from 1980 to 2000, cotton exports from the West African bloc increased from about 4 percent to 15 percent of total world exports over the same period (Larsen 2003:17; Tshirley et al. 2008:123). Their success, however, is largely attributed not to the export-oriented development model per se but rather to the comprehensive and efficient provision of inputs, credit, support services, infrastructure, and region-specific research established (p.106) by the French state trading company that maintained a monopoly over cotton marketing in most of Francophone Africa until the late 1990s (Larsen 2003:13).

Beyond the promotion of export-oriented agriculture, in many countries structural adjustment meant the liberalization of the cotton market and the privatization of STEs that had organized, marketed, and governed national cotton sectors and their links to the global market. For example, throughout the 1990s and early 2000s, countries in Eastern and Southern Anglophone Africa—Tanzania, Uganda, Zambia, and Zimbabwe—privatized STEs and cooperatives, leaving private companies to dominate input supply, ginning, and marketing (Baffes 2004; Tshirley et al. 2008).

In the Francophone West African countries, privatization and liberalization moved more slowly as both the STEs themselves and the French state's cotton trading company, the CFDT (La Compagnie Française pour le Développement des Fibres Textiles, or the French Company for the Development of Textile Fibers), resisted these policies. The roots of this resistance lie in the nature of the West African cotton sectors after independence (see Box 3.1). While colonial ties ended upon independence, neocolonial economic ties were forged as the CFDT negotiated with new state trading enterprises (STEs) in West African countries to establish a trade monopsony in the region. As the IMF and World Bank pressured West African states to privatize their STEs as part of structural adjustment programs, both the STEs and the CFDT opposed these policies.

For the CFDT, privatization would eliminate their monopsony in the region (Bingen 1998, 2006). Given France's important role as a bilateral donor to the region, the CFDT's resistance did successfully delay privatization for a time (Baffes 2007; Bingen 1998). Cotton producers had a different view of the situation. As Bassett (2001) demonstrates in Côte d'lvoire, cotton growers' incomes declined from the 1970s to the 1990s as profits were accruing not to them but to input suppliers, cotton merchants (CFDT), and the state marketing board. Cotton producers saw the CFDT system as unaccountable. They worked to build up their producer organizations. This allowed them to stage strikes and protests to call for greater decision-making power and a greater share of the export price of cotton (Bassett 2001; Bingen 1998).

For their part, West African states saw the privatization agenda reflecting the broader hypocritical nature of the US-led project of market liberalism. The World Bank argued that the “managed monopoly” system in place in West African countries was vulnerable in a highly competitive global market, claiming it limited incentives to minimize costs and stifled entrepreneurial (p.107) decision-making. From this view, the managed monopoly model was stymieing economic growth and depressing cotton producers' incomes (Bingen 2006:223). From the perspective of the African STEs, however, it was US cotton subsidies, not their managed monopolies, which were depressing cotton producers' incomes.

Moreover, West African states argued that the key to their global competitiveness was not firm structure—public vs. private—but rather the maintenance of an integrated model to ensure cotton quality, which yielded price premiums on the international market. In West African states, the cotton sectors were organized in a so-called filière intégr ée, in which a single STE managed all of the steps in the production process, from the provision of seeds and other inputs to the purchase of the crop at the end of the growing season, the separation of the cotton lint from the seeds at ginneries, and the marketing of exports (Bassett 2001; Siaens and Wodon 2008). For West African states, privatization risked the deterioration of this quality control (p.108) system and thus of the reputation for high quality that their cotton producers enjoyed on the global market.

Indeed, the experience of Southern and East African countries in the wake of privatization reinforced these concerns. Liberalization in many countries did provide farmers with prompter payment for their cotton and a higher share of the final market price for lint than they had previously received (Baffes 2004; Tschirley et al. 2008). However, in a number of countries, evaluations of the privatization and liberalization of cotton sectors suggest that quality control systems broke down and cotton quality deteriorated. Farmers no longer had the infrastructural support or incentives to produce high-quality cotton (Gibbon and Ponte 2005:182). In Uganda, Tanzania, and Zimbabwe, for example, privatization and liberalization brought an influx of relatively small, private ginner-merchants attempting to gain a slice of this new market. As a result, intense competition emerged among newly privatized gins to purchase seed cotton, which destabilized input provision and quality (Tschirley et al. 2008:134–5). In Tanzania and Zimbabwe, market liberalization brought intensified competition among small-scale ginner-merchants and the elimination of grading at the first point of sale, thus eliminating quality premiums for farmers (Gibbon and Ponte 2005:147; Tshirley et al. 2008:140). In Tanzania, private ginners purchased cotton across the country, mixing local seed stocks. Few private firms provided inputs. Those who did offered them on a cash-only basis (Gibbon and Ponte 2005:147). These disruptions to quality control and input distribution mirrored the consequences of privatization and liberalization in other sectors such as the coffee industry (Gibbon and Ponte 2005:136).

Quality control was best maintained in countries like Zambia, which reproduced a single-channel system that mirrored the operation of the state trading enterprises but in private form. In these systems, privatization took the form of geographical monopsonies. Private firms were given the sole right to purchase cotton from farmers within a given region and were obligated to provide inputs and technical assistance on credit to farmers. With their monopsony control, private firms have had an interest in increasing production, as well as in defending national reputations for quality by maintaining grades as incentives for farmers (Gibbon and Ponte 2005:146–7). These systems, however, amounted largely to handing over public companies to private, and often transnational, firms (Larsen 2002, 2003; Poulton et al. 2004). In Zambia, the state trading enterprise, LINTCO, was sold in two parts, one to the private firm, Lonrho Cotton, and the other to Clark Cotton, a South African firm (Tschirley et al. 2008). The Zambian sector (p.109) soon became a staging ground for two of the largest transnational merchants' competition in acquisitions. Dunavant purchased Lonrho Cotton while Cargill purchased Clark Cotton (Larsen 2003). In cases like this, quality control remained, but the responsibilities for grading cotton and determining the distribution of profits were turned entirely over to private firms with little public oversight. Small cotton producers effectively sold directly to a vertically integrated transnational conglomerate that had a monopoly over ginning in their geographical zone.

Despite the concerns West African states held regarding the experiences of East and Southern African countries, reforms in West Africa proceeded under continued pressure from the IMF and World Bank. Mali was the only country in which ginning and export marketing continued to be managed by a single STE. In Côte d'Ivoire, Benin, and Burkina Faso, STEs still existed, but their monopsonistic positions were broken up by privatizing some gins and/or allowing private companies to enter the market. The initial years of these reforms brought instability to input provision and quality control systems (e.g., in Benin, see Siaens and Wodon 2008:161). Quality control fared best where public regulatory structures were left in place (Gibbon and Ponte 2005:182).

For transnational merchants, privatization of STEs thus opened up access to new cotton supplies through the different ways of “[substituting] private for government trading organizations” (Bassett 2008:53). Between 1994 and 2009, the number of state trading enterprises involved in selling significant quantities of cotton (between 50,000 and 200,000 tons annually) fell from fifteen to just three, due mostly to privatization in African countries (ICAC 2010b:2). This allowed the largest transnational merchants to expand their operations significantly during the 1990s in terms of the number of countries from which they purchased cotton (Larsen 2003:9). In some cases, as we have seen, this has meant vertical integration into ginning. Cargill, for example, came to own cotton gins in five countries in southern Africa: Malawi, Tanzania, South Africa, Zambia, and Zimbabwe. Moreover, firms have also attempted to develop long-term relationships with reliable local merchants/ginners, which involves prefinancing seasonal purchases (Larsen 2003:9). While about 17 merchants were active in the West African cotton trade by the mid-2000s, just five handled 70 percent of production (Estur 2004a:11; see also Fok 2005).

Privatization thus enhanced transnational merchants' power in the cotton trade and ability to impose contract terms and rules for dispute settlement. As a range of small, private ginner-merchants took the place of (p.110) state-level collective bargaining, transnational merchants purchasing cotton from these smaller firms enjoyed formidable bargaining power to impose contract terms and dispute settlement mechanisms. As one analyst suggests, in their relations with private ginners in African countries, transnational merchants were often capable of “taking advantage of the poor know-how of ginning companies regarding contractual disputes” (Estur 2004a: 22).

In short, the 1980s to the 2000s was a period in which transnational merchants were gaining increasing power vis-à-vis both their clients and suppliers and began to use this powerful position to impose their private authority to govern contracts and settle disputes. Yet, it was also a period in which tensions were building in opposition to their position as de facto transnational rule-makers within the US-led neoliberal project.

The Limits of Private Authority

The private authority of transnational merchants over dispute settlement could not be achieved simply through growing economic dominance and coercion. Rather, while transnational merchants were increasingly able to impose their authority, they also faced limits to this strategy. These limits were rooted in part in intensifying price volatility from the 1970s to the 2000s and in the privatization and commoditization of risk that accompanied the privatization of state trading enterprises.

While declining prices and price volatility had long been a concern of countries in the global South, the end of US buffer stocks and the oil shocks of the 1970s significantly increased price volatility in cotton and other primary commodity markets. Cotton prices were relatively stable in the 1950s and 1960s. Through its support programs for domestic producers, the US government effectively operated a world buffer stock that kept the international price of cotton high and stable (Finlayson and Zacher 1988). In the late 1960s, however, the United States sold off its cotton reserves, destabilizing the cotton market and making the price of cotton more volatile. This, paired with the successful oil cartel of OPEC (Organization of Petroleum Exporting Countries), led to significant fluctuations in commodity prices for much of the decade. Cotton prices reached significant highs in 1973, 1976, 1980, and 1983, from which producers benefited (Finlayson and Zacher 1988). But these price highs were preceded and followed by significant lows. For example, international cotton prices jumped from 30 cents a pound in the spring of 1973 to 90 cents that November (Busch 1982:425). Numerous efforts were made to establish an interstate agreement to stabilize cotton prices but to no avail (see Box 3.2). (p.111)

The trend toward falling prices and price volatility persisted into the 1980s and 1990s. The real price of cotton continued to decline, from more than $3 per pound of lint in the early 1950s to $1–$2/lb. in the 1970s to just $0.50 –$0.60/lb. in 2004 (all in 2004 dollars) (Townsend 2004). At the same time, price volatility continued in the 1980s and 1990s. Baffes (2004) argues that, if volatility is measured as variability from one year to another, price volatility was the highest from 1973 to 1984 during the oil crises and commodity price boom. However, volatility was 2.5 times higher during the period from 1985 to 2002 compared with the period from 1960 to 1972 (Baffes 2004). Part of this volatility came from changes in supply and demand due to weather, producers shifting to other crops, the price of synthetics, and shifting government policies (Estur 2004a; Finlayson and Zacher 1988). However, two additional factors were of critical importance. In the 1980s and 1990s, China increasingly opened up to the international market. Information on China's cotton production, consumption, stocks, and policy agenda was difficult to obtain. As a result, China's net trade (p.112) with the rest of the world was a key factor in year-to-year price fluctuations (Estur 2004a).

Price volatility also increased alongside the growth in speculation in cotton markets. Until the 1980s, the major participants in cotton futures trading were merchants. Merchants used futures to hedge or protect themselves against sudden price changes. In the 1980s, however, there was a general proliferation of financial instruments and derivatives. For cotton, this meant the introduction of options trading on cotton futures contracts on the New York Cotton Exchange in 1984. Options contracts are cheaper than futures contracts and thus available to smaller merchants and spinners in the cotton trade. In addition, options contracts were available to smaller speculators, who were investing in financial markets but could not afford a diversified portfolio of commodity futures. By the early 2000s, about 40 percent of the average daily trading on the New York cotton futures exchange was speculative activity—that is, not connected in any way to sales of physical cotton (Estur 2004a).

As more speculators entered the market, this also brought a shift in the type of speculative activities. Traditionally, speculators chose to buy or sell cotton futures on the basis of fundamental analysis, or the actual projections of future supply and demand for cotton. Some cotton merchants participated in this type of speculation. In the 1980s and 1990s, financial speculation based on technical analysis became more common. Using technical analysis, speculators chose to buy or sell cotton futures and options not on the basis of actual supply and demand forecasts but rather on the basis of past market movements. This meant that money was being shifted in and out of cotton futures markets with little relation to the actual changing conditions “on the ground” in cotton-producing and -consuming countries (see also Talbot 2004: chapter 5).

Intensifying price volatility did not pose a problem to transnational merchants' operations per se. To the contrary, it was the management of price volatility that gave transnational merchants their position as market-makers—that is, those actors willing to absorb risks to allow exchange to occur. Transnational merchants managed price risk through tools like futures markets. However, price volatility did affect their relationships with their clients. Spinners and cotton producers in the global South, as well as cotton producers in the North and even many smaller and medium-sized merchants, were much less able to manage price risk. Hedging instruments did not (and still do not in many cases) exist for many cottons outside the United States (Baffes 2001). Moreover, the lines of credit necessary to suecessfully (p.113) and competitively hedge were increasingly beyond the scope of all but the largest transnational merchants, as the executive director of the ICAC explained:

For traders and those who have the financial resources to weather price fluctuations, meet margin calls and take advantage of volatility, there is opportunity for profit. But there are very few firms that are large enough, savvy enough, and have the expertise and financial resources to take advantage of the profit possibilities that come from volatility. There are some who can, but for the vast majority of market participants, it's a negative…. For those without a virtually unlimited line of credit behind their hedge position, they simply can't do it. (cited in Cotton International 2010c)

If actors could not hedge or otherwise manage their price risks, they would be much more likely to default on contracts in the case of significant price fluctuations.

This was particularly the case given the growing privatization and commodification of risk through structural adjustment programs. A key role that state trading enterprises had played was the socialization of price risk. Through their monopolies on cotton buying and selling, STEs could average out the highs and lows of price fluctuations to help cotton producers secure a stable income and to help textile manufacturers avoid radical shifts in the prices of their raw materials. With privatization, the management of price risk was privatized and individualized. Moreover, price risk management was commodified in the sense that it was no longer a public service but a form of private insurance that should be bought on the futures market. As many textile manufacturers in particular lacked the access, expertise, and credit availability to manage their risk on future markets, the decision to simply default on a contract when the price moved against them was an alternative risk management strategy.5 One transnational merchant explained to me the rule of thumb that their firm used: if the international price moved more than 10 percent below the contract price, they expected the chance of default to increase by 50 percent. By the early 2000s, Estur (2004a) noted that it was no longer unusual for monthly prices to vary from the annual average price by a margin of 15 percent.

If textile manufacturers in the United States defaulted on their contracts as a way to manage price risk, this would not pose a significant problem for merchants as they could arbitrate the dispute in their private arbitral body, which would be upheld in US state courts, forcing the defaulting party to pay damages. On the global stage, however, and particularly in countries (p.114) in the global South, the growing risk of contract defaults posed by rising price volatility posed more complicated problems. Even if transnational merchants had the economic power to impose their contract terms and dispute settlement authority on, for example, textile manufacturers in India, they could still face significant losses if the textile manufacturer defaulted on the contract. This was due to the fact that merchants' private authority was contingent on the support of states and a transformation in states' organizing logics to serve transnational firms rather than domestic producers or spinners.

Private arbitral authority was supported by Western governments through the 1958 New York Convention (see chapter 2). Many countries in the global South, however, had resisted adoption of this Convention as they felt that private transnational arbitration prioritized the interests and customs of transnational firms and overrode national laws, such as those designed to protect domestic sectors and facilitate national economic development (Sempasa 1992; Sornarajah 1989). Latin American countries, for example, adhered for many years to the Calvo Doctrine, named after an Argentine jurist, which held that disputes arising from international contracts, especially foreign investment contracts, should be settled only by domestic courts and tribunals applying domestic law. Developing countries further upheld this doctrine through resolutions in the New International Economic Order (NIEO) (Sempasa 1992; Sornarajah 1989). Many countries were finally compelled to sign on to the 1958 Convention as a condition for IMF and World Bank loans and/or as part of a regulatory competition for mobile capital (Robé 1997). Yet, this still did not ensure the enforceability of private arbitral authority. Despite being signatories to the Convention, some state courts would reopen cases to apply domestic laws and procedures when an arbitral decision was brought to them to be enforced.

Facing this context, merchants were forced to litigate disputes in state courts to get contracts enforced, which was a lengthy and costly process. The costs of litigating a contract default could be as high as 60 percent or 70 percent of the value of the contract one was trying to recover (Busch 1982:42). Thus, it was not feasible to take every textile manufacturer who defaulted on his/her contract to court in his/her respective country. Moreover, problems emerged even among merchants and spinners who accepted private arbitration and had well-developed private contract rules. One US firm reported losing arbitration, and subsequently litigation in court, to a German firm, due to the differences between their respective trade associations' rules (Busch 1982:435). That is, problems emerged due to the fact that (p.115) merchants, textile manufacturers, and other actors in the cotton trade were organized in separate and largely nationally-focused trade associations with nationally focused rules. As one commentator remarked, “There was soon no predominant culture in the cotton trade, and as trade became internationalized and variegated, so did the standards of doing business” (Busch 1982:437).

In this context, transnational merchants explored new possibilities for addressing the contract enforceability problems that ensued from price instability. In 1976, ACSA and the LCA established a new transnational organization: the Committee for International Cooperation between Cotton Associations (CICCA). This organization brought together private trade associations around the world with the express purpose of promoting contract sanctity in the transnational cotton trade (CICCA 2009; ICAC 1996). In addition, merchants tried to construct a new culture for the transnational trade. They aimed to reconstruct the kinds of thick personal relationships that allowed gossip networks and threats to reputation to serve as powerful disincentives to contract defaults (Bernstein 2001:31). To this end, the LCA began to supplement their yearly convention with social events to help build new relationships and invited their new clients to these events. In the United States, ACSA began to hold events such as the Sourcing Summit, in which they invited spinners from around the world to make connections with US merchants and tour US cotton production and ginning facilities. Such events worked not only to sell more cotton but also to create a “community” in the cotton trade.

Merchants considered other possibilities for minimizing contract default problems, but as competition to expand transnationally at this time was intense, they were unable to forge cooperation around proposed strategies. A member of one merchant association recalled that they had considered rating countries on risk factors and hiring credit investigators and asset trackers to help enforce arbitral awards. However, some of the largest firms were reluctant to put up the money to protect firms that took on too much risk. As one merchant representative stated to me: “Why should I pay for an army of lawyers so someone can go and do high-risk business and then I'd have to pay to bail him out?”

In general, deeper cooperation to stabilize transnational contract governance was overshadowed by new opportunities to maneuver in a transnational market without clearly defined rules. Expanding one's operations into new markets while avoiding significant losses due to contract defaults became a new axis of competition for merchants. One transnational merchant (p.116) explained to me that one of his competitors had been selling into the Indian market when prices fluctuated dramatically. This competitor had not practiced due diligence to ensure the credit-worthiness of his buyers. As a result, when prices fluctuated significantly, the competitor faced a rash of contract defaults and spent years trying to settle them in Indian courts. “Knowing Your Buyer” represented a key risk management strategy, and merchants competed to establish relationships with reputable buyers. In these ways, merchants chose to compete over who could manage this risk in the plays of the game, rather than to negotiate new transnational rules of the game. As one observer commented, it threatened to be “a situation akin to anarchy—a game whose participants must be as adept at changing rules as at classing cotton” (Busch 1982:437).

Private Authority Over Dispute Settlement in Comparative Context

The growing power of transnational cotton merchants and their efforts to extend their dispute settlement authority globally from the 1970s to the 1990s is part and parcel of the broader rise of private authority in the world economy. Indeed, for many, the rise of private authority is seen as a defining feature of the current era (Cutler et al. 1999; Gereffi 1994; Hall and Biersteker 2002; Haufler 2001; Kahler and Lake 2003; Büthe and Mattli 2011). The rise of private arbitral authority to settle disputes for transnational commercial transactions in particular is a trend that extends far beyond cotton and other agricultural commodities. Over 90 percent of cross-border contracts for the trade of goods as diverse as electronics, oil, or solar panels contain a private arbitration clause (Berger 2010). While bulk commodity sectors like cotton have long been the mainstay of private arbitration, since the 1970s the number of more general arbitration forums has grown from around a dozen to more than one hundred in the 1990s, and the caseload of major arbitral institutions, such as the International Chamber of Commerce, have more than doubled in the same period (Mattli 2001a:920).


Contrary to macrolevel approaches that tend to overstate the coherence of projects of market liberalism led by powerful coalitions of states and firms, the transformation of quality governance in the cotton trade from the 1970s to the 2000s reflects a more contested and piecemeal process (see (p.117) also Krippner 2011). As their competitive positions in the capitalist world economy shifted, US textile manufacturers, cotton producers, the USDA, and transnational merchants strategized to reconstitute the rules governing the cotton trade. This was a highly contested and uneven process, given divergent interests within the United States itself regarding the trajectory of market liberalization.

It was out of both the new opportunities and the new competitive pressures created by the United States' trade liberalization project that a new system of quality classification and dispute settlement was forged. Responding to growing competition from East Asia, US textile manufacturers pushed for protectionism through the MFA and for technological upgrading support. The new open-end spinning technology they adopted, however, sacrificed quality for economies of speed. This, paired with cotton producers' competition with synthetic fibers, stimulated cotton producers, textile manufacturers, and the USDA to retool their quality classification system to increase the competitiveness of the cotton-textile sector as a whole. As the rise of giant retailers in the 1990s heightened competitive pressures across the industry, the USDA's mechanized quality classification system increasingly came to be seen as the de facto system for global adoption. In short, as institutionalists suggest, it was a process of strategizing and problem-solving that generated a new quality classification system. At the same time, however, we can better understand why the retooled classification system took the form that it did by situating actors within broader contests over the organization of the global capitalist system. Actors sought to solve the particular competitive problems they faced given their historically and spatially specific position within the shifting world economy.

For transnational merchants, this was a period of expansion as both the protectionist compromise of the MFA and the privatization agenda pursued through IMF and World Bank structural adjustment programs created opportunities to extend the global reach of their sourcing and distribution networks. As their bargaining power vis-à-vis relatively small and geographically dispersed trading partners strengthened, transnational merchants were more able to impose their preferred contract and arbitration terms. This rise of private authority, however, was not as unproblematic as much of the literature suggests. The authority of transnational merchants over dispute settlement did not depend simply on states retreating to allow private firms and associations to take over their responsibilities, such as through the privatization of state trading enterprises. State authority continued to underpin private authority. Particularly in a volatile market, private dispute (p.118) settlement continued to depend on state courts to enforce arbitral awards as if they were decisions of the courts themselves. We see that a key limit to transnational merchants' authority was the degree to which state courts around the world accepted this denationalized function as legitimate. To allow private authority over contract governance to function, there must be at once privatization of state functions and the denationalization of the organizing logic of state institutions. In the cotton sector at the end of the twentieth century, this involved shaping the role of state courts such that they no longer guarded public policy and public welfare concerns and instead facilitated transnational trade (see also Sassen 2006). Yet, many states resisted such transformations.

Ultimately, the status of the USDA and transnational merchants as de facto transnational rule-makers would not go unchallenged. Rather, from the 1970s to the 1990s, tensions began to build in opposition to these dominant actors, their rules for quality governance, and the United States' uneven neoliberal project as a whole. Much as in the conflict over standards at the turn of the twentieth century, the contemporary transnational rule-makers would face both a redirection strategy on the part of emerging rivals and a protection strategy waged by marginalized actors by the turn of the twenty-first century.


(1) . The AAA represented a policy of supply management that gave farmers guaranteed minimum prices for targeted commodities. If market prices dropped below the government's minimum price support levels, the Commodity Credit Corporation purchased the commodities from farmers at the price support level. This raised farm income and served to manage supply by pulling the commodities off the market. Farmers also had to adhere to the acreage allotments set through production controls or they would not receive price supports (Winders 2009).

(2) . Winders (2004, 2009) notes that, as cotton and wheat producers became more exportoriented, they supported changes in agricultural policies that maintained subsidies but reoriented them toward market expansion rather than supply management through production controls. However, supply management did not end completely until 1996. A version of price supports remains in place.

(3) . The MFA was preceded by the Short-Term Arrangement of 1961 and the Long-Term Arrangement of 1962, which had similar aims.

(4) . Color and trash (leaf grade) continued to be evaluated manually until 2000 when HVI color measurements became the official basis for the color grade. Leaf grade continues to be manually evaluated by a classer.

(5) . The problem of contract defaults due to price volatility first became a major issue in the 1970s. During this period, many cotton producers and textile manufacturers were compelled to break their contracts to buy or sell cotton in order to keep afloat (Busch 1982). Throughout the decade, producers in the United States, Mexico, Nicaragua, and Guatemala, and importers and textile manufacturers in Thailand, the Philippines, Korea, and Taiwan, began to default on their contracts (Busch 1982:429–35). Among merchants, this became known as the “contract crisis” of the 1970s. US textile manufacturers did not default in large numbers during the crisis in the 1970s. However, Bernstein (2001) reports that, even though much trading in the United States was based on nothing but a handshake for many years, merchants increasingly demanded that US textile manufacturers sign legally enforceable contracts in the 1980s and 1990s. With the decline of the US textile sector, many US manufacturers who were teetering on bankruptcy became more likely to default.

Copyright © 2021. All rights reserved.
Privacy Policy and Legal Notice