The twentieth century has seen huge differences in productivity performance both among European countries and between Europe and the USA. Moreover, although much quantitative research has been devoted to the performance of individual countries, surprisingly little is known about comparative productivity levels and their movement over time. This raises important issues concerning both the measurement and explanation of the differences across countries, and indicates the scope for international cooperation in research to fill out the global picture.
Differences in productivity levels and/or growth rates have traditionally been seen as a challenge to orthodox economic theory, since they appear to suggest that opportunities for profit are being missed in a competitive framework. But recent theoretical work in industrial economics (on barriers to entry, strategic behaviour of firms, agency problems, etc.) and in growth theory (on external effects in human capital formation, research and development, etc.) suggests that persistent cross-country differences in productivity may now be explicable within an optimizing framework. These issues were discussed at a CEPR Workshop on `European Productivity in the Twentieth Century', held in London on 24/25 March 1990. The workshop was organized by Stephen Broadberry (University of Warwick and CEPR), and financial support was provided by the Department of Trade and Industry and the Nuffield Foundation.
The workshop opened with a paper on `Explaining Anglo-American Productivity Differences in the Mid-Twentieth Century' by Stephen Broadberry and Nicholas Crafts (University of Warwick and CEPR), who is Co-Director of CEPR's Human Resources since 1900 programme. Broadberry and Crafts provided evidence on the scale of the UK's `productivity gap' with the USA since 1907, and noted that the productivity of US industry taken as a whole was more than double that of the UK as early as the 1920s. There were considerable variations, however, in the size of this productivity gap for different industries. It was generally much larger for heavy industries, particularly metals and engineering, than for light industries, particularly textiles, food, drink and tobacco, in which the UK maintained its comparative advantage. Broadberry and Crafts reported that the results of both a cross-sectional econometric analysis of the productivity gap for the mid-1930s and of a number of case studies for individual industries suggested that the traditional reasons offered for Britain's economic failure are often valid for particular industries, but that these cannot be generalized to the manufacturing sector as a whole. They further argued that variations in the ability of poorly performing firms to survive in different industries contributed significantly to variations in the productivity levels between industries.
Bart van Ark (NIESR and Universiteit Gröningen), in a paper on `Comparative Levels of Labour Productivity in Postwar Europe: Some Evidence for Manufacturing', provided an overview of the evidence from a number of studies based on the industry-of-origin approach on the comparative performance of industrial productivity across countries. Taking benchmark levels of comparative labour productivity from a number of binary studies of the UK and four other countries (France, Germany, the Netherlands and the USA), van Ark then used indices of output and employment for each country in order to analyse changes in their comparative productivity levels over the period 1938-88. He concluded that the USA had maintained the highest productivity in manufacturing throughout the period, and that France, Germany and the Netherlands had all overtaken the UK by the early 1970s, although French productivity levels remained substantially lower than either German or Dutch levels throughout the 1970s and 1980s. Further, the relative productivity of British manufacturing was shown to be currently at least as low compared with the other European countries and the USA as it was in 1975.
Rolf Dumke (Universität der Bundeswehr, München) presented a paper on `Reassessing the Wirtschaftswunder: Reconstruction and Postwar Growth in West Germany in an International Context'. He argued that the very high levels of post-war growth in Germany and in a number of other countries (including Austria, Italy and Japan) are best explained by considering not only the effects of `catch-up' growth arising from their technological gap with the USA but also the scale of the wartime destruction they suffered. In an econometric analysis of real per capita GDP growth for 16 major industrialized countries, he showed that the fall in GDP between 1938 and 1948 contributed significantly to the explanation of post-war growth, even when allowance was made for the productivity gap with the US as measured in 1950. Dumke concluded that the reconstruction thesis should therefore be regarded as complementary to the catch-up hypothesis, and noted that the reconstruction effect remained important into the 1960s. He also reported more direct evidence of the significance of wartime destruction for Germany, as shown by the ratio of physical to human capital. While there was much destruction of physical capital, Germany's human capital (measured in terms of education and R&D) continued to grow throughout the war, so that the social capacity for growth was maintained.
In a paper on `Productivity in Scandinavia: A Comparative View' Olle Krantz (Lund University) used a similar methodology to van Ark, by establishing relative levels of per capita GDP for 1985 and extrapolating back to 1860 using national data on GDP and population (since data on employment were judged to be unreliable). Denmark's per capita GDP was distinctly higher than Sweden's until the 1930s, fell below Swedish levels between 1930 and 1950, and did not rise to match Swedish levels until the 1980s. Danish productivity also remained substantially above Finnish levels until the 1930s, but this ratio also fell steadily from the 1930s onwards, as Finland's productivity converged to match Danish and Swedish levels in the 1980s. Thus the history of Scandinavian manufacturing industries appears to offer support for the convergence thesis.
Isabelle Cassiers (Université Catholique de Louvain) presented a paper on `Wages and Productivity in Belgium, 1910-1960', written jointly with Peter Solar, in which they sought to explain the transformation of Belgium from a low-wage to a high-wage economy in the middle of the twentieth century. They showed that the behaviour of productivity and real wages were significantly different in the basic industries that were open to international competition and in those industries that catered mainly for the domestic market. Whether union pressure for higher wages in the `international' sector led firms to choose a higher point on a given labour demand curve, or stimulated their adoption of superior technology and work practices, remained, however, an open question.In a paper on `Productivity in the UK Services Sector: Historical Trends 1856-1985 and Comparisons with the USA 1950-1985', Robert Millward (Manchester University) maintained that the measurement of output in public and professional services has led to a distorted view of productivity in the service sector over the last century. The common practice of using the growth of labour input as a proxy for the growth of the sector's output means that the growth of labour productivity is implicitly set equal to zero, and hence that total factor productivity growth is simply capital productivity growth multiplied by the share of capital. Since capital productivity is equal to output divided by the capital stock, and output is proxied by labour, this approach implicitly equates productivity growth with the rise in the labour intensity of technique. For other sectors of the UK economy, it is the growth of capital productivity, rather than labour productivity, that has tended to be zero, at least over the longer term. Millward suggested that productivity growth in the non-marketed sector should therefore be equated instead with growth in the capital intensity of technique, i.e. using capital growth as the proxy for output growth, and that this will lead to a more favourable view of the contribution of services to productivity growth in the UK. Millward also argued that, contrary to conventional wisdom, the record of productivity growth in the transport and fuel sectors since they were brought into public ownership in the UK bears comparison both with those of the privately owned manufacturing sector in the UK and with those of the transport and fuel sectors in the USA, where private ownership is much more extensive than in the UK.
Patrick O'Brien (St Antony's College, Oxford) presented a paper written jointly with Gianni Toniolo (Università degli Studi di Venezia and CEPR) on `The Poverty of Italy and the Backwardness of its Agriculture Before 1914', in which they argued that rural poverty could not be equated with agricultural backwardness. A comparison of the levels of labour productivity and land productivity in Italy and the UK circa 1910 showed Italian output per worker to be about 60% of the UK level, while output per hectare in Italy was about 40% higher than in the UK. O'Brien and Toniolo draw the conclusion that Italian agriculture had adapted to cope with rural over-population by concentrating upon a mix of crops that maximized the utilization of labour and achieved high productivity from scarce land. Rather than viewing industrialization and urbanization as having been held back by agriculture, O'Brien and Toniolo maintained that continuing rural poverty was a result of the failure of Italy's industrial and urban economy to absorb the underemployed labour from the countryside.
In the final paper of the workshop, on `Comparative Productivity in British and German Industry 1907-37', Stephen Broadberry and Rainer Fremdling (Universiteit Gröningen) provided data on physical output per worker for 23 industries, which showed that, contrary to popular belief, the productivity of labour in German industry had not forged ahead of that in the UK by the 1930s. Further, their estimates for labour productivity disaggregated by industry confirmed the pattern of Britain's comparative advantage as shown in the Anglo-American results presented by Broadberry and Crafts. German productivity was substantially higher in heavy industry, while British productivity was above German levels in light industry, and relative plant and market size were important proximate determinants of German/UK productivity levels. They emphasized further that neither the UK nor Germany had responded positively to the challenge of the second industrial revolution in the USA, so that a substantial productivity gap remained between Europe and America.

A selection of the papers presented at this workshop will appear in the November 1990 issue of the Oxford Bulletin of Economics and Statistics.