The knowledge tree: CSIRO in Australia's innovation systems.

This special issue of IMPP is devoted to presentation and consideration of the important aspects of CSIRO's roles in Australia's innovation systems. The recent Productivity Commission review of public support for science and innovation in Australia included CSIRO in its review but only briefly considered the nature of the context in which CSIRO operates. The review was broadly favourable to the Organisation as it has altered over the last few years, although it made no judgement on the level of funding appropriate to enabling CSIRO to maximise its contribution to innovation in Australia. The recent Budget, however, was not wholly sympathetic to CSIRO: on the one hand, it increased the funding cycle to four years but, on the other, allocated a small rise of 2% over the four year period which in real terms is a decrease.

This ambivalence towards CSIRO and its work has characterised most of the Organisation's history, to the extent that several commentators over the last decade have referred to restructuring CSIRO and its funding as 'a strange national sport' (New Scientist, quoted Ewer 1995 Foreword) or a 'peculiar national past-time' (Johnston 1995: 46). There always seems to be a thought in the heads of governments of whatever political persuasion that somehow the CSIRO, as recipient of considerable amounts of public monies, should be doing 'more' for the nation. This attitude has intensified or retreated at different times but remains a theme in policy development for research systems in Australia. CSIRO has probably been put under a not always friendly microscope more often than most.

The reasons for this attitude are several. Some derive from ambiguities written into the CSIRO Acts. The central mission of the Organisation is essentially to undertake scientific and technological research which is of benefit, both as public good and economically, to Australia. These admirable aims do not always translate cleanly into neat forms of organisation and clearly visible results, however. In some ways paradoxically, because the CSIRO is so large an element of the Australian research and innovation systems, the more visible CSIRO work becomes to some sections of the community, the more others complain.

But there are other issues. Some relate to the capacity of public policymakers, especially politicians but also officials, to understand the broad importance of science to economic development through innovation. Policymakers, and often scientists themselves, whether in CSIRO or the universities, have struggled to comprehend the nature of innovation itself and the processes by which major innovations are developed, adopted and managed.

The argument of this introductory paper is that sensible policies for and judgements about the roles, actual and potential, of the CSIRO in Australia's innovation systems cannot be made in an environment where policies are not coordinated and complementary across relevant areas of economy and society. Many performance measures in present circumstances are inappropriate because key elements of the CSIRO's operating environment and the constraints they impose as well as the opportunities they should provide are left out of the equation. The second argument is that CSIRO must be considered in the context of higher education and research, knowledge generation and distribution systems in Australia, not as a single player.

INNOVATION, SCIENCE AND PUBLIC POLICY

Innovation and the knowledge tree

Innovation as a concept has several aspects relevant when thinking about the role of the CSIRO in Australian innovation systems. The first is the difference made between product and process innovation and the second the well-known division between radical and incremental innovations. Using the accepted convention, innovation is defined here as a novelty of economic value, a change in what is successful in the market place, which distinguishes it from invention, a new product whose economic value has not yet been tested. If innovation involves a radical change in product, it often has associated process changes and radically new products become process innovations, as with IT and with biotechnology and nanotechnology when they come fully on stream. It is important for a country to develop and retain some proficiency in platform technologies of this kind, if only to assist in their ultimate diffusion.

Innovation studies approach their subject from many different perspectives. Schumpeter emphasised technological changes and industry transformation through the 'gales of creative destruction' (1934), others the importance of 'development blocks' (Dahmen 1988) as providing the scale and scope and competitive but supporting environment in which innovations, especially radical ones, could develop and transform an economy. Some focus on the significant role of 'clusters' as innovation beds that are customer-focused but include many companies that both compete and collaborate and the importance of lead users (see eg Porter 1990) while others have emphasised entrepreneurship and the creation and growth of new firms, especially those based to a large degree on scientific and technological discoveries. All these debates affect expectations about innovation roles played by science and technology.

At the organisational level, debates focus on whether small or large firms are more innovative, about the need for scientific and technological knowledge but also for possession of the 'complementary assets' that comprise, for instance, the firm's capabilities in marketing or information collection and management (Teece 1987). Questions are asked about whether investment in new processes is as important as searching for new products, about whether firms should focus on families of products of products and incremental changes or go all out for the radical technologies that enable not just firms but whole industries to regroup and renew themselves. And many more debates surround the degree to which science is used and at which stages of product and process development (Faulkner and Senker 1995; Rosenberg and Nelson 1994) and so on.

These debates were slow to gain currency in Australia, full recognition of the importance of innovation to competitiveness coming with a rush only in the late 1990s. In the 1980s especially, policymakers increasingly came to rely on 'economics' for policy choice and legitimation and were slow to recognise the existence of differing views within that discipline on important matters; the emerging evolutionary economics segment of economics which deals with innovation was hardly taught or written about in Australia until very recently. The circuitous nature of paths to innovation and industrial renewal was also ill-understood among scientists, even those with relatively close links to industry.

The one innovation model taken up in policy arenas was the linear model which depicts science as leading to innovation in the private sector along a one-way street, leading from public sector science, whether undertaken within the CSIRO, the universities or other public sector research (PSR) institutions, to industry which then did the final engineering and put the new products to market.

This is a very crude model and has been described even more crudely here to make the point that the reality is much more complex. Since the 1980s, it has been known to be inadequate, although occasionally accurate, and largely replaced by 'fifth generation' understandings that include feedback mechanisms, interactive processes, concurrent engineering and development, and recognize that science and technology play different roles in different industries, that in some, even high tech fields, science is used only at certain points to solve specific problems (see Dodgson, Gann & Salter 2005 for presentation of innovation models). There is now also evidence that the most important sources of innovation knowledge are customers, followed by suppliers, and that competitors may come before public sector research organisations (see the European Community Innovation Surveys), even in high tech industries such as IT (Hyland et al. 2006). Even biotechnology, perhaps seen by observers as the area most closely dependent on new science, has been shown to become rapidly more dependent on other factors for commercial success, such as partnering, marketing, capital availability and government regulations (Vitale 2003 quoted Marceau 2007). Science is always somewhere in the equation but not necessarily at the forefront.

We also know that the organisation of knowledge generation and related activities matters. Some countries are better at innovation not only because their science and technology production organisations are better aligned with each other and with industry 'needs' but because they are also better aligned with other aspects of public policy, notably skill development and labour market policies, IP regimes, tax policies and company internal organisation. Differences observed between countries in the extent and economic effectiveness of their innovation activities led observers to see that countries have 'innovation systems' of different kinds or levels--national, regional, local, technological, sectoral--and these must function well for innovation success.