groups. Baldwin and Robert-Nicoud (2002) argue that the government will generally pick losers, because
losers need to lobby to survive, whereas winners do not. Furthermore, policymakers have to take into
account that the financing of policy initiatives require distortionary taxation. Finally, rent seeking and
high transaction costs increase the risks of government failure
The effectiveness of public support for specific technologies or sectors in particular is difficult to
determine ex-ante, because the government usually has limited information about future technological
developments and comparative advantages. The size of external effects is also unknown and differs with
each technology. Because innovation processes and technologies tend to become more complex, the
information problem increases. A strong policy focus on technology is also not completely clear from a
welfare point of view. For example, based on the composition of our export the present comparative
advantages of the Netherlands probably lie in more traditional sectors. Trying to influence these
comparative advantages in the direction of new technology is a risky business. Finally, implementation
costs are higher in the case of specific policy and can amount to 30%
There are thus several potential market failures present with regard to innovation policy that may
legitimize policy intervention. However, policy solutions to address these market failures are not without
risk. Government failure, the risk of making the wrong selections and the terms of trade effects may
lower the effectiveness of government intervention. Authorities should therefore only interfere when
market failures are most severe and when the risk of government failure is small.
4. The innovative score of the Netherlands
In previous paragraphs we have argued that stimulating productivity growth can be an important element
in achieving economic growth and that there is a theoretical basis for innovation policy. In the following
two paragraphs we look at the innovative position of the Netherlands and the design of innovation policy.
A commonly used indicator with regard to productivity and innovation policy is the total of R&D
expenditure. The Barcelona goal is to spend 3% of GDP on R&D. This should not be interpreted as a
goal in itself, but as a benchmark (AWT, 2002: 14). R&D is after all only an input, whereas in the end
only the results are relevant. Furthermore, just as other investments, R&D will have diminishing
marginal returns and the social optimal level of R&D-expenditure depends on several different factors
and should not necessarily be equal to 3%.
Presently, the Netherlands private R&D expenditure of Dutch companies is 1.2%, which is below OECD
average (figure 2). The relative low expenditure on R&D can partly be explained by the sectoral
composition of the Netherlands. The Dutch economy has few sectors of industry with a high R&D-
intensity such as the pharmaceutical industry, ICT-industry and airplane industry. The sectoral
composition explains circa 50% of the relative R&D gap compared to a set of high performing OECD
countries (Hollander and Verspagen, 1999). R&D investment of seven large companies largely
determines total R&D expenditure.