European economic integration is widely expected to raise productive efficiency through scale economies and may also raise growth if increased efficiency is associated with higher returns to investment and faster capital accumulation. Some regions may not share in the integration windfall, however, if non- accumulated productive factors (such as land and labour) are fixed and production grows in proportion to the stock of accumulated capital. In Discussion Paper No. 651, Research Fellow Giuseppe Bertola develops a model of increasing returns to geographical concentration of factors. Returns to capital accumulation are constant in two non-integrated localities with potentially different levels and growth rates of production. The potential scale economies from integration are achievable only by concentrating factors in the most productive locations. Since not all factors can move, and owners of immobile factors in underprivileged regions will not share in the integration windfall, fiscal policy should be designed to provide lump-sum compensation to such `losers'.

Bertola notes that this simple model's predictions are largely refuted by the data, since countries and regions do converge over time and capital does flow into capital-poor countries that become integrated with more developed ones. He develops a model in which returns to capital accumulation are constant in the aggregate of two distinct localities but decreasing in each considered separately. The increasing returns required for steady endogenous growth with non-accumulated factors are incompatible with the perfect and complete markets that ensure the integrated economy's optimal production, and returns to an investment decision in one locality may spill over to the other in ways that are not reflected in equilibrium factor rewards.

Bertola finds that free factor mobility need not optimize the integrated economy's overall performance, so there may be a privately optimal (but socially harmful) concentration of production: if factors move to one location and spillovers from the other dry up, the overall distribution of production may become less efficient. Geographical linkages between activities should be internalized to individual agents' decisions through tax instruments to prevent such side-effects. In conclusion, Bertola notes that the increasing returns that sustain endogenous growth need not be concentrated in a specific location, so both geographical and intertemporal dimensions of realistic distortions must be considered in planning the tax structure of a united Europe. If economic growth in Europe after `1992' is driven by private investment decisions, production concentration is to be expected, and policies should aim to limit such concentration.

Models of Economic Integration and Localized Growth
Giuseppe Bertola

Discussion Paper No. 651, March 1992 (IT)