What future for the regions of Europe as the continent pursues ‘ever closer union’? The first of four articles on their prospects explores the ‘new economic geography’ and its attempts to explain industrial location and the origins of inequalities between regions.

CEPR researchers have been among the pioneers of the ‘new economic geography’, which offers a new approach to the location of firms. Three important questions are at stake: why do firms often cluster together? What leads regions with similar underlying characteristics to turn out so very different? And how does the propensity of industry to ‘agglomerate’ change as regions become more integrated? The latest research is surveyed by Gianmarco Ottaviano and Diego Puga in a recent CEPR Discussion Paper.

Agglomeration happens on a number of different levels. At one end are the small scale agglomerations of finely defined sectors, such as US carpet production in the city of Dalton, Georgia and the Italian textile industry in the city of Prato. At the other end are large scale agglomerations cutting across state and national boundaries. These include the US ‘manufacturing belt’ (the parallelogram with corners in Green Bay, St Louis, Baltimore and Portland) and the European ‘Hot Banana’ (the area between Milan and London, containing Northern Italy, Southern Germany, South East France, the Ruhr area, the Île de France, Belgium, the Netherlands and South East England).

The new models of industrial location explain the large scale phenomena as the result of the tension between two opposing forces. Agglomeration forces encourage firms to concentrate in a few locations: firms benefit from locating near each other because, for example, they have access to larger pools of skilled labour. But wage differences (or other factor supply issues) push in the opposite direction: if too many firms locate close together, the competition for labour forces wages up. This encourages the dispersion of industry. The agglomeration forces may overcome the dispersion forces if workers migrate easily or if there are strong vertical linkages between firms operating in the same or related industries. For example, suppose that regions have identical technology and endowments and contain two sectors: agriculture and manufacturing. If industry is imperfectly competitive and there are some trade and transport costs, firms want to be close to other firms supplying intermediate goods to reduce their production costs – a cost linkage. At the same time, the presence of firms using intermediate goods raises the sales and profits of intermediate goods suppliers – a demand linkage.

When the manufacturing sector is small relative to the rest of the economy, these cost and demand linkages can induce it to cluster in a few regions. A combination of scale economies and transport costs encourages firms to locate close to large markets, which in turn are those with relatively many firms. Wages in these regions will be higher than elsewhere but the positive ‘pecuniary externalities’ created by linkages will compensate for the higher wage costs and the congestion.

The hallmark of these location models is that agglomeration forces tend to encourage the concentration of industrial activity via ‘cumulative causation’. In other words, spatial concentration itself creates an environment that encourages spatial concentration. But by affecting the balance between agglomeration and dispersion forces, the process of regional economic integration can also have a decisive influence on industrial location.

What happens in these models is that the early stages of integration bring larger gains for more industrialized regions, as firms exploit scale economies by concentrating production close to markets where they have more customers and suppliers. Agglomeration forces tend to be more sector-specific than dispersion forces: for example, a given firm will only have buyer and supplier relationships with a handful of sectors, but would suffer from having to compete for office space with almost any sector. As a result, integration encourages increasing specialization even when regions are a priori very similar. This tends to increase the differences between rich and poor regions, even if there are overall gains.

Income disparities across the regions of the European Union (EU) are much wider than across the United States, yet manufacturing is less geographically concentrated. Will closer integration make the economic geography of Europe more similar to that of the United States, with higher industrial concentration and narrower income differentials? It may not if inter-regional and inter-EU migration remain as minimal as they are at present. Despite the single market for goods and labour and large intra-country wage differences, there is very little migration across countries: only 1.5% of EU citizens live in a member state other than the one in which they were born. This contrasts with the high mobility of US workers.

If labour does not move, for low enough trade costs, it will be firms that move. Eventually the same forces that foster divergence can reverse it. When regions become sufficiently integrated, firms in labour intensive sectors increasingly relocate to the regions where factors that cannot be easily transported (notably labour) cost less. Growth in these sectors then creates demand for capital and intermediate goods, and can lead to convergence by the less favoured regions.

What about the empirical evidence? Are pecuniary externalities relevant? And does economic integration shift the balance between these agglomeration and dispersion forces? A recent CEPR Discussion Paper by Marius Brülhart and Johan Torstensson studies the evolution of industrial employment patterns in 11 EU members over the period 1980–90. These authors find support for the models’ main implications: manufacturing is concentrated in regions close to the geographical core of the EU and the degree of concentration is higher in sectors with larger scale economies. The degree of concentration has, however, fallen in the 1980s.

The ‘new economic geography’ has a number of important policy implications, notably in trade policy and infrastructure policy: the design of trade agreements and the building of infrastructure networks can play an important role in shaping regions’ locational advantages in terms of access to world markets. Another promising application is European regional policy, the subject of the next article.