In recent years, the issue of industrial policy has received relatively little attention from the economics research community in Europe. European governments and the European Commission have introduced a variety of policies concerning manufacturing industry; examples include state aids, anti-dumping policies or research joint ventures (RJVs). The main objective of an April CEPR/WZB conference was to stimulate debate between economists and policy-makers on the design and implementation of industrial policy. The first part of the conference focused on existing policies, and attempted to provide an analytical framework with which to evaluate industrial policy. The second part of the conference focused in more detail on specific policy issues such as the role and performance of RJVs and expenditure on infrastructure. Lastly, the conference concluded with a panel discussion, with representatives from the European Commission, industry, member states and the transition economies. The conference took place at the Wissenschaftszentrum, Berlin (WZB) on 19/20 April 1996 and was coordinated by Damien Neven (Université de Lausanne and CEPR) and Lars-Hendrik Röller (WZB and CEPR). The conference formed part of a CEPR research programme on ‘Market Structure and Competition Policy’, supported by the European Commission’s Human Capital and Mobility Programme. Additional funding was provided by the WZB and DGIII of the European Commission.

In ‘Technology Policy and National Institutional Frameworks: UK and Germany’ David Soskice (WZB) argued that the institutional framework will have a major influence on the broad type of innovation strategies that firms develop. He characterized different types of innovation strategies, and derived the corresponding requirements for the firm’s organizational structure and the institutional framework. Germany, Sweden and Switzerland, he argued, are economies in which newly emerging technologies and large complex systems (for example, telecommunication or airline systems) are not easily developed. This contrasts strongly with the UK and US. Germany is strong in incremental product and process innovation, often at the scientific leading edge in established technologies such as machinery. Soskice concludes that German technology policy during the 1980s and 1990s broadly reflects the characteristics of the institutional framework.

Wendy Carlin and John van Reenen (both University College London and CEPR) presented ‘Competitiveness and Trade Performance: German Industry in Comparative Perspective’, written with Andrew Glyn (University of Oxford). They analyse the impact of cost competitiveness on trade performance across 14 OECD countries and 12 manufacturing industries for the 1970s and 1980s, using the OECD’s STAN and ANBERD databases. The econometric results show that changes in relative unit labour costs are a robust determinant of changes in export market shares. Nevertheless, there remain significant trends even when changes in costs are controlled for. This suggests that non-price determinants of export performance, such as R&D intensity and investment, are also important. Additionally, the analysis highlights a ‘German paradox’: German industries have maintained export market share in spite of unfavourable trends in costs and in relative investment. Lastly, the results indicate that nominal exchange rate changes can play a role in offsetting a weak underlying export performance.

Considering existing policy, Michael Kende (INSEAD) analysed the evolution of ‘Government Support of the European Information Technology Industry’. The basic policy objective was to narrow the technology gap with the US, and more recently, Japan. During the 1960s, each member state government focused on creating a national champion, in the belief that scale would allow effective competition with IBM. By the end of the 1970s, it was evident that this policy was failing. Policy objectives shifted, focusing more on EU-wide research programs such as ESPRIT. Support was shifted from individual firms to basic IT research. Kende argues that the problem for European firms may not necessarily be the inability to innovate but rather, a relative inability to commercialize technology (making new products using existing technology). He argues that greater attention should be paid to providing an environment in which firms, particularly start-ups, are able to succeed commercially with both existing and new technologies.

Frank Lichtenberg (Columbia University, New York) presented his paper on ‘The European Strategic Program for Research in Information Technologies (ESPRIT): An Ex-post Analysis’. ESRIT is by far the largest of the Research and Technological Development (RTD) programs established by the European Commission, and is scheduled to last at least 15 years, involving over 9000 organisations of various types throughout Europe. Using data contained in several Community RTD databases, Lichtenberg evaluates some important aspects of the structure, conduct and performance of the program. In particular, he focuses on the characteristics of organizations receiving support from ESPRIT. The amount and kind of European industrial cooperation in IT, and the research outcomes (including processes and prototypes, publications, and market applications) are analysed. Lichtenberg’s analysis revealed that the most common type of organizations involved are universities and educational organisations and that there is no bias with respect to organization size. Lastly, although contractors must collaborate with at least one organization from another member state to receive funding, national preferences matter. In other words, the strongest tendencies are to form partnerships with other organizations in the same country.

What can the EU hope to achieve in the field of industrial policy Peter Holmes (University of Sussex) and Paul Seabright (Churchill College, Cambridge, and CEPR) addressed this question in ‘Industrial Policy after Maastricht: What is Possible?’. They note first that the EU does not have a single objective function, due to the differing objectives of the member states’ governments. Within the framework of competition law, decision-making is easier to accomplish because the Commission (subject to the Court) is the sole actor. It is argued that there is no EU industrial policy per se, but rather a range of policies affecting industry. However, in the absence of EU intervention, there would either be inter-governmental competition or cooperation. The authors’ basic argument is that there should be more competition between firms, but less competition between governments (in the fields of standard setting, competition policy, etc.). As a general rule, the more intervention, the stricter are the procedural requirements for action.

Looking at specific policy issues, Bruno Cassiman (Northwestern University, Illinois) and Patrick Greenlee (Antitrust Division, Department of Justice, Washington DC) consider ‘Firm Objectives and the Formation of Research Joint Ventures (RJV)’. Partners rarely disclose clearly the nature and degree of cooperation within a joint venture. Possible objectives for the firm include coordinating research and investment in development, sharing research and development output, and the degree of collusion in the final product market. For each of the two research-related objectives, profits can be written as a function of the partition of firms into joint ventures under different assumptions regarding cooperation in the final product market. In contrast to most of the literature, Greenlee and Cassiman allow multiple ventures to form. They then compare industry structures that arise as equilibria of coalition formation games, and discuss joint venture policy.

In the paper ‘The Centralization of R&D Resource Allocation’, Rune Stenbacka (Helsinki School of Economics) and Mihkel Tombak (University of Helsinki) develop and analyse a model of the R&D budget allocation under both centralization and decentralization. Decision rules for the optimal allocation of resources are suggested. They examine the effects of errors in the assessment of R&D proposals. An analysis is made of the incentives of both centralized and decentralized authorities to improve these assessments through monitoring. The results show that centralized authorities would have greater (less) incentives, on average, for improving the assessment of good (bad) projects. Jurisdictions with a high proportion of good (high probability of success) R&D projects, however, are likely to be dissatisfied with the monitoring efforts of centralized authorities.

A two-stage duopoly game characterized by a one-way spillover structure in process R&D is considered in ‘Endogenous Innovator/Imitator Roles and RJV Performance’ by Rabah Amir (WZB). Know-how only flows from the more R&D intensive firm to its rival. In equilibrium, a priori identical firms always engage in different levels of R&D, giving rise to an innovator/imitator configuration where the innovator is the more R&D intensive firm. Thus, in equilibrium firms are of different size. In view of this endogenous firm heterogeneity, the social benefits of, and the firms’ incentives for, RJVs are somewhat different from the case of ex-post symmetry.

Claude D’Aspremont (CORE, Université Catholique de Louvain-La-Neuve) presented ‘Knowledge as a Public Good: Efficient Sharing and Incentives for Development Effort’ written with Sudipto Bhattacharya (London School of Economics), and Louis-André Gérard-Varet (GREQAM, École des Hautes Études en Sciences Sociale, CNRS). They consider the question of RJVs where adverse selection arises from knowledge sharing (information exchange). Additionally, there is moral hazard involved in the choice of private development efforts which are aimed at translating privately acquired and/or shared knowledge into valuable market innovations, as actions remain individual. The model of Bhattacharya, Glazer and Sappington (1992) is extended to situations where one cannot identify a ‘most knowledgeable partner’. The authors define the conditions for transfers that implement both efficient first-best knowledge sharing and subsequent development efforts.

In ‘Production Function Estimates of the Rate of Return on Public Infrastructure’ Manuel Balmaseda (CEMFI, Madrid) examines the typically high output elasticities and high rates of return of public sector capital estimated from Cobb-Douglas production technologies. Balmaseda replicates the high rate of returns but then contrasts these results with the low elasticity estimates obtained if disaggregated state-level data are used. These low elasticities are corroborated under alternative technology specifications with regard to factor input mobility and under disaggregation of the public infrastructure capital according to its functions. The importance of state-specific effects is highlighted in all specifications. Lastly, the state-level data suggests that public capital makes no significant contribution to private state output.

A panel discussion closed the conference. The panel was chaired by Alexis Jacquemin (Université Catholique de Louvain-La-Neuve and the European Commission). He posed two questions to the panel: What type of industrial policy will work? What is the role of EU policy, if it exists?

Helmut Sihler who is the former CEO of Henkel AG, and Chairman of the Board of several firms including Porsche AG, Degussa AG and Deutsche Post AG, discussed the issue of market failures. He asserted that there is certainly a role for the government to intervene, but only if the projects would not have taken place without a subsidy. He argued that government action would be better aimed at fostering the appropriate political/legal framework, rather than simply giving out subsidies.

Dominique Strauss-Kahn (Ministry of Industry, France) started by observing that France has a somewhat strong reputation in the field of industrial policy. He focused on three areas: cooperative R&D, trade policy, and competition policy. He argued that that research policy should be focused on helping both small and large companies to take risks and gave some examples of successful Franco-Italian R&D.

He discussed the issue of anti-dumping procedures, arguing that it takes a relatively long time in the EU to stop a firm, whereas in the US, the procedure is almost instantaneous. Lastly he believes that competition policy shows the EU at its best. Strauss-Kahn concludes that more attention should be paid to those sectors where there is a social dimension (i.e. declining manufacturing sectors), and also to hi-tech industries, that after Maastricht not much is possible, unless there is a political step forward.

Károla Attila Soós (Ministry of Industry and Trade, Hungary) explained the objectives of Hungary’s current (1996) industrial policy: to increase investment and restructuring (with the aim of joining the EU); to increase value-added in production and exports; to make the legal framework closer to the EU framework, again, with the aim of integration; and, to create a trade surplus.

In terms of industrial policy, to achieve the objectives set out above, he asserts that there are no sectoral priorities. The tax system should be rationalized, accompanied by the privatization of the banking sector. The capital market will be developed. There will also be a focus on increasing the share of small- and medium-sized enterprises in industrial production, plus the development of infrastructure, and participation in European networks.

To date, FDI in Hungary has gone to Budapest. There is increasing investment in North East Hungary, due to the strong government incentives offered. Currently, FDI in Hungary totals $12-13 billion. Areas that have been specifically targeted include microelectronics, IT, telecommunications, component parts for electronics and cars, and waste recycling. Soós also mentioned a few problems: EU subsidisation of the steel industry has caused problems for Hungarian steel, as their capacity is unsellable. EU competition policy has been unsuccessful here. EU labour market rigidities may cause problems for Hungary.

Stefano Micossi, Director General of DG III, closed the panel discussion. The EU is often seen as a ‘protectionist’ bloc, and some sectors are indeed protected. The EU has, according to Micossi, been slow on the issue of anti-dumping, and he observed that there are some grey areas in EU trade policy, such as the car industry. On the topic of R&D, he said that 5% of the total public research budget goes to the ESPRIT and Eureka programs. Airbus, for example, was never an EU-sponsored research project, but instead an inter-governmental one. He agreed that some sectors are characterized by a substantial amount of procurement bias (e.g. defence). But while the EU has tried to apply general principles, some of its member states have resisted the application of these principles.

According to Micossi, the EU is moving towards strengthening competition, enhancing innovation, and providing a supportive environment for EU firms. A strong IT industry is seen as important. He argued that markets should be open to competition, with a rule-based policy, where the degree of intervention is lowered. The nature of this process is such that there must be a consensus.