Incentives and Structures
There are several scales of interaction between national laboratories and universities, from the investigator scale, to new models around joint institutes, to user facilities, and at the highest level, laboratory management. Each of these interactions is based on different incentives and faces structural barriers to collaboration that must be considered. In addition to the scale of interaction, the differences between the expectations, missions, and institutional structures of the laboratories and universities must be understood and then addressed to enhance the ability to engage in more effective collaborative activity, particularly at the institutional scale. Strategic partnering at multiple levels across institutions provides a formula for success. This session of the workshop focused on understanding the issues relative to different incentives and institutional structures that can provide barriers to collaboration between the laboratories and the universities and on identifying some potential paths to solution.
While the incentives for collaboration were seen by many participants as differing across scales, most participants voiced the importance of these collaborations for meeting their mission and research goals. At the individual investigator level, collaborations are developed to support key proposals or to meet specific capabilities needed for research programs. According to Gerald Stokes, director of the Joint Global Change Research Institute, some DOE programs encourage the participation of university researchers as a competitive factor in some solicitations. At the very highest levels, institution to institution, strategic interaction with universities in targeted areas is viewed by some laboratories as critical to their goals for innovation. Alton Romig, of Sandia National Laboratories, spoke to the
strategic focus at Sandia relative to building collaborations with universities, with an emphasis on select university programs that support laboratory priorities, such as the Nanoscale Science and Engineering Center or the research area of microelectronics. This view was echoed by Jeffrey Wadsworth from Oak Ridge National Laboratory, who described a set of partnership activities between ORNL and targeted universities that is helping to build collaborative research programs and new joint institutes and to strengthen their staff capabilities and recruiting at the laboratory.
The importance of relationships at higher levels between the institutions was reinforced by presenters and some participants. At Oak Ridge, for example, Wadsworth pointed to the key role that Georgia Tech played in ORNL’s successful proposal for a nanoscale science research center. Stokes described the model for joint institutes, an emerging concept for collaboration, as an option for developing a strategic collaboration between a university and a laboratory at the institutional level. Yet at all levels, significant issues or challenges were identified that create disincentives for collaboration. These fall into three general areas:
-
Institutional structures, including contractual mechanisms;
-
Financing models and resource availability; and
-
Research models and expectations.
Key points raised by participants in the workshop in each of these areas are described below.
INSTITUTIONAL STRUCTURES
Institutional structures are optimized for each type of institution and there are differences among them. These differences present challenges for collaborative activity between institutions. The way in which each institution handles contracting, for example, is driven in part by its overall mission, its work culture, and the regulations under which it must operate. Contractual mechanisms were identified by many participants as extremely problematic. The breakout session identified at least three layers of regulations for the laboratories: (1) those stemming from DOE (which in itself is historically a composite of several agencies), (2) those stemming from practices in DOE field offices, and (3) those mandated by the contractor institutions running the laboratories. On the university side, the accounting and contract or grant language is dictated largely by federal Office of Management and Budget (OMB) circulars.
As a result of their different historical and sectoral origins, contractual language requirements of national labs and universities were often simply incompatible. For example, a laboratory might require an indem-
nification clause in all of its contractual agreements, yet state-run universities are expressly prohibited from signing an indemnification clause. Different accounting and audit requirements between the two types of institutions can also drive bizarre behaviors. For example, the apparent “cost of a person” difference between universities and national labs can motivate labs to offload people onto universities, while providing the same nominal salary. These same differences in accounting procedures can make it extremely difficult for contractor-run laboratories to make joint faculty appointments with universities.
Many participants in the breakout session echoed the challenges in the current system, where project funding must generally be obtained for collaborations either through a subcontracting mechanism or by both institutions submitting proposals to different funding agencies. The problem with the first model is that it does not represent the equal partnership that many collaborators would prefer to see. The “work for hire” framework used by most laboratories to secure research collaborations was mentioned as being fundamentally antithetical to collaboration between equals. The problem with the second model, two separate proposals, is that one party may succeed in obtaining funding through the proposal process, and the other may not, resulting in no collaborative program. Several participants also raised the point that there is some inconsistency within DOE in the way collaborations are funded—each part of DOE works a bit differently. The models being used by NIH were raised as good examples to look at for thinking about university-laboratory collaborations. The NIH glue grant was raised as a specific example.1 Several university representatives stated that NSF does not have a good model for these collaborative arrangements, and given the importance of NSF to university researchers, having such an option at least within a portion of the NSF research budget would be helpful. Most participants voiced the view that a separate source of funding specifically targeted at collaborations would help significantly to address this broader issue.
The problem in sharing human and financial resources between laboratories and universities was attributed to the differences between DOE procurement regulations and the OMB circulars under which universities operate. Research administrators in attendance suggested that an OMB-
1 |
The purpose of the NIH Glue Grant Initiative is to make resources available for currently funded scientists to form research teams to tackle complex problems that are of central importance to biomedical science and to the mission of the National Institute of General Medical Sciences (NIGMS), but that are beyond the means of any one research group. A high level of resources may be requested to allow participating investigators to form a consortium to address the research problem in a comprehensive and highly integrated fashion. For more information on glue grants, please visit http://www.nigms.nih.gov/funding/collab.html. |
blessed universal agreement for university-laboratory collaborations or a reworking of the original OMB circulars would be a potential solution.2
The differences in contracting processes and priorities, time to move paperwork through systems, and accounting challenges make collaboration difficult. The bottom line is that collaborative programs require extra effort to address these structural differences between the organizations. Participants suggested that we need either the resources required to run this gauntlet or the opportunity through specific funding sources at the national level to compete for programs that would require and support a collaborative university-laboratory team (see discussion on finances below).
Another institutional challenge is presented by joint appointments. Particularly for the joint institute construct, the ability to provide joint appointments for researchers at both the university and the laboratory is important to achieving the desired goals of the collaboration. Several participants voiced the opinion that joint appointments are extremely beneficial to collaborative research, but they can be extremely difficult to achieve. At a nonuniversity-managed laboratory, real joint appointments are much easier to establish since the individual actually does have a single employer. At nonuniversity-managed laboratories however, joint appointments are much more difficult, if not impossible, to achieve. The problem seems to lie in the two different human resource systems and overhead systems that can operate in universities and national laboratories. It really seems necessary to have high-level formal commitments with universities, not only for joint hiring, but also to enable more collaborative programs and joint institutes, for example. These joint research centers were mentioned several times as a worthwhile method of establishing joint collaborative programs.
Finally, the issue of foreign nationals working on DOE-funded projects or at DOE sites was raised by both university and laboratory participants. This issue was also addressed in several of the other sessions. A significant proportion of the student body in science and engineering today consists of foreign students. Universities have difficulty accepting contracts or research opportunities that restrict foreign students from participation. Yet DOE laboratory funding increasingly carries some restrictions on participation, requiring U.S. citizenship in many cases to either work on the project and/or to visit the government facility for meetings, and so forth. Some universities simply will not take on projects that may restrict student participation. Although both laboratory and university participants discussed how they have addressed this issue in the past, it is clear that a widespread understanding of these options was not held by all partici-
pants and that it would be useful to develop a set of recommended procedures for facilitating broader student participation in DOE (and now DHS, as well) research programs.
FINANCING MODELS AND RESOURCE AVAILABILITY
Participants raised a number of issues relative to finances and the availability of resources for supporting collaborative activity, some of which were viewed as providing major barriers to collaboration. A major issue is the difference between the two institutions in overhead rates. Laboratories must fully recover their costs on a project, and their staff has relatively high overhead rates compared to those at a university. For groups wanting to develop a joint research project, this has led to some bizarre workarounds, such as putting some of the research team on the other institution’s payroll to reduce the nominal overhead rates.
Another point raised is that there is no source of research funds specifically designed to support joint activities. LDRD (laboratory-directed research and development—a type of overhead-generated funding with fewer use restrictions than programmatic funding) money was used to sponsor many collaborations with universities, but that pot is shrinking greatly. The current situation, in which small amounts of money are drawn from many different types of sources within the laboratory, leads to erratic support of collaborative programs and graduate students, which is particularly damaging to the latter. In addition, the different sources of money usually have different criteria for use and rules of accounting, which can hamper their intended purpose.3 For example, an issue raised by Alvin Kwiram at the breakout session is that at some laboratories, seminar money cannot be used to support foreign attendees or speakers, yet many university researchers and U.S. scientists are not U.S. citizens. Some laboratories additionally lack a category of funding that would allow them to be members of university consortia.
In terms of external, programmatic funding, breakout session participants expressed the view that the situation was not much better. Each institution and/or research team ends up competing for the same pot of money, and one of these groups must be in a “lead” position. This lack of joint money leads to distortion of the peer relationship between lab and university researchers. Not only does this situation create the need for subcontracting agreements between institutions (leading right back to the contractual difficulties and different rate structures), but it also sets up a
“rich-uncle, poor-nephew” phenomenon, where the lead organization controls the funding, scope, and direction of the research and the “supporting” organization is in a position simply of support. The Accelerated Strategic Computing Initiative (ASCI) program seems to be a good example of a program in which the funding is shared between laboratories and universities. Participants were intrigued by the possibility of extending this model to other programmatic sources of funding.
RESEARCH MODELS AND EXPECTATIONS
The primary mission of a university is the creation and dissemination of knowledge; the primary mission of a national laboratory is providing results that support DOE’s research program. These differences in focus result in differences in the way in which the accompanying research is conducted. Universities have the ability to bring a significant number of graduate and undergraduate students to a project. However this can lead to different time scales for the expectation of project results, as well as different working models for conducting research. A doctoral student may need 4-5 years on a project to complete both research and educational activities. Master’s students may need a clearly defined 1-year project to complete the thesis requirement. Laboratory projects may have more of a 1-2 year time frame for completion, with continuous progress expected and reported on to DOE. Additionally, the large, complex nature of many laboratory projects requires them to be team driven, whereas most universities reward individual (single PI) achievements.