The widening and deepening of economic integration in Europe and the development of regional blocs elsewhere have brought income convergence to the forefront of policy debate at a time when the academic literatures on `new growth' and economic geography have produced a variety of new theoretical and empirical insights. A joint conference with the Institut d'Anàlisi Econòmica and the Universitat Pompeu Fabra, on `Regional Integration, Trade and Growth', held in Barcelona on 18/20 November, brought together leading participants in this academic debate from Europe and beyond. Papers presented focused on the determinants of integration at the regional level, the convergence of European regions, trade and regional specialization, stabilization policy in integrated markets, and policies to promote regional growth. The conference was organized by a committee chaired by Ramon Marimon (Universitat Pompeu Fabra, Barcelona, and CEPR) and also comprising Richard Baldwin (Graduate Institute of International Studies, Geneva, and CEPR), Daniel Cohen (Université de Paris I, Ecole Normale Supérieure, Paris, and CEPR), Angel de la Fuente (Institut d'Anàlisi Econòmica and CEPR), Massimo Motta (Universitat Pompeu Fabra and CEPR) and Xavier Vives (Institut d'Anàlisi Econòmica and CEPR). The generous hospitality of the Institut d'Anàlisi Econòmica and the Universitat Pompeu Fabra and the sponsorship provided by the Generalitat de Catalunya are gratefully acknowledged.

Jacques Drèze (Center for Operations Research and Econometrics, Université Catholique de Louvain, and CEPR) presented the first paper, `Regions of Europe', at the Palau de la Generalitat de Catalunya. This lecture also served as the inauguration of the `Centre de Recerca d'Economía Internacional', a joint venture of the Universitat Pompeu Fabra and the Generalitat de Catalunya. Drèze introduced the concept of `regions' that belong to no EC member state but rather to the Community directly, discussed various aspects of its economic desirability, and investigated how assets and liabilities should be apportioned at the time of such regions' secession.

Economic Geography

The first two papers focused on location decisions of firms and workers in environments where increasing returns produce an impetus for agglomeration. Anthony Venables (LSE and CEPR) opened the first full day with `Integration, Specialization, and Adjustment', joint with Paul Krugman. In their dynamic theoretical model, trade costs and agglomeration effects work in opposite directions to produce several possible patterns of industrial location. Their main result was that different countries tend to specialize in different industries as trade costs fall. Also, the dynamic process through which a stable equilibrium is attained may reduce the real wages of a high proportion of workers in declining industries.

Alasdair Smith (University of Sussex and CEPR) pointed out that the costs of specialization may be lower if these trade patterns reflect specialization within industries than in cases where particular countries lose entire industries. Smith and others suggested extending the dynamic analysis of labour mobility and the corresponding distributional and welfare effects.

Luis Rivera-Batiz (Universitat Pompeu Fabra) then presented `A General Interregional Equilibrium Model', joint with Carlos Asilis. In their two-sector, general equilibrium model, spatial elements relate to transport costs, location-specific wages, land rewards and proximity to pollution-generating industrial centres. They used this model to study product variety and land use in equilibrium in a framework which considers firms' decisions on location and sales and workers' choice of location and sector.

Thomas Gehrig (Universität Basel and CEPR) pointed out that different equilibria may have very different welfare implications. Edward Prescott (University of Minnesota and Federal Reserve Bank of Minneapolis) enquired about the potential for using this type of model for empirical analysis.

Jacques-François Thisse (Ecole Nationale des Ponts et Chaussées, Paris, and CEPR) presented his joint paper with Masahisa Fujita and Yves Zenou, `Firm Location and Job Creation in Cities'. This developed a standard urban model in which workers compete in land and labour markets and a new firm is free to set its wage. He showed that nationally- and city-oriented firms and those that develop land can all create new jobs by locating in parts of a city in which they enjoy some monopsony power on the labour market.

Massimo Motta noted that some important conclusions depend on the authors' assumption that firms in the city centre do not respond to the new entrants to the labour market.

In `Cities and Growth: Theory and Evidence from France and Japan', Jonathan Eaton (Boston University) and Zvi Eckstein (Tel Aviv and Boston Universities) noted that historical data on the growth of population centres indicated that their size distribution has been relatively constant, while its distribution and evolution have been very similar for both countries. Eaton then presented a theoretical model consistent with these empirical findings which also accounted for the tendency for larger cities to exhibit higher levels of human capital, rents and wages.
David Canning (Queen's University, Belfast) commented that the criterion for moving to a larger city based on higher lifetime utility may no longer be appropriate in a world in which mobile labour may move several times. Ramon Marimon and Jacques-François Thisse questioned the sample selection and techniques used in the empirical section.

Patterns of Growth

Edward Prescott opened the second session with `Barriers to Technology Adoption and Development', joint with Stephen Parente, which proposed a theory of development quantitatively consistent with observed disparities in the wealth of nations and their experience of development. Focusing on firms' technology adoption decisions and the barriers often placed in the path of entrepreneurs, they calibrated their model to observations of US balanced growth and the post-war Japanese development miracle. For this structure they found that the disparity in technology adoption barriers required to account for the huge observed income disparities across countries was within plausible limits.

Angel de la Fuente agreed that technology adoption is critical but showed that interpreting the barrier parameter as a measure of the efficiency of investment instead allowed a standard neoclassical model to fit the data equally well. Hugo Hopenhayn (Universitat Pompeu Fabra) noted that changes in barriers are fairly permanent and argued that countries' changes in rank order should therefore exhibit some persistence.

In `One Business Cycle and One Trend from (Many,) Many Disaggregates', Danny Quah (LSE and CEPR) analysed the co-movements of aggregate economic fluctuations with those embedded in an extensive cross-section of regional disaggregates for the US, to shed light on whether regional disturbances cause macroeconomic fluctuations which then propagate across regions or regional disturbances are prior and cumulative; in the latter case, such disturbances may be driving the aggregate business cycle rather than the more `conventional' candidates: aggregate technology, demand and money. His mixed empirical evidence indicated that aggregate fluctuations were related only to very particular characteristics of disaggregate dynamics.

Albert Marcet (Universitat Pompeu Fabra and CEPR) stressed the need to account for the experience of countries in which some regions have always been poorer than others, such as Spain. The issue is not why regions perform differently over the cycle but whether convergence is feasible, even if it is very slow.
Steve Knack (University of Maryland) presented the final paper of this section, `Institutions and the Convergence Hypothesis: The Cross-National Evidence', which considered the role of institutions in determining the potential for `catch-up' growth. His statistical results suggested that `quality' of institutions is a better predictor of countries' ability to exploit the potential for catch-up than their initial levels of income or human capital. He found strong convergence in per capita incomes for countries with institutions such as secure property rights, which are conducive to savings, investment and production.

Giorgia Giovannetti (Trinity College, Cambridge, and Università di Cassino) pointed out that explicit modelling of institutions and contract enforcement is essential to ensure that possibly large deterrence effects are captured; she also expressed concern that measurement errors would affect the qualitative data.

Political Economy

Torsten Persson (Institute for International Economic Studies, Stockholm, and CEPR) opened the second day with `Federal Fiscal Constitutions. Part II: Risk Sharing and Income Redistribution', joint with Guido Tabellini. They studied the political and economic determinants of regional public transfers to show that realistic restrictions on the policy instruments that provide interregional risk-sharing introduce a trade-off between efficiency and redistribution. They clarified how this trade-off is resolved in political equilibrium under alternative fiscal constitutions, and their model predicted that federal social insurance schemes chosen by voting will oversupply regional risk-sharing, which federal intergovernmental transfer schemes chosen by bargaining will undersupply.

Ramon Marimon objected that the set of contracts from which the government may choose is arbitrarily restricted and that the `pooling' contract upon which voters agree is not efficient: a `separating' solution would be a better choice for the median voter.

In `Uneven Technical Progress and Job Destruction', Daniel Cohen and Gilles Saint-Paul (Département et Laboratoire d'Economie Théorique et Appliquée, Paris, and CEPR) sought to reconcile some micro stories, which view technical progress as a source of job destruction, with the macro observation that historically the labour force has shrunk by less than productivity has increased. Assuming that technical progress takes place unevenly and that goods are complements for the final consumer, they showed that uneven technical progress leads to job destruction in the sector which benefits from it and job creation in the least productive sector.
Ramon Caminal (Institut d'Anàlisi Econòmica) noted that this model provides no explanation of workers' opposition to technical progress and asked whether empirical evidence indicates a reallocation of work between sectors such as manufacturing and services. Cohen answered that if opposite and repeated shocks are expected, the incentive to reallocate is small; wages may therefore fall and unemployment rise. Raquel Fernández (Boston University) supported this argument by noting that repeated shocks in the same direction are unlikely to occur. Zvi Eckstein objected that finite-lived workers will always associate lower welfare with technical progress. Jacques-François Thisse stressed that myopic politicians will focus on the transitional phase in which workers are worse off.

José-Victor Ríos-Rull (University of Pennsylvania) presented a joint paper with Per Krusell, `Distribution, Redistribution, and Capital Accumulation', which investigated the role of the initial wealth distribution in determining an economy's capital accumulation path if the political mechanism allows redistributive taxation. For a model calibrated to US growth data, he found that redistribution of initial capital has large effects on subsequent capital accumulation, while the time-period for which the current tax rate is voted is also important: institutions that allow taxes to change frequently lead on average to higher taxation and lower capital accumulation.

Yannis Ioannides (Virginia Polytechnic Institute and State University, Blacksburg VA) objected that very few existing fiscal systems approximate the one modelled here and stressed that the lack of such data makes evaluation of the reported numerical experiments difficult. Hugo Hopenhayn observed that discussion of `tiny' changes in the initial distribution is misleading since they have discontinuous effects on the wealth of the median voter.

Raquel Fernández concluded the second morning session with `Public Education and the Evolution of Income Distribution', joint with Richard Rogerson, in which they developed a dynamic general equilibrium model of public education provision in a multi-community setting which they calibrated to US data. They found that a pure system of national financing raises average income in the steady state, average spending on education, intergenerational income mobility and welfare.

Teresa Garcia-Milà (Universitat Pompeu Fabra) objected that the model abstracted from many issues that critically affect the relationship between education and income distribution. The bench-mark case in which education is financed solely from property taxes is too extreme, while modifying the model to consider peer and parental effects on the quality of education, together with a more realistic characterization of the housing market, would reduce the measured impact of the policy reform.
Labour Mobility

Pietro Reichlin (Università di Napoli `Federico II') began the afternoon session with his joint paper with Aldo Rustichini, `Diverging Patterns in a Two-Country Model with Endogenous Labor Migration', which developed a theoretical model with increasing returns to scale, free trade and perfect labour mobility to account for various migration patterns and growth paths. Modelling the size of the labour force alone as the stock of human capital, they showed that any wage differential is sufficient to establish a migration flow from the `low-wage' to the `high-wage' country; this differential will persist so that the former country remains permanently underdeveloped in relative terms. Extending the model to allow heterogeneous inputs of skilled and unskilled labour, Reichlin showed that both divergence and reversals of migration flows were then possible.

Jacques-François Thisse remarked that this model could be viewed as a special case of some of his previous work, while various participants discussed how to model endogenous labour mobility that is neither ad hoc nor completely instantaneous.

In `Migration and Growth: The Experience of Southern Europe', joint with Alessandra Venturini, Riccardo Faini (Università degli Studi di Brescia, Innocenzo Gasparini Institute for Economic Research, Milan, and CEPR) developed a model in which labour's response to relative income differentials and public good provision generates a potential migration flow which is then constrained by policy. Their tests on data for internal and international migration flows from Southern Europe indicated that wage differentials alone cannot account for such flows. The income level in the sending country is an important determinant, however, and Faini concluded that the downward trend in migration from Southern to Northern Europe is therefore likely to continue.

The discussion, led by Iain Begg (Department of Applied Economics, Cambridge), focused on the importance of social policies in sending and receiving countries as determinants of migration.

In the final paper, `Regional Labor Market Dynamics in Europe and Implications for EMU', joint with Jörg Decressin, Antonio Fatás (INSEAD) compared the labour market responses to national shocks in Europe with regional shocks in the US. They found that employment shocks are less persistent in Europe and concluded that this is because European countries specialize less than individual US states. Also the bulk of adjustment to employment shocks reflects changes in labour market participation rather than mobility.
Juan Francisco Jimeno (Fundación de Estudios de Economía Aplicada, Madrid) suggested that small changes in their wage equation would produce drastically different simulation results.