Henry Ford, that proud no-nonsense engineer, once famously said that history was bunk [ie an absurdity]. He would have been right if he had said that some history is bunk. The popular and elite understanding of the history of science and technology is a case in point: too much of that understanding gains its authority from repetition rather than evidence. These histories are important in the policy imagination, endorsing particular approaches to science and technology: because of this, evidence-based history can change minds not just about the past but also about what one might do for the future. In highlighting the lack of originality of so much thinking about science and technology, history can have a powerful salutary effect.

Take the Lisbon Strategy, which called for the European Union to move towards spending 3% of GDP on R&D by 2010. Of course, the Commission's policy is directed at increasing R&D rather than decreasing GDP, and reflects a view that Europe is deficient in R&D. For the historian, that view and that target create a distinct sense of déjà vu.

I recently appeared in a couple of BBC radio programmes on the history of science advice to the British government. The producer used wonderful recordings of the key advisors, and two bear directly on the relationship between R&D levels and economic performance. The first was a recording of Sir Henry Tizard, who became the UK's most senior scientific adviser in both the civil and military fields from 1945. The second was of Sir Solly Zuckerman, who was the UK's chief scientific adviser in the 1960s. Both confronted the problem of the relationship of R&D to economic growth. In Tizard's time, the country faced the problems of economic reconstruction after the war. In Zuckerman's time, there was great concern that the British economy was not growing as quickly as that of other European countries. Many put this down to a lack of R&D.

From the usual histories of science policy one would have expected Tizard and Zuckerman to demonstrate a belief in that 1980s invention, the 'linear model of innovation'. This is the view, allegedly dominant since the second world war, that academic research, or organised R&D, is the main source of new technology, and thus central to economic growth. One would have expected the advisors to call for more R&D spending, if only as representatives of the scientific community. But they did no such thing. Sir Henry Tizard, speaking at the 1948 meeting of the British Association for the Advancement of Science, said that Britain needed to exploit its existing knowledge rather than produce more. As evidence he pointed to international comparisons. He claimed that Britain had become a 'scientific nation' only in the inter-war years, but had done very well scientifically since. However Sweden and Switzerland, which he saw as not being in the same scientific league as the UK at the time, had higher levels of productivity. They had better technology, and better production, but inferior science. What a sobering message to the assembled scientists, and to the public listening to his radio broadcast!

In Tizard's time there were no systematic R&D spending scoreboards, though he would have had good data for Britain. By the time Zuckerman was chief scientific advisor, internationally comparable R&D data were available from the OECD. The data showed that in the mid-1960s two countries led the OECD in terms of R&D spend as a proportion of GDP. They were the United States, as might be expected, and Great Britain. The level they had reached should give advocates of the Lisbon Strategy pause: it was that very same 3% of GDP.

For Zuckermann it was clear, for he had comparative economic data too, that the ratio of R&D spending to GDP did not correlate with national rates of economic growth, at least not positively. The British government had been advised by Bruce Williams, a pioneering economist of innovation, that there was no such positive correlation. This key observation, which still holds today, is not appreciated widely enough - it is drowned out by claims that imply the opposite. The countries that spent a lot on R&D as a proportion of GDP grew slowly, while the fast-growing were modest or low spenders. Zuckermann called for restraint in R&D, not its expansion.

That there was and is no positive correlation should, however, be the starting point for any analysis of national or European R&D policies. It is simply another way of saying that R&D done locally is a much less important source of new technology than all the other national and global sources of new technology. We should not expect a simple positive relation even if technology were the only determinant of growth, which it isn't.

That is just the beginning of the problem. Why should R&D spending sum to a total representing total innovative activity? After all, isn't much of R&D done in competition, thus cancelling out a lot of the overall effort? And why should this already dubious total be divided by GDP? Does a higher ratio give more innovation than a low one? There are answers to these questions, but they have nothing to do with any sensible arguments for setting levels of R&D.

The tragedy is that we have known all this since the early days of science and technology policy. High ratios of R&D to GDP characterise the past as well as the present; the arguments for increasing R&D to GDP ratios are leftovers of naïve techno-nationalism, not the work of informed futurology. But rich critiques of these arguments about R&D are there to be found in the historical record too. For those of us who want to see a novel future, these richer arguments need to be the basis of policy not, as too often happens, the bad ideas that were being criticised.

Professor David Edgerton
Hans Rausing Professor, Centre for the History of Science, Technology and Medicine
Imperial College
d.edgerton@imperial.ac.uk