The Elite, the Natives, and the Outsiders: Migration and Labor Market Segmentation in Urban China Author(s):

C. Cindy Fan Source: Annals of the Association of American Geographers, Vol. 92, No. 1 (Mar., 2002), pp. 103-124 Published by: Taylor & Francis, Ltd. on behalf of the Association of American Geographers Stable URL: http://www.jstor.org/stable/1515270 Accessed: 21/05/2010 11:11
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=taylorfrancis. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Association of American Geographers and Taylor & Francis, Ltd. are collaborating with JSTOR to digitize, preserve and extend access to Annals of the Association of American Geographers.

http://www.jstor.org

The

Elite,

the and in

Natives, Labor Urban Market

and

the

Outsiders:

Migration

Segmentation

China

C. Cindy

Fan

Department of Geography, University of California, Los Angeles

Established migration theories are mostly based on capitalist market economies and downplay the role of institu? tions in internal migration and labor market processes. In socialist and transitional economies such as those in Russia and China, however, investigations of migration and the labor market must begin by examining the nature and con? sequences of state institutions. In this article, I argue that the migration and labor market processes in Chinese cities are deeply influenced by an institution-based opportunity structure. The household registration (hukou) system, in particular, is interwoven with distribution of services and job opportunities. Most peasants who enter cities in re? sponse to increased demands for cheap labor are not granted urban citizenship and are treated as "outsiders" to the urban society. The experiences of these "temporary migrants" contrast with those of "permanent migrants" who are state-sponsored or have access to institutional resources. Using qualitative accounts from a 1995 village-level survey in Sichuan and Anhui and quantitative data from a survey I conducted in Guangzhou in 1998, this article examines the most salient differences among the three subpopulations with different resident statuses: nonmigrant urban na? tives, permanent migrants, and temporary migrants. I show that resident status is central to explaining migration processes and labor market segmentation in the Chinese city. The findings indicate that in terms of human capital attributes, mobility resources, and labor market entry and shifts, permanent migrants are the most privileged and successful elite, followed by nonmigrant natives, and finally by temporary migrants at the bottom of the hierarchy. These results hint at a new social order of stratification in Chinese cities, underscore the compelling relations be? tween internal migration and labor market development in transitional economies, and suggest that in these econ? omies the state deepens the bifurcation effects about which labor market segmentation theory is concerned. Key Words: China, institutions, labor market, migration, transition.

that address the relations approaches internal migration and the labor market are primarily based on capitalist market economies. Theoretical between and labor market segmentation Neoclassical, behavioral, theories can partially explain in transitional changes as they fall increasingly under the forces of economies, and rural-urban urbanization, globalization, migration. these theories do not emphasize state institu? However, tions, which happen to be central to spatial and social processes in socialist and transitional researchers write about institutions, concerned less tangible with international institutional internal studies economies. When they are primarily or migration, immigration,

other hand, market forces are exerting for strong "pull" peasants to migrate to cities and to accelerate the development of an urban labor market (Cao ments; 1995; Knight and Song 1995; Yang and Guo 1996; Cook and Maurer-Fazio 1999). By the 1990s, estimates ofthe "floating population"?rural people leaving the countryin the range of 1999a, 18). The vastly differing fates among migrants and between peasant migrants and urbanites epitomize the awkwardness of this transitional system. "Permanent migrants" are sponsored by the state 100 million (Solinger and are selective and privileged; compared with them, "temporary migrants" are "outside of state plan," on their are out? own, and inferior. Migrants' varied opportunities comes of differentials in human capital, but above all they are products of socialist policies such as the house? hold registration and segmented tional barriers (hukou) system. The young, distorted, labor market fully reflects the institu? side in search of work elsewhere?were

on the

factors related examine

occupation-wide and theoretical state, such

to ethnicity and labor markets. Few empirical policies enforced by the in Russia and

registration systems China, and their effects on intranational migration. China's "socialist market economy" is marked by an uneasy blend of "plan" and market. On the one hand, hallmark institutions of the command con? economy tinue to control urban permanent residence and entitle-

as the

blocking peasant migrants from prestias well as individuals' gious jobs, positions in a rigid insti? tutional hierarchy.

Annals oftheAssociation ofAmerican 92(1),2002,pp.103-124 Geographers, ? 2002byAssociation of American Geographers Published 350 MainStreet, MA02148,and108Cowley UK. byBlackwell Malden, OX41JF, Publishing, Road, Oxford,

104

Fan

premise is that resident status?whether one has local urban hukou?is at the center of migration and labor market segmentation in China. I processes show that a hierarchy of opportunity structure in which permanent migrants are at the top, followed by urban and finally temporary at the nonmigrants migrants bottom, characterizes China's urban labor market. not form, respectively, siders." I argue from the elite, the natives, that this hierarchy does in human differentials capital They and the "out? result attributes

This article's

cordingly, demanded

is selective of those whose labor is migration of neoclasby the host areas. Later extensions

sical migration and longtheory to including expected term returns (e.g., Todaro 1976) also seek to explain why rural-urban migration flows in developing countries con? tinue despite urban unemployment. Whereas the neoclassical perspective dividual approach modifying human or in? emphasizes the behavioral characteristics, capital focuses on contextual or psychological factors decision-making over the neoclas? that individuals

solely ofthe such as education, but is also clearly an outcome state's control over resident status and access to insti? resources, stratification. While relations tionships derstood. rolevel existing between tutional which research hukou reinforces on China a new social order of has dealt with the

rational constraining (Wolpert 1965). The latter s advantage lies in the recognition sical perspective do not behave the same

and migration, with labor market segmentation studies

their interrela-

exactly way and that human and rational economic calculations only partially capital More recent works explain individuals' decision-making. Hanson and on job search (e.g., Granovetter 1974,1981; this argument 1995) extend by of informal the roles personal communica? highlighting in the ways in which tion and nonwage job attributes Pratt 1991; people go about finding jobs. Focusing on gender, Han? are fully em? son and Pratt (1991) argue that employees bedded in social relations that drive men and women into different labor market segments. Like the neoclassi? cal approach, however, the behavioral perspective pays little attention to institutions. Both neoclassical supply-side factors approaches examine and the migrants' and workers' per? and behavioral De Haan

most Furthermore, data that are limited for understanding processes such as migration job search, and job decision-making, of qualitative This article uses a combination change. information in Sichuan from a published study of village households in 1995 and quantita? and Anhui provinces

are poorly un? employ only mac-

in the city of Guangtive data from a survey I conducted the experizhou in 1998. My analysis aims to examine and the salient differences ences of peasant migrants and tempo? migrants, nonmigrants, permanent city. The next section re? rary migrants in the Chinese views migration and labor market theories and how they among deal with institutions. I then describe the institutional and the ur? internal migration bases for understanding ban labor market in China and discuss the inadequacies the Chinese case. of existing theories for understanding The tions articles analysis begins with interpreta? empirical of selected accounts from the Sichuan and Anhui by descriptive survey. and statistical analyses of

The structural which spectives. approach, highlights and market processes, stresses demandwider institutional side factors. the structural The bulk of recent focuses research that articulates perspective into two sectors: the formal and infor? ket is segmented mal sectors, or the primary and secondary sectors (e.g., Piore 1979). To put it simply, labor market segmentation exists on how the labor mar?

survey, followed the Guangzhou

when people of similar backgrounds are clustered in specific occupations or jobs (De Haan 1995). Explanations range from market imperfections (Amsden 1980) to Marxian notions of dualism and polarization (Reich et

Migration,

Labor

Market, between

and

Institutions and labor market science

al. 1973). Research on third world cities has focused on rural migrants who are unable to find work in the formal urban capitalist sector and spill over to the informal sec? tor, one that is marked by labor-intensive production, unskilled labor, low productivity and income, and poor (Harris and Todaro 1970; job security and protection McGee 1982; Gupta 1993). At the same time, research on capitalist industrial economies has identified a pri? wherein is in an in? mary sector, employment organized ternal labor market with nonprice mechanisms and welldefined career ladders (Kerr 1954; Sakamoto and Chen 1991; Gordon 1995). In the secondary sector, jobs are less stable and employers tend to minimize their commit? ments and responsibilities toward labor. The indicators

The intersection has been research. The

migration one of the most fruitful areas of social

most popular theoretical that approaches deal with migration, labor market, and/or the interaction the two emphasize one or more of the neoclassiand structural perspectives. In the neocal, behavioral, classical view, migration is an outcome of the geographi? cal differences in the supply and demand of labor, and at the individual level is an investment in human capital and a result of rational calculations of costs and returns (e.g., Schultz 1961; Sjaastad 1962; Todaro 1969). Acbetween

The Elite, the Natives,

and the Outsiders

1Q5

for the primary-secondary dichotomy vary greatly, but common them are industrial sector, among occupation, firm size, and union/nonunion divide (Sakamoto and Chen 1991). Gordon (1995) shows that primary sector jobs are associated with contracted migration?migration secured with employment?and sector jobs secondary with speculative moves where job search follows migra? tion. Despite the variety of settings in which research theory is conducted, a common thread running through it is the identification of two distinct types of labor market experiutilizing segmentation In this light, the formal-informal can dichotomy be subsumed under the rubric of the primary-secondary in the primary sector have better dichotomy. Employees ences. desirable bargaining power, benefits, job security, and career development than those in the secondary sector. portunities skills, come, Applications have highlighted The first deals portant. labor markets of the labor market segmentation two ways in which institutions more and more in? op? the labor market

industrialization and drives, campaigns, recruitment rural-urban organized promoted migration 1991). Faced during the late 1920s and 1930s (Hoffmann with rapid urban in-migration and overcrowding in cit? an internal passport sys? ies, the Soviet state established tem in 1932, requiring individuals aged sixteen or above to obtain a residence permit (propiska) (Buckley 1995). Accordingly, and became dustrial tration was subject to official approval migration another area of the society to be organized and planned?just like the effect of five-year plans on in? Most importantly, production. was interwoven with government of social such services, and health care residence regisguarantees and as state-subsidized

lectivization

distribution

education, housing, Plane 1995a). Though successful in controlling

and (Mitchneck the registration system was not rural-urban migration, it served

theory are im?

an important in controlling function access to social benefits and programs (Buckley 1995). Since the breakup of the former Soviet Union, most researchers' attention has shifted to the massive from newly tion of Russians and Russian-speakers such as the Ukraine, migra? inde?

of internal with organization at the firm or occupation level that is usu? with large firms and unionization (Saka? ally associated moto and Chen 1991). Empirical studies have found that this institutional and ethnic factor works minorities, into the secondary sector (Gordon 1995). Sec? the ond, studies on international migration emphasize laws and regulations of immigration and refugee migra? of disadvantage who are more likely to be to the

states pendent Uzbekistan to Russia

women

crowded

(e.g., Pilkington institutions continues to less, the legacy of Soviet-period be powerful. s remark summarizes 915) (1995, Buckley most effectively the stickiness of socialist institutions: "In any large scale sociopolitical the institu? transition, tions of the previous regime are not always compatible with the process of change and reform." Despite transi? tion, residence permits are still in use in Russia, symbolizing the state's guarantee of access to official systems of distribution. The notion of a social contract between the state and individuals to affect migrants' calcu? continues so that individuals' and migration processes decisions lus, are influenced not only by human capital and economic but also by access to services and re? tied to the registration and system (Mitchneck Union features and post-Soviet of institutions economies. and the enmigration is

and Kazakhstan, 1998). Neverthe-

tion (e.g., Farer 1995), foreign worker programs (e.g., Calavita 1992; Solinger 1999b), the incorporation of im? 1992), migrants into the labor market (e.g., Waldinger and the less tangible forms of segmentation categories, such as ethnicity and nativity (Wright and Ellis 1997, In both it is the 2000). cases, migrants and the marginal? ized subpopulations in society that are subject to institu?

constraints?firmand occupation-wide internal labor markets, immigration and discrimina? restrictions, tion?so that they are more likely to enter the secondary sector. These constraints further illustrate the inadethat rely solely on neoclassical quacy of investigations and human capital explanations, and suggest that an in? stitutional is the perspective necessary for understanding interaction between Despite the many market segmentation have migration recent studies and and the labor market. that examine labor bases, few market capitalist

tional

considerations sources Plane

1995a, 1995b). The case of the former Soviet illustrates two distinct

Russia

its institutional

among in socialist and migration, labor markets, and institutions transitional economies (Mitchneck 1991). Yet it is in these societies that institutions, state institu? especially tions, play most powerful roles in shaping migration labor market processes. In the former Soviet Union, and col-

investigated settings other than economies. We know little about the interaction

migrants bear the consequences of fines and refusal of social ser? vices. In capitalist these institutional effects societies, exist, but usually dom for internal institutions for undocumented migrants. are interwoven immigrants and selstate and nonstate Second, with one another, making it for the effects of a particular institu-

found in socialist and transitional commonly state institutions are tangible entities, First, forcement of regulations means that internal to official approval. Undocumented subject

difficult extremely tion to change until the entire array of related institutions also changes. In large Russian cities, the registration sys-

106

Fan

tem

is steadfast

because

stricted

and because

the housing market is still re? the distribution of benefits is admin-

istered through registration (Buckley 1995). This feature at least partially explains the lasting legacies of socialist policies. By contrast, in capitalist societies the provision of and access to services nal migration. These tional perspective focusing central While tional are largely independent of inter? differences suggest that an institu? on state institutions must be of migration and labor mar? and transitional economies.

A paramount policies. policy that kept this model operating was the hukou system, which was in the late 1950s and has divided Chinese implemented citizens into two unequal tiers?the privileged urban and 1992; Shen (Christiansen and Tong 1992; Cheng and Selden 1994)- Specifically, hukou is a household record of an individual's (1) regis? tration classification and (2) registration and location, underprivileged to the next. Regis? is usually passed from one generation and tration classification refers to the "nonagricultural" location refers to "agricultural" categories; registration where a person s hukou "resides," which essentially records where he/she belongs. This article primarily focuses on in cities?not location just having a non? registration but also having classification registration in a in the past pro? residing city?which and even today vided access to food and other necessities to subsidized benefits. Prior to entitlements guarantees agricultural one's hukou the economic reforms, where a labor market was almost for peasants who it was next to impossible nonexistent, did not have local hukou to survive in cities. only did the hukou system control rural-urban it has in essence reinforced a dualistic system migration, Not a deep divide between the city and 1990; Cheng (Christiansen countryside Huen The Selden 1994; Cao 1995; Wong and 1998). sys? tem reflected the state's bias toward urbanites, to whom that has fostered and the it pledged work, on the other hand, were expected to The rural Chinese, and were shut out from in the countryside be self-reliant state support. In this regard, the hukou system is effective with powerful institutions that because it is coordinated social functions such as public secu? important rity, housing, and welfare (Chan and Zhang 1999). In an? other symptom of urban bias, the state used the "unified purchase and marketing" (tonggou tongxiao) system that oversee prevailed until recently to set the prices of agricultural goods low and the prices of industrial goods high?the so-called scissors gap?thus value from agri? extracting cultural production in the countryside to accomplish in? dustrialization on the cheap (Tang, Chu, and Fan 1993; Cheng and Selden 1994; Wang 1997; Chan and Zhang 1999). At the same time, the hukou system guaranteed that peasants would be anchored to the countryside and would contribute to the geographical transfer of value to urban areas. it was clear that the preBy the 1980s, however, reform model was incompatible with reform efforts aimat and ing increasing efficiency productivity. Gradually, labor market features were introduced into therefore, China's urban economy (Knight and Song 1995; Maurerin terms of food, housing, full responsibility all and sorts of welfare entitlements. education, in China the rural

and its austere

to the investigations in socialist ket segmentation the neoclassical, are all useful, spectives economies must state institutions

and structural per? behavioral, research on socialist and transi?

tive complicates In the next contexts. tional bases transitional segmentation is concerned.

pertinent begin by examining and policies. An institutional perspec? market analyses that assume capitalist

section, I elaborate on the institu? and urban labor markets in of migration

the China, where the states role deepens effects about which the structural approach

Institutional Urban Labor

Bases Markets

of Migration in China

and

of a labor market are and the development Migration related forces and intricately the two most compelling urban China in the post-Mao pe? that have transformed riod. Yet both continue by the state. In pre-reform to be constrained and monitored

was highly cenChina, labor allocation tralized and tightly controlled by the state. The system of "unified state assignment" (tongyi fenpei) assigned school and regions ac? to specific sectors, occupations, graduates cording workers Likewise, blueprints. to the jobs according As a result, job mobility state's plan of labor allocation. was low, labor market media were not necessary, and unwas rare. With an "iron rice bowl," workers employment for improving Though productivity. it was was absorbed the state's low, efficiency by subsidies to cities. The state sector dominated the urban urban economy, while the nonstate sector was small and weak. In addition, the urban economy was relatively homogeneous because heavy industry was designated as the key sector. In fact, pre-reform Chinese cities were regarded as stratifi"producing" entities where social and residential cations the tertiary sector was kept minimized, small, and city size was strictly controlled (Lo, Pannell, and Welch 1977; French and Hamilton 1979; Lo 1994). The not have model of urban development could pre-reform been possible without a powerful government were lacked incentives to the state's development were transferred to new

The Elite, the Natives,

and the Outsiders

1Q7

In large cities such as Shanghai the tongyi remained in a residual manner fenpei system only by the mid-1990s School still have ac? (Davis 1999). graduates cess to state-sponsored channels for employment, but they now have the option of using market channels such as advertisements, job fairs, and employment agencies. reforms Enterprise replaced the "iron rice bowl" with performance-based hiring, firing, and compensation. State-owned have begun to use labor (SOEs) enterprises and decentralize labor man? contracts, adopt wage reform, The number of private enterprises, agement. foreignfunded firms, and other non-state-controlled businesses has significantly begun to evolve At the same time, cities have from purely productive agents to becom? entities characterized by consumerism, ing consuming markets, a division of labor, a thriving ser? burgeoning vice sector, a growing middle tional and Western outlook class, and a more interna? and Welch (Lo, Pannell, increased.

Fazio 1995).

make

their survival the notion

exemplify readjust and revise themselves mands and circumstances?an transitional lation

in the city possible. These changes that state institutions constantly in response important to new feature de? of the

1999a). Whereas popu? economy (Solinger movements to the were mostly urban 1980s prior to urban or rural to rural (Yang 1994, 120), the above have given impetus to rural-urban migration the two decades. during past the state maintains its gate-keeping role However, over urban permanent residence rural by denying changes workers considered porary cluded in the city local urban hukou. Their migration is and they are considered tem"self-initiated,"

Without a local hukou, they are exmigrants. from the more prestigious and desirable jobs? residents. jobs that are reserved for the urban permanent Instead, rural migrants are considered outsiders to the ur? ban society, picking finding and most are relegated to the bottom rungs, and low-paying jobs and up dirty, dangerous, and underclass existence in the a marginalized

1979; Lo 1994; Yang and 1977; French and Hamilton Guo 1996; Wang 2000). Services such as domestic work, hotels and restaurants, repair shops, and hair salons have side by side with factories expanded sumer goods for the world market. Both the transformation the infusion of foreign the especially investment that produce con-

and economy on labor put pressure supply of cheap labor, in urban supply, areas. Many new jobs created, such as nannies, restaurant servers, and garment factory workers, are at the lower stratum and are not desired by end of the occupational At time, regular urbanites. of the nonstate sector have forms and the expansion more flexible and facilitated the made labor allocation the same labor market re?

of the urban

city (Knight and Song 1995; Yang and Guo 1996). On and selected the contrary, state-sponsored migrants to and all the advanthe city are given urban citizenship it. These include migrants that tages that accompany have close who move ties with the state, such as those to work in jobs assigned by the state and re? turn migrants from previous state-sponsored migrations (Gu 1992), skilled and educated migrants such as profesinstitutional and university students, who are among the most and privileged competitive groups in China. Their mi?

sionals,

1992). While a hiring of "outside" labor (Christiansen new "pull" for labor is exerted by urban areas, the "push" from the countryside has always existed, and has been further of surby the increasing magnitude labor to in due plus agricultural improvements agricul? tural productivity (Shen 1995; Shen and Spence 1995).l Since the 1980s, in response to the pull and push for labor, the state has created a variety of additions to the hukou system and new identity statuses that facilitate the "temporary" migration of peasants to work in urban areas (see, e.g., Wang 1997; Wong and Huen 1998; Chan and Zhang 1999). Most notable among them are the food grain hukou" (zili kouliang hukou) in "self-supplied 1984, the "temporary residence permit" (zanzhu zheng) in 1985, and the increasing "identification usage ofthe card" {shengfen zheng) since the mid-1980s. In essence, to migrate to towns and cities peasants are permitted without obtaining urban hukou, but they will not in that case have access to subsidized benefits. The expanding labor market and the marketization of goods and services exacerbated

by a local hukou, and they are gration is accompanied considered Not only do these permanent migrants. selective and Guo 1999) have the highly migrants (Yang in and to the stay city, they also have ac? legitimacy right cess to an array of jobs closed to temporary migrants, in? with full benefits. They cluding high-paying positions and the native urbanites are given priorities in the labor market. In many ways, the hukou system serves as an internal passport system in China (Chan, Liu, and Yang 1999; Chan and Zhang 1999), granting urban citizenship to deemed the but not to migrants deserving by city peasant migrants brought in because of the cheap labor they pro? vide. Indeed, researchers have compared peasant mi? with undocumented in grants' experiences immigrants the U.S. (Roberts migrants ping into and foreign workers in Germany and Japan 1997; Solinger 1999b). Though some peasant have become successful entrepreneurs by tapfamily and resources native-place stories are by far the of migration (Ma and exception restrictions,

Xiang 1998), success rather than the rule. Despite the

relaxation

108

Fan

hukou

continues

versus urban) tus, and, above and constraints

to symbolize origin, connotes all, defines

one's geographical one's socioeconomic one's access,

(rural sta?

opportunities, (Christiansen 1990; Cheng and Selden 1994). In the view of the state, permanent migrants are official, orderly, and "within state plan," whereas tempo? rary migrants are unofficial, haphazard, and "outside of state plan." A variety of other terminologies have been this dichotomy?hukou to describe versus nonhukou migration, "plan" versus "nonplan" (or self-initiated) and de jure migration, formal versus informal migration, used versus de facto migration (Gu 1992; Yang 1994; Li 1995; the Chan, Liu, and Yang 1999; Fan 1999). However, terms "permanent migrants" and "temporary migrants" are by far the most descriptive of these two distinct tracks of migration in China, and are the preferred terms in this and Goldstein 1991; Goldstein paper (e.g., Goldstein and Guo 1992; Yang 1993, 2000; Yang and Guo 1996). "blue stamp Another addition to the hukou system?the in and has endorsed hukou"?was 1992 formally (lanyin) become 1990s. popular in large cities since the mid- and late It aimed at attracting the most desirable elements

jobs in the city. The supply of labor is not only a function of agricultural surplus most importantly?regulated labor, but is also?and by that control who can and cannot gate-keeping policies cheap be hired. and less desir? By making only the low-paying able jobs available to peasant migrants, the state has, in essence, deepened the formation ofa secondary sector by a continued maintaining supply of cheap labor to it. Be? cause temporary migrants lack permanent resident sta? tend to focus on short-term monetus, their calculations tary gains rather than capital, which deviates investment of human long-term from the Todaro-type (1976) as? that to and longsumption migrants respond expected term returns. In addition, the neoclassical logic treats as a and does not migrants homogenous group explain the bifurcation disadvantaged Though economic between elitist permanent migrants and temporary migrants. the behavioral rational approach downplays like the neoclassical decision-making, per?

its their access to desirable

institutional factors. In it does not emphasize spective the Chinese institutional context, migrants' permanent affiliations with the state translate into greater opportu? nities in the urban labor market; migrants' lack of these affiliations conversely, temporary renders them margin?

of the migrant population by providing them with right in cities (Wang of abode and certain welfare provisions Chan and and Huen 1998; 1997; Wong Zhang 1999). a blue stamp hukou include a The criteria for obtaining large investment and skills. Most cannot essence monied creams or home purchase and age, education, rural migrants are not eligible for and afford this relatively new type of hukou, which in for the benefit off the highly educated, skilled, of the local government's of local economies. and/or coffer

alized labor in the city. Individuals' job search process is not only a function of social network and personal rela? tions; it also reflects the state. The processes structural the degree of their affiliations with

approach's focus on wider contextual makes it a more and labor market segmentation

and the stimulation

the blue stamp hukou does little hukou inherited from the decade-old

In this regard, to mediate the dualism system. Because new and is still of small

for China. In the Chi? alternative appealing theoretical nese city, temporary are indeed being chanmigrants neled into jobs similar to those in the secondary sector in other third world cities. However, the labor market seg? theory assumes homogeneity among migrants in developing countries (Breman 1976) and does not ad? dress their differentiations by their affiliations with the state. In the Chinese city, temporary migrants are not spillovers from the primary sector; rather, they are blocked from entering the primary sector. by state institutions The of an institutional incorporation perspective, espe? one state institutions, would complicate cially involving labor market analyses and highlight the role of the state labor mentation

the blue stamp hukou is relatively it is not included in this article's analysis. magnitude,

Inadequacies

of Noninstitutional

Approaches

Most migration and labor market theories developed for capitalist market economies ignore how state institu? tions shape the opportunity structure in the labor mar? ket. Here, I follow Reskin and Hartmann (1986, 75-80) and define opportunity structure as the nature of employ? ment opportunities that are available, known, and open to people. The neoclassical and labor market allocation views migration perspective of labor purely as functions labor mobility and opportu?

supply and demand, assuming unrestricted and unrestricted access to job information nities.

ap? is their neglect of the interaction between mi? gration and labor market processes on one hand and the of social services on the other. The cases of provision proaches both Russia and China terwoven market illustrate that institutions are in? with one another processes and that migration and labor are intricately related to housing, edu-

in shaping the opportunity structure and deepening market segmentation. Another common drawback of noninstitutional

In China, peasant migrants do not have unre? stricted labor mobility, and their lack of local hukou lim-

The Elite, the Natives,

and the Outsiders

109

cation, health care, public security, and other services in the city. Though this article does not examine in detail the provision of these services in China, the empirical evidence reveals that peasant migrants are out? analysis siders to the city not only in terms of access to jobs but also in terms of access tion I discuss to other services. In the next sec? experiences evidence of these qualitative based on a survey in Sichuan and Anhui.

a man working in a state-owned in Hefei, Anhui comments:

construction

enterprise

Our work is dangerous and difficult. I work at least 12 hours a day and make only 11 yuan (U.S.$1.30).3 There is practically no compensation for work-related injuries. Our con? tract states that compensation is provided only for workrelated deaths and lifetime disabilities. Even in those cases the compensation is a one-shot deal and is low?maybe But you are com10,000 to 20,000 yuan (U.S.$l,200-2,400). own if on our arms or only your eyes are injured . . . pletely so we are very careful not to get injured. (NNJYZ 411-13) The inferior institutional 1995,

Experiences

of the

"Outsiders"

of peasant and temporary migrants The opportunities in China are deeply influenced by state institutions. reflect and labor market experiences Their migration their outsider lected in Sichuan provinces information status in the city. To illustrate this, I use se? from a 1995 study of 300 households and Anhui (NNJYZ 1995),2 two major origin accounts

positions of peasants legitimize of their labor (Knight, Song, the widespread exploitation and Huaibin 1999b). The story of this 1999; Solinger Sichuanese woman is typical among rural-urban migrants:

migrants. The study produces This inof experiences. migration first-person data can shed and rare source of qualitative valuable of rural-urban in re? meanings and experiences light on individual-level such as institutions lation to large-scale sociopolitical I select quotes that capture the hukou. In the following, sentiments migrants expressed by the interviewed typical struction in that survey. This Sichuan man who works as a con? stresses the worker in Shenzhen, Guangdong of hukou: importance . .. factories and enterprises always recruit workers with lo? cal hukou first. We from the countryside have few options; construction is about the only work available to men from rural areas. (NNJYZ 1995, 264-67) He is joined by this Anhui man in Nanjing, Jiangsu: man who works as a handy-

I work in a toy factory in Dongguan, Guangdong ... I work eight hours everyday, including weekends, and get 1 yuan extra for four hours of overtime. We are not allowed sick leave. Even if I am sick, I must still go to work and would not let my boss know because I am afraid to. We eat and to a room . . . it's too crowded . . . sleep at the factory?12 We are paid by the piece; I don't know how they calculate my wage?the factory doesn't tell us how our wages are cal? culated ... I make about 200 yuan a month, and I pay 60 yuan for my meals in the factory. (NNJYZ 1995, 85-86) of peasant migrants as outcategorization barriers against them and from institutional siders results is especially blatant in the workplace. As outsiders, they must apply for and carry all kinds of permits. For exam? The social ple, a worker in a plastics factory in Shenzhen describes:

Peasant migrants are all in menial work?work else wants to do. (NNJYZ 1995, 355-57) and by this Sichuan factory in Shenzhen, woman who works

that nobody

in an eyeglass

Guangdong:

Before I started working in the factory, I had to obtain an identity card, a health certificate, and a "single" (unmarried) certificate. At the factory, I obtained a temporary res? idence permit, a factory permit, and another health certifi? cate. The factory deducted the permit fees from my salary .. . If you are caught in the streets without a temporary res? idence permit, the Public Security Bureau can fine you 300 to 500 yuan. (NNJYZ 1995, 94-96) Most factories have gates, which in addition to providing also between the security symbolize the impermeability temporary migrants and the world outside the factory. A migrant woman in Dongguan describes:

We [peasant migrants] are always the frontline production workers. Better jobs like office secretaries are reserved for the locals. (NNJYZ 1995, 59-61) These three cases illustrate that peasant migrants are disin the urban labor market and that they are into jobs that are considered less desirable by local urbanites?in in Chinese cities. essence, the sec?

advantaged channeled

and are shunned ondary sector

I seldom leave everyone of us you cannot get go to bed early. The notion

the dorm; the factory has a security guard; has a factory permit?without the permit in . . . after work I am usually too tired and (NNJYZ 1995, 85-86) is also rooted in the cultural

SOEs do employ rural migrants, such mi? Though grants are mostly hired as contract workers and are de? nied the benefits to which permanent workers are entitled (Maurer-Fazio 1995; Solinger 1999b). For example,

of outsiders

in the Chinese society. The con? practices of membership and vocabularies of cepts migration connote one's mem? with to the host community rather than bership respect

110

Fan

tions

to the act of migration. of migrant?qianyi

In their places lack of a permanent home?for example, or renkou?and their population" liudong used.

For example, the literal translazhe or qianyi di ren?are rarely come terms emphasizing migrants' the "floating lack of mem?

any where

and will adjust under almost 1995, 225-27) (NNJYZ

any situation.

The lack of job security also results in high job turnover and promotes peasant migrants' emphasis on short-term monetary gain, as illustrated worker in Shenzhen: by another construction

bership in the host community?for example, "people from outside" (wailairen or waidiren) (Duan 1998). Be? cause the bulk of peasant migrants work in industrial and services with called sectors, dagongmei their migration "being (young is commonly employed"?and associated they are dagong?literally

Bosses can fire you or fine you as they like. I will immedi? ately switch to a new job if it offers higher wage. (NNJYZ 1995, 264-67) The above quotes support the argument that the state and its institutions and policies, hukou and including those that distribute urban services, have formalized an structure opportunity market segmentation that induces based on one's and legitimizes labor resident status in

(young working men). in a hardworking, city is mainly tied to their membership and labor force (Zhou 1998; tolerant, cheap, disposable the ur1999). In contrast, Knight, Song, and Huaibin banites are the "insiders" or the natives?bendiren?who in in the city. Differences cultures further segregate are entitled to the membership cuisine, and regional

or dagongzai working women) Their meaning to the Chinese

dialects, the peasant

the city. Though this opportunity has been structure in place for decades, changes in state policies that permit for peas? rural-urban migration but not urban citizenship ant migrants have brought its segmenting effects to the of fore. The rest this article focuses on a survey I con? The analysis is guided by a com? ducted in Guangzhou. parative approach differences among ent resident grants, grants that seeks to identify three subpopulations in the the most salient that have differ?

(Duan 1998). migrants from the urbanites from peasant migrants illustrate their These comments with respect to the positions precarious and peripheral urbanites: When we waidiren walk or bike we give way to the locals. They'll give you a hard time if you're not careful. (NNJYZ 1995, 267-70) We don't interact with the locals in Guangzhou . . . we don't understand their dialect. They look down upon us. (NNJYZ 1995, 234-35) When we have problems we would never go to the local government or Public Security Bureau to get help. Once they notice your accent they will ignore you and protect the interests of the locals. Regardless of whether you are right or wrong, you are doomed. (NNJYZ 1995, 132-35)

statuses

the permanent

(Guangdong Li 1995; 1997).

Chinese city: the nonmimigrants, and the temporary mi? wailai nongmingong lianhe ketizu

The

Guangzhou

Survey

Background The most appropriate sites for studying migration and in China are cities that have received large numbers of permanent and temporary and migrants labor market where changes in the urban economy have accelerated demands for labor. For these reasons, I selected Guang? zhou for this articles quantitative analysis (Figure 1). One of the largest cities in China, Guangzhou is the pro? vincial capital of Guangdong. to the 1990 According was the most attractive destination Census, Guangdong of interprovincial migration and had one of the highest levels of intraprovincial mobility among all provinces (Fan 1996). The census records a total of 35.3 million in? 1985 and 1990, of whom 12.5 migrants between or were in Guangdong 4.4 million, percent, (Table l).4 The in-migration rate in Guangdong was 7.8 percent, twice the provincial average (3.9 percent) across China. in the Temporary migrants were especially prominent province, as illustrated by a rate of 4-6 percent compared with the provincial The 1995 average of 1.9 percent. One-Percent (SSB 1997) Population Sample Survey ternal

without local hukou may have Though migrants in the for an extended of time, they are stayed city period considered "temporary" migrants in the eye of the state. Many migrants have few options other than returning to their home villages in the future. A man from Sichuan comments: working in a repair shop in Guangzhou Will I be staying in Guangdong for good? Absolutely not. We don't have [local] hukou there, which makes life very difficult. My children will be going to school soon. Without a hukou in Guangzhou we cannot afford sending them to school there. (NNJYZ 1995, 30-34) Peasant migrants are unlikely to have long-term plans in the city, which is simply viewed as a place of employment and income as one construction gain (Yang 2000), worker in Shenzhen comments: We [migrant workers] are here to make money, and we will try to be as tolerant as possible. We are tough and will go

The Elite, the Natives,

and the Outsiders

111

Figure 1. The city of Guang? zhou, the Pearl River Delta, and the Guangdong province. See Lin (1997, 80) for the delineation of the Pearl River Delta.

again documents significant between and Guangdong (Table 1). The 1990 Census

rate gaps in the in-migration China's provincial average a total of 16.2 million

vestment

into China

since

the late 1970s.

Such

invest?

estimates in China

(Table 1). It is widely temporary migrants that known, however, census-type surveys underestimate the actual volume of migration, especially that of tempo? rary migration.5 The 1997 Temporary Population 1997) reports (Gonganbu huzheng guanliju who had registered 37.3 million temporary migrants6 with local Public Security Bureaus, of whom 28.7 per? cent or 10.7 million were in Guangdong (Table 1). How? ever, even the 1997 survey underestimates temporary mi? those who did not register grants, because it excludes with local authorities. The standard estimates of tempo? were 70 rary migrants by scholars and journalists in million the late 1980s and in the range of 100 million by the mid-1990s (Solinger in estimates, discrepancies cities are clearly prominent nent 1999a, 18). Regardless ofthe and its however, Guangdong destinations of both perma? made Survey a total of

jobs requiring cheap labor and has been mi? a major reason for the large number of temporary in Within the grants Guangdong. province, migrants are in the Pearl River Delta, which prominent especially consists of special economic zones, large cities, and rapas Hong urban and rural areas serving hinterland (Sit 1989; Lin 1997). production is part of the Pearl River Delta, is close to Guangzhou Hong Kong (150 km), and has received large amounts of idly growing Kong's from and via Hong Kong (Li 1993). foreign investment of Guangzhou's and Both the diversification economy the infusion of foreign investment into the city have exof migration and labor market for information that census-type call in-depth processes surveys lack. In order to examine these processes, in 1998 I conducted that a questionnaire survey in Guangzhou7 includes three 300 permanent larger number nonmigrants, types of respondents?305 911 and migrants, temporary migrants. A of temporary migrants were included be? erted a pull for migrants. Detailed examinations

ment has created

and temporary migrants. has benefited immensely from its ties and Guangdong to Hong Kong, a major source of foreign inproximity

cause they are the newest and most dynamic migrants in and labor Chinese cities and because their migration

112

Fan

Table 1. Estimates of Migration

Volumes in China and Guangdong

Sources: Gonganbu huzhengguanliju (1997); SSB (1997). Also, see endnote 4. Notes: Please see endnotes 5 and 6 for the variationsin definitions among the three sources.Migrationrate is defined as the proportionof migrantsamong the population ages 5+ in 1990 for the 1990 Census, which only recordsmigrantsages 5 + . Since the 1995 One-Percent Population Sample Survey includes Pop? migrantsof all ages, the respective migrationrates are computed as proportionsofthe entire population in 1995. Migrationratesfor the 1997 Temporary ulation Surveycannot be computed because the surveyincludes largeproportionsof migrantsexcluded from the officialpopulation estimates. In all cases, the migrationrate for China refersto the provincial average.

market

processes

are less well

understood.

Rather

than

migrants and temporary mi? examining only permanent was the survey designed to reveal a more com? grants, the ur? urban labor market involving of the plete picture ban nonmigrants permanent as different and Though nonmigrants migrants both have local hukou, I treat them migrants' selecgroups because of permanent as well.

framework (see the appendix). sampling Despite is a major magnet for mi? these limitations, Guangzhou grants of all kinds, it has a large, diverse, and changing and it has a rapidly expanding urban labor economy, the market. is among the best For these reasons, Guangzhou field sites for studying the relations between migration and the urban labor market in China.

tivity The survey sample was arrived at using stratified quota across both major occupawith stratification sampling, tional categories and geographic districts in Guangzhou. The appendix to this article describes the sampling frame? the work and the sampling process in detail. Specifically, aims a framework at wide variety of including sampling and at adjusting for the expansion of com? occupations merce and services in the city between 1990 and 1998. In the survey, "nonmigrants" refers to individuals who had lived in Guangzhou for at least fifteen years and whose hukou hukou in Guangzhou; migrants" are "permanent those who had moved to Guangzhou since 1990 and whose were

and their close ties to state institutions.

Demographic

Characteristics

between the Guangzhou Comparisons survey sample and the 1990 Census's Guangzhou sample8 are instrumental for assessing the quality of the survey data. The details are reported elsewhere (Fan forthcoming) not repeated in this article, but the similarities and are between

the two samples in age and sex structure, migration oriattainments are worth noting. As in gin, and educational the 1990 Census, in the migrants Guangzhou survey were younger than nonmigrants (Table 2), and both per? and temporary migrants were heavily concen? in the 15-39 age group. Migrants had higher sex ratios than nonmigrants, and permanent migrants had higher sex ratios than temporary migrants. The vast ma? and permanent held jority of nonmigrants migrants manent trated classification; only a minorso. did the Also, migrants place-of-birth data show that compared to nonmigrants and permanent who are largely of urban origin (cities and migrants, towns), temporary migrants are mostly of rural origin. Fi? nonagricultural ity of temporary registration nally, the educational attainments of the survey sample

were in Guangzhou; and "temporary migrants" re? fers to those who had stayed in Guangzhou for at least three months but whose hukou

were not in Guangzhou. The survey included only individuals ages 15 or older. The Guangzhou that must survey has two limitations be taken into consideration when interpreting the find? ings. First, a survey of one site, albeit a major city, is not of other urban areas in China. necessarily representative Second, gender the census, the size and occupational and distributions ofthe survey sample are functions of unlike

The Elite, the Natives,

and the Outsiders

113

Table 2. Demographic and Migration Characteristics: The Guangzhou Survey

and agri? industrial, and agriculture. While professional culture are respectively at the highest and lowest ends of the occupational ser? stratum, the statuses of commerce, work are not clearly established industrial Xu, and Wei 1993). This is partly a result ofthe transformation of Chinese cities, where new ongoing and industrial commerce, services, jobs have rapidly emerged, and partly due to the large variations of job sta? vices, and (Stinner, tuses within framework different each category. Though the survey's sampling arrived at occupational distributions that are and less concentrated than those of the Census

sample, the former is probably a more realistic represen? tation of Guangzhou's economic structure in the late 1990s (see the appendix) than is the latter. In addition, the sampling framework Census sample?namely, maintains the modes from the industry (32.8 percent) for non? mi? (46.0 percent) for permanent migrants, professional for and grants, industry (33.9 percent) temporary migrants (Table 2), as well as the respective occupational ranking, with permanent migrants at the bottom. grants at the top and temporary mi?

In short, comparisons of the Guangzhou survey sam? that the indicate ple and the Census' Guangzhou sample good source. Most importantly, survey is a reasonably differences the data highlight important socioeconomic underscore hukou the three and among subpopulations, of social and economic as an important definition stratiin Guangzhou. Contrasts in terms of human cap? ital are marked by a hierarchy in educational attainment: are the most educated and permanent migrants highly fication temporary migrants the least educated. These discrepancies in human capital and residence status have probably in labor market opportunities. reinforced differentials The next two sections deal more specifically with migra? tion and labor market processes. Processes: Migration and Decision-Making * In the moves are considered the same as "selfsurvey,"industry/business" initiated"moves. Permanent migrants Patterns, Reasons,

Spatial

have temporary migrants and In the vastly differing migration processes patterns. the of Guangzhou survey, majority permanent migrants came from cities and the majority of temporary migrants from counties (Table 2), again underscoring originated the urban origin of permanent migrants and the rural or? 49.5 percent igin of temporary migrants. Respectively, and 53.9 percent of permanent and temporary migrants were interprovincial migrants. The central region was most popular source of interprovincial permanent and temporary migrants, but a significant of proportion came from the east? (37.2 permanent migrants percent) ern region.9 This pattern suggests that permanent mithe

and

are similar to those of the Census sample, essentially repeating the latter's educational ranking, where perma? nent migrants were the most highly educated and tempo? distribution of earlier, the occupational the Guangzhou results from a sam? survey predetermined pling framework. I used the five occupational categories consistent fessional with conventional (including definitions in China: administrative), commerce, pro? services, rary migrants As described the least educated.

114

Fan

grants are more likely than temporary migrants to come from more developed origins. The contrast between permanent migrants and tem? is even more in their reasons for marked porary migrants The standard nine "reasons for migration" migration. listed in Table 2 have their origins in Chinese censustype surveys. It is important to note that these "reasons" connote the means and types of migration and the de? of state rather than the motives be? involvement, gree hind population movement (Fan 1999).10 "Job transfer" and "job assignment" refer to the state's allocation of human resources by assigning school graduates and transferring workers to specific jobs and regions. "Study/training" refers to migration to attend schools or training pro? Since admission to institutes is grams. higher-education to state and is subject by local planning accompanied hukou, "study/training" volvement. By contrast, initiated moves or trade sectors also reflects "industry/business" in industrial, for engaging state in? significant refers to self-

Table 3. Migration

Considerations:

The Guangzhou

Survey

and is considered

state plan." "Friends/relatives" the help of friends and relatives, and is mostly associated In short, while "job trans? with self-initiated migration. and denote the fer," "job assignment" "study/training" state's significant and the labor market, "industry/business" and "friends/ relatives" are most closely identified with market forces. role in monitoring human resources

commercial, "outside of migration refers to migration to seek

* Multiple responsesare permitted. All responsesare included in the per? centage computation.

for their moves. Their reasons for leaving key incentive their origins?63.5 percent citing "low income" and 19.1 percent "few jobs," compared with 28.0 percent and 5.7 percent, respectively, among permanent migrants?fur? ther highlight the important role of short-term economic return. as Finally, the reasons for choosing Guangzhou the destination11 depict the economic pull of the city. Specifically, "higher wages" and "ease in finding jobs" acfor, respectively,

and "job transfer" "job assignment," "Study/training," were the most prominent reasons migration among the surveyed permanent migrants, while temporary migrants were overwhelmingly or "self-initiated" "industry/business" The of reflects migrants. importance "study/training" the concentration zhou. local Entrants The urban hukou of educational to universities institutions are almost in Guang? always offered

counted

55.3 percent of the responses and 49.8 percent of those by per? by temporary migrants manent migrants, whereas "proximity to origin" was not an important reason among either group. The above analysis of migration considerations, com? bined with earlier observations about migrant origins, that temporary in Guangzhou are suggest migrants to the large gaps in mostly "upward" movers responding income and employment between the city opportunities and their origins (Ma, Liaw, and Zeng 1997). In other words, to temporary migrants the push from the origins is as strong as the pull from Guangzhou. Though perma? nent migrants may also be "upward" movers, their migra? tion origins and considerations suggest that, to them, pull forces are stronger Labor Market and Ownership Comparisons nonmigrants, Processes: Sectors of the permanent job-search migrants, experiences among and temporary mithan push forces. Job Search

grants. the varied whose

and are by definition mi? permanent in migration differences reason highlight

available to migrants. Migrants opportunities jobs are allocated by the state and migrants for education state plan" migrants?are purposes?"within for local hukou in cities, while "outside of state that move on their own initiatives and

eligible

plan" migrants are not sponsored Unlike

by the state are by and large not eligi? ble for local hukou and must appeal to market resources.

census-type surveys, the Guangzhou survey was designed to analyze migration motives and decisionmaking (Table 3). "Job search" was the most important and temporary among both permanent surveyed, and "study" and "increase income" second motive migrants were the

for permanent mi? leading motive respectively and The of grants temporary migrants. prominence "job search" and "increase income" mi? among temporary grants indicates that immediate monetary return was a

The Elite, the Natives,

and the Outsiders

115

grants can shed some light on the processes by which the urban labor market is segmented and how that segmenta? tion is related to resident status. Table 4 shows that "in? was the leading job search criterion of all three and that it was most prominent groups among temporary to migrants. "Nature of work unit," referring specifically come" the "ownership sector" of the employment (see below), was a criterion for respectively 18.6 percent and 23.0 per? cent of nonmigrants and permanent migrants, but only 4-7 percent of temporary These differences migrants. for again suggest that income was an important incentive all migration but was especially important for temporary such as job attributes migrants, whereas nonmonetary sector were also high among nonmigrants' ownership In addition, and permanent considerations. migrants' but "location" was ofthe "stability" job quite important, and "benefits" were of low importance Nonmigrants, permanent and chanmigrants differ in their sources of information in "relatives nels for employment (Table 4)- Though in? of the source labor market was Guangzhou" leading formation for all groups, "relatives outside Guangzhou" to all three groups. and temporary migrants,

was a prominent source of information for temporary mi? grants but not for permanent migrants. Social networks, those from home villages, were the dominant including of information for temporary migrants. On the of nonmigrants other hand, higher proportions and per? manent migrants than temporary migrants used informa? tion via "advertisement" and "work unit/school." Re? did this to "how find indicate that the job?" you sponses source majority grants, percent of all three relied mi? groups, especially temporary on their own searches. Respectively, 42.6

and 30.7 percent of nonmigrants and permanent their or found "recruitment" migrants present jobs by "work unit assignment," reasons that accounted for only 8.3 percent of temporary migrants. social networks and self-initiation three more groups, nonmigrants to institutional connected In summary, though are important to all and permanent migrants are

and organized sources than are temporary migrants, who must and channels social networks in rely on informal resources, including the origin and in the destination. Besides indicator another important distribution, occupational is the distribuof labor market segmentation

Table 4. Job Search and Sector: The Guangzhou

Survey

116

Fan

tion

of ownership

sectors?state-owned, and

collectiveBoth the

Modeling Evidence

Permanent

and Temporary

Migration

owned, "new-economy," state- and the collective-owned socialist-type components shrunk in size in recent

self-employed. sectors are traditional sector. This

of the state

has

in large and years; however, older cities such as Guangzhou, it remains prominent. At the same time, recent reforms of SOEs and changes in the urban economy have promoted shifts of the labor force to the nonstate sector, which is relatively new and in enterprises and includes employment other complex than This state-owned and collective-owned to which I refer in this subsector, sector, is mainly represented "new-economy" by foreignand and invested enterprises by private, family-owned, It is owned characterindividually enterprises. especially ized by jobs in industry and services, and is very important to temporary migrants who have little access to institu? tional resources and whose opportunities mainly lie out? is another rela? side the state sector. Self-employment tively new means of livelihood and outside involving businesses Both one the have employers as opposed individuals the state sector, who own their enterprises. article as the

presented thus far indicates that substantial exist among the three subpopulations, espe? cially between permanent migrants and temporary mi? between resident grants. To quantify the relationships status and the migration process, I analyzed these differ? contrasts ences via a logistic regression analysis (Table 5). The de? variable is coded one for permanent pendent migrants and zero for temporary migrants. In essence, the analysis evaluates gration that how well the independent variables predict mi? is accompanied urban hukou versus by

migration not accompanied by urban hukou. The advanmethods, such as logistic regression, tage of multivariate over descriptive statistics is that an independent variable's contribution dent variables are held constant. indepen? For example, in order to of institutional factors, it is important factors. included in and access. can be assessed while other

assess the strength to isolate the contribution

in enterprises. to being employees sector and the self-employed new-economy since the 1980s rapidly gained prominence for 43.3 percent

of human capital variables are Four groups of independent the analysis: demographic, location, motive, It is expected that individuals who are more

and of nonmigrants but 16.4 only per? percent migrants, cent of temporary migrants (Table 4). The majority of sector. temporary migrants were in the new-economy These data again support the notion that nonmigrants 57.1 of permanent and permanent migrants and institutional established temporary migrants and less institutionalized have greater labor market access to wellprocesses, to newer

(Davis 1999). SOEs accounted

cated, from urban backgrounds, cerned with monetary return and who have greater ac? and institutional cess to state-sponsored opportunities are more likely to be permanent migrants, while indi? viduals who are less highly educated, from rural areas, and highly motivated return and who rely by monetary on their own resources are more likely to be temporary migrants. A relatively of high p2 (0.65) and a high percentage classified observations (93.2 percent) correctly suggest that the independent variables as a whole are successful in identifying the salient differences between permanent and (Table 5). Because migrants temporary migrants odds ratios are unit-dependent, they are reported for ref? erence purpose only. The size of standardized regression coefficients liable indicators tests are more re? significance of the relative importance of the inde? variables in predicting the permanent migrant and associated migrant outcome. results support the expectations described two dummy variables, SENIOR HIGH and

highly edu? and relatively less con?

while

are mostly

channeled

of the urban econ? segments in of all nonmigrants, one five roughly migrants, and temporary migrants were in the permanent that self-employment sector, indicating self-employed omy. In addition, has emerged labor market. Finally, as an important rate of job segment turnover of Guangzhou's was the

highest migrants and lowest among nonmi? small number of jobs grants (Table 4). The relatively their older held, despite nonmigrants age (Table 2), sug? a level of job stability among them, further gests high among temporary in their current job by their long duration 9.9 years) at the time ofthe survey. Tempo? rary migrants' higher job frequency and shorter duration at the present job depict a relatively high level of job illustrated (averaging not only their association turnover, which underscores with the more insecure and fluid segments of the labor market but also their highly income-driven job search approaches.

the

pendent versus temporary The overall earlier. The ABOVE tively of ABOVE related

SENIOR

HIGH, are both significant and posito permanent The coefficient migration. SENIOR HIGH is the second largest among

all independent which suggests that higher variables, education is especially influential in increasing the likeli? hood of permanent migration versus temporary migration. The variables?INTRAPROVINCIAL, and PLACE OF BIRTH?are all INTERPROVINCIAL, and positively related to permanent significant migrathree location

The Elite, the Natives,

and the Outsiders

117

Table 5. Logistic Regression on Permanent

and Temporary Migrants: The Guangzhou Standardized

Survey

Independent Variable Demographic EDUCATION (reference: below senior high) *SENIOR HIGH *ABOVE SENIOR HIGH Location ORIGIN (reference: central and western regions) *INTRAPROVINCIAL *INTERPROVINCIAL (eastern region = 1) *PLACE OF BIRTH (urban = 1) Motive (reference: non-income- or job-related) *INCOME/JOB Access *SELF-INITIATED JOB INFORMATION (reference: non^Guangzhou or institutional sources) *GUANGZHOU *INSTITUTION Model chi-square ?2 log likelihood with intercept ?2 log likelihood of model p2 Percentage correctly classified Number of cases Degrees of freedom

Regression Coefficient

Wald Statistic

Odds Ratio

Notes: 1. Dependent variableis coded 0 for temporarymigrantsand 1 for permanentmigrants. 2. * Indicatesdummyvariable. 3. Significance levels are indicated as follows: ** is 0.05, *** is 0.01, and **** is 0.001. 4- p2 ? 1 ? (?2 log likelihood of model / ?2 log likelihood with intercept).

that migrants from within Guangdong, tion, indicating from the eastern region, and from urban origins are more likely to be permanent migrants. is coded one for migrants who consid? INCOME/JOB ered income increase or job search as the main motive for migration, and zero otherwise. Though it is not signif? icant, driven that migrants most negative sign suggests by income and job-related motives are more likely to be temporary SELF-INITIATED is coded migrants. its

"relatives pared with the reference group of variables outside outside Guangzhou," "agencies Guangzhou," and "other" (Table 3). The results suggest that migrants in Guangzhou and having greater access to information to institutional migrants. The logistic sources are more likely to be permanent

one for "industry/business" or "self-initiated" migration and zero for other types of migration, and is the most influential independent variable in the analysis. Its nega? government that "self-initiated" migrants without are more likely to be temporary sponsorship variables migrants. Finally, both JOB INFORMATION are positively and significantly related to permanent GUANGZHOU is coded one if "relatives migration. in Guangzhou" or "agencies in Guangzhou" were the of information is coded about the labor market, one if "advertisement" They and or tive sign indicates

regression results point to the important roles of institutional factors such as government sponsorfor employment ship and channels independent of mi? such as their human grants' attributes, capital and their origins. In other words, high education and urban back? ground only partially explain permanent migrants' abil? ity to obtain urban hukou. Likewise, temporary migrants' low education and rural backgrounds are not the only reasons for their not obtaining urban hukou. This finding underscores structural mechanisms whereby hukou sta? tus is not simply associated with the abled and competitive but is, above all, a gate-keeping mechanism that stratifies the urban labor market and relegates peasant migrants, regardless of their abilities, cure, and inferior positions. to peripheral, inse-

main source INSTITUTION

"work unit/school"

was the main source.

are com-

118

Fan

Labor Market, The

Market Sectoral

Processes: and

Entering Occupational

the

Labor

First-timejob seekers

Shifts

first job one gets and whether one stays in the same ownership sector and occupation during job change structure in the can shed further light on the opportunity To this end, I analyzed the shifts that acstatus the change from a previous nonworking company or job to the present job. Because very few respondents engaged in agriculture during the time of the Guangzhou survey (see Table 2), they are omitted from this analysis. labor market. whose By "first-time job seekers," I mean individuals first of the were their at the time jobs?who survey jobs of all respondents?plus accounted for 30.1 percent those who were agricultural workers prior to holding the present job.12 In other words, first-time job seekers in? clude all those who entered the urban labor market for and of nonmigrants the first time. The high proportions testiSOEs a entering comprise migrants permanent resources mony to these migrants' access to institutional and highlight such as job assignment by state agencies, and lo? between state-sector the association employment 2). On the other hand, the majority of entered the new-economy sector, temporary migrants further supporting the argument that they have little ac? cal hukou (Figure cess to state-sponsored employment. who were not first-time For respondents computed the ownership tination rates. Retention sectors' retention rate defined job seekers, I rates and des-

in one sector of respondents sector after job change, is an indicator of the ownership to retain labor. The rates are sectors' relative strengths and do not add up to 100 (Figures 2 and 3). sector-specific of sectoralDestination rate, defined as the percentage who shifted into a specific sector, de? shift respondents notes the relative abilities of sectors to attract labor from rates are group-specific and add Destination rates mea? up to 100. Both retention rates and destination to labor. Most of sure the respective sectors' attractiveness the the illustrate varied all, however, they opportunities statuses. with different resident offer to workers sectors other sectors. retention rates were in SOEs and the The highest the self-employed sector for nonmigrants, new-economy sector and SOEs for permanent migrants, and the new2). (Figure temporary migrants economy rates Among all three groups, however, the destination and self-employed sectors were the of the new-economy highest. These results indicate that while SOEs are still a favorable who and permanent migrants, sector for nonmigrants to to access state-sponsored employment enjoyed and self-employment In particular, the high sector for

as the percentage that remained in the same

Non-migrants Permanent migrants Temporary migrants Figure 2. Ownership sectors: The Guangzhou survey.

retention tor among

and destination

permanent work state-sponsored among such migrants.

sec? rates of the self-employed from that a shift migrants suggest

has occurred to self-employment To temporary migrants, the new-

sector is clearly the most prominent?large economy of them entered that sector, stayed in that proportions sector despite job change, or shifted to that sector from the differing opportu? other sectors. Besides highlighting the find? among the three groups of respondents, of ings underscore the increased diversity and complexity labor market. Guangzhou's nities of first-time job seekers, retention distributions are not func? and destinations rates by occupation rates, For the tions of nonmigrants, survey design (Figure 3). for first-time job seekers the most popular occupation The the industrialized econ? work, reflecting it was of For migrants, pro? Guangzhou. permanent omy fessional work, which was partly a result of their high and partly a result of their access to in? level of education was industrial stitutional resources. Industrial and services employment

begin with, jobs in the new-economy attractive. sectors are increasingly

The Elite,

the Natives,

and the Outsiders

119

First-timejob seekers

itable, occupation The prominence

for those already in the labor market. of permanent migrants in commerce group and are suggests that they are the most competitive and temporary migrants to more likely than nonmigrants succeed ization The above in new and thriving data highlight occupations. the increasing commercial-

of its of Guangzhou's economy, the heterogeneity of the labor market, and self-employment. popularity struc? Above all, the findings underscore an opportunity ture defined nonmigrants when they by resident status. Permanent enjoy state-sponsored first enter the urban migrants and resources employment

labor market; as they in are such especially competitive migrants change jobs, in or moving to thriving sectors and occupa? remaining In and commerce. tions, particularly self-employment contrast, temporary migrants have limited access to in? when resources and prestigious stitutional occupations therefore and are labor the enter market, highly rep? they resented in new-economy jobs and in services and indus? temporary migrants are During job change, able than those with local again less competitively such as commerce. hukou to shift to popular occupations trial work. once

Noiwnigrants Permanent migrants Temporary migrants Destination rates

Summary

and

Conclusion

n Non-migrantsPermanent migrants Temporary Figure 3. Occupations: The Guangzhou survey.

theoret? In this article, I have argued that established between mi? ical approaches for studying the interaction in general gration and labor market neglect institutions and state institutions omies, econ? in particular. In transitional such as those in Russia and China, that are negothe difficult match between plan and market, in?

were

for temporary migrants leading occupations that were first-time job seekers. These differences repre? for the argument that entrance and sent strong evidence the access to the labor market status. retention is heavily dependent on one's

tiating The are especially of institutions revealing. vestigations ofthe opportunity Chinese state is a critical determinant structure that shapes individuals' their participation and resident status varied opportunities between institutions. duced and legitimized by gate-keeping hukou nism system that allocates functions

migration to cities and in the urban labor market. The link is inThe

resident

rates suggest high occupational of respondents that large proportions stayed in the same despite job change. Among the four occupa? occupations rates. In par? had the highest retention tions, commerce The overall ticular, 92.6 percent of permanent migrants in that occupation remained commerce previously in job despite with their to be associated which high appears change, sector that was ob? rate in the self-employed retention served earlier lar occupation migrants most popular retention was the most popu? (Figure 2). Commerce to which to shift, especially for permanent was not the (Figure 3). Even though commerce and sector among destination first-time rates Guang? and more commercialized is becoming and the most profis the most attractive, job seekers, that indicate its high

as a society-wide sorting mecha? to "within state urban citizenship with the state. These

plan" migrants

with close affiliations

migrants are, in essence, the elite in the city permanent than even the ur? and are in more prestigious positions ban natives. In contrast, most peasant migrants are de? nied local hukou data from a in the city. The qualitative Anhui and 1995 survey in Sichuan support the notion institutional outsider that peasant positions migrants' render them at the bottom of the urban socioeconomic and labor market Results strata. from the 1998 Guangzhou survey I conducted structure de? the argument that an opportunity together with human capital-type factors

zhou's economy that commerce

support fined by hukou,

120

Fan

hierarchy migrants are the most prestigious, are in the middle, and temporary mi? urban nonmigrants struc? least This opportunity are the grants prestigious. between ture has led to varied migration per? processes in which permanent manent market and temporary migrants and differing labor Per? among the three subpopulations. to begin with, as manent migrants have the advantage they are more highly educated and come from more de? processes resources veloped locations. Their access to institutional further privilege them in the channels and government labor market. The net result is that they have the most in they are the most competitive prestigious occupations, the labor market, and they are more likely than nonmi? grants and temporary migrants to succeed in new sectors. also have urban hukou and Though urban nonmigrants associated tions benefits, they in the labor market posi? mi? hierarchy than permanent from institu? migrants are dislocated occupy somewhat lower

such as education,

results

in a socioeconomic

Appendix A sampling framework with stratification both across

and the eight urban districts in categories occupational was used to guide the sampling process. Guangzhou in Guang? of major occupations Using the distribution zhou from the 1990 Census ments the were made. as a basis, four types of adjust? I made the first adjustment to reflect structure between in the city's economic

changes of 1990 and 1998, by increasing the relative proportions and services. Then, I adjusted the occupa? commerce in ofthe three types of subpopulations tional proportions order The to reflect actual their likely three occupational official because are not in known, Guangzhou groups data do not include migrants that have not registered I relied on informants with local authorities. Therefore, of these Bureau survey in Guangzhou, including State Statistical a of as well as scholarly and journalisvariety specialists, of occupa? breakdowns the tic sources that estimated tional (e.g., management, categories ment workers, etc.) among migrants 2001). Third, in order to facilitate the three that tional upward subpopulations, numbers sufficient categories the proportions street vendors, garin Guangzhou (Fan occupational distributions distributions.

grants. Temporary tional resources, are crowded into peripheral segments, in the labor market. and are less capable of advancing leave them few options Their marginalized positions other than relying on informal sources and striving immediate

for

monetary gains. and backin human differences capital Though grounds are important factors, statistical analyses in this article have shown that individuals' resident status exerts effects on their labor mar? and substantial independent ket processes. Their labor market outcomes, especially in differ? terms of income and benefits, are also significantly see Fan 2001). Guangzhou's ent (as reported elsewhere; to migrants, its diverse and commercialand its rapidly growing labor market sug? ized economy, in may well be repeated gest that these observations other large Chinese cities that have also received numer? attractiveness ous migrants. This article's findings suggest that in the not only shape migration Chinese city, state institutions but are also enmarket labor and segmentation processes a new social order marked gendering stratifications. backed socioeconomic The Chinese by institution-

among comparison to ensure I made adjustments of them in respective occupa? were included. For example, I adjusted and permanent of nonmigrants under commerce. Finally, I al-

migrants located roughly equal except for occupational

in the categories

to men and women, that are clearly domi? categories nated by one sex (e.g., nannies, who are predominantly and construction workers, who are predomi? women, proportions nantly male). for the initial sampling The geographic proportions framework across the eight urban districts of Guangzhou were derived based on existing data of the geographic distributions of population, population, nonagricultural as well as the settlement and migrants in Guangzhou, history of individual urban districts. For example, Yuexiu because and Haizhu were allocated larger proportions are among the oldest and most urban parts of and they are known for increasing concen? Guangzhou, smaller tration of migrants during the 1990s. Conversely, were allocated to suburban districts, such as proportions they Baiyun, that are less known an initial selected for migrant framework, concentration. a team of six inUsing terviewers sampling the quota-sampling employed individuals

Its com? urban labor market is complex. ofthe role the small to in no is due persistent part plexity the way in which plan and market state in monitoring coexist. It is this institutional of existing factor that undermines The neoclassical the and applicability behavioral approaches do not specifically address the role of institutions. The labor market segmentation theory offers a useful guide, but it needs to expand to incorpo? rate the role of the state and state institutions. tutional and theories.

An insti?

in socialist is especially necessary perspective in which segmentation is economies transitional institutional affiliations and status. much defined by very

randomly tional and geographical numbers or proportions work were reached. questionnaire

and technique who satisfied the occupa? criteria until the predetermined in the initial

sampling frame? respondent was given a to fill out with the help of the interviewer. Each selected

The Elite, the Natives,

and the Outsiders

121

were made to the the survey, fine adjustments in cases where the predetermined sampling framework numbers or proportions were judged too high or too low For example, based on our field observations. it was exDuring in Guangzhou tremely the propor? who were permanent migrants. Accordingly, tions of permanent migrants in the industrial category were adjusted downward. By and large, however, we were difficult workers closely the initial sampling frame? breakdowns in Table 2 summawork. The occupational quite rize the final sampling framework by major occupation. is more In this study, the quota-sampling technique traditional random than the sampling appropriate method based on databases provided by local authorities, First, official databases inevitably overlook temporary migrants who did not obtain proper paperwork such as the temporary residence permit. My in? for three reasons. formants in Guangzhou and the survey both suggest that a large proportion of temporary migrants choose not to and obtain proper permits because of the fees involved not be because the permit requirement may strictly enforced. For example, among the temporary migrants in? cluded in the survey, 19.0 percent did not obtain tempo? residence technique rary permits. The quota-sampling allows us to examine this important group, which is easily missed using official databases. Second, random sam? pling based on official databases makes it difficult to satcriterion stratification isfy the occupational above. Third, a survey not based on official permits outlined databases able to follow to find industrial

in December study. Finally, a pilot survey was conducted 1997, six months before the actual survey, which tested the questionnaire and identified and resolved possible inconsistencies among the six interviewers.

Acknowledgments The research on which this article is based was funded

Science Foundation (SBR-9618500), by the National the Luce Foundation, and the UCLA Academic Senate. I am grateful to Ling Li for her major role in administerZhou and ing the Guangzhou survey, and to Daming for me access from the Luo the data Zhaoyuan helping Sichuan and Anhui survey. I would also like to thank Kam Wing Chan and Yunyan Yang for their contribution to a collaborative project, Chase Langford for carto? and assistance, John Paul Jones III and the five graphic Ear? for comments and suggestions. referees anonymous of this article were presented at the Interna? on Resource Management, Urbaniza? Workshop in Hong Kong and Zhujiang Delta, tion and Governance of Hong Kong, 23-24 Chinese May 2000, University Lewis Center for Re? and at the Pacific Rim Workshop, lier versions tional gional Policy Studies, UCLA, 3-4 August 2000.

Notes 1. Estimates of surplus rural workers in China vary considerably; most fall in the range of 100 to 220 million (Solinger 1999a, 18). 2. This survey was administered by the Ministry of Agriculture and included 300 peasant households in twelve villages in Anhui and Sichuan. It was conducted during late January and early February of 1995, when many out-migrants had returned to their home villages for the Spring Festival. The accounts by the interviewed households were transcribed word for word; the average length is approximately 3,000 Chinese words. 3. In the mid-1990s, when this survey was taken, monthly wages in the range of 200 to 500 yuan (U.S. $25-40) were quite typical of peasant migrants in industrial or services sectors. By the late 1990s, similar jobs could offer wages up to 1,000 yuan. 4- The statistics are based on a sample of the 1990 Census pro? vided by the National Information Center. It is a onepercent clustered sample containing information about every individual in all households of the sampled village-level units (villages, towns, or urban neighborhoods in cities), and consists ofa total of 11,475,104 records. 5. The 1990 Census defines a migrant as an individual five years or older whose usual place of residence on 1 July 1985 was in a different city, town, or county than on 1 July 1990 and (1) whose hukou was in the 1990 place of residence or (2) who had stayed in the destination for more than one year or had left the hukou location for more than one year.

more in-depth questions to be asked. described above did not yield a Though the technique and random sample, specific strategies representative were employed to increase as much as possible the repand quality of the survey

resentativeness, randomness, data. First of all, the occupational and geographical stratifications outlined earlier reduced possible occupational and spatial skewedness of the data. Second, several guidelines were used to reduce possible biases ofthe data. For example, no more than one person per household

and one person per work place were interviewed, except under special circumstances. If significant in variations resident (nonmigrants, permanent migrants, and and in other demographic dimentemporary migrants) sions such as age, income, and education existed within or workplace, the household we allowed a maximum of and four interviewees per household per workplace. Third, I employed a total of only six interin order to minimize viewers inconsistencies possible had during the interviewing process. All six interviewers had extensive in conducting prior experience questionnaire surveys in Guangzhou, and all were trained specifithe questionnaire used in the cally for administering two interviewees status

122

Fan

The first type of migrants had obtained local hukou at the destination and were by definition permanent migrants, whereas the second type of migrants did not obtain local hukou and were temporary migrants. The Census excludes moves within cities or counties, migrants younger than five years old, migrants who died between 1985 and 1990, mul? tiple moves, and return migrants between the two years. Furthermore, the "more than one year" requirement ex? cludes temporary migrants who had stayed in the destina? tion for less than one year and those who had left the hukou location for less than one year, who in the Census were grouped under their hukou locations rather than their destinations (Banister and Harbaugh 1992). The 1995 OnePercent Population Sample Survey uses definitions similar to the 1990 Census, except that it examines the period includes migrants of all ages, and employs a 1990-1995, "more than six months" requirement. 6. Approximately three-quarters of these migrants had stayed in the destination for less than one year, partly accounting for the large discrepancies between the survey's estimates and those from the Census (Table 1). 7. The survey was designed in conjunction with other mem? bers of a collaborative project. See this article's acknowledgements. 8. The 1990 Census' Guangzhou sample includes Guangzhou's city proper (shiqu) and suburban districts (jiaoqu), but not the surrounding counties Guangzhou administers. 9. The most popular regional delineation in China is one es? tablished during the Seventh Five-Year Plan (1986-1990), involving the eastern (coastal) region, which is the most de? veloped and has grown rapidly since the economic reforms, and the central and western regions, which have lagged in economic growth (Figure 1). 10. Detailed definitions of the nine reasons for migration can be found in SSB (1993, 513-14, 558). 11. Multiple responses were permitted for this question, which partly explained the scattered distribution of the responses. 12. Former agricultural workers accounted for a significant pro? portion of temporary migrants (24.3 percent).

References Amsden, H. A. 1980. The economies of women and work. Harmondsworth: Penguin Books. Banister, J., and C. W. Harbaugh. 1992. Rural labor force trends in China. Agricultural and Trade Report (Situation and Out? look Series, Economic Research Service, United States De? partment of Agriculture). Washington, DC: United States Department of Agriculture. Breman, J. 1976. A dualistic labour system? A critique ofthe in? formal sector concept. Economic and Political Weekly 11 (48): 1870-75. Buckley, C. 1995. The myth of managed migration: Migration control and market in the Soviet period. Slavic Review 54 (4): 896-914. Calavita, K. 1992. Inside the state: The Bracero Program, immigra? tion, and the I.N.S. New York: Routledge. Cao, X. 1995. Zhongguo nongcun laodongli liudong yu renkou qianyi yanjiu zongshu (Summary of research on rural labor flows and population migration in China). Nongcun Jingji Yanjiu Cankao (Rural Agricultural Research) 2:23-33.

Chan, K. WMT Liu, and Y. Yang. 1999. Hukou and non-hukou migrations in China: Comparisons and contrasts. Interna? tional Journal of Population Geography 5:425-48. Chan, K. W., and L. Zhang. 1999. The hukou system and ruralurban migration in China: Processes and changes. China Quarterly 160:818-55. Cheng, T, and M. Selden. 1994- The origins and social conse? quences of China's hukou system. The China Quarterly 139:644-68. Christiansen, E 1990. Social division and peasant mobility in Mainland China: The implications of the huk'ou system. Issues and Studies 26 (4): 78-91. -. 1992. "Market transition" in China: The case of the Modern China 18 (1): Jiangsu labor market, 1978-1990. 72-93. Cook, S., and M. Maurer-Fazio. 1999. Introduction to the spe? cial issue on "The workers' state meets the market?labor in China's transition." Journal of Development Studies 35 (3): 1-15. Davis, D. 1999. Self-employment in Shanghai: A research note. China Quarterly 157:22-43. De Haan, A. 1995. Migration in Eastern India?A segmented labor market. Indian Economic and Social History Review 32 (1): 51-93. Duan, J. 1998. Nongcun waichu wugong qingnian di shequ dingwei (Social positions of young peasant migrants). In Waichu Wugong Diaocha Yanjiu Wenji (Collection of papers on surveys of peasant out-migration), ed. Zhongshan Uni? Department, 74-81. Guangzhou, versity Anthropology China: Zhongshan University Anthropology Department. Fan, O O 1996. Economic opportunities and internal migra? tion: A case study of Guangdong Province, China. The Pro? fessional Geographer 48 (1): 28-45. -. 1999. Migration in a socialist transitional economy: Heterogeneity, socioeconomic and spatial characteristics of migrants in China and Guangdong province. International Migration Review 33 (4): 493-515. -. 2001. Migration and labor market returns in urban China: Results from a recent survey in Guangzhou. Envi? ronment and Planning A 33 (3): 479-508. Forthcoming. Permanent migrants, temporary mi? grants, and the labor market in Chinese cities. In Resource management, urbanization, and governance in Hong Kong and the Zhujiang Delta, ed. K. Y. Wong and J. Shen. Hong Kong: Chinese University Press. Farer, T. J. 1995. How the international system copes with in vol? untary migration: Norms: institutions and state practice. Human Rights Quarterly 17 (1): 72-100. French, R. A., and E E. I. Hamilton. 1979. Is there a socialist city? In The socialist city: Spatial structure and urban policy, ed. R. A. French and E E. I. Hamilton, 1-21. Chichester: John Wiley and Sons. Goldstein, A., and S. Guo. 1992. Temporary migration in Shanghai and Beijing. Studies in Comparative International Devebpment 27 (2): 39-56. Goldstein, S., and A. Goldstein. 1991. Permanent and temporary migration differentials in China. Papers ofthe East-West Pop? ulation Institute no. 117. Honolulu: East-West Population Institute. Gonganbu huzheng guanliju (Public Security Bureau). 1997. Quanguo zanzhu renkou tongji ziliao huibian (Tabulations of the national temporary population statistics). Beijing: Zhongguo renmin gongan daxue chubanshe.

The Elite, the Natives,

and the Outsiders

123

Gordon, I. 1995. Migration in a segmented labour market. Transactions, Institute of British Geographers NS 20:139-55. Granovetter, M. 1974. Getting a job: A study of contacts and careers. Cambridge, MA: Harvard University Press. -. 1981. Toward a sociological theory of income differ? ences. In Sociological perspectives on labor markets, ed. I. Berg, 11-47. New York: Academic Press. Gu, S. 1992. Two types of population migration in China: A comparative study. Chinese Journal of Population Science 4 (1): 75-84. lianhe ketizu (Joint Study Guangdong wailai nongmingong Group on Migrant Workers in Guangdong). 1995. Jingying yimin yu xinxing da chengshi zhanlue (Elite migration and the strategies of newly developed large cities). Nongcun Jingji Yanjiu Cankao (Rural Economic Research) 2:1-15. Gupta, M. R. 1993. Rural-urban migration, informal sector, and development policies: A theoretical analysis. Journal of De? velopment Economies 41:137-51. Hanson, S., and G. Pratt. 1991. Job search and the occupational segregation of women. Annals ofthe Association of American Geographers 81 (2): 229-53. Harris, J. R., and M. P. Todaro. 1970. Migration, unemployment, and development: A theoretical analysis. American Eco? nomic Review 60:126-42. Hoffmann, D. L. 1991. Moving to Moscow: Patterns of peasant in-migration during the First Five-Year Plan. Slavic Review 50 (4): 847-57. Kerr, C. 1954. The Balkanization of labor markets. In Labor mo? bility and economic opportunity, ed. E. W. Bakke and P. M. Hauser, 92-110. Cambridge, MA: MIT Press. Knight, J., and L. Song. 1995. Towards a labor market in China. Oxford Review of Economic Policy 11 (4): 97-117. Knight, J., L. Song, and J. Huaibin. 1999. Chinese rural migrants in urban enterprises: Three perspectives. Journal of Devel? opment Studies 35 (3): 73-104. Li, R. 1993. Nan zhongguo diqu laodongli shichang zhuangkuang yu zhujiang sanjiaozhou laodongli shichang di fazhan (Labor market situation in southern China and develop? ment in the Zhujiang Delta). Renkou Yanjiu (Population Research) 1993 (6): 13-18. Li, S. 1995. Population mobility and urban and rural develop? ment in Mainland China. Issues and Studies 31 (9): 37-54. -. 1997. Population migration, regional economic growth, and income determination: A comparative study of Dongguan and Meizhou, China. Urban Studies 34 (7): 999-1026. Lin, G. C. S. 1997. Red capitalism in South China: Growth and de? velopment ofthe Pearl River Delta. Vancouver: University of British Columbia Press. Lo, C. P. 1994. Economic reforms and socialist city structure: A case study of Guangzhou, China. Urban Geography 15:128? 49. Lo, C. P, C. W. Pannell, and R. Welch. 1977. Land use changes and city planning in Shenyang and Canton. Geographical Review 67:268-83. Ma, L. J. C, and B. Xiang. 1998. Native place, migration, and the emergence of peasant enclaves in Beijing. China Quar? terly 155:546-81. Ma, Z., K.-L. Liaw, and Y Zeng. 1997. Migrations in the urbanrural hierarchy of China: Insights from the microdata ofthe 1987 national survey. Environment and Planning A 29:70730. Maurer-Fazio, M. 1995. Building a labor market in China. Current History 94 (593): 285-89.

McGee, T. G. 1982. Labor markets, urban systems, and the ur? banization process in Southeast Asian countries. Papers of the East-West Population Institute 81:1-28. Mitchneck, B. 1991. Geographical and economic determinants of interregional migration in the USSR, 1968-1985. Soviet Geography 33:168-89. Mitchneck, B., and D. A. Plane. 1995a. Migration patterns dur? ing a period of political and economic shocks in the Former Soviet Union: A case study of Yaroslavy' Oblast. The Pro? fessional Geographer 47 (1): 17-30. -. 1995b. Migration and the quasilabor market in Russia. International Regional Science Review 18 (3): 267-88. Nongyebu nongcun jingji yanjiu zhongxin (NNJYZ) (Ministry of Agriculture). 1995. Zhongguo nongcun laodongli liudong: gean fangtan ziliao (Labor Migration in China: Accounts of Individual Interviews). Beijing: Ministry of Agriculture. Pilkington, H. 1998. Migration, displacement, and identity in postSoviet Russia. New York: Routledge. Piore, M. 1979. Birds of passage: Migrant labor in industrial socie? ties. New York: Cambridge University Press. Reich, M., M. Gordon, and R. Edwards. 1973. A theory of labor market segmentation. American Economic Review 63:359-65. Reskin, B., and H. Hartmann. 1986. Womens work, mens work: Sex segregation on the job. Washington, DC: National Acad? emy Press. Roberts, K. D. 1997. China's "tidal wave" of migrant labor: What can we learn from Mexican undocumented migra? tion to the United States? International Migration Review 31 (2): 249-93. Sakamoto, A., and M. D. Chen. 1991. Inequality and attainment in a dual labor market. American Sociological Review 56 (3): 295-308. Schultz, T. W. 1961. Investment in human capital. American Economic Review 51:1-17. Shen, J. 1995. Rural development and rural to urban migration in China 1978-1990. Geoforum 26 (4): 395-409. Shen, J., and N. A. Spence. 1995. Trends in labor supply and the future of employment in China. Environment and Planning C 13:361-77. Shen, Y M., and C. Z. Tong. 1992. Zhongguo renkou qianyi (China's population migration). Beijing: Zhongguo Tongji Chubanshe. Sit, V. F. S. 1989. Hong Kong's new industrial partnership with the Pearl River Delta. Asian Geographer 8 (1/2): 103-15. Sjaastad, L. A. 1962. The costs and returns of human migration. Journal of Political Economy 10 (Supplement): 80-93. Solinger, D. J. 1999a. Contesting citizenship in urban China: Peas? ant migrants, the state, and the bgic of the market. Berkeley: University of California Press. -. 1999b. Citizenship issues in China's internal migration: Comparisons with Germany and Japan. Political Science Quarterly 114(3): 455-78. State Statistical Bureau (SSB). 1993. Zhongguo 1990 nien renkou pucha ziliao (Tabulation of the 1990 population census of the People's Republic of China). Vol. 4. Beijing: Zhongguo Tongji Chubanshe. -. 1997. Zhongguo l % renkou chouyang tiaocha ziliao (Tab? ulation of the 1995 One-Percent Population Sample Survey ofthe People's Republic of China). Beijing: Zhongguo Tongji Chubanshe. Stinner, W. E, W Xu, and J. Wei. 1993. Migrant status and labor market outcomes in urban and rural Hebei Province, China. Rural Sociology 58 (3): 366-86.

124

Fan

Tang, W. S., D. K. Y. Chu, and C. C. Fan. 1993. Economic re? form and regional development in China in the 21st cen? tury. In Pacific Asia in the twenty-first century: Geographical and developmental perspectives, ed. Y M. Yeung, 105-33. Hong Kong: The Chinese University Press. Todaro, M. P. 1969. A model of labor migration and urban unemployment in less developed countries. American Eco? nomic Review 59:138-48. -. 1976. Internal migration in developing countries: A review of theory, evidence, methodology, and researchpriorities. Geneva: International Labour Office. Waldinger, R. 1992. Taking care ofthe guests: The impact of im? migrants on services. International Journal of Urban and Re? gional Research 16:97-113. Wang, F. 1997. The breakdown of a great wall: Recent changes in the household registration system of China. In Floating population and migration in China: The impacts of economic re? forms, ed. T. Scharping, 149-65. Hamburg: Institute of Asian Studies. -. 2000. Gendered migration and the migration of genders in contemporary China. In Redrawing boundaries: Work, household, and gender in China, ed. B. Entwisle and G. Henderson, 231-42. Berkeley: University of California Press. Wolpert, J. 1965. Behavioral aspects ofthe decision to migrate. Papers and Proceedings of the Regional Science Association 15:159-69. Wong, L., and W.-P. Huen. 1998. Reforming the household reg? istration system: A preliminary glimpse of the blue chop

household registration system in Shanghai and Shenzhen. International Migration Review 32 (4): 974-94. Wright, R., and M. Ellis. 1997. Nativity, ethnicity, and the evo? lution of the intraurban division of labor in metropolitan Los Angeles. Urban Geography 18:243-63. -. 2000. The ethnic and gender division of labor com? pared among immigrants to Los Angeles. International Jour? nal of Urban and Regional Research 24 (3): 583-600. Yang, Q., and F. Guo. 1996. Occupational attainment of rural to urban temporary economic migrants in China, 1985-1990. International Migration Review 30 (3): 771-87. Yang, X. 1993. Household registration, economic reform, and migration. International Migration Review 27 (4): 796-818. -. 2000. Determinants of migration intentions in Hubei province, China: Individual vs. family migration. Environ? ment and Planning A 32 (5): 769-87. Yang, X., and F. Guo. 1999. Gender differences in determinants of temporary labor migration in China: A multilevel analy? sis. International Migration Review 33 (4): 929-53. Yang, Y 1994. Zhongguo renkou qianyi yu fanzhan di chanqi zhanlue (Long-term strategies of population migration and de? velopment in China). Wuhan, China: Wuhan Chubanshe. Zhou, D. 1998. Zhongguo nongmingong di liudong: nongmingong shurudi yu shuchudi bijiao (Migration of peasants in China: comparison of origins and destinations). In Waichu Wugong Diaocha Yanjiu Wenji (Collection of papers on sur? veys of peasant out-migration), ed. Zhongshan University Guangzhou, China: Department, 3-23. Anthropology Zhongshan University Anthropology Department.

Correspondence: Department of Geography, University of California, Los Angeles, CA 90095-1524, e-mail: fan@geog.ucla.edu