An Economic Gamble

For a recent example, I looked to the California Institute for Regenerative Medicine, now run by Robert Klein, chair of CIRM's Independent Citizens' Oversight Committee. Several years ago, Klein was facing a big task: Identify a figure that California voters would agree to pay for stem cell research. Rather than start from scratch, Klein looked to the NIH. In 2003, NIH was spending $220 million on stem cell research. Adjust for inflation, stretch it out over 10 years, and you get $3 billion, a figure that 59% of California voters approved.

NIH has some economic evidence to support what it does. In 2000, a report on the benefits of NIH research by the Joint Economic Commission of the US Congress found high economic returns from investing in research. In other words, besides the obvious health benefits of NIH funding for biomedical research, it also saves Americans money by lengthening their lifespan and improving healthcare. In their meta-analysis of a number of economic studies, the authors concluded that if even a minor fraction of the healthcare savings from healthier, longer-living people were due to medical research, the payoffs from that research would be many times the initial investment.

Klein used the findings of this report to make a good case that taxpayers would get something significant back from their investment, arguing that, overall, federally-funded biomedical research pays back to the economy to a considerable degree. This is another reason Klein chose to imitate NIH - what the institutes spend is working. To gain voters' approval, "it was important to have a responsible economic plan that had been shown on a portfolio basis to have some results,"

Klein says. The "Yes on 71" campaign (based on CIRM's proposal, called Proposition 71) helped deliver the message to voters.

Initially, at least, some say the gamble California citizens agreed to take will pay off. An economic impact report conducted by Bruce Deal at the Analysis Group and Laurence Baker at Stanford University found that Californians could expect returns of at least 120% to 236% on their investment in stem cell research over thirty years. At the high end, if Proposition 71 leads to "major advances in health care treatments," the authors say the state could get back more than seven times the cost of the initiative.

If we get so much back from biomedical research, why not invest $30 billion? $300 billion? Certainly there are limitations on how many dollars are available to invest, as well as competing investments such as Medicaid and education (on the state level), and the war in Iraq (on the national level). But if science's returns are so economically robust, why don't we put more into it?

Twenty-eight percent. This is the figure Edwin Mansfield, a now-deceased economics professor at the University of Pennsylvania, obtained after wrestling with an army of assumptions to pinpoint a likely return on research payoffs.¹ In 1991, Mansfield estimated that the rate of return on investing in academic research (across all disciplines) was 28%, meaning each dollar put into research would yield $1.28 in social and economic benefits within about a decade.

As part of the study, Mansfield estimated that 27% of drug industry products would not have been developed, except with significant delay, had academic research been eliminated from the pipeline. James Adams, an economist at Rensselaer Polytechnic Institute in Troy, NY, has been able to pick apart some of the relationship between academic research and industrial innovation. He surveyed the research and development laboratories of 200 companies to measure the amount they invest in learning about research at universities, such as attending conferences, hiring and meeting with consultants, and purchasing publications. On average, companies spent about six percent of their research and development budgets on learning efforts, and with each 10% rise in federal funding at universities, the learning budget at companies rose by more than one percent.²

"What happens next is we found the learning share [of R&D budgets] to be positively correlated with more patents," Adams says. So when the federal investment in university research increases, companies spend more money on learning about academic research, and companies produce more patents, an indicator of future economic impact. "There's no question that [research] is an incredibly large contributor to all advanced countries' economies," Adam says.

Other groups agree. Looking at a group of 16 developed countries, Dominique Guellec and Bruno van Pottelsberghe de la Potterie (former and current chief economists, respectively, at the European Patent Office) found that when publicly funded research at universities and government laboratories increased by one percent, countries experienced a 0.17% increase in productivity, measured as the ratio of industry's domestic product to labor and capital.³

Economists have also found that medical research can, not surprisingly, have an enormous impact on human health, especially longevity. Kevin Murphy and Robert Topel at the University of Chicago found that, from 1970 to 1990, the economy earned $1.5 trillion each year solely from reductions in heart disease death rates. "These values are truly enormous," the authors write. Though such changes could be due to improvements in public health or lifestyle, "if even a small fraction of this improvement is due to medical research, the economic return to that research could be substantial."⁴

Often the relationship between science and savings is hard to tease apart. A 2003 study by the Europe-headquartered Organization for Economic Cooperation and Development found that private R&D appears to have "high social returns," which includes economic benefits, but noted no clear-cut relationship between publicly funded research and economic growth.⁵

Joe Cortright, an economist and vice president of Impresa, an analysis company in Portland, Ore., has found that the benefits of federal academic funding on biotechnology vary by region (see "The biotech contrarian.") For example, Johns Hopkins University receives the most federal money, but is "not a particularly good performer in terms of commercialization," says Cortright. Likewise, Chicago, St. Louis, Houston, and Detroit are leading research centers with little to show in terms of bringing their work out of academia and into biotechnology companies.

Mansfield considered his estimate tentative and loaded with caveats, but still conservative. For example, he looked at the impact of academic research on only seven industries, and only as far out as 15 years. "Of course, the roughness of this figure should be emphasized," Mansfield wrote.

For every estimate of the returns on scientific investment, there are many reasons why that estimate could be wrong. Each economic study of the impact of science carries its own assumptions and other potential confounders. "These claims can always be demolished," says Terence Kealey at the University of Buckingham, often because they are loaded with assumptions that greatly affect the figure. For instance, Mans-field's 28% does not account for the cost of development, marketing, or the cost of building factories, some of which could lower his estimate of the return on science funding. "He's assumed in some magical way that scientists do their research and produce the facts and these instantly become products," Kealey notes.

"There are all kinds of methodology and measurement problems," agrees Iain Cockburn at Boston University. Congress' 2000 economic impact report on NIH, for example, assumes the agency is responsible for about 10% of health advances in the United States. "The question of what's the rate of return to NIH budget is fundamentally a very difficult one," Cockburn says.

Some estimated returns on investing in R & D

The bottom line: No one knows what the actual returns of science are. "I'm not aware of any metric that people can agree on, particularly for basic research. It's so hard to trace the scientific origins of a product that's actually making money," says John Marburger, director of the Office of Science and Technology Policy, which advises President Bush on science.

Part of what makes it difficult to estimate the return on scientific investment are the long lags between input and output, says Adams. "What does come out of the scientific community has to be filtered through applied research and development, and products have to be developed and marketed, and it's easy to see why it takes a long time for this to happen." Economists have found time lags to be anywhere between a few years and many decades. The mathematical innovations that influenced the development of aircraft occurred in the 19th century. Adams estimates that the peak effect on the growth of industrial productivity occurred between 20 and 30 years following the publication of new research findings.⁶

Moreover, when looking at the economic benefits of research, what should you measure? The impact of research leaks out of the laboratory and into the economy in a number of ways, says Ben Martin from the University of Sussex. There's the obvious linear scenario, in which a scientific discovery gets published in a journal, is picked up by a company, becomes an innovation, and earns that company money. But most scenarios aren't so straightforward.

Since every estimate is fraught with potential caveats and limitations, economics rarely drives budget decisions.

"There are no tools currently for assisting in making research investments for maximum return," according to an e-mail from a spokesperson for the US House Committee on Science and Technology. "All the studies staff is aware of are retrospective and vary considerably in their estimates of the amount of economic return from past research investments."

Instead, according to Martin, deciding "how big the scientific cake should be and how it should be cut up" is largely "a political process." Lobbying by science groups, competition with other countries' spending, input from agencies and experts, and an assumption that scientific investments positively impact society and economy are all part of budget decisions, says Martin.

Even though CIRM's Klein followed NIH's lead and considered the evidence that spending on science helps the economy when crafting CIRM's budget, economics plays no role in the agency's granting decisions, says Dale Carlson, spokesperson for CIRM. "We're focused strictly on the scientific merit of each proposal, its feasibility, and the potential contribution to the field from the results," Carlson wrote in an e-mail.

Likewise, Liz Allen, a senior policy advisor at the nongovernmental Wellcome Trust, which spends about £00 million on funding research each year, says economic return is not the driving factor in deciding grants. "If all [organizations] only fund things with a fairly clear and direct economic impact...that may, in turn, have a detrimental effect on the progression of science," because of the long and unpredictable route of basic science.

Mary Woolley, president of Research!America (where Eugene Garfield, founder of The Scientist, is a board member), says making funding decisions without considering economics has not hurt the research enterprise. "Historically investing in science just for accomplishing the next step in science has paid off in achieving better health outcomes and economic returns," Woolley says. For instance, basic research decades ago on the function of prostaglandin led to numerous therapeutic applications now on the market, Cockburn says.

"The whole notion of trying to funnel government money into areas you're likely to get the biggest economic payback is ill founded," says Greg Conko, a senior fellow at the Competitive Enterprise Institute, "because areas where you're likely to get the biggest economic payback, government funding is not needed."

Instead, Conko says, industry ought to be stepping in to invest the research. According to a 2003 report from the US Department of Commerce, companies spent $16.8 billion on biotechnology-related research and development in 2001.⁷ Conko says government's role in providing money is for basic science research that is so far removed from an identifiable technology or product, it is considered a public good without obvious use. "The challenge for government-funded research is to identify areas that are not and cannot be funded by the private sector," Conko says.

Marburger says that in recent years industry has been playing a larger role in research and development investments. The US gross domestic product (GDP) spent on R&D has remained stable at approximately 2.7% for more than a decade, but the proportion of public investment in research to GDP has decreased, meaning that the private sector has been picking up the slack. "I think it's possible to tap into private support to a greater extent than we have in the past and to find ways to encourage that," says Marburger. The president's American Competitiveness Initiative, for example, provides tax incentives for research and development.

Relying on the private sector, however, for greater research investments (and their subsequent economic rewards) isn't a safe bet, since biotechnology companies tend to have high failure rates in getting products to market, cautions Cortright. "It's like a lottery," he says. "What's the economic impact of the lottery?"

Marburger says that public and private funding agencies would benefit from reliable estimates of return rates, since evidence-based funding decisions will be that much better. Though reliable metrics are currently unavailable, measuring the economic impact of investments is "one of the first things to do when you look at whether you're doing it right," he says.

Pierre Azoulay, an economist at the Massachusetts Institute of Technology, says although it is pretty clear that returns from science are good, "from a political perspective we want to know if we are underinvesting or overinvesting in life sciences research. And the question rests on the rate of return on those investments." Martin says that the scientific community would benefit from having solid estimates of economic returns. "We know if we can get hard numbers, it will be that much easier to persuade politicians to spend more money," Martin says.

Governments are eager to get reliable estimates of where scientific dollars are going. The US Department of Commerce is assembling a committee to measure the economic impact of innovation in the private sector. E.R. Anderson, the deputy undersecretary for economic affairs at the Economics and Statistics Administration of the US Department of Commerce, explains the rationale: "We can't [provide incentives] if we can't measure it."

Despite the Wellcome Trust's position that economic considerations have no place in its funding decisions, the organization, along with the Medical Research Council and the UK Academy of Medical Sciences, has launched a new research program dedicated to providing quantitative estimates of the economic benefits of UK government funding for medical research. Allen says that it's important to consider economic returns from government-based funding organizations, "whose broader remit is to ensure the development of science bases and national economies."

In 1958 Zvi Griliches at the University of Chicago found a 700% rate of return on research investments into hybrid corn.⁸ "But that does not mean that we should spend any amount of money on anything called 'research,'" Griliches urged in his paper. "The moral is that, though very difficult, some sort of cost-and-returns calculation is possible and should be made." Marburger has also asked experts to develop better economic metrics. "The judgments about how much money to spend will be informed by that," Marburger says.

Diana Hicks, chair of the School of Public Policy at the Georgia Institute of Technology, cautions that measurements to pin down returns might work against research. The impacts of research are varied, she says, and to simplify, we will have to leave out important details, which may underestimate science's impact. "The question is, should agencies be managed only for short-term, easily measurable research impact? This probably would be bad for the long-term benefits that science brings society," Hicks says.

So where has that brought government's investment in science? In the United States, it's at about 2.7% of GDP. Marburger says the number is essentially arbitrary. "You could regard it as the collective expression of our society's judgment about the value of research in competition with everything else," Marburger wrote in an e-mail.

It is interesting that the arbitrary 2.7% of GDP has spread to global R&D goals, Marburger says. The European Commission has made it a goal of member countries to reach three percent of GDP in R&D spending "based largely, if not exclusively, on the fact that this is roughly what the US spends, despite the fact that there is no conscious policy within the US that leads to this figure."

In a recent study, Kenneth Manton at Duke University and his colleagues found that achieving an optimal rate of return (which they pegged at stock market returns of 6.3%) would require the United States to quadruple its investment in research to 12.8% of GDP.⁹ The model "clearly says there needs to be significant increases for a period of time to get near an optimal funding level," says Manton.

Concern in the United Kingdom, says Martin, is that other countries are feeding research more than they have in the past. The United Kingdom spends about 1.9% of its GDP on research and development, which is less than that spent by nine European countries, the United States, Japan, and Israel. "Unless we do something about that, we'll get behind in the technology race, and companies will go where they're spending more on research," Martin says.