Quantifying success of science stimulus spend is challenging
June 5, 2009
The US economic stimulus package was intended to both shore up short-term employment figures and lay the groundwork for future economic growth through infrastructure and technology improvements. Science agencies were among the beneficiaries, with the National Institutes of Health getting $10.4 billion dollars, the National Science Foundation another $2 billion, and various other agencies like the Department of Energy picking up additional funding for research efforts. Obviously, the scientific community was pretty excited about the money but, since it was allocated, it’s beginning to realize that it will eventually be asked to account for how well the spending worked. As discussed in a Policy Forum that appears in today’s issue of Science, however, measuring the impact of scientific spending is anything but straightforward.
The article was written by Julia Lane, who is in the NSF’s Office of the Science of Science and Innovation Policy. No, that’s not a typo—the group focuses on trying to provide quantitative measures of the impact of changes in science policy. And Lane argues that the public will be expecting something of the sort, writing, "within 2 years, the public will want to be informed about the impact of the stimulus on the economic recovery."
On a certain level, providing these numbers are simple. Funding more science employs more people, and they require additional facilities, equipment, and consumables; they also spend money on things like lunch. It’s possible to make estimates of all of these and combine them and, indeed, the paper notes that someone already has: $20 billion in research investments are projected to create 400,000 US jobs next year.
But Lane implies that this is a very simplistic figure that nobody should be satisfied with, writing, "this approach functionally equates the impact of science to that of building a football stadium or an airport." Science and technology investments, much like infrastructure investments, are generally promoted as providing economic benefits that extend well beyond the direct impacts of the money spent. Science and technology leadership is often equated with the long-term economic health of nations and individual regions within them. So, Lane takes a look at that argument, and finds that the evidence supporting it is mixed.
On a national level, she notes that three-quarters of the efficiency gains the US has enjoyed in recent years have been traced back to technology improvements, which would seem to support the contention. But, in contrast, Japan spent heavily on science and technology during its period of economic stagnation, and hasn’t reaped obvious benefits from that. In the same vein, Sweden embarked on an R&D spending program in the 90s, but unemployment hasn’t budged. So, clearly, the economics of research spending aren’t simple or linear.
Instead, Lane makes the case that they’re episodic, unevenly distributed, and may take decades to bear fruition. So, for example, she cites a study of the development of high-tech industry in the San Diego area that shows that over 50,000 jobs in the biotech and electronics fields can be traced back to the efforts of only four faculty members at UCSD. The uneven distribution also applies to which fields have the largest impact, as the majority of efficiency gains in the general economy have been attributed to IT technology and processes.
The lag time can also confound the analysis of science’s wider benefits. So, for example, the biotech industry developed out of research accomplishments that date to the 1950s, and the Internet can be traced back to academic networks of the 70s. Anybody performing an analysis of that research spending a decade after the checks were cashed might have thought things looked promising, but they would be pretty unlikely to have guessed the sorts of economic benefits that would eventually come out of them.
So, is it pointless to think we’ll ever know how to direct at least some of our research dollars for economic purposes? Although we’re not at the point where we’d find it easy to identify which professors to fund in order to get effects similar to the ones San Diego has seen, Lane sees a few reasons for optimism. She notes studies that have looked at therapies based on monoclonal antibodies and RNAi, which are just now coming to market, even though the basic concepts were identified decades earlier. These have identified some of the problems that caused the delays (a lack of technology that scaled for monoclonal production, and commercialization ahead of a full understanding of the system for RNAi), which might help direct future research and translational dollars.
But, mostly, Lane seems to think that the stimulus may turn out to be our best way to generate the empirical data needed to better understand the relationship between scientific and economic development. Over the past several years, all federal agencies have made the transition to an all-electronic grant processing system, which should make tracking the flow of ideas and money a bit easier. In response to the opportunity, the NSF is funding a grant program to track the impact of the stimulus.
This may not get anyone—scientists, the public, or government—the answers they’ll be looking for a couple years down the line. But, as Lane points out, this will undoubtedly not be the last time we look to government spending to stimulate the economy, and it would be a shame to waste an opportunity to learn how to do a better job of it.
Science, 2009. DOI: 10.1126/science.1175335
This post has been written by John Timmer on June 4, 2009 9:29 PM couresy of arstechnica.com.