The Rise of Ethnicity under China's Market Reforms

Authors


  • The author would like to thank Dr Catherine Ingram for her constructive comments and editorial help. The migration survey in Kaili city was conducted in conjunction with Dr Mark Wang and funded by the Australian Research Council (ARC Project DP0880244).

Abstract

This article investigates the dynamic relationship between economic development and the identification of ethnic minorities and argues that identification of China's ethnic minorities manifests itself at various levels. At the national level, the introduction of market mechanisms and economic growth initiatives have been concentrated predominantly in the coastal areas and metropolises, and are thus increasingly distant from ethnic minorities, a disproportionate majority of which reside in the western parts of the country. This growing regional disparity has placed ethnic regions and populations in a distinctly unfavourable position in terms of economic engagement and development. Regional development in the ethnic-minority homelands has been characterized by the representation and reinvention of ethnic cultural traditions and the production of cultural economies. Unequal economic growth has resulted in a massive migration of ethnic minorities to the cities. Simultaneously, urban development has reinforced ethnic identity, particularly through urban labour-market development. Urban and regional development has, in turn, led to the production, activation and magnification of ethnic identity at individual and group levels.

Introduction

The dynamic relationship between ethnic-minority identity and modernization has been a recurrent theme in social and economic literature. The debate has gained currency as a result of the expansion of global capitalism and the subsequent rise in ethnic conflict and violence (Olzak, 2010). China, which officially recognizes 55 ethnic-minority groups, in addition to the Han Chinese, has also experienced a growth in ethnic unrest, conflict and violence. Violent conflicts between ethnic groups especially between ethnic-minority groups and the Han majority — are increasing rapidly, challenging the nation's social and political order. The most prominent clashes have been those in 2009 between Uyghurs and Han Chinese in Guangdong, and then in Urumqi, Xinjiang, as well as the 2008 clashes between Tibetans and Han in Lhasa. Although ethnic minorities represent a small part of China's population, altogether accounting for only 9.4% of the nation's population in 2005 (NBSC, 2009a), their unique characteristics and strategic geographical positions make them politically, socially and economically important for China's transformation.

Growth in ethnic unrest and conflict in China is directly related to the country's economic development. From 1978 to 2009, the country's gross domestic product grew by 9.8% annually, reaching US$ 5.15 trillion in 2009 (NBSC, 2010a). Such remarkable growth has lifted hundreds of millions of Chinese out of poverty. However, the distribution of economic growth is unequal across regions, notably between the eastern and western parts, and between urban and rural areas. Growing regional disparity paired with urban development has significant implications for ethnic minorities, particularly for the way in which these groups engage with market-economy developments. The changing ways in which minority groups engage with the economy, in turn, reflect the construction of their identity, thus complicating social and economic transformation in China. However, economic literature often portrays China as a homogeneous ethnic state with a few ethnic minorities in the border areas, while the literature on social development assumes that economic reform and development will lead to prosperity for all ethnic groups (IOSC, 2009). The Chinese government (Deng Xiaoping, 1993: 364) and Chinese scholars (such as Zhang and Dong, 2009; Wang, 2010) are strong proponents of this view. Both these groups believe that economic development through market expansion will solve — and is the only solution to (Xia, 1995) — the problem of ethnic inequality, and will eventually lead to ethnic de-differentiation. In other words, they believe that ethnic and cultural differences among ethnic groups will disappear. However, observations in Western contexts point to the opposite: market expansion appears to activate rather than mitigate ethnicity, thereby enhancing existing ethnic differences (Sassen-Koob, 1979; Olzak, 1983) and produces new ethnicities (Chai, 1996). China's approach to economic development has been increasingly similar to that of Western economies, both being characterized by the expansion of capitalism. This raises some questions:

  • In what ways have market reform and resultant economic growth had an impact on ethnic groups, especially on the relationship between Han Chinese and ethnic minorities?
  • How have ethnic minorities reacted to and engaged with market reform in the processes of urban and regional developments?
  • And what are the consequences of such impacts and reactions on ethnic-minority identity?

This article employs a combination of qualitative, quantitative and comparative methodologies to investigate the interactions between ethnic minorities and economic reform at different geographical scales and attempts to answer these questions.

The article begins by reviewing how ethnicity — and especially its constructed aspects — is conceptualized in market economies, and by outlining the development of China's ethnic minorities. It then discusses China's emerging spatial economies over the past three decades, and how these underpin the geographic mismatch between ethnic-minority concentration and regional economic development. The subsequent discussion concentrates on two interrelated questions. The first question is how the relationship between China's ethnic minorities and economic development has recently changed. The analysis highlights the fact that ethnic minorities are becoming increasingly more removed from the country's growing economic opportunities as a result of their geographic and cultural distance and are being discriminated against by labour-market institutions. This leads to the second question: How are these changes reflected in ethnic identity at both individual and group levels? These questions are further examined by means of a case study, which investigates social, economic and spatial production in an ethnic-minority area, Qiandongnan in south-west China, and elucidates how ethnic-minority identity has been redefined, activated and reproduced here. The discussion is further supported by a field survey among migrant workers in the region and a quantitative analysis of the surveyed data to compare the labour-market experiences of migrant workers from different ethnic groups and to estimate the effect of ethnic identity on labour-market outcomes.

Ethnicity, market economies and ethnic minorities in China

People's identity is fundamental to their social behaviour and economic performance (Akerlof and Kranton, 2000; Bauder, 2008). Ethnicity as an important dimension of identity is politically, socially and economically ‘constructed’ (Barth, 1969; Olzak, 1992). It can be self-ascribed or externally imposed, reflecting the salience of race or culture as people define themselves and others in ethnic ways during social and economic transformation. The degree of ethnic intensification is conditional and dependent upon external social, economic and political processes and actors ‘as they shape and reshape ethnic categories and definitions’ (Nagel, 1994: 152) — a process in which ethnic individuals and groups themselves play an active role.

Researchers have identified numerous conditions or ‘sites’ in a market economy that motivate the construction of ethnicity. These include politics, labour markets, residential spaces and urbanization, as well as people's cultural practices and daily experiences (Olzak, 1983; Chai, 1996; Cornell, 2007). Wimmer (2008) contends that ethnic minorities tend to set themselves apart from the majority population group as ‘ethnic others’ when economic and political power differentials between ethnic groups are prominent. This separation leads to the development of new cultural diacritics, which reinforce ‘the taken for grantedness’ of ethnic boundaries and result in ‘further and ongoing cultural differentiation, and so forth’ (Wimmer, 2008: 1002). Padilla (1986) illustrates the emergence of a Latino ethnicity among Mexicans and Puerto Ricans in Chicago as a strategic response to policy change. Olzak (1992) indicates that ethnic status is a key determinant of labour division and that labour-market incorporation can strengthen ethnic boundaries. Ethnicities affect both the supply- and demand-side factors of labour-market segmentation. On the supply side, ethnic factors function as ‘allows or denies’ to manoeuvre within the particular contexts of labour-market segments (Bauder, 2001). On the demand side, an employer's preference to draw employees from the majority and aversion to employing ethnic-minority workers often pushes ethnic labour into the secondary segment of the labour market (Olzak, 1992). As a result, ethnic-minority members face two choices or consequences. One is that ethnic-minority members activate their ethnic identity for labour-market competition. This enhances ethnic boundaries and leads to a rise in ethnicity in the form of ‘ethnic enclaves’ and ethnic residential segregation. The other choice or consequence is that ethnic-minority members soften or erode their ethnic identity so as to be better positioned in the urban labour market. This brings about a decline in ethnic solidarity, ‘yet ethnicity does not disappear’ (Olzak, 1983: 362). Throughout the twentieth century, the two simultaneous processes of the waxing and waning of ethnicity have also been observed in many Western countries.

China's ethnic-minority groups (shaoshu minzu) are indigenous to China, and each group has historically been established in particular regions. However, the ethnic-minority groups as ‘nationalities’ were only recently defined by the Chinese state, and many of these ‘nationalities’ literally did not exist until their creation in the 1950s. In the early years of the People's Republic of China, the Chinese Communist Party launched a project of ethnic identification (minzu shibie). This project led to the official recognition of 55 ethnic-minority nationalities, in addition to the (majority) Han Chinese. The categorization was arguably based on subjective perceptions and on Stalin's definition of nation (common territory, language, economic mode and psychological nature manifested in a common culture) and was implemented by officials and scholars (Yang, 2008). Each group was registered according to individual and group characteristics, emphasizing their cultural differences from the Han Chinese (Zhu and Blachford, 2006). This ethnic recognition allowed for the political and economic construction of minority nationalities, based on the Han nationality as the model of civilization (Heberer, 1989; Harrell, 2001).

While China's ethnic identification project acknowledged the differences between ethnic minorities, its aim was to set ethnic boundaries by targeting specific policies at ethnic minorities. However, the whole of China was soon directed towards a ‘class struggle’ campaign that lasted from the late 1950s to the 1970s. Mao's goal with the ‘class struggle’ was to eliminate class differences and to mobilize the whole population towards a homogeneous communist society. According to the doctrine of Marxism that Mao followed, ethnicity was a class category (Fishman, 1989) and needed to be eradicated. Thus ethnic minorities were forced to discard their linguistic, religious and cultural practices, which ‘were believed to potentially slow the progress of socialism’ (Hoddie and Lou, 2005: 12). Campaigns of ‘class eradication’ became more radical during the Cultural Revolution (1966–76) and had a disastrous effect on ethnic culture. Ethnic traditions were seen as part of the ‘four olds’ (old ideas, customs, culture and habits; in Chinese, sijiu) that had to be destroyed (po sijiu). However, ‘the differences in the nature and boundaries' of China's minority groups were not eliminated (Harrell, 2001: 26).

The utopia of ‘class equality’ was abandoned after Mao's death in 1976, and the economic reform initiated at the end of the 1970s shifted the nation towards a market economy. Policy liberalization and the introduction of market instruments have since released great forces to generate growth in investment, consumption and individual wealth. These developments have been unequal and uneven. Market competition between regions and between people has become a new feature of social and economic development, which highlights the political economy of ethnic identity. Likewise, these forces have drawn isolated ethnic-minority groups into a system of individual and group competition in the context of unequal economic development, creating a new rationale for ethnic identification and leading to a search for ethnic identity. Olzak (1992) suggests that individuals tend to identify with their own ethnic group when members of two ethnic groups encounter one another in a competition situation. Important material and psychological rewards are derived from this affirmation of membership. Thus, the search for ethnic identity among China's ethnic minorities has become an important aspect of their social and economic development under economic reform.

Economic development and ethnicity remaking in China

In 1953, the population of ethnic-minority groups was 35.3 million, accounting for 6.1% of China's total population. This number increased to 40 million by 1964. Between the 1982 and 1990 census, the growth rate of the ethnic minority population was 3.9%, compared to a growth rate of 1.3% for the Han Chinese. Thus, the proportion of ethnic minorities in China's total population rose from 6.6% in 1982 to 8.4% in 1990 and to 9.4% in 2005 (NBSC, 2008). In the 2000 census, 18 of the 55 minority groups had a population of more than one million, compared to only 10 such groups in 1964. This dramatic increase in the ethnic-minority population was a result of differential policy treatment. Such policies have included different family-planning policies: minority members can have two or more children per couple, while the Han (urban residents) are normally allowed to have one child only. Policy that permitted members of ethnic minorities with lower academic requirements to enrol in colleges and universities has also had some influence (Mackerras, 2003). This triggered a phenomenon known as ‘ethnic switching’, with members of ethnic groups renegotiating their membership. Between 1978 and 1990, about 20 million people formally switched from a Han identity to an ethnic minority identity. This accounted for about half of the net increase in the minority population (Zhang and Huang, 1996). The economic development that occurred as a result of market reforms differs remarkably from that which took place during the preceding period of central control. The transformation produced new economic conditions for ethnic minorities to view themselves and to be viewed by others in terms of their economic development and integration.

The focus in this section of the article is on the spatial economic changes as conditions or ‘sites’ for activating ethnicity and magnifying ethnic differences in China.

Central allocation, which dominated social and economic development during the planned period, was determined by government policies. Ideologically, such policies were aimed at removing what was known as the ‘three main differences’ (between industry and agriculture, between manual and mental labour, and between the cities and the countryside) rather than being focused on market efficiency. Spatially, the highest concentration of investment occurred in the central and western areas, aimed at eliminating regional disparity. While disparities between west and east did exist, they gradually declined (Sun, 1983; Ma and Wei, 1997). By contrast, market reforms and related preferential policies after 1978 concentrated on the coastal regions and large metropolises (Vogel, 1989). These included an open-up policy, huge state investment in infrastructure, tax subsidies and the expansion of the foreign and private sectors, all of which are key drivers of China's current remarkable economic growth and regional inequality. For example, almost 90% of inward foreign direct investment in China is concentrated in the coastal areas. In 2009, foreign-investment enterprises contributed to 30% of total employment in Fujian province, 22% in Tianjin city and 18% in Guangdong province (NBSC, 2010b). As a result, economic growth varied greatly at regional and provincial levels. Average annual growth rates range from below 9% in most western provinces (such as Qinghai, Gansu, Ningxia, Tibet and Guizhou provinces) to over 13% for the southern coastal provinces of Zhejiang, Jiangsu, Fujian and Guangdong. The per-capita income gap has also widened. For example, in 2009 the per-capita income in Zhejiang province was 8,805 yuan. This was 3.36 times that of Guizhou (2,475 yuan). However, the per-capita income difference between the two provinces was only 2.35 times in 1990 (NBSC, 2010b).

In terms of ethnic-minority development, China practises a form of so-called ‘regional autonomy’, a practice that has involved the creation of a variety of Ethnic Minority Autonomous Regions (EMARs) in the west, particularly in areas with a high concentration of ethnic minorities. The various designated EMARs include 5 provinces, 30 prefectural regions, 121 counties and a number of townships, which together represent 64% of China's land area. Government subsidies and financial assistance granted to the EMARs are aimed at lessening ethnic inequalities (Maurer-Fazio et al., 2007; IOSC, 2009). However, these benefits have had little effect, given the huge impact of market forces. Gustafsson and Li (2003) reported that the minority–majority gap in average per-capita income in China almost doubled as a result of the coastal preferential policies implemented between 1988 and 1995. This regional disparity has created a distance that has prevented ethnic-minority communities from participating in rapid-growth sectors and has limited benefits from economic growth. This imbalance has caused ‘serious discontent among minorities’ (Zhu and Blachford, 2006: 342).

Lack of capital investment and a desire for establishing ethnic identities have led cities and areas in ethnic regions to resort to cultural assets and resources of ethnic identity for social and economic development. Ethnic regions and cities have thus begun packaging and ‘selling’ their ethnic culture — developing a ‘symbolic economy’ (Zukin, 1995) and an ‘experience economy’ (Pine and Gilmore, 1999). This includes the utilization of ethnic culture and the development of ethnic cultural festivities in urban and regional planning and development, and encompasses art, music, dance, crafts, museums, exhibitions and sports. It is now acknowledged that ethnic tourism has become a major industry in ethnic areas (Oakes and Schein, 2006; Donaldson, 2007). Such developments have promoted economic growth whilst magnifying ethnicity in the ethnic regions, and have made both Han Chinese and ethnic minorities more aware of the differences that separate them (Oakes, 1997; Hasmath, 2008).

Increasing regional disparity has also had a second, more important consequence. A growing number of ethnic-minority people have, voluntarily or reluctantly, been pulled from the west to the east and from rural to urban areas. The urban concentration of ethnic minorities has risen phenomenally, and has been facilitated by a gradual easing of mobility control in terms of the hukou policy, which was a key apparatus of the planned economy. Hukou policy categorized the Chinese population as either rural or urban, and the movement of people between rural and urban areas was strictly controlled. The hukou policy has since been liberalized, resulting in massive rural-to-urban migration and making a huge number of workers available to the urban labour market at low wages. However, hukou-based discrimination against migrants has remained a major obstacle to their full participation in China's urban labour markets (Solinger, 1999; Knight and Song, 2005; Fan, 2008). Despite this, migration from rural to urban areas continues to increase (Wang and Wu, 2010); the rate of ethnic migration was calculated at 8.5% in 2000 (NBSC, 2008), representing nearly ten million ethnic-minority people on the move that year. As Mackerras (2004) notes, this migration is mainly economically driven, in contrast to the politically mobilized migration initiated by the government during the 1950s and 1960s. The ethnic population in Guangdong province, for instance, increased by 583% from 0.18 million in 1982 to 1.23 million in 2000, while the population of Han Chinese grew by less than 60% during the same period (NBSC, 2008). The ethnic population in Shenzhen, a city that was established on the basis of the open-up policy, increased from a few hundred in 1982 to 760,000 in 2010 (Li, 2010). Ethnic minorities are increasingly moving out of their homelands and their distribution has spatially expanded (Zhang and Zeng, 2005). For example, in 2000 the number of Tibetans living outside of Tibet and other Tibetan EMARs (located in Gansu, Qinghai and Sichuan provinces) reached 400,000, representing 7.4% of the total Tibetan population in China an increase of 90,000 from 1990.

While the ethnic population in coastal cities is growing rapidly, an unprecedented level of urban development has been taking place in the EMARs, while in-migration of Han Chinese is growing too. There were only 10 cities in the 5 provincial EMARs (Xinjiang, Inner Mongolia, Ningxia, Tibet and Guangxi) in 1978; by 2008 this figure had increased to 72 (22 in Xinjiang, 20 in Inner Mongolia, 7 in Ningxia, 2 in Tibet and 21 in Guangxi) (NBSC, 2009b). In many of these cities the ethnic population outnumbers the politically and economically dominant Han Chinese. Several important observations can be made regarding the implications of urban development and ethnic migration for ethnic identity. First, the increase in ethnic migration has led to an increase in the level of ethnic consciousness of both ethnic minorities themselves and the urban population, as well as ethnic identification of ethnic-minority migrants. Many ethnic migrants I interviewed especially those who had left their village for the first time — became aware of their ethnic identity only upon moving to cities and coming into contact with other ethnic groups. Secondly, ethnic migration has also increased minority groups' awareness of the economic disadvantages of their home regions and has increased their restlessness. Importantly, urban settings and the development of an urban labour market are key ‘sites’ for the construction of ethnic identity (Amin, 2002). The urban workplace and its employment activities differ vastly from ethnic minorities' social and productive activities in the rural areas. This has led to the intensification of ethnic contacts. This intensification, in turn, leads to more conflict, especially in an unequal situation (Forbes, 2004). There was no urban labour market in China during the centrally planned period. The government provided the urban population with ‘iron rice bowl’ employment (that is, guaranteed lifelong employment and housing). Over 90% of urban employment was within the state sector (including collective ownership), and all urban residents were allocated employment and were paid according to an egalitarian system of wage grades and scales. Labour was not regarded as a commodity. Ethnic differences had little or no meaning in terms of allocation of employment, nor social and economic entitlements such as housing. By contrast, the nation's economy has now rapidly changed from a state monopoly to a multi-ownership market economy. The state-employed workforce has declined remarkably: from 81% in 1995 to less than one third in 2009, with private and foreign investment firms now dominating (NBSC, 2010b). Employers have been granted more leeway in employment practices and in most cases, a labour contract system is in place. Employment is based on the skills, qualifications and social capital of labourers, and employees' identities therefore become manifest. People's memberships and identities have been recounted and categorized to designate ‘insiders’ and to exclude those who are considered ‘outsiders’ (Fan, 2002). Institutions and policies have been developed to restrain ‘outsiders’ from accessing social services and from taking the jobs of ‘insiders’. This re-categorization of differences has led to increased separation and has made identity an economic distinction. The labour-market experience of individuals and groups has become diversified and is now based on their different identities; an obvious stratification has developed, which clearly differentiates farmers, industrial workers and government officials (Solinger, 2006; Wu, 2006; Song et al., 2011).

In a multi-ethnic labour market, discrimination is ‘the rule rather than the exception’ (Lee, 1966: 53), something that is reflected in the development of China's urban labour market too. Discrimination against ethnic minorities has been reported, and there are income differentials between ethnic and Han urban residents (Bovingdon, 2002; Zang, 2010). Hasmath (2008) found that in Beijing only 20% of the ethnic-minority population earned above the average gross income. Compared to similarly qualified Han Chinese in the same occupations, ethnic-minority workers earned less and had fewer career prospects. In a recent case study, Zang (2008) concludes that ethnic status has emerged as a key determinant of labour-market discrimination and is more important than gender in job attainment. In conclusion, market transition has led to marked change in China's cities since Mao's era. It has also aligned them with Western cities in the sense that they are sites for generating, negotiating and contesting ethnic identities (Amin, 2002; Uitermark et al., 2005). Particular ways in which ethnic groups adapt to economic developments have emerged, and these groups contribute to maintaining ‘strong ethnic interaction networks and institutions’ (Olzak, 1983: 357). An example of these developments was the emergence of ‘Xinjiang Village’, an ‘ethnic enclave’ in Beijing (Ma and Xiang, 1998).

A case study

The case study that follows illustrates the resurgence of ethnicity through cultural strategies in regional development, spatial representation of ethnic minorities and differential labour-market outcomes between ethnic minorities and Han Chinese. The region considered in this case study is the Qiandongnan Hmong (Miao) and Kam (Dong) Autonomous Region, a prefecture in the south-east of Guizhou province, in south-west China. The development of the region has typically reflected China's market-transition process: during the planned period it relied heavily on central allocation, but it has since embarked on a journey of ethnic reconstruction, creating a symbolic economy to promote regional development. This study is based on the author's long-time involvement in the region's local and ethnic life, as well as on observations and a recent migration survey in the region.

There are two reasons why the region was selected as a case study. First, while there have been some studies comparing Han Chinese and ethnic minorities in northern China — notably focusing on Hui, Uyghur or Mongolian communities (Zang, 2008; Cao, 2010; Zang, 2010; Han, 2011) — to date there has been no similar research on China's south-western ethnic minorities. There have been few studies of their experience of market transition because they have not drawn attention to their economic disadvantages in the way that their north-western cousins did. Secondly, the region is home to an indigenous population consisting of two large ethnic groups: the Hmong and Kam. About 20% of China's 8.94 million Hmong people and 43% of its 2.96 million Kam people live in the region. The region is also a major source of ethnic migrant workers.

Qiandongnan covers 30,300 square kilometres. When the ethnic identification project was completed in 1956, the region had a multi-ethnic population of 1.92 million, comprising Han (36.1%), Hmong (35.4%), Kam (22.8%) and other ethnic minorities (5.7%) (QBS, 2008). The region's ethnic composition changed little until 1974. Since then, the growth rate of Han Chinese and ethnic-minority groups have started diverging. The Han Chinese group was surpassed by the Hmong in numbers in 1979 and by the Kam in 1989. In 2006, the region had a population of 4.45 million, of which 42.1% were Hmong and 31.9% Kam. Spatially, the Hmong people were mainly concentrated in the north-western part of the region, in Leishan, Taijiang, Majiang, Huangping, Sibin counties and Kaili city. The majority of Kam people were distributed in the southern and eastern parts of the region, particularly in Liping, Rongjiang, Congjiang, Tianzhu and Jinping counties (see Figure 1). Both Hmong and Kam traditional cultures are — like those of many indigenous peoples — strongly attached to the land, but they do differ from each other. As noted elsewhere (Wu, 1992), most Kam people live along river valleys and have adapted their culture to valley environments. By contrast, most Hmong people live in mountainous areas, thus their culture and social organization have been shaped by life in the mountains. At village level, the intermixing of two groups is rare, and individual villages usually consist of only one ethnic-minority group.

Figure 1.

Qiandongnan region and Kaili city

Kaili is the capital of the region and its only city. In 1953, it was a small town with a population of 2,000 Hmong people; it was not until 1984 that Kaili officially received city status. In 2006, its total permanent population was 464,800, ethnic minorities comprising 75.5%. Of these, 64.8% were Hmong people, 24.5% Han Chinese, 2.3% Kam and 8.4% belonged to other ethnic-minority groups (QBS, 2008) (see Table 1). Altogether 186,000 people were residing in the urban area (jie dao). The Kam and Han people who live in Kaili are migrants who are highly urbanized, and most of them live in the city itself. Though small in number, the Han Chinese have dominated the urban development of the city.

Table 1. Population by ethnicity in Kaili city and Qiandongnan region
 Qiandongnan RegionKaili City
Population%Population%
Source: QBS (2008)
Hmong1,871,20042.1301,00064.8
Kam1,416,30031.910,5002.3
Han806,10018.1114,00024.5
Other minorities351,2007.939,0008.4
Total4,444,800100464,500100

Population development in Kaili city has been an interesting process of ethnic and cultural mixing. The first wave of Han Chinese moved to Kaili in the late 1950s, drawn by newly established modern schools, government institutions and modern industries, which accompanied the city's designation as the capital of Qiandongnan prefecture in 1958. Incoming Han Chinese included government officials and industrial workers who were relocating there mostly from Guiyang, the capital city of Guizhou province (shown in Figure 1). Limited numbers of Hmong and Kam people from within the Qiandongnan region were also recruited. This process was accelerated by ‘third-front’ construction (sanxian jianshe), which introduced modern industry to the city for the first time. (During the Cold War, Mao divided China into ‘three fronts’, which roughly encompassed the coastal, middle and western areas; see Naughton, 1988.) Third-front construction occurred in the 1960s and the 1970s and involved the relocation of military-based factories and their workers from major coastal cities to the ‘third front’, namely remote parts of western China. Eleven of these factories, including over 40,000 employees and their dependents from cities such as Beijing, Shanghai, Dalian, Suzhou and Hangzhou, were relocated to Kaili. The project was a key driver of industrialization and modernization in the region. By the end of the 1970s, the relocated workers, with their families, accounted for over 40% of the city's urban residents. However, this development was confined to the city and, more precisely, to the factory areas.

Despite these developments in Kaili city, the Qiandongnan region remained one of the poorest and least developed regions in China. Manufacturing output accounted for 27.9% of the region's GDP, up from 23.7% in 1978 and 8.6% in 1956. Most of the third-front factories were relocated or closed down in the 1990s, as a result of increasing competition from industrial development in coastal areas. Traditional agricultural activities dominated regional economic development. Official data show that the rural population living below the official ‘absolute poverty’ line (668 yuan per year) numbered 390,000, while an additional 729,000 rural people survived on incomes above that but below the official ‘low income’ line (860 yuan per year). Together, these groups accounted for 27.3% of the region's total population in 2006 (QBS, 2008).

Ethnic tourism, characterized by the commodification of ethnic-minority culture, is rapidly developing in this region. Commodification is accomplished through representing, packaging and selling images and costumes, and includes singing and dancing performances by Hmong and Kam people. Ethnic traditions are newly invented and promoted, and village ethnicities are being distinguished. For example, the Kam people's ‘big song’ (dongzu dage in Chinese, or ga lao in Kam), a singing tradition originally found only among Kam villagers in a small, limited area, is now performed in international concert halls in New York, Vienna and Paris and at global and national events like Shanghai's Expo 2010 and The World in Paris 2011. This song genre has also increasingly started appearing in national and regional television programmes, at university events and in school classes.

Kaili has become a major centre for staged big song’ performances. Performance stages are being built, and many other cultural elements of the Kam and Hmong villages have been introduced to support the urban and economic development of the city. Various types of miniature Hmong and Kam villages have emerged in Kaili, which act as both places of business (restaurants, hotels and tourism resorts) and performance stages for ethnic culture. These ‘villages’, labelled either miao xiang (Hmong villages) or dong zhai (Kam villages), attempt to showcase the ethnic cultures of the region in an authentic way (yuan sheng tai) by presenting Hmong and Kam traditions, including foods. The presentation of ethnic cultures in these ethnoscapes is ‘interactive, theatrical, omni-sensory and adaptive to audience reaction’ (Richards and Palmer, 2010: 30), and such presentations have become a business model for, and an important theme of, urban and regional development. The workers (including chefs and serving staff) are from Hmong and Kam villages respectively, wear Hmong and Kam costumes and perform traditional Hmong and Kam songs and dances. The development of these villages is exactly what Amin and Thrift (2002: 125) portray as the ‘experience economy’: ‘a set of living, embodied geographies which provide a new source of value through their performative push’.

The village development has been extended to rural areas through what is known as a ‘gateway village’ strategy. The gateway village for Hmong tourism is Xijiang (Leishan county), and for Kam tourism it is Zhaoxing (Liping county) (see Figure 1). These are said to be the largest villages of the two ethnic groups in the region, respectively. In these villages, houses are redecorated in the traditional style. The aim is to attract tourists to the villages through this, as well as to promote interest in the villagers' daily-life activities. Activities range from day-to-day productive activities to birthday celebrations and weddings, as well as various religious practices, rituals and ceremonies. Traditionally, some of these activities were secret, but are now performed to tourists, though with modifications. Thus, the whole village becomes a stage for ‘performing’ the group's ethnic culture.

These developments are reinforcing the ethnic consciousness of the members of different ethnic groups within the region, creating a sense of otherness for observers from the outside. They lead to a new form of regionalism that distinguishes geographical spaces based on regional ethnicity and, in turn, to the production of ethnic identity. For example, the Kam language includes many different lects within two main subdialects, and ‘big song’ was originally only sung within certain villages in Liping, Congjiang, Rongjiang and Sanjiang counties within southern subdialect-speaking communities who shared a common lect. These innovations and expansions have changed the ‘big song’ into a ‘collective representation’ (Durkheim, 1965 [1912]: 471) for the entire Kam people. It has become what Fredrik Barth (1969) terms an ‘identifying marker’ — a phrase which he uses to describe ‘overt signals or signs — the diacritical features that people look for and exhibit to show identity’ (ibid.: 14). Through this, the Kam group's collective identity has been reproduced.

Like elsewhere in rural China, labour emigration (laowu shuchu) is crucial to the livelihoods of rural households and is used as a strategy for regional development in the area. Large-scale labour migration commenced in the mid-1980s and has rapidly intensified, with a 9.1% annual growth rate in the region since 2000. In 2006, 43.7% of Qiandongnan's 2.03 million-strong rural labour force moved elsewhere to seek employment (QBS, 2008). Wage remittances from migrant workers accounted for an average of 38.8% of total household income in 2006, exceeding the average income from agricultural activities by 4%. Because of the limited capacity of the urban sectors in Kaili city, 88.5% of the region's migrants moved to the large labour markets outside the region (QBS, 2008). The emigration of ethnic people within this labour force brings money, innovation and knowledge back to their home regions. However, the incorporation of these workers into the urban social and economic system has emphasized their ethnic identity, affecting both the host society and the workers themselves.

In order to examine the impact of market transition on ethnic labour and, in turn, how workers' ethnicity has been activated, a questionnaire survey was conducted in Kaili city between 22 October and 26 November 2008. The survey focused on ethnic migrant workers, who face a double set of disadvantages: first, the disadvantages associated with being in possession of a rural hukou; and secondly, the disadvantages associated with the workers' minority identity. It attempted to collect data to answer two questions: whether or not a worker's ethnicity affects his or her labour-market engagement, and whether or not the development of the urban labour market discriminates against ethnic labourers. Surveys included detailed information on the jobs each person had held in the cities since leaving their home village, such as location, job description, duration, income and related training and welfare benefits, along with the respondents' ethnicity characteristics, such as home language, motivation for migration and episodes of discrimination.

In China, migrant workers need to register with the local authorities in order to obtain work permits and temporary residence permits. A total of 29,200 migrant workers were registered with the Kaili Bureau of Public Security in 2007. However, this record was reliable only at the time the migrant workers registered. They could leave the city at any time without reporting to the bureau.

Migrant workers in Kaili could not be approached through traditional household-based sampling methods such as a random sampling, as only a small proportion lived within the urban-neighbourhood communities, in contrast to the situation in other Chinese cities. Many lived either in factory dormitories, at the back of the restaurants or on the construction sites where they were working, or in the surrounding rural areas. For this reason, an innovative sampling strategy that combined stratified sampling with judgement sampling was employed to access the desired information. A stratified sample of 22 elementary units was initially obtained from an inventory provided by the Qiandongnan Bureau of Statistics, which listed all businesses and industries in Kaili. This was necessary to ensure the geographic and industrial representativeness of the sample distribution. As some migrant workers did not work in the listed businesses and organizations, the assistance of local people who knew the area was enlisted to distribute questionnaires within stratified samples and other informal sectors.

A total of 350 survey questionnaires were distributed, with the intention of covering 1% of migrant workers in the city of Kaili. The ethnicity of the migrant workers was not considered during sample distribution, which ensured that the original survey was unbiased in terms of industrial distribution, obtainment of occupation, income and ethnic identity. A total of 308 valid samples were collected, covering jobs from a wide range of industries and organizations in the city. This range was considered important: as Becker (1957) and others have noted, in competitive labour markets, ethnic discrimination by employers, workers and customers does not result in wage differentials for ethnic workers in identical jobs, but rather in segregation of workers according to different jobs.

Table 2 summarizes the survey samples, drawing on the responses by 164 male and 144 female respondents. Ethnic minority migrant workers (EMMWs) accounted for 63.1% of total respondents, with Han migrant workers (HMWs) representing the remaining 36.9%. Of the EMMWs, 30.5% belonged to the Hmong minority, 27.3% were Kam and 5.5% belonged to other ethnic-minority groups. There was no notable difference in average age, industrial distribution, migration experiences and job mobility in the three groups of EMMWs in the city. On average, the migrant workers surveyed had spent 8.6 years working in the urban labour market and had been employed in 3.3 different jobs. However, it is worth noting that 43% of the respondents had spent a period of migration elsewhere, and that 57% of the total number of jobs in which respondents had worked had been in locations outside Kaili city, mainly in Guangdong, Jiangsu, Zhejiang, Shanghai and Guangxi. So the surveyed migration experiences also reflected the urban labour-market condition of China in general.

Table 2. Profile of surveyed migrant workers
 HanHmongKam
  1. a5.5% of migrants come from other ethnic-minority groups
Source: author's research
Ethnic structure (%)a36.930.527.3
Male (%)54.945.751.2
Female (%)45.155.348.8
Average age31.131.830.3
Average length of migration experiences (years)8.88.48.5
Average job mobility (no. of jobs held)3.53.23.3
Mandarin competency (self-assessed, 1 to 10)5.935.525.84
Education   
• Illiterate and primary school (<6 years of formal education)14.214.921.4
• Junior high school (7–9 years of formal education)44.258.558.3
• Senior high school (10–12 years of formal education)34.521.316.7
• College and above (>12 years of formal education)7.15.33.6

Table 3 reports the average monthly incomes of migrant workers over three time periods in accordance with the process of labour-market reform. It shows that there was little difference between the wages of the three groups before 1996. This is similar to the findings from a household survey in Guizhou at that time: as Gustafsson and Li (2003: 818) comment, ‘there was scarcely any minority–majority income gap’. The average monthly wage of HMWs increased by 131.4% from the first period to the period between 2002 and 2008, far exceeding the wage increase figures of 80.7% and 95.2% for Hmong and Kam workers, respectively. This can be attributed to the introduction of many ‘tough’ market policies in the mid-1990s and especially after the 15th Party Congress in September 1997 (Solinger, 2002). State-owned enterprises (SOEs) were granted the authority to dismiss labourers who had joined them before the labour contract system was instituted in 1986. These policies, and the resulting industrial restructuring, led to over 45 million workers being laid off (xia gang), who thus joined the migrant workforce and competed with migrant workers. It is reported that ethnic workers were more likely than their Han counterparts to be laid off when an SOE downsized its workforce (Becquelin, 2000). Majority SOEs completed their restructuring through the reduction of their labour forces in 2001 (Wong and Ngok, 2006). In the period from 2002 to 2008, Hmong and Kam migrant workers respectively earned only 76.6% and 83.5% of HMWs incomes. This shows that income differentials between Han and ethnic-minority migrant workers widened in the wake of China's market reform. The income gap between Hmong and Kam migrant workers increased too, but not to the same extent. This suggests that the ethnic identity of labourers had become a factor that widened the labour-market outcome.

Table 3. Change in monthly income over time, and income differentials by ethnicity
 Pre-19961996–20012002–08
Source: author's research
Han423.6817.4979.6
Kam419.0725.4818.1
Hmong415.4699.7750.4

Furthermore, to establish the effect of ethnic status on the income of migrant workers, regression analyses were conducted to estimate the effect of ethnicity on labour-market outcome. The average monthly income (avminc) of the surveyed migrant workers' first jobs in the urban labour market was selected as the dependent variable. The model specification was based on discussions on ethnic labour-market segments and the development of China's urban labour market in previous sections, bearing in mind the potential problems in analysing data on the earnings impact of ethnic identity that may be attributed to a wide range of exogenous variables. Based on this, the changing avminc is assumed to be a function of three factor sets: migrant workers' human capital, geographic factors and the ethnic characteristics of the migrant worker. The latter include level of education, gender, ethnic identity (Ethnicity I) and language competency (Mandarin). Members of ethnic-minority groups and Han Chinese in China are not always easy to distinguish by physical appearance only. So a further ethnic variable (Ethnicity II) is introduced, namely the ethnic identity of migrant workers as known by their employers. In terms of Ethnicity II, 1 denotes migrant workers who reported that their employer knew her or his ethnic minority status; otherwise it is shown as 0. A worker's ethnic identity might be known for various reasons and is assumed to be a disadvantage in terms of labour-market engagement. For job location, 1 denotes jobs held in Kaili city, while 0 represents employment outside the Qiandongnan region.

Pearson correlation and a scatter-plot matrix of ethnic variables were done to detect the possible presence of heteroscedasticity and effects of nonlinearity. Results indicated that reasonable linear fits could be obtained by regressing avminc against the ethnic variables individually. The null hypothesis of no heteroscedasticity was further tested by a White's test, carried out by running an OLS regression. It could not be rejected at the 5% level of significance. Four individual regressions were run based on the jobs migrant workers had held in the three periods: pre-1996, between 1996 and 2001 and between 2002 and 2008.

The regression results are reported in Table 4. For all models, the joint effects of the selected explanatory variables on avminc were significant at the 5% significance level, but the effects of some individual variables were not significant. The largest coefficient is −0.286 (model 1) for job location. Regional income inequality continued driving ethnic-minority people in the region towards emigration. This result confirms that increasing inequality in regional development is the most crucial factor affecting emerging ethnicity in China, and specifically the ethnically based income gap (Hannum and Xie, 1998; Gustafsson and Li, 2003). Variables of Ethnicity I, Ethnicity II and Mandarin competency have the expected signs, and the coefficients are statistically significant in the regressions. These indicate that ethnic identity has emerged as a significant negative factor affecting migrant workers' wage level. The variable Ethnicity II has a negative sign and high coefficient (model 1, Beta = −0.220), which is second only to the effects of geographic factors. Ethnicity I supports the assumption; it has a positive sign in the regression in the period prior to 1996 (model 2, Beta = 0.132) and is negative for the market transition period between 1996 and 2001 (model 3, Beta = −0.336) and between 2002 and 2008 (model 4, Beta = −0.186). This result indicates that wage discrimination against ethnic-minority workers has emerged in the urban labour market.

Table 4. Ethnic income effects and market transition: repression results
 Model 1Model 2Model 3Model 4
Source: author's calculations
Gender (male)0.0590.043.0860.094
Educational attainment0.1590.1580.1250.105
Job location (Kaili)−0.268−0.298−0.311−0.348
Ethnicity I−0.1300.132−0.336−0.186
Ethnicity II−0.220−0.151−0.134−0.331
Mandarin competency0.1160.1090.0940.113
R20.3480.4430.3910.378
F17.76.910.312.3
N30864101143

Discrimination has increasingly pushed urban ethnic migrants into self-employment, mainly the selling of traditional arts, crafts and foods. Although the surveyed data did not specify jobs in this sector, ethnic commercial streets (minzu jie) and ethnic marketplaces (minzu jiaoyi shichang) that are characterized by ethnic goods and vendors are emerging as a distinct characteristic of the city. In fact, the level of self-employment of ethnic migrants has not only grown in ethnic cities such as Kaili, but also in other major metropolises. Wang et al. (2011) report that Shanghai had 15,200 ethnic migrants in 2009 (12% were Hmong and 4.4% Kam). They accounted for 2.8% of the city's 5.42 million migrant population, that is, those who lived in the city for longer than six months without a Shanghai hukou in that year. Of the ethnic migrants, 31.4% were self-employed (for example, selling silver jewellery for the Hmong and shish-kebab for the Uyghur). The high presence of self-employed migrants and their businesses has become an integral part of China's urban landscape and plays a significant role in the negotiation of their ethnic identity and representation (Baranovitch, 2003).

Conclusion

Ethnic identity and its impact on the economic development of ethnic minorities has always been an issue of scientific and political concern. In the past three decades, China has experienced exceptional economic growth under its economic reforms, such as the expansion of market institutions. These reforms have transformed China fundamentally from a socialist into an economically stratified society consisting of advantaged and disadvantaged groups. In this article, I have explored the ethnic impact and consequences of China's transition to a market economy, and have argued that economic development has underlined the political economy of ethnic identity. Ethnicity, which was relatively hidden for three decades under socialist development, has risen, either by being adopted or imposed, in both urban and regional developments in China.

The expansion of market institutions under China's economic reform policies has increased regional inequality while encouraging competition for social and economic opportunities at individual and group as well as regional levels. This has led to ethnic identification at various levels. At the national level, unequal economic developments have led to policy efforts and investments being concentrated primarily in the coastal areas and large metropolises, while becoming increasingly distant from ethnic minorities, a disproportionate number of whom reside in the western parts of the country. This has placed ethnic regions and its people at a clear disadvantage in terms of economic incorporation and development. In the homelands of ethnic-minority groups, regional development has been characterized by the representation and reinvention of ethnic cultural traditions and the production of cultural and experience economies. This regional development, in turn, has led to the production, activation and magnification of ethnic identification at both individual and group levels, a development exemplified by the development of ethnic tourism and the experience economy, and the ethnic construction of the Hmong and Kam groups in the Qiandongnan region. The ‘big song’ tradition has been used as a new ‘identifying marker’ to unify and amplify Kam people's social, economic and cultural identity, and has made the Kam region ethnically prominent.

Unequal economic growth has resulted in a massive increase in ethnic migration to the cities. The urban concentration of ethnic groups from different cultures, and especially their competition in the urban labour market, has had consequences for urban development. It has also highlighted the cultural identity of ethnic minorities, increasing the ethnic awareness of the Han majority, and making the ethnic minorities more self-consciousness. This development mirrors ethnic developments in Western countries, where the city is the key ‘site’ for ethnic identity negotiation, contesting and remaking. The disadvantages and advantages of being identified as a member of an ethnic-minority group have intensified and have been reinforced by market developments. The case study on ethnic migrants showed that discrimination against members of ethnic-minority groups in terms of urban labour-market outcomes is growing.

The resurgence of ethnicity challenges both urban development and ethnic policies in China. The country is increasingly arranging the production of urban spaces for market-economy success, while seldom considering the interests and culture of the different populations, especially those of ethnic-minority groups. In addition, ethnic policy, which has been central to the EMARs, has failed to facilitate the social and economic development of ethnic minorities in a meaningful way, as an increasing number of them are migrating out of the EMARs. Ethnic development, identification and contestation are now occurring, which are concentrated in the urban areas both within and outside the EMARs.

Ancillary