Migrants in Shanghai's manufacturing companies: employment conditions and policy implications

Abstract

The abundance of cheap labour associated with mass migration contributes to the global competitiveness of Chinese manufacturing. Yet migrants are concentrated in low-skilled production line work with inferior working conditions. Drawing on data from fieldwork on 21 manufacturing companies in Shanghai in 2005 and 2006, this article compares working conditions of migrants and local workers with reference to statutory labour regulations. Such comparison throws light on inequality in the workplace. Migrants from both the countryside and other urban areas are studied, which enhances our understanding of the similar institutional arrangements confronted by both rural and urban migrants, as well as the differences resulting from the rural and urban dichotomy. Policy recommendations to improve equality are then outlined.

Keywords:

• migrant workers
• migration
• employment conditions
• China

1. Introduction

China ranks third as a global manufacturing power behind the United States and Japan, producing the largest outputs of at least 100 products, such as toys, shoes, bicycles, TV sets and computers (Commercial Bureau 2006). One of the main contributors to the competitiveness of Chinese manufacturing lies in the abundance of cheap labour associated with the vast migration from rural to urban areas, and from the interior to coastal regions (Tuan and Ng 2001, Chan 2009). The country is experiencing the largest migration in history; over 150 million people, more than 10% of the total population, have been on the move since 2005 (National Statistical Bureau 2007). Most of these migrants seek job opportunities in cities, with over one-fifth employed in the manufacturing sector (Liang and Ma 2004). A 2004 national survey indicated that migrants accounted for 68% of the sector's workforce (Wei 2006). The supply of such a vast number of migrants keeps labour costs down. The average hourly compensation ¹ for a manufacturing worker in China was just US$0.67 in 2004, about 3% of that in the United States (Lett and Banister 2006).

Research has revealed a segmented/divided labour market between migrants and local workers in manufacturing factories (Wang et al. 2002, Tan 2003, Knight and Song 2005). Local workers receive preferential treatment in jobs and resources provided by government. According to the discriminatory labour market policies in the 1990s, jobs such as accountants, clerks and security guards were reserved for local residents to assist their re-employment after the redundancy program in the state sector (Li 2004). As a result, migrants were concentrated in low-skilled manual jobs rejected by local workers. Many migrants suffered inferior working conditions including long hours in dirty, noisy workshops and frequent wage delay or defaulting (Tan 2003, Pun 2005).

Migrants' employment conditions were further deteriorated by the household registration (hukou) system. The system, originated in 1958, is a social control measure, which prevented self-initiated migration by linking individuals' hukou status with legal residence and entitlement to rationed food, jobs, housing, land, and benefits during the central planning era (1958–1978) (Cheng and Selden 1994). As the Chinese economy liberalized after 1978, a large number of people ignored the hukou system and moved to cities. The links between hukou status and food and legal residence were gradually dismantled. However, without local hukou status, migrants were denied access to many employment-based benefits such as a pension and unemployment insurance, because employers were only required to contribute to benefits for people with local hukou status (Wang 2004, Li 2006).

The twenty-first century has witnessed central government's policy shift from regulating migration and restricting migrants' job choices to encouraging migration and improving migrants' situations. The Circular No. 1 issued by the State Council in 2003 ended the limitation of migrants' job choice and required employers to provide migrants with basic social benefits such as a pension and medical insurance (Li 2004). In 2004 two other Circulars (No. 92 and No. 1405) were issued, calling for a simplified procedure for migrants' job application, and provision of training by both receiving and sending governments (Davies and Ramia 2008). Then, in 2005, the State Council required all levels of government to urge employers to pay off wage arrears accumulated before 2003. The 2006 regulations further required companies with records of wage arrears to deposit money in banks as wage insurance before they could legally employ workers (Li and Peng 2007). Because of these recent government initiatives, it is expected that migrants' working conditions might be improved.

This study examines migrants' employment conditions after the policy shift. Much of the existing literature only describes migrants' exploitative work in factories and sweatshops (Chan 2001, Pun 2005, Jacka 2006). By comparing working environments, working hours and pay and benefits of both migrants and local workers, this article throws light on inequality in the workplace. Employment conditions are discussed with reference to statutory regulations. Thus the analysis enables us to examine local implementation of labour regulations and suggest policies to improve equality.

Migrants, in this paper, are defined as people who migrated to Shanghai but whose household registration (hukou) remains at their place of origin. ² Two groups of migrants, rural and urban migrants, are identified, referring to those who came from the countryside and other urban areas, respectively. Without local hukou status, both rural and urban migrants are subject to social and economic constraints. Yet, urban migrants differ from their rural counterparts in that urban migrants have higher educational attainments, a result of the huge and persistent rural/urban divide in China (Knight and Song 1999). By comparing employment conditions of both rural and urban migrants with those of local workers, this analysis enhances our understanding of the similar institutional arrangements confronted by both groups of migrants, as well as the differences resulting from the rural and urban dichotomy.

The following sections begin with an introduction of the data used in this study. This is followed by a discussion of working conditions of migrants and non-migrants, together with the discrepancies between employment practices and labour regulations. The reasons for such discrepancies are subsequently examined. The article concludes with policy recommendations.

2. Data

The data came from fieldwork in Shanghai at the end of 2005 and the beginning of 2006. A mixed method was employed, incorporating a questionnaire survey of both migrant and non-migrant workers in 21 manufacturing companies, site visits, interviews with personnel managers in the surveyed companies and additional 20 interviews with migrants in their residential areas. This article draws data from the survey, supplemented by interviews that provided deeper information illustrating the survey results.

Shanghai is the industrial centre of the Yangtze River Delta, which along with the Pearl River Delta represent the two leading manufacturing regions in China (Chen 2007). Since Deng Xiaoping's southern tour in 1992 confirming the government's determination to liberalize the economy, Shanghai has experienced rapid economic development. This has been commensurate with 2-digit GDP growth rates (SA 2006). At least 5 million migrants from various parts of the country have been attracted to the city, about a quarter of the migrant workers finding jobs in the manufacturing sector (Shanghai Statistics Bureau 2004). The study of employment conditions in Shanghai throws light on migrants' work experience in other big cities in China, especially those in the coastal areas with economic booms. Moreover, as one of the most developed cities in China, Shanghai's experience may portend the patterns emerging in other cities. Thus, findings based on Shanghai may have a leading role in influencing policy initiatives on migration and labour market at the national level.

Supported by the Shanghai Trade Union, the survey was conducted in Minhang and Putuo districts, which have vibrant manufacturing sectors and provide home to many migrant workers. Through a stratified random-sampling method, companies in both districts were first classified by ownership (public, private or foreign). This is because employment practices differ in companies with different ownerships; public companies tend to be regulated by the state, whereas private and foreign ones follow market forces (Hassard et al. 2006). Nine private, ten foreign and two public companies were chosen proportionately, producing products from clothes, furniture and plastic bottles, to electronic goods and machines. The total employees in these companies varied from 27 to 1700.

In each company, respondents were randomly selected from employee lists to complete questionnaires that asked for detailed information on migration and work experiences. In total, 855 valid questionnaires were collected, comprising 527 migrants (61.6%) from both rural and urban areas and 328 (38.4%) local workers. The proportion of migrants is broadly consistent with the percentage of 68% in the 2004 national survey in the manufacturing sector (Wei 2006).

2.1. Background information

The survey data show that 82.8% of migrant workers were between the ages of 15 and 34 years. This is consistent with both Chinese and Western migration literature, which demonstrates that young people are more likely to migrate (Boyle et al. 1998, Hare 1999). Chinese migration literature in the 1980s and early 1990s indicated that men dominated migration, and women were left behind in their places of origin (Cai 1997). Left-behind women in rural areas were seen as a ‘reserve army of labour for agriculture’ (Davin 1999, p. 85). However, an increasing number of women, including many married ones, have migrated (Roberts et al. 2004). The 2000 census revealed that 42.4% of the migrants in Shanghai were women, an increase of 14% on 1993 (Wang et al. 2002, Shanghai Statistics Bureau 2004). In this survey of manufacturing workers, female migrants accounted for 44.8% of the migrant sample; 65.5% of all migrants and 60.2% of female migrants were married. One explanation for an increasing number of female migrants concerns the demand for production line work such as sewing, weaving, electronic assembly and packaging in export processing industries because of China's active participation in global supply chains (Jacka 2006). Women are considered to be better than men and therefore preferred by employers for these jobs that ‘require’ patience, care and nimble fingers (Lee 1998, Piper and Roces 2003, Pun 2005).

Migrants came from 201 different counties in 23 provinces. The largest number is from Anhui Province, accounting for 25.5%. This is followed by Jiangsu Province (21.5%) and Sichuan Province (15.2%). Both Anhui and Jiangsu are geographically close to Shanghai. This spatial pattern is consistent with the literature demonstrating that people migrate to places near their hometowns because of lower migration costs (Shen 1995, Liang 2001). Many people originating from Sichuan migrate to eastern coastal areas, as Sichuan is home to numerous surplus rural labourers, with a low level of economic development (Fan 2004, Liang and Ma 2004).

Research has been undertaken relating migrants' origins with their occupations (Lee 1998, Ma and Xiang 1998, Xiang 2000, Roberts 2001). In Beijing, migrants from Wenzhou county of Zhejiang Province were engaged in the clothing business in the 1990s; those from Henan Province collected garbage; and women migrants from Anhui Province acted as maids (Xiang 2000, Zhang 2001). In South China, studies of factory workers revealed that many migrants from the same provinces clustered in the same production lines; some even shared the same dormitory rooms (Lee 1998, pp. 117–123, Pun 1999). Such job specialization was explained by the finding that migrants searched for jobs through their limited network, which was based on their place of origin (Ma and Xiang 1998, Roberts 2001).

However, no strong evidence supports clusters of migrants from the same origins in this survey. Migrants from the same localities (at both province and county levels) were employed in different companies producing diverse goods; each company employed migrants from various counties and provinces. The lack of migrant clusters could be partly explained by migrants' newer, diverse ways of securing jobs including job agencies and the open labour market. A second explanation concerns the recruitment practices of some companies that did not employ workers through employee referrals, because of extra costs of migrant clustering, such as potential collective actions that are easier to organize among workers from the same locality (Pun 2005, Smith and Pun 2006). This provides evidence of an interesting change of practice in recruitment.

3. Migrants' working conditions

This section provides a profile of migrants' working conditions in Shanghai's manufacturing companies, compared with those of local workers. Topics covered are labour contracts, occupations, working environments, working hours, wages, payment terms and benefits.

3.1. Labour contracts and deposits

The 1995 labour law (Articles 16 and 17) requires both parties to sign labour contracts abiding by Chinese laws and regulations. As contracts seek to aid justice, whether workers have signed a written contract is one of the criteria of distinguishing between formal and informal work (International Labour Office 1972, Bangasser 2000). The informal sector is characterized by prolonged working hours and inferior working environments in developing countries (Nand 2004).

Table 1 lists the percentages of people signing contracts for rural migrants, urban migrants and local workers. The table shows that most workers sign contracts. As research revealed that many migrants worked without contracts in the 1980s and 1990s (Ma and Xiang 1998), this finding in Shanghai could be regarded as evidence of positive effects of the recent policies that required employers to sign contracts with workers. It is also noted that the surveyed companies were registered with the government, therefore more likely to comply with regulations. An interesting finding from Table 1 is the different percentages for workers with different residential status. Local workers had a higher ratio of signing contracts (88.3%) than both urban migrants (80.6%) and rural migrants (73.3%). Moreover, 15.8% of rural migrants and 9.3% of urban migrants signed contracts with agencies as a result of their companies' outsourcing of human resources. Such workers were more likely to receive lower wages than those who signed contracts directly with companies. Rural migrants were at a disadvantage as they had the highest ratios of no contracts and signing contracts with agents.

Table 1. Percentages of recruits signing labour contracts for rural migrants, urban migrants and local workers

	Rural migrants (%)	Urban migrants (%)	Local workers (%)
Contract with companies	57.5	71.3	85.2
Contract with agencies	15.8	9.3	3.1
No contract	26.7	19.4	11.7
Total	100	100	100
Numbers of workers	412	108	324
Source: The Shanghai survey.

Table 2 shows that 1 year is the popular contract period. About 1.1% of urban migrants and 0.7% of rural migrants obtained permanent contracts, far smaller proportions than local workers (13.3%). Urban migrants had contracts lasting longer than rural migrants. Variations existed about respondents' preference of contracted period. Because some companies asked for a fine from those who left during the contracted period, certain workers preferred a short contract to keep flexibility.

Table 2. Tabulation of the length of contracts for rural migrants, urban migrants and local workers who signed contracts

	Rural migrants (%)	Urban migrants (%)	Local workers (%)
Less than 1 year	8.8	5.6	0
1 year	69.2	55.6	60.4
Above 1 year but not permanent	5.0	18.8	16.6
Permanent	0.7	1.1	13.3
No designation of contracted period	16.3	18.9	9.7
Total	100	100	100
Numbers of workers	302	87	286
Source: The Shanghai survey.

Such a fine might take the form of a deposit required from migrants at the beginning of employment. A deposit was also used as a guarantee that migrants would return tools or uniforms when they left their companies. According to labour regulations, employers have no legal right to collect deposits (Yang and Chen 2006). However, migrants in 5 out of 21 companies, about 12.4% of the total migrant respondents, reported that they were obliged to pay deposits. The amount could be as high as a month's salary.

Inconsistency existed in three companies where some migrants were asked to pay deposits, while others were not, even though they did the same job. Further analysis found that those who paid deposits had longer tenures in their companies. This provides evidence for a positive change in employment practices that these companies no longer asked for deposits from newcomers.

3.2. Occupations, working environments and hours

The majority of rural migrants (79.0%) were employed in production line work, compared with 36.7% of urban migrants and 38.2% of local workers. The finding corresponds to the literature demonstrating that rural migrants make up the majority of production line workers (Meng and Zhang 2001, Roberts 2001). Meanwhile, 54.2% of urban migrants and 47.4% of local workers had managerial or professional jobs, compared with only 12.5% of rural migrants.

Urban migrants made up a higher proportion of managerial or professional workers and a lower proportion of production line workers than rural migrants. Since both rural and urban migrants are confronted with the same institutional constraints without local hukou status, such occupational difference reflects their differing capabilities including education. The survey found that urban migrants had an average educational level of 12.5 years, which was significantly higher than the 9.5 years for rural migrants. This educational differential mirrors the huge rural/urban divide in China, with its impact on schooling opportunities (Zhang and Kanbur 2005, Chan 2009).

Significantly different occupations between rural migrants and local workers suggest the important role of local hukou status in securing good jobs. Despite the removal of migrants' job constraints in 2003, migrants continue to be excluded from certain jobs. For example, a personnel manager stated that their company only recruited local workers as accountants, out of habit; most peripheral jobs in factories such as security were conducted by local workers. These findings reflect the ongoing impact of institutional arrangements in the late 1990s when certain jobs were reserved for local workers (Li 2004).

The differing occupational attainment arguably influences working environments and working hours. As noise and dirt are common in workshops, production line workers work in inferior environments. Table 3 shows that lower proportions of rural migrants thought their working environment was clean, quiet, safe or spacious. For example, only 27.4% of rural migrants stated that they worked in a clean environment, while the corresponding percentages for urban migrants and local workers were 46.2% and 51.8%, respectively.

Table 3. Workers' evaluation of their working environments

Workers' perceptions about their working environment	Rural migrants (%)	Urban migrants (%)	Local workers (%)
Clean	27.4 (390)	46.2 (104)	51.8 (309)
Quiet	13.4 (365)	29.1 (103)	35.3 (306)
Safe	43.3 (379)	61.8 (102)	57.8 (306)
Spacious	31.9 (367)	51.5 (101)	47.8 (299)
Note: The numbers in parentheses refer to the numbers of respondents on which the percentages are based.
Source: The Shanghai survey.

It is also not surprising that long working hours are particularly associated with production line work, as line workers frequently face deadlines. The labour law describes normal working hours as 8 hours a day and less than 44 hours a week; overtime should not exceed 36 hours a month (Article 36). However, these regulations were not implemented in all companies, a result consistent with other migrant surveys (Chan 2001, Li 2002, Cooke 2005, Pun 2005). Table 4 shows that rural migrants worked 237 hours per month, 24.7 hours more than urban migrants and 48.6 hours more than local workers. Also, 60% of rural migrants, 34% of urban migrants and 14% of local workers worked more overtime than the statutory 36 hours a month. Furthermore, about 5% of migrants worked extremely long hours, 12 hours a day, 7 days a week.

Table 4. Working hours for rural migrants, urban migrants and local workers

	Rural migrants	Urban migrants	Local workers
Average working hours per month (hours)	237.0	212.3	188.4
Average overtime working hours per month (hours)	61.0	36.3	12.4
Percentage of workers with no overtime work	17.1	34.0	38.4
Percentage of workers with overtime hours exceeding 36 hours per month	60	34	14
Percentage of workers who worked 12 hours a dayand 7 days a week	5.3	5.0	0
Numbers of workers	416	109	327
Source: The Shanghai survey.

It is possible that migrants were willing to work overtime to augment their incomes, since seeking higher incomes was one of their main reasons for migration (Cai and Wang 2003). Another important explanation of long working hours concerns forced work by employers. In three private labour-intensive companies producing clothes and plastic goods, migrants were required to work at weekends, unless they got special permission from supervisors. However, all local workers in these companies, regardless of their jobs, had weekends off as regulated by labour law. This is another example of differential treatment of local workers and migrants.

3.3. Wages and payment terms

Both rural and urban migrants received lower levels of pay than local workers (5.5 yuan per hour for rural migrants, 9.8 yuan for urban migrants and 9.9 yuan for local workers). However, the pay gap between urban migrants and local workers was much narrower than that between rural migrants and local workers.

The Oaxaca–Blinder decomposition method was employed to examine the existence, and the extent, of wage discrimination against migrants by comparing the wage structure of migrants with that of local workers (detailed in Chen and Hoy 2008). The technique identifies the sources of wage differentials as differences in individual characteristics, such as education, together with differences in returns to these characteristics. The differences in returns are described as labour market discrimination, because they denote the pay gap that migrants and local workers received for the same productivity-related characteristics (Gunderson 2006). The results show that both rural and urban migrants would get a 27.9% and 14.9% increase in average pay, respectively, if their characteristics had been evaluated in the same way as local workers. This indicates wage discrimination against both rural and urban migrants, but rural migrants experienced more discrimination in the surveyed manufacturing companies.

Such wage discrimination can be illustrated by the following statement in the questionnaire: ‘we, migrants, are casual workers in the factory. We have 700 yuan each month, while local workers get 880 yuan for exactly the same job’. Clearly, this is against labour law, which supports equal pay for equal work (Article 46). Another explanation for wage discrimination concerns the fact that rural migrants were less represented in managerial or professional jobs and concentrated in low-paid production line work. The decomposition analysis revealed that rural migrants would have received wages 15.3% higher if they had the same occupational distribution as local workers, controlling for education (Chen and Hoy 2008).

The labour law also states that wages should be paid to workers each month on time (Article 50). In contrast to the reports about widespread wage arrears before 2003 (e.g., Li 2002), most respondents in the survey received their salary on time. This could be the result of recent policies calling for timely payment. However, an exception was found in a public company where migrant workers had not been paid for 5 months by the time of the survey. These migrants lived at their workplace and got a minimum subsidy (200 yuan) from the company for survival in unpaid months. The company was about to go bankrupt. This served as the reason for defaulting wages. Yet, local workers there were paid each month on time, although their wages were lower than a year previously, when the company had been solvent. This discrepancy is further evidence of inequality between migrants and local workers.

Overtime is another issue concerning payment terms. The labour law (Article 44) states that overtime should be paid at 150% of basic wages during working days, 200% during weekends and 300% during statutory holidays. Yet its implementation is problematic. Piece rate wage is a common way to circumvent paying overtime wages as the piece rate never changes. According to a survey in private small firms in Guangzhou in 2002, 62% of workers mentioned that they only got a normal rate of pay for overtime work (Cooke 2005, p. 162). In this Shanghai survey, 68.0% of migrants reported that they were paid for overtime work; 26.9% never got overtime pay; and 5.1% got overtime pay under certain circumstances, such as statutory holidays. As piece rates were used in several private companies, it is likely that the above percentage of being paid for overtime work overestimated the percentage of being paid at the statutory overtime rate.

3.4. Benefits and dormitories

Migrants are not entitled to the same social benefits as local residents simply because they do not possess the local hukou status (Li 2006). In the 1980s and 1990s, most migrant workers did not receive any social benefits when working in cities. The Shanghai government took the initiative in 2002 to implement a ‘comprehensive insurance’ scheme, which protects migrants against workplace accidents, serious diseases and old age. Such provision was promoted at the national level in 2003 (Reutersward 2005). However, companies' contribution to a migrant's benefits was much lower than that to a local worker. For example, in Shanghai in 2005, companies were required to pay 176 yuan per month for a migrant, while their average contribution for a local worker reached 685.3 yuan. Local workers' benefits are also more extensive, including compulsory unemployment insurance, medical insurance, a pension and optional housing provident fund.

It is not uncommon for companies to avoid contributing to workers' benefits as a means of reducing costs. However, most local workers got pensions (91%) and medical insurance (88.9%). About 76.5% of local workers obtained unemployment insurance and 57.1% had housing provident fund. As for migrants, 87% obtained comprehensive insurance, while 10.1% did not receive any benefits. Five out of 21 companies contributed to comprehensive insurance only for migrants who had worked there for over 1 year. This leads to the disadvantaged position of newly recruited migrants.

Dormitories, which were provided by 7 out of the 21 companies, could be regarded as an extra benefit to poorly paid migrants. Several workers of the same gender shared a room, and each had a bed. While research in the Pearl River Delta revealed the pervasiveness of dormitory life in manufacturing companies in the 1990s (Lee 1998, Fan 2001, Pun 2005), this Shanghai survey showed that only 7.3% of the migrants lived in dormitories; private renting was instead the common choice. One explanation for this apparent inconsistency concerns the new migration pattern indicating that a growing number of people migrate to join their spouses (Roberts et al. 2004, Connelly et al. 2007). Indeed, three quarters of married migrants in this survey lived with their spouses in the city. These migrants had to or preferred to rent private housing. Company dormitory is argued to be ‘a tool of discipline’, which allows control of workers' life by employers outside of the workplaces (Hershatter 1986, p. 165, Pun 2005). As fewer workers in this sample lived in dormitories, and migrants could choose between dormitory or private renting, such control was reduced.

4. Problems in implementing labour legislations

In China, labour bureaux are the only official institutions in charge of enforcement of labour regulations (Reutersward 2005). They carry out regular checks in randomly selected companies and additional inspections prompted by workers' complaints. In contrast to many industrialized countries where trade unions are important institutions protecting workers' rights through collective action, trade unions in China are not independent (Chen 2003, Cooke 2005). The establishment of a company trade union is encouraged by the government, but subject to approval by employers. Trade union directors are usually appointed by company managers. Collective action is not legally permitted, despite its rapid rise in recent years due to labour disputes (Lee 2007). All these factors weaken unions' role of protecting workers' rights.

The weak implementation of labour law may partly be due to the abundant labour supply associated with vast migration. With a net migration flow of 10 million each year and many potential migrants in the countryside, the urban labour market continues to be a buyers' market (Reutersward 2005). This results in huge competition for jobs and workers' weak bargaining position. Moreover, migrants' disadvantaged positions and their depressed job expectations make them even more vulnerable to exploitation (Knight and Song 2005).

Another interpretation of the weak implementation of labour law concerns local labour bureaux' tolerance of companies' illegal practices, facilitating the investment and tax revenues brought by companies to the local economy. Since the financial and administrative decentralization of the mid-1980s, governments at provincial levels have attained financial autonomy, which allows them to keep revenues after turning in quotas to the central government (Lee 2007). This has provided a huge motivation for developing the local economy. Moreover, GDP has been a major measurement of local governments' performance (Li and Zhou 2005). Therefore, local labour bureaux can overlook employers' breaches of regulations because they put priority on attracting investment. This explanation was confirmed during fieldwork where the local labour bureaux ignored violations of labour law. For example, in a labour-intensive company where migrants worked 7 days a week, the personnel manager made the following statement:

As you know, it is impossible for our workers to work only eight hours a day and five days a week. We have lots of orders. We need our workers to produce, and workers are also willing to work overtime to top up their incomes. The labour bureau officials understand the situation in our company. They told me, ‘your employment practice is not legal. Your workers work too many hours. But, as long as we do not receive any complaints from your workers, your company is fine. But we will take further actions if we get complaints.’

Personnel manager, private company producing plastic products

The above testimony shows that a labour bureau's inspection prompted by complaints could be more efficient. It is therefore crucial for workers to be aware of their working rights and the procedures for seeking assistance. Yet, many workers lacked such knowledge. For example, in the case involving 5-month wage arrears for migrants in the public company described before, the migrants continued working there without pay because they had no idea about any institution which could assist them. Hence, it is important for the labour bureaux to make their services more transparent and accessible.

5. Conclusions

This study has found migrants to be a heterogeneous group. Due to the huge rural–urban divide in China, urban migrants differed from their rural counterparts in that urban migrants had significantly higher educational attainments and more experience of urban life. Urban migrants had a higher probability of signing labour contracts with employers and longer contract duration. They also had better occupations and working environments, and shorter working hours but a higher level of pay. Consequently, the rural/urban dichotomy was reflected in the working conditions of rural and urban migrants.

However, compared with local workers, both urban and rural migrants were at a disadvantage. Local workers were more likely to sign labour contracts with their employers; they were not required to pay deposits or work regularly on weekends; they got higher wages and were paid on time even when their company was in the process of bankruptcy. Migrants received some social benefits, an improvement on the 1980s and 1990s. However, their benefits remained fewer than those for local workers. These findings show that a local hukou continued to influence people's work prospects as local workers receive privileges and protection unavailable to migrants.

Positive effects of recent policies aiming for better treatment of migrants exist, such as provision of social benefits to migrants and fewer cases of defaulting wages. However, migrants, especially rural migrants, continued to suffer exploitation and discrimination in Shanghai's manufacturing companies. The discrepancy between labour regulations and companies' employment practices towards migrants remained evident. Further policies are needed to ensure equal treatment of migrants and local workers.

Equal treatment requires abolishing the privileges associated with the local hukou status. This is because the hukou system continues to support a dualistic social structure, which results in migrants' disadvantaged position, despite changes in the past three decades (Wang 2004, Chan 2009). Apparently, there is inertia in the system, partly because the system benefits both urban residents and employers by protecting locals from fully competing with migrants and by setting migrants' wages lower.

Reform on the hukou system has proven to be a complex and enormous task, as the system is deeply embedded within other systems of resource distribution such as labour, land, social benefits and public services. Hence, further hukou reform requires systematic reforms in all these systems, which would ultimately disconnect hukou status from resource distribution and promote free migration.

In light of this research, two recommendations are made to remove the distortions resulting from the hukou system in the labour market. First, the 2003 policy of abolishing all restrictions of migrants' employment should be implemented consistently. Job advertisements requiring a local hukou, which were found during fieldwork, should be eliminated. Besides the equality argument, this practice would improve efficiency by removing the artificial hukou barrier and promoting job mobility between different occupations, industries and regions.

Second, the same level of social benefits should be available to all workers. Difficulties exist, especially in big cities, which attract millions of migrants from all over the country. Unlike countries where provincial or national governments are responsible for benefit provision, benefits in China are provided locally by city governments (Tao and Xu 2007). Governments in big cities resist providing migrants with benefits and services similar to local residents, partly because of the huge financial costs (Liaowang 2007). A practical way could be to gradually lift the barriers to accessing benefits. For example, migrants who have worked in the city for a certain period should be entitled to a full package of benefits.

Besides reforms on the hukou system, labour market policies should protect migrants' working rights and ensure equal treatment. A new labour contract law was implemented on 1 January 2008, which requires all employers to sign contracts with workers, and demands a substantial fine from those who do not. This law aims particularly to protect migrants, as migrants are more likely to work without contracts. However, as this study suggests, it is not uncommon for companies to violate labour regulations. Hence, labour bureaux should be empowered to make regular inspections in a wider range of companies. It is also important for migrants to be aware of their rights and capable of protecting them.

Urban migrants in this study had significantly better occupational attainment, working conditions and pay levels than rural migrants. This is because of the better education and skill-learning opportunities in cities. The urban-biased policy has, over decades, resulted in a huge rural/urban divide in China, with consequent impact on income, health and education. Hence, policies addressing rural development, especially on investment in schooling and training programmes, are crucial for improving individuals' life chances.

This study focuses on migration and employment conditions in the manufacturing sector in Shanghai, China. Hence, its findings might not be generalized to other migratory populations; its policy recommendations cannot be applied directly to other Asian countries. Indeed, migration and migrant worker issues need to be understood within specific historical, demographic, political and socio-economic contexts. Despite these limitations, the study highlights the significance of migrant-friendly policies in welfare and equality. This is relevant to other Asian countries where labour migration is rising due to rapid urbanization.

Notes

1. The compensation measures the total cost of employing a worker, including wage, training and employer's contribution to benefits.

2. Other names for this group include ‘temporary migrants’, ‘floating population’ (liudong renkou), ‘floaters’, ‘transients’ or ‘population coming from the outside’ (wailai renkou) (Ma and Xiang 1998; Roberts 2001).