GOVERNMENTS should not fund science. Everybody believes they should, but they
are mistaken. Government money for science only displaces private funding, and
private funding works better.
The first person to argue for government funding of science was the English
lawyer and politician Francis Bacon, in his Advancement of Learning,
published in 1605. Bacon argued that governments had to fund science because no
one else would. He also propounded the idea that, as economic growth depended on
applied science, and as applied science depended on pure science, the funding of
pure science by government was crucial to the creation of wealth. He invented,
in fact, the 鈥渓inear model鈥�:
Government funded pure science 鈫� Industrial technology 鈫� Economic growth
Advertisement
Bacon, of course, was educated at Cambridge and, I鈥檓 afraid, like so many
Cambridge-educated lawyers, deeply dishonest鈥攊n 1621 he was convicted of
corruption. While practically everybody still believes in his linear model, it
is actually as crooked as its creator.
Let us start by examining the last link in the chain: does technological
development really lead to economic growth? Here, the answer is clear. Economic
growth is technological development; the two terms are synonymous. In
the 1950s, the economist and Nobel laureate Robert Solow reported on the factors
that underpinned the doubling in gross output per hour of work that the US
enjoyed between 1909 and 1949. He found that seven-eighths of the economic
growth was attributable to 鈥渢echnical change in the broadest sense鈥�. Solow went
on to show that only one-eighth of the growth stemmed from increased capital
injection. None of the other standard economic parameters he studied鈥攕uch
as the growth rate of the workforce鈥攑layed any important part.
So, technology is crucial鈥攁nd according to the linear model it springs
out of basic science. But does it? One of the first scholars to answer this
question, in his book Wealth from Knowledge, published in 1972, was
Frederic Jevons. Like all the best people, Jevons was originally a Cambridge
biochemist, and he went on to become professor of science and technology policy
at Murdoch University in Western Australia. But during the 1960s, Jevons led a
team studying the origins of 84 technical innovations in Britain that had been
important enough to win a Queen鈥檚 Award for Industry. He found that 鈥渁lthough
scientific discoveries occasionally lead to new technology, this is rare鈥�.
Generally, he concluded, technological development takes place in the research
and development departments of industry. 鈥淭echnology builds on technology,鈥� he
said.
The economist Edwin Mansfield of the University of Pennsylvania extended
these findings. In 1991, he tried to quantify the impact of university science
on the development of industrial technology. For the period from 1975 to 1985,
Mansfield surveyed 76 major American firms which, collectively, accounted for
one-third of all sales in seven manufacturing industries: information
processing, electrical equipment, chemicals, instruments, drugs, metals and oil.
He discovered that 鈥渁bout 11 per cent of new products and about 9 per cent of
new processes could not have been developed, without substantial delay, in the
absence of recent academic research鈥�. Thus some 90 per cent of new technology
arises from the industrial development of pre-existing technology鈥攏ot from
academic science.
But the implications for university science are even grimmer, because
Mansfield also found that those 10 per cent of new products and processes tended
to be economically marginal, accounting for only about 3 per cent of sales. In
practice, then, around 97 per cent of commercially useful industrial
technological development is generated by in-house R&D. So Bacon鈥檚 linear
model needs a fork and lots of arrows:
Basic science 鈫愨啋 Industrial technology 鈫� Economic growth
Pre-existing technology 鈫� Industrial technology
It also needs a reverse arrow, because advances in basic science depend as
much on advances in technology as vice versa. Consider radio astronomy. What
could be purer? Yet the discovery that stars emit radio waves emerged out of
industry when Karl Jansky, an engineer at Bell Labs in the US, was investigating
the source of static in telephones. And what about high-temperature
superconductors? University departments to study these materials are popping up
all over the globe. But they were discovered not in academia but by IBM
engineers, Georg Bednorz and Karl M眉ller, in 1986.
Let us now ask another crucial question: does basic science need to be funded
by the government? People believe that industrialists will not fund pure science
because it can be easily 鈥渃aptured鈥� by competitors, who will get a free ride.
Further, most pure science is unlikely to be of immediate use. Because
industrial technology depends, at least in part, on pure research, the
government has to supply it as a public service.
That is the dogma. The practice is different. Companies fund pure science
very generously. In a 1980 study of 16 major American oil and chemicals
companies, Mansfield showed that they all invested in pure science. Indeed, the
more a firm invested in basic science, the greater its productivity became.
Support for this finding has come from Zvi Griliches, an economist at Harvard
University. In a study of 911 large American companies, published in 1986, he
showed that the firms that engaged in basic research consistently outperformed
those that neglected it. And the more basic research a company performed, the
greater its profits鈥攁nd vice versa.
True costs of research
It is worth looking at the quality of industrial science, too. In July 1994,
when Current Contents magazine listed the institutions that produced
the largest number of cited papers in biology, two of the top seven were private
companies, Genentech and Chiron. (Another was the totally private Howard Hughes
foundation, and three more were private foundations that accept government
money, the Salk and Whitehead Institutes and the Cold Spring Harbor Laboratory,
all in the US.)
Why do companies fund so much pure science, and of such quality? Well, it is
true that companies find it hard to capture the bulk of the benefits of their
own basic science, and that most of these benefits are exploited by competitors.
As an example, take the 1992 study by the Japanese economists Hiroyuki Odagiri
of Tsukuba University and Naoki Murakimi of Kushiro University. When they
examined Japan鈥檚 10 largest drugs companies, they found that each company
recovered an average annual return on its investment in R&D of only 19 per
cent. However, each company also obtained the equivalent of a 33 per cent return
on R&D done by the other nine companies. So each company was apparently
free-riding on the other nine.
But there is no such thing as a free ride in R&D. The great myth in
science funding is that published research, or other people鈥檚 R&D, is freely
available. It is not. Access to it is very expensive. Only highly skilled
scientists can capture other people鈥檚 research. It takes years of training
before a scientist can read research papers properly, and hours every week to
read all the new papers, assimilate them and integrate them into a research
strategy. That reading is crucial. Basic science is unpredictable, and no
company can hope to meet all its own needs. Indeed, the main function of basic
scientists employed by companies is not to do in-house research, but to read,
interpret and capture the global literature.
The best scientists, of course, only like to do research鈥攖hey do not
want to spend their careers in libraries. So to retain good scientists,
companies essentially bribe them with laboratories, money and the freedom to
publish. Naturally, companies will exploit any fundamental discoveries made by
in-house scientists, but those companies are more likely to capture useful basic
science when their researchers are reading in the library or chatting at
conferences. Yet there is no conflict, because the best scientists read the
literature assiduously, and love discussing it at conferences. Thus the free
market rewards companies that foster pure research, and vice versa.
The irrelevance of the government funding of science is illustrated by the
comparative statistics. People think that the funding of science is a political
decision, dependent on the whims of politicians. It is not. It is economically
determined, and politicians and industrialists are merely responding to their
economic circumstances. Consider the first Graph (
top left), which relates the
wealth per capita of members of the OECD with the quality of their scientific
papers, as judged by the number of times their scientific papers were cited. It
shows a very strong correlation. A plot of national wealth per capita against
the numbers of papers published per capita, or the numbers of patents published
per capita, would look very similar.FIG-20364301.jpg
This, to me, is a fundamental economic law: the quantity and quality of a
nation鈥檚 science and of its R&D is determined by its wealth per capita. This
is because, as Solow showed, a nation can only get richer by introducing new
technology. For an advanced country, this means pushing back technological
frontiers with its own research; a poor one can simply copy what is being done
elsewhere. These imperatives translate into research in the following way.
In advanced countries, companies spend increasing amounts of money on
research to develop new products. If a company is situated in a country where
taxes are low, like Japan or Switzerland, it simply invests its own money. If it
is in a country like France or Germany, with high taxes, then it lobbies hard
for its government to fund science. Either way, successful companies in rich
countries ensure their research needs are met.
In poor countries, companies borrow their technology, so they do not do their
own research, and do not lobby governments for it either. They are more
interested in basic needs such as roads and the education of their workforce.
Thus we see that the funding of science is economically determined, and that
governments are irrelevant. This is demonstrated by a little history.
Britain led the world through the agricultural and industrial revolutions
without government funding for science. And it still produced scientists like
Charles Darwin, Humphry Davy, Michael Faraday and many others. The British
government only started to fund science because of war: the predecessors of the
research councils and of the higher education funding councils were created
between 1913 and 1919 to help train the defence scientists of the future.
Spoils of war
A similar pattern emerged in the US, which overtook Britain around 1890 as
the richest country in the world鈥攁gain in the absence of government
funding for science. And once again, it was war that changed everything. The
National Academy of Sciences was created in 1863, at the height of the Civil
War, to help build ironclad ships to beat the South. The Office of Scientific
Research and Development, which ultimately spawned the National Science
Foundation and the National Institutes of Health, was set up in 1941.
Then, in 1957, the Soviet Union gave the West a dreadful shock when it
launched Sputnik, the first artificial satellite. In response, in 1958, the US
government created NASA, Congress passed the National Defense Education Act to
pour money into higher education and science, and the British government created
seven new universities.
What were the consequences of all this sudden government largesse? Nothing
very much. The second Graph (
left) shows that the American economy has grown
inexorably at around 2 per cent a year since 1820, and an almost identical graph
can be drawn for most Western economies. There are short-term fluctuations, as
politicians play with the currency, do irresponsible things with tax or, God
help us, go to war. But in the long term, economic growth is in the hands of the
market, scientists and technologists. It is steady and beneficent.FIG-20364302.jpg
The graph also shows what happened to long-term economic growth rates in the
US when, after 1941, the federal government cranked up its funding of basic
science. Neither the initial expansion of such funding nor the plateau that
followed in the 1960s made a difference to the economic growth rate. Why not?
Because all the federal government did was to displace private funding.
Let us illustrate this with some comparative statistics. Civil R&D
includes both applied science performed in industry and basic science. The third
Graph (
top right) shows that national expenditure on civil R&D depends on
national wealth. But the points on this graph are more scattered than they are
for pure science (
first Graph). Why is this?
One reason is that there is more 鈥渘oise鈥� for civil R&D. A nation鈥檚
scientific papers, the citations they attract or the number of patents published
are relatively easy to count. But civil R&D is performed by hundreds of
firms, so accurate data can be hard to collect. Further, not all economies are
the same. Some countries, for example, specialise in R&D-intensive
industries.FIG-20364301.jpg
But there is another source of noise 鈥攖he source of funding. In
countries like Australia and New Zealand, the government may pay for up to 80
per cent of national, civil R&D. Others, like Japan and Switzerland, believe
in laissez faire, and the private sector pays for up to 80 per cent of it. Other
countries fall somewhere in between.
Now in the third Graph, Australia and New Zealand fall below the linear
regression line, while Japan and Switzerland rise above it. A comprehensive
survey of all OECD countries shows a highly significant correlation between a
nation鈥檚 total budget for civil R&D and how much of that money comes from
private sources. Countries with governments that do not fund civil R&D end
up devoting a greater proportion of their national spending to it. Thus we see
that government funding of civil R&D damages it, because it displaces more
money than it supplies itself.
The same economic argument seems to apply to basic science, though it is more
difficult to substantiate by looking at OECD countries. This is because Japan is
the only OECD country to run a largely laissez faire system for basic science.
Less than half of Japan鈥檚 university science is funded by the state, compared
with more than 85 per cent in most OECD nations (yet, as the first graph shows,
Japan鈥檚 record in basic science is creditable).
Flourishing philanthropy
There is, of course another significant source of funding for science,
besides government and industry, and that is private donation. Historically, the
evidence is clear: philanthropy flourishes under laissez faire. In a free
market, capitalists make a lot of money and the extraordinary thing is that they
then give it all away. This year, for example, it was announced that David
Packard, cofounder of Hewlett-Packard, had left $4 billion to his
research foundation in his will. Packard is only the most recent in a long line
of generous donors. His predecessors include men such as Howard Hughes, William
Keck, John D. Rockefeller and Andrew Carnegie.
When Britain, too, enjoyed low taxes, its rich men also endowed science.
Witness Henry Wellcome鈥檚 vast trust, or the foundations set up by Samuel
Courtauld and Lord Nuffield. From these examples we see that when people are
endowed with economic power, they fund science. So do ordinary people. The
American Heart Association spent $105 million, much of it from small
donations, on research in 1991 and the American Cancer Society $94
million.
Where laissez faire survives today, so does the private support of pure
science. Consider Hong Kong, where tax rates are tiny. The island now needs
high-tech science, so the Royal Hong Kong Jockey Club has just donated
$220 million to create a new university modelled on MIT. Extrapolate that
$220 million for a population of 6 million to Great Britain鈥檚 56 million
people or the 260 million Americans, and one begins to see what the private
funding of science would look like under global laissez faire.
It is clear that industry will fund basic science for base reasons, and
philanthropists for nobler reasons鈥攕o long as they are left alone. If
governments intervene, to levy tax and use the proceeds to support science, they
not only displace private funders, they also fail to put back as much money as
they displace. Eventually, this effect will frustrate the development of
research and damage economic growth. Already, this trend can be detected. In
Economic Freedom of the World 1975-1995, published last year, the
economist James Gwartney of Florida State University shows that countries like
Switzerland and Japan, with the lowest taxes and the most laissez faire economic
policies, including those for science, are growing richer faster than more
interventionist states like Germany, France and Britain.
It can be difficult for central planners to abandon science to the vagaries
of the free market and philanthropists. But the empirical evidence is clear: in
research, as in so many other areas of life, the sum of innumerable personal
vagaries is healthier than the central planning of bureaucrats.