China ’ s Rural-Urban Migration : Structure and Gender Attributes of the Floating Rural Labor Force

This article focuses on Chinese female rural migrant workers. Based on the survey data collected in Anhui and Sichuan provinces of China, the article investigates gender aspects of Chinese rural-urban migrants in the context of the household migration model. Results of the examination indicate that the rapid economic development that China is undergoing makes it possible for rural women to go from traditional villages to modern and post-modern cities and gain employment in urban areas and VTEs (village and township enterprises). The young, unmarried, and well-educated rural women are much more likely to migrate or to get off-farm jobs, and the occupations held by these rural female workers are more varied. There are gender features among the rural-urban migrant workers. The proportion of the returnees among married women is much higher than that of their male counterparts. Income also differs between female and male laborers. While the proportion of women in the lower income group is higher than men, the proportion of female workers in the higher income groups is lower. The lack of social security for rural residents proves to be a highly signifi cant institutional barrier for rural women to access the new opportunities offered by a more open economy and the modernization process. On a more general level, the results of the study suggest that market-oriented economic reform brought about diverse effects on Chinese women in terms of labor market status. Though the institutional barriers put Chinese female rural workers in a position of disadvantage, the performance of female rural-urban migrant workers suggests that they are active beings rather than passive victims merely adapting to the social transformation. Female rural-urban migrant workers have been and still are playing important and specifi c roles at the crossover between the emerging capitalist economy and the traditional rural society. In doing so they are positive participants of globalization in a wider development perspective.


Introduction
According to the new economics of labor migration theory, migration is hypothesized to be an effort made by households to overcome market failures that constrain local production (Taylor 1999).This approach challenges the conclusions of neoclassical economic theoretical models and argues that migration does not simply involve individual rational choices to migrate or not to migrate.This theory assumes that migration decisions are typically made by social units such as families or households.For instance, Oded Stark (1982) argues that, in less-developed countries, rural-urban migration might be undertaken primarily to improve an individual's or a household's comparative income position relative to the reference individuals or households.On a general level, J. Edward Taylor (1999) argues that, 'market failures force rural farming households to self-fi nance their production and to self-insure against income risk … migration provides their households with liquidity, in the form of remittances, which may be used to fi nance new production technologies, inputs and activities.They also offer income insurance, by providing households with access to an income source'.In this sense, in the new economics of labor migration theory, the households are able to make more complex migration plans due to the possible diversifi cation of labor among the household members.
Empirical research on China's rural-urban migration has suggested that Chinese migration reveals certain characteristics that are postulated by mainstream internal and international migration theories (Chan 1999, 65).The common understanding of China's rural-urban migration model is that the Chinese case, to a certain extent, can be explained by developmentalist theory which assumes that migration decisions are a part of family strategies to raise income and insure against risks (Huang 1996;Cai 2000;Guan and Guo 1997).Under the so-called 'household responsibility system', a farmer's household is an integrated decision-making unit that seeks household rather than individual utility-maximization (Meng 2000).Migration decisions are made by households instead of individuals in order to benefi t whole families' particular household economic gains (Bai and He 2002;Du 1997b;RTMA 1995;Zhou 2001).In a similar vein, the principle of maximizing family interests affects the decision to return.In particular, females returning to the countryside represents the same decision-making pattern employed by the household in favor of collective well-being (Luo 2001; In this study, the concept of the 'rural migrant worker' refers to a group in the Chinese labor force who move between rural and urban areas, seek and fi nd temporary jobs in non-agricultural sectors in urban areas, and when out of work, return to their home villages and engage in farming again.According to China's Hukou household registration system, this group of laborers maintain permanent rural resident status whenever and wherever they work.In Chinese literature, both in scientifi c analyses and policy documents, the group of rural-urban migrants without a Hukou transfer is considered to be the so-called 'fl oating rural laborer' (Nongcun liudong laodongli or Nongmin gong) instead of the 'migrant laborer'.In this paper, I use the term 'migrants' in line with the mainstream term in migration studies.
The paper will apply the family migration approach in order to explore who migrates to urban areas and who stays or returns to the home villages under the household migration decision model; in particular, whether the attributes of individuals affect rural workers' migration experiences and labor market participation opportunities in the context of family strategies.Among those individual attributes, I focus on the usual demographic classifi cations used in social science analyses: age, gender, educational attainment and marital status.The migration experiences and labor market opportunities are indicated by migration duration, income and non-agricultural occupations.The article will examine the relationship between these attributes and rural female workers migration and tries to answer the following questions: • Who are the rural-urban migrants?What are the common features and differences among the migrants?
• Do rural female workers have access to the new opportunities?Who benefi ts most and least from the possibilities to migrate?
• Are there differences between genders in the moving patterns?If there are, in which aspects do they exist and what is the impact on the migration decision?
• Are there gender-specifi c factors that prohibit rural female workers' migration and employment in urban areas under the family migration model?
The data used in this paper were collected during 1999 and 2000 in Anhui and Sichuan provinces1 of China as a part of a research project on rural-urban migrant workers in China -'Study on Out-Migrants and Return Migrants' (SOMRM).The project was conducted by the Research Center for Rural Economy, Ministry of Agriculture, from 1997 to 2001.The author of the paper was a member of the research team.The data for the project was collected using case interviews and sample surveys.The analysis of this paper relies primarily on the survey data.
The survey was based on a sample.The sampling system was developed by the National Bureau of Statistics (NBOS) and the Ministry of Labour and Social Security (MOLSS).
The method was designed in order to collect comprehensive information on rural labor force and mobility.The systematic samples covered 31 provinces, 857 counties, 5,000 townships, 8,000 villages, and all in all 70,000 rural households (MOLSS & NBOS 1999).The system started in 1988, and is still the only national panel survey on rural laborers in China.Yet this national sampling system has some weaknesses.For example, the questionnaire defi nes some variables in too-broad categories.For an extensive analysis on some specifi c issues, it is still, at the moment, the best available survey approach for academic and policy studies on rural labor force.In order to overcome the limitations of the national sampling system's data, complementary questionnaires were attached in the SOMRM project.This was expected to support a more detailed analysis of the situation and trends of rural-urban migrant labor's outgoing and return (Bai, N. and He, Y. 2002).The survey was based on rural household registration records in Anhui and Sichuan provinces, from which individual demographic and rural household economic data were taken.This strategy allowed the SOMRM to conduct a scientifi cally valid study on the subject.The SOMRM data base contains two separate parts: the fi rst is the data set on total rural population and labor force which covered 28,957 individuals from 71 counties in Anhui and Sichuan provinces (including the 62 counties presented below); the second is the data set on migrant workers, the returnees, and the locally transferred rural labor2 .The second data set covered 15,101 workers from 62 counties in the two provinces.The SOMPM data base represented the rural labor force in the two provinces relatively well, in particular considering the size of the samples3 , and it provided ample material for reliable research on rural-urban migrant workers.
The cross-tabulation and chi square test are employed in the paper to analyze the structure and gender attributes of the fl oating rural labor force.This method is based on the difference between the expected and observed observations.Many observations were deleted from the estimation procedure due to missing values, resulting in purged data sets ranging a wide span, depending on the exact set of variables chosen for analysis.
The main source of missing values was rural-urban migratory information in detail and other relevant variables in the data set of 71 counties.In those cases, the 62 counties data set was utilized (with the sign (*) in the text).This strategy insured the validity of the data because a relatively big sample size still remained for fl exible research, and it also allows analysis of the gender aspects of Chinese rural-urban migrants.
The paper will fi rst draw out some background features of the fl oating labor force at a macro level, and will then present the general profi le of rural-urban migrants in the two provinces and describe the features of female rural workers in terms of occupational status.Following that it will examine whether there are gender-based differences among migrants.The penultimate section identifies and explains the gender-specifi c factors that affect families' migration decisions.The fi nal section will draw conclusions and offer discussion.

Background of the surge of a fl oating labor force
Since the late 1970s, China has been undergoing a gradual transformation from a centrally planned command economy to a market-based system; this great transformation brings rapid economic growth and social changes.One signifi cant aspect is that there has been an increasing number of rural-urban migrants, including massive numbers of rural women moving from their home villages to distant urban areas to fi nd jobs.By 2003, the amount of rural-urban migrant laborers was as high as 114 million, which accounted for over 20% of the total 500 million rural labourers (NBOS 2004)4 .Due to a lack of offi cial national statistics, various surveys on female rural-urban migrant workers suggest estimates ranging from 30% to 35% of total rural-urban migrant laborers during the 1990s (Bai and He 2002;Zhou 2001).At the beginning of the new millennium, about 50% of all migrant workers in China were estimated to be women (UNRISD 2005, 83).Rural-urban migrants shape the largest population movement during peacetime in China and perhaps the largest movement in the whole of world history (Robert 2000).
During the era of the centrally planned economy, the Chinese government rigidly controlled labor mobility and migration, especially the residence changes from rural to urban areas and job transfers from farming to non-agricultural work.It was almost impossible to migrate from rural to urban areas, except when permission was rarely granted by the authorities.This control has functioned through the household registration system (Hukou system), which is a unique institutional arrangement that strictly segregates rural and urban areas (Cai 2003).The Hukou system was set up in the mid-1950s.It functioned as a domestic passport preventing rural residents from entering cities, and rural laborers were excluded from working in non-agricultural sectors 5 .Under this system, an urban Hukou membership was required in order to stay in cities and obtain employment.The urban Hukou status included a series of social entitlements (Song, Huang and Liu 2002) like food quotas, jobs assigned by the government, as well as associated occupational welfare benefi ts such as free housing and free health care.Hukou membership also guaranteed privileged access to urban public services such as education.
Since economic reform started in the late 1970s, the system of the centrally planned economy has changed in many respects.Up to the mid-1980s, the rigid controls on rural-urban migration were gradually eased.As a consequence of the market-oriented reform in labor policy in particular, food and housing provision in urban areas, employment, housing conditions and food supply regulations in cities were gradually changed.
Hence, more and more rural labourers are able to move to cities to fi nd employment.This movement led to the so-called phenomenon of 'the surge of the fl oating rural labor force' (Mingong Chao) (Du and Bai 1997, 2;Cai 2000).
The surge of the fl oating labor force (Mingong Chao) is a result of various push and pull factors (Du 1997a;Huang 1996;Bai and He 2002).It is widely accepted that the surge of China's rural-urban migration is a grass-roots reaction to the enduring poverty in rural areas.Poverty is a strong push factor in changing location.Other push factors in China consisted mainly of two phenomena.The fi rst was that tens of millions of surplus rural laborers were released from agricultural production when rural economic reform started in the late 1970s.The current number of surplus rural workers is estimated to range from 150 million to 170 million people, which accounts for 30-35% of the total rural labor force (Liu, J. 2002, 74).By 2017, the surplus rural laborers will still exceed 100 million workers (Du 1997b).Such massive unemployment and under-unemployment among the rural labor force creates a vast group of job-seekers looking for employment in cities.The second push factor was the decreasing income from agricultural activities.The proportion of farmers' agricultural income to their total income has continuously declined since the 1980s.In 1985, agricultural income accounted for 75% of farmers' total income, whereas the proportion had decreased to 50% in 2000 (Hu, A., Hu, L. and Chang 2003).An attempt to fi nd non-agricultural employment has become the main device for rural residents to try to maintain their income level.
The main pull factor is the demand for rural migrant laborers in urban areas.This demand has continued to increase since the economic reform started.One important reason for this relates to the construction of Chinese social policy.As rural-urban migrants are not covered by the formal social security system, urban employers do not need to pay mandatory social insurance contributions for the migrant workers.Employers' social security contributions total 30%-35% of payroll (Song and Gao 2001, 15); therefore, migrant labor is much cheaper than urban labor.Consequently, the new unskilled and low-paid jobs created in the market economy reform period were basically fi lled by rural migrants (Du 1997a;Meng 2000, 143;Bai and He 2002).
The increasing rural-urban income disparities, which have expanded during the last two decades, have fortifi ed the process.According to some estimates, the calculated ratio of per capita disposable income of urban households to that of rural households was 2.6 in 1978, increasing to 2.9 in 2001 (Hu, A., Hu, L. and Chang 2004).Offi cial estimates indicated that if various social welfare benefi ts and various subsidies paid to urban residents are taken into consideration, the actual per capita disposable income of the urban population is 6 times higher than that of rural households (Chen 2002).Justin Y. Lin and his colleagues' study (Lin, Wang and Zhao 2004) indicated a smaller disparity: the income ratio between urban and rural people was 1.8 in 1985, and increased to 2.4 in 2000.Although there are signifi cant differences between various estimates, the trend is the same in all statistics.Disparities have increased according to the World Bank: rural-urban disparities accounted for more than 50% of the overall inequality in China in 1995 and they explained 75% of the total increase in inequalities between 1984and 1995(World Bank 1997c).In the same period, migratory activities also increased tremendously.Higher incomes in the place of destination compared to the place of origin accelerate geographical mobility and raise migration rates.
Since the mid-1980s, the central government restrictions on the overall rural-urban migration were gradually eased (He and Bai 2002, 5).However, offi cial permission for rural laborers to move freely to urban areas has never been given.Institutional barriers to rural-urban migration also remain effective in some ways.Local urban governments have introduced various policy measures to block rural migrants from getting jobs in these localities (Song, Huang and Liu 2002, 168).For example, cities usually have rigid regulations that restrict urban employers from recruiting rural migrant workers; hence they only qualify for certain jobs for which it is impossible to hire local urban residents (Du 1997b).Until 2000, the central government tried to introduce certain policy changes and made attempts to abolish local discriminatory regulations (Song et al. 2002).However, the pervasive legal restrictions for rural-urban migrants still exist.Under current policies, rural-urban migrants have access to certain occupations in urban areas, but are nevertheless much like 'foreign laborers' (Wailai mingong) working in the cities, and only in a very few cases can rural resident status be changed to urban status.All in all, this means that most rural migrants remain excluded from the urban social security system (Luo 2000; Song, Huang and Liu 2002), which means that they have to move from the city to the countryside when unemployment, work injury, sickness, or anything that may disenable them to work strikes.This situation forces rural-urban migrants to move frequently between rural and urban areas.

The general characteristics of sampled rural labor
The rural workers studied were divided into four groups on the basis of their migration experience: the existing migrant worker, the returned worker, the off-farm worker in localities, and the non-migrant worker.In this investigation, the 'existing migrant worker' refers to workers who had gone out to work for at least three months in the year 1999 when the investigation was conducted.'Return workers' refers to those who had migrant experiences before 1999 but had returned to their home villages in 1999 (and worked in either agricultural or non-agricultural sectors in the countryside).The 'off-farm workers' (Jiudizhuanyi laodongli) in the localities includes workers who were engaged in non-agricultural jobs in their home villages in 1999 and those who had migrated for less than three months in 1999.'Non-migrant worker' refers those who without any migration experience.
Table 1 shows that the share of rural workers with migrant experiences is more than 20% of total rural labor force in the two provinces.6Meanwhile, the numbers of rural workers who are engaged in non-agricultural sectors constitutes 30% of the total rural labor force studied; the number of return workers makes up only 6% of the total rural labor.
Regarding the age among the various groups, existing migrant workers are on average approximately 10 years younger than the others, while the return migrants are closer in age to the total rural workers as well as non-migrant workers.The average level of education of the total rural labor targeted is 7.3 years, which is equivalent to the fi rst grade of Chinese secondary school.It is clear that the rural workers who engaged in non-agricultural jobs are well-educated compared to the others.The average educational attainment of existing migrants and off-farm workers in the localities is equivalent to the third grade of secondary school, whilst that of return migrants is barely above fi rst grade.

Local distribution of outgoing and return rural laborers
The results show that there is a remarkable imbalance in the distribution of out-migrant workers across the two provinces.Among the 62 counties investigated, there are seven counties in which out-migrant workers constitute over 25% (as a proportion of the total rural labor force in the county), with the highest proportion being 36%; fi fteen counties at 20%-25%; eleven counties at 15%-20%; ten counties at 10%-15%; ten counties at 5%-10%; and nine counties at less than 5%.There is also a great imbalance in the distribution of return migrant workers across different parts of the two provinces.Huang and Carrel 1998).In the localities that are rich in natural resources, agricultural production and income can maintain a basic level and fewer villagers migrate, and in the localities that have few natural resources more people migrate.
In addition, this study reveals that there is a new cultural pattern of migration in the villages where there is a long history of migration and greater numbers of people who migrate.In those villages, migration, as the main tool in combating poverty, has become part of the local tradition.This argument is supported strongly by the case study material.We learned from interview data that in these villages, even during the Mao period, there were still some rural workers who tried to evade the rigorous restrictions in order to fi nd jobs in urban areas.Nowadays, a kind of 'migration culture' has developed among the villagers.Migration is regarded as a necessary experience for young people: an expression/representation of individual capabilities, and of course the most available measure to improve their own and their family well-being.Thus, out-migration has to some extent turned into an indicator of one's individual capability and moral standpoint; if persons are not too old or in ill health to move to cities, they are expected to migrate, or they should at least try to do so.As a matter of fact, during the interviews, the respondents who never migrated usually expressed abashment over that fact and consistently tried to offer excuse for it.
With regard to the difference in the number of returnees in the counties under investigation, our research indicated that the main factors are the employment situation and government policies in migration destinations (Bai and He 2002;Song and Zhao 2002).This implies that the share of return migrants is more or less evenly distributed across the counties, or more evenly than the numbers of those who migrate, which is more dependent on local circumstances.The comparison of out-and return migration to some extent supports this hypothesis.The range of variation in out-migration between the 62 counties is from 2% to 36%, whereas it is only between 2% and 15% in return migration.However, one can contemplate why there is still such a large variation, if the working situation in the destination rather than the home village is decisive.Based on the fi ndings of our case study and previous theoretical model and empirical analysis on mass migration, this study offers an answer to this question.Due to the nature of the informal migration network, villagers within the region usually move to the same urban areas and the jobs concentrate in certain sectors (see Fields 1975;Du 1997a;Song 2001).Therefore, changes in the economic situation and employment policies present in migration destinations turn out to be the determining factors for the total amount of out-migrants in that area during the entire period.The interview data also revealed that urban local government labor and social policies are the most crucial factors to affect the employment situation of migrants in these destinations.The more exclusive or constraining the policies for rural migrants are, the more migrant workers return to their home villages.

Gender distribution
Table 2 shows that among the total rural labor force, the proportion of female and male laborers is more or less even (48% vs. 52%).However, a clear imbalance appears when we look at migrant laborers.Among the existing migrant laborers, the proportion of female to male workers is 30% to 70%, and the proportion of female to male return laborers is about 40% to 60%.The differences are statistically signifi cant.The chi squared (χ²) value for a six-fold table -female and male vs. migrant, returnee, never-migrated -is 265.081 at 2 degrees of freedom (df).This means female migrant workers are much fewer in number than male migrant workers, but female migrants return home more than male migrants.The interesting question is why such an imbalance exists.Are working conditions for the female employees much worse than for males in the migration destination?Is the home village more appealing for women than men?These questions will be answered shortly.

Rural female laborers: employment and mobility
I now turn to another theme of this article, rural women's labor market status in the context of household migration strategy.I will answer the question of why there is an imbalance in the gender distribution in migration.In this section, the analysis treats the rural household as a decision unit capable of choosing who among the family members should migrate.Here the total sample of rural female labor force as described above was regarded as the unit for analysis, and it was divided into four groups according to their migration and labor market status: (1) Existing female migrant labor force (2) Female workers in village and township enterprises (VTEs) -among this group, some women had been migrant workers in a city for a period but currently are not: when the survey was conducted they were employed in the locality (3) Return female migrants (4) Non-migrant female workers (without any migrant or non-farming work experience).
In order to give a general picture of rural female laborers' occupational situation, particularly the migrant female workers, these four groups will be analyzed according to the following three considerations: age distribution, educational situation and occupational situation.The fi rst two fi rst factors reveal whether these individual attributes affect rural female workers' labor market status when the household makes the decision on who should move out and get off-farm employment, and who should remain in family-based agricultural work.The latter serves as a characteristic to display rural women's labor market participation opportunities as a consequence of the family migration model.

Age distribution among different rural female worker groups
Table 3 presents the main demographic characteristics in each rural female labor group.We can draw some preliminary and intuitively reasonable hypotheses based on the characteristics of migrant women.Since there is huge disparity in the developmental level between the origin village and the destination city, we can suppose that younger and well-educated rural women are the most prone to migration.In contrast, the age of the return female labor group resembles that of the non-migrant female labour group.The majority of the workers in these two categories are between 26 and 46 years old, and the numbers of the returnees and non-migrant female workers increase as the age-span increases.On the other hand, while 50% of rural women above 46 years of age are engaged in farming work, only less than 15% of them have jobs in non-agricultural sectors.This means that rural-urban migration is a juvenile phenomenon and that relatively old women are left in the countryside in charge of agricultural work under the household migration decision model.This is quite a reasonable and not very surprising result.In the modern and post-modern surroundings to which the migrating women must adapt, younger people are better at adapting.Furthermore, and more importantly, young workers are usually more mobile due to their family situation.

Educational level of rural female workers
We can also fi nd some interesting traits in educational attainment distribution among the four groups.Table 4 shows that fi rstly, the degree of education received by existing female migrant workers and VTE employees is signifi cantly higher than that of the return and non-migrant labor groups.In other words, in the former two groups, workers with more than seven years of education make up about 70% and 80% of the total number of female workers in each group respectively, whereas in the latter two groups the majority received only less than six years of education, and about 15%-20% of them never attended school.Secondly, about one quarter of the female VTE employees have received more than ten years of education.In this category of rural female workers, the percentage of the rural female labor force with more than 10 years of education is much higher than in the other three groups of laborers.Thirdly, in all of the four groups, there is a considerably higher percentage of illiterate persons, particularly in the return and non-migrant groups, with the proportion of illiterate and semi-illiterate individuals approaching one quarter of the non-migrant female workers and one-fi fth of the return group.The low educational level of the returnees may indicate that it is not only the employment situation in the destination city that is of importance.The migrants with lower qualifi cations cannot cope in new circumstances that demand various skills that they do not have.

The occupational situation of rural female workers
To provide more information on the occupational situation of rural female workers, Table 5 presents the occupational distribution for fi ve groups in the female rural labor force.It is unfortunate that the survey questionnaire defi nes occupations in very broad categories; however, within each of the broader categories, occupational differences can nevertheless be observed.
Table 5 shows certain features of the occupation distribution: (1) The occupations that are taken up by female migrant workers and VTE employees consist of a comparatively wide variety.Among these occupations, industry and service jobs hold the largest percentage of employment.In contrast, an overwhelming majority of the return labor group (return 2, 98%) and the nonmigrant labor (97%) are concentrated in agricultural work, including farming and rearing livestock.
That means after returning to their home villages, most female migrants go back to traditional work and life.However, does this mean that nothing changed?It may be that the form of occupational structure is preserved but the content of the old occupations might be changed.The survey data cannot answer this question at this juncture, but it will be examined using interview data.And in fact, the interview data will indicate that despite 'non-changes' in occupational classifi cations, there are important differences in activities within occupational groups.
(2) There is an observed difference between return migrants and existing migrants; that is, the proportion employed in construction and transportation industry among the female returnees is much higher than among the existing female migrants (36% vs. 6%).A possible explanation for this result is that for migrant workers, employment in the construction and transportation industries in cities was highly uncertain.One reason is that in China's case, these two categories of industry as well as the associated job opportunities are directly affected by government policies and economic circumstances (Song et al. 2002).Therefore, if negative changes in policies and economic circumstances take place in the destination location, great numbers of migrants employed in the construction and transportation industry would lose their jobs and would have to go back to the countryside.Here the situation tallies with the imbalance in the share of return migrants across migrant-sending counties.It demonstrates that again, on the whole, non-agricultural employment opportunities for female migrants depend chiefl y on state policies.
Are there differences among migrants that are determined by gender?

The duration of migration
Table 6 shows that in general, the duration of the return female workers is close to that of the male group.The most common duration is two years.However, there are two slight distinctions.
One is that among men, 17% stay one year and 16% stay 5-14 years in cities while amongst women the fi gure is 22% and 11%, respectively.Another slight difference exists in the average overall migration period, which is 3.1 years among men, while it is only 2.7 years among women.All in all, this shows that women stay for shorter periods outside their home villages than men.
Here the migration period appears to be connected to the issue of the cost of migration.Some labor migration analysts have pointed out that when migrants make migration decisions, they usually weigh expected benefi ts and possible costs (Mundlak 1979;Todaro [1969Todaro [ ]1984;;Sjastaad 1962).In the Chinese case, rural-urban migration is also supposed to be a rational decision-making process following the principle of maximizing benefi ts.Due to the household register system and some local restraining rural-urban migrant polices, the migration costs for rural workers are quite high; for example, migrants must buy various migration permits in both the countryside and the cities, therefore the rural workers must take measures to minimize the migration costs and maximize the benefi ts during the migration (Guan and Guo 1997;Zhao, Y. 1997;Bai and He 2002).This study suggests that there are two forces pulling the migrant workers in opposite directions: on the one hand, this situation forces the rural migrant workers to try to stay in urban areas at least for a certain period to earn enough money to make the expenses and the income earned in the city worth it -otherwise their migration would not 'get more kicks than halfpence' (Du 1997b;Bai and He 2002).
On the other hand, the extremely poor working and living conditions the rural-urban migrants suffer during their migration make most of the migrants return home as soon as possible, when they have earned 'adequate' money to somehow cover expenses and fulfi l expectations.It appears that two to four years is the 'best' duration for migration and the appropriate length of time to balance the costs and benefi ts of migration for most rural workers, regardless of their gender.

Income and economic contribution to the household
In respect to migrants' income, there are two possible questions to be clarifi ed.Firstly, due to the lack of comparable income data for different rural labor groups, in this study I use the 'amount of money the worker sent home or took home', i.e. remittance, as an indicator to display the mobile worker's income, as it is very common for rural-urban migrants to send most of their wages back home -almost all of the savings besides their minimum basic live expenses -to support families, build a house and/or prepare for marriage etc. (Zhao, C. 1997;Li, X. 2001).Therefore, 'the money they send home' is more or less an accurate proxy for the relative income level, and is in any case a direct indicator in evaluating female workers' economic contribution to their families.Secondly, only two of the rural labor groups were represented for a comparison of the degree to which female laborers' wages contribute to the household income.This is because in the return and non-migrant female labor groups, the majority are only engaged in family-based agricultural work and it is impossible to identify those women's individual income in relation to the entire household income.Therefore, the comparison was conducted between the existing migrant group and the locally transferred labor group, both of which have identifi able individual wages.
The fi ndings presented in Table 7 show that among these two female groups, those who sent home on average between 100 and 400 Yuan/month (approx.12-50 euros), occupy the majority (over 70%), and within this majority, those who sent 200-300 Yuan/month (about 25-37 euros) make up the highest percentage of the two groups, i.e. about 30% of the two female labor groups.(female vs. male migrants: χ² = 601.891;df = 6; sig.= .000;female migrants vs. female locally transferred labor: : χ² = 31.02;df = 6; sig.= .000* 62 counties data. In the meantime, the income data also demonstrate that both of the two aforementioned female groups have a much higher percentage of low-income earners (who sent home less than 100 Yuan/month) than their male counterparts.In the existing migrant labor groups, the percentage of lower-income earners among females is 10% higher than among males; in the two locally-transferred labor groups it is 6% higher.In contrast, in terms of the high-income earners (over 500 Yuan sent home per month, approximately 63 euros), both of the female groups have a percentage lower than that of the two male groups.In particular, in the groups who sent home 500 Yuan/month, the percentage of the existing female migrant group is 7% lower and the locally-transferred female group 8% lower than among their male counterparts.

The barriers for rural female workers' migration: within and beyond the household migration model
Some researchers have argued that the labor market in China is characterized more and more by its gender equality, has stronger, younger and more educated workers working off the farm and fewer barriers for women (Rozelle, Brauw, Zhang et al. 2001).Perhaps it is true in terms of the law or government regulations, as no offi cial statement that directly impedes rural women's migration and getting off-farm employment can be found.However, this study fi nds that there are still institutional barriers that hinder rural female migrants' labor market participation.The obstructive effect of these factors is magnifi ed particularly in the household migration model.The information presented in subsequent chapters will demonstrate this argument.Here I explore the relationship between marital status and rural-urban migration, as well as the effects of family responsibilities on female workers' migration.
This issue is fi rst examined by looking at the differences between male and female rural workers in terms of their proportion of all migrants.It is indicated by the sample data, which refl ects that among rural female workers, the proportion of migrants is much lower than among male workers.The number of existing and former female migrant laborers is 15% of total rural female labor, remarkably lower than of male migrant labor (28%).At present, the number of existing female migrant laborers is still remarkably lower than that of male migrants, 10% vs. 21%.On the other hand, among the female workers the proportion of returnees is higher than among the male workers.The male returnees constitute only one quarter (25%) of the total number of the male laborers who have migrant experience, whereas the female returnees make up more than one-third (35%) of the total female workers with migrant experience.The difference in proportions of male and female migrant workers illustrates that it must be much harder for women to decide to leave in the fi rst instance.Hence, it is clear that obstacles exist that hinder rural female workers from moving to urban areas in some way.

Explanation for different proportions: women's family obligations
To understand the key factors that infl uence female migrants' labor market participation under the household migration model, I fi rst present the reasons for migrants making the decision to return, and then analyze the relationship between the decision to return on selected background attributes/variables such as gender, marital status and the mobility of rural workers.
Eight options were given in the questionnaire as the reasons for moving back: 1) too old to keep on migrating, 2) preparation for marriage, 3) bearing and rearing children, 4) taking care of parents or other family members, 5) illness or work injury, 6) diffi culty fi nding jobs in urban areas, 7) went back to home town to invest, 8) other reasons.The respondent had to name one to three of the most important reasons for his/her return.The primary reason for returning was diffi culties in fi nding employment in cities -as many as 57% of the respondents gave this as their primary motive for moving back.
The second most common reason was the responsibility of taking care of parents and family members (51%), and the third was bearing and bringing up children (24%), followed by getting married (22%), getting old (15%) and being injured or sick (7%).
In addition, there were about 3% of the total number of return workers who made an investment in their hometown (Source: RCRE, MOA, Survey data set of 'A Study on Out-Migrants and Return Migrants', Result hh 99p+lab_b_new.xls, 2000, Beijing) Obviously, the decision to return is not affected by one single factor but is a consequence of many combined factors.However, upon closer inspection, we fi nd that except for the lack of available jobs in urban areas, the attempt to invest in the hometown, and getting married, all of the other reasons are related to a situation of lack of social protection, indicated by family responsibilities.Sickness and work injury also belong to this category.If we suppose that the responsibility for children -i.e. it being impossible to combine work and family -is an indication of social security problems, we are up to 80%, which means that in most cases problems with social security are important reasons for people to move back to their home village.
Table 8 shows that among the total rural labor force, the majority are married; unmarried workers constitute less than one-fi fth (18%).In contrast, unmarried workers correspond to almost half (49%) of the existing migrant workers.Migrants vs. non-migrants: χ² = 1182.051;df = 2; sig.= .000;returnees vs. non-migrants: χ² = 7.531; df.= 2; sig.= .050 The proportion of unmarried persons is a little lower among the return group (12%) than the non-migrant labor group (15%).The difference between non-migrants and existing migrants is statistically highly signifi cant, whereas the difference between returnees and non-migrants is somewhat signifi cant.When both the existing and former migrant labor force is taken into consideration, the unmarried proportion is about 37%.
In addition, among the existing and former migrant group, about 47% of the married rural workers have returned, whereas one-tenth (11%) of the unmarried workers have returned.This means that marital status has an important impact on the migrant's decision to move back, and migration relates strongly to the unmarried (see also table 9).When the gender variable was introduced to answer who is the most likely to return, it is clear that married female workers would belong to that group.Table 9 shows that among the unmarried labor groups, the gender structure is very similar among migrants, returnees and non-migrants, and there are almost no differences in gender distribution in these groups.However, when we look at the married groups, substantial differences appear.Table 10 indicates that married female workers constitute only one quarter of the total married existing migrants, while they make up over half (53%) of the non-migrant labor group.This means that marital status is a crucial factor in the mobility of female migrants.This attribute can be observed more clearly in Figure 1.χ² = 307.913;df = 2; sig.= .000.
Among the unmarried, regardless of gender, there are about 10% return workers, with unmarried male return workers making up 10% and unmarried females 12%.In contrast, among married migrants, while only 41% of the married males returned, more than half (60%) of the married female workers had already returned.Both among the male and females, differences between married and unmarried groups are statistically very signifi cant.
This pattern shows that marital status is an important factor affecting migrants; in other words, family responsibilities draw migrants back to the countryside under the household migration model.The fact that there is a higher proportion of married female workers who return to the countryside indicates that marriage is a hindrance for rural women's migration.On the one hand, this is a consequence of the family migration decision model, because rural women increase their family obligations when they get married.On the other hand, the extreme lack of any formal social security in rural China proves to be one of the main institutional barriers to prevent rural females from migrating.As a result of this institutional barrier, a large number of married women have to remain in or return to their home village to fulfi l the traditional family role.

Conclusions and discussion
This paper investigated gender aspects of Chinese rural-urban migrants in the context of the household migration model.It was found that the rapid economic development that China is undergoing makes it possible for rural women to go from traditional villages to modern and post-modern cities and gain employment in urban areas and VTEs with better wages than they get from farm work.Among our sample of rural laborers in Anhui and Sichuan provinces, female workers accounted for one-third of the total number of rural-urban migrants.This non-agricultural employment provides opportunities for rural women to get out of the previous situation of extreme economic poverty from which both they and their families suffer.This suggests that, generally speaking, rural female workers benefi t from China's contemporary social transformation and modernization.Female: χ² = 220.384;df = 1; sig.= .000;Male: χ² = 176.217;df.= 1; sig.= .000.
Among the different groups of rural females, depending on their age, educational attainment, and marital status, there are differences in terms of access to the new opportunities offered by a more open economy and the modernization process.Under the household migration model, young, unmarried, and well-educated rural women are much more likely to migrate or to get off-farm jobs, and the occupations held by these rural female workers are more varied.In contrast, the majority of those who remain in the villages and continue in farming are a little older, married and either possess no education or only basic educational qualifi cations.
It seems that there are gender-related differences among the rural-urban migrant workers.Compared to the quite balanced gender structure of the total rural labor force, there is an obvious imbalance among the migrants.Although the duration of migration among females is the same as that of the male laborers, the proportion of returnees among married women is much higher than among their male counterparts.Income as measured by the amount of money sent to the families differs between female and male laborers.While the proportion of women in the lower income group is higher than men, the proportion of female workers in the higher income groups is lower.Hence, rural women as a whole are in a disadvantageous position in the labor market compared to other groups.
Existing institutional barriers place heavy constraints on rural women's access to the new opportunities to emerge out of poverty in terms of both economics and agency.The lack of social security for farmers proves to be a highly signifi cant institutional barrier.These are some of the detrimental consequences of the sorry performance of social policies during the transformation period.
The results of the study reveal some social, economic, and political implications on a more general level.First of all, the article provides a specifi c perspective to evaluate the social consequences of the marketization taking place in contemporary China.Rural women's increased labor force participation has been one of the most signifi cant aspects of social change in China during recent decades.Since China's rural economic reform began in the late 1970s as a result of a 'grass-roots' reaction to extremely poor living conditions in the countryside (Meng 2000, 21;Du1997a), rural women have played an increasingly important role in household-based agricultural production activities.In the 1990s, the labor force participation rate of rural female workers was consistently at 95%, whereas the urban female participation rate decreased from 76% to 64% in the same period (ACWF 2001).
This study proves that female rural-urban migrant workers constitute an important part of the surge of rural laborer mobility; those rural-urban migrant women have been and still play important and specifi c roles at the crossover between the emerg-ing capitalist economy and traditional rural society.This implies that market-oriented economic reform brought about diverse effects on Chinese women in terms of labor market status.Under these circumstances, examining the changes in rural female workers' role and status is a way to understand the 'meaning' of industrialization and marketization taking place in China.In this sense, the fi ndings of the study contribute to previous theories on the social consequences of the augmentation of capitalism in the west (e.g.Ritter 1986;Hoven 2002).
Secondly, the study views Chinese rural female workers as active beings rather than passive victims merely adapting to the social transformation.In world history, rural women's agency capacity in traditional society was in a way limited, and it has been argued that the huge transformation that is currently taking place puts female migrant workers in a position where they can be seen as victims of the transformation process.Though institutional barriers put Chinese female rural workers in a position of disadvantage, I argue that the performance of female rural-urban migrant workers suggests that they have become one of the most important groups in the making of China's economic reform and social transformation, and in doing so, are also positive participants of globalization in a wider development perspective.Female rural-urban workers are engaged in various non-agricultural activities in cities and special economic zones, and prove to be a vital force in boosting China's economic growth.
The industrialization and globalization process will possibly turn China into the 'workshop of the world'.The increasing demand for a cheaper rural work force in labor-intensive sectors will maintain the immense number of female rural-urban migrants, and consequently the importance of rural female workers on the labor market, and their contribution to China's and the world's economic development will continue to increase.This means that rural women have undoubtedly affected the economic and social transformation process, and vice versa: they are shaping China's social transformation and their social status is in turn shaped by this process.While their experience refl ects the salient characteristics of the modernization process underway in China, it also represents some aspects of the general trend of social change taking place in later-developing societies (cf. UNCTC 2004;UNESCAP 2003, UN 2003).
The employment of young, rural migrant women in export-oriented manufacturing industries has been studied extensively in some developing countries (UNRISD 2005;Gills 2002a).The fi ndings on Chinese female rural-urban migrant workers will contribute to the understanding of the changing role of rural women under different circumstances of globalization.

Table 1 .
General characteristics of the rural labor force of Anhui and Sichuan provinces in 1999 (N: 14,561)* Source: RCRE, MOA, Survey data set of 'A Study on Out-Migrants and Return Migrants': Result hh 99p+lab_b_new.xls,2000, Result hh 99p+lab99b_new_note.xls, Beijing.* 62 counties data As migration employment information is mainly communicated through relatives and neighboring villagers, migrants rely heavily on an informal network.The earlier migration emerged in a given locality, the more pioneer rural-urban migrants they have, and the better developed the informal out-migration network is.Hence, as a consequence, there are more migrants, and vice versa.The third reason is the situation regarding natural resources in any given county(cf.Bai 1997; Huang and CarreZhou 2001)g 2001;Zhou 2001)ant workers (to the total county's labor force) of 23%, which is the highest of all the 62 counties; thirteen counties have a percentage of 10%-15%; seventeen counties 5%-10%; seventeen counties 2%-5%; and fourteen counties less than 2% (Source: RCRE, MOA, Survey data set of 'A Study on Out-Migrants and Return Migrants', Result hh 99p+lab_b_new.xls,2000,Beijing).Various socio-economic factors explain these results.Regarding out-migration, this study confi rms the explanations common to many previous studies on China's ruralurban migration (see Bai 1997;Huang and Carrel 1998;Song 2001;Zhou 2001).The imbalance in the distribution of return can be explained by at least three factors.The fi rst one is the development level of non-agricultural sectors in the localities.In the counties in which the VTEs and other non-agricultural activities are developing well, more off-farm employment opportunities are created, therefore farmers can gain relatively good pay from those jobs instead of moving to distant urban areas (e.g.Du 1997b).The second reason is the history and informal network of migration in given localities(cf.Song 2001;Zhou 2001).

Table 2 .
Gender distribution by migrant experiences in the total rural labor force of Anhui and Sichuan provinces in 1999 (%), (N = 14,561)*

Table 3 .
Age structure of rural female workers (%), (N = 8,548)* Data show that the majority of female migrant workers and VTE employees are concentrated in the 20 to 35 years age group.The number of existing migrants decreases as the age-spans increase after the age of 25, and the number of VTEs decreases as the age-spans increase after the age of 35 years.That means that the rural women who are below the age of 35 have relatively more opportunities to get off-farm jobs.The age difference between female migrants and female rural enterprise workers is very small; the only distinction between them is in the age group of below 20 years.The migrant women under 20 make up more than one-third of the total female migrant workers while the share of this age-bracket is only about 10% among VTE employees.The data indicate that, as a rule, female migrant workers are younger than other rural women.

Table 5 .
The top occupations of rural women (%), (N = 8,947)* Source: RCRE, MOA, survey data set of 'Study on Out-Migrants and Return Migrants, RCRE, MOA, Survey data set of 'A Study on Out-Migrants and Return Migrants', labb99 by sex.xls, 2000, Beijing.Note: Return 1 represents the return female migrant's occupation, which they had for the longest period when they stayed in the urban area.Return 2 represents the return female workers' occupation after they return to the village.* 62 counties data except the Non-migrant group.

Table 7
Source: RCRE, MOA, Survey data set of 'A Study on Out-Migrants and Return Migrants', Labb99 by sex.xls, 2000, Beijing.

Table 8 .
Marital status and the mobility of rural workers (%)(N = 14,688)