Skip Navigation

The Congress has approved the initial funding of several large nondefense research and development projects that are expected to cost $23 billion in the period 1992 through 1996. The large size of these projects has caused some concern as to whether they may be funded at the expense of other R&D activities in the years ahead. At the request of the Senate Committee on the Budget, the Congressional Budget Office (CBO) undertook to examine the funding of large nondefense R&D projects in the past in an effort to determine whether such "megaprojects" tend to crowd out federal spending on smaller R&D programs and on other government-sponsored activities. In keeping with CBO's mandate to provide nonpartisan analysis, the paper makes no recommendations.

David Moore and Philip Webre of CBO's Natural Resources and Commerce Division wrote the paper under the supervision of Elliot Schwartz. The budget staffs of the Department of Energy and the National Science Foundation provided valuable assistance in compiling R&D project budgetary histories and forecasts. Thomas Lutton of CBO provided computational assistance. The authors wish to thank William Boesman, Daryl Chubin, Peter Fontaine, Leon Lederman, Frances Lussier, and other reviewers for their valuable suggestions.

Francis Pierce edited the manuscript. Kathryn Quattrone assisted with the figures. Donna Wood typed the many drafts and prepared the report for publication.
 

Robert D. Reischauer
Director
July 1991
 
 

SUMMARY

• Background
• Trends in Spending for Large R&D Projects
• Budgetary Options

I INTRODUCTION

• Civilian Science and Technology in the Federal Budget
• Large and Small R&D Projects in the Budget
• Approaches to Measuring the Budgetary Implications of Large R&D Projects

II MEASURING THE BUDGETARY EFFECTS OF LARGE R&D PROJECTS

• Measures of Spending for Large R&D Projects in the Budget
• The Spending Aggregates

III COMPARING PAST AND PROJECTED SPENDING ON LARGE CIVILIAN R&D PROJECTS

• The Trend in the 1980s
• Projections for the 1990s
• Large Civilian R&D Projects and Other Spending

IV BUDGETARY OPTIONS

• Large R&D Projects and Productivity
• Budgetary Options

APPENDIX Measures of "Big Science" Spending Aggregates
 

TABLES
1.	Major Components of Federal Funding for Civilian R&D In 1991
2.	R&D In Selected Agencies for 1991
3.	Measures of Spending for Large Civilian R&D Projects
4.	National Science Foundation Projects Included in the CBO Inventory
5.	Department of Energy Projects Included in the CBO Inventory
6.	National Aeronautics and Space Administration Projects Included in the CBO Inventory
7.	The Three Largest Projects, 1980-1986
8.	Large Civilian R&D Project Share of Agency Budgets
9.	Inflation-Adjusted Spending on Large Civilian R&D Projects
10.	Federal R&D-related Spending
A-1.	Measures of Spending for Large Civilian R&D Projects by Agency
A-2.	R&D Spending Aggregates
FIGURES
S-1.	Spending on Large Research and Development Projects as a Percentage of Spending on All Nondefense Research and Development, Fiscal Years 1980-1996
S-2.	Alternative Projections of Spending for General Science, Space, and Technology Minus the President's Request for the Three Largest Projects, 1990-1996
1.	Expenditures on Nondefense Research and Development, 1961-1991
2.	Spending on Large Research and Development Projects (Inventory Measure) as a Percentage of Budget Functions for General Science, Space and Technology, and Energy and of Nondefense Research and Development, 1980-1996
3.	Spending on the Largest Three Research and Development Projects as a Percentage of Budget Functions for General Science, Space and Technology, and Energy and of Nondefense Research and Development, 1980-1996
4.	Effects of the Space Shuttle on the Inventory Measure of Large Research and Development Spending, 1981-1991
5.	Spending on Research and Development Plant by Three Agencies, 1980-1991
6.	Spending of Fields of Research Dominated by Large Instruments and Facilities as a Percentage of All Nondefense Research and Development, 1980-1991
7.	President's Request for Spending on Large Nondefense Research and Development Projects as a Percentage of Domestic Discretionary Budget Authority, 1990-1996
8.	President's Request for Spending on Large Nondefense Research and Development Projects as a Percentage of All Nondefense Research and Development Spending, 1990-1996
9.	Alternative Projections of Spending for General Science, Space, and Technology Minus the President's Request for the Three Largest Projects, 1990-1996
10.	Spending in Budget Functions 250 and 270, 1981-1996
BOX
1.	The Human Genome Project

 
 

The Administration's 1992 budget request calls for increases in federal support for nondefense research and development (R&D), and for the agencies and budget areas that fund these activities. These expenditures include funding for the National Aeronautic and Space Administration's space station and 'its Earth Observation System, and the Department of Energy's Superconducting Super Collider. Concern has arisen that these and other large R&D projects will be funded at the expense of smaller-scale activities, including many projects initiated by single investigators.

This study examines the budgetary history of large nondefense R&D projects during the 1980s, as a background to the Administration's proposal for the 1990s. Three specific questions are addressed:

• Are large nondefense science and technology projects currently consuming more budgetary resources, in either absolute or relative terms, than during the 1980s?
  
  
• Under the Administration's program, will large nondefense science and technology programs increase their share of science and other spending aggregates relative to current levels and to the trend during the 1980s?
  
  
• Are large nondefense science and technology projects funded at the expense of all other R&D spending (including "little science") or other types of spending?

BACKGROUND

Advances at the frontiers of science and technology have required ever more complex and expensive facilities, instruments, and experiments. The proliferation of these projects in the federal budget has raised a number of issues. Large R&D projects are expensive. Outlays on the space station and the Earth Observation System (EOS) could run to $35 billion and $17 billion, respectively, before the year 2000. The Superconducting Super Collider could require an investment of between $8 billion and $12 billion by the same year. Cost, of course, is not the only standard by which to judge an investment. Supporters of these large R&D projects believe they will deliver benefits that justify the expense, although not all share their confidence.

Large R&D projects are also risky. Their costs of development and operation are difficult to estimate, and it is also difficult to be certain of the capabilities of a system once developed, and the importance of its mission. Failure in a large R&D project may cause serious setbacks to scientific or technical projects that depended on its success. The very riskiness of large R&D projects, however, is an argument for federal support, since the government is the only institution able to bear the cost and the risk of these ventures.

Finally, some critics of large R&D projects raise a political concern: such projects, particularly in their development and construction phases, provide income for localities and businesses. Also, large projects increase the budgets of the federal agencies that sponsor them. These factors are not particularly important so long as they do not intrude heavily on the scientific and technical criteria used to make decisions about the mix of large and small projects. But if cost overruns or budget constraints require choosing between large and small projects, it would be against the national interest to support large projects at the expense of small ones for nonscientific and nontechnical reasons.
 

TRENDS IN SPENDING FOR LARGE R&D PROJECTS

After examining a variety of budgetary measures, CBO finds that the Administration's plan would increase the share of funding devoted to large nondefense R&D projects to levels not seen since the early 1980s. As Summary Figure 1 shows, the share of total nondefense R&D funding devoted to large nondefense R&D projects peaked sharply in the early 1980s, then fell in 1984 when spending for the development of the space shuttle ended. The share of nondefense R&D spending accounted for by an inventory of 80 large R&D projects and facilities rose from around 10 percent in the mid 1980s to over 15 percent by 1991 (the lower panel in Summary Figure 1). If the Administration's program was enacted, the share of large R&D projects would rise even more during the first half of the 1990s to 22 percent by 1996. The three largest projects in the inventory alone would double their share of nondefense R&D spending from the current level of 8 percent to 15 percent under the Administration's plan. The three largest projects would also increase their share of all domestic discretionary spending from 1.1 percent in 1990 to 2.8 percent by 1996.

The Administration's proposal calls for increases in overall R&D spending large enough to maintain the shares going to both "big science" and "little science." This approach would avoid the situation of the early 1980s when total spending in R&D-related budget functions remained flat or declined while large projects consumed a greater portion of the total.

If the cost of large projects increases, or overall funding for general science and space turns out to be less than requested, the Congress could confront a choice between funding large R&D projects and other science and technology spending. Summary Figure 2 shows what could happen to spending on other science and technology if the very largest projects were funded as the Administration proposes, but the Congress placed constraints on overall space and science spending. If the largest three projects were funded as proposed and budget function 250 was permitted only to keep up with inflation-CBO's baseline projection-funds available to support other science and technology spending would be 25 percent below the Administration's plan. If spending for general science, space and technology (function 250) was frozen at the 1991 level through 1996, and the largest projects permitted to proceed as planned, other science spending would be reduced by 45 percent relative to the levels the Administration proposed. The National Science Foundation, a prominent source of funding for small R&D, would certainly be among the agencies considered for reductions from the Administration's plan under either scenario.
 

BUDGETARY OPTIONS

Current knowledge about the relative benefits of large projects as compared with smaller ones is not sufficient to provide much guidance as to how funds ought to be allocated between these types of activities. To keep better informed, the Congress may wish to initiate a biennial "cross-cutting review" of science and technology spending as suggested recently by the Office of Technology Assessment. The review would compare such spending by all agencies with broad national objectives and specific technological goals to see whether the distribution of current spending, including that between large and small R&D efforts, best meets these objectives and goals. Such a review could be useful even though it would partly duplicate the annual budget review several Congressional committees already conduct.

The Congress could also try to balance spending for large projects and spending for small projects by using multiyear appropriations and arbitrary annual spending caps. Multiyear appropriations might be effective in controlling the total cost of big projects by allowing agencies to proceed on an optimal schedule without tailoring their programs to fit annual budgetary requirements. But multiyear appropriations will not be effective in controlling cost if technical uncertainties lead to cost overruns. In defense programs, multiyear appropriations have been more successful in procurement projects than in developing technology.

Arbitrary annual caps could be placed on spending for large projects, set at levels that would assure adequate funds for other science spending. Caps offer the advantage of being in current use and easily understood, but they would raise the total costs of big projects.

Canceling one or more of the largest projects could offer immediate and sustained budgetary savings. Canceling the space station program, for example, could free up between $2.0 billion and $2.6 billion each year for other space and science projects, if Congress chose to appropriate funds for these purposes rather than other federal priorities.

Finally, the Congress may wish to pursue partnerships with other countries in large science and technology projects on a more equal basis than at present. Foreign partners would have more say in managing such projects in exchange for carrying a larger share of the costs. U. S. contractors would have to give up some of the procurement business, however. Multiyear appropriations could help to make more equal international partnerships effective. Thus, a secondary cost of these ventures could be the loss of Congressional oversight and funding flexibility granted by annual appropriations.

This document is available in its entirety in PDF.