Taken as a whole, the industrial competitiveness activities at the national laboratories are unfocused and lack a firm policy foundation. There is uncertainty within the laboratories about how large and broad-ranging these activities ought to be, about how to fund them, and about how they should relate to the other main areas of activity that the laboratories are engaged in - in particular, whether industrial competitiveness should be viewed as a primary or a derivative function. The uncertainty within the laboratories partly reflects an inconstancy in DOE policy and Congressional intent regarding these issues. One of the options that the Department has considered, elevating industrial competitiveness to a primary objective or mission, risks distorting existing programs and diverting resources towards activities that are unrelated to either DOE's competencies or its other missions. Such activities are unlikely to produce results that will benefit either the agency's industrial partners or the public in the long run. Recent indications are that DOE headquarters staff have recognized the need for greater focus in this area and are moving to develop criteria toward that end. We welcome these developments and encourage early discussions of the evolving situation with laboratory leadership.
Another issue requiring clarification concerns the appropriate division of labor among the national laboratories, industrial research laboratories, and the research universities regarding industrial R&D. The notion that the national laboratories should serve as a `bridge' between the research universities and industry in this arena - a notion presented to the Task Force - is not in general a useful way to think about this division of labor. It does not reflect what is currently happening, it is not consistent with the fact of the research universities' superior record relative to the laboratories as agents of technology transfer, it is difficult to reconcile with any plausible view of the future evolution of these institutions, and it appears to be based on an outdated picture of industrial innovation as a linear process that originates in basic research laboratories.
A more useful picture is of the laboratories serving as nodes in a national network of research and development institutions, with knowledge flowing in both directions along the links between the laboratory, university, and firm nodes. This network is simultaneously pursuing fundamental knowledge for its own sake, innovation for the purposes of private wealth creation, and public missions (national defense, public health, environmental quality, etc.), while at the same time educating and training the next generation of scientists and engineers. Because it is a true network, each type of R&D institution is involved in some way in all of the network's activities, but there is specialization among them. While the national laboratories are one of the primary nodes for public missions like national security, their role in wealth-creating innovation is necessarily secondary (and probably even tertiary in sectors far removed from the DOE's mission areas). The question is whether, how, and to what extent they can add value to the primary role of industry in this arena.
The Case For Industrially-Relevant R&D At The Laboratories
A Critical Distinction
However, what seems to have been lacking at both the DOE and laboratory levels
is a clear understanding of the distinction between the application of
laboratory-developed technologies to industries where such application is an
essential part of the primary mission areas of DOE (e.g., the development of
energy saving technology for industry use, or the development of methods for
analyzing and predicting the behavior of nuclear waste repositories, or the
development of advanced techniques for oil field simulation) and, on the other
hand, the application of these technologies to industries which have not had a
relationship with DOE and whose activities lie well outside the mission domains
of the agency and its laboratories. Without a clear understanding of this
important distinction, the risk is that DOE will allocate public funds and the
technical and human resources embodied in the laboratories in unfruitful
To clarify and sharpen this distinction, we have found it useful to categorize the range of possible laboratory activities in industrially-relevant R&D according to their position in the simple matrix shown in Figure 2. In practice the boundaries between the classes of activity shown in Figure 2 are imprecise, but the basic distinctions are important nonetheless. The four categories are as follows:
Figure2 Expected DOE National Laboratory Contributions to Industrial Research and Development
Expansion Outside DOE Mission Areas
We are concerned about the implications of expanding the laboratories'
industrial R&D activities outside the existing DOE mission areas. If the
DOE is to get into programs whose primary mission is to meet a need of private
industry (as contrasted with programs where private industry needs are
complementary to government needs, but not dominant), private industry will
need to have a large say in allocation decisions and in evaluation, since only
industry has the intimate knowledge of the marketplace that is critical to the
success of such efforts. The difficulty comes in reconciling this need for
industry direction with the parallel need for public influence over the
disposition of public resources at government laboratories.
We are also concerned that the expansion of the laboratories' roles in serving the technology needs of private enterprise will create additional managerial problems within DOE. For any organization to be effective, the activities it manages need to be associated with a coherent set of objectives. Otherwise, it is virtually impossible to allocate resources rationally, or to evaluate the various activities and programs in terms of how they contribute to the performance of the organization as a whole. This is amply borne out by experience in private enterprise which indicates that most conglomerates do badly, especially in managing technological innovation. Under statute, the DOE is faced with the considerable challenge of managing an already diverse set of missions. Adding to this complex task the requirement to consider the technology needs of the private sector in areas not related to the Department's traditional domain of activity is likely to distract DOE from its public missions and lessen its impact while undermining the effective pursuit of those industrial objectives. We are concerned that `porkbarrel' criteria for program funding might increasingly replace more rational resource allocation, and that the laboratories might be more likely to propose industrial programs merely based on `make work' criteria.
We are further concerned about the possibility that DOE and its laboratories, in engaging in industrial R&D, may find themselves competing with private firms in providing technical services or new technological developments. In such a situation, the laboratories' access to public funds would give them an inappropriate advantage. Such situations are more likely to arise the less DOE and laboratory management know about the commercial application in question. We are already aware of some instances in which such competition appears to have occurred, and we are concerned that the problem may become more serious in the future.
All of these problems can be ameliorated if the industrially-related R&D done by the laboratories is focused on industries and applications which are themselves instrumental to achieving the DOE's public missions. That is not to suggest that laboratory directors should be denied the flexibility to initiate new technical projects at the periphery of current activities. Making room for individual and small group initiatives of this kind is an important way to keep the laboratories lively and exciting places to work. If such a project becomes large enough, though, a judgment must be made as to its fit with the rest of the laboratory's activities. This will depend not only on technical compatibility but also on the project's relevance to the laboratory's (and hence the DOE's) missions.
The current industrial partnership activities of the laboratories have brought them into contact with a number of industries that do not lie within the Department of Energy's traditional domain. In addition to the drawbacks specified above, these new involvements also suffer from a number of other problems:
Technology Partnership Mechanisms
Cooperative Research and Development Agreements (CRADAs) currently occupy pride
of place among the array of mechanisms employed by DOE to encourage
laboratory-industry cooperation in technology development and transfer.
Introduced to the laboratories in 1989, the CRADA mechanism reflects the
intent of Congress, developed and refined in legislation over the preceding
decade, to ensure a greater degree of laboratory involvement in industrially
relevant activities. Industry criticism of the CRADA mechanism has focused on
the slow and uncertain nature of the negotiation process. The DOE and the
laboratories have responded to these concerns. Nevertheless, there remain wide
variations across the laboratory system in the speed with which CRADA
negotiations are being concluded. Ironically, recent contractual reforms
intended to encourage more businesslike relations between DOE and its
management and operations (M&O) contractors at the laboratories risk
complicating laboratory/industry negotiations because the contractors are more
strongly motivated than before to secure rights and to take direct profits from
laboratory-generated intellectual property. The Task Force believes that the
contract-driven M&O motivation to profit from laboratory intellectual
property will act as a significant barrier to industry acquisition and
subsequent commercialization of laboratory technology in the future.
Other criticisms focus on the lack of independent review of the operation and outcomes of CRADAs, and also the process used to select industrial partners for these agreements. Regarding the latter, when a CRADA proposal is submitted there is typically no solicitation of competing proposals, nor an announcement that public resources are available for such work and will be awarded exclusively to the applicant company if others do not step forward. Nor is the peer review process as rigorous as other DOE programs; independent expert evaluations of the validity of the proposed work and its relevance to the DOE mission are not routinely solicited. These practices can be traced to earlier legislative and executive branch attempts to reorient the laboratories towards industrially relevant activities as rapidly as possible. However, their practical effect today is to leave the laboratories vulnerable to charges that the selection process is flawed and that the competitive playing field is being unfairly tilted towards the laboratories' chosen partners. In fact, DOE routinely and successfully implements several different models of independent peer review, at least two of which are relevant to the present situation -- the allocation of DOE beam line resources, and the allocation of funds under the Small Business Innovation Research (SBIR) program. The DOE should consider applying these practices to CRADAs as well.
One of the most common metrics used by DOE and others to assess the technology
transfer performance of the laboratories is the rate at which new CRADAs are
being signed. An obvious limitation of this metric is that it measures inputs
rather than outcomes, and the latter are likely to vary greatly from one CRADA
to another. It also fails to distinguish among different classes of CRADAs.
Some pertain to work that the laboratories would have undertaken anyway, even
if no CRADA had been signed. Others entail a significant departure from a
previous program of work. In some cases technology transfer activities have
reportedly drawn the laboratories away from their primary missions. In others,
resources appear to have been diverted away from fundamental research to
support them. In such cases, a cost-benefit metric may be the only meaningful
measure of performance. We also note that none of the laboratories appears to
be measuring the effectiveness with which technology is being transferred in
from industry or the universities to support their missions. In general, we
recommend that greater emphasis be placed on outcomes in the measurement of
technology transfer performance. The question of performance metrics for the
laboratories in discussed further in Section VII of this report.
To Table of Contents