A Fair Deal for Federal Research at Universities

The government has gone too far in its efforts to shift the costs of federally funded academic research to the universities. This transfer has reduced researcher productivity, led to inadequate research management, and has almost certainly prevented access to research universities by qualified students who happen to be poor. In spite of these negative side effects, there are some who advocate additional constraints on reimbursement of facilities costs, constraints that would make universities reluctant to do research in critically important fields where the facilities are expensive.

Instead of shifting research costs to the universities, government officials should be sitting down with university leaders to develop an indirect cost reimbursement system that is fair and creates the most productive environment for research. The Federal Demonstration Partnership (FDP) has already begun this work, but the government needs to expand the scope of this effort and begin implementing its findings. The system for indirect cost reimbursement must be equitable if the research universities are to meet the nation’s expectations that they will continue to lead the world in basic and long-term applied research, respond to new national needs associated with homeland security, and fulfill their fundamental responsibility of educating our nation’s most capable young people.

The distinction between direct and indirect research costs is codified in the Office of Management and Budget (OMB) Circular A-21. Direct costs are those associated with a specific research project. They include salaries for scientists and wages for project team members, as well as materials and supplies, travel, project-specific equipment, and subcontracts to other organizations. Indirect costs, referred to as facilities and administrative (F&A) costs, are those that cannot be associated with a specific project. Facilities costs include the operation, maintenance, and depreciation of buildings used for the research, research equipment for which the university has paid, interest on debt associated with buildings placed into service after 1982, and library expenses. Administrative costs include the central office of the university president, financial management, departmental administration (including clerical assistance associated with the research), and activities related to meeting environmental, safety, and health standards. In 1991, reimbursement of administrative costs was capped by OMB at 26 percent of total direct costs. At the same time, OMB declined to permit faculty to classify secretarial and clerical salaries as direct costs, even those directly related to the research.

In the United States, indirect costs are reimbursed because the structure for performing basic and long-term applied research differs markedly from that in most other industrialized countries. In those countries, a large fraction of the basic and long-term applied research is performed in national laboratories and universities that are funded directly by government. It pays not only direct research costs, but also for construction, operation, renovation, and maintenance of buildings and laboratories, as well as for the research projects’ administrators and clerical support. That is, government pays up front for virtually all the research costs, both direct and indirect.

In the United States, research universities are either private or funded by the 50 states. The universities put up their own money for construction, operation, and maintenance of almost all buildings in which the research is performed. In addition, they take primary responsibility for faculty salaries and administrative services. They are reimbursed for the portion of the facilities costs associated with federally funded research, but they get this money only after they have invested in the facilities and hired staff. Also, they receive money to defray the costs of facilities and research administration only when their faculty members are successful in competing for research funds. Thus, universities take on a significant financial risk when they decide to pursue an area of research.

The United States’ unique system accounts, in large measure, for this country’s leadership in almost every research field, as well as for the exceptional quality of U.S. graduate education. The combination of funding from the states and from private sources has built a research infrastructure and a graduate education system unmatched anywhere in the world. It is unlikely that the federal government would have paid for the development of such a system. Indeed, as will become clearer, the federal government has not even been providing its negotiated share of the indirect costs of research. In spite of that, there are frequent calls for further cuts in indirect cost reimbursement.

For example, the National Science Foundation’s (NSF’s) Authorization Act of 1998 requested the White House Office of Science and Technology Policy (OSTP) to analyze indirect costs, including “options to reduce or control the rate of growth of the Federal indirect cost reimbursement rates, including options such as benchmarking of facilities and equipment cost, elimination of cost studies, mandated percentage reductions in the Federal indirect cost reimbursement…” OMB had already proposed benchmarking of facilities in the Federal Register in September 1997. In that proposal, indirect costs associated with buildings that cost less per square foot than a mandated amount would be reimbursed readily, whereas significant justification would be required for those that cost more.

OMB’s intention was to decrease the cost of constructing university research buildings and, consequently, indirect cost reimbursement rates. But in 1998, OMB recognized that benchmarking is likely to lead to undesirable changes in the types of research performed in universities. Buildings that cost less than average tend to be ones that provide electrical outlets for computers and not much else. Research in areas such as semiconductor processing or biomedical research–areas that require extensive environmental, safety, and health safeguards because of the toxic materials under study–require buildings that cost significantly more than the average. A likely consequence of benchmarking would be to build fewer of the costlier types of buildings because reimbursement would be uncertain. Given the importance of research and student training in fields such as semiconductor processing and biotoxins, these buildings are vital. Reimbursement policies should not discourage that construction.

Finding Facts

OSTP engaged RAND’s Science and Technology Policy Institute to perform a preliminary study in response to Congress’s 1998 request. The findings in RAND’s mid-2000 report were striking. In sum:

• “What evidence we have indicates that the underlying cost structures in universities have lower F&A costs than federal laboratories and industrial research laboratories. Because of specific limitations on university F&A reimbursement, such as the administrative cap, the actual amount awarded to universities for F&A costs is likely to be even lower than what cost structure comparisons would indicate.”
• University F&A rates in total have remained stable for at least a decade, but those for administration declined and those for facilities increased. The decline in administration rates was likely due to OMB’s 26 percent cap.
• RAND estimated “that universities are providing between $0.7 and $1.5 billion in facilities and administration costs that would be eligible for reimbursement based on their negotiated F&A rates. We estimate that universities are recovering between 70 and 90 percent of the facilities and administrative expenses associated with federal projects.”

In short, not only was there no evidence that the government was being overcharged for indirect costs, the evidence suggested the opposite: that the universities were being prevented from recovering a significant percentage of what they had spent.

In its July 2000, response to Congress, which built on the RAND report, OSTP found no acceptable method for reducing indirect cost rates, except for cooperative analysis by the government and universities to reduce unnecessary indirect costs caused by less-than-thoughtful policies and procedures. The FDP, which has representatives from the government and the universities, is performing this analysis at present. It is vital that OMB and OSTP participate actively in this endeavor and ensure that the partnership’s recommendations are implemented.

Since the RAND report was written, much has happened to influence considerations of F&A cost reimbursements. The need for university investment in research buildings has increased dramatically. Fiscal year (FY) 2002 appropriations were consistent with the plan for doubling the budget of the National Institutes of Health (NIH) in five years, by 2003. It seems likely that the FY 2003 appropriations will complete that process and initiate corresponding increases, perhaps as much as a doubling, for research in the physical sciences and engineering. These increases will necessarily increase the number of academic researchers, which will, in turn, dramatically increase the need for research space. Yet, early in the NIH doubling process, a 1998 NSF report stated: “Sixty-five percent of institutions with existing or needed research space in the biological sciences and 52 percent of institutions with existing or needed research space in the medical sciences reported that the amount of biomedical research space they had was inadequate to meet their research commitments . . . In order to meet their current research commitments, the biomedical institutions reported that they needed an additional 9.0 million net assignable square feet (NASF) of research space in the biological sciences or 23 percent more than they currently have. At the same time, they reported that they needed an additional 7.1 million NASF of research space in the medical sciences or 21 percent more than they currently have.”

It is unlikely that universities have acquired even the additional 16.1 million NASF required to meet their 1998 commitments, let alone that implied by the subsequent NIH budget doubling. At a very conservative $150 per NASF, the cost of the space needed in 1998 would be $2.4 billion. Since most of that money is not being supplied by the federal government, it is vital to adopt federal policies and practices that encourage, rather than discourage, universities to construct or renovate research space. Among the most important steps would be clear statements by the administration that it has no intention of introducing arbitrary limitations or caps on reimbursement for facilities and that it will discourage Congress from mandating them. The government must recognize the enormous task the universities face in providing research space.

Even with such assurances, the universities will have a hard time providing the space, because the economic downturn and the stock market decline have led to decreased endowment yields, decreased state tax income, and decreased contributions from donors. The assurances will, however, encourage some universities to borrow construction money, taking advantage of low interest rates and the potential for reimbursement of the interest via indirect cost recovery.

Since I returned to Stanford’s faculty in 2001, the impact of both the 26 percent cap on administrative reimbursement rates and the prohibition on direct charging of the salaries of administrative and clerical staff supporting faculty in their research has become painfully evident to me. Because most research universities are spending more than 26 percent on administrative endeavors related to federally sponsored research, the net result of the cap and the prohibition has been a significant reduction in clerical support throughout these institutions. Faculty at research universities spend a considerable amount of time on clerical and administrative tasks directly linked to our research. This is an inappropriate use of faculty time. These tasks do not require our special expertise; they could be performed much more effectively by clerical staff with special skills. OMB should eliminate the A-21 restriction against direct charging of clerical and administrative support that is linked to specific research projects.

Similarly, it is difficult to accept the voluntary nature of environment, safety and health (ES&H) responsibilities resulting from the 26 percent cap. For example, Institutional Review Boards (IRBs) are supposed to be protecting human subjects in federally sponsored biomedical research. If these IRBs are to perform their work properly, they need large amounts of attentive faculty and clerical time. In limiting administrative cost reimbursement to 26 percent, which is below existing costs for most research universities, the federal government has walked away from cost of clerical and faculty time. We appear to be relying on faculty volunteerism and university largesse to carry out the extremely important work of protecting research subjects. NIH has attempted to fix this problem with a temporary allowance of a direct expenditure. IRB costs, however, are just one example of a significant, mandated, indirect ES&H expenditure that should be documented and included in an administrative cost reimbursement structure–one that is not capped arbitrarily.

Shifting Costs

Perhaps the most troubling conclusion of the RAND report is this one: “universities are providing between $0.7 and $1.5 billion in facilities and administration costs that would be eligible for reimbursement based on their negotiated F&A rates.” Recent statistics from NSF indicate that the proportion of academic R&D expenditures deriving from the funds of the institutions themselves has risen from 14 percent in 1980 to 20 percent in 2000. During the same time, the fraction provided by the federal government has declined from 67 percent to 58 percent. Thus there is no question that federally mandated cost sharing and limits on indirect cost recovery have shifted academic R&D costs from the federal government to the universities.

The total amounts are significant and have ramifications far beyond the research enterprise itself. In 2000, the academic institutions’ expenditures for R&D were $5.6 billion. Of these, approximately one-half of the institutional funds, or about $2.8 billion, were used to pay for unrecovered indirect costs and cost sharing. Most of these funds were provided by Carnegie Institution Type Research I and II universities, which in 1997 had a combined enrollment of 2.7 million students. Thus, the average year 2000 expenditures for unrecovered indirect costs and cost sharing at these institutions were on the order of $1,000 per student. That’s about one-quarter of typical tuition at selective public research universities.

Over the same 1980-2000 time period in which the costs of federally funded research have been shifted to the universities, these universities have become less accessible to those who might use that education to break out of the lowest income stratum. The recent report of the National Center For Public Policy and Higher Education, Losing Ground, states: “Increases in tuition have made colleges and universities less affordable for most American families.” In particular, the share of family income required to pay tuition at public four-year institutions has increased from about 13 percent in 1980 to about 25 percent in 2000 for families in the lowest income quintile. The report also stated that accessibility is decreased still further because “Federal and state financial aid to students has not kept pace with increases in tuition.” The $1,000 per student of unrecovered indirect costs and cost-sharing amounts to more than one-half these average tuition increases.

Some university administrators contend that these expenditures do not play a role in tuition increases. Yet it seems likely that cost sharing has a direct effect. This money frequently comes from discretionary university funds that could be used in many different ways, among them a halt to tuition increases. Given the importance of accessibility to public research universities, all federal agencies that fund academic research should adopt the NSF Cost-Sharing Policy of 1999 that limits severely the circumstances in which cost sharing is expected.

Making the case that the failure to recover indirect costs leads to tuition increases is more difficult. In many states, indirect cost reimbursement goes directly to the state treasury, not to the university. Thus, one could argue that these funds would not be available to the universities to prevent tuition increases. There is evidence to the contrary, however. Losing Ground noted that: “The steepest increases in public college tuition have been imposed during times of greatest economic hardship,” a disquieting conclusion. For example, average public tuitions grew more than 12 percent in 1991, when there was a marked economic downturn. In the same year, however, outlays by academic institutions for R&D rose by 8 percent–with public institutions presumably providing the major portion. Thus, at the same time that states were using large tuition increases to offset their income problems, R&D expenditures increased significantly. It is hard to avoid the hypothesis that tuition increases could have been smaller had the public institutions been able to recover a greater fraction of their indirect costs and provide fewer cost-sharing dollars.

The decreasing availability of four-year colleges to talented and accomplished students who lack only money is a fundamental societal problem. Part of the American dream is the knowledge that hard-working and capable people can improve their economic status through a college education. That mobility helps maintain our society. There is good reason to believe that this mobility is threatened by indirect cost reimbursement and cost-sharing policies and practices. The federal government should change its view of them. Rather than seeking to shift R&D costs to the universities, it should fully fund the cost of the academic research that it sponsors. Public research universities should join their private counterparts in seeking full indirect cost recovery.

In summary, there appears to be little merit in federal policymakers’ efforts to drive down indirect cost recovery and mandate cost-sharing. Instead, OMB, OSTP, and university representatives should ensure that a revised A-21 leads to full indirect cost recovery, including allowing researchers to charge secretarial and clerical salaries that are directly related to their research. The NSF Cost Sharing Policy of 1999 should be adopted by all research-sponsoring agencies. These actions would enhance the universities’ abilities to construct and renovate research space for both basic and long-term applied research, as well as to provide the student training the nation needs. Researchers could then concentrate on research and education instead of clerical work. It would enhance the institutions’ abilities to meet their fiscal, property control, environment, and health and safety responsibilities, as well as safeguarding human subjects. It is likely at least to slow the rate of tuition increases and perhaps could even lead to tuition decreases and easier access to college.

These actions should be accompanied by continued analysis of both federal and university policies and practices that drive up indirect costs, with a goal of achieving more cost-effective approaches. Government and university officials should, therefore, ensure the continued effective functioning of the FDP.