Gilles Spielvogel, Sciences Po Paris, IRD-DIAL and University of Lille 2, Paris, France
Patterns of specialization and the internal geography of regions (assigned to theme
Understanding how increasing economic integration between regions may contribute to transform their specializations is a growing concern of economic theory. Several recent papers have focused on the possible outcomes of further economic integration between countries on the geographic concentration of economic activity within countries, the issue of European Union enlargement being a more or less explicit background (Monfort and Nicolini, 2000; Behrens, Gaigné, Ottaviano and Thisse, 2003). In a quite different perspective, Krugman and Livas (1996), with the Mexico–United States case in mind, have shown that increasing economic integration between the North and the South may decrease geographic concentration in the South. However, in many developing countries, urban primacy remains the rule, and geographic decentralization does not appear to be a common outcome of increased international integration. Rather, it may be observed that regional inequalities are often extremely wide in African countries and that if urban primacy is sometimes decreasing or halting, it is mainly a consequence of the slowing demography. In more advanced countries, such as Brazil, regional inequalities seem to increase as a direct consequence of trade integration (Haddad et al., 2002). Clearly, a detailed empirical assessment of the impact of international integration on internal geography of developing countries is needed. But it may also be interesting to understand how extremely polarized spatial structures have emerged in numerous developing countries, and why it seems unlikely that trade integration may lead to more balanced geography in some circumstances. In order to address this question, it is necessary to build a model encompassing various geographic scales. The emergence of agglomerations at the local level and rural-urban interactions constitute the first level of geographical concentration: even if some were modest in size and others have not survived, proper urban systems existed at early times in various parts of what is now the developing world (as far as Africa is concerned, see Coquery-Vidrovitch, 1991). Then, interregional trade and migration have long been – and still are – powerful engines of spatial differentiation in developing countries, even when these areas where not directly connected with Europe (one can for instance think of trade networks between coastal areas and the desert ports in West Africa prior to colonization; see Hopkins, 1984). Finally, international or intercontinental trade has also played an extremely important, and very early, role in the formation of spatial structures in most South countries. Of course, in a vast majority of cases, the first significant experience of long distance “trade” between a South region and the North (say, Europe) took place within some kind of colonial context. The objective of this paper is therefore to build a model encompassing the first two geographic levels described above and to explore the role of changes in interregional and intraregional transport costs on the evolution of local and regional specializations. The model will borrow features from various trends of the economic geography literature. For instance, the local level part is built upon a von Thünen like model, related to the work of Fujita and Krugman (1995), while the interregional level is closer to a core-periphery model. The first part of the paper is devoted to the presentation of the basic model. The economy is made up of two initially identical regions. The typical region consists of two sectors. The manufacturing sector produces several distinct goods with increasing returns to scale, which are induced by specialization of workers. The agricultural good is produced with decreasing returns to scale. The production of manufacturing goods is assumed to take place in a single central location, while farming is spread over the hinterland. Trade between the city and the hinterland occurs with transport costs. Workers may move costlessly across sectors and locations and settle where they enjoy the highest possible welfare, though it is assumed that the migration process is sequential: migration occurs firstly within sectors and/or within locations and secondly between the city and the rural zone. When interregional transport costs are high enough, both regions remain in autarky and produce all goods. If interregional labor mobility is not allowed, significant welfare differences between the two regions are possible if endowments and/or technologies are different. On the other hand, if workers can move freely from one region to the other, migration may either lead to convergence or induce divergence. In the latter case, a small initial welfare difference between the two regions can degenerate into the total disappearance of the less favored region and sometimes to a decrease in the welfare level of the most favored one. In the second part of the paper, we expose the different trade patterns that are susceptible to arise between the two regions once interregional transport costs are sufficiently low. For instance, one region may specialize in manufacturing goods while the other produces only the agricultural good. Of course, less extreme patterns are also possible, the manufacturing sector being split between the two regions. Which specialization pattern is more likely to emerge for a given level of interregional transport cost depends on multiple parameters, notably endowments and technologies, but intraregional transport costs play a key role.
 | Conference organized through conf-vienna
(copyright Gunther Maier) |