تجارت، مهاجرت از روستا به شهر، و اختلاف درآمد های منطقه ای در کشورهای در حال توسعه: یک مدل تعادل عمومی فضایی با الهام از مورد چین

Inspired by the case of China, this paper develops a spatial agglomeration model to explain the increasing regional disparity in China and explore several policy implications. The model shows that the improving trade condition and the increasing rural-to-urban labor mobility in China may be the reasons for the enlarging income gap between the coastal area and the hinterland. With a geographical advantage in international trade, the coast becomes the initial location for industrial agglomeration and its leadership becomes strengthened by the positive feedback mechanism from increasing returns to scale. The necessary labor supply for industrial agglomeration in the coast comes from intraregional rural-to-urban migration instead of interregional migration. As a consequence of the industrial agglomeration, the income disparity between the coast and the hinterland increases. The location disadvantage of the interior comes from higher transportation cost in international trade. However, the model suggests that increasing domestic accessibility can actually make the interior worse off.

Along with new opportunities in economic growth has come the problem of increasing regional disparity in a growing number of developing countries. In China, for example, the income gap between the coastal provinces and the hinterland has grown fast since the country opened its doors in the early 1980s. At the same time, there was a rapid agglomeration of manufacturing in the coast. Empirical studies have suggested that the surge in international trade and foreign direct investment (FDI) and the significant increase of rural-to-urban labor mobility may be important factors for manufacturing agglomeration and regional income disparity (Fujita and Hu, 2001). In this paper, we set up a spatial general equilibrium model to explain how globalization and rural–urban migration may affect regional disparity in developing economies. We focus on exploring the policy implications of the model. Early literature in economic development (Hirschman, 1958) has described the changes in regional disparity as an inverted U-curve (the Kuznets curve). As stressed by Williamson (1965), ‘regional income differentials increase in early development stages, then stabilize, and then diminish in mature periods of growth.’ One essential assumption behind the inverted U-curve is that increasing returns to scale lead to the fast growth of certain regional centers of economic strength, which was termed the ‘polarization process’ in Hirschman (1958): ‘On the other hand, several unfavorable or polarization effects are also likely to be at work. Comparatively inefficient, yet income-creating, Southern activities in manufacturing and exports may become depressed as a result of Northern competition… A most serious, and frequently observed, polarization effect consists in the kind of internal migration that may follow upon the economic advances of the North. Instead of absorbing the disguised unemployed, Northern progress may denude the South of its key technicians and managers as well as of the more enterprising young men.’ Because of the difficulties in modeling market structures of non-perfect competition, these ideas were not formalized into a general equilibrium framework until recently. Starting from Fujita (1988) and Krugman (1991), recent literature in spatial agglomeration modeling tries to formalize these ideas in a spatial general equilibrium framework. In this line of research, production agglomeration can be explained as a result of increasing returns in manufacturing or services sectors. For a detailed review of the literature, see Fujita and Thisse (1996). The agglomeration models were extended into three-location frameworks to explore the effects of international trade on regional development (see, for example, Krugman and Elizondo, 1992). However, most existing literature focuses on production agglomeration while little of it addresses the income disparity issue directly. Note that production agglomeration does not necessarily imply regional income disparity in per capita terms. Even in the core-periphery structure that is intensively modeled in existing literature, the difference between the core and the periphery is just specialization rather than disparity. Two exceptions were made by Krugman and Venables (1995) and Puga (1999), which addressed the international or interregional income disparity issue from the view of production agglomeration. Krugman and Venables (1995) developed another application model to explain the relationship between globalization and national inequality. They assumed a perfectly elastic supply of labor from agriculture to manufacturing, so that industrial agglomeration has no effect on agricultural wages in their model. Following Krugman and Venables, Puga (1999) modeled the rise and fall of regional inequalities driven by changes in transport cost. Puga considered agglomeration both with interregional migration and without interregional migration. However, Puga also assumed a free shift of labor from agriculture to industry, which is not true in China. This paper models the increasing income disparity as a result of manufacturing agglomeration that is caused by both increasing returns to scale and first-nature advantage in location. Differentiating from the above literature, this model introduces intraregional rural-to-urban migration, rather than interregional migration, as an important factor in manufacturing agglomeration. The model contains a developing economy with two domestic regions and the rest of world. Within each region, there is an urban center and a surrounding rural area. The rural area is characterized as a traditional agricultural society, with low labor productivity and low income. The urban area is associated with industrialization, modern sectors, higher productivity, and higher wage rate. Regional disparity is closely related to the urban–rural disparity. In advanced regions, urbanization levels are high, while in backward regions the urbanization level is relatively low. Frequently subjected to policy restrictions, rural–urban migration is a common phenomenon in developing countries. In our model, rural–urban migration provides a labor supply for manufacturing growth; furthermore the supply of labor from rural to urban manufacturing is not perfectly elastic due to policy restrictions. Agglomeration does not necessarily lead to per capita income differences; however, when cross-region migration is restricted, production agglomeration can directly cause regional income disparity. Intra-regional rural to urban migration makes the manufacturing agglomeration possible without inter-regional labor migration. In addition, our model embraces several considerations that are specific to the Chinese economy: (1) the domestic market is relatively small and the international market plays an important role in the economy (China’s export and import volume was about 45% of its GDP in 1994), (2) there is a large population but a very limited amount of arable land, which results in low marginal labor productivity in the rural area, (3) interregional labor migration is highly restricted by both government intervention and labor market imperfections, (4) laborers are heterogeneous in terms of skill, and the mobility from low-skilled labor to high-skilled labor is rather difficult, and (5) the model introduces rural industry township/village enterprises (TVEs), which play an important role in the Chinese economy. The model first explains the mechanism of a trade-led agglomeration process. Globalization, which integrates developing economies into the world market, gives border locations a geographical advantage in terms of saving transportation costs in trade. This first-nature advantage breaks the symmetry between the two regions, and drives the manufacturing sector to agglomerate in the coast. Further agglomeration is generated through positive feedback mechanisms. Manufacturing agglomeration, however, is not associated with interregional labor migration. The necessary labor supply for fast manufacturing growth in the coast comes from intraregional rural–urban migration. As a result of the manufacturing agglomeration, more agricultural labor shifts to the manufacturing sector in the coast than that in the interior. Because the marginal labor productivity is low in rural areas, the more labor shifts to the manufacturing sector, the faster the regional per capita income grows. Numerical simulations are used to explore policy implications of the model. The location disadvantage of the interior comes from higher transportation costs in international trade. However, the model suggests that increasing domestic accessibility can actually make the interior worse off. The model also shows that despite the lower per capita income, the interior may not have a lower production cost for manufacturing because government restrictions on rural–urban migration works against the scale of manufacturing. Easing the restriction can help to generate a lower-wage cost advantage for the interior manufacturing, which can prevent manufactures moving to the coast and reduce the regional income disparity. We also explore the possible results of freeing interregional migration. Introducing interregional migration will strengthen the production agglomeration in the coast while reducing regional income disparity. The next section provides some facts of the regional disparity in China that motivate the model and help to understand the assumptions. Section 3 presents the formal spatial equilibrium model. 4 and 5 examine the effects of globalization, transport costs, rural–urban migration, and inter-regional migration. Section 6 summarizes our main findings and discusses policy implications.

This paper applies a spatial equilibrium model to explain increasing regional disparity in developing countries, with a special focus on China. We model a three-location spatial economy, consisting of a coast region, an interior region, and the rest of the world. Within each region, there are rural and urban sectors. Rural–urban migration is subjected to restrictions, and there is no interregional low-skilled labor movement. Under this spatial structure, we analyze the relation of international trade, production agglomeration, and regional income disparity, and discuss several policy effects. With the existence of transport cost, globalization provides a location advantage to the coast, promoting production agglomeration. Increasing returns to scale in intermediate-goods production and increasing returns to varieties of intermediate goods provide a positive feedback mechanism, which can strengthen the trade-led agglomeration. When interregional labor migration is highly restricted, production agglomeration can directly cause a regional income disparity between the coast and the interior. The model shows that along with increases in trade accessibility, regional income disparity between the coast and the interior is also affected by high-skilled labor mobility and degree of increasing returns. A striking finding from the model is the inverse effect of transport cost. Although high transport cost is the direct reason for the interior’s disadvantage, the model suggests that reducing transport cost can actually make the interior worse off. This result depends on the degree of scale economy, the relative size of the domestic and international market, and the existing level of transportation cost. Numerical simulation suggests that the inverse effect will occur when there exist strong increasing returns to scale and a large domestic market. The inverse effect provides an economic justification for the local protections that have been widely practiced among local governments in China. We use the model to investigate the effects of rural–urban migration and policies restricting such migration. With the absence of interregional labor mobility, rural–urban migration provides the necessary labor supply for urban manufacturing growth. Manufacturing agglomeration in the coast region shifts more rural labor to the manufacturing sector. The interior rural area contains a larger excess (relative to the arable land) labor pool, which can be a potential lower-wage-rate advantage for the interior’s manufacturing. Numerical simulation shows that removing the rural–urban migration restriction in region 2 will indeed make the interior more attractive to manufacturing, and will reduce the regional income disparity. The conclusion is robust to different trade costs and transport costs. The model and simulations can help us have a clear view on the trends in regional disparity and manufacturing agglomeration in China. There are two factors for us to consider. First, the trend toward globalization is becoming stronger. The Chinese economy will become more open and trade conditions will be further improved after China joins the WTO in the near future. Trade and foreign direct investment will continuously generate a great impact on the Chinese regional economy. As shown in our model, we can expect that along with the improving trade conditions and FDI, manufacturing agglomeration in the coast will get stronger. Second, the Chinese economy is in a transition toward market economy. Economic liberalization will promote greater factor mobility and a stronger scale economy. As Fig. 7 shows, a stronger scale economy will increase manufacturing agglomeration further. If the rural–urban migration restriction and the interregional migration restriction are retained, regional income disparity will stay high, and even increase along with rising manufacturing agglomeration. To prevent the income disparity from rising further, increasing labor mobility is necessary. As our simulation shows, removing the rural–urban migration restriction will be effective, and it will not hurt the national economy as a whole. In the long run, increasing interregional migration freedom is a necessary condition to reduce the regional disparity in real income.