The `catch-up' perspective has transformed economists' and economic historians' views of patterns of growth: economies that grew very rapidly following World War II are now seen to have benefited from a `catch-up bonus'. Because of the war's devastating effects, countries like Germany and Italy had a lot of lost ground to make up, but catch-up may not be automatic, so economies can remain backward. A CEPR workshop held in Munich on 12/13 June re-examined the productivity experience of post-war Europe in the light of recent work on catch-up and convergence. This was the first workshop of a CEPR research initiative on `Comparative Experience of Economic Growth in Postwar Europe'. This is supported by the Commission of the European Communities under its SPES programme and led by Nicholas Crafts, Professor of Economic History at the University of Warwick, Gianni Toniolo, Professor of Economics at Università degli Studi di Venezia, and Rolf Dumke, Professor of Economic History at the Universität der Bundeswehr-München, all Research Fellows in CEPR's Human Resources programme. This workshop was organized by Rolf Dumke and hosted by the Ifo-Institut für Wirtschaftsforschung.

Nicola Rossi (Università degli Studi di Venezia) and Gianni Toniolo (Università degli Studi di Venezia and CEPR) opened the workshop with a paper on `Catch-Up Growth in Postwar Italy'. Drawing on their revised historical statistics for Italy, Rossi and Toniolo measured `true productivity growth' since 1890 by modifying the traditional Solow growth accounting methodology to allow for returns to scale, market power and the capital stock's adjustment to its long-run optimum. Their provisional results suggested that these factors may explain much of the Solow residual.

Knut Borchardt (Universität München) questioned their attributing the dramatic rise in measured real GDP across World War I to the rapid growth of unproductive public expenditure, which Nicholas Crafts suggested could have strong implications for inter-war growth in the catch-up framework. Albrecht Ritschl (Universität München) was concerned about making allowance for the black economy in the Italian data. Steve Broadberry (University of Warwick and CEPR) suggested allowing for a different set of biases in the Solow residual, such as externalities in physical and human capital accumulation, which have featured in recent work on endogenous growth. Bernd Görzig (Deutsches Institut für Wirtschaftsforschung, Berlin) argued that treating capital and labour as fixed and variable factors respectively has become less appropriate as labour markets have become more organized and the importance of human capital increased.

In his paper, `Spanish Postwar Backwardness', Leandro Prados de la Escosura (Universidad Carlos III, Madrid) used the Chenery- Syrquin approach to establish common patterns of development in Europe and analyse Spain's deviations from the `European norm'. His results pointed to the importance of low productivity in agriculture, which held back Spain's development by maintaining a low level of demand for industrial goods and retarding the release of labour and capital from the agricultural sector. Furthermore, Spain's catch-up process was clearly affected by political factors, with periods of stagnation in the 1930s and 1940s and again during 1975-85.

Rolf Dumke noted that the search for historical patterns of development along these lines continued the tradition of the German historical school. Nicola Rossi, in contrast, suggested that a common pattern of development might be captured adequately by the regression approach: common parameters in a common functional form could result in divergence if the explanatory variables took different paths in different countries. Bernd Görzig suggested looking at different regions within Europe rather than imposing a single European norm.

In a paper on `Convergence, Growth and Policy in Germany in International Comparison', Rolf Dumke questioned the role of government policy in Germany's `economic miracle'. After a critical survey of the literature on the role of policy in Germany, he adopted a comparative econometric approach following Kormendi and Meguire, allowing for a number of policy effects in a growth equation for a cross-section of 18 OECD countries during 1950-80. This found support for convergence, but no role for direct policy variables such as government spending and money supply growth, or even general demand-side variables such as exports, investment and inflation. Dumke found periods of deviation from convergence at the regional level, however, which suggested a limited role for policy in the medium run.
Nicholas Crafts noted that Dumke's regression analysis dealt only with demand-side policies and still left a potential role for supply-side policies. David Soskice (Wissenschaftszentrum Berlin für Sozialforschung) argued that institutions were more important than policies as narrowly defined. Nicola Rossi noted that assessment of the impact of policy requires a knowledge of how far policy was from a neutral stance, which raised problems of endogeneity. Jörg Roesler (Forschungsschwerpunkt Zeithistorische Studien, Berlin) noted that a comparison of West and East German post-war growth indicated the importance of both institutions and policy.

The recent debate on convergence is largely based on figures relating to GDP per hour worked, which are available from 1870 onwards for a number of advanced industrial countries. In his paper, `Manufacturing and the Convergence Hypothesis: What the Long Run Data Show', Steve Broadberry considered data on long- term productivity trends in manufacturing. While estimates for the economy as a whole suggest global convergence, his estimates for manufacturing indicated local convergence of labour productivity levels within Europe together with a large, persistent productivity gap between Europe and the US. Although this productivity gap has been stationary since the mid- nineteenth century, there have been substantial swings in comparative productivity over the medium term, which suggest the need for caution in interpreting trends over short periods. In particular, the widening of the gap between Europe and the US across World War II cautions against an over-optimistic interpretation of the rapid European productivity growth of the 1950s and 1960s.

Leandro Prados de la Escosura questioned the consistency of the figures for manufacturing and the whole economy, citing in particular his own work on comparative productivity in agriculture. David Soskice was concerned about the treatment of quality in comparative productivity studies and stressed that labour productivity is an incomplete indicator of economic performance. Wendy Carlin (University College, London) suggested that additional data on relative labour costs might throw light on the issue of international competitiveness.

In his paper, `Productivity Catch-Up and Convergence in European Manufacturing Since 1950', Bart van Ark (Universiteit Gröningen) studied post-war comparative productivity levels for seven European countries, including two in Eastern Europe (Czechoslovakia and East Germany). He noted that patterns of convergence differed for manufacturing productivity between sub- periods and between manufacturing and the whole economy; they were not significantly different at the level of six major groups of manufacturing branches, however, while the observed pattern in manufacturing in Europe differed from that for a larger sample of countries including the US and Japan.

Steve Broadberry suggested trying to confirm the apparently low level of Spanish productivity in the early 1960s by another bench-mark study, since Spain appeared as an outlier in the statistical analysis. Wolfgang Gerstenberger (Ifo Institut für Wirtschaftsforschung, München) was sceptical of data at this level of aggregation and stressed the differences between comparative productivity and competitiveness. Noting the apparently poor performance of German manufacturing in the 1980s, Bernd Görzig suggested that the slow growth of real wages had reduced pressure on firms to raise their productivity, which emphasized the need to consider the direction of causality.

Bernd Görzig brought the first day's proceedings to a close with his paper on `Productivity Comparisons: East versus West Germany', which calculated East German labour productivity levels as a proportion of West German levels in the periods immediately before and after German unification. He found that estimates for 1988 based on physical production indicated a much higher figure than those for 1991 based on sales.

Jörg Roesler noted that such differences between estimates should not be surprising: there is a great difference between what can be produced in a planned economy without consumer sovereignty and what can be sold in competitive markets. Nicholas Crafts suggested trying to measure the relative decline in East German productivity since 1950, when differences with the West were probably still small. Gianni Toniolo also pleaded for the longer view: just as economic historians now debate whether US slavery was economically viable when it was abolished, future economic historians will debate whether central planning was a viable economic system. David Soskice suggested that the East German system was probably well suited to the Fordist system of production in the early post-war period, but increasingly ran into difficulties from the 1960s as more flexible systems of production emerged. Bart van Ark stressed the importance of quality in comparing the performance of East and West European economies.

Wendy Carlin opened the workshop's second day with her paper, `Wages Policy in West Germany', which examined the contribution of wage restraint to Germany's rapid post-war growth. She found that wage restraint had exerted direct effects on growth as improved profits enhanced investment and competitiveness and hence increased exports; its indirect effects operated through the lack of inflationary pressure, which reduced the need for restrictive macroeconomic policies.

Knut Borchardt questioned Carlin's assumptions about the direction of causality: rather than adopting wage restraint to allow rapid productivity growth, West German unions in the early post-war years were surprised by rapid productivity growth and took time to increase their wage demands, bearing in mind their experience after World War I, when wage growth had outstripped productivity growth. Gianni Toniolo reported a similar experience in post-war Italy, while Steve Broadberry stressed that unions and firms bargained over effort as well as wages: there was also wage restraint in the UK, but it was accompanied by a defence of restrictive practices, which reduced levels of effort and productivity. Nicholas Crafts asked to what extent Marshall Aid had helped to establish the politics underlying Germany's post-war productivity by alleviating distributional conflicts.

In his paper on `Convergence and the Role of Institutions', David Soskice argued that national institutional systems had diverged during the 1980s even though labour productivity levels had apparently converged. He distinguished two `families of capitalism': `Anglo-Saxon', liberal market economies exhibit little coordination among businesses and between businesses and banks, and independence of government from business; `North European/Japanese' economies are characterized by close coordination among businesses and between businesses and banks and close relationships between government and business. Soskice maintained that these two systems had recently diverged: North European/Japanese economies had responded to the development of microprocessor technology with `flexible specialization', placing greater emphasis on internal labour market relations and more sophisticated, innovative products, which improved their international competitiveness. Anglo-Saxon economies, in contrast, had responded by enhancing their `financial flexibility', using pre-set machinery whose performance did not depend on the skills of workers, and using the new technology to monitor individual performance with very tight, short-run financial controls. Since these changes in the Anglo-Saxon economies were a response to improved North European/Japanese competitiveness, the resulting convergence of labour productivity resulted from the divergence of the institutional systems.

Albrecht Ritschl questioned the classification of Germany and Japan together, since there may be more differences than similarities between the two economies, but he welcomed the revival of the idea of `economic style' from the German historical school. Knut Borchardt questioned whether the convergence of institutions could be avoided, citing the case of capital markets in a world of capital mobility as an example. Wendy Carlin wondered how Soskice accounted for Germany's poor economic performance during the 1980s. Bart van Ark welcomed his classification of the UK as separate from the rest of Europe, maintaining that its inclusion often distorted the identification of common European patterns.

The workshop concluded with a final round table session, led by Nicholas Crafts, in which participants identified issues for further research in three main areas. First, measures of labour productivity should focus on the accuracy of the data, the identification of patterns of convergence, and relationships among labour productivity, competitiveness and overall economic performance. Second, the key issues in growth performance concern the identification of potential growth rates, which link back to the measurement of initial productivity gaps, growth during reconstruction, and the establishment of long-run trends and common patterns across countries. Third, assuming that we can establish whether economies have performed above or below expectation, explanations of their growth should consider the links between performance and institutions such as industrial relations, training systems, banking structures and technology policy.