Making sense of the national labs
Limited by Design: R&D Laboratories in the U.S. National Innovation System, by Michael Crow and Barry Bozeman. New York: Columbia University Press, 1998, 321 pp.
In a previous professional life as a systematic biologist, my taxonomy professors drilled into me the concept that taxonomy is not an end in itself. To be meaningful, systems of classification have to elucidate deeper relationships that might otherwise go unnoticed, they have to be in some sense predictive for as-yet-undiscovered relatives of the things being classified, and they have to represent some “natural” ordering of those things rather than being a convenient catch-all dreamed up by an armchair systematist.
Michael Crow and Barry Bozeman’s book recalls those taxonomy lessons, because Limited by Design is above all else a taxonomy–a reclassification of the nation’s 16,000-odd R&D laboratories. Indeed, one of the glowing dust-jacket blurbs refers to the authors as the Linnaeus of U.S. R&D labs, and in fact Crow and Bozeman have brought a new critical eye to how the national innovation system is organized and how it is evolving. The book brings this great bolus of data together in one fairly hefty gulp, permeated with the minutiae of the nation’s R&D labs. At one point, the authors even admit to their delight in “how easy it was to dazzle our colleagues with such arcane information as the approximate number of R&D laboratories.” But there are many points at which the dazzle wears thin and a number of times when it actually begins to detract from the book’s major purpose. Readers may find themselves more impatient than impressed by the time Crow and Bozeman actually begin drawing inferences from the data.
The authors begin by trying to define what exactly is a national laboratory. Clearly, the 700 or so federal laboratories, which account for more than a third of federal R&D spending, should be included. But then the issue of classification becomes tricky: What about the hundreds of university labs competing for federal funds and churning out Ph.D.s? Or the 14,000 or more industrial labs, some of which dwarf even the most ambitious federal facilities? It’s no wonder that when the authors asked a cross-section of stakeholders how many national labs there were, the answers ranged from 500 to 20,000.
Crow and Bozeman define a national laboratory as a facility focusing on science or engineering and employing at least 25 personnel. Using this standard, the sheer numbers are overwhelming–between 16,000 and 17,000. The authors argue that for the past 50 years federal R&D policy has essentially ignored the diversity among these institutions, crafting policies that focus almost exclusively on the federal labs. And even there, it’s been mostly a one-size-fits-all prescription.
To illustrate the undesirable effect of ignoring diversity, Bozeman and Crow focus on the Department of Energy labs. “These labs are not moving forward according to an assessment of their strengths and weaknesses or based on a plan for building national strength,” they warn. “They are moving forward in political competition with each other. There is no meaningful plan allocating responsibilities.”
The underlying diversity has been uncovered incrementally over the past 15 years through the National Comparative Research and Development Project (NCRDP), with which both Bozeman and Crow have been intimately involved almost since its inception. The NCRDP has been a massive undertaking to try to understand scientific innovation in the United States and other countries; it’s spawned hundreds of research papers and treatises. If some of the data and conclusions in Limited by Design seem stale or twice-told, the likelihood is that you read one of the 60 or more papers of their own that the authors cite in the book’s bibliography.
What taxonomy tells us
The primary original contribution of the book is what the authors call their “environmental context” taxonomy. Rather than simply categorizing laboratories according to their major source of funding–university, federal, industrial–Bozeman and Crow factor in the economic character of a lab’s output along with its “publicness.” The latter is a measure of the extent to which government, through funding decisions or other means, influences an institution’s research agenda. Publicness also sounds a little like current measures of a lab’s audience, but in Limited by Design the concept is modified to reflect the extent to which the lab is oriented toward proprietary output, public domain output, or some mix of both. The confluence of economic character and publicness yields a matrix of nine cells, each of which contains a lab type and mission uniquely different from the other eight and, the authors suggest, requiring separate policy considerations.
When the authors test this matrix on actual labs, most of the characterizations are not very surprising: More than 90 percent of all the government labs, for example, cluster as “pure public science” or “public science and technology,” which means moderate to high market influence and high government influence. Likewise, 70 percent of the university labs cluster in cells denoted “hybrid science” or public science, which implies low market influence and moderate to high government influence. In fact, one of the principal conclusions one can draw from this analysis is that, by and large, these apples don’t fall very far from their stereotypical trees. That an overwhelming majority of government labs fall out in very similar categories should lend some confidence to overarching policies for their management. And that more than two-thirds of university labs also fall in related categories argues likewise. If anything, these data suggest that the evolution of the government laboratories from government missions over the past 50 years has been rather glacial and that university laboratories, despite all the current angst about the change of culture and attitudes in university settings and the commercialization of the academic enterprise, don’t seem to have changed enough to radically alter their classification.
Unfortunately, however, the policy recommendations in Limited by Design are few, and most are applicable only to the federal labs. Congress might take note of some of the principal recommendations, such as the suggestion to resurrect a labs-closing commission to weed out dead wood. As for the basic research laboratories, Bozeman and Crow argue that they should stick to basic research and not engage in the commercialization so popular with contemporary lab management. Intermediate-range applied research labs, such as Argonne National Laboratory, the authors believe, should be provided with discretionary funds to further the development of technology to the point of market impact.
There are some very annoying aspects to Limited by Design. The foreword, for example, gives an entirely different set of facts and figures for Los Alamos National Laboratory than does chapter 1, just three pages removed. More troubling, however, is that fully a quarter of the book isn’t written by the authors themselves but is given over to promotional pitches written by dozens of the labs surveyed. From Los Alamos we hear, “Along with a compelling mission, we focus on continuing to do great science in the service of the nation.” The Applied Physics Laboratory at Johns Hopkins University reports that it “has maintain[ed] an umatched breadth of capabilities for more than 50 years as a dedicated vital Navy resource and as a university laboratory providing research, development, and engineering support to the Department of Defense to apply effective technical solutions to complex national defense problems.” Such effusions do not evoke the reader’s confidence.
Overall, however, Bozeman and Crow do yeoman’s duty by reclassifying the national laboratories. Still, this is not the book that will spur policymakers to action. After all, since 1978 some 20 blue-ribbon task forces, review boards, and special commissions have looked at the management problems of the federal labs without having a significant effect on the operation of the laboratory system. We are still waiting for the book or report that will be compelling enough to move policymakers to act on the raft of sensible recommendations for reform that have come from respected observers of the national labs.