POVERTY THE WORLD BANK REDUCTION AND ECONOMIC MANAGEMENT NETWORK (PREM) Economic Premise MAY 2012 • Number 84 JUN 010 • Numbe 18 69899 The Promise and Peril of Post-MFA Apparel Production Gladys Lopez-Acevedo and Raymond Robertson For anyone concerned about the effects of globalization on poverty in developing countries, the apparel sector in general and the end of the Multi-Fibre Arrangement (MFA) and the Agreement on Textiles and Clothing (ATC) in particular are key areas of interest. As an important first step toward industrialization, the apparel sector continues to provide an alternative for workers in low-wage agriculture or service jobs (especially less-skilled workers and women), even after other manufactur- ing sectors are established. By providing formal labor experience, these jobs hold the promise of lifelong participation in the labor market, which in the long term can help workers move out of poverty. Therefore, understanding how employment, wage premiums, and the structure of the apparel industry have changed after the end of the MFA and ATC is important to appreciate the effects of this significant policy change on poverty. Why Apparel Matters for Poverty ies and found that if worker characteristics are considered, ap- parel workers earn a sector-specific “premium� over the econo- Developing countries, and especially the poorest countries, mywide average. This means that apparel workers earn more in have high rates of unemployment and underemployment and a apparel than they could elsewhere in the country. Further- large number of workers in agriculture. The pattern of econom- more, Robertson et al. (2009) found that working conditions ic development has historically involved growth in manufactur- are better in globally engaged industries. In other words, appar- ing. One of the first manufacturing sectors to emerge in poor el jobs, relative to domestic alternatives, are often “good� jobs countries is often apparel. Due to relatively low start-up costs, that directly contribute to poverty reduction. De Hoyos, Bus- relatively low skill requirements, and the feasibility of small- solo, and Núñez (2008) found that poverty would have been scale production, apparel provides manufacturing jobs to low- 1.5 percentage points higher in Honduras without the apparel- skilled and female workers that have few other paid employ- centered assembly sector (maquilas). Results are similar in oth- ment options. er countries. Prior to the end of the quota system, entry-level There is a popular perception that average wages are low in workers in Cambodia’s apparel factories earned wages well apparel, and, by international standards, they certainly are. But above the poverty line (Yamagata 2006). although wages are low, the workers in apparel—predominantly The fact that apparel workers in most countries are typi- young, less educated women—are workers that earn lower wag- cally women is another reason why it is important to include es everywhere in the domestic labor market. To compare global- the apparel sector when considering the link between globaliza- ization, wages, and working conditions across countries and tion and poverty.1 Boosting economic opportunities for wom- industries, Robertson et al. (2009) compiled five country stud- en is increasingly accepted as necessary step for economic devel- 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise opment. Paid employment opportunities for women are low the same transition path out of apparel that the now-devel- particularly important for poverty reduction. When women oped countries previously traveled. work, fertility rates fall, and their labor productivity generates The first step to understanding the changes induced by the efficiency gains and higher per capita growth rates. These ef- end of the MFA is documenting how production shifted be- fects are widely recognized and help explain why the third and tween countries when the agreement ended (figure 1). The end fifth United Nations’ Millennium Development Goals2 involve of the MFA was followed by rising apparel exports, falling pric- improving the status of women. es, and a reallocation of production and employment between Therefore, in many ways, what affects the apparel sector countries. Global apparel exports increased from US$252 bil- has the potential to have a significant effect on poverty. Until lion in 2004 to US$366 billion in 2008. As global supply in- 2005, apparel exports were governed by the MFA and ATC. creased, world average apparel unit values fell. The ending of these agreements relaxed global constraints and Many expected that the shifts in apparel production across allowed a radical shift in apparel production between coun- countries after the MFA would be mainly driven by wage differ- tries. The relative success of China is well known—even to those ences because apparel is labor intensive. Wage differences, how- not interested in apparel, poverty, or development. What hap- ever, only explain 30 percent of the variation in the change in pened in the rest of the developing world is a story that had not exports across countries. For example, China was predicted to been told until now—Sewing Success: Employment, Wages, and gain while nearly all other countries lost. But, in fact, other large Poverty following the End of the Multi-Fibre Arrangement (Lopez- Asian apparel exporter countries, such as Bangladesh, also in- Acevedo and Robertson 2012) describes how wages and em- creased exports. Other countries, such as Hong Kong SAR, ployment changed after the end of the MFA and ATC in nine China, Mexico, and Honduras, experienced falling exports and key apparel-producing developing countries: Bangladesh, Cam- market shares. Therefore, export gains were not simply due to a bodia, Honduras, India, Mexico, Morocco, Pakistan, Sri Lanka, shift from higher-wage countries to lower-wage countries. and Vietnam. These countries paint a picture of the rising and Countries that gained the most, including India, Bangladesh, falling fortunes of poor workers after this important policy Vietnam, and Pakistan, implemented proactive policies specific change. to the apparel industry. While wage differences explain some of the production shifts, domestic policies targeting the apparel Production Shifts and Labor Demand sector, ownership type, and functional upgrading of the indus- The link between the end of the MFA and ATC and workers is try perhaps played a more important role. straightforward for those familiar with international economic Wages, Employment, and Working theory. MFA quotas restricted exports and therefore shaped Conditions production patterns: first, MFA quotas restricted exports for those countries with the capacity to produce and export more Significant changes within countries followed these produc- than they were allowed. Second, countries with less production tion shifts. Bangladesh, India, Pakistan, and Vietnam experi- capacity were considered investment and sourcing opportuni- enced rising apparel employment between 2004 and 2008. ties for countries whose exports were restricted. This invest- Honduras, Mexico, Morocco and Sri Lanka, however, experi- ment expanded the otherwise small apparel sector. enced falling apparel employment. Sri Lanka and Bangladesh The resulting production patterns affected the demand for workers. In the bound countries, the quotas limited the de- Figure 1. Changes in Apparel Exports in Selected Countries mand for, and therefore employment of, apparel workers. On log change in total apparel exports the other hand, countries not meeting their quotas received 1.5 �tted values foreign capital and sourcing from the supplier firms to take ad- 1 Vietnam vantage of unfilled quota limits, which created jobs. Removal of Cambodia the MFA quotas would potentially have the opposite effect: in- .5 India Bangladesh B creasing demand for workers in the expanding countries and PakistanSri Lanka Morocco L 0 the risk of job loss in countries no longer attractive because of Honduras H Ar Mexico Rus their quota allocation. -.5 The complexity of the apparel sector, however, means that the story may be more multifaceted than a simple shift of labor -1 demand. Countries that expand exports and production could -1.5 do so through either labor-intensive or capital-intensive tech- 4 6 8 10 12 niques, potentially creating opposite effects on employment. log GDP per capita Countries that lose apparel jobs may simply move those work- Source: Authors’ calculations using data from the World Bank Development ers into other manufacturing jobs and thereby effectively fol- Indicators and OTEXA. 2 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise moved up the apparel value chain in ways that helped explain chain, or “modernizing� production techniques) may support their changes in employment. Other countries remained as sustainable competitiveness in the long run, but may not neces- suppliers of low-end products with almost no decision-making sarily help the poor. Workers without the skills or access to power at other stages of production. Wage premiums rose (in training to transition up the value chain can easily be left be- most cases) in countries that experienced an increase in ex- hind, such as in Honduras. Therefore, policies that support up- ports, and fell in countries with falling exports. The results grading need to be complemented with targeted workforce de- therefore suggest that workers were affected by changes in both velopment to make sure that the most vulnerable workers are employment and wages. not left behind. As the apparel-specific wage premium rose, the male–fe- Concluding Remarks male wage gap fell throughout the economy, not just in the ap- parel sector, in several countries. The falling wage gap shows Exports and employment after MFA/ATC did not necessarily that women’s wages were rising relative to men’s wages, suggest- match predictions, but the changes in many cases were signifi- ing that the entire economy may be affected by the change in a cant. While it is well known that China’s gains in apparel were policy affecting a single sector. This result is consistent with the impressive, the experiences of other countries have received idea that the increase in apparel exports coincides with an in- less attention. Some large Asian countries that were expected to crease in the demand for the factor used intensively in apparel— experience falling exports actually increased their exports. women—that may then have significant long-run poverty-re- Countries with active policies gained the most, including India, ducing effects. Bangladesh, Vietnam, and Pakistan. Policy may have also influenced wage premiums. Coun- These changes corresponded with significant changes in tries that implemented proactive adaptive policies experienced both employment and earnings. Since the apparel premium is a rising wage premiums, but wage premiums declined in coun- critical component of wages that help workers move out of pov- tries with less proactive responses. For example, Sri Lanka took erty (De Hoyos, Bussolo, and Núñez 2008; Robertson et al. steps to focus on higher value production. Total apparel ex- 2009), the fact that premiums rise (in most cases) in countries ports rose, but total employment fell as Sri Lanka shifted into that “gain� and fall in countries that “lose� represents an impor- less labor-intensive stages of the value chain. Similar shifts oc- tant dimension of the workers’ experience after the MFA/ATC curred in Vietnam and Bangladesh. One important lesson that complements employment. from these results is that moving up the value chain may in- Finally, to the extent that upgrading may correspond with crease exports but reduce jobs for women and workers most falling demand for labor (as in Sri Lanka), having a vision for the likely to be closest to poverty. At the same time, Mexico’s expe- evolution of the textile and apparel sector that incorporates de- rience suggests that shifting out of apparel may not necessarily veloping workers’ skills seems crucial. Otherwise, moving too be negative. If the country is moving into more advanced man- quickly up the value chain could result in a lost opportunity for ufacturing, moving out of apparel may be a sign of economic more vulnerable workers to gain valuable work experience in development. manufacturing. One policy-related concern is that the opportunities for Challenges and Prospects pro-poor apparel production may be short lived and are highly One of the key lessons from this experience is that export pro- sensitive to changes in the global economic environment. It is motion alone may not be sufficient to help the poor. While rising important to identify how the change in policy is affecting apparel exports correlated with rising wages and employment in workers and through which channels they are being affected so large Asian countries, they coincided with falling employment that the appropriate policy responses can be applied.3 For econ- in Sri Lanka. Along the same lines, falling apparel exports may omies that are gaining, it is important to know who is benefit- not necessarily be bad for the poor. If rising global competition ting and what conditions are necessary for the poor to benefit provides the impetus to shift toward higher-value products and from the change in policy. If the poor are not benefitting, policy services, as in Mexico, then falling apparel exports may simply be makers need to ask why and determine what, if anything, might a sign of development. But if shifting into higher-value goods be done to help the poor capture some of the gains. and services is not possible, rising competition can generate real About the Authors losses. This seems to be the case in Honduras, where falling ex- ports correlated with falling wages and employment. Therefore, Gladys Lopez-Acevedo is a Senior Economist in the World Bank policies that facilitate movement into higher-value economic ar- Poverty Reduction and Equity Department. She is a fellow in the eas, such as education opportunities, are important. International Initiative for Impact Evaluation and has published Even within the apparel sector, promoting “upgrading� extensively in the areas of poverty, labor markets, and evalua- (defined as shifting to higher-value goods, shifting up the value tion in academic and policy journals. Raymond Robertson is 3 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise Professor of Economics at Macalester College, a nonresident fel- Loayza, Norman, and Claudio Raddatz. 2006. “The Composition low at the Center for Global Development, a member of the of Growth Matters for Poverty Alleviation.� World Bank Policy Research Working Paper 4077, Washington, DC. State Department’s Advisory Committee on International Eco- Lopez-Acevedo, Gladys, and R. Robertson. 2012. Sewing Success: nomic Policy (ACIEP), and a member of the U.S. Department of Employment and Wages Following the End of the Multi-fibre Ar- Labor National Advisory Committee for Labor Provisions of Free rangement. Washington, DC: World Bank. Trade Agreements. Oostendorp, Remco H. 2009. “Globalization and the Gender Wage Gap.� World Bank Economic Review January 23(1):141–61. Notes Özler, Şule. 2000. “Export Orientation and Female Share of Em- ployment: Evidence from Turkey.� World Development 28(7): 1. Several recent papers have focused on the gender dimensions 1239–48. of globalization including Rendall (2010), Aguayo-Tellez, Ai- ———. 2001. “Export Led Industrialization and Gender Differences rola, and Juhn (2010), Oostendorp (2009), Özler (2000, in Job Creation and Destruction: Micro Evidence from the 2001), and Sauré and Zoabi (2009). Turkish Manufacturing Sector.� Working Paper 116, Economic 2. The United Nations established eight Millennium Develop- Research Forum, Cairo. Rendall, Michelle. 2010. “Brain Versus Brawn: The Realization of ment Goals for 2015. The third promotes gender equality and Women’s Comparative Advantage.� SSRN eLibrary. female empowerment, while the fifth focuses on maternal care. Robertson, Raymond, Drusilla Brown, Gaëlle Pierre, and Laura 3. The recent World Bank (2011) report, More and Better Jobs in Sanchez-Puerta, Eds. 2009. Globalization, Wages, and the Qual- South Asia, suggests that “…South Asian countries would bene- ity of Jobs: Five Country Studies. Washington, DC: World Bank. fit from reorienting their labor market policies from protecting Robertson, Raymond, Rajeev Dehejia, Drusilla Brown, and Debra Ang. 2010. “Labour Law Compliance and Human Resource jobs to protecting workers.� Management Innovation: Better Factories Cambodia.� Better References Work Discussion Paper No. 1, ILO, Geneva. Sauré, Philip, and Hosny Zoabi. 2009. “Effects of Trade on Female Aguayo-Tellez, Ernesto, Jim Airola, and Chinhui Juhn. 2010. “Did Labor Force Participation.� Swiss National Bank Working Trade Liberalization Help Women? The Case of Mexico in the Papers 2009-12. 1990s.� NBER Working Paper 16195. World Bank. 2011. More and Better Jobs in South Asia. World Bank, de Hoyos, Rafael, Maurizio Bussolo, and Oscar Núñez. 2008. “Can Washington DC. Maquila Booms Reduce Poverty? Evidence from Honduras.� Yamagata, Tatsufumi. 2006. “The Garment Industry in Cambodia: World Bank Policy Working Paper 4789, Washington, DC. Its Role in Poverty Reduction through Export-Oriented Devel- Frankel, Jeffrey, and Paul Romer. 1999. “Does Trade Cause opment.� Institute of Developing Economies Discussion Paper Growth?� American Economic Review 89(3) (June): 379–99. No.62, JETRO. The Economic Premise note series is intended to summarize good practices and key policy findings on topics related to economic policy. They are produced by the Poverty Reduction and Economic Management (PREM) Network Vice-Presidency of the World Bank. The views expressed here are those of the authors and do not necessarily reflect those of the World Bank. The notes are available at: www.worldbank.org/economicpremise. 4 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise