The past few years have witnessed many major infrastructure initiatives from European governments together with a new wave of trade integration associated with the completion of the single market and the Maastricht Treaty. In Discussion Paper No. 909, Research Affiliate Philippe Martin and Carol Ann Rogers develop a model of market integration in which `public infrastructure' is broadly defined to include not only transport and telecommunications but also law and order and general public administration, and it forms an input in the production function. Poor infrastructure impedes both internal and international trade by preventing the consumption of a proportion of goods produced and traded by their purchasers.

Martin and Rogers find that trade integration causes firms in the sector with increasing returns to locate in the countries with the best infrastructure, where prices are lower and relative demand higher for domestic goods. Integration of goods and capital markets between two dissimilar countries has ambiguous effects: it attracts firms into the country with the poorer infrastructure and the lower capital/labour ratio, but its low income level and low level of demand drive them away. In general, a higher level of infrastructure, whether brought about by technological progress or public policy, will magnify the concentration effects of differentials in infrastructure, country size and capital/labour ratios. Improving infrastructure in the poor country relative to the rich country may even increase industrial concentration in the latter and induce regional divergence. Policies to foster convergence between rich and poor countries should favour infrastructure that facilitates intra-regional rather than international trade, to ensure that increased domestic demand for goods produced in the poor country will lead firms to relocate there rather than simply amplifying its disadvantage.

In Discussion Paper No. 910, Martin and Rogers relate these results to the negotiations over the regional aid provisions of the Maastricht Treaty, when the Commission sided with Spain, Greece, Ireland and Portugal, on the grounds that divergence could endanger Europe's social and political cohesion and hence its trade and monetary integration. Trade integration was expected to lead to convergence of per capita GDP levels in the 1980s, but there has been some divergence of the poorest EU regions' per capita GDP from the average during 1983-8, despite large capital inflows into the `poor four'. For some categories of infrastructure, however, regional disparities within large countries Germany, Spain, France, Italy and the UK are larger than those found at the country level.
Martin and Rogers apply their model to show that the European Union's structural funds, which reduce disparities in levels of public infrastructure and induce relocation of firms from rich to poor countries, may be interpreted as the price the richest countries pay for the benefits of full trade integration. The removal of all trade barriers by the poor countries could drive firms to relocate from South to North, and regional policies provide a strong instrument to reverse this undesirable effect. Intra-industry trade, patterns of industrial location and per capita GDP are all well explained empirically by regional differences in the provision of education and telecommunications. They are only poorly correlated, however, with transport infrastructure, which EC regional aid programmes have tended to emphasize.

Industrial Location and Public Infrastructure
Trade Effects of Regional Aid
Philippe J Martin and Carol Ann Rogers

Discussion Papers Nos. 909-10, February and March 1994 (IT)