Population migration pattern in China: present and future

Population migration pattern in China: present and future Lu Qi 1), Leif Söderlund 2), Wang Guoxia 1) and Duan Juan 1) 1) Institute of Geographical Sciences and Natural Resources Research, CAS, Beijing China, 100101, Luq@igsnrr.ac.cn, Wanggx@igsnrr.ac.cn 2) MTT Agrifood Research Finland, FIN-31600 Jokioinen, Finland, leif.soderlund@mtt.fi Abstract This paper explores the population pattern in China since the 1990s, as it plays a significant role in the process of urbanization and modernization. At present, China is experiencing nationwide population redistribution. There have been big increases in both intra-province and inter-province migration relative to the situation in the 1980s, especially in the case of cross-region migration. The data show that Eastern China is the main destination for migration, particularly the Pearl River Delta, Yangtze River Delta and Bosea Economic Circle. It is the increase in surplus rural labour, the heightened disparity between rural and urban incomes and between inland and coastal incomes, and institutional changes that have contributed to the current population pattern of migration. In the last twenty years there has been only a small change in the migration pattern, with the Markov Chain, that is the Pearl River Delta taking a smaller share of the migrants and more of them being concentrated in the Yangtze River Delta. Key words: migration, migration patter, China 247 MTT Agrifood Research Finland

Introduction Population migration can be defined as the movement of people from one place to another, largely in pursuit of employment. In the developing countries the majority of this takes place from rural to urban areas, as is inevitable in a market economy where the opportunities for work are greater in the cities. Since 1978, China has experienced a profound economic transition, from a planned economy to a market economy, which has had significant impacts on the various aspects of life. Among these, rural-urban migration is the most evident change. After the stage of leaving the land but not the home town in the 1980s, which means that most of the rural population moved to local township and village enterprises, more and more rural inhabitants have undertaken cross-region migration since 1990s, and it can be predicted that this will be the irreversible secular trend in the coming decades. This paper aims to explore the population migration pattern since the 1990s and its causes, and to predict the pattern in the future. Population migration pattern Data resources Nationwide investigations into migration in China were carried only in connection with the fourth and fifth censuses, applying to the periods 1985-1990 and 1995-2000, respectively, which contributed greatly to research into migration. By analysing these census results, we can obtain a picture of provincial migration and set up a matrix for province-to-province migration. Another important source of data consists of the China Statistical Yearbooks. Intra-provincial and inter-provincial migration China is experiencing great population redistribution in the context of rapid urbanization, with accelerations in both intra-provincial and inter-provincial migration. It is a worldwide rule of migration that more migrants will gravitate to a place that is more developed. We know that the eastern region of China is the most developed area, so it is obvious that this will be the most active region of in-migration, particularly Guangdong and the provinces in the middle of China but not attached to the coastal provinces or Sichuan in the west, as seen in Figure 1. Mainly chose intra-provincial migration, which is understandable considering the cost of migration, e.g. transportation, in all the provinces. The structure of migration varies between the regions, however, more people choosing intraprovincial migration in Eastern China, where this accounted for 85% of total migration, whereas this form amounted to about 37% of all migration in Middle and Western China. The standard error attached to the structure of 248

migration shows that there was no big difference in Middle China, but that the situation is different in the east and west (Table 1). This means that the distribution between intra-provincial and inter-provincial migration in Middle China was even in every province, i.e. they all had the same migration structure. The high standard errors in the Eastern and Western China indicate that there existed some areas that were either the main destinations for inmigration or the main origins of out-migration. Fig. 1. Gross migration in China in 1995-2000 Table 1. Standard error of migration structure East Middle West Standard error 16.87 3.60 14.62 The change in inter-provincial migration from the 1980s to the 1990s Comparison of the data given by the fourth and fifth censuses shows that the proportion of cross-provincial migrants in Middle China increased, from 31.8% to 37.4%, i.e. by about 6%, while the proportion in the western provinces declined by about 3%. As for Eastern China, the proportion of intraprovincial migration increased greatly, by about 15%. As for the distribution of inter-provincial migrants, these accounted for 56% of the total in the relatively well-developed eastern area in 1985-1990, 26.2% in developed middle 249

region and 17.8% in the less developed western region, while in 1995-2000 the figures were 79.18% in the eastern region, 9.03% in the middle region and 11.8% in the western region. Obviously, the trend for a concentration of migrants in Eastern China has strengthened further. The table 2 shows interprovincial migration in 31 provinces or municipalities. Here the average was used to reflect the centralization tendency and the standard deviation to reflect the degree of dispersal. The average indicated that the possibility for migration is gradually expanding among the population of China and the flow is increasing constantly. Thus the average in-migration rate in 1995-2000 was nearly 3 times that for 1985-1990. As for the standard deviation, the degree of dispersal in the spatial distribution of in-migration and outmigration increased, which means that the destinations and origins of the migration were becoming more and more concentrated. What is more, the degree of concentration in in-migration was obviously greater than that in out-migration. The trend in migration is also shown in Figure 2. Table 2. Parameters of inter-provincial migration 1 Average Standard deviation 1985-1990 in-migration rate 1.4120 1.3613 1995-2000 in-migration rate 4.0115 5.1263 1985-1990 out-migration rate 1.1510 0.4865 1995-2000 out-migration rate 2.8122 2.3306 1985-1990 net migration rate 0.2610 1.4222 1995-2000 net migration rate 1.1993 6.5597 Data source: China, Fourth Census, China, Fifth Census 3 10 Out-migration rate( %( 2. 5 2 1. 5 1 0. 5 Out-migration rate( %( 8 6 4 2 0 0 2 4 6 8 1985-1990 In-migration rate (%( 0 0 3 6 9 12 15 18 21 1995-2000 In-migration rate ( %( Rural Fig. 2. mobility The migration of labour rate distribution of 31 provinces in 1980s and 1990s Since rural-urban migration is the main type in the developing countries, it 1 It included 30 provinces during 1985-1990, because Chongqing Municipality was established in 1997. 250

Rural mobility of labour Since rural-urban migration is the main type in the developing countries, it was the rural population that formed the main proportion of the migrants. It is necessary to discuss the mobility of rural labour. Considering the cost of migration, most rural labourers choose to move within their own province, so that these make up 64% of the total rural out-migrants. Of the three regions, the first in terms of the rural population choosing the destination in the same province is the eastern region, then the middle region and finally the western region, thus coinciding with the level of economic development in China. As for inter-provincial migration, the rural population of Middle China took first place and then that of Western China (Table3). Undoubtedly the higher wages paid in the east of China form the main pull factor. Since the reform and opening-up of the economy, the economic disparity between three major parts of the country has been gradually widening all the time. Although a strategy of "development of the western regions" has been practised, the accumulation effect of economic development has made the disparity between the eastern area and the central and western regions still greater, causing more active migration at the same time. Table 3. Intra-provincial and inter-provincial migration of rural population in three regions Interprovince Percentage of total rural interprovincial migrants Intra-province Percentage of total rural innerprovincial migrants Total rural migrants Percentage of total migrants in the country East 5693790 21.98% 24489000 51.81% 30182790 41.25% Central 12771770 49.32% 12169760 25.75% 24941530 34.09% West 7434310 28.70% 10603750 22.44% 18038060 24.65% Total 25899870 35.40% 47262510 64.60% 73162380 100% The mobility of rural labour varies among the provinces, although the rural population made up a large percentage of migration over the whole country. On the basis of the fifth census, we calculated and plotted the scatter of the percentages of intra-provincial and inter-provincial migration for the 31 provinces (Fig. 3). Taking the averages for the two forms of migration as the reference lines, we were able to place the 31 provinces in four sectors. From the figure 3 we can see that: (1) There were 17 provinces in the top right-hand part of the diagram, with above-average percentages of both forms of migration. These provinces were located not only in western and Middle China but also in the eastern region. It is reasonable for the rural labour to migrate because of poorer economic 251

development, lower per capita income, smaller area of cultivated land per capita and relatively convenient transport in the west and central regions. Although the income of rural labourers in the eastern region was higher, they had extensive further possibilities to move in search of a higher income. Another important reason was their more open ideology. (2) There were five provinces in the top left-hand sector, with low interprovincial migration and higher intra-provincial migration. The rural population of Tibet, Qinghai and Ningxia, for instance, will have found migration far more costly due to the distance from the centre. whereas for people in Guangdong and Hainan, who had a higher development level, it was not necessary for most rural labourers to move out of their own province, in addition to which there was a natural obstacle for those in Hainan. As for Shanxi, there were many township and village enterprises available on account of the abundant natural resources, and the rural labourers in any case had a relatively conservative ideology. (3) The nine provinces lying in the bottom right-hand sector can be classified into three types in terms of their economic and natural conditions. Beijing, Tianjian and Shanghai are the oldest municipal areas in China, and even their rural labourers had higher incomes, higher, in fact, than the urban incomes achieved in the western provinces. For Liaoning, Heilongj and Jilin, located in the northeast of China, the natural conditions placed some obstacles on inter-provincial migration. Furthermore, these areas form the heavy industrial base of China and the state-owned enterprises play a considerable role, operating less labour-intensive industries. In addition, as economic development is relatively low, the tertiary, service sector is still relatively poorly developed reducing the area s capacity for absorbing rural labour. All these factors reduced intra-provincial migration in the three provinces, too. The third type includes Xinjiang and Inner Mongolia, which are minority provinces, where the cultural difference is the key factor for migration, in combination with distance, of course. As discussed above, it was clear that the migration of rural labour, especially the choice between intra-provincial and inter-provincial migration, depends on many factors. Not only general factors such as income and distance, but also cultural differences, the industrial structure and especially the existence of enterprises were key factors. In any case, migrants in China have always been rational in choosing their options in terms of costs and benefits, even in the institutional transformation stage. 252

70 % of Inter-provincial migration 60 50 40 30 20 10 0 Tibet Xinjian Beijing Tianjin Shangh 20 Ningxia Hainan Shanxi Guangd Qingha Innermog Heilongj Jilin Liaoning 40 60 80 100 Fujian, Hebei, Shandong, Zhejiang, Jiangsu, Hubei, Henan, Jiangxi, Anhui, Hunan, Yunan, Gansu, Shanxi, Guizhou, Chongqing, Sichuan, Guangxi % of Intra-provincial migration Fig. 3. Distribution of the 31 provinces by rural labour migration structure General Causes of Migration Increase in surplus rural labour Both the heavy industry strategy and unreasonable management system that prevailed in agriculture in the pre-reform days induced many rural labourers to stay in the rural areas. With the reform of 1978, which was introduced together with the household responsibility system in rural areas, many surplus rural labourers concealed by the original communistic system (Daguofan) became more and more evident. At the same time, the imbalance between land and population in the rural areas became more acute. On the one hand, the rural population increased greatly, while on the other hand the area of cultivated land decreased continuously. According to the State Statistical Bureau in 2002, the total area of cultivated land had declined to 13.04 billion ha and the area per capita was no more than 0.11 ha, less than half the world average at the end of 2001. Both of the reasons discussed above pushed the rural labourers into migrating to the urban areas. Furthermore, progress in science and technology improved agricultural productivity and make it possible for the farmers to cultivate land with machines on a large scales, which released more labourers as a result. It is estimated that the current rural population of China is 800 million and that there are 500 million rural labourers. If we deduct the 150 million labourers required in agriculture and the 70 million people employed in township and village enterprises, there leaves 280 million rural labourers at least who would need to migrate to urban areas. At present, there are some 120 million rural labourers employed in the cities, 253

which means that there will be a further 160 million moving there in the next 20 years or so. Hence it can be demonstrated that the surplus rural labourers concealed by the old systems in the past or generated by the natural growth in population will be the main sources of future migration. They will be a great tidal wave in the coming rapid urbanization period as the reform of institutions and systems proceeds further. Urban-rural and inland-coastal income disparities The increasing gap in the regional economies is the main factor promoting cross-region migration in China. Because of geographical location, climatic conditions and the influence of western colonists on the history of China and so on, a major gap has existed between Western China and Eastern China since 1840. When the People s Republic of China was established in 1949, the regional gap shrank for some years under the effects of state planning. By the 1980s, however, with the introduction of a special economic region, the coastal provinces in the east enjoyed substantial development. Especially after Mr. Deng had delivered his address following a visit to the south of China in 1992, the gap between Eastern and Western China became wider again. Although the central government tried to adopt a strategy of development in Western China to reduce the gap, the cumulative effect of economic development did not have much impact. It is clear from Figure 4 that the gap is becoming increasingly wide. 12000 10000 8000 6000 4000 2000 Eastern China Middle China Western China National Average 0 1970 1975 1980 1985 1990 1995 2000 Fig. 4. Per capita GDP in Eastern China, Middle China and Western China Rural-urban migration is currently the main form, accounting for 73% of total migrants according to the census of 2000. The main reason for the increasing migration is the growing income gap between the urban and rural areas. It can be seen from Figure 5 that though incomes kept on increasing, the income gap was also increasing. The income ratio between urban and rural residents increased from 1.86 in 1985 to 3.11 in 2002, and the real gap may 254

taken to be still larger than that shown in the statistics if such payments to the urban population as housing subsidies, welfare benefits, capital bonuses and medicines paid for out of public funds are considered and set against the expenditure required of the rural population in the next round of agricultural production. RMB Rur al Ur ban 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 1985 1987 1989 1991 1993 1995 1997 1999 2001 Fig. 5. Trends in per capita annual income in urban and rural households, 1985-2002 Institutional changes affecting rural-urban migration Migration and the distribution of population in China have always been closely related to the government s policies. To some extent institutional changes, especially when they affect the systems for controlling migration, can reduce the costs arising from this migration. With the introduction of the household register system, rural-urban migration came under rigid state government control, causing the employment structure to lag behind the economic structure. After the reform, control over migration was relaxed and many new policies were introduced. First the state government permitted farmers to enter the cities, provided they were able to feed themselves and had a fixed place of residence or employment in the town or village enterprises. In 1992 most of cities cancelled the grain coupon system, which had provided insurance for the rural labourers entering the cities. These changes in the system had some influence on migration, as can be proved from the statistics. From 1985 to 2000, migrants of all kinds increased in number rapidly, from 34 million to 110 million. Meanwhile, non-public ownership enterprises, including foreign capital enterprises, joint venture enterprises, privately-operated enterprises and other non-governmental enterprises, accelerated the migration. The change in the employment system gives younger people with knowledge and ambitions more choices. The central government also tried to adopt positive financial policies to create more opportunities for 255

the surplus rural labour. According to the nearest statistics, the state added RMB 300 billion to its national debt to create labour-intensive projects. Some hydrological, transportation, energy and environmental protection projects have been established, and others are also being carrying out which supply lots of opportunities for rural labourers, contributing to their migration, improving their incomes and relieving the employment pressure caused by the slowing down of the township and village enterprises. Future distribution of migrants The scale forecast for inter-provincial migration We adopt here the method of time-trend extrapolation to estimate total interprovincial migration in the last twenty years. Since population migration is affected by many factors, it can be observed that the numbers of migrants fluctuate greatly if we depict the migration data for every year on the graph, and forecasting becomes difficult. Hence we should introduce an index, the cumulative value of migration. This is the sum of migration in the past years, which makes the trend in migration uniform and continuous. We can then obtain the final result in reverse by deduction. The fitting curves and their equations are shown in Figure 6. By calculation, we obtain the following result for inter-provincial migration over the period 1992-2020 (Table 4). 10000 persons 12000 10000 y = 3.949(t-1991) 2 + 214.6(t-1991) + 23.261 R 2 = 0.9997 8000 6000 4000 2000 y = 6.4782(t-1991) 2 + 147.34(t-1991) + 90.736 R 2 = 0.9995 0 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019 Year In-migration Trendline of out-migration Out-migration Trendline of in-migration Fig. 6. The growth trend line of in-migration and out-migration 256

Table 4. Cumulative inter-provincial migration over the period 2001-2020. unit: 10,000 persons Year 2001 2005 2010 2015 2020 Total Migrants 289 1527 3251 5174 7293 It may be seen that total inter-provincial migration over the period of twenty years will amount to about 73 million persons. By reference to the proportion of inter-provincial migrants in the fifth census, total inter-provincial migration in the following twenty years should be 0.22 billion, which means that there will be about 11 million inter-provincial migrants every year. If the state population policies were to become more flexible, the economy more active and the obstacles to inter-provincial migration be abolished step by step, this form of migration could prove to be more pronounced than in the prediction. Projection of the inter-provincial migration pattern for China Many empirical studies have proved that a Markov Chain can be used to predict the distribution of population. A Markov process can be described by a TPM (transition probability matrix), provided that the initial vector 0 0 0 0 W = { p1, p2 L, p n } is known, where the transition probability matrix is P s and the state W, after the phase s, is as follows: s s 1 0 s 1 W = W P = W P. Employing the statistics for the fifth census, held in 2000, we can establish the inter-provincial in-migration probability matrix PMI (31*31 matrix) and out-migration probability matrix PMO (31*31 matrix)and the initial probability vectors WPMI (1*31 matrix)and WPMO (1*31 matrix), respectively. Using the Matlab software, we can easily obtain results for the probabilities of migration by province in 2005, 2010 and 2020. As the transition probability matrix is convergent, the migration of population will reach a stationary stage after a certain period. Hence, we can calculate the probability of a stage at which the inter-provincial migration is stationary. The Markov result shows that the probability of inter-provincial outmigration will be stable after 33 years, and that of in-migration after 17 years. In addition, the Markov Chain indicates that the spatial pattern of interprovincial migration will gradually become stationary from 2010 onwards. Changes in the probability of net migration by province are indicated in figure 7, which shows that the net migration provinces will include Beijing, Tianjing, Hebei, Shanghai, Jiangsu, Zhejiang, Fujian, Guangdong, Hainan 257

and Xinjiang. Compared with migration in the late 1990s, the nation-wide distribution of net migration will imply a slight adjustment in the long run. The proportion of in-migration in Guangdong will decrease, while Beijing, Shanghai, Tianjing and Jiangsu will attract more migrants. Probability of Inter-provincial migration(%) 15 10 5 0-5 -10 Beijing Tianjin Hebei Shanxi InnerMongolia Liaoning Jilin Heilongjiang Shanghai Jiangsu Zhejiang Anhui Fujian Jiangxi Shandong Henan Hubei Hunan Guangdong Guangxi Hainan Chongqing Sichuan Guizhou Yunnan Tibet Shaanxi Gansu Qinghai Ningxia Xinjiang Fig. 7. Prediction of inter-provincial migration probabilities in China by province in 2010 Conclusions China is experiencing rapid urbanization and modernization. Migration in Eastern China has been more active on account of the developed economy and aspiring employment institutions, especially in the Yangtze River Delta, Pearl River Delta and Bosea Economic Circle, which have not only absorbed large numbers of migrants from other provinces but have also had intraprovincial migrants. The current pattern of migration shows a universal abundant migration of population from the western and central regions to the eastern regions, that is, from the less developed regions to more developed regions. Compared with the migration experienced in the 1980s, this interprovincial migration tended to be more concentrated in the 1990s. The increases the rural-urban and inland-coastal income disparities are the driving forces for migration in China and will continue to act in this way. The institutional reform speeded up rural-urban migration in its time. It is estimated that there will be 0.22 billion inter-provincial migrants in the twenty years from 2001 to 2020, and that the present pattern will be retained except for a slight adjustment. Various patterns of migration may be detected in China, including Random Migration (from the countryside to the big cities), Chain Migration (through introductions from relatives and friends, moving to all levels of cities), Step Migration (from the countryside to the towns and on to the small and middle- 258

sized cities and then to the big cities), Organized Migration (under government control) and Forced Migration (reluctant migration by the socially destitute) and so on. Chain Migration is the most widespread pattern in China, being associated with a strong sense of locality, which is the migrants arriving at a destination are mainly from certain provinces nearby, or else outmigrants of a certain origin may mainly move to certain destinations. Some recent social investigations have found that the migrants choose the destination according to what their relatives or friends choose to do, even if they have complete information on the possibilities for migration. Hence the predilection for Chain Migration may lead to little change in the migration pattern in the coming years as a result. In order to create sufficient employment for migrants in the main destinations, it will be necessary to actively develop tertiary-sector industries in the urban areas and labour-intensive manufacturing industries and urban agriculture. Meanwhile, it will be necessary to encourage and promote extensive applications of modern technology in traditional agriculture in order to improve local agricultural productivity and liberate the surplus rural labour force to accelerate urbanization. Acknowledgments This research is a part of Work Package 8 Scenarios of harmonizing ruralurban interaction, within the project Sustainable Agroecosystem Management and Development of Rural-Urban Interaction in regions and cities of China (SUSDEV-CHINA, ICA4-CT-2002-10004). The participants of the discussion and data collection and analysis for this paper are Wu Peilin, Zhang Junyan, Li Juan, Duan Juan, Liu yang and He Jinlan. Special thanks are due to Leif Söderlund, coordinator of the project, for his instruction and coordination work. 259

References Cai, F. 2001. The Invisible Hand and Visible Feet: Internal Migration in China, Working Paper Series No.5 Institute of Population Studies, Chinese Academy of Social Sciences. Cai, J. 1990. The provincial migration and its impact on the urbanization in China, Geographical research, vol.2: 122-129. Chan, K.M. 1999. Internal migration in China: a dualistic approach, In: Pieke, F. & Mallee, H. (eds), Internal and International Migration: Chinese Perspectives, Richmond, Surrey: Curzon Press. 49-71. Hu, D. 2002. Trade, rural-urban migration, and regional income disparity in developing countries: a spatial general equilibrium model inspired by the case of China, Regional Science and Urban Economics, vol.32: 211-338 Lu, Y. 2001. Reformation of Huji system and harmonious development of rural-urban relations, Journal of Xuehai, No.6: 57-61 Xu, J. 2002. Mathematical Methods In Contemporary Geography, High Education Press, China Zhu, C., Gu, C. & Ma, R. 2003. The influential factors and special distribution of floating population in China. ACTA GEOGRAPHICA SINICA Vol.56 No.5. 260