Report No. 70809-LA Lao PDR Labour standards and productivity in the garments export sector A survey of managers and workers Poverty Reduction and Economic Management Sector Department East Asia and Pacific Region July 2012 CURRENCY EQUIVALENTS Currency unit = Lao Kip US$ 1 = 8,000 Kip ABBREVIATIONS AND ACRONYMS AFTA ASEAN Free Trade Area ALGI Association of Lao Garments Industries ASEAN Association of Southeast Asian Nations ATC Agreement on Textiles and Clothing BFC Better Factories Cambodia BSCI Business Social Compliance Initiative ES Enterprise Survey FGD Focus Group Discussion GATT General Agreement on Tariffs and Trade GDP Gross Domestic Product IFC International Finance Corporation ILO International Labor Organization LECS Lao Expenditure and Consumption Survey LGSS Lao Garment Sector Survey LNCCI Lao National Chamber of Commerce and Industry MFA Multi Fiber Agreement MoIC Ministry of Industry and Commerce MoLSW Ministry of Labour and Social Welfare NSEDP National Socio-Economic Development Plan NTR Normal Trade Relations SocSO Social Security Organization WRAP Worldwide Responsible Accredited Production WTO World Trade Organization Labour standards and productivity in the garments export sector 3 Acknowledgements This report was prepared by Richard Record (Trade Specialist, EASPR), Stephanie Kuttner (Social Development and Gender Consultant) and Kabmanivanh Phouxay (Researcher, National University of Laos), with assistance from Konesawang Ng- hardsaysone (Trade Analyst, EASPR) and Florian Le Bris (Data Analyst Consultant). Overall guidance was provided by Genevieve Boyreau (Senior Country Economist, EASPR), Mathew Verghis (Lead Economist, EASPR) and Keiko Miwa (Country Man- ager, EACLF). Field surveys were carried out by Indochina Research (Laos) Ltd. Ad- ministrative assistance was provided by Phet Udom Mainolath (Program Assistant). The report was peer reviewed by Ximena Del Carpio (Senior Economist, EASHS) and Elisa Gamberoni (Economist, PRMGE) and helpful comments were also received from Daniel Lederman (Lead Economist, PRMTR) and Jill Tucker (Chief Technical Advisor, ILO/Better Factories Cambodia). Marzia Fontana (Institute of Development Studies, University of Sussex) provided useful guidance to the study team. The authors gratefully acknowledge the cooperation of Mr. One-Sy Boutsyvong- sakd (President of the Association of Lao Garment Industries) in providing overall guidance to the analysis, as well as comments and suggestions from the Ministry of Industry and Commerce (Foreign Trade Policy Department-National Implemen- tation Unit and Trade and Product Promotion Department), as well as the many members of the garments private sector who provided inputs to the study. A large number of helpful comments were made at a stakeholder consultation workshop conducted on October 21, 2011 in Vientiane. Finally, this analysis would not have been possible without the contribution of many young Lao garment sector workers who shared their personal experiences and expectations with the research team. This research was made possible by resources from the World Bank’s Gender Ac- tion Plan, and from the Lao PDR Trade Development Facility, a multi donor trust fund financed by Australia, the European Union and Germany, and administered by the World Bank. The findings and interpretations expressed here are those of the authors and do not necessarily reflect the views of the World Bank, its Executive Directors, or the countries they represent. 4 Labour standards and productivity in the garments export sector Summary This report describes labour practices in the Lao garments export sector as seen by both management and labour. It is based on an analysis of primary data collected through a representative survey of firm managers; focus group discussions among current and former garment workers; and, key informant interviews. To date there has been only limited docu- mentation of these issues in Lao PDR and this research seeks to begin filling this gap. The Lao garments industry is comprised of approximately 100 firms, employing just over 20,000 workers. Annual exports are around US$ 200 million, most of which go to the Eu- ropean Union. The industry is modest by most international standards, but is the largest formal sector of manufacturing employment in a country which is trying to manage the effects of a natural resources boom on the rest of the economy. Workers are mostly young women from rural and peri-urban areas seeking to earn income to secure a better life for themselves and their families. Many would have liked to continue their education, but their families could not afford it and a remarkable number of work- ers send remittances home so that siblings can gain a better education (and avoid factory work). Factory work, although difficult and demanding is considered less arduous and of- fers better income opportunities than alternatives such as household farming. Most workers see garments work as an interim strategy to build savings in order to start their own small business and/or family. Workers have very little knowledge of factory work and living conditions prior to recruitment and very limited understanding of what it means to work in a modern industrial setting. Hours are long (8hrs/day, 6 days a week plus over- time) and workers complain about compulsory overtime, sometimes until early morning hours. Even in relatively “good� factories, workers complain of insufficient drinking water, poor air quality and heat, inadequate sanitation facilities, difficult living conditions, strict factory rules and difficult relations with supervisors. Firms identify the supply of labour as their most significant constraint. Attrition from large and medium firms is around 3.5 percent of the workforce every month, while small firms lose on average just over 6 percent of their workers every month. Firms also report that as many as half of their workforce has been with the factory for three years or more, suggest- ing very high drop-out rates and churning primarily among new labour market entrants. As a result, the sector is operating at levels below full capital utilization due to labour shortag- es. Firms want to invest in improving labour productivity, but are afraid of losing their return on investment because of high staff turnover. While high attrition rates plague garments industries in many countries, managers of multi-national companies operating in Lao PDR report that this rate is high, even by regional industry standards. Labour standards and productivity in the garments export sector 5 Reasons for workforce departures are multiple, but include difficult transitions, especially for rural migrants with insufficient information on working and living conditions or understand- ing of their contractual rights and obligations. Wages, both rates and a perceived lack of con- sistency and transparency (due in part to poor numeracy among new workers), and working hours, particularly excessive overtime demands, are identified as the primary reasons work- ers quit. Lao workers also resent pressure and conflicts with supervisors that can be further exacerbated by cultural clashes with foreign male supervisors. The lack of effective systems of worker representation or dispute resolution mechanisms provide limited means of resolv- ing problems, so workers feel they simply have to ‘put up’ with difficult situations and talk of quitting when ‘they can no longer stand it’. The sector appears to be stuck in a low equilibrium trap with worker dissatisfaction driving high worker turnover, which in turn drives low capacity utilization, investment in workforce skills upgrading and training and productivity. Workers cannot voice their interests and there is no effective system of workforce representation or collective bargaining. Thus, when faced with problems, workers feel that they have little alternative but to withdraw their labour, fur- ther exacerbating the labour supply problems that the industry faces. Lao PDR is also strug- gling with access to premium buyers due in part to a perceived lack of credible information on the industry’s labour standards and management practices. In order for the garment sector to progress, this study points to possibilities to increase worker wellbeing, reduce turnover and improve productivity through mutually reinforcing actions. Possible interventions include: a transition support program and improved labour market information systems to support labour force entry; better worker representation and dispute resolution mechanisms; and, a factories improvement program that would include monitor- ing/certification systems to provide incentives for firms to improve conditions as a means of accessing elusive premium buyers. 6 Labour standards and productivity in the garments export sector Contents 1. Introduction 8 2. The garment sector in Lao PDR 9 3. Business environment and industrial relations in Lao PDR 11 4. Garments workers and the labour market in Lao PDR 14 5. Data and methodology 15 6. Firm characteristics and workforce demographics 16 6.1 Firm size, workforce shares and annual sales 16 6.2 Market orientation and international certification 17 6.3 Nationality and gender of ownership 17 6.4 Employee contract status 18 6.5 Gender, age and origin of workers 18 7. Managers’ perspectives 20 7.1 Labour supply 20 7.2 Labour productivity 24 7.3 Labour relations 24 7.4 Working and living conditions 25 8. Workers’ perspectives 26 8.1 Recruitment expectations and experiences 26 8.2 Working conditions 28 8.3 Labour relations 30 8.4 Living conditions 31 8.5 Social status 33 8.6 Future aspirations 33 9. Implications for improving labour standards and productivity 33 9.1 Firm level 33 9.2 Industry level 34 9.3 Government level 35 9.4 Development partners: Better Factories Lao? 35 References 38 Annex A: Methods of quantitative data collection and analysis 40 Annex B: Regression analysis of the determinants of worker departure rates 43 Annex C: Methods of qualitative data collection and analysis 49 Labour standards and productivity in the garments export sector 7 1. Introduction This report describes labour practices in the Lao Our main objective in this report is to present PDR garment sector as seen by both manage- the differing perspectives on working conditions ment and workers. It is based on an analysis of between managers and workers, and to explore primary data collected through a representative ways in which the Lao garment sector could break survey of firm managers, focus group discussions out of this unsustainable situation. More specifi- among current and former garment workers, and cally, insights on management practices, work- key informant interviews in the Lao Garment Sec- ing conditions and worker motivation, as well as tor Survey (LGSS) coordinated by the World Bank industry technology and workflow practices, are between March and May 2011. To date there has drawn from the survey and recommendations been only limited documentation of these issues made on what firms, industry, government and in Lao PDR and this research seeks to begin filling development partners can do to boost productiv- this gap. In particular, it seeks to help increase un- ity in the sector. Given the particular constraints to derstanding of the dynamics and drivers of worker the sector in Lao PDR, there is a particular focus on flows into and out of the industry. how labour standards might be improved upon to reduce the high levels of attrition which are cur- The Lao industry, while modest by international rently inhibiting investments aimed at enhancing standards, is the largest source of employment in labour productivity. the manufacturing sector, and a significant con- Section 2 provides a general overview of the gar- tributor to annual exports. Though the industry ment sector in Lao PDR and its characteristics, has managed to survive the international liberal- situating it in the context of other developments ization of the garment trade, firms are constrained both within the Lao economy and in international by high landlocked transport costs and increas- trade. Section 3 offers a brief description of the ingly trapped in a cycle of low labour productiv- institutional setting, in terms of the business en- ity and high staff turnover rates. While reported vironment, industrial relations, labour laws and instances of extreme violations of international mechanisms for workers’ representation. Section 4 labour standards appear to be few, employees, then situates Lao garment workers in the country’s many of whom are young female migrants, suffer economic transition and shifting labour markets from often difficult working and living conditions, and links this to broader evidence on the gender long hours and compulsory overtime, and, for the division of labour in the global economy, and in most part, have a very limited understanding of particular, on the contradictory effects of working their contractual rights and obligations. They tend in the manufacturing sector for women. Section 5 to see work in the garment sector as a temporary describes data and methodology of the LGSS. Sec- phase in their life undertaken to gain much-need- tion 6 gives a brief overview of firm characteristics ed income to help parents and siblings make ends and workforce demographics as revealed through meet, and to improve their own future prospects. the LGSS. Sections 7 and 8 present and contrast the Thus, firms struggle to raise productivity when perspectives of managers and workers on labour faced with a perpetual need to replace experi- practices and productivity in the Lao garment sec- enced workers with new unskilled recruits, and tor; together, these sections constitute the main the sector remains stuck in a sub-optimal equilib- contribution of the report. Finally, section 9 offers rium of low-productivity and high turnover rates. reflections on possible ways forward. 8 Labour standards and productivity in the garments export sector 2. The garment sector in Lao PDR While modestly sized by the standards of neigh- Indeed, the LGSS found garment managers to be bouring country garment producers1, the Lao gar- generally positive about business prospects for ment sector is the country’s largest manufacturing their firms and the sector generally. The major- employer. The industry employs more than 20,000 ity of managers across all firm sizes are expect- workers (or about 1 percent of the country’s total ing sales to increase in 2011, although a smaller labour force) and produces almost exclusively for majority (especially amongst large firms) believed international markets—some firms sell directly this would correlate to increased profits and some to international buyers while others subcontract. were less optimistic about the industry’s pros- Similar to workforce demographics in garment in- pects over the medium term. Indeed, findings dustries globally, most factory workers are young from a recent global study of the post-Multi-fibre women between the age of 16 and 25, about half Agreement (MFA) apparel industry would suggest of whom have migrated from regions outside of that the small Lao industry will continue to face the capital, Vientiane (where almost all factories downward wage pressures from larger competi- are located). tor countries unless it can successfully upgrade its apparel sector, including through investment in Most Lao garment firms were established in the workers (Lopez-Acevedo and Robertson 2012). early to mid 1990s as manufacturers sought to establish a production base in a country that was Annual garment exports have hovered just under not quota constrained and had relatively low la- the US$ 200m mark in recent years4. This currently bour costs. Many anticipated a severe contraction represents the fourth most significant export sec- in the Lao garment industry following the expira- tor in terms of revenue after minerals (mostly cop- tion of the Agreement on Textiles and Clothing in per and gold), wood products and hydropower. 20052, but the sector has managed to confound However, while the sector has seen reasonable expectations by surviving and indeed continuing rates of growth during the last decade, this has to grow3 . Similarly, the sector managed to escape been outstripped by faster growth in the mining the significant drops in orders that so badly af- and hydropower sectors. Thus the share of gar- fected neighbouring Cambodia during the 2008- ment exports as a percentage of total exports 09 global financial crisis and drops in world trade. has declined from around 24 percent in the early As of mid-2011, garment producers in Lao PDR 2000s to just over 6 percent in 2011 (Figure 1). Val- were reporting strong orders with a number of ue addition appears to be growing steadily also as firms now unable to meet demand from buyers. the average value of each piece is rising. 1 Compare, for example, with 290,000+ garment workers in Cambodia. 2 The Agreement on Textiles and Clothing (ATC) was enacted in 1995 as part of the establishment of the WTO. Previously trade in textiles and garment was controlled via a complex system of quotas under the Multi Fibre Agreement (MFA) and lay outside the scope of the General Agreement on Tariffs and Trade (GATT). The purpose of the ATC was to provide WTO members with a ten year adjustment period to dismantle quotas. 3 This, notwithstanding research showing that it is increasingly difficult for smaller suppliers to compete with leading supplier countries and firms who have strengthened their positions in the post-MFA world (Staritz 2010). 4 Gross garments exports reached a record US$ 206m in 2008; and in an August 2011 article in the Vientiane Times newspaper, the head of the Association of Lao Garment Industries reported that export value of Lao garments had reached about US$ 100m during the first six months of 2011, rising about 29 percent from the same period of the previous year. Labour standards and productivity in the garments export sector 9 Figure 1: Lao garment exports have been rising Figure 2: The EU is by far the largest destination steadily, but the sector’s share of total exports has for Lao garment exports declined 250 30 Others Japan Canada 2% 3% 3% 25 200 20 150 US 15 17% 100 10 50 5 EU 0 0 75% 2003 2004 2005 2006 2007 2008 2009 2010 2011 Exports (US$m) (LHS) Quantity (Million pices) (LHS) % of total exports (RHS) Source: ALGI and World Bank staff estimates Most garments (around 75 percent) are exported a source of demand for relatively low skilled labour) to the EU. This is primarily resulting from the addi- has played an important role in the transition of tional tariff advantages associated with Lao PDR’s the population working in the agriculture sector duty-free access to the EU market under the terms to industry. of the “Everything But Arms� Agreement, but also due to the small size of Lao garment producers However, the Lao garment industry struggles to and their inability to meet the larger orders required compete given low productivity, high transport to compete in the US market. The US market has costs and long lead times associated with a land- grown in importance since the establishment of locked supply chain6 . Exporting a standard 40ft Normal Trade Relations (NTR) in 2005, but still remains container from Vientiane adds as much as 45 per- of limited importance to the Lao garment sector cent to total shipping costs to the final destination (Figure 2). and can double total shipping time, compared to exporting a container from Bangkok7. Nonetheless, the garment industry is an impor- tant component of the current Lao Government’s Firms also have difficulties in attracting and retain- trade integration and export diversification strat- ing workers. Many garment firms would like to be egy. It is also one of the few major export sectors running more production lines and report having that has the potential to generate signficant em- to turn away orders because they do not have suf- ployment, with the other primary exports in the ficient machine operators. This problem appears natural resource sectors generating comparatively to be pervasive regardless of enterprise size or few jobs5 . Similarly, in many other developing ownership8 . Reasons cited by the industry man- countries, the garments manufacturing sector (as agers for the existing shortage include: a recent 5 See Record and Nghardsaysone (2010) for a wider discussion on Lao PDR’s export economy. 6 A 2006 study by the Lao National Statistics Centre focused on three main aspects of competitiveness in the Lao garment sector: labour costs, labour and capital productivity, and lead times. The study found that the hourly wage in Lao PDR was considerably lower than in China (US$ 0.12 vs. US$ 0.68) but that Lao labour productivity in the sector was also substantially lower. (NSC 2006) 7 World Bank (2010) ‘Lao PDR Trade Facilitation Performance’, Trade Development Briefing Note No.2, The World Bank Vientiane Office, January. 8 World Bank (2011) Lao PDR Investment Climate Assessment: Policies to Promote Growth in the Non-resource Sectors, The World Bank, Washington, DC. 10 Labour standards and productivity in the garments export sector expansion of this sector; competition from other for this survey with current and former garment sectors like tourism (now growing at 16 percent workers suggest that other issues are also be at per annum); higher relative wages for unskilled play: namely, that long hours and low pay rates, jobs in Thailand; and the negative social status as- difficult working and living conditions, as well as sociated with being a garment worker in Lao PDR. tense workplace relations are key reasons that However, findings from interviews conducted workers leave the industry. 3. Business environment and industrial relations in Lao PDR Lao PDR, with its small population, low rate of ur- comparator countries and lower than one would banization, and high inland transport cost to re- expect for a country at this income level. Indeed, gional sea ports, will perhaps inevitably struggle productivity in Lao PDR is slightly lower than Cam- to compete on cost with regional competitors bodia and Vietnam and significantly lags behind in labor-intensive manufactured exports. The the leading countries in the region such as China, economically active population is estimated at Thailand and Malaysia (Figure 3). 3 million, half of the total population. A low ur- banization ratio of 27 percent puts the total non- Figure 3: Labour productivity in Lao PDR is lower agricultural labor force in 2010 at around 650,000 than in the most successful exporting economies persons. Out of this, only about 260,000 are em- ployed by firms in manufacturing and service Value added per worker (2005 US$) sectors9 . In contrast unofficial estimates put the Malaysia number of migrant workers working in Thailand Chinea at about 250,000. This small employment base in Thailand Armenia the country certainly contributes to the shortage Moldova of labor faced by garment manufacturers. This Cambodia study seeks to go behind these figures and find Vientnam Kyrgyz Repulic out more about what motivates workers to enter Laos the garments sector and why they leave at such Mongolia high rates. Uzbekistan Nepal Tajikistan Research from the new Investment Climate As- sessment for Lao PDR also suggests that labour $0 $5,000 $10,000 $15,000 productivity of Lao firms is lower than in most Source: World Bank Enterprise Surveys (2009) 9 Ibid Labour standards and productivity in the garments export sector 11 New research also finds that unlike in most coun- and small firms are more productive than large tries, exporting firms in Lao PDR show lower lev- firms. Average value added per worker for all firms els of productivity than non-exporters (Table 1)10 . in Lao PDR is US$ 1,554, while the average for gar- This finding is counterintuitive as in most coun- ment firms is US$ 1,009. In addition to sectoral dif- tries exporters are more productive than non- ferences between exporters and non-exporters, it exporters, large firms are more productive than is also possible that low levels of competition in small firms and foreign-owned firms are more pro- domestic markets affect measured productivity. ductive than domestic firms11. This is not the case This study seeks to investigate further possible in Lao PDR. World Bank analysis indicates that al- factors contributing to low-productivity including though the median foreign-owned firm is slightly management practices, working conditions and more productive than median domestic firm, non- worker motivation, as well as industry technology exporters are more productive than exporters12 and workflow practices. Table 1: Firm performance in Lao PDR, by firm type Number of Value Unit Labor Labor Costs Capital per Capital Workers added per Costs per Worker worker Productivity worker (book value) All 11 $1,554 36% $439 $1,110 134% Internationalization Exporters 30 $415 60% $197 $1,233 72% Non-exporters 10 $3,808 32% $533 $888 194% Ownership Foreign-Owned 89 $1,775 38% $710 $1,686 131% Domestic 10 $1,516 36% $432 $1,057 136% Size Small (5-19) 7.5 $2,190 36% $355 $740 214% Medium (20-99) 30 $1,606 33% $666 $2,466 17% Large (100 and up) 220 $1,246 39% $631 $711 131% Sector Garments 30 $1,009 68% $601 $841 103% Furniture and wood 8 $451 60% $333 $2,561 93% Other 11 $3,916 32% $444 $888 194% Source: World Bank Enterprise Survey (2009) 10 Ibid. 11 Note that size categories (small, medium, large) used in the Enterprise Survey (which includes firms from all sectors) differ somewhat from those used in this Garments Sector Survey. However, the relationship between firm size and labour productivity across the larger formal private sector found in the Enterprise Surveys appears to also hold in the garments sector. 12 This is a result taken from the 2009 Lao Enterprise Survey (and discussed in the 2011 Investment Climate Assessment), and holds for all manufacturing firms in the ES sample, of which garments manufacturing represents around a third. It is a counterintuitive result, but likely driven by the emerging signs of Dutch disease associated with the Lao natural resources boom, as well as low competition in the domestic market. This is driving up the returns to non-exporters, while exporters struggle with a costly supply chain and low productivity which erodes competitiveness. 12 Labour standards and productivity in the garments export sector When asked to name the most binding constraint However, among manufacturers (of all products) to business development, businesses in Lao PDR in Lao PDR, workforce education is identified as (including both manufacturers and services firms) the most significant constraint (Figure 5). identify taxation, access to finance and an inad- equately educated workforce (in that order, see Figure 4)13 . Figure 4: Tax rates, access to finance and lack of Figure 5: Tax rates are the main obstacle for firms skilled labor as the top obstacles in service, poor skills and finance lead in to growth (% of firms) manufacturing firms (% of firms) Tax Rates Tax Rates Access to Finance Access to Finance Workforce education Informality Workforce education Electricity Informality Access to Land Transportation Access to Land Customs & Trade Reg Electricity Licenses & Permits Laos Transportation Tax Administration Services Courts Customs & Trade Reg Manufacturing Corruption Tax Administration Licenses & Permits 0 10 20 30 40 0 10 20 30 40 Source: World Bank Enterprise Survey (2009) The garment industry saw some of the first private economies. During the 2000s, political and trade capital investment in Lao PDR following the intro- union leaders in China and Vietnam (which are duction of the New Economic Mechanism in 1986, the models often used by Lao economic policy which signalled the country’s transition towards a makers) began to recognize that without prop- market economy. The Association of Lao Garment erly functioning trade unions to represent work- Industries (ALGI) is therefore one of the longest es- ers in collective bargaining and social dialogue, tablished members of the Lao National Chamber it would be hard to maintain the stability and co- of Commerce and Industry (LNCCI). However, the hesion needed for sustained economic develop- Lao private sector continues to struggle with lega- ment. As in both China and Vietnam, Lao national cies of a heavily regulated and planned economy. trade union structures are fully incorporated into the political structure of communist party (Yoon The Lao government has also struggled to adapt 2009). The representation of workers in the Lao from its role as ‘central planner’ to more one of garment and all other sectors of the economy re- ‘regulator and facilitator’. This is evident in the mains solely mandated through the official mass area of industrial relations which have not evolved organizations of the Lao People’s Revolutionary apace with the emerging private sector and lag Party (the Lao Federation of Trade Unions, the Lao even in comparison to other regional ‘transition’ Women’s Union, and the Lao Youth Union). 13 World Bank (2011) op. cit. Labour standards and productivity in the garments export sector 13 While the 2006 Lao Labour Law establishes basic ing do not appear to be sufficient to counter some national labour standards (such as minimum wag- of the unfair labour practices and difficult working es, restrictions on overtime, and entitlements such conditions reported by workers. Absent a system as maternity and sick leave) 14 and incorporates el- of collective bargaining or effective representa- ements of ILO core labour standards15 , these ap- tion through official mass organizations, garment pear to be only weakly enforced and there is very workers have limited effective bargaining power. limited opportunity for third party arbitration or As a result, when individuals are dissatisfied with dispute resolution. Labour standards established working conditions, they have little alternative under contracts with foreignbuyers or interna- than to simply quit - contributing to the industry’s tional trading regimes seem to create stronger vexing constraints of high turnover rates and in- inducements for compliance, although these sys- sufficient labour supply. tems of voluntary certification and periodic audit- 4. Garments workers and the labour market in Lao PDR Women’s participation in paid employment has authors focus on exploitative working conditions, risen in most countries over recent decades (ILO health hazards, and lack of training and promo- 2007) and Lao PDR is no exception16 . While many tion, others argue that such images do not always factors contribute to this trend, trade liberaliza- hold true and that factory work is the only option tion clearly drives the feminization of manufac- most women have for earning ‘legitimate’ cash turing in many developing countries: the greater income and ‘working with dignity’. Furthermore, the share of garments, textiles and electronics in while working conditions are generally poor a country’s exports, the greater the employment- across labour-intensive manufacturing sectors, it is creating impact of trade for women. This has led not clear they are significantly worse than in other to a proliferation of studies and debates around non-skilled or semi-skilled jobs open to women. the situation of women workers within globaliza- However, patterns are varied and changing, with tion, their struggles for rights, and the implica- differences, for example, between semi-industri- tions of the rise of female waged employment for alized countries and lower income countries, and women’s autonomy and self-esteem. While some between young single women (who used to be 14 Labour Law (Amended), Passed by National Assembly 27 Dec 2006; Promulgated by Presidential decree No. 05/PO, 16 Jan 2007. 15 Lao PDR has ratified five out of the eight ILO Conventions corresponding to core labour standards, and the revised Labour Law of 2006 includes key principles from these ratified conventions. Lao PDR has ratified ILO Conventions as follows: C29 on Forced Labour; C100/C111 on Equal Remuneration and Employment Discrimination; C138 on Minimum Age; and C182 on Worst Forms of Child Labour. Although Lao PDR did not ratify the Conventions on Collective Bargaining (C98) and Freedom of Association (C87) elements of these core labour standards are included in the 2006 Labour Law (although within the unitary socialist state model of a national workers’ trade union). The only Convention that has neither been ratified nor reflected in conforming legislation is C105 (Abolition of Forced Labour), but a study has been undertaken by MOLSW in partnership with ILO to support a decision on this issue (http://www.ilo.org/ilolex/english/index/htm accessed on 05/05/2011). 16 Lao Population Census data show that between 1995-2005, the proportion of women in wage employment nationwide increased from 38 to 44 percent, rising at a rate of about 1 percentage point per year (NSC 2005). 14 Labour standards and productivity in the garments export sector the preferred workforce, at least initially, in Asia) ited. Nonetheless, garments sector work is one of and older, married and better-educated women the few opportunities available to young women who are increasingly perceived as better workers with limited skills or education to access wage (in Mexico for instance, because they are more employment. mature, reliable and less apt to jump from one job to another) (Fontana 2009). Lao garment workers can also be situated in the context of broader shifts in Lao labour market dy- The labour market in Lao PDR, although changing, namics. In particular, they are part of the grow- remains highly informal and agriculture-based. ing phenomenon of rural-urban migration linked More than 71 percent of households—including to shifts away from agriculture to service and in- 40 percent of urban households—report agri- dustrial sector work in the context of the country’s culture, forestry, or fishing as their main activity rapid economic transition. Female rural–urban and most of this work is on family-run smallhold- migration to factory work is a relatively recent ings17. The latest Population Census data suggests phenomenon in Lao PDR, and it has successively that only around 6 percent of the population, gained acceptance as an economic strategy in ru- or around 13 per cent of the active labour force, ral communities. However, young women experi- works in the formal sector; and that public sector ence many tensions between traditional respon- workers18 still outnumber private sector workers . sibilities towards their families, their low-paid and Thus, notwithstanding Lao women’s high labour low-status positions as garment factory workers, force participation rates (amongst the highest in and their aspirations to participate in modern the region)19 most of this is as ‘own account work- urban consumer culture (Phouxay and Tollefsen ers’ or ‘unpaid family workers’ in the agriculture 2011). This study seeks to understand better the sector (World Bank and Asian Development Bank, experiences of Lao garment factory workers as forthcoming). Thus, the labour market in Lao PDR they try to balance these competing demands remains very thin and capacity for labour market and aspirations. regulation or standards enforcement remains lim- 5. Data and methodology The analysis presented in this report is based on of the garments workforce. Data from these primary data collected by the World Bank in Lao surveys provide a broad picture of basic firm PDR between March and May 2011 through a va- characteristics, workforce demographics, la- riety of methods as follows: bour standards and practices, and informa- tion related to labour productivity in the Lao 1. Structured firm surveys conducted with garment sector.20 owners/managers of garment factories us- ing stratified random sampling of more than 2. Focus group discussions (FGDs) conducted half of all firms, representing over 70 percent with garment factory workers from small, me- 17 National Statistics Center, Lao Economic and Consumption Survey 2007/08 (LECS4). 18 National Statistics Center, Population Census 2005. 19 UNDP, Asia Pacific Human Development Report, 2010 (calculations based on ILO, Key Indicators of the Labor Market, 2009, Sixth edition, Geneva). 20 For more details, see Annex A on quantitative methods and Annex B on qualitative methods. Labour standards and productivity in the garments export sector 15 dium and large-sized factories. Participants social status and expectations; personal well- were selected to represent a range of demo- being and future aspirations. graphic characteristics and work-experiences (long-term and shorter-term employees, cur- 3. Individual interviews conducted with work- rent and former workers, migrants and Vien- ers, supervisors (head of line) and dormitory tiane residents, ages, marital/family status, managers from the same factories selected etc.). Discussions touched on the following for FGDs, as well as with some key industry topics: recruitment experiences and expecta- leaders. These were used to triangulate the in- tions; working and living conditions; workers’ formation collected through the surveys with managers and FGDs with workers. 6. Firm characteristics and workforce demographics The following section gives a brief overview of firm small firms (with less than 100 employees), me- characteristics and workforce demographics21 as dium firms (with between 100-499 employees), revealed through data collected in the LGSS: and large firms (with 500 or more employees)22. The following chart (Figure 6) describes these cat- 6.1 Firm size, workforce shares and egories of firms in terms of their share of all Lao annual sales garment firms (share of total firms) as well as their share of the total Lao garment workforce (work- The LGSS drew from a sample of 89 garment force shares). Thus, as can be seen below, large manufacturing firms employing a total of 19,540 firms constitute only 10 percent of all firms, but workers. Firms were grouped according to size: almost half of the total garment sector workforce. Figure 6: Larger firms are few in number, but account for the bulk of the workforce Workforce Shares (%) Source: World Bank LGSS (2011) 21 The section on workforce demographics is based on a sub-set of the sample because three firms did not complete the survey (1 large + 2 medium firms) and were excluded. 22 Please note that this scale of garment firm size was developed specifically for this study of the Lao garments sector to reflect observed seg- mentation of the industry which in turn allows for analysis of relationships between specific firm characteristics/profiles and labour practices and/or productivity outcomes. This scale differs from the Lao Government’s categorization of firms as micro/small/medium/large as those hav- ing respectively, 1-4, 5-19, 20-99,100+ employees. According to this classification almost all Lao garments firms would be considered either medium or large. 16 Labour standards and productivity in the garments export sector Medium firms constitute another 43 percent of primarily for indirect export, and, to a lesser extent, the workforce, while small firms constitute only 8 for the domestic market24 (Figure 7). Most exports percent of the workforce, although they account are destined for Europe (62 percent), US/Canada for over half of all firms. (18 percent), Japan (12 percent) and Thailand25 (4 percent)26 . Two-thirds of all large firms report Although not all firms reported and there were participating in an international social compliance some difficulties in interpreting the figures de- certification scheme, while no small firms and only clared by those who did, average annual sales (Jan one medium-sized firm did27. – Dec 2010) ranged from around US$ 67,000 for small firms, to US$ 640,000 for medium firms and 6.3 Nationality and gender of US$ 6.8 million for large firms23. ownership 6.2 Market orientation and The ownership structure of garment firms also var- international certification ies significantly by size. Small firms are overwhelm- ingly Lao-owned, while a majority of large firms Firms across all size-categories tend to have a have foreign or mixed foreign and Lao ownership. ‘dominant market orientation’; that is, they tend Medium firms are more evenly split between Lao- to produce almost entirely for either direct export, owned and foreign or mixed-ownership struc- indirect export (i.e. sub-contracted by a direct ex- tures. Foreign owners are predominantly regional porter), or for the domestic market. Direct exports nationals with Thais constituting the largest group represent the vast majority of total sales declared of foreign owners (50 percent) followed by Japa- by large firms and about two-thirds of total sales nese (20 percent) and Australian (10 percent). by medium firms, while small firms are producing Figure 7: Market orientation of Lao garment manufacturers (% sales for export/domestic market) 100 75 Domestic market 50 Direct export 25 indirect export 0 Small Medium Large Source: World Bank LGSS (2011) 23 Most of the firms declared their annual sales in Lao Kip (LAK) which was converted in USD taking the average exchange rate over the period, but some inconsistencies remained in the data and 3 firms had to be dropped from the sample. Thus, these figures remain largely indicative and useful more for understanding differences in orders of magnitude of market shares across firm sizes. Keep in mind also that there are sig- nificant differences within categories in terms of firm size/market shares (e.g. large firms ranged from 600 to over 2000 employees). 24 As the domestic market in Lao PDR is extremely small, some portion of these reported ‘domestic’ sales are likely being sold to larger export- ing firms. 25 Many exports reported to Thailand are likely sub-contracted production destined for onward export to other markets. 26 Note that these figures differ slightly from those reported by the Association of Lao Garment Industries in Figure 2. However, the overall trend is broadly the same. 27 Firms most commonly participate in WRAP (Worldwide Responsible Accredited Production), a US-based non-governmental organization established in January 2000 as a certification program for labor-intensive consumer products manufacturing and processing, based on 12 principles focusing on local law and workplace regulations, generally accepted international workplace standards, and the environment. One firm participates in SA 8000 created in 1997 by US-based non-governmental organization Social Accountability International for improving working conditions with eight core elements including: health and safety, working hours, child labor, forced labor, discrimination, freedom of association and collective bargaining, wages, and discipline. Another participates in BSCI (Business Social Compliance Initiative) of the European-based Foreign Trade Association created in 2002. Labour standards and productivity in the garments export sector 17 Women represent nearly a third of Lao garment der stereotypes of being more docile and more firm owners and over half of all managers28. Both compliant than male workers and whether this female ownership and management presence vary in turn leads to less favorable terms of contract in by firm size with higher levels of both amongst this sector. medium and small-sized firms as compared with large firms. 6.5 Gender, age and origin of workers 6.4 Employee contract status Women represent 85 percent of the total Lao gar- The majority of workers in medium and large firms ment sector workforce. While there are more wom- surveyed hold permanent ‘open-ended’ or ‘term’ en than men employed across all job categories in contracts (of at least 1 year), while only 36 per- garment firms, the distribution of male workers cent of those in small firms do. The remainder of and female workers across job categories differs all workers are either short-term employees (less somewhat: men are more likely than women to than one year) or day labourers. There was no gen- be hired as professional or technical staff, or as der-bias found in employment contract status as unskilled workers (Figure 8). Women are also the revealed through this survey within the garment majority (79 percent) of supervisors at the produc- sector; that is, male and female garment workers tion level across all firm sizes, although this figure were found to be as likely to hold permanent or is somewhat lower amongst medium-sized firms short-term contracts across small, medium and (65 percent). large firms. We note, however, that there is a high level of Figure 8: Female workers are under-represented gender segregation across sectors in Lao PDR (as in professional, technical and unskilled in most developing countries) with women con- positions relative to industry as a whole. stituting the vast majority of workers in the gar- % of total workforce % of female workforce ments sector. Therefore, it would be useful to do 85.2 87.9 a gender analysis of contract status across differ- ent types of industries where men and women are concentrated to better reveal any systemic or 6.3 4.4 8.5 7.7 structural biases in contract status between men and women. It would also be useful to delve deep- Professional Skilled or semi-skilled Unskilled worker er into garment manager perceptions and hiring (managers, (sewing machine (cleaners, guards, accountants, etc.) and operators, cutters etc.) packers, etc.) practices, for example, to determine whether em- Technical (engineers, ployers prefer young female garment workers in mechanice, etc.) part because they are believed to conform to gen- Source: World Bank LGSS (2011) 28 Note that amongst smaller firms, owners are often also acting as managers. 18 Labour standards and productivity in the garments export sector About half of all female production workers mi- Female professionals tend to be younger than grate from outside of the capital city Vientiane, their male counterparts and the youngest seg- and thus they are more likely to be migrants than ment of the workforce (workers less than 17 years their male counterparts, 62 percent of whom are old) is predominantly female (Table 2). Vientiane residents (Figures 9 and 10). Male pro- fessional and technical staff are also slightly more likely to be from foreign countries (22 percent) than their female counterparts (19 percent). Figure 9: Most male garment workers are Figure 10: …while female garment workers are Vientiane residents… more likely to be migrants Male Semi-Skilled Female Semi-Skilled Foreign Migrant 1% 51% Lao 49% VTE VTE Migrant resident Lao resident Migrant 62% 37% Source: World Bank LGSS (2011) Source: World Bank LGSS (2011) Table 2: Distribution of garment workers across job categories by age and gender 14-17 yrs. 18-25 yrs. 26-40 yrs. 40+ yrs. Job category %M %F %T %M %F %T %M %F %T %M %F %T Professional 0.3 0.0 0.2 9.9 12.9 13.0 62.8 72.6 67.8 27.0 14.5 18.9 (managers, accountants, etc.) and Technical (engineers, mechanics, etc.) Skilled or semi-skilled 0.5 2.0 1.7 51.1 50.9 50.7 38.3 42.3 42.3 10.1 4.8 5.7 (sewing machine operators, cutters etc.) Source: World Bank LGSS (2011) Labour standards and productivity in the garments export sector 19 7. Managers’ perspectives This section reviews a series of key concerns raised majority cited labour-related constraints as their by managers regarding labour practices and pro- main concern. This finding was largely confirmed ductivity in the Lao garment sector, namely: la- in the 2011 LGSS in which large and medium firm bour supply (recruitment and retention challenges), managers overwhelmingly identified ‘labour con- labour productivity (skills and training), labour straints’ as their biggest challenge in doing busi- relations (workplace dynamics and dispute reso- ness in Lao PDR (while smaller firms had more lution), as well as working and living conditions difficulties with raising capital and/or securing at their factories. The discussion seeks to reveal financing; Figure 11). Across all firm sizes, the ma- managers’ perspectives on the nature of these jority of managers identified labour supply as their challenges and what strategies firms are using to primary labour-related constraint and labour skills address them. and productivity as their second biggest (Figure 12). 7.1 Labour supply The 2009 World Bank Enterprise Survey for Lao PDR29 included 50 garment factories among the sample. Among these firms the overwhelming Figure 11: Labour is the top constraint for Figure 12: and in particular the supply garment firm. of labour constriant constriant Large Medium Small Large Medium Small labour labour supply labour skills and pro gov’t taxes & admin labour disputes infra’ture labour costs import / export other none 0 20 40 60 80 100 0 50 100 Source: World Bank LGSS (2011) 29 Raw data available at www.enterprisesurveys.org. 20 Labour standards and productivity in the garments export sector All firms appear to have difficulty attracting and In addition, firms across all size groups appear to retaining workers with attrition from large and be losing more workers than they are recruiting. medium firms at around 3.5 percent of their work- This would indicate that the sector is continuing force every month, while small firms lose over 6 to operate at levels below full capital utilization. percent of theirs. In contrast, firms report that as many as half of their workforce has been with the Managers report that their employees leave for a factory for three years or more. This suggests that number of reasons. They believe that those who there is a core longer- term workforce in the in- leave the industry completely are most likely to do dustry, and then a substantial proportion of work- so in order to return home to their communities or ers that are constantly entering and exiting. While due to family responsibilities such as getting mar- high attrition rates plague garments industries in ried, having children, etc. Workers at larger firms in many countries, managers of multi-national com- particular seem more likely to move across facto- panies operating in Lao PDR report that this rate is ries within the garments sector, looking for better high, even by regional industry standards. pay or working conditions. Figure13: A high percentage of workers leave Figure 14: …but it is larger firms that are most the factory every month, especially so among likely to highlight the turnover level as being small firms… high (and a constraint to business operations) 7 80 low ok \ satisfactory high 6 60 5 4 40 3 2 20 1 0 0 Small Medium Large Small Medium Large Source: World Bank LGSS (2011) Interestingly, few managers report that workers and the Lao garments sector as initial alternatives leave to work in Thailand. This finding is somewhat when seeking employment, rather than switch- surprising given the large number of Lao migrants ing between the two. Certainly, the Lao garment (both legal and illegal) in Thailand and the strong sector attracts fewer migrants from the southern “pull factor� that Thailand is regarded to exert on provinces of Lao PDR (bordering Thailand) than the Lao labour market. A possible explanation from northern provinces. This reflects Lao migra- may be that migrants choose between Thailand tion patterns generally which show migration to Labour standards and productivity in the garments export sector 21 Thailand is strongest from bordering southern have tried recruiting through local authorities or provinces. sending company representatives on recruitment drives, and a small number have tried using the Given these high turnover rates, firms are con- offices of the Ministry of Labour and Social Wel- stantly seeking to recruit new workers using a va- fare (MoLSW) and private recruitment agencies. riety of strategies (Figure 16). The most commonly About half of all firms offer a ‘finder’s fee’ bonus used recruitment strategies are simply through of between 1.25 – 25 USD depending on the skill notices at factory gates and through informal net- level and retention of the new recruit. works of workers’ friends and relatives. Some firms Figure 15: Most commonly reported reasons for Figure 16: Firms employ a variety of different workers leaving the factory (percentage strategies in order to recruit new and replacement frequency by firm size) workers none Large Medium other (please specify) Other 5 to work in thailand Small Ministries (Labor, Educ...) 12 to pursue further education or training Recruitment agency 12 opportunities Company representatives 23 because of family responsibilities (marriage, childbirth, ca Contacts w/local authorities 35 to take another job in another industry or Informal 88 sector to take another job in another garment factory Announcement-ad 89 to set up own business or help family business Source: World Bank LGSS (2011) to return to home community (seasonal migrants) 0 20 40 60 80 100 Source: World Bank LGSS (2011) About half of all firms reported minimum age Firms reported average starting wages varying of recruitment of 18 years or older, while half re- from approximately USD 48 at small firms to USD ported ages between 14-17 years30. 90 percent 60 at large firms. However, workers report some of all firms claim they required some proof of age, firms’ practices of paying reduced wages for up to although several reported difficulties with regis- six months during a probation period, notwith- tration systems (e.g. family book, identity cards) standing Lao Labour Law requirements to pay 90 used by local authorities and 78 percent of re- percent of salary during a probation period not to spondents from large firms believe that, despite exceed 30 days31 . The minimum wage in Lao PDR proof of age requirements, some workers may be (in force at the time of the LGSS) provided for a less than minimum age. Workers in FGDs said that total minimum monthly compensation package documentation was not required from all firms of LAK 569,000, approximately USD 71 in total (Table and that false documents were sometimes used 3). However, this salary structure of basic salary for those which did. plus allowances has apparently been difficult 30 Under Lao Labour Law (Article 41) 14- 17 year olds are permitted to work a maximum of 8 hours per day and are not permitted to work at night (10pm-5am). 31 Lao Labour Law, op. cit., article 27. 22 Labour standards and productivity in the garments export sector to standardize across firms with some firms pay- base rate of LAK 626,400 (approximately 78 USD) ing the full amount in cash and others paying the from January 201232. This compares to minimum various allowances in kind (e.g. allocations of rice wages of USD 75 in Vietnam, USD 68 in Cambodia, instead of the meals allowances etc.). Following USD 61 in Bangladesh and USD 50 in Myanmar (all recent tri-partite negotiations, ALGI expects the figures from ALGI). minimum wage to be fully integrated into a single Table 3: Legal minimum wage in Lao PDR in early 2011 Minimum wage (LAK) 348,000 Monthly allowance: meals plus attendance (LAK 8,500/day) 221,000 Total worker wage per month (26 days/month) 569,000 LAK wage per day 21,885 USD wage per day 2.74 LAK wage per hour (9 hrs/day) 2,432 USD wage per hour (9 hrs/day) 0.30 Source: ALGI Wages for factory workers are calculated differ- celebrations for national holidays. According to ently depending on jobs performed by different workers, managers also make it very difficult for garment workers. Those in general units (packing, workers to resign: withholding partial salary of stock-room, etc.) usually earn a set rate based on new recruits and workers’ final month salary. the minimum wage plus allowances as described above. Those who work in production lines get a “When someone wants to resign, the factory re- salary based on production wages calculated ei- fuses and puts pressure on us - often allocating ther on the basis of their individual production that person to the areas that no-one likes such rates or on the basis of their unit’s production rates. as general section. Some workers could not con- Managers set piece rates and targets depending tinue under these conditions and they decided on the demands of different tasks. Workers then to quit without approval - they just packed their get compensated based on meeting and/or ex- bags and left without any termination pay.� (Lao ceeding these targets with piece rates set based garment factory worker) on levels of difficulty and average time spent on each piece. FGD participants seemed aware that garment fac- tories are facing labour shortages and believed In terms of strategies to improve retention rates, that it is easy for experienced workers to move in many firms report offering annual bonuses of in- and out of jobs. They explained that after with- creasing amounts for numbers of years worked, drawing from the workforce for a short period (3-6 promises to move into higher paying positions months) workers tend to return to the same fac- in the production line and various other informal tory so they don’t have to ‘start all over again’. encouragements such as recreation events and 32 Interview with President of the Association of Lao Garments Industries, Vientiane, 17 Oct. 2011. Labour standards and productivity in the garments export sector 23 7.2 Labour productivity by firm size. While training on ‘merchandizing and production tracking’ were identified as top priori- When large and medium-sized firms were asked to ties for managers across all firm sizes, ‘marketing rate their company’s willingness to make various and sales’ was a top priority for small and medium investments in order to improve labour produc- firms, while large firms also prioritized training in tivity and performance, they expressed strongest social compliance. Medium-sized firms were more interest in the introduction of new production likely to prioritize training on information technol- technology and line management techniques and ogy for their managers. were more sceptical of increasing wages, overtime pay and/or bonuses. However, more than half ex- Although managers identify skills deficits as a ma- pressed some or strong interest in providing more jor contributor to low productivity levels, and de- benefits to workers and this jumped to 80 percent spite the fact that managers had clear ideas about for non-monetary incentives. Over 90 percent of their workforce’s training needs, many managers firms were interested in offering more training and expressed reservations about investing in training career development opportunities to workers. In for workers. When asked to identify the primary FGDs, workers also repeatedly expressed the de- constraint for increasing their firm’s investment in sire for greater training opportunities to improve skills training, half of LGSS respondents cited ‘loss their performance within the garment sector. of return on investment because of high turnover All firms provide some basic initial skills training rates’ and another quarter cited that ‘useful or ap- upon recruitment and new hires are quickly put propriate training was not available’. Interestingly, into production lines to continue skills develop- only four respondents cited insufficient resources ment ‘on-the-job’. Two-thirds of large firms and as the primary constraint to greater investment in about half of medium-sized firms provide workers training. Several managers called for greater pub- with opportunities for limited skills-upgrade train- lic investment in skills development for garment ing mostly related to sewing, marking and cut- workers, some suggesting the establishment of a ting skills, but also for production management, basic skills training centre to provide a skilled pool line balancing and administrative/finance-related of applicants to the industry. However, firms were tasks. more reticent to send their existing workforce ‘off- site’ for skills-upgrading citing opportunity costs When asked to identify priority training needs for and logistics expenses, as well as the fear of losing their firm’s workforce, production skills (cutting, workers to other factories. stitching, etc.) and basic literacy/numeracy were identified as priorities for production workers, 7.3 Labour relations while production management skills (e.g. line bal- ancing) and quality assurance (control, inspection, Overall, managers painted a fairly positive picture etc.) were most commonly identified as top priori- of attitudes and relations in the workplace, some- ties for production supervisors by small and me- what in contrast with workers’ accounts described dium firms. Large firms also identified team man- further in the section 8.3. Almost 90 percent of agement and motivation as a top priority. At the managers surveyed from large firms felt motiva- management level, a diversity of training priorities tion of workers was good, while only half of those was identified and responses varied significantly from small firms and a quarter from medium firms 24 Labour standards and productivity in the garments export sector did. Managers from large firms were also more 7.4 Working and living conditions likely to rate relations between workers and their supervisors/managers as good, than did those While managers of large firms generally rated from small and medium firms. However, these workplace health and safety as ‘good’, managers generally positive responses, especially from large of small and medium-sized firms were more likely firms, seem to mask a more complex reality on the to rate health and safety as ‘ok or satisfactory’. This factory floor, with both current and former work- contrasts with FGD accounts from workers from ers in FGDs describing much more tension and some factories who described their working condi- conflict in the workplace. tions as being very hot and crowded with poor wa- ter and sanitation facilities, and inadequate safety Large firms are twice as likely as small and me- standards (exit doors, protective clothing, etc.). dium firms to report that problems or disputes in the workplace or in the dormitory affect produc- When asked to assess different aspects of work- tion ‘sometimes’. Complaints or dissatisfaction place conditions, there again were variations with pay and compensation, working conditions amongst different groups of managers. The major- as well as with meals and accommodation were ity of all managers assessed quality of lighting as also commonly cited as causes of dispute. In large ’good’; but less than half assessed air temperature firms, disputes are most commonly resolved di- or air quality (exposure to dust, fibres, toxins, etc) rectly between the parties or with the involve- as being ‘good’. Around three-quarters of man- ment of a worker representative; while in small agers from large firms rated access to protective and medium firms there is more commonly also clothing or equipment as ‘good’ while only half of involvement of a supervisor or manager. Almost those from medium firms and a quarter of those no firms reported involvement of an outside arbi- from small firms did. trator or other third party. All large firms, two-thirds of medium firms and half Worker representation varies by firm size with of small firms provide on-site accommodation for nearly all large and medium firms reporting dor- a proportion of their workers. Overall, 54 percent mitory representatives as well as representatives of the Lao garment workforce are living in factory of the Lao Federation of Trade Unions, while work- dormitories, about one-fifth of whom have small ers in small firms were more likely to have informal accommodation fees deducted from their wages. representatives. The Lao Women’s Union was also About one-third of all firms provide other off-site present in just under a third of all firms. A major- accommodation to workers (especially married ity of large and medium firms receive inspections workers or male workers who are ineligible for by the Ministry of Labour and Social Welfare on la- female dorms) and/or accommodation subsidies bour practices, working conditions, social security and allowances. and other matters within the ministry’s mandate. However, many firms complained that labour reg- ulations and social welfare system requirements were often unclear and arbitrarily enforced. Labour standards and productivity in the garments export sector 25 The number of workers reported per dormitory ported providing the following facilities described room varied from 2 to 20, as did the facilities pro- in Figure 17 below. vided in different dormitories, but overall, firms re- Figure 17: Facilities available in garment factory dormitories (percentage of firms reporting) Other Security guards Recreation facilities Cooking facilities Laundry facilities Separate sanitation Shared room Private room 0 25 50 75 100 Source: World Bank LGSS (2011) Very few workers in focus groups and interviews safety, insufficient cooking and laundry facilities. were satisfied with their living conditions in fac- Yet many new migrants found their only source of tory dormitories, complaining in particular of comfort and support from fellow dorm residents unclean and inadequate water and sanitation fa- (discussed further in section 8.4 below). cilities, lack of privacy, lack of ventilation and fire 8. Workers’ perspectives This section presents key concerns and issues 8.1 Recruitment expectations and raised by a cross-section of garment sector work- experiences ers during focus group discussions and individual interviews. While neither statistically representa- In FGDs, workers explained that most dropped tive nor exhaustive, these perspectives offer valu- out of school because their families could not af- able insight into the experiences of workers from ford the cost of education and needed older chil- a range of backgrounds, focusing on their recruit- dren to contribute to household income or family ment expectations and experiences, working and farming activities. living conditions, labour relations, social status and future aspirations. 26 Labour standards and productivity in the garments export sector “All parents want to support their children to at- Some also migrated to ‘see the big city’ and broad- tend school but their economic situation means en their experiences. that children have to drop out of school to help on the family farm. In my family there are six “When I first came here I just wanted to see the children. We all attended primary school when big city. I had no intention of working in a gar- we were too small to work in the farm, but when ment factory... Then I felt I needed to get some we were able to work, we left school. Therefore, money as I wanted to stay permanently in Vien- in my family, none of us have completed prima- tiane, so I applied for a job here.� (Lao garment ry school level.� (Lao garment factory worker). factory worker). Some of their families saw little benefit in continu- Most had friend(s) and/or relative(s) working in the ing education for girls: garment sector (or living in Vientiane with knowl- edge of factory work) who alerted them to some “In my family there are 13 children. My parents of the difficult living and working conditions, yet always said that girls do not need secondary most still had expectations of ‘nicer supervisors, education: learning to read and write is enough more income, less overtime, cleaner and safer for girls but boys should complete secondary working conditions with better accommodation education. Even with secondary education, we facilities’. can’t get good jobs so I left school at Grade 3 and moved with my friends to find work in Vien- For many workers, the reality of factory work is tiane.� (Lao garment factory worker). very different from their expectations: hard work and long hours of compulsory overtime, poor liv- Most took jobs in the garment sector for economic ing conditions, and insufficient income (especially reasons: seeking income as a pathway out of pov- when first starting) to cover basic living expens- erty, to support siblings’ education and to secure es. In particular, demands of industrial working a better future for themselves. Many believed fac- routines and pressure from managers were com- tory work to be preferable to working in rice fields pletely new and very difficult for many. But most or other options in their villages: felt they had little alternative having neither the skills needed for more attractive ‘office’ jobs, nor “I am working in the garment factory because I sufficient capital to start a small business. do not have a job in my village apart from work- ing very hard in the rice fields for my family.� (Lao “For a person who was born to a poor family garment factory worker). then ended up with no qualification and no job opportunity, only garment factory or construc- “I worked in [another] garment factory for one tion worker jobs are available to them.� (Lao year but my family asked me to come back and garment factory worker). help in the rice field. I stayed there three years but it is harder than working in a factory, so For most workers, recruitment is swift: many spoke I returned to Vientiane.� (Lao garment factory of being hired ‘on the spot’ without any complex worker). hiring procedures (beyond submitting proof of age) and little explanation of their contractual rights and obligations. Labour standards and productivity in the garments export sector 27 “When I applied for a job here they asked me “When I first came here I was full of fear, scared only a few questions, no introduction of names, that I would make mistakes then the supervisor no information about work and they did not would shout at me as he did to other workers. even tell me about how much I would get for On my first day of work I saw the supervisor yell- my salary. After they decided to offer jobs [to me ing very loudly at a woman; I was so scared that and the other new recruits] they called the line when he walked past, my hands were shaking. manager to bring us to the workplace. There After that I got more used to the situation and was no time to settle down or prepare - we just working style. and when I could not, I just talked started working immediately. We had to leave to myself or cried alone.� (Lao garment factory our personal belongings at the security box un- worker) til we finished work; later, they took us to the dormitory.� (Lao garment factory worker). According to several FGD participants, rigid work- place rules and harsh management make the Recruitment procedures are more formalized in workers lose loyalty and enthusiasm for their jobs; some larger factories: workers must be 18 years and is a major reason why workers quit. old or above, able to read and write, and provide proof of age and identification (ID card, fam- “When we had a problem in the factory the only ily book, permanent residential certificate) and a solution is to be patient; if we can’t then we quit. letter of guarantee from their village headman. Therefore, when facing this situation, many of the However, most participants in the FGDs reported workers who had never worked in garment work starting to work in the factory while they were 15 before and never had anyone yelling at them like and 16 yrs old and providing falsified documents. this just ran away.� (former Lao garment factory Conditions of employment (factory rules, working worker) hours, salary and overtime requirements) are also more clearly explained in some larger factories. 8.2 Working conditions Workers complained of the practice in some fac- Overall, workers assessed their working conditions tories of withholding partial salary of new workers to be difficult and demanding, and in some cases, as a ‘guarantee’ to prevent them from quitting dur- complained of being subjected to harsh treatment ing first 6 months of employment; and that these by supervisors. However most chose to continue earnings were never re-paid. working in the garment sector in order to earn much needed cash to help their families and im- New workers, especially those migrating from ru- prove their own prospects. Some expressed hope ral areas, have difficulty adapting to a formal work that their siblings and (future) children would not environment and working under strict rules (no have to struggle to earn their incomes under simi- speaking or eating during working hours, fines for lar conditions; and that this was in fact a motiva- being late or absent, etc.), long hours and intense tion for them to continue working in the garment supervision. Therefore a large number chose a fac- sector. tory where they knew someone already working. 28 Labour standards and productivity in the garments export sector Hours Instead, many workers report very low ‘fixed over- Managers report average working hours per time rates’ that do not reflect Labour Law guide- week of between 56.5 (for small firms) and 59.3 lines. They also complain of extreme exhaustion (for large firms), including an average of 9 hours after being pushed to work ‘all night shifts’ and overtime. This finding contrasts significantly with punitive measures against those who do not re- testimonies from focus group discussions and in- port to work the next morning. dividual interviews in which workers complained that overtime is often much longer, poorly paid “Overtime work is compulsory and everyone and often compulsory. has to attend. If we are absent they will deduct a penalty that is larger than OT pay from our sal- Factories in Lao PDR do not appear to operate in ary. …If we have to do overtime from 6pm till shifts. Regular working hours are reportedly simi- 6am, we still have to come back to work again lar across the industry: a regular working day from the next day at 8am - even we did not sleep. If 8-5pm (with one hour for lunch), six days per week we do not attend work that day, the manager with several daily hours of overtime throughout will be very angry - we will receive a warning let- the week and on weekends, particularly during ter and a fine.� (Lao garment factory workers) ‘crunch times’ (such as the months prior to Lao new year when workers generally have one week of While most workers seem happy to work reason- annual leave). Although Lao Labour Law requires able regular overtime to earn extra income, many factories to provide every worker with at least one complain that excessive overtime for little com- day’s rest per week and to seek authorization from pensation drains their energy and motivation and the MoLSW (and approval of the workers’ repre- is often cited as a key factor in the decision to quit. sentative) for workers to undertake more than Those who had a more positive assessment of 45 hours overtime per month (or 3 hours on any their working conditions often referred to a ‘rea- single day)33, this does not seem to be an impedi- sonable amount of overtime’ as a key factor. ment to extending overtime in practice34. The Law also clearly provides that workers are meant to be Benefits compensated at 150 percent for regular overtime; The majority of medium and large firms are en- 200 percent for overtime at night (after 10pm) and rolled in the state’s social security program which at rates of 250 percent and 300 percent for over- provides workers with public health insurance, time at night, on holidays or during weekly rest maternity leave, and pension amongst other ben- days respectively35. However, according to work- efits (see Box 1). However, still roughly half of large ers, these overtime rates are not consistently ap- and medium-sized firms report paying for work- plied in practice. ers’ health care costs directly. Box 1: Lao Social Security System The social security scheme was introduced in 1999 to expand insurance coverage from public sector employees to those working in private and state-owned enterprises. It is a comprehensive package of health care and other benefits including medical care, sick leave, maternity leave, death ben- efits, employment injury or occupational disease benefits, retirement pensions, life insurance, and disability insurance. The scheme is managed by the Social Security Organization (SocSO), a semi- autonomous organization within the Ministry of Labor and Social Welfare. Enrollment in social se- 33 Lao Labour Law, op. cit., articles 18 and 19. 34 It remains unclear whether authorizations are sought and granted by the MoLSW, or whether this requirement is simply not respected. 35 Lao Labour Law, op. cit., article 48. Labour standards and productivity in the garments export sector 29 curity is mandatory for all enterprises with at least 10 employees and funding is generated from a combination of employee and employer contributions: 4.5 percent and 5 percent of employees’ salaries, respectively, up to an income ceiling of LAK 1,500,000. Health insurance, which also re- ceives government subsidies, is the largest of all benefits provided under the social security system and finances outpatient and inpatient care, and prescription drugs through public hospitals. On the whole, workers in FGDs seemed to be un- Workers’ views on their wages varied – with those clear about their entitlements under Lao Labour able to achieve higher earnings through produc- Law (in relation to annual leave, maternity leave, tion ‘piece rate’ system more likely to express satis- pension, etc.) or even under their current con- faction with their salaries. On the lower end of the tracts. They were also very sceptical about the salary scale, and for those with dependants, many benefits of participating in the state’s social securi- report finding it increasingly difficult to cover their ty program. Those who had direct experience with basic expenses due to the rising cost of living. seeking health care using their public health in- surance cards complained that service is poor and Many workers complained that piece rates seemed medicine provided is of cheaper quality than the to be set arbitrarily and some suspected manag- medicine provided to those who pay for services ers of driving these rates down unfairly. In particu- ‘out of pocket’. They also resented the deduction lar, new entrants appear to struggle to understand from their salaries and did not really understand how piece rate coefficients are calculated depend- why they were required to pay monthly insurance ing on the relative difficulty and number of steps contributions even though they ‘never got sick to produce an individual garment. and never used it’. This would suggest the need for strengthening of the health insurance component “Working in the tailoring unit is good because of the social security system and better dissemi- you can get more money if you produce more, nation of information regarding its requirements but it is very hard work... and it is not fair because and benefits. the piece rate is not stable: it is often changed depending on supervisor.� (Lao garment factory Wages worker). Factories report that base monthly wages varied from USD 46 per month for unskilled workers in “I observed that if we are able to produce to tar- small factories to USD 71 per month for skilled or gets for about 2-3 months then the piece rate semi-skilled workers in large factories. However, gets reduced.� (Lao garment factory worker). those working in production lines can earn sig- nificantly more (upwards of double these base 8.3 Labour relations amounts) and most workers want to be in these units. The salary for most production line workers Participants in FGDs from different factories spoke is calculated based on piece rates and targets set about harsh behaviour by supervisors and man- either for individual workers or production units. agers as being the most difficult aspect of their Only a small percentage of small and medium- working conditions. According to one group, the sized firms (12 percent and 16 percent respec- main reason that workers resign is ‘supervisors in- tively) allow workers to earn additional income by sulting them and yelling at them when they made taking extra jobs home. mistakes’. They feel they are being treated as low- skilled workers and not as valuable staff of the fac- tory: 30 Labour standards and productivity in the garments export sector “The most difficult aspect of working in the gar- “I would like the social welfare department ment factory is that you have too many bosses come to check our factory once a month, I want and they criticize you in front of everybody and them come to talk to workers in each line direct- make you feel very bad. Particularly when you ly - interview workers not interview managers, are very tired from work and hoping to earn because they won’t tell the truth.� (Lao garment some cash that you could send to your family… factory worker). then when they shout at you like you are not a human.� (Lao garment factory worker). “I want to attend Labour Law training.� (Lao garment factory worker). There also appear to be cultural clashes between some Lao workers and foreign managers. One unexpected finding from FGDs suggests a slightly different picture from what is reported in “There are many times that foreign staff use im- other countries, with many workers complaining polite and inappropriate words with Lao staff of strict rules and harsh supervision in some larger - humiliating us, scolding us and as a result we export-oriented factories, while those working in lose motivation and spirit while the pressure smaller, Lao-owned operations have a better as- rockets.....� (Lao garment factory supervisor). sessment of their working conditions because of ‘softer’ family-style management. However this Workers feel they have limited options for improv- finding would need to be further explored as ing their situation or persons they can talk to. FGDs included workers from only one small, two medium, and one large factory. “If there will be a counsellor or someone sup- posed to help us when we have problems, I 8.4 Living conditions would like to suggest that person should not be factory staff and should be friendly person... Migrant workers generally depend on factories and it has to be a woman, otherwise it is useless to provide them with accommodation in factory again.�(Lao garment factory worker). dormitories as they cannot afford to pay rent for private accommodation. Many said that they had Many workers’ accounts suggest a lack of effective expected better living conditions when they mi- representation of workers in the garment sector: grated to work in garment factories, although there is variation in the quality of accommodation “In our factory there is a trade union officer who provided by different factories. Workers regularly is supposed to provide support for workers but complain of lack of privacy or secure storage for he cannot help with anything. In the past I used personal belongings, being cramped in hot rooms to report to him about the piece rate having with poor ventilation, no windows, no electrical been reduced, and he just told me: never mind, sockets, no place to eat, insufficient and unclean I will talk to them for you. But he never takes ac- toilets/baths, and insufficient water for drinking tion, I guess he is worried they will cut his salary and bathing. too.� (Lao garment factory worker). “We have to buy our own cooking pot and keep Several others wanted to know more about labour everything in a very tiny bedroom. A few days standards and their enforcement: ago there was a power outage as we were us- Labour standards and productivity in the garments export sector 31 ing too much electricity at the same time. Some women - in my home town talk about garment nights I am too tired to wait for my turn to cook, workers very badly. They had a negative image so I just eat dry instant noodles then go to bed about us: for instance, they said factory girls al- and get stomach aches.� (Lao garment factory ways go with married men and some girls sell worker). sex to earn more money. Actually they know we do not all behave the same.� (Lao garment fac- “Some nights I feel very thirsty but there is no tory worker). drinking water in the dorm so I have to wait till morning to get drinking water at the workplace; Male workers recognized that negative social im- those who have more money can buy drinking ages and stigma attached more to female gar- water… they treated us as we do not have any ment workers than males, but they nonetheless feeling and are not human.� (Lao garment fac- complained that some people (including manag- tory worker). ers) treated garment workers generally as unedu- cated, low-status, rural people. Conditions are particularly bad where dormitories have been set up in old storage areas. Others explained that attitudes in the local com- munity where their factory is situated understand “The dormitory where I stay is three storeys and their situation and are more sympathetic: it is divided into rooms… some rooms have up to 20 people. Before being a dorm, it was a stor- “I heard people selling food in front of the fac- age room: they just put fans in, then put us in it.� tory say that they felt so sorry for factory work- (Lao garment factory worker). ers who have left their families hoping to earn some money and they have to live in poor con- Some managers expressed sympathy with the dif- ditions and eat low-class food�. (Lao garment ficult living conditions of workers in factory dorms factory worker). and made useful suggestions during the LGSS on how these could be improved, including through in- What both male and female workers in FGDs did spection and support from local health authorities, enjoy about their status was the freedom from tra- expansion of water and electricity supply, increased ditional social constraints, to choose friends and space for cooking, eating and recreation, etc. have an income. Yet most workers expressed res- ervation when asked whether they would recom- 8.5 Social status mend garment factory work to their friends and relatives. Many thought they might find the work- Workers reported that attitudes towards them are ing and living conditions too difficult, although it mixed: some in the local community and in their was at least an opportunity to earn income. How- home villages ‘looked down’ upon factory work- ever, they often did not want to see close family ers, but not all. They explained these negative ste- and friends doing such hard jobs or suffering the reotypes as follows: worst of their bad experiences. Those workers who would recommend family/friends take em- “In reality, garment factory girls consist of both ployment at their factory explained that this is ‘be- good and bad but the society just decided that cause it is better than living in the village without we are bad girls. I heard people - both men and a job’. 32 Labour standards and productivity in the garments export sector 8.6 Future aspirations When asked about their future aspirations most hoped to get married and open their own small Most workers began working in the garment sec- business. It seems that participation in the gar- tor with the expectation to be able to put aside ments sector may contribute to delaying age of regular monthly savings. However, those with marriage and first childbirth. Certainly most FGD their own children report having difficulties meet- participants expressed the desire to bear fewer ing basic expenses; while single migrant workers numbers of children (2 or 3) than their mothers generally report trying to put aside about 20 per- had borne, explaining that this was because of the cent of their earnings into savings or to send home, rising cost of living and raising children, in particu- but often find this difficult, particularly when they lar, the need to invest in their education to ensure have been recently hired (and are subject to the better opportunities. withholding of their wages) or earn only basic wages (without additional piece rate pay). Remit- When asked where they saw themselves in five tances are mostly sent home to help cover costs year’s time, most did not expect to be working in of family agricultural inputs or siblings’ education the garment sector, but had only vague plans – and other basic needs. Some workers were hoping often contingent on whether their marital status to save enough to return to their studies (includ- would change. Building savings is their key strat- ing to complete higher education diplomas) but egy for building a better future, however this ap- most were either using their income to cover basic pears to be difficult for many. Most wish they had living expenses for themselves and their families, enough savings to move out of the garment sec- and/or trying to build enough capital to start their tor sooner, explaining that ‘only those who marry, own small business (such as opening a food stand, return to family agriculture or have enough sav- tailoring shop, market stall or other small retail ings can leave permanently’. There was very little business). mention in FGDs of migration to Thailand as an alternative to their current jobs. 9. Implications for improving labour standards and productivity This section identifies some preliminary implica- particularly during the key entry and transition tions at the levels of the firm, the industry and the phase for new workers, greater career continuity public sector in Lao PDR, as well as possible sup- among the workforce (and therefore higher pro- port strategies for development partners. ductivity) as well as to improvements in workers’ well-being. 9.1 Firm level Increased transparency in conditions of em- Findings from the survey and focus group discus- ployment, contractual rights and obligations. sions suggest a number of key areas where chang- While most workers have been informed about es in policy and practices, and/or targeted invest- their basic conditions of employment (hours, pay, ments could lead to higher staff retention rates benefits) they do not always understand their Labour standards and productivity in the garments export sector 33 contractual rights and obligations in any detail, Improved working and living conditions. While nor how to enforce them. Furthermore, they re- conditions vary between firms, there are consis- port that even basic conditions (hours, pay, ben- tent reports of inadequate provision of water and efits) are not always followed in practice and can sanitation, as well as inadequate ventilation and be changed arbitrarily. Therefore firms could take cooling, in both workplaces and dormitories at steps to ensure that conditions of employment are many garment factories. Investments to improve clearly understood by workers, including through these conditions, possibly through innovative increasing the transparency of systems for setting public-private partnerships, are a key strategy for all rates (overtime, piece work, etc.) and any ex- factories wishing to increase worker well-being ceptions to these. and retention rates. 38 Better workflow management. Many workers 9.2 Industry level identify overtime demands as a key area of dissat- isfaction and reason for resigning36. While firms The LGSS found that all large firms and the vast struggle to respond to buyer demands for quick majority (95 percent) of medium firms are mem- turn-around, they risk increasing dissatisfaction bers of the Association of Lao Garment Industries and turn-over in their workforce pushing down (ALGI). When asked what information and support productivity. Firms should be encouraged to anal- they would like to receive from the Association, yse their workflow management practices and im- many firms asked for more market information plement improvements based on best practices and information on new technologies, as well as drawing on experiences from the garment sectors support for labour skills development and inter- in comparator countries. national certification. Improved workplace relations. Workers iden- In addition to its role as industry representative tify both the physical and social aspects of their and interlocutor with other stakeholders on la- working environment as difficult and demanding. bour issues, this study suggests four areas where In particular, this study finds dynamics between ALGI can play a key role in assisting firms improve workers and managers is a key factor in workers’ labour standards and productivity. well-being, productivity and retention rates. While managers indicate that disputes are generally re- Skills development. ALGI recently established a solved directly between the parties; workers ex- Garment Service Centre offering training and con- plain that, in practice, they simply do not have any sultancy services to garment factories, targeting effective means to defend their interests to man- supervisors and middle managers. While many agers or resolve disputes, and therefore, ultimate- LGSS respondents expressed some scepticism as ly resort to simply quitting. Improving workplace to whether the GSC could meet their company’s relations is therefore a key area for further study training needs, the majority wanted to learn more and action for garment firms37. about services and training it offers. 36 Various studies reviewed for a recent ILO/IFC Better Work Discussion Paper find that while certain levels of overtime can be both desirable for workers seeking extra pay and for managers seeking to increasing outputs, excessive overtime can lead to poorer worker health, reduced efficiency, high rates of accidents and worker dissatisfaction, resulting in high worker turnover and lower productivity (Seo 2011). 37 A 2009 study by the IFC and Harvard University provides robust evidence of the positive impact of supervisory skills training on both work- ing conditions and productivity in Cambodia. Analysis of factory records showed that workers whose supervisors received training demon- strated increased productivity and improved relations with their supervisors (IFC 2009). 38 Recent evidence collected in Cambodia suggests that factory investments in improved working conditions do not lead to higher risk of plant closure (as might be suspected in a highly competitive market); but rather, there is some evidence that improvement in standards relating to compensation actually increase plant survival rates (Robertson et.al. 2011). 34 Labour standards and productivity in the garments export sector Sharing best management practices and mini- Labour market information and skills develop- mum standards. ALGI could take a lead role in ment. Both firms and workers suffer from a lack the dissemination of best practices and minimum of regular information on labour markets in the standards for human resources management in garment sector and in the private sector gener- the garment sector in areas such as workflow and ally. The resultant information asymmetries and overtime management, standard contracts, inter- gaps are leading to sub-optimal employment out- national certification, etc. comes that could be improved with public sector support. Skills development is also a key concern Strengthening industrial relations systems. ALGI to both industry and workers that will require in- could play an important role in supporting the tensified partnership between the public and pri- development of effective systems for worker rep- vate sectors. resentation and collective bargaining at firm and industry levels; and support good practice on the Labour standards compliance. Labour standards dissemination of information on worker contrac- compliance in the Lao garment sector (as else- tual rights and obligations, as well as on dispute where) is a key challenge and requires transparent resolution mechanisms. and robust systems for their effective monitoring and enforcement. This study suggests scope for Improving public image of garments work. In co- improvements in these systems in which the Lao ordination with the MoLSW, and based on agree- government has an important role to play. ments for improved monitoring of labour stan- dards and factory working conditions, ALGI could Strengthening industrial relations systems. The drive a public information campaign to counter Lao government should be encouraged to act as a negative stereotypes and improve the image of catalyst and arbiter for the development of a mod- garment factories and workers. ern industrial relations system, adapted to the Lao context, and which draws from the experiences of 9.3 Government level the region and of other transitional economies. When asked what support garment firms would Strengthening social security systems. While like to receive from the Lao Government on labour firms and workers in the garment sector have only issues, the most common requests were for sup- limited experience with the social security system port to increase labour supply and labour skills. that was recently extended from the public to Firms also requested more information and com- the private sector in Lao PDR, initial findings from munication on government regulations from the both the firm survey and focus group discussion Ministry of Labour and Social Welfare (regarding with workers would suggest the need for improve- compliance with labour regulations and social ments to the operation of the system generally, welfare systems) and the Ministry of Industry and and in the provision of quality health care through Commerce (on import-export tariffs and regula- the health insurance system, in particular. tions). Meanwhile, many workers sought more information regarding their rights and obliga- 9.4 Development partners: Better tions under Lao Labour Law. While this study did Factories Lao? not survey public institutions, findings from firm surveys and focus group discussions with workers This section points to opportunities for develop- suggest some key areas for further consideration. ment partners to support and strengthen labour Labour standards and productivity in the garments export sector 35 standards compliance and monitoring systems and monitoring of impacts. building on regional experiences. In particular, it suggests drawing on experiences from the well- A significant proportion of the sector’s workforce established Better Factories Cambodia initiative39 are migrant workers who face many challenges (see Box 2). It also looks to emerging experiences transitioning into the urban industrial sector and from the new Better Work program in Vietnam therefore represent a larger trend in labour force where there are strong similarities to Lao PDR in transition in Lao PDR. Indeed, this is a key transi- terms of the model of economic development (i.e. tion for many developing countries where the transition from centralized planning to market- apparel sector often acts as a gateway into manu- based systems while retaining unitary socialist facturing for workers whose alternatives might be state institutions, including for the representation in agriculture, the informal sector, or low-produc- of workers through centralized trade unions) and tivity service work(Lopez-Acevedo and Robertson under-developed systems for modern industrial 2012). Furthermore, one can point to the particu- relations. lar demographics of this young female workforce who are at a critical life cycle stage for the devel- Making the case that an initiative to improve work- opment of life-long skills and for the inter-gener- ing conditions (both social and economic upgrad- ational transfer of improved human development ing40) is in the interests of Lao garment workers, outcomes (i.e. better health and education out- is relatively easy. When asked why focus on work- comes of children as the result of their mothers’ ers in the garment sector, which represents only delayed age of marriage/childbirth and increased a small portion of the Lao labour market, one can income). Thus, focusing on a sector that employs point to the fact that unlike other broader-based young female workers taps into an important driv- sectors, the garment sector workforce is highly er for long term economic growth and poverty re- concentrated and relatively easy to access; it is duction in Lao PDR (Kuttner 2008). therefore well suited for targeted interventions Box 2: Better Factories Cambodia (BFC) BFC grew out of a Cambodian-US trade agreement promising better US market access in exchange for improved garment industry working conditions (this, in the lead up to the end of preferential treatment that the Cambodian industry had previously enjoyed under the Multi-Fibre Agreement, which was due to expire in 2005). BFC runs a program of unannounced factory visits to check on working conditions using a checklist based on Cambodian labour law and the standards of the ILO, and endorsed by the government as well as by employers and unions involved in the garment industry. To ensure accuracy, workers and management are interviewed separately and confiden- tially. BFC began originally in 2001 as a project supported by ILO and was later joined by the IFC to help expand its compliance-focus to market-driven service provision for the industry. Subsequently more international buyers have been attracted to Cambodia and have been convinced to drop individual auditing requirements, relying instead on the BFC semi-annual synthesis reports acces- sible to all industry stakeholders. Participation is a condition of export licensing for manufacturers and therefore the program includes all exporting garment factories in Cambodia. BFC is guided by a tripartite committee from Cambodian ministries, the Garment Manufacturing Association of Cambodia and the Cambodian union federations. Source: www.betterfactories.org 39 See for example, randomized impact evaluation which found that training of garment supervisors in Cambodia led to higher productivity, lower reject rates, and lower costs (IFC 2009). Employee relations training for garment supervisors improved productivity through improved working climate - especially communication between foreign supervisors and Cambodian workers reducing conflicts. 40 See discussion of distinctions between ‘social and economic upgrading’ and variations amongst types of workers in Barrientos et. al. (2010).. 36 Labour standards and productivity in the garments export sector Making the case to improve labour standards course, also eager to increase the productivity and and productivity through a Better Factories-type profitability of its members. partnership for the Lao garment industry, one can point to the persistent problems of high turnover Finally, supporting the labour intensive manufac- rates and insufficient labour supply to meet de- turing sector is both in keeping with the govern- mand from buyers. While margins are tight, Lao ment’s export diversification and pro-poor growth garment firms are still benefiting from relatively strategy. In making the case that a Better Facto- low labour costs which could be translated into ries pilot be a good idea for the Government of greater productivity through efforts and invest- Lao PDR, one can point to the direct benefits of ments to improve labour skills and labour stan- such a program in the garment sector as well as dards41 as suggested by a recent global study of its demonstration effects to other sectors. The Lao the apparel sector (see Box 3). The industry could government is facing the challenge of a growing also benefit from spill-over from neighbouring labour force with insufficient non-farm labour markets struggling with rising labour costs42. Fur- market opportunities43. Understanding why un- ther, the industry appears to be struggling with der-employed rural workers are having difficulty negative perceptions of some buyers regarding transitioning from agricultural to industrial la- labour standards that could be overcome through bour, and finding solutions jointly with industry a strong system of certification such as through a and workers’ representatives to overcome these Better Factories or similar intervention. Garment challenge through a Better Factories pilot, could manufacturers are also represented by a strong as- provide valuable lessons for strengthening the sociation (ALGI) which has demonstrated commit- Government’s pro-poor growth and employment ment to providing more and better employment generation policies under the country’s 7th NSEDP opportunities in the Lao garment sector and is of (National Socio Economic Development Plan). Box 3: Sewing Success? “Having a vision for the evolution of the textile and apparel sector that incorporates developing worker skills is import... countries that had larger increases in apparel exports [during the post-MFA period] were those that promoted apparel sector upgrading...[Furthermore] in an industry driven by reputation-sensitive buyers in importing countries, concern for labor conditions and worker treatment may be not only a labor rights issue but also a competitive advantage, as the case of Cambodia suggests.� (pp. 3-4). Findings from the a global World Bank study Sewing Success? Employment, Wages and Poverty fol- lowing the End of the Multi-fibre Agreement (Washington, DC, 2012) 41 See for example ‘Productivity in Cambodian Garments Sector: Investing in skills development of workers and also in better industrial rela- tions is key to improving productivity’, BFC Newsletter No. 17, March 2011. 42 ‘Rising labour costs bring benefits’ by Song Huong (April 05, 2011) at http://www.vtgvietnam.com/marketnews_detail.asp?serno=222 (ac- cessed 05/05/2011). 43 See ILO projections that fastest labour force growth is to occur in countries with greatest numbers of people who are poor and the largest informal economies: Lao PDR (25.2 percent), Cambodia (22.1 percent) and Philippines (20.2 percent) followed by Malaysia and Vietnam (19.7 percent and 15.8 percent respectively (ILO 2008). Labour standards and productivity in the garments export sector 37 References   Barrientos, Stephanie, Gary Gereffi, and Arianna Rossi (2010) “Economic and Social Upgrading in Global Production Networks: Developing a Framework for Analysis�, Capturing the Gains working paper 2010/3, www.capturingthegains.org. Davies, Simon and Record, Richard (2010) “Gender and Entrepreneurship in the Lao PDR: Evidence from the 2009 Enterprise Survey�, Investment Climate Assessment background paper, The World Bank Vien- tiane Office. Fontana, Marzia (2009) ‘The gender effects of trade liberalization in developing countries: a review of the literature’ in Bussolo, M. and de Hyos, R. (eds.) Gender Aspects of the Trade and Poverty Nexus: A Macro-Micro Approach, Palgrave Macmillan and World Bank, Washington DC. International Finance Corporation (2009) Supervisory Skills Training in the Cambodian Garment Indus- try: A Randomized Impact Evaluation, IFC Advisory Services in East Asia and the Pacific. International Labour Organization (2007) Global Employment Trends for Women, Geneva. International Labour Organization (2008) Labour and Social Trends in ASEAN 2008: Driving Competi- tiveness and Prosperity with Decent Work., Regional Office for Asia and the Pacific. Kuttner, Stephanie (2008) Economic Empowerment of Adolescent Girls and Young Women: Opportuni- ties and Constraints in Lao PDR, Report prepared for the World Bank Adolescent Girls Initiative, Vientiane Office. Lopez-Acevedo, Gladys and Robertson, Raymond, eds. (2012) Sewing Success? Employment, Wages and Poverty following the End of the Multi-fibre Arrangement, The World Bank, Washington, DC. National Statistics Centre (2006) “Addressing the impacts of the phasing out of textiles and clothing quotas in Lao PDR�, Vientiane, Lao PDR. National Statistics Centre (2005) “Population Census 2005�, Vientiane, Lao PDR, http://www.nsc.gov.la/ PopulationCensus2005.htm. National Statistics Centre (2006) “Addressing the impacts of the phasing out of textiles and clothing quotas in Lao PDR�, Vientiane, Lao PDR. National Statistics Center (2008), Lao Economic and Consumption Survey 2007/08 (LECS4), Vientiane, Lao PDR. Phouxay, Kabmanivanh and Tollefsen, Aina (2011) ‘Rural-Urban Migration, Economic Transition, and Sta- tus of Female Industrial Workers in Lao PDR’ in Population, Space and Place, Vol.17, Issue 5, September/ October. 38 Labour standards and productivity in the garments export sector Record, Richard and Nghardsaysone, Konesawang (2010) “Export Dynamics and Diversification in Lao PDR: An Analysis of the Product Space�, Lao PDR Development Report 2010: Natural Resource Manage- ment for Sustainable Development, background paper, World Bank, Lao PDR. Robertson, Raymond, Dehejia, Rajeev and Brown, Drusilla (2011) “Working Conditions and Factory Sur- vival: Evidence from Better Factories Cambodia�, paper presented at ILO/IFC Better Work conference ‘Workers, Businesses and Government: Understanding Labor Compliance in Global Supply Chains’, Oc- tober 26-28, Washington DC. Seo, Ji-Won (2011) Excessive Overtime, Workers and Productivity: Evidence and Implications for Better Work, Better Work Discussion Paper Series: No. 2, International Labour Organization and International Finance Corporation, Geneva. Staritz, Cornelia (2010) ‘Making the Cut?: Low-Income Countries and the Global Clothing Value Chain in a Post-Crisis World’, World Bank, Washington, DC. United Nations Development Program (2010), Power Voice and Rights: A Turning Point for Gender Equal- ity in Asia and the Pacific, Asia Pacific Human Development Report, Regional Center for Asia Pacific, Co- lombo Office, Sri Lanka. World Bank (2010) ‘Lao PDR Trade Facilitation Performance’, Trade Development Briefing Note No.2, World Bank Vientiane Office, January. World Bank (2011) Lao PDR Investment Climate Assessment – Policies to Promote Growth in the Non Resource Sectors, The World Bank, Washington, D.C. World Bank and Asian Development Bank (forthcoming) Reducing Vulnerabilities and Increasing Op- portunities: A Country Gender Assessment for Lao PDR, World Bank Vientiane Office. Yoon, Youngma (2009) “A Comparative Study on Industrial Relations and Collective Bargaining in East Asian Countries�, ILO Working Paper No. 8, November.   Labour standards and productivity in the garments export sector 39 Annex A: Methods of quantitative data collection and analysis The primary method of quantitative data collection was through a firm survey carried out by Indochina Research (Laos) Ltd. between March and May 2011. This was done under the guidance of the World Bank study team which was also responsible for the survey design and analysis of results. It drew on models of firm surveys previously undertaken in Lao PDR (2009 Enterprise Survey, 2010 Social Health Insurance Survey) and labor-related firm surveys undertaken elsewhere by the World Bank. The Lao Garment Sec- tor Survey covered questions grouped under the following headings: 1. Firm characteristics (ownership structure, export-orientation, years of operation, etc.) 2. Workforce demographics (gender, age, origins, etc.) 3. Labour practices (hours, compensation, benefits, etc.) 4. Human resource management (recruitment, training, retention rates, etc.) 5. Labour relations and standards (health and safety, worker representation, etc.) 6. Labour productivity (rating labour-related constraints, use of performance indicators, etc.) 7. Business prospective (perspectives of managers on industry prospects over short and medium terms) Details of survey implementation are as follows. Population identification Based on data provided by the Association of Lao Garments Industries (ALGI), there were 89 garments firms identified in Lao PDR (all located in the Vientiane area) employing 19,540 reported workers at the time of the survey (March – May 2011). Firm distribution (by size) is illustrated in the chart below. Figure 14: Firm distribution by size 2,500 2,000 Nb of workers 1,500 1,000 500 0 0 20 40 60 80 100 Firms (from smallest to largest) Source: World Bank LGSS (2011) For the purpose of the survey, firms were grouped according to size: 47 small firms (with less than 100 employees), 33 medium firms (with between 100 - 499 employees), and 9 large firms (with 500 or more employees). The following table describes these categories of firms in terms of their share of all Lao gar- 40 Labour standards and productivity in the garments export sector ments firms (share of total firms) as well as their share of the total Lao garments workforce (workforce shares). Thus , as can be seen below, large firms constitute only 10 percent of all firms, but almost half of the total garments sector workforce. Medium firms constitute another 43 percent of the workforce, while small firms constitute only 8 percent of the workforce, although they number more over half of all firms. Actual numbers of firms in each size category are indicated in brackets. Figure 15: Firm and workforce shares Workforce shares (%) 53 49 43 37 10 8 Source: World Bank LGSS (2011) Sampling A stratified random sampling approach was used based on the sampling frame in Table 4 below. Adjust- ing for actual numbers of employees reported in the survey44 (slightly increasing total workforce and causing one firm to ‘graduate’ from medium to large category), and removing firms no longer in exis- tence from the sample, resulted in a sampling rate of 60% overall. Table 4: Sampling distribution Strata45 Number Share of Workforce Number of Sampling of firms total firms share (%) firms rates (%) (%) interviewed Large 9 10.1 49.0 9 100 Medium 33 37.1 42.8 19 57.6 Small 47 52.8 8.2 25 53.2 TOTAL 89 100 100 53 59.6 Response rates Response rates were generally good although there were some difficulties securing interviews due to managers being unavailable. For small and medium firms, this challenge was dealt with by selecting the next random firm from each group. However, non-responses and refusals from managers of small and medium firms did result in slightly lower sampling rates than originally planned for these groups. 44 ALGI figures for number of employee per firm were therefore used for all non-respondents: that is, firms which were not surveyed and three firms from whom survey responses on number of employees were incomplete. 45 This scale of garment firm size was developed specifically for this study of the Lao garments sector to reflect observed segmentation of the industry which in turn allows for analysis of relationships between specific firm characteristics/profiles and labour practices and/or productiv- ity outcomes. This scale differs from the Lao Government’s categorization of firms as micro/small/medium/large as those having respectively, 1-4, 5-19, 20-99,100+ employees. According to this classification almost all Lao garments firms would be considered either medium or large. Labour standards and productivity in the garments export sector 41 Other challenges The complexity and length of this survey made it a challenge (ensuring comprehension of questions and accuracy of responses) for both enumerators and respondents operating in a relatively ‘low capac- ity environment’. There were also challenges in ensuring the accuracy and appropriate translation of the survey and responses from Lao into English. In order to minimize these risks, extensive training of enumerators and piloting of the survey was undertaken under the guidance of the World Bank research team. Method of data analysis After extensive data ‘cleaning’, standard statistical analysis was undertaken using STATA. In the first in- stance, this was mostly tabulations of responses to all questions of the survey, sensitive for differences between small, medium and large-sized firms (with weightings based on sampling rates above).   42 Labour standards and productivity in the garments export sector Annex B: Regression analysis of the determinants of worker departure rates In this sub-section we attempt to investigate causality between firm characteristics and/or practices with turnover rates as a proxy both for firm productivity (assuming high turnover reduces productivity) and for worker well-being (assuming that workers resign to improve their well-being). Although we try to account for possible measurement errors and risks of bias due to sample selection, reverse causality, etc, these results need to be interpreted with care given the small sample size. Nonetheless, they add some weight to the qualitative findings from focus group discussions on the determinants of worker departures. However, there were a number of key issues raised by workers during FGDs that could not be captured in the firm survey data: such as accounts of frequent and excessive overtime (beyond what managers reported), firm practices of withholding wages during probationary periods or as fines for mi- nor infractions (contrary to Lao Labour Law provisions and therefore under-reported), as well as reports of harsh (sometimes abusive) treatment and conflict with supervisors. With these caveats in mind, the firm survey data results suggest that firm size, overtime frequency, working environment conditions, wages or workers’ origins are possible explanations for what is described as one of the main constraints to the garments sector performance, that is, high worker turnover rates. Description of variables (variable name in parenthesis) Our variable of interest is the attrition rate at the firm level. It is calculated as the proportion of workers who left the factory in the year (e_29_512) over the firm’s total workforce (c1_155). We aim to regress this variable on a number of independent variables that we suspect to be possible determinants of turnover rates: Description of variables (variable name in parenthesis) Our variable of interest is the attrition rate at the firm level. It is calculated as the proportion of workers who left the factory in the year (e_29_512) over the firm’s total workforce (c1_155). We aim to regress this variable on a number of independent variables that we suspect to be possible determinants of turnover rates: (i) Firm size: we account for firm size by including size dummies (for small, medium and large) as de- fined in the quantitative analysis. We also account for a firm’s overall performance by including the amount of total sales in USD (b3_63-us). (ii) Earnings: low wages may provide strong incentives for workers to leave the industry. We account for this by including either the monthly average minimum wage for both unskilled and skilled workers (d4_312 and d4_322) or the average starting wage for a new worker (e7_42a). However, this indi- cator needs to be interpreted with caution as the data on wages with various components (base wage, piece rate pay, attendance bonuses, in kind food and accommodation benefits, etc.) were dif- ficult to capture as well as the probable under-reporting of wage reduction/withholding practices. (iii) Working environment: we include a set of dummy variables indicating whether the firm provides var- ious social security benefits (d8_339 to d8_355), transportation benefits (d9_359), meals (d10_364), on-site and/or off-site accommodation (d11_366, d11_371) and whether the firm implements a Labour standards and productivity in the garments export sector 43 human resources plan (e1_400). As an alternative, and because our number of observations is very small, we build up a simple summary indicator ranging from 0 to 1 in which all mentioned variables enter equally. Note however that this does not allow for any consideration of the quality (actual or perceived) of any particular benefit; merely whether it is provided or not. (iv) Overtime work: according to workers responses in the qualitative analysis, OT hours are one of the main factors explaining departure. We account for OT including either the total number of hours worked (d2_302) or the overtime hours only (d2_303). However, this indicator needs to be inter- preted with caution as there were significant discrepancies between the amount and frequency of overtime as reported by workers in FGDs and by managers in the firm survey. (v) Workers’ origin: being a Lao migrant or a foreign migrant may induce workers to leave the factory to return in their home community. Within each firm we compute the fraction of workers being Lao migrants and foreign migrants (based on variables from part C3 of the survey). Empirical strategy We turn to the estimation of the following linear equation: attri = β0+ β1 sales + β2 wageun+ β3 wagessk + β4 over + Xi + zi + ε where: attri is the logarithm of the attrition rate, sales is the logarithm of annual sales in Lao kip (LAK); wageunand wagessk are respectively the monthly minimum average wage in LAK for unskilled and skilled workers; over is the average number of overtime hours; Xiis a set of indicator variables accounting for the working environment; and, zi are time dummies. Note that we cannot use a Probit regression model in that setting: even though our dependent variable is a ratio, it is not necessarily bound between zero and one. Indeed, some firms (in particular small ones) have attrition rates larger than 100 percent in one year. The total number of workers leaving and enter- ing the factory over a year is greater than the number of workers employed at any one time. Hence our use of a linear regression model instead. Possible sources of bias and caveats Sample size: we are able to compute attrition rates for only 50 firms of our sample, as three firms did not provide the required information. Including a wide set of observables restricts our sample. Similarly, full observations are not available for all firms. As a consequence, we don’t expect to find particularly sig- nificant estimates in the following regressions. Even when estimates are indeed statistically significant, their magnitude and subsequent interpretation should be taken with caution. 44 Labour standards and productivity in the garments export sector Sample selection bias: as described in annex A, only a portion of the targeted firms responded to the survey. Non-response may create a sample selection bias. We attempt to account for this possibility us- ing a Heckman sample selection procedure (or Heckit procedure), in which the only variables included in the selection equation are size dummies. Indeed, the quite different sampling rates we had according to our three size categories indicate that size may be what drives mostly selection. Besides, the number of workers is almost the only information we have for non-respondent firms (ALGI). Omitted variables / endogeneity: Many factors which may possibly have an impact on attrition rates and be correlated to some of our independent variables cannot be controlled for: workers’ ability to bargain, overall socio-economic environment, unemployment rate, etc. Reverse causality: high turnover rates may have an impact on a firm’s performance, size and working environment. All these potential biases affect the strength of our results and suggest that conclusions should be drawn with significant care. Table B1: Determinants of attrition rates Dependent variable: log attrition rate OLS (1) (2) (3) (4) (5) (6) (7) (8) Firm’s size -0.309 0.032 (0.165) (0.263) Medium -0.517* -0.266 -0.179 0.368 0.424 0.738* (0.271) (0.316) (0.314) (0.325) (0.363) (0.419) Large -0.508 0.147 0.294 0.904 0.943 1.350** (0.350) (0.559) -0.129* (0.555) (0.559) (0.577) (0.654) Ln sales (LAK) -0.113 (0.076) -0.145* - - -0.162* (0.076) (0.077) 0.173** 0.168** (0.090) (0.078) (0.081) Overtime work 0.044* 0.043* 0.084*** 0.085*** 0.115*** (0.023) (0.026) (0.030) (0.030) (0.036) Ln unskilled wage - - -1.594** 1.377** 1.385** (0.622) (0.599) (0.607) 0.833* Ln skilled wage 0.746* 0.739* (0.436) (0.425) (0.431) -0.077 Work environment -0.034 (0.098) (0.094) Lao migrant (%) -0.007 (0.005) Foreign migrant (%) -0.064 (0.102) Work environment No No No No No No No No dummies N 50 50 50 50 50 41 41 41 R² 0.06 0.08 0.09 0.15 0.18 0.32 0.32 0.37 Note: p<0.10*, p<0.05**, p<0.01*** Estimates for log variables should be taken as elasticites while others give the attrition rate returns to scale. Labour standards and productivity in the garments export sector 45 Results: OLS regressions From the above regressions it appears that the amount of sales (in LAK), the number of overtime work hours and the average minimum wage for both skilled and unskilled workers are the most important factors for attrition rates. Indeed they remain statistically significant and with expected signs to the in- clusion of other controls such as working environment proxy or workers’ origins. Size dummies turn out to have a positive and significant impact only in the last column (8) when workers’ origins are included. Interestingly, while a higher minimum wage for unskilled workers leads to a substantial decrease in attrition (a 1 percent increase in unskilled minimum wage reduces attrition by more than 1 percent in all specifica- tions) as would be expected, we find the reverse effect for skilled workers. Using a summary indice or a set of dummies (not shown in the table) for the work environment never gives a statistically significant impact. Overtime work hours are the most robust determinant across specifications – as expected from the qualitative survey – though with a small magnitude: an additional OT work hour suggests an in- crease in the firm attrition rate by 0.044 - 0.115 percent. In general, implementing the Heckit procedure using firm size in the selection equation changes neither the magnitude nor the significance of most estimates, suggesting that selection is not incidental in that setting. Results: 2SLS regressions OLS estimation does not prevent us from the risk of reverse causality in the model. High attrition rates may impact a firm’s size, economic performance or need for OT work hours. In order to isolate the causal impact of our dependent variable on attrition rates, we need to find a set of robust instruments for all possible determinants. Unfortunately, both sample size and the limited amount of information available prevents us from identifying and using good instruments. Looking at the covariance matrix between our dependent variables and other firm characteristics such as its ownership structure, percentage of output to export, etc., shows almost no statistically significant relationship. The only significant correla- tion (at the 10 percent level) we find relates positively a firm’s size and its number of years of operations in Lao PDR. Furthermore it is unlikely that a firm’s “age� may directly impact attrition rates. We therefore attempt to instrument firm size by age in a 2SLS model: Table B2: Instrumenting firm size by age Dependent variable: log attrition rate 2SLS (IV 1) (IV 2) (IV 3) (IV 4) Firm size -0.913 -0.878 0.278 0.155 (1.399) (1.421) (1.387) (1.065) Ln sales (LAK) 0.098 0.081 -0.135 0.078 (0.317) (0.331)) (0.312) (0.193) Overtime work 0.029 0.082** 0.091* (0.038) (0.039) (0.055) Ln unskilled wage -1.318 -1.272 (0.955) (0.853) Ln skilled wage 0.703 0.655 (0.699) (0.540) Work environment -0.018 (0.136) Lao migrant (%) -0.003 (0.010) Foreign migrant (%) -0.055 46 Labour standards and productivity in the garments export sector Work environment No No No No dummies N 50 50 41 41 R² 0.06 0.08 0.31 0.18 Note: p<0.10*, p<0.05**, p<0.01*** Using such an instrument, we should be aware that we mechanically increase the imprecision of our estimates – interpreting the results with such a small sample is then an even harder task. Thus, unsur- prisingly most of our estimates lose their significance and display large standard errors. The only factor which seems partially robust to this 2SLS strategy is the number of overtime work hours: in the two last columns an extra OT hour increases the attrition ratio by 0.081-0.091 percent, results in the range of those found previously. Results: Looking at the deviation of attrition rates to the sample mean, OLS Instead of looking at the elasticity / returns to scale of attrition rates with respect to certain variables, we look at the impact of the latter determinants on the deviation of attrition rates to the sample mean. In- deed, what we are most interested in is what cause a firm to have a higher rates of departures compared to cover firms, rather than what causes attrition per se. Table B3: Looking at the deviation of attrition rates from the mean Dependent variable: log attrition rate OLS (1) (2) (3) (4) (5) (6) (7) (8) Firm’s size -0.189** -0.104 (0.090) (0.149) Medium -0.323** -0.244 -0.228 0.094 0.108 0.256 (0.147) (0.175) (0.178) (0.148) (0.165) (0.190) Large -0.307 -0.101 -0.074 0.302 0.312 0.524 (0.190) (0.309) (0.315) (0.255) (0.263) (0.297) Ln Sales (LAK) -0.036 -0.032 -0.041 -0.066* -0.065* -0.067* (0.042) (0.043) (0.044) (0.117) (0.037) (0.040) Overtime work 0.008 0.008 0.037*** 0.038*** 0.052*** (0.015) (0.032) (0.013) (0.014) (0.016) Ln unskilled wage -0.663** -0.665** -0.766** (0.273) (0.277) (0.282) Ln skilled wage 0.383* 0.381* 0.420** (0.193) (0.196) (0.198) Work environment -0.008 -0.028 (0.043) (0.044) Labour standards and productivity in the garments export sector 47 Foreign migrant (%) -0.019 (0.046) Work environment No No No No No No No No dummies N 50 50 50 50 50 41 41 41 R² 0.08 0.11 0.12 0.10 0.13 0.17 0.32 0.37 Note: p<0.10*, p<0.05**, p<0.01*** Overall Table B3 gives a rather similar conclusion than previous estimations. In the full specification, OT work hours, total sales and average minimum wage for both unskilled and skilled workers seem to be the best predictors of deviations for attrition rates from the sample mean, with consistent signs. Conclusion Though such a regression analysis is subject to a wide range of bias and caveats that cannot be totally ruled out (small sample size, reverse causality, omitted variables), we can nonetheless gain increased confidence in the finding from worker interviews that, in addition to difficult working condition and relations between workers and supervisors (which were not adequately captured in firm survey data), high overtime frequency and low levels of minimum wage for unskilled workers may be the most im- portant determinants of the observed high turnover rates in the Lao garments sector. 48 Labour standards and productivity in the garments export sector Annex C: Methods of qualitative data collection and analysis Qualitative data was collected using two methods: focus group discussions (FGDs) with workers, and in- dividual in-depth interviews (IDIs) with workers and their supervisors. All FGDs and IDIs were conducted away from the factory and measures were taken to protect the anonymity of participants. All discus- sions and interviews were conducted in Lao language by a moderator using semi-structured guidelines focusing on the following topics: 1. Recruitment experiences and expectations 2. Working and living conditions (including a group ranking exercise) a. Pay and compensation b. Benefits c. Workplace relations and environment d. Living conditions 3. Social status and expectations 4. Personal well-being 5. Aspirations and alternatives Focus group discussions Six FGDs were conducted with workers in groups of 6-8 participants. Four FGDs were held with cur- rent female workers in four purposely selected garment factories to represent a range of firm sizes and ownership structures (one large foreign-owned, one medium foreign-owned, one medium Lao-owned, and one small Lao-owned). Participants were identified by the research firm (i.e. without factory man- agement involvement) in order to represent a range of demographic characteristics (migrants/non-mi- grants, marital status, shorter and longer-term employees, living in factory accommodation or in the lo- cal community). A fifth FGD was conducted with former workers (identified by participants in previous FGDs) and a sixth FGD was conducted with male workers from these same four factories. Prior to con- ducting the FGDs, the research team compiled a matrix of demographic and background information on the participants including: gender, age (current and at time of recruitment), ethnicity, migration and marital status, level of education, household characteristics and length and position of employment in garment sector. FGDs lasted on average between 2.5 and 3 hours. Individual interviews From each of the four factories, two FGD participants (workers) were selected for further individual in-depth interview. One supervisor (head of line) and one dormitory manager was also identified from each of these factories for individual interview. Thus a total of sixteen individual in-depth interviews were conducted. These were used to triangulate the information collected through the surveys with managers and focus group discussions with workers. Interviews lasted on average between 45 minutes and 1.5 hours. Labour standards and productivity in the garments export sector 49 Analysis of FGD and interview data All FGDs and IDIs were recorded and transcripts compiled and reviewed by the lead Lao researcher (World Bank consultant). Summaries of FGDs and IDI transcripts were prepared by the moderator/in- terviewer and these were translated into English. Summaries were then analyzed by the lead interna- tional researcher (World Bank consultant) for patterns and/or recurrence of responses and these were compiled in a document of ‘main messages’. Summaries of groups with particular characteristics (males workers, former workers) as well as individual interviews were reviewed to identify any specific insights or particularities in their responses or contradictions with those of other groups. 50 Labour standards and productivity in the garments export sector