North American Free Trade and Changes in the Nativity of the Garment Industry Workforce in the United States*

Transcription

1 International Journal of Urban and Regional Research Volume 25.2 June 2001 North American Free Trade and Changes in the Nativity of the Garment Industry Workforce in the United States* DAVID SPENER AND RANDY CAPPS Globalization of social and economic life at the end of the twentieth century has been characterized by the rise of export-oriented manufacturing in Asia and Latin America coupled with the limited deindustrialization of the United States and western Europe, massive international labor migration from the former regions to the latter, and the negotiation of free trade agreements to facilitate cross-border flows of commodities and capital. The relationships among these phenomena are particularly interesting with regard to the garment industry in the United States and Mexico in the aftermath of the implementation of the North American Free Trade Agreement (NAFTA). Garment manufacturing is a quintessentially global industry. On the one hand, worldwide the industry is dominated by third-world production for first-world consumption (Murray, 1995). On the other hand, the most important garment districts in the United States feature workforces that are predominantly immigrants from Asia and Latin America. These immigrants willingness to apply their considerable skill and effort to the sewing of garments for below-average wages has helped keep important segments of US garment production competitive with Asian and Latin American imports. The implementation of NAFTA has accelerated the decades-long decline in garment production and employment in the United States by increasing the competitiveness of Mexican-sewn imports and fueling the expansion of maquiladora production in Mexico. In this article, we address the question of the extent to which US producers reliance on cheap immigrant labor can continue to retard the march of apparel manufacturing out of the country as garments produced by even cheaper labor overseas flood the US market in the post-nafta period. The article is divided into five sections. In the first section, we introduce concepts that are key to our discussion, including the new international division of labor thesis, dual labor market theory, the state s role in boundary management and the implications of these with regard to industrial development and migration in Mexico and * Partial support for this research was provided by the Tom and Mary Turner Faculty Fellowship of Trinity University, by grants from the Ford Foundation and the National Institute of Child Health and Human Development (P-30 Grant #5 P30 HD06160) to the University of Texas at Austin, the Center for the Study of Western Hemispheric Trade (at the University of Texas at Austin), and by a grant from the Hewlett Foundation to the University of California, Irvine. We would also like to acknowledge the technical support of Starling Pullum and other computer-services staff at the University of Texas at Austin Population Research Center and the University of California, Irvine. In addition, we owe thanks to Roland Chanove, Marilyn Espitzia, Kelly Fenton, Betsy Guzman and Allison Newby for their assistance in conducting our survey of garment workers in El Paso. Finally, Jennifer Bair and three anonymous reviewers assigned by IJURR made helpful comments on an earlier draft of this article.. Published by Blackwell Publishers, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA.

2 302 David Spener and Randy Capps the United States. In the next section, we examine changes in the regulatory regime governing international trade in garments and the subsequent shifts that have occurred in Mexican apparel exports to the United States on the one hand, and in Mexican and US garment employment on the other. In the third section, we review the role immigrants play in the US garment industry and the debates about if and how immigrant workers and entrepreneurs contribute to its international competitiveness. We then turn our attention to a case study of garment production in El Paso, Texas, where thousands of Mexicanimmigrant and Mexican-American women have lost their jobs as seamstresses since the implementation of NAFTA began in In the fourth section of the article, we analyze data from US County Business Patterns, the decennial US Census of Population and Housing, the annual March US Current Population Surveys, and the US Department of Labor s records of certified NAFTA-related layoffs to ascertain the extent to which El Paso s experience of heavy immigrant garment job losses is typical of the rest of the country. In the concluding section we discuss the implications of our analysis for the future of the US garment workforce. The new international division of labor, immigration and free trade: implications for the US-Mexico relationship The internationalization of manufacturing industry is a central feature of the globalization of society at the end of the twentieth century. During the last 35 years, lightmanufacturing export zones in low-wage newly industrializing countries (NICs) have come to produce a substantial portion of world output in industries such as electronics and apparel (Fernández-Kelly, 1983; Portes and Walton, 1981; Sassen, 1988; Sklair, 1989; Gereffi and Wyman, 1990). A powerful, though much-contested concept in the globalization literature concerning industrial development has been the new international division of labor (NIDL) thesis, originally advanced by Frobel et al. (1980). 1 This thesis contends that capitalist industrialists, in their drive to maximize profit, move laborintensive manufacturing operations from high-wage regions at the core of the world economy to low-wage regions on the periphery. In NIDL s starkest formulations, this movement constitutes a lose-lose transaction for the international working class, devastating core-region communities as workers lose their jobs (Bluestone and Harrison, 1982; Harrison and Bluestone, 1988), while in peripheral regions workers suffer exploitation in the form of low wages, hazardous working conditions and various forms of harassment, persecution and repression (see, for example, the Lorraine Gray documentary film entitled Global assembly line). Complementary to the NIDL hypothesis, dual labor market theory posits that manufacturers who continue to conduct labor-intensive operations in core countries are obliged by market competition to exploit vulnerable segments of the workforce, typically women, immigrants and members of racial and ethnic minority groups. Employers expect these more vulnerable workers to work for lower wages than white, male, native-born workers (Piore, 1979; Portes and Walton, 1981; Sassen, 1988; Fernández-Kelly and García, 1989; Portes, 1991; Petras, 1992). Portes and Walton (1981) argue that manufacturing firms employing such a low-wage workforce domestically do so because they lack the capital and global reach necessary to automate their operations or to avail themselves of even cheaper labor abroad. The operation of the new international division of labor and of dual labor markets within nations depends upon the way in which a state structures its internal markets and its participation in the world market. Especially crucial here is the state s role in managing the 1 See Gereffi (1994a) for a summary critique of the new international division of labor hypothesis.

3 Changes in the garment industry workforce in the United States 303 international boundary through trade, investment and migration policies (Block, 1994). The present conjuncture in US-Mexico relations is particularly interesting in this regard, because both nations have made concerted efforts to regulate the functioning of the boundary between them. On the one hand, the 1993 North American Free Trade Agreement (NAFTA) and Mexico s unilateral opening policies in the years preceding NAFTA s ratification have rendered the US-Mexico border considerably more permeable to the movement of goods and capital in both directions. On the other hand, since 1993 the United States government has heightened control of its southern border and increasingly restricted the movement of Mexican laborers into its territory (Andreas, 1996; Spener and Staudt, 1998). While it is too soon to speak definitively of the empirical consequences of either of these boundary management policies, two likely outcomes are especially relevant to this discussion. First, by increasing access to the US market and providing a stable production and investment environment in Mexico, NAFTA will substantially deepen US-Mexico economic integration. A central aspect of this integration is the expansion in Mexico of laborintensive manufacturing of goods destined for export to the United States. In a process consistent with the NIDL hypothesis, firms will shift production resources into Mexico from both the United States and other countries in order to gain preferential access to the US market. Second, if the US government dedicates sufficient resources to its current border control efforts (this remains an open question), the flow of unauthorized migrant labor into the United States, from both Mexico and other countries, is likely to be curtailed. A boom in the export of Mexican manufactures to the United States While the data needed to confirm reductions in the flow of unauthorized labor migrants across the United States southern border are thus far unavailable (see United States General Accounting Office, 1999), data from the 1990s show that US-Mexico trade has grown dramatically following Mexico s economic opening of the late 1980s, the implementation of NAFTA and devaluation of the Mexican peso in Total trade between the United States and Mexico grew from $59.2 billion in 1990 to $174 billion in 1998 (United States International Trade Administration, ). Mexico s most dynamic manufactured exports to the United States were mainly commodities whose production was labor-intensive, led by electronics goods, automobiles and autoparts, machinery and capital equipment (including computers and computer parts and components) and apparel. Exports in these industries grew especially rapidly following the creation of the North American Free Trade Area in 1994 (see Table 1). Much of the growth in these exports was fueled by expansion of in-bond assembly or maquiladora production, which grew from $14 billion in 1990, to $23 billion in 1993, to $53 billion in Concomitantly, the number of maquiladora plants in Mexico grew from 1,938 plants with 460,000 employees in 1990, to 2,166 plants with 541,000 employees in 1993, to 3,143 plants with over one million employees by the end of 1998 (Instituto Nacional de Estadística, Geografía e Informática, 1990; 1993; 1998). By the end of the 1990s, maquiladora production accounted for over half of Mexico s total exports to the United States of $95 billion (Instituto Nacional de Estadística, Geografía e Informática, 1998; United States International Trade Administration, 1998). NAFTA, together with the devaluation of the Mexican peso in , has helped cement Mexico s emergence as the low cost manufacturing center of North America (Martin, 1995). Free trade and Mexican migration to the United States NAFTA s proponents argued that over the long term the agreement would create manufacturing jobs in Mexico, raise Mexican incomes and thereby reduce unauthorized migration to the United States (Orme, 1996). Even before debate over NAFTA began, export-oriented manufacturing in Mexico was discussed as a development strategy to reduce pressure on Mexican workers and peasants to migrate to the United States (Cornelius, 1991; Gregory, 1991; Portes, 1991; Weintraub, 1991). Although most analysts

4 304 David Spener and Randy Capps Table 1 US imports from Mexico, Dollar value of imports (thousands) Annual rate of growth (%) Electronic goods 8,216,161 11,051,276 25,783, Automobiles and autoparts 4,253,745 6,075,461 16,734, Machinery and capital equipment 2,527,115 3,615,310 11,629, Apparel 831,692 1,315,882 6,702, Professional instruments 819,467 1,348,738 3,325, Furniture 742,716 1,041,312 2,698, Articles of iron or steel 369, ,814 1,215, Plastics and articles thereof 300, , , Toys, games & sports equipment 270, , , Beverages, spirits and vinegar 233, , , All commodities 31,194,293 39,929,656 94,708, Source: United States International Trade Administration (1999). of migration concluded that NAFTA would actually increase short-run pressures for Mexicans to migrate to the United States (see, for example, Weintraub, 1992; Cornelius and Martin, 1993; Martin, 1995), there can be little doubt that migration pressures would increase much more if the approximately one million people working in maquiladora plants lost their jobs. While scholars of migration have paid some attention to the potential effects of NAFTA and similar free trade agreements on push factors from migrant-sending nations, they have been relatively mute regarding the possible impacts of free trade on pull factors related to the employment of migrants in receiving countries. Nonetheless, a considerable number of immigrants to the United States, including Mexicans, find employment in manufacturing sectors which are vulnerable to free trade. Relative to their share of the total US working population, immigrants are heavily over-represented in several large, labor-intensive manufacturing industries, including apparel, food and kindred products, computers and professional equipment (see Table 2). Of special theoretical interest to us here is the considerable overlap between the manufacturing industries in which immigrant workers are concentrated in the United States and those in which Mexico and other NICs specialize as exporters. Indeed, the Mexican manufacturing exports which have grown most dynamically in absolute and relative terms in the 1990s correspond to the US manufacturing industries with the highest foreign-born shares: apparel, electronics and computer-related goods. This correspondence suggests that trade liberalization measures such as NAFTA may have the consequence of bringing third-world export-platform workers into increasing competition with immigrant workers in low-wage manufacturing in the United States. If this is the case, then the US workers losing their jobs in the process may be less American than suggested by NAFTA critics such as Ross Perot, who predicted a giant sucking sound as good American jobs began to move south of the border (Orme, 1996: 13). Deregulation of US-Mexico trade in apparel Although it does not contribute as much to GDP as other sectors, the apparel industry 2 was one of the largest providers of manufacturing jobs in the United States throughout 2 We use the terms apparel industry and garment industry interchangeably throughout this article to refer to the design, assembly and finishing of clothing. We exclude from our analysis any consideration of

5 Table 2 Foreign and Mexican-born shares of manufacturing workers a in the United States, 1900 and b Total employees % Foreign-born % Mexican-born b b b Apparel and other finished textile products 969, , Office/accounting machines, computers etc. 99, , Food and kindred products 842, , Electrical machinery, equipment and supplies 556, , Professional/photographic equipment, watches 209, , Toys etc. 78,295 67, Rubber and miscellaneous plastics products 445, , Chemicals and allied products 409, , Furniture and fixtures 341, , Leather and leather products 111,724 64, Metal industries 1,102,423 1,047, Textile mill products 476, , Paper and allied products 378, , Stone, clay and glass products 318, , Lumber and wood products 419, , Motor vehicles and motor vehicle equipment 686, , Other manufacturing industries 2,504,967 1,746, All manufacturing industries 9,950,203 8,498, Sample size 520,154 12,051 a Tabulations include only manufacturing workers in operator, fabricator and laborer occupations (occupation codes ). b In order to increase the sample size and generate more precise estimates, we have pooled the samples from the March 1997, 1998 and 1999 Current Population surveys. Data shown in this column should be interpreted as the annual average for these three years. We do not display data for 1993 in this table, because the US Current Population Survey did not begin reporting nativity until 1994 and we would need to pool 1994 through 1996 to get accurate estimates. Therefore, we rely on a comparison between the 1990 US Census and the most recent Current Population Survey data to show trends in nativity. Changes in the garment industry workforce in the United States 305

6 306 David Spener and Randy Capps most of the twentieth century, employing slightly over one million workers in 1990 (United States International Trade Administration, 1990; Murray, 1995; Mittelhauser, 1996; 1997). Since the 1960s, however, apparel has become one of the sectors most affected by international trade, due to rapid increases in export-oriented production among NICs, especially those in East Asia and Latin America (Appelbaum et al., 1994; Bonacich et al., 1994; Gereffi, 1994b; Christerson and Appelbaum, 1995; Murray, 1995; Mittelhauser, 1996; 1997). By the mid-1990s, imported garments accounted for over half the value of apparel purchased in the United States (Gereffi, 1994b; Mittelhauser, 1996). More than any other industry, apparel manufacturing exemplifies the emergence of a new international division of labor, as third-world imports have risen alongside declining US apparel employment. US employment in the industry peaked at around 1.4 million workers in By 1990, however, almost one-third of apparel jobs had disappeared or been transferred to overseas production sites, and this job loss trend continued into the 1990s (Bonacich and Waller, 1994; Murray, 1995; Mittelhauser, 1996; 1997). Job loss in the United States undoubtedly would have been even greater during the 1970s and 1980s were it not for the protections it received under the Multifibre Arrangement (MFA). Under MFA auspices, the US government made bilateral agreements with developing countries that included both tariffs and quotas designed to protect its domestic apparel industry. 3 At the same time, however, the United States also pursued trade policies that promoted the limited outsourcing of apparel firms sewing operations to Mexico and the Caribbean basin under Item 807 of the US Tariff Code. By promoting maquiladora production in nearby developing nations, Item 807 had the effect of diverting a substantial portion of total US apparel imports from Asia to the Americas, while continuing to protect domestic design, cutting and finishing operations (Bonacich and Waller, 1994; Murray, 1995). Since the late 1980s, MFA quotas limiting developing country apparel exports to the United States have been progressively lifted. NAFTA subsequently superseded the MFA with regard to the North American garment trade. It has accelerated the growth of Mexican imports of clothing by allowing clothing assembled from North American components to enter the United States duty and quotafree by the year 2001, with the most important restrictions on garment maquiladora imports lifted in the first years of implementation of the agreement. Then, at the world level, the MFA was replaced in 1995 by the Agreement on Textiles and Clothing (ATC), negotiated under the auspices of the World Trade Organization. The ATC enlarges quotas on fabric and apparel by small percentages each year until all quotas are eliminated in Not surprisingly, total US garment imports have more than doubled in the 1990s, rising from $23.7 billion in 1991 to nearly $50 billion in 1998 (see Table 3). Most recently, African, Caribbean and Central American countries have been granted NAFTA-parity by the United States for textile and apparel goods, a move which is expected to further increase US apparel imports (Odessey, 2000). Growth in Mexican apparel exports to the United States In conjunction with the 1994 peso devaluation, deregulation of the international garment trade has encouraged many large US manufacturers and retailers to source more of their production in Mexico. Strategies to outsource production include opening new production facilities, contracting to Mexican producers and associating with Mexican firms in textile manufacturing and other industries (Bair and Gereffi, 1997; Gereffi, 1997; Kessler, 1997; National Interfaith Committee on Worker Justice, 1998; Patten, 1999). Complementary to employment in textile manufacturing, by which we mean the production of the fabric from which clothes are cut and sewn. 3 According to one study, by the mid-1980s MFA had prevented the loss of 214,000 US garment jobs (Cline, 1987, cited in Murray, 1995).

7 Table 3 US Apparel imports by country, Rank Total imports (in thousands of dollars) in Mexico 1 831,692 1,095,469 1,315,882 1,785,901 2,778,721 3,742,897 5,245,470 6,702,462 China 2 3,434,833 4,475,009 5,296,844 5,085,577 4,653,080 5,025,522 5,998,006 5,670,678 Hong Kong 3 3,941,828 4,266,991 3,940,633 4,323,634 4,260,860 3,927,665 3,972,714 4,447,842 Dominican Republic 4 899,697 1,163,033 1,367,106 1,537,944 1,697,694 1,720,112 2,189,089 2,310,148 Taiwan 5 2,430,974 2,228,182 2,072,899 2,022,287 1,916,836 1,825,157 1,947,789 1,997,997 Korea, Republic of 6 1,929,489 1,926,199 1,945,076 1,884,599 1,660,519 1,423,819 1,554,700 1,903,538 Honduras 7 196, , , , ,236 1,239,814 1,686,241 1,902,527 Philippines 8 983,324 1,156,050 1,235,366 1,316,445 1,488,529 1,458,724 1,573,698 1,725,370 Indonesia 9 565, , ,789 1,029,572 1,189,006 1,332,179 1,602,838 1,663,417 India , , ,820 1,184,030 1,163,491 1,249,951 1,400,789 1,560,434 Bangladesh , , , , ,516 1,020,745 1,331,468 1,497,946 Italy , , , ,861 1,076,198 1,272,202 1,356,449 1,483,313 Thailand , , , ,961 1,041,637 1,058,924 1,275,773 1,468,662 Canada , , , , , ,300 1,204,968 1,419,079 Sri Lanka , , , , ,753 1,001,523 1,197,957 1,300,044 El Salvador 16 90, , , , , ,607 1,052,564 1,170,942 Guatemala , , , , , , ,853 1,137,469 Macao , , , , , , , ,188 Costa Rica , , , , , , , ,168 Turkey , , , , , , , ,267 Other countries - 4,000,064 4,703,825 5,010,805 5,570,572 6,157,679 6,134,958 6,658,030 7,833,800 World total - 23,745,435 28,255,047 30,492,211 33,250,408 36,103,421 37,929,586 44,645,824 49,795,291 Source: United States International Trade Administration (1999). Changes in the garment industry workforce in the United States 307

8 308 David Spener and Randy Capps Figure 1 US apparel imports from top 5 countries (source: United States International Trade Administration, ) this process has been the shift of Mexican garment plants away from production for the domestic market (where consumption of garments plummeted following the peso devaluation see Pozos et al., 1997) and toward exports to the United States as maquiladoras (Carrillo and Hualde, 1997). As shown in Table 1, between 1991 and 1993, US apparel imports from Mexico rose 58% from US $832 million to $1.3 billion, and then more than quintupled to $6.7 billion in In addition, Mexico s rank as an apparel exporter to the United States rose from seventh in 1991, to fifth in 1993 and to first in 1998, surpassing the Dominican Republic, China, Hong Kong, Taiwan and Korea, and its share of total US imports grew from 3.5% to 13.5% (see Table 3 and Figure 1). Changes in garment employment in Mexico and the United States Mexico s surge in apparel exports to the United States in the 1990s has been accompanied by significant changes in employment in both countries. Especially dramatic has been the growth of export-oriented apparel employment in Mexico: in 1993, there were 404 maquiladoras employing 64,510 workers in garment production, but by the end of 1998, there were 224,023 workers producing garments in 912 maquiladoras (Instituto Nacional de Estadística, Geografía e Informática, 1993; 1998). Concurrently, total apparel industry employment in the United States declined from 1.03 million in 1990, to 972,060 in 1993, and then to 835,219 in 1997 (United States Bureau of the Census, 1990; 1993; 1997). Thus, the absolute decrease in US apparel employment of around 195,000 from 1990 to 1997 was nearly half the total loss between 1974 and 1990 (400,000). Job loss among operators, fabricators and laborers in the industry was more severe, falling 36% from 969,391 in 1990 to 622,475 in (see Table 2). 4 Not surprisingly, 30% of the 4 We rely on several different sources for data on apparel industry employment in this article, each of which has its strengths and limitations. Data from the 1990 US Census of Population and Housing s 5 Percent

9 Changes in the garment industry workforce in the United States ,254 workers certified by the US Department of Labor (USDOL) for NAFTA-induced job losses between 1 January 1994 and 1 July 1999 worked in the garment industry, although the garment industry accounted for just 7.3% of all US manufacturing employment in While some US jobs may have been lost due to improvements in pattern-making, cutting and spreading techniques, most of the losses appear to be due to shifts towards imported clothing (Mittelhauser, 1996; 1997; Murray, 1995). Immigrants in the US garment workforce Garment manufacturing has become the United States most internationalized industry, not only in the sense that US companies now contract most of their manufacturing overseas, but also because it has the country s most immigrant-dominated industrial labor force. The industry has been recognized as a magnet for the undocumented immigrant workers that have so concerned government policy-makers since the 1970s (Hill and Pearce, 1987; Long, 1987; Sassen, 1988; Kwong, 1997). Mexican immigrants, in particular, play a significant role in feeding the industry s demand for low-cost labor, and they have been crucial in building the industry in Southern California and the border region of Texas (Rungeling, 1969; Taylor, 1980; Cornelius et al., 1982; Coyle et al., 1984; Muller and Espenshade, 1985; Márquez, 1995). At the end of the twentieth century, Mexican workers continued to dominate workforces in the nation s largest garment district in Los Angeles, as well as in El Paso, which was home to the sixth-largest district (Bonacich, 1994; Appelbaum and Christerson, 1997; Spener, 1998; Light et al., 1999; Kessler, 1997). In both cities, up to two-thirds of operators in the apparel industry were born in Mexico. By , Mexican immigrants made up 20% of garment workers in the United States and 42% of all immigrant garment workers in the United States. At the century s end a large portion of garments purchased in the United States were being produced by 349,000 Mexican workers, 224,000 of them working south of the US- Mexico border and another 125,000 north of it. 5 Sewing the same types of garments for Public Use Microdata Sample (PUMS) are taken from household surveys that permit us to establish the occupation and nativity of employed individuals at the national, state and metropolitan levels for The principal limitation of this source is the infrequency of its publication every 10 years. At the national level, figures derived from the PUMS are comparable with those derived from the March sample of the US Current Population Survey (CPS), which is administered annually. The principal disadvantage of the CPS is that the sample size is too small to use for estimating worker characteristics below the national level. We also make use of data from the US Census Bureau s County Business Patterns (CBP), published annually. CBP employment data are taken from surveys of registered business establishments and reflect employment during the same week in March each year. Because CBP data are based on establishment surveys, they exclude workers who are not employed in registered establishments. Another disadvantage is that CBP data contain no information on the characteristics of workers such as occupation and nativity. As a consequence, the CBP is only useful for establishing overall levels of apparel industry employment at the national, state and county/metropolitan area level. In this article we use the CBP to monitor annual changes in apparel industry employment at the subnational level. The reader should bear in mind that figures derived from the CBP include both white-collar and blue-collar workers in the apparel industry, while figures we present from the PUMS and the CPS include only manual workers, meaning that they are not precisely comparable. An added disadvantage of the CBP data is their tardy publication. At the time of this writing, the most recent year available for analysis is For this reason, we have relied upon data from the Texas Workforce Commission and the California Employment Development Department to establish more recent changes in apparel industry employment levels in El Paso and Los Angeles. Data from these two sources are drawn from monthly surveys of registered business establishments. Like the CBP data, these sources contain no information on individual worker characteristics. While similar to those derived from the CBP, figures from these two state agencies are not strictly comparable to the CBP insofar as they are based on year-round tracking of employment levels while the CBP provides data for only one week out of each year. 5 At the beginning of 1999, 224,023 Mexicans labored in garment maquiladoras (Instituto Nacional de Estadística, Geografía e Informática, 1998). There were 124,495 Mexican immigrant workers in apparel

10 310 David Spener and Randy Capps an integrated North American product market, Mexican workers on opposite sides of the border increasingly enter into direct labor market competition with one another. Analysts who argue that immigrant employment in the US apparel industry has helped retard the movement of production overseas have offered three general explanations exploitation, succession and entrepreneurship for this phenomenon. Exploitation is the most straightforward and is a corollary to both the NIDL hypothesis and dual labor market theory. According to this explanation, immigrants especially the undocumented are vulnerable due to their newcomer status. Employers prefer them to native workers because they can consign them to work for low wages, under substandard conditions in loosely regulated non-union sweatshops (Fernández-Kelly and García, 1989; Sassen-Koob, 1989; Petras, 1992; Blumenberg and Ong, 1994; Bonacich, 1994; Pessar, 1994; Figueroa, 1996; Kwong, 1997; Ross, 1998). Other authors provide support for the succession thesis, which holds that Asian and Latin American immigrants succeed rather than displace natives, as natives move up the employment ladder into better-paying jobs (Waldinger, 1986; 1996; Long, 1987; Gill and Long, 1989; Waldinger and Lapp, 1992). A third explanation centers on the role played by immigrant entrepreneurs in garment manufacturing in Los Angeles, New York and Miami since the 1960s. According to this thesis, immigrants find niches as garment subcontractors in the women s fashion segment of the industry, where they are competitive due to the strength of co-ethnic ties of cooperation and reciprocity between owners and their employees (Waldinger, 1986; Wong, 1987; Morokvasic et al., 1990; Bailey and Waldinger, 1991; Zhou, 1992; Portes and Stepick, 1993; Logan et al., 1994; Palpacuer, 1997). Light et al. (1999) offer a variant of the immigrant entrepreneurship explanation in which Asian immigrant entrepreneurs in Los Angeles hire Mexican and Central American immigrant seamstresses to compensate for shortages of co-ethnic employees in the city. Kwong (1996; 1997) recognizes the important role played by Chinese entrepreneurs in building a garment enclave for fellow immigrants, but nonetheless finds the relationship between owners and workers to be highly exploitative. Bonacich and Appelbaum (2000) found similarly exploitative relations between immigrant entrepreneurs and their employees in the Los Angeles garment district. Thus, the exploitation, succession and entrepreneurship explanations are not mutually exclusive. Other analysts, however, are skeptical that the employment of immigrants by itself can keep the garment industry in the United States, especially if reliance on immigrants means merely securing a cheap source of domestic labor (Christerson and Appelbaum, 1995; Appelbaum and Christerson, 1997; Bonacich and Appelbaum, 2000). These authors insist that apparel manufacturing in the United States has no future if producers pursue low-road strategies at the expense of technological and organizational innovations. In their work on the Los Angeles garment industry, they note that immigrant labor is not sufficiently cheap to prevent job loss because even undocumented workers are paid wages far higher than garment workers in developing countries. It is not just the ready availability of cheap immigrant labor that makes Los Angeles an attractive production site. Just as importantly, Los Angeles remains attractive due to its role as a national design center for cultural products flowing from the synergy between the entertainment and fashion industries, as well as the external economies arising from the agglomeration there of hundreds of firms dedicated to garment design, production, marketing and ancillary business services. Thus, the authors conclude that Los Angeles producers should manufacturing in the United States in , according to our analysis using a weighted average of the March 1997, 1998 and 1999 US Current Population Surveys. The figure for garment maquiladora employment probably significantly underestimates the number of Mexicans sewing apparel for export to the United States since it includes only those workers who are directly employed in an in-bond plant registered as a maquiladora. Because much maquila production is carried out by subcontractors in smalland medium-sized firms working for registered maquilas, these data underestimate both the number of firms and workers involved in export-oriented apparel production (personal communication from Jennifer Bair, 19 June 2000).

11 Changes in the garment industry workforce in the United States 311 not attempt to out-compete China by lowering costs, but should instead compete by raising quality. Strategies to improve quality include: upgrading employees skills, increasing work-schedule flexibility, guaranteeing short turn-around times, keeping up with fashion trends and focusing production on high-price fashion garments. Palpacuer (1997) makes similar arguments with regard to New York s garment industry. 6 Low-cost immigrant labor and entrepreneurial immigrant enclaves may have prolonged the life of the US garment industry through the end of the twentieth century, but the industry may become a casualty of the forces of global competition once they are fully released in the twenty-first century. Waldinger and Lapp noted this possibility at the beginning of the 1990s, well before NAFTA and the ATC were on the table: Ethnic succession, the build-up of immigrant networks, and an immigration system that has privileged kinship relations have made the penetration of the industry by immigrants a selffeeding process... The industry s current labor needs have attracted immigrants with low skills and low levels of education. Today, they complement native workers engaged in higher-level, related activities that depend on the availability of a local labor force. But what of tomorrow? The industry s decline is continuing and, indeed, accelerating. Can today s immigrant garment workers adapt to an economy in which similar, low-skilled manufacturing jobs will have dwindled further? And if not, what are the implications for the legal immigration system? (Waldinger and Lapp, 1992: 105). The future is already evident in El Paso, the nation s sixth largest garment district. We now turn our attention to that city on the border between Mexico and Texas, where thousands of Mexican-American and Mexican immigrant workers have lost their jobs in the years since the creation of the North American Free Trade Area. The displacement of Mexican workers in El Paso The garment industry has played a prominent role in the El Paso economy since the 1920s when Lebanese immigrants Mansour and Hannah Farah began to manufacture work clothing for rail, ranch and mine laborers in the Southwest. After the second world war, production continued to grow as the city emerged as one of the country s major garment districts, specializing in the cutting and sewing of men s and boy s trousers, especially blue jeans (van Dooren, 1997). On the eve of NAFTA, El Paso was in the midst of a boom in garment production brought about by the resurgence of denim fashions and the rise of stonewashing as part of the production process (Bary, 1993). Several large blue-jeans manufacturers, including Levi Strauss & Company, Lee Company, Wrangler and Sun Apparel, had expanded production in their existing plants and/or opened new plants in El Paso in the late 1980s and early 1990s (interview, El Paso Office of Economic Development, October 1996; Medaille and Wheat, 1997). These large companies invested in advanced technologies and initiated flexible organizational approaches in El Paso in order to boost productivity (Crimmins, 1995). As a consequence of this expansion in blue-jeans production, average annual garment employment in El Paso grew from 18,400 in 1990 to 23,600 in In addition to the large-scale blue-jeans factories, El Paso was home to a number of smaller garment firms engaged in cutting, sewing, laundering and finishing blue jeans and 6 Similarly, authors who have analyzed the garment industry outside major immigrant-dominated districts suggest productivity may be improved and costs contained through modular production and partial automation of the assembly of standardized, non-fashion garments (Bailey, 1993; Taplin, 1994a; 1994b; 1995; Berg et al., 1996; Dunlop and Weil, 1996; Mittelhauser, 1996). 7 All employment figures cited in this section of the article, unless otherwise noted, derive from the authors calculations using covered employment and wage data from the Texas Workforce Commission. Apparel employment includes SIC category 23, apparel manufacturing and SIC 721, garment services. SIC 721 includes the laundering and finishing plants so vital to the denim-wear industry in El Paso.

12 312 David Spener and Randy Capps a variety of women s and girls outerwear garments. Some of these smaller garment enterprises, especially the cutting and finishing rooms, actively collaborated in subcontracting networks with Mexican maquiladoras across the border. These firms exhibited many of the characteristics typically associated with immigrant-owned firms in enclaves in Los Angeles, New York and Miami, including a skilled workforce, interfirm collaboration based on social networks, and intra-firm social capital based on ethnicity and networks shared by entrepreneurs and their employees (Bull et al., 1993; Portes, 1995). At the same time, pay in these smaller firms was seldom higher than the minimum wage. Employment was unstable and a sweatshop work environment often prevailed. An industry dominated by Mexican immigrant women In the winter of , we conducted a survey of 300 employees laboring in eight garment plants in El Paso, ranging in size from a major Levi s blue-jeans factory to small subcontracting shops. 8 As indicated in Table 4, the vast majority of sewing-machine operators surveyed were women (82%) and Mexican immigrants (81%). On average, these immigrant women were about 40 years old, having first moved to the United States 15 years earlier (in the early 1980s) at around age 25. Over 30% were single mothers. Around 10% of the sample reported that they continued to live in Ciudad Juárez, commuting to their El Paso jobs across the border on a daily basis. Less than 30% were high-school graduates and most (86%) had attended primary school only in Mexico. Of special interest to us here is the finding that over 30% had previously worked in a Mexican maquiladora and more than one-fifth had worked in the garment industry in Mexico before migrating to the United States. 9 Post-NAFTA job loss With the expansion of maquiladora production of garments in Mexico brought about by the virtual elimination of MFA quotas in 1989, NAFTA and the peso devaluation, Mexican workers in El Paso came into direct competition for their jobs with their co-nationals across the border. As illustrated in Figure 2, Mexico s skyrocketing apparel exports to the United States in the 1990s were led by precisely the same garments in which El Paso specialized men and women s blue-denim trousers and other men s cotton trousers. NAFTA eliminated 807 restrictions requiring garments to be cut, laundered and finished in the United States in order to receive special tariff treatment, thus undermining the potential for El Paso to continue to complement maquiladora production by carrying out these operations. After NAFTA went into effect, US companies sourced increasing amounts of their production in Mexican cities such as Torreón, Chihuahua City, Aguascalientes, Querétaro and Tehuacán (Bair and Gereffi, 1997; Gereffi et al., 1997; Kessler, 1997; Medaille and Wheat, 1997; 8 The sample for this survey was drawn on a convenience basis. In-depth interviews were conducted with the owners/managers of El Paso garment establishments listed in the Texas Directory of Manufacturers. These interviews explored multiple aspects of the operation and organizational history of the informants enterprises. Of the 25 owners/managers interviewed, 10 consented to allowing their direct production workers to be surveyed on-site regarding the composition of their households and their educational, employment and migratory histories. Workers in 8 plants were actually surveyed. Because of the sensitivity of the issue, workers were not asked questions regarding their citizenship or legal immigration status. In order to minimize biases related to the occupational distribution of workers in different types of plants, we have presented results from a 184-person sub-sample of the survey consisting solely of sewingmachine operators (cutters, bundlers, inspectors, launderers and pressers were also interviewed). Although our sample was not drawn purely at random, variation in the sociodemographic profiles of the workers among the 8 plants surveyed was not pronounced. Moreover, our results are generally consistent with data on El Paso garment workers collected in the 1990 US Census of Population and Housing (see Spener, 1997). 9 In our survey, separate questions were asked regarding previous maquiladora employment and previous employment in the Mexican garment industry. Thus, not all respondents indicating maquiladora employment had necessarily worked in a garment maquiladora, nor had all respondents reporting garment industry employment necessarily worked in a garment maquiladora.

13 Changes in the garment industry workforce in the United States 313 Table Selected sociodemographic characteristics of El Paso sewing-machine operators, Winter Number = 184 Percent female 81.5 Average age 40.4 Percent who are of Mexican ancestry 98.9 Percent interviewed in Spanish 91.9 Percent born in Mexico 81.0 Percent born in state of Chihuahua 54.9 Percent born in Ciudad Juarez 37.5 Percent completed high school 28.2 Percent who attended elementary school only in Mexico 85.8 Percent of graduates who attended high school in Mexico 64.8 Percent living with minor child and no spouse present 30.1 Percent currently residing in Ciudad Juarez 9.8 Average age of arrival of Mexican born 25.6 Average years residing in USA of Mexican born 15.2 Percent of total who were born in USA but raised in Mexico 9.9 Percent who ever held a job in Ciudad Juarez 44.8 Percent who ever worked in a maquiladora in Mexico 30.4 Percent who worked in the Mexican garment industry 22.8 Source: Survey conducted by Spener, November 1996 and February Spener, 1997; Chanove, 1998; National Interfaith Committee for Worker Justice, 1998; Patten, 1999). Since peaking at 23,581 jobs in 1993 on the eve of NAFTA, average annual garment employment in El Paso fell to 13,201 in 1999, representing a decline of 44% in just six Figure 2 US imports of Mexican-made garments, (source: US Department of Commerce, National Trade Data Base)

14 314 David Spener and Randy Capps years. Through July 1999, the United States Department of Labor (USDOL) certified 10,037 El Paso workers as displaced by NAFTA-induced production shifts, most (6,166) in the garment industry. In fact, El Paso had the largest USDOL-certified garment job loss of any city in the United States. Of the garment jobs certified by USDOL, 2,106 (34%) resulted from companies shifting their own production to Mexico, while the remainder were accounted for by a surge in Mexican imports. Major companies that laidoff significant numbers of workers as a result of moving portions of their operations out of El Paso to Mexican locations included Farah, Sun Apparel, the El Paso Apparel Group, Action West and Sierra Western. 10 The Department of Labor s certifications list does not include, however, what is by far the largest single group of laid-off garment workers in El Paso in the post-nafta period: the 2,700 employees of Levi Strauss & Company who labored in four sewing plants and one finishing center that have been permanently shut down by the company in three separate actions beginning in November Prior to the layoffs, Levi s had been El Paso s largest private employer, with 4,600 workers (one quarter of its total US labor force). The company had been praised by the state for its investments in training and state of the art technologies (Crimmins, 1995) that were to preserve its ability to produce jeans competitively in the United States. 11 Although Levi s did not initially cite pressures arising from NAFTA as the cause of its lay-offs in El Paso, by 1999 the company had acknowledged that in the future its manufacturing of jeans would rely mainly on its network of international contractors, many of whom were located in Mexico. Various reports indicate that the company s vaunted automation, team production innovations and repeated re-engineering failed to produce hoped-for productivity gains and instead led to higher overhead costs, inconsistent management and bewildered, discouraged workers (Chanove, 1998; King, 1998; Munk, 1999). Thus far, V.F. Corporation s Wrangler and Lee Divisions, which together employ around 4,000 workers in El Paso, have not laid off large numbers of workers. Unlike Levi s, V.F. has successfully used state-of-the art technologies not only in assembly, but also to develop computerized market response systems to track what V.F. garments retail chain clients sell in order to initiate shipment of replacement garments within hours. This enables V.F. to restock its jeans on retail shelves within just three days (Chanove, 1998). At the same time, neither Lee nor Wrangler has been adding employees in El Paso at the levels they promised in order to receive tax breaks from the city. 12 More significantly, the Wrangler division has opened, or is in the process of opening, state of the art facilities in Mexican cities such as Torreón and Chihuahua City and has been using El Paso plants as a site to train new personnel for the Mexican facilities (Gereffi et al., 1997; Chanove, 1998; Patten, 1999). 13 It may only be a matter of time, therefore, before V.F. cuts back on production in El Paso and begins to lay off employees, a possibility that officials with the 10 Sun Apparel, for example, now has 12,000 indirect employees in a network of 72 cutting, sewing and laundering plants throughout the Mexican republic (personal communication from Jennifer Bair, 17 July 2000). 11 A report in Fiscal Notes, the newsletter of the Texas Comptroller, credited Levi Strauss with the following: pioneering a flexible production approach with alternative manufacturing systems, under which a 35 member team make garments from start to finish. Employees are trained to perform different tasks on each garment and trained in such non-tangible skills as team-building and decision-making... [The company] has automated several sewing functions such as thread design, back packets, belt loops and labels. The company has also renovated work stations and equipment to reduce repetitive motion injuries. For example, compact cutting machines reduce fatigue and make units more worker friendly (Crimmins, 1995: 12 13). 12 Personal communication, Roberto De Franco, Director, El Paso Office of Economic Development, 20 July In addition, Wrangler now operates sewing and laundering plants in Yucatán state (personal communication from Jennifer Bair, 17 July 2000).