Friday, July 21, 2006

Countries might converge, unlike provinces

Albert van der Horst
CPB, Netherlands Bureau for Economic Policy Analysis
Van Stolkweg 14
2585 JR The Hague
The Netherlands
Phone: +31 703383402

Countries might converge, unlike provinces

The European Union is one of the most prosperous parts of the world. Yet there are large regional disparities in productivity, wages, and employment, and they have only increased with the recent enlargement. Thus, there seems to be ample justification for helping lagging regions to catch up, both at the national and the European level. Unfortunately, regional policy appears to be ineffective. Well known are the fruitless attempts of Italy to bridge the gap between the Mezzogiorno and the North, of Germany to bridge the gap between the Neue Länder and the West, and of the European Commission to reduce regional disparities in general.

This study validates one explanation of the ineffectiveness of regional policy. We set off from the New Economic Geography (NEG), a relatively new branch of economics that incorporates agglomeration advantages and location choice in a formal general equilibrium framework. By estimating the key parameters of NEG models with European regional data, we are able to underpin the conclusion that dogged attempts to make lagging regions catch up are often doomed to fail. Lagging regions do not stand alone, but pertain to local core-periphery systems. Economic activity lured to the periphery by subsidies will in the long run end up
in the core. This is because the periphery lacks the critical economic mass.

Regional policy demands a theory that explains the location of production and consumption. Since the early nineties economists have such a theory at their disposal: the NEG unites within a consistent general equilibrium framework older insights from international trade theory and spatial economics.

Firms and workers are subject to centripetal and centrifugal market forces. The owner of a firm must choose a location for his plant. If he chooses a core region, i.e. a region with a large market, then he saves on trade costs. Less goods will have to `exported' to other regions. Therefore, he can set a lower price, and thus capture a larger share of the market. If he chooses a peripheral region, i.e. a region with a small market, then he faces less competition from other local firms. Moreover, he evades urban costs such as congestion and high land prices. Similarly, a worker must choose a location where to live and work. If he chooses a core region, then he gets a higher real wage. If he chooses a peripheral region, he evades urban costs.

The list of centripetal and centrifugal forces working on firms and workers can be extended, and differs between NEG models. The common denominator of the models is, however, that location decisions depend on the balance of these forces, which in their turn depend on trade costs. A range of trade costs supports an even distribution of economic activity. We call this state the `dispersion equilibrium'. A complementary range supports an uneven distribution of economic activity, in which one region hosts a disproportionate amount. We call this state the `agglomeration equilibrium'.

The agglomeration equilibrium is characterised by a spatial wage structure: agglomeration advantages materialise as higher wages in core regions. Moreover, the causality of location choice underlying the equilibrium is circular: firms and workers prefer the core since it has the largest market; the core has the largest market since it host many firms and workers. Thus, if agglomeration equilibria are the rule, regional disparities are difficult to counter with regional policy.

What does the European economic geography look like? A satellite picture of Europe reveals banana shaped beam of light running from London to Milan that indicates a large cluster of economic activity. Zooming in reveals, moreover, that similar core-periphery structures repeat themselves at lower levels of aggregation. Agglomeration is ubiquitous.

Descriptive statistics confirm the eyeball analysis of the satellite picture. It indicates that
agglomeration is especially pronounced at a (sub) provincial level of regional aggregation. Moreover, the little movement in the location of economic activity that can be discerned is attributable to the increasing importance of services as compared to agriculture and manufacturing. Agglomeration is local and stable.

The observation that agglomeration is global is confirmed by the econometric analysis of the spatial wage structure in European regions. We show that wages are relatively high, not only in, but also in the neighbourhood of agglomerated regions. The estimates imply a strong agglomeration advantage that quickly peters out over distance.

It is unsurprising that regional policy often fails to reduce regional disparities in productivity and income. Core regions tend to be better off than their surrounding peripheries. And core-periphery structures are hard to upset since agglomeration advantages pull economic activity to the cores.

In this light the recent shift in thinking on regional policy makes sense. Policy makers seem more willing to recognise that disparities between provinces, and between regions within provinces are persistent. They increasingly target the available funds on regional growth poles.

There is, however, an equity-efficiency trade off. Agglomeration is positively related to overall productivity and growth, but negatively to wage equality. Moreover, improving the market access of peripheral regions with infrastructure - or any other initiative that promotes economic integration - may in fact increase regional disparities. It increases the incentive to locate in the core once the peripheral market can be supplied from here with
more ease. If one is to improve the fate of lagging regions, then a focus on large regions that contain their own core periphery structures, as well as a focus on labour - the least mobile production factor - is desirable.

No comments:


© Copyright 2006, European House Skopje. Blog developed by Zoran Dimitrov