At a Brussels joint lunchtime meeting with the European Centre for Advanced Research in Economics (ECARE), Université Libre de Bruxelles, on 1 June, Michael Burda discussed recent research on European labour markets. Burda is Professor of Economics at the Humboldt-Universität zu Berlin and at the Institut Européen d'Administration des Affaires (INSEAD), Fontainebleau, and a Research Fellow in CEPR's International Macroeconomics and Human Resources programmes. His talk was based in part on his CEPR Discussion Paper No. 868, `Gross Worker and Job Flows in Europe', written with Charles Wyplosz and produced as part of CEPR's research programme on `Market Integration, Regionalism and the Global Economy', supported by the Ford Foundation, and also on his forthcoming paper, `Locational Competition versus Cooperation in Labor Markets: An Implicit Contract Reinterpretation', written with Antje Mertens. The views expressed by Burda were his own, however, not those of the above institutions nor of CEPR, which takes no institutional policy positions.

Burda first noted that high and persistent unemployment in Europe in fact masks workers' considerable mobility across labour market states of employment, unemployment and non-participation. Total exits from unemployment also display a countercyclical pattern, as found in the US and Japan, are highly coherent with unemployment inflows, and most strikingly mainly take the form of job findings rather than exits from the labour force. These empirical patterns pose a challenge both to conventional models of the business cycle and to `search' models in which unemployment derives from workers' misperceptions of relative wage offers. While all these models assume the paradigm of a `representative agent', theory must allow heterogeneity of workers and/or firms in order to fit the facts: either some firms fire while others hire during downturns, or the same firms hire and fire different workers.

Burda distinguished flows of employment (which depend on workers' and firms' satisfaction with the match) from flows of jobs (which are assigned profitabilities independent of the match quality). The destruction of `jobs' in this sense is strongly countercyclical, while job creation may be mildly procyclical or even acyclical. While flows of jobs and workers are obviously related, models adopting this simple framework to differentiate between them (which can also distinguish unfilled jobs from `planned positions', for which work-places have yet to be created), can account for the stylized facts of European aggregate unemployment patterns. This also points to the need for further research into firm and worker heterogeneity in aggregate models.

Burda then compared and contrasted European and US experience to focus on the implications of mobility across regions for their patterns of wages and unemployment. The inequality of income distribution across US regions has increased markedly during the last 30 years, while the regional distribution of income in Western Germany (which proxies for Western Europe as a whole) has barely changed. This increased income inequality across the US does not really operate through wage flexibility, however, since incomes of individual occupational groups in different regions are remarkably similar. It reflects rather a combination of increased geographical specialization by individual industries and changes in the inter-industry wage structure. At the same time, high inter-regional labour mobility has led to a constant variance of regional unemployment in the US, which contrasts markedly with the secular upward trend displayed in Germany since the 1960s. Since the inter-regional variance of productivity increased over the same period in both countries, only differences in their regional mobility can account for their different patterns of regional unemployment.

Burda proposed an `implicit contract' model to account for these different experiences. If labour is relatively immobile across regions and cities, regions or nations engage in `locational competition' by bidding to attract capital, real wages become more variable. This gives rise to demands for insurance against such idiosyncratic, insurable risk. This insurance may be provided by collective bargaining institutions or government legislation on minimum wages, labour regulations or progressive taxation. If regional labour mobility is high, on the other hand, as in the US, shocks are mediated by migration instead, and insurance of this type is unnecessary. This model is therefore able to account for the dramatic variation of regional unemployment patterns observed in Western Europe since the 1960s.