BOOK REVIEWS

Agricultural Underemployment and Rural Migration in China: Facts and Figures

At the end of February 2001, the State Statistical Bureau (SSB) published the first results of the general population census that had been carried out on November 1st 2000. This announcement came as a considerable surprise: while the urban population (389 million people in 1999) had before been estimated at 31% of the total population, it was now (at 458 million) up to more than 36%. In earlier years, the increase of the urban population (through natural growth and the absorption of suburban peasants) had averaged only 8 to 9 million people per year. So, what had brought about this leap of 60 million in the number of urban citizens, as shown by the census?

Quite simply, rural migrants, living in towns at the time of the census, had been officially discovered and counted by the statistical authorities. Of course, the phenomenon of rural migration is not new; both the Chinese authorities and the urban population have long been familiar with these “peasant workers” (mingong) who flocked to the cities to sign up for the hardest and least well paid jobs. The phenomenon already existed by the early 1980s and acquired its full momentum during the 1990s, as reflected in the press and in sociological studies. Even so, the extent of the problem remains poorly understood, even as it radically changes the employment landscape both in the countryside and in towns and cities, leading to a re-examination of the system for civil registration of the population (hukou) in terms of rural and urban inhabitants.

Accordingly, this article will attempt to take stock of the statistical data concerning both agricultural underemployment and rural migration, with the aim of evaluating their significance at a time when China may be facing an acceleration in the rural exodus in the wake of its accession to the World Trade Organisation.

Agricultural underemployment,a purely relative notion

The official Chinese press often quotes the figure of 150 million “surplus peasants” threatening at any moment to swamp the urban employment market (1). But what does this figure mean?

Firstly, we should be aware that the agricultural labour force, after rising from 285 million people in 1978 to over 340 million by the early 1990s, is now stable or in fact in decline (at about 330 million in 2000), whereas the rural labour force is in constant growth (305 million in 1978, 480 million in 2000). This development, which is explained by the growing importance of rural (non agricultural) enterprises (Town and Village Enterprises, or TVEs) marks a fundamental turning point, one that is often ignored. It means that workers in agriculture amount to less than half China’s total workforce: this has been the case since the mid 1990s (see Table 1) (2). In other words, China, though still profoundly rural, is now a society in which the “farmers”, strictly speaking, are in the minority.

Table 1 - Rural and agricultural labour force, 1978-2000 (million of persons)

Sources: China Statistical Yearbook 2001 (rural and agricultural labour, pp. 363-364) and Labour Statistical Yearbook 2001 (total labour, p. 8)

All the same, even though agriculture employed no more than 46% of the Chinese labour force by the year 2000, it accounted for only 16% of GDP, which underlines the lower productivity of the agricultural sector. And, with nearly 330 million available workers, the oversupply among the peasants is only too visible, and evidenced by the small size of family holdings. If one takes into account only agricultural households (classified as such in the 1997 agricultural census, applying to 1996) (3), 193 million families (out of the 214 million rural households) were sharing some 130 million hectares of cultivated land, which gives an average of 0.67 hectares per farm (4). How can full employment be maintained among the available agricultural labour force, an average of almost two people per farm, on so little land?

Herein lies the problem of agricultural underemployment in China. But how is it to be measured?

To help us we have survey data, available since 1985 and gathered annually by the Pricing Bureau. It provides, in addition to cost and profit figures for the various agricultural products, details from the households included in the survey on the working hours associated with each type of crop or livestock. On this basis, and by applying this data to the land areas worked each year and, for instance, to the numbers of livestock reared, we have been able to estimate the number of working days required for agricultural production from 1985 to 2000 (Table 2). If we suppose that a full time agricultural worker works for 280 days a year (the figure adopted by the Pricing Bureau) (5), we can calculate the number of agricultural workers needed each year for the agricultural work available and, by subtracting this from the official total for the agricultural labour force, the number of workers seemingly without occupation.

This calculation brings to light an apparent surplus of agricultural labour of a considerable size, varying between 30% and 40% of the available workforce; and which, moreover, has been on the increase in recent years, rising from 85 million in 1995 to over 135 million in 2000. Our figures for the period under consideration are lower, however, than the estimates published in China, which rise from 120 million (6) to 152 million (7).

This surplus is in fact quite relative. Firstly, our estimates are based on various coefficients and are only approximations. Defining a full time job in agriculture is a very difficult task in itself – all the more so in China where the great majority of farms are mixed and where time is devoted each day to quite diverse activities (crop cultivation, domestic animals and so on) that defy accurate measurement. Nor do our calculations take account of the numerous related tasks (picking up firewood, feeding the oxen used for ploughing, and so on) that take up a large part of farmers’ working day (or that of their children) (8). Neither have we counted important activities such as fishing (even though the numbers of “fishermen” or those involved in full-time aquaculture are small). The real workload is therefore heavier than what can be calculated.

Furthermore, a “supernumerary” reserve of labour is needed to manage the seasonal peaks in demand. From this point of view, the notion of a standard eight-hour working day as used by the Pricing Bureau is illusory considering that, according to requirements, farmers can work only two or three hours a day in some seasons (such as winter, on days of frost or rain) but more than ten hours at other periods of the year. In these peak periods it often happens that the family’s labour – though surplus on the whole – is not enough; and farmers have to help each other out on an exchange basis (9). Indeed, agriculture is seasonal in the extreme. For example, in the rice-growing area of central China, the period from May to July is a peak season; and our field observations have shown that, even with a 30% average rate of supernumerary labour for the year as a whole, families lacked the muscle during a brief spell to maintain the alternating rhythm of harvesting and transplanting (10). From this point of view, a margin of 20% to 30% of supernumerary labour appears quite normal.

A closer study of this type of agricultural employment shows that farmers adapt the provision of their labour according to its opportunity cost. Generally speaking, the number of working days put into each crop has diminished considerably over the past fifteen years, falling by one third in the cases of rice and wheat (see Table 3): farmers mechanised certain operations (ploughing, for instance) whereas more plentiful job opportunities arose outside agriculture. In economic terms, in certain years the provision of labour diminished for crops for which prices had become unfavourable: this was the case with soya beans, for instance, where labour provision fell by 35% from 1996 to 2000, and with apples. When these crops, such as cotton, require much labour, a reduction both in the provision of labour and in the areas under cultivation can considerably affect the total demand for agricultural labour; and where prices improve the reverse is true.

These considerations help us to explain the sudden increase in the rate of surplus labour that took place over the past three years, rising from 30% to more than 40% of the available farm workforce (see Table 2). The fall in the demand for farm labour was certainly not due to any technological advance, but quite simply to the collapse of agricultural prices, resulting in a continuous fall in the rate of growth of peasant incomes (11). The reduced number of working days devoted to individual crops, particularly marked from 1997 onwards, reflects a perfectly rational reaction on the part of the farmers.

But what has become of the labour—the equivalent of 40 million full-time jobs—that was thus withdrawn from agriculture? To answer that question, we must first analyse the development of non-agricultural work in the countryside.

The ceiling on non-agricultural jobs in the countryside

The number of non-agricultural jobs has risen sharply since decollectivisation and the rural reforms, from 67 million in 1985 to more than 150 million in 2000 (see Table 4). Yet, this extraordinary advance seems to have run out of steam since the mid-1990s. Here, the official figures can be deceptive. While the census showed 151 million non-agricultural jobs in 2000, when the total is broken down into categories it shows a stagnation of workers in industry, and little advance between 1995 and 2000 in other sectors (up from 22 to 27 million in construction, from 10 to 12 million in transport, from 12 to 18 million in catering and trade). This slowdown in rural non-agricultural employment is reflected in the staffing of TVEs, which, counted on a different basis from that of the employment figures registered in the counties (xian) has stagnated at 128 million people. Only one category—“others”—shows a big increase, from 44 to 54 million people.

What does this last category represent? In fact, the farming Agricultural Yearbooks for the years 1985-1992 give totals between six and eight million for the number of administrative staff (in health, education and management, for instance) included in “Others”; the remaining unspecified jobs rose from 13 to 23 million (out of an “Others” total rising from 19 to 31 million (see Table 4). For the three years 1997-1999, these same Yearbooks no longer give details on administrative staff, but indicate that the “Others” category includes those working part-time away from their place of origin: 25 million in 1997, 27 million in 1998 and 20 million in 1999 (12). In other words, the official count of non-agricultural rural labour includes migrant workers, whose number in recent years can be estimated at from 20 to 30 million and probably much higher by 2000.

This fact is confirmed by the findings of the First Agricultural Census of 1997. Reflecting 1996, this census indicated that 76% of labour in the rural context was employed in agriculture (13). If, from the official total of rural workers reported in the Statistical Yearbooks (453 million people, see Table 4), we subtract the 23 million assumed to be migrants, the 323 million farm workers represented 75% of those really present in the countryside (14).

So, our examination of the figures for employment in the countryside, imprecise and sometimes contradictory though they may be, makes it clear that the reduction in working hours and in farming employment since decollectivisation has, in the first stage, been compensated for by a large increase in non-agricultural jobs in the countryside, particularly in TVEs where the workforce has almost doubled between 1985 and 1995. In recent years, this source of new jobs seems to have dried up, and surplus farm labourers have been forced into becoming migrants. Some of their number is officially accounted for at the district level among the 20 million to 30 million workers temporarily absent or under contract. However, this number is only the tip of the iceberg. We shall see, indeed, that seasonal migrants are in fact far more numerous, being made up precisely of these 30% to 40% of labourers designated as agricultural, but who cannot find full time work in agriculture, especially during the off-peak season.

The migrations of rural labour

We have seen that the rural workers adapted their provision of labour in agriculture according to its opportunity cost: on the one hand according to the expected farming incomes and on the other according to the opportunities for non-agricultural employment.

So, it is not the surplus of agricultural labour that forces underemployed workers to leave the sector. It is the gap between city and rural incomes that attracts migrants towards the towns and cities. Now, this difference in incomes, which had diminished slightly in the period immediately following decollectivisation, has been growing steadily since 1985.

In 1978, per capita rural incomes amounted to only 39% of urban ones. By 1985 they were up to 54%, but fell back again in 2000 to only 36% (see Table 5). The average salary of a city dweller is in effect over 170% higher than that of a peasant worker; that of a worker in a rural enterprise is 60% higher than that of a farm worker. We should note, moreover, that when it comes to rural incomes, more than half are now made up from non-farming family activities plus wages from rural enterprises or of migrant labourers.

So, city workers earn nearly three times what peasants do. Indeed the gap between town and country is greater still, in that this earnings ratio does not take into account the public services provided in cities (health and education, for instance). In this regard, it would seem that China’s countryside population, whose living conditions have deteriorated in recent years (under the growing pressure of collective deductions, taxes, health and education expenses, and so on) (15), are in a relatively less favourable situation that the average in countries covered by UN statistics (the ratio between town and country being mostly below 1.6, as against 2.8 in China) (16).

These being the conditions, it is not surprising that the towns have become increasingly attractive. Before 1978, any migration (and, even more so, any rural exodus) was prevented by the population registration system (hukou) that fixed a resident in his/her commune of origin, backed up by the state monopoly over the provision and sale of agricultural products (tonggou tongxiao) (17). It was only with decollectivisation, the abolition of People’s Communes and, in the towns, the liberalisation of trade in staple foods that people could begin to migrate from rural areas. In 1984, the state, in an official document, authorised rural workers to work in small townships on the condition that they “made their own arrangements for their grain rations” (zili kouliang, in other words, they had to seek out their own supplies on the market and not from the state-subsidised shops, which were reserved for the townspeople). During the same period, the central government encouraged the development of TVEs. Soon afterwards, the reforms began to affect urban sectors; their economy, particularly in the developed, coastal regions, was about to experience a spectacular rate of growth. Thus, conditions were ripe for the start of an exodus from farming into rural non-agricultural jobs and from the countryside into the towns (18).

The migration of rural labour in the strict sense began half way through the 1980s (19). During the 1990s, the phenomenon spread more widely and the “waves of peasant workers” (mingong chao) now attracted the attention of numerous researchers both foreign (Solinger, Rozelle, West) (20) and Chinese (Cai Fang, Lu Xueyi).

The studies that were then carried out enable us to trace the features of these peasants on the move: young, better educated, working mainly in the developed cities and regions, and supported in their migration by their own networks of family or village connections (see Box p. 55).

Some characteristic features of migrants  Age. In most cases, migrants are young: according to a survey by the All China Federation of Trade Unions (ACFTU) covering 4,400 peasant workers, 86% are under 35, of whom 59% are under 25. Their average age is 27.5, which is 10.7 years younger than the average for workers in general (1). Many young people in the countryside join the migrant “wave” straight from school, a third of them having no experience in farming. Education. Migrants are better educated than the average. According to the survey by the ACFTU, their educational range is: 1.4% illiterate, primary school 11.2%, junior middle school 65%, junior high school 21.6%, specialised and higher education 0.7% (the corresponding averages for rural workers as a whole are 16.9%, 47.2%, 30.7%, 4.9% et 0.3%)(2). Occupation. Generally speaking, migrants will accept the jobs that city dwellers turn down (work that is dirty, for instance, or dangerous or very hard), especially those in the sectors of construction, services and craft workshops. According to the survey by the Agricultural Bank of China (ABC), migrants are employed in the following sectors: industry 22%, construction 33%, transport 10%, trade and catering 31%, agriculture 4%(3). Behind these figures is concealed the fact that, for many migrants, working conditions are appalling. Days worked over the year. Owing to the seasonal nature of agriculture and to the demand of employers, migrants are generally unable to work full time, but only for some months during the year. According to the ABC survey, migrants worked on average for 207 days in the year (in 1993)(4). Another survey of migrants from Henan came up with a similar figure: 215 days a year in 1993 (5). In fact, a sizeable number of migrants go home to their villages to work on the land during the peak seasons: 26% of them according to the ACFTU (6), and as many as 58% according to a survey by Cai Fang, go home at least once in the year during the peak farming season (7). On this important point, surveys of migrants from Sichuan and Anhui arrive at similar conclusions: in 1994, the migrants from Sichuan stayed in their workplace for 8.12 months (6.48 months for migrants from Anhui), and in 1995 for 8.57 months (7.19 for Anhui); 29% of them stayed for 7-10 months, 54% for more than 10 months. Number of years as a migrant. Migrants work outside their birthplace; but they do not settle down permanently in their place of work. With time, nonetheless, the number of years spent as migrant workers does increase: 28% of migrants have done these outside jobs for over 10 years and 54% for over 5 years (8). The ACFTU survey shows that the average period of stay in outside work for migrants is 6.51 years. According to the China’s Youth Daily (9), a survey carried out in 1999 in the big cities covering 1,000 migrants showed that 19% had been in the city for over 2 years, 17% for over 3 years and 30% for over 4 years; 46% of the migrants questioned had no intention of going home to their villages to settle down. The method of migration. Fellow villagers and family networks are the main channel of migration. According to a survey of 300 families, 9% of migrants accompanied members of their family, 15% had been invited by relatives from the same village, 51% by fellow villagers, 9.2% by relatives from other villages and 5% had managed on their own. Only 6% had been helped by the village collective or by the township administration and 3% had been directly invited by the enterprises that employed them (10). Reverse migration. While more and more peasants are leaving their village to work outside, the opposite phenomenon is nevertheless encountered, a proportion of migrants returning to their villages. The Ministry of Agriculture carried out a survey in August and September 1999 among more than 2,000 rural workers in 6 provinces (Hebei, Shaanxi, Anhui, Hunan, Sichuan and Zhejiang). The survey identified, apart from 78% non-migrants and 13% migrants, more than 8% of the workers as former migrants now back in their villages (11). The return of these migrants is explained by the harshness of working conditions in the town: as they get older, migrants become less competitive within the labour market. Above all, the present hukou system prevents migrants from settling their families in the city, and obliges them sooner or later to return to their hometown in the country (where the money they have put by, together with the experience acquired in the cities, helps them to find, or create, decent non-agricultural jobs). Translated from the French original by Philip Liddell 1. ACFTU, op. cit., 1998, p. 4. 2. China Population Statistical Yearbook 1998, op. cit., pp. 34-35. 3. Han Xiaoyun, “Earnings, Consumption and Sector Characteristics” in Zhongguo nongcun jingji (Rural economy of China), No. 5, 1995, pp. 40-44. 4. Li Fan, “Extent of the Migrations and Other Questions” in Zhongguo nongcun jingji, No. 9, 1994, pp. 31-35. 5. Qi Tongxian, “Present Situation of the Wave of Peasant Workers in Henan”, in Zhongguo nongcun jingji, No. 2, 1995, pp. 39-41. 6. ACFTU, op. cit., 1998. 7. Cai Fang, “Study of the Mobility of Rural Labour in Shandong: Characteristics and Consequences”, in Zhongguo nongcun guancha (China Rural Survey), No. 2, 1996, pp. 51-56. 8. Du Ying, “Analysis of Group Characteristics and the general background to the Migrations of Rural Labour in China at the Present Stage”, in Zhongguo nongcun jingji, No. 6, 1997, pp. 4-11. 9. China’s Youth Daily, April 14th 2001. 10. Du, op. cit., 1997, p. 8. 11. Zhao Yaohui, Causes and Consequences of Return Migration: Recent Evidence From China, CCER Working Paper, No. E2001010, Peking University, 2001.

But while the features of these migrants have become increasingly familiar, the actual extent of their migrations is still difficult to evaluate. A lack of statistics on an intrinsically fluid phenomenon, has forced us to rely on surveys carried out in the countryside and on evidence about migrants in the villages or families that were investigated. Such surveys or studies lead us to rather disparate estimates.

For 1993, these estimates vary between 10.7% of the rural labour classified as migrant (survey by the Agricultural Bank of China, carried out in December 1993 and January 1994 and based on a sample of 12,673 families in 442 districts) (21) and 14.5% (survey by the Research Centre on Development of the State Council, RCD, carried out at the level of 28 villages in 28 districts of 18 provinces, between May and August 1994) (22). On the basis of the rural labour force of that year (443 million people), that would give a range of anything between 47 and 64 million migrants.

For 1994, we have the detailed results of a survey by the Ministry of Agriculture, carried out between November 1994 and April 1995 in 318 villages across 29 provinces, based on a sample of 7,677 peasant families. The migrant rate (as a percentage of the rural labour force) was then identified as 14.41% (23), which comes to 64 million people, a similar figure to that obtained for 1993 by the RCD.

For 1995, the census carried out by the State Statistical Bureau on 1% of the population, indicated a smaller total of migrants (because a certain number of migrants had escaped the census, only 54 million people were included, whose hukou was different from their place of residence at the time of the census) (24).

So, we may estimate that, in the mid-1990s, the number of migrants was probably somewhere between 50 million and 60 million people (25).

The survey by the Ministry of Agriculture also showed that six agricultural provinces (Sichuan, Anhui, Jiangxi, Hubei, Hunan and Henan) contributed by themselves nearly half the total number of migrants. Still more interestingly, the great majority of migrants, 71%, went to work in “real” towns (shi), the “townships” of the countryside (jianzhi zhen) being excluded, which amounts to 45 million people on the basis of a total of 64 million migrants, 22% of them in the cities (14 million people) and 49% in small and medium-sized towns (31 million). Yet, most remained in their province of origin: only 30% worked outside their province (19 million people) (26), while 28% (18 million) stayed in the same province (but outside their county of origin) and 42% (27 million) remained in the same county (but outside their township). The manufacturing, building and service sectors between them absorbed two-thirds of this migrant labour (28%, 23% and 15% of the migrants respectively) (27).

These survey results, at least those concerning migrations between provinces, were confirmed by the 1995 census, from which one could deduce that 18 million people came from other provinces (28). Five of the six provinces mentioned above (Sichuan with three million emigrants, Anhui 1.7 million, Henan with 1.6 million, Hunan with 1.5 million, Jiangxi with one million) contributed nearly half of these inter-provincial migrants.

In 1996, the number of migrants seems to have gone up again. The First Agricultural Census indicates that 24 million rural workers (whose families were surveyed in the villages and townships) worked, in fact, outside their province of origin. Besides, 21 million worked in their province but outside their county, and 27 million worked in their county (but not in the same township). Thus, the census gives a total of 72 million migrants (29).

Therefore, from1993 to 1996, the number of migrants would have gone up from 50 million to more than 70 million. Has this seasonal exodus gained momentum since, as would seem to be suggested by the recent drop in real agricultural employment, as calculated above? Unfortunately, a lack of precision in the most recent statistics prevents us from giving an answer to that question.

This uncertainty, often caused by basic differences in the way migrants are defined, is well illustrated by the results of the survey conducted jointly in 1999 by the Ministry of Employment and the National Bureau of Statistics, among the 66,000 families of the State Statistical Bureau sample (30).

According to this survey, the most recent available up to now, in 1998, migrants represented 20.56% of the working people in the sample, a proportion which, if applied to China as a whole, would be 95 million people. Yet, 48% of these migrants (46 million people) were simply commuters, working outside the village but without leaving their township. The number of real migrants, comparable to those identified in earlier surveys, was put at no more than 49 million people! The number of those working outside their province (19 million people) had not changed since 1994, while those working outside their county (but in the same province, 13 million people), or in the same county (but not the same township, 17 million people), were fewer than had been estimated in earlier years. In fact, this survey, distorted no doubt by the very nature of the sample established for the requirements of farming statistics, and including only those migrants working for more than six months outside their place of origin (31), reaches conclusions that contradict those of other surveys showing that a majority of migrants went into the towns (contrary to these other surveys, the 1999 survey by the SSB indicates that 57% of migrants including commuters, 54 million, went into “towns”, but 19 million of them actually went to the administrative centre of their township, jianzhi zhen, in a milieu that in fact was rural, leaving only 35 million who went into the cities, as against 45 million in 1994 according to the survey by the Ministry of Agriculture, and 42 million according to the 1995 census) (32).

This same survey, covering the same number of inter-provincial migrants in 1998 as those surveyed in 1995, tends to accentuate the spatial polarisation of the migrations, with 83% of the migrants beyond their province of origin making their way towards the eastern provinces or the cities of Shanghai and Peking (49% towards Guangdong, 7% towards Zhejiang, 7% towards Peking, 6% towards Shanghai, 5% towards Fujian, 5% towards Jiangsu, and so on), as against 55% at the time of the 1995 census (only 26% towards Guangdong, 8% towards Jiangsu, 8% towards Peking, 6% towards Shanghai, and so on). Conversely, six of the provinces exporting labour provided two-thirds of the migrants: Sichuan and Chongqing (17%), Jiangxi (13%), Hunan (12%), Anhui (11%), Hubei (7%) and Henan (7%), whereas in the census they provided only 53% (Sichuan 17%, Anhui 10%, Henan 9%, Hunan 9%, Jiangxi 5%, Hubei 3%).

Looking beyond all these uncertainties, and basing ourselves on the increase observed between 1993 and 1996, it is not unreasonable to think that the total number of migrants, including those living for less than six months outside their place of origin, was likely to reach nearly 100 million people by the year 2000, probably 60 million of them working in the cities and 30 million leaving their provinces.

What do these figures mean? If one considers that nearly 20 million migrants are already included in the non-agricultural labour of the countryside, there remain 80 million people, officially counted under the heading of farming labour, and working in fact, for at least a part of the year, outside their villages. This figure must be set alongside the 100 million to 140 million “surplus” farm workers in relation to the work really absorbed by agriculture. In other words, in the low season, a majority of the reserve of surplus labour actually leaves to seek employment in the towns, making up the bulk of those legions of seasonal migrants identified by the census.

Integrating with the urban milieu: a difficult process for the migrants

So, while easing underemployment in the countryside, the migrants send back a significant proportion of their income for their families in the villages. According to a survey by Li Qiang, more than 70% of migrants send money to their families (33). Thus, in 1998, a migrant earning nearly 5,900 yuan a year was sending an average of nearly 2,600 yuan to his family, and consuming for his own local needs only 2,300 yuan (much less than a city dweller, about 4,300 yuan per person per year, but more than a person living in the countryside, 1,600 yuan) (34).

So, for the peasants, the chance to migrate, even temporarily, is welcome. But are the migrants welcome in the cities?

At first, public opinion, as passed on by the media, focused above all on the inconveniences, the “trouble” caused to public order by these waves of mingong: for instance, the congestion on the railways during the New Year migrations, and the growth in crime; these “troublesome migrants” were accused of disturbing the peaceful life of the cities.

Even so, the Chinese cities were experiencing rapid development, particularly in the fields of construction and services; and these migrants satisfied the growing need for labour (35). Since the mid-1990s, the tensions between migrants and city dwellers have been taking a new form. The restructuring of old industries and the reforms to the state enterprises have led to a rise in unemployment (the unemployment rate, officially 3.5%, with fewer than 10 million unemployed, is in fact put at nearly 9%, with 16 million people out of work). The number of licensed employees (xiagang or, since 1998, bu zai gang), of whom a certain number find new employment, has apparently increased from some three million in 1993 to nearly 22 million in 1999 (36). Are the potential conflicts between migrants and city dwellers extending now to the field of employment with this rise in urban unemployment?

In fact, as noted above, the migrants do not take the same jobs as the townspeople. The regulatory restrictions they face when seeking work in collective enterprises – and even more so in state enterprises – mean that 90% are employed in the private sector, or self-employed (according to a survey of 619 peasant workers in Peking, 18% were employed in private enterprises and 53% self-employed (37); according to a study in Shanghai, less than 8% of the posts taken up by migrants had any impact on the employment prospects of local people (38)).

Far from being a burden on the public finances, as were the licensed employees in 1999, each of whom was paid an average of more than 1,500 yuan in subsidies (taking a total of 3.2 billion yuan out of the unemployment funds), migrant labour costs hardly anything to the collectivity. And this is where the Chinese authorities face the main social problem: how are they to integrate these migrants with urban society, without making them outcasts in their own country?

Unlike city workers, migrants are very badly insured (less than a quarter of those migrants in work benefit from medical insurance and less than a third are covered against accidents at work (39); according to a survey by the Ministry of Agriculture, these proportions are lower, at 12% and 25% (40)). When they do sign contracts with their employers (in less than 10% of cases in private enterprises), these contracts are usually invalid or even illegal (with clauses such as “Death or injuries at work are not covered”) (41). Tragic fires in working premises, resulting in the death of migrant employees, are regularly reported in the press. And we need hardly speak of the working conditions to which migrants are subjected.

Still more important, from the point of view of the future of the rural exodus, is the fact that education expenses are prohibitive for migrants if they do wish to bring their family with them: this denies them any chance of settling down in the long term. With no urban hukou, a child must pay “compensation expenses” (jie du fei, or zhangzu fei) rising to 2,000 or 3,000 yuan a year, in order to attend a public primary school in the city. Some migrants have even tried to establish their own “people’s” schools, independent of the urban authorities and much less expensive (320 yuan in fees per half-year); but such schools are still illegal (their number in Peking is put at more than a hundred, while the number of migrant school-age children is put at more than 100,000) (42).

These considerable differences in status between city dwellers and rural migrants are the effect of a deliberate policy of segregation (43). This policy, in the big cities, takes the form of regulations aimed at limiting the migrants’ entry to the labour market (definitions for three categories of jobs in Shanghai in 1995, then in Peking in 1997: authorised to migrants, restricted to migrants, barred to migrants) (44). Even in average-sized towns, nearly forty sectors or professions are barred to migrants (45). Consequently, the number of migrants in the very large cities is declining (in Peking, after reaching a high point of 3.3 million, the number of migrants has fallen back to fewer than two million) (46).

Rather than simply banning certain types of jobs, most localities use systems of “certificates” or “cards” to try and control the flood of migrants: working licences, family planning cards, temporary residents’ permits... In reality, these systems are more like a form of charge imposed on migrants by the local authorities than a real tool for limiting immigration: according to some sources, at the average rate of 600 yuan in various expenses per person per year, the migrant-peasants contribute nearly 60 billion yuan to the running costs of various public institutions in China.

What happened to the reform of the hukou?

The hukou system with its family record book was promulgated in 1958. It differentiated between the family registrations of countryside inhabitants (nongcun hukou) and those of “cities and townships” (chengzhen hukou). It resulted in setting the “agricultural population” (nongye renkou) against the “non-agricultural” (fei nongye renkou). The aim had been to keep people in the area of their original People’s Commune, thus preventing any migration and favouring city dwellers who enjoyed numerous advantages, among them low-priced food rations (and thus enabling “primitive accumulation” to take place at the expense of agriculture).

Since decollectivisation, the reforms have removed this role as a brake to all mobility for rural dwellers. As the earlier examples show, the hukou has become principally a discriminatory weapon against migrants from rural areas, denying them access to the privileges that city dwellers enjoy (47). Is this system—this discrimination—viable over the long term? This question has a special relevance now that China’s entry to the WTO is supposed to accelerate the mobility of the labour force, and the rural exodus, intensifying the problems of integration faced by migrants in urban areas.

In fact, the need to reform the hukou had been apparent ever since, in 1984, rural workers were authorised to settle in the townships, on condition that they provided their own food (cf. above, the system of zili kouliang). This reform brought a certain flexibility into the employment field by encouraging mobility, geographical and vocational, among people from the countryside. But, since the city hukou conserved its numerous advantages, speculation and corruption spread: a city hukou could be bought on the black market (a hukou in Peking could be sold for as much as 100,000 yuan) (48).

Over the following years, new systems for registering the population were tried, mainly aimed at limiting the numbers of migrants in the big cities. There were, for instance, systems allowing temporary residence – in Peking in 1985, in Tianjin, Shanghai and Wuhan in 1988. During the 1990s, many cities introduced a special category of hukou, known as the “blue-sealed” hukou, or the quasi-urban hukou (zhun chengzhen hukou), which gave access to some advantages normally reserved for city-dwellers only (49). In fact, the acquisition of this hukou was subject to very strict conditions as to levels of qualification, and the possession of spacious accommodation, all of which allowed access only to a minority of workers whose wealth or skills were sought-after. Those equipped with this precious key remained numerically marginal (in Liaoning, their number rose from 88,000 at most in 1993 to 320,000 in 2000; but in Shanghai no more than a thousand people were qualified in 1997) (50).

In 2001, more far-reaching reforms were announced, arousing great interest in the media. In the province of Guangdong, the distinction between the agricultural hukou and the non-agricultural hukou was to disappear, with the creation of a single “residence hukou” (juming hukou) (51). The government of the province of Fujian also decided to put a similar system in place in 2002 (52), while abolishing the different types of provisional hukou (for instance, zili kouliang hukou, or the blue-sealed hukou). In other cities, the various quotas applied to residence permits for migrants would be replaced by systems offering “conditions for admission” (zhunru tiaojian).

These spectacular announcements, however, were not to be followed by any significant changes. The conditions for obtaining the new hukou in Shenzhen and in Guangdong province have not yet been announced. Generally speaking, the admission conditions in the big cities remain, as in the past, prohibitive: in Peking, you have to be the head of (or a shareholder in) a private enterprise that has taken on one hundred Peking citizens per year for three years, or to have paid 800,000 yuan in taxes per year in three consecutive years (53). The only exception seems to be the city of Shijiazhuang, where the conditions imposed were far less onerous (only two years’ residence, no special skills required, no restrictions as to type of employment) (54). In reality, in this capital city of Hebei, which is racked with unemployment, job opportunities for migrants are very poor, and the reforms have had scarcely any effect (55).

More significant without a doubt is the total opening up of small towns and townships, which has been in place since October 1st 2001. On condition they have a fixed place of residence and stable employment rural workers can now change their hukou without charge (while retaining the possibility of keeping their rented land in their village of origin, as for instance at Fenghua, in Zhejiang) (56). This reform, however, is more the recognition of a fait accompli: for a long time now, peasants have been settled in the townships, and have started their own businesses and bought their own homes. The reform carries less weight in that the residents of these townships do not enjoy the same advantages as residents of big cities: for example, qualified school-leavers are no longer guaranteed appointments in the administration or in state enterprises, and they have no priority access to the universities, social security is practically non-existent and medical treatment is as expensive as in the countryside.

The most important reform will perhaps be the decision by the Planning Commission to abolish the “seven expenses” to which migrants have hitherto been subjected, starting before the end of February 2002 (57). For migrants, such a measure, if really applied, will have greatly more significance than the “reforms to the hukou” that are currently under discussion.

 

The uncertainty attached both to the future status of the hukou and even to the extent of the rural migrations relates to the challenges facing the Chinese authorities on the eve of a radical transformation: the urbanisation of a country that is still mainly rural, along with the decline – even, in the long term, the disappearance – of the world’s greatest concentration of peasants.

The divide between “rural” and “urban” people, maintained by their respective status, and aggravated by the widening gap between their earnings, is seemingly not about to disappear. China is a developing country and already has great difficulty in financing the new social security cover for its city dwellers, which must be substituted for the protection offered by the old “work units” (danwei). It is inconceivable that peasants migrating to the cities should be able to benefit from this cover, at least in the short term. So, the former segregation between peasants and city dwellers will probably last through the discriminatory measures that turn “peasant workers” into second-class citizens.

To more fully appreciate this phenomenon one would need more precise knowledge about the flow rates of migrants, the situation in one type of city being quite different from that in another. If, as the figures seem to indicate, the largest cities, such as Peking or Shanghai, are succeeding in containing the stream of these “immigrants” from the interior, then the phenomenon of the shanty towns that proliferate around cities, as experienced in South Asia or Latin America, will be avoided.

What will remain is the problem existing in small and medium-sized towns. If the development of the townships, contrary to the authorities’ expectations, are compromised by the growth crisis of the TVEs, the migrants are going to turn towards the cities. How they integrate will depend partly on what measures the local authorities may take and equally on the practical details of the migrations: do the peasant workers make their way first of all to the administrative centres of their own districts, or to the towns in other districts, or in other provinces? What connections do they still keep with their villages of origin, and how do the structures of peasant families adapt to the stress caused by these migrations? These are some of the questions which will determine the characteristics of the coming rural exodus. It also means that China will in time no longer resemble the country it was for so many centuries. Answering such questions is a task requiring far more information, and far more research, than are available to us at present.

Translated from the French original by Philip Liddell