The current arrangements for the EU's main regional policy instrument – the Structural Funds – expire in 1999, at which point new arrangements will be needed to ensure effective regional policy until at least 2006. At a lunchtime meeting hosted by ECARE, and jointly organized by CEPR and ECARE in Brussels on 25 November 1997, Diego Puga (LSE and CEPR) described the lessons afforded by recent location theories for the design of the required reforms.

The context for the reform of regional policy will have three components. The first is the expected enlargement of the EU over the next few years. The second is the profound regional income disparity that exists within the Union and which has persisted despite an intense process of regional specialization in different sectors in the EU, and continuing resource transfers. Third, in spatial terms, regional problems have come to be viewed less as large disadvantaged regions and more as concentrated pockets of poverty and unemployment.

Puga reflected on the lessons for all three of these elements. On the challenges posed by enlargement, he noted first that the applicant countries with which negotiations were likely to begin in 1998 have a GDP per capita which is only 30% of the EU average. By contrast, the GDP per capita of Greece, the poorest member of the current union, was 63% of the EU average. Although enlargement will open new opportunities to firms and consumers with potentially large benefits for both East and West, these disparities will have significant implications. In particular, although EU membership must involve the same status for new entrants as for existing members, different aspects of membership may need to be phased in at different speeds.

For example, if full labour mobility were introduced immediately, too large a gap in wage levels between new and existing members could drain the newcomers of their more skilled workers. But although it may be tempting to try to bring applicant country wage levels close to those of the EU as rapidly as possible, such a process should not be rushed. A long transition period before full labour mobility is achieved would be preferable to unjustified wage increases in the new entrants, since this would push the objective of a more cohesive Union further away. German reunification had provided a clear example of how integration can homogenize wages across regions that offer very different degrees of attraction to firms. Between the first quarter of 1990 and October of that same year wages in former Eastern Germany rose by 42%. Half that increase was accounted for by the harmonization of social security contributions alone. Timing the transition to enlargement was thus a crucial issue to be addressed.

On the issue of regional disparities, Puga claimed that over the past 15 years income differences across member states had fallen, but that inequalities among regions within individual member states had risen. This had led the European Commission to propose decentralizing the administration of Structural Funds to local governments, on the basis that they would have better information on local needs and on the costs of addressing them. The optimal degree of decentralization would involve a compromise between exploiting better information on local conditions – by delegating the decision to a small local jurisdiction – and taking proper account of broader repercussions by vesting powers in a larger jurisdiction. Funds that could be used to attract particular investments should be administered centrally and be subject to clear rules, otherwise firms would probably shop for aid and regional governments compete for activities they consider particularly important. That would not only be inefficient – it would also undermine the credibility of European regional policy.

Controlling the use of European regional funds would not be sufficient, however. Subsidies offered to firms usually conform to the aid practices of individual member states. The EU has set differential ceilings on state aid; in practice, though, the less prosperous states face more restrictive budgetary constraints of their own. Unless tighter ceilings are introduced for wealthier states, a tool justified as a means of helping less developed regions to escape from 'underdevelopment traps' could end up trapping them all the more firmly.

The consequences of increasing regional specialization should also be considered. Over the past 15 years, European regions have become increasingly specialized in different industrial sectors, and this process is likely to intensify with further integration. Specialization can yield large overall gains, as firms can better exploit the positive externalities of being close to firms in related activities and can avoid having to compete with firms in unrelated activities for local factors and services. But workers employed by locally declining sectors would suffer as regions adapted. Given the low propensity of European workers to migrate, adjustment would have to take place mainly through the movement of workers between sectors within each region rather than through the movement of workers between regions within each sector. Greater regional specialization would therefore increase the need for schemes designed to enable workers to move from locally declining to locally expanding sectors. Although training in skills which are not sector-specific might help ease the transition from one sector to another, it is sector-specific skills which could give a region the 'depth' of comparative advantage necessary to withstand shocks.

Finally, Puga drew on work in location theory to convey the importance of understanding the wider effects of individual projects, especially in infrastructure. He took as an example the effects of a road project connecting regions with different levels of industrialization. Some growth economists would regard infrastructure as merely another input into production and, hence, would view the road project simply as a growth-enhancing addition to the infrastructure stock of affected regions. But, in at least two respects, this approach understated the role of transport infrastructure in facilitating the movement of goods and people across space.

First, location theories indicate that better communications can make a less developed region a more attractive location for firms by giving them improved access to the inputs and markets of more developed regions. A rise in the level of activity of one industry in such a region could, in turn, induce another supplying industry to produce at a more efficient scale. Cumulative causation could result in the road project providing wider and stronger positive effects for less favoured regions than would be estimated by traditional cost-benefit analysis. Against this, however, improvements in transport infrastructure would also make it easier for firms in richer regions to supply poorer regions at a distance, and could thus harm the industrialization prospects of less developed areas.

Second, the overall impact of a project depends not only on the nature of the project itself but also on the wider economic environment. For instance, the combination of minimal inter-regional migration with wage-setting at the national sectoral level arguably had led infrastructure improvements to worsen the convergence prospects of the Italian Mezzogiorno. Lacking the industrial base and market size of Northern regions, but having similar factor costs, local firms had lost out to Northern competitors as better communications had lowered the natural protection they enjoyed. Traditional cost-benefit analyses therefore needed to be complemented by a quantification of these wider effects to assess more accurately the impact of projects.