A great deal of discussion has occurred on which policies improve the scientific and technology base of a country and so promote innovation and growth. It is necessary to analyse the theoretical frameworks that inform policy decisions. We shall argue that evolutionary framework best captures the nature of the innovation process and that the central problem of policy is how to connect together the different institutions that are involved in the innovation system. Here, the main task of policy-makers is to create technology support system.

Usually science and technology are treated in a hierarchical manner. However, science and technology are largely independent and mutually beneficial bodies of knowledge. Science and technology are created by different processes within distinctly different communities located in different institutional contexts. The two communities respond to different incentive mechanisms.

In broad terms, science is open and codified; its outputs are widely diffused. Its primary incentive is knowledge for its own sake through international publications. Conversely, technology is naturally practical and tacit; its outputs are artifacts. It is closed and is concerned for secrecy or patent protection.

Technology can be defined in a number of complementary ways: as knowledge, as skills, as artifacts. The traditional rationale for policy intervention is market failure.

Three sources of market failure in innovation can be identified:

• technological knowledge is a public good - a non-rival and non-excludable good: actors has incentives not to reveal their preference for knowledge in hope of avoiding to pay for it;
• divergences between private and social costs and benefits; many beneficiaries (free riders) do not contribute to the cost of technological knowledge;
• uncertainty - unable to make estimations of technical and commercial returns to innovations; it is inappropriate to use probability calculus.

Much decision making about technology creation is an act of faith with unpredictable time delays between creation and application. One immediate consequence of this is that information will be asymmetric between firms and potential suppliers of capital, and between the R&D managers and the firm’s board of directors.

The consequence of these various market failures is the absence of markets that promote the efficient creation and exchange of technological knowledge. One broad set of market failures relates to the nature of the commodity itself. A second broad class of market failures relates to the indivisible cost incurred in creating the knowledge behind an innovation.

To overcome these problems, publicly-funded laboratories and subsidies are required.

While this perspective has considerable influence in policy debate, it has major limitations.

• The approach fails to provide any details about the design of policy instruments, or their appropriate method of implementation, or the areas that are most in need of support.
• The approach conceives competition in terms of equilibrium market structures. Yet the evolutionary theorists argue that competition is a process of change.

Innovation implies the existence of differences between firms and of information asymmetries. Innovating firms have different knowledge, ‘imperfections’ should be seen as integral and necessary aspects of the production and dissemination of knowledge.

According to evolutionary economics, effective competition depends on diversity in behaviour, and this can only be maintained by continual introduction of new and better products, and new and better methods of production. Competition is a process of introducing and diffusing diverse discoveries. Both innovation (the generation of variety) and market selection are essential parts of the competitive process.

The natural tendency of the evolutionary process, however, is to concentrate production, to destroy the diversity that drives competition, and to tend towards monopoly. Consequently, if competition is to be preserved, mechanisms must be in place to recreate diversity. Innovation mechanism, and technology policy in particular, has to ensure and recreate diversity.

Consequently, technology policy becomes an essential component of competition policy. Within this perspective, the emphasis of innovation policy shifts from correcting ‘market failures’ to encouraging experimental behaviour. Thus, the attention of policy makers shifts from efficiency towards creativity. Rather than supporting a few firms on selected projects, policy is designed to maintain the conditions of open and diverse technological opportunities.

As firms have different capabilities, technology policy needs to be sensitive to their difference in the innovation opportunities. There are two types of policy:

• Primarily concerned with resources and incentive, and takes the technological possibilities and capabilities of firms as given. Such policies encourage firms to change the scale, and focus of their effort. For example, tax allowances for R&D, specific innovation subsidies, public purchasing of innovative products, and terms and duration of patent protection.
• Another type of policy aims to change and enhance the innovation possibilities that firms face by improving their access to knowledge, and by improving managerial capabilities. Such policies aim to improve the links between the firms and other institutions and agencies in the technology support systems.

This latter policy emphasise the development of the science and technology infrastructure in the economy; an infrastructure that facilitates the intercommunication of existing research, and mutually shapes the future research agenda of different organisations. This infrastructure is a set of interconnected institutions that create, store and transfer knowledge and skills that define technological opportunities.

In this infrastructure, many institutions are involved: universities and other institutes, private firms, private consultancies, government laboratories and industrial research associations. Networks, public committees, and other non-market methods coordinate this strong division of labour, ensuring that knowledge flows more easily. Two routes can achieve greater connectivity.

• One way of greater connectivity has been to emphasise the ‘national’ domain of the science and technology infrastructure. Yet this can be extended downwards and outwards. Because different technologies have different supporting infrastructure, a sectoral perspective becomes essential. Also, science has to be recognised as an international system, so that governments collaborate in many areas, such as defence and aviation.
• A second route to building infrastructure has been to create new ‘building institutions’ between industry and academia to facilitate the effective transfer of knowledge, and to shape the research agenda. Identifying potential areas of knowledge involves difficult matters of judgement, drawing together various knowledge inputs, establishing communications between science and industry, and mobilising resources according to strategic objectives.

The role of policy is to facilitate the ongoing development of innovative variety. It creates the conditions for the emergence of new opportunities (i.e. infrastructure).