The commission operated under the auspices of the U.S. Department of the Treasury. Assisting the commission were staff members of the Office of Management and Budget and the Department of the Treasury, as well as individuals from organizations with which some of the commissioners are affiliated. The commission is grateful for all this support, and wants to acknowledge the following people who made contributions to the overall effort and to this report:

     The budget of any organization, private or public, is a statement of both the resources to be made available to the organization and the priorities of those who manage it. The budget that the President submits to the Congress, which in fiscal year 1999 covered expenditures of nearly $2 trillion, tells the American people how the administration proposes to spend their taxes and, until recently, the proceeds of federal debt issued to finance the shortfall between total expenditures and revenues. The budget is thus inherently a political document, but in the best sense of the term. This is because it reflects the collective judgment of the individuals in a democracy about how much public funds are to be raised and how they are to be used.

     This commission has devoted its attention to one particular kind of expenditure in the federal budget: spending on "capital." Although this term has been defined in various ways for different purposes, a common element among all of the definitions is that capital spending--whether undertaken by the private or public sector--is intended to generate benefits over the long run.

     In this report, we have concentrated on capital spending by the federal government because it is our charge. But we cannot emphasize too strongly that capital spending at all levels of government, as well as by the private sector, provides important benefits to the nation as a whole in significant part because those benefits are delivered over the long run. It is easy in the day-to-day battles over budget policy to forget that such spending helps determine the kind of society that we and our children will live in--not just this year but many years from now as well. We therefore encourage this president and future presidents to help educate American citizens about the importance of devoting current resources toward future needs--in the form of spending on capital by both the private and public sectors.

     Most firms in the private sector, as well as many state and local governments, recognize the importance of capital expenditures by making decisions about them separately from decisions about how much to spend on annual operating expenses. By contrast, the federal government has never done this.

     This commission has been directed to examine whether this practice ought to be changed--that is, whether the federal government should adopt a "capital budget"--and, if not, what other steps, if any, should be taken to improve the federal decision-making process as it relates to spending on capital or "investment" expenditures.

     Capital budgeting is a process that takes explicit account of capital spending levels. In this report, we primarily examine versions of a capital budget in which either: (1) the size of the deficit or surplus is made to depend, in part or in whole, on the amount of expenditures defined as "capital," or (2) a single decision is made about how much to spend on "capital," under some definition. A variation of the first definition is what we label the "simplistic" version of the capital budget, one in which capital spending may be financed, in part or in total, by borrowing. We treat the second definition as the equivalent of imposing a separate "cap" on expenditures defined to be capital, or in the alternative, a process whereby the depreciation of capital is explicitly taken into account in the budget process. We briefly note in a concluding section that there are other, perhaps less formal, variations of a capital budget that we do not extensively analyze here.

     The commission had its origins during the Congressional debate about whether to amend the Constitution to require the federal government to have a balanced budget every year. Nothing in this report should be construed as support for the balanced budget amendment considered by the Senate in 1996. ^(a) Nor does the commission endorse the adoption of the simplistic version of the capital budget. Furthermore, a majority of the members of the commission does not support, at this time, adopting a budget procedure that would impose a separate cap on capital spending.^(b) The reasons for reaching these conclusions are spelled out in the body of the report.

     At the same time, we have concluded from our study of existing practices and after gathering evidence from a wide range of experts, that the existing federal budget process--as it affects decision-making about capital expenditures as well as other types of spending--has significant weaknesses. Insufficient attention is paid to the long-run consequences of budget decisions. Capital spending in particular is inefficiently allocated among projects. Moreover, the current process shortchanges the maintenance of existing assets. ^(c)

     Accordingly, the commission urges the Congress and the executive branch to undertake a thorough examination of how the budget process may be improved beyond addressing capital-related needs. Toward this end, it may be productive for both branches to create a new Commission on Budget Concepts to aid them with this task. ^(d)

     In the meantime, we believe there are a series of constructive responses to the shortcomings we have identified, though they do not include adopting any particular form of a capital budget as we have just defined the term. These responses are aimed at improving each of the component parts of the budget process: setting priorities currently and for the long run, making budget decisions in the current year, reporting on those decisions, and subsequently evaluating them in order to make improvements in future years. Key to achieving these improvements is ensuring that the appropriate information is made available to decision-makers and the public throughout the process so that policy makers (1) are properly informed when deciding how to spend taxpayers' money and (2) can be held accountable by the public for those decisions.

     The recommendations we summarize below take account of two important features of federal budgeting.

     First, many government efforts have objectives, such as the management of foreign affairs or the defense of the nation, that cannot be readily measured in monetary terms. In stark contrast, it is relatively easy to keep score in the private sector, where firms are often judged by a single metric, such as current profitability, return on equity, or the dollar value of their shareholders' equity.

     Second, borrowing is subject to less discipline at the federal level than it is at lower levels of government. States and localities cannot "print money" to cover the debts they issue, whereas one arm of the federal government--the Federal Reserve--has the ability to "monetize" debt issued by the Treasury. A related difference is that federal debt is viewed by the marketplace as practically free of default risk, whereas states and localities have a strong interest in maintaining high credit ratings, which constrains borrowing at the state and local level.^(e)

     These considerations necessarily imply that federal budgeting rules should not simply replicate rules that may be used in the private sector or at the state and local levels of government. But at the same time, because the existing federal budget process has the weaknesses we have noted, certain improvements are appropriate. We have concentrated on suggestions for the executive branch; however, as will become evident below, certain of these require the cooperation of and concurrence by the Congress.

     We also recognize the essential role of the American people as monitors, advocates, and parties whose interests ultimately are at stake during the budget process. For this reason it is important to increase the transparency of that process--not only to enhance the quality of inputs to the Congress from the private sector and other levels of government, but also to increase the federal government's accountability to the American people.

     To facilitate the setting of priorities among all programs, not just those involving capital expenditures, the commission recommends:

     Recommendation 1: Five-Year Strategic Plans.--Although federal agencies are now required (under the Government Performance and Results Act) to prepare strategic plans every three years and performance plans annually, this process should be improved in several respects:

• The strategic plans should (1) be prepared annually, (2) be integrated with the annual performance plans and the agencies' five-year budget projections that are now submitted to the Office of Management and Budget (OMB), and (3) be included as an integral part of the budget justifications sent to the Congress.



• The strategic plans of the agencies and their annual budgets should be tied to the life-cycles of their capital assets.



• OMB should standardize the formats of these plans, in consultation with GAO and CBO, to make them more useful to policy makers.



• OMB should expand its efforts to evaluate the plans and facilitate the Administration's use of them for government-wide planning.



• Congress should take such plans into account in deciding on annual agency appropriations. It should also consider how it might improve its own procedures so that it can pay more attention both to the longer-run implications of its current year decisions and to issues with longer-run consequences. In undertaking this task, Congress might find it useful to take advantage of the wide range of institutional expertise available to it, including resources within the Congressional Budget Office, the General Accounting Office, and the Congressional Research Service.

     Recommendation 2: Benefit-Cost Assessments.--There should be an ongoing effort within the federal government to analyze the benefits and costs of all major government programs (whether or not related to capital spending), so that they can be adjusted, refashioned, or eliminated, as appropriate. OMB, the agencies, and the Congress (through GAO and CBO in particular) should be given the resources to carry out this important function.

     To improve the process by which annual budget decisions are made, the commission recommends:

     Recommendation 3: Capital Acquisition Funds.--To promote better planning and budgeting of capital expenditures for federally owned facilities, Congress and the executive branch should experiment by adopting for one or more agencies separate appropriations for "capital acquisition funds" (CAFs). Budget authority would be lodged in the CAFs for federally owned capital assets. The CAFs would "rent out" their facilities to the various programs within each agency, charging them the equivalent of debt service.

• CAFs would help ensure that individual programs are assessed the cost of using capital assets.



• By spreading capital costs across entire agencies, CAFs would help smooth out the lumpiness in appropriations sometimes associated with large capital projects.



• If the CAF experiment proves successful, the CAF approach should be adopted throughout the government.

     Recommendation 4: Full Funding for Capital Projects.--All capital projects, or usable segments thereof, should be fully funded before the work begins. In this way, Congress can fully evaluate their likely costs and benefits before appropriating funds for them.

     Recommendation 5: Adhering to the Scoring Rules for Leasing.--Existing rules that govern the scoring of leases should be strictly followed by both agencies and the Congress. This will discourage the signing of short-term leases when it is cheaper over the long run to construct or purchase a facility.^(f)

     Recommendation 6: Trust Fund Reforms.--Although trust funds for highways, airports, and other uses insulate certain types of spending from the balancing process that is inherent in the rest of the budget, they can be useful if the funds going into them truly represent charges or fees for the use of the government services they support. But this purpose is fulfilled only if the monies raised by earmarked taxes or fees to support infrastructure or other types of capital--averaged over some reasonable period, such as three years--are actually spent on the dedicated uses.

• To ensure that this is done, the President's budget should disclose the earmarked taxes or fees and spending of these various capital-related trust funds. This will allow policy makers to make informed decisions about whether to increase spending on the authorized activities or reduce the charges now being assessed purportedly to finance those activities.

• State and local governments that are recipients of capital-related grants from the federal government should be required to maintain their capital--such as highways--as a condition to receiving any additional federal aid (unless those governments can demonstrate that there is no longer a need for the assets the federal government initially supported).

     Recommendation 7: Incentives for Asset Management.--The executive branch and the Congress should experiment with incentives to encourage agencies to manage their assets efficiently. One possibility might be to allow, on an experimental basis, one or more agencies to keep a limited portion of the revenues they raise from selling or renting out existing assets.

     Steps must be taken to improve the methods that are used to give the results of those decisions (and the programs they support) to the public and policy makers. In particular:

     Recommendation 8: Clarification of the Federal Budget Presentation.--The President's annual budget should contain a breakdown of proposed current and projected federal spending over the budget year and the subsequent four years among the following categories: investment, operating expenditures, transfers to individuals, and interest. Such a breakdown would make available to policy makers and the wider public the President's long-run vision for federal spending. This information might also encourage Congress to find ways of taking a longer-run view in its annual budget deliberations.^(g)

     Recommendation 9: Financial Statement Reporting.--Reporting on financial activities and asset positions of the federal government should be enhanced in a number of ways to better inform the Congress and the public about the ways in which the federal government's assets are being used and maintained:

• Federal agencies should be required to issue to policy makers and the public more detailed information (both in print form and on their websites) about the composition and condition of the federally owned or managed capital assets under their control. OMB should consolidate these reports, which should continue to be based on independently developed accounting standards, and report on them in summary fashion in the annual budget.



• There should be enough information in the consolidated reports to provide Congress and the public with accurate benchmarks for making appropriate comparisons both in the current year and over time.



• The calculation of depreciation in various government reports should be standardized.

     With more comprehensive, objective information on how the federal government as a whole, as well as individual agencies and programs, have used resources, increased or depleted assets, and undertaken new investments, debates over critical national policies would be better informed. Private corporations report audited financial results and asset and liability positions to investors. By the same token, the federal government should make available to the American people audited financial statements and underlying detail that go well beyond the information shown annually in the unified budget. Just as corporate decision-makers have accurate accounting data to help them assess past performance and make decisions about the future, Congress and the public should also have accurate accounting on federal assets and investments.

     Recommendation 10: Condition of Existing Assets.--Work is planned at the federal level for agencies to begin developing standardized methods for estimating deferred maintenance. The commission strongly supports these efforts and encourages OMB to work with the agencies to complete this task promptly and to implement its results. In addition, the federal government, working with states and localities, should endeavor to report on the condition of assets owned at these lower levels of government, or at least those that have received federal support. In combination with the rest of the information provided in the audited financial statements, data on deferred maintenance will enable policy makers to develop sound plans for maintaining existing assets and spending on new ones where that is advisable.

     Finally, steps should be taken to improve the process used in evaluating the impact of past budgetary decisions, so that policy makers can be in a position to make improvements, if warranted.

     Recommendation 11: Federal "Report Card."--Under OMB guidance, agencies should assess the extent to which major investment projects have produced returns in excess of some benchmark cost of capital, such as the prevailing interest rate on long-term federal debt, the average cost of capital expected by private market investors, or some other threshold that OMB believes the public would find useful. This federal "Report Card" could be included in the President's annual budget. The commission recognizes that the projects for which it might be feasible to provide a monetary analysis may account for a relatively small fraction of total spending; nonetheless, it believes that over time advances in estimating techniques may permit a larger fraction of total spending to be evaluated in this manner. Where benefits and costs cannot be expressed in monetary terms, the evaluations should identify project objectives and assess outcomes qualitatively.

     The foregoing recommendations are summarized in the table on the following page. The columns in the table refer to three different classes of capital, which are discussed in the body of the report: the federal government's own assets (such as buildings in which federal agencies are located), the federal government's investment in assets owned by state and local governments (such as highways), and the federal government's investment in what we have labeled intangible national assets that are financed but not owned by the government (such as benefits accruing from federal expenditures on research and development and or on education). Our recommendations are then classified both by the stage of the budget process at which they are directed and by the types of capital that they are likely to affect. Because a number of our recommendations are designed to improve decision-making with respect to one or more categories of capital, they are listed in multiple columns.

     While the primary responsibility for initiating most of the foregoing recommendations rests with the executive branch, in certain cases Congress also has an important role. Indeed, virtually all of the recommendations require active Congressional cooperation if they are to have a positive effect on the budget process and budget decisions.

     Although the commission as a whole does not endorse setting a separate cap on capital spending, it nonetheless discussed the technical details of such a change in budget procedure. The concluding section of this report contains our findings on these issues, outlines the key pros and cons of subjecting capital spending to its own limit, analyzes proposals to reflect depreciation of capital assets in the budget process, and briefly describes some alternative versions of a capital budget.

     In sum, the federal budget process can be and should be improved. The commission believes the recommendations outlined in this report would help accomplish this objective.

     This commission has been charged with examining capital budgeting in other countries, states and local governments, and the private sector, and, in the process, with addressing a number of questions about capital budgeting. It is only appropriate, therefore, to begin with the threshold issue: what is "capital" (or its annualized counterpart, "investment")?

     The commission has not settled on, nor does it endorse, a single definition of capital.⁽¹⁾ Instead, a series of distinctions between different types of capital or "investment" spending, both by governments and by firms in the private sector, seem warranted for different purposes (and different commissioners place varying amounts of emphasis on alternative definitions of capital).

     One distinction relates to the functions of capital. At its broadest level, any spending that yields benefits beyond the typical reporting period (such as a year) should be considered to be investment, and "capital" refers to the assets created by this spending. Such a definition would encompass spending not only on physical or fixed assets, such as structures and equipment, but also on human and a variety of intangible assets. "Human capital" consists of the skills imparted to individuals through training and education that enable them to increase their earnings not just in a single year, but potentially throughout their lives. Intangible assets can cover a very broad class of items. In private sector financial accounting, for example, intangibles are often measured by the expenditures required to gain patents, copyrights, trademarks, or other intellectual property protection. Certain types of public spending--including research and development (R&D), defense, nutrition, disease prevention, police protection, and drug treatment and prevention programs--may also produce intangible assets that deliver, or are at least designed to deliver, benefits over years, if not lifetimes.

     Broad definitions of investment or capital could be useful for several purposes. For example, to the extent citizens and policy makers are interested in enhancing economic growth, the definition should count both private and public sector spending on buildings, equipment, research and development (including some defense-related R&D), and education and training. An even broader definition would be justified if the goal were to measure capital aimed at improving social welfare--one that included expenditures on national defense and police to enhance security as well as spending on childhood immunization, maternal health, nutrition, and substance abuse, to improve the health and well-being of citizens over many years. ^(h)

     The accounting standards used in the private sector do not take such an expansive approach to the definition of capital. Generally speaking, they limit capital to physical and certain intangible assets (such as investments in intellectual property). Similarly, the National Income and Product Accounts (NIPA)--the federal government's statistical system for collecting and reporting data on overall economic activity--define capital to be spending only on physical assets.⁽²⁾ It is important to keep in mind, however, that while these accounting standards may be conservative, they do not necessarily constrain the way managers think about spending that provides longer-run benefits. For example, although private sector accounting standards define employee training expenditures as an expense, this spending typically generates longer-term benefits to the firm (and to the employees). The fact that these expenditures are written off during the course of a year does not stop managers or investors from considering them as investments in the future well-being of the firm.

     A second distinction relates to who owns capital: specifically, whether it is owned privately or publicly (and if publicly, by federal, state, or local governments). Individuals and firms reap most of the benefits from the spending on capital they undertake; however, the public benefits when government is making the expenditures. For example, government spending to educate each generation of citizens benefits the entire public by ensuring that the population continues to be literate, cognizant of the benefits of our system of government, and able to work in an ever-changing economic environment. Similarly, when the government spends money on the nation's defense or finances basic scientific research, the benefits accrue to all citizens. Appropriately enough, economists call investments that confer benefits on a wide class of parties "public goods," because no private person or firm can capture all of their benefits. Identifying and funding those programs that produce returns to society well above the cost of capital is especially important for enhancing economic growth.

     These points highlight the different criteria that are used to decide whether to add to private and public capital. In the private sector, capital spending decisions are made based primarily on how they affect shareholders, and are evaluated predominantly in monetary terms. In the public sector, decisions about capital take into account the impact on the public at large and rest on both monetary and non-monetary considerations.

     A third distinction is between federal government capital and national capital. Federal government capital, as we use the term, refers only to those assets the government owns, such as federal buildings or federal military hardware. National capital is a broader term, including all government spending aimed at delivering long-term benefits to any portion of the nation, whether or not it is owned by the federal government. So, for example, using the broad functional definition of capital discussed above, national capital would include spending at all levels of government on roads and other physical assets, research and development, and education and training, among other items. At the federal level, what OMB labels as "federal investment outlays," illustrated in Table 1, represents federally financed national capital regardless of who owns it.⁽³⁾ As the table shows, nearly half of the federal government's investment outlays in fiscal year 1997 were devoted to physical capital, about one-third to research and development, and the balance to education and training--roughly the same proportions that were prevalent during the earlier part of the decade.⁽⁴⁾

     Federal government capital, in contrast, can be defined as including only assets owned by the federal government, so it can be accounted for in a fashion similar to the way capital is measured in the private sector. For example, OMB's Capital Programming Guide, which provides guidance to federal agencies on capital planning, procurement, and management, defines "federal capital" to include land, structures, equipment, and intellectual property (including software) belonging to the federal government that has an estimated useful life of at least two years. Consistent with this definition, Table 2 illustrates how the federal government provided almost $66 billion of budget authority for fiscal year 1997 on "major capital acquisitions": government buildings, information technology, and "other items" (weapons systems in the case of the Department of Defense, and facilities and equipment for other agencies). The table shows that the major part of the federally owned investment was for defense-related purposes.

     This distinction between "national" and "government" capital is of more than academic interest. As discussed below, the government of New Zealand has adopted a separate capital budget but only for government capital. In contrast, the General Accounting Office has suggested defining a budget target that is a variation of national capital: public investments that promise "to raise the private sector's long-run productivity," which would include spending on infrastructure, non-defense R&D, education and training, and some defense activities, but would specifically exclude what GAO calls "federal capital," such as government-owned buildings, weapon systems, and land [GAO, 1993].

     A fourth definitional distinction is between capital created by (1) direct government spending and (2) public and private capital spending induced by government policies. The advantage of confining any definition to direct spending is that measurement is relatively easy. Nonetheless, if the objective is to measure the impact of overall government policy on national capital (narrowly or broadly defined), then a definition based only on the government's direct expenditures is too limited. A full accounting would also require inclusion of capital spending at the state and local levels and by the private sector that may be brought about by such policies as federal deficit reduction (through lower interest rates), and targeted tax incentives, as well as regulatory mandates such as those requiring or inducing expenditures on pollution control or occupational safety.⁽⁵⁾ Granted, such induced spending may be very important; however, the operational problem with adding induced expenditures is that they cannot be directly measured, but instead must be estimated, using economic models or survey responses.

     The different definitions underscore the proposition that "capital" is not a single, uniform concept, but one that varies according to why the term is being used. Indeed, this is one reason that most members of the commission are opposed to recommending that a separate capital budget using one single definition of capital be adopted for decision-making purposes. Nonetheless, definitional issues should not stand in the way of illuminating the consequences of choosing among different government programs, whether or not they are labeled as capital. Nor should debate over definitions distract attention from (1) the need to improve planning and evaluation for whatever expenditures policy makers may choose to label as capital, or, (2) in the case of federal capital in particular, the need to identify the assets the government has and report them in a coherent way.

     Finally, one important characteristic of much (but not all) capital spending is that its value declines over time. Buildings and machines wear out. Patents and copyrights have limited lives. Even the value of basic education and training may decline in a world of continuing technological change, which requires many workers to upgrade their skills constantly to maintain their earnings.

     Accounting standards in the private sector, as well as the concepts reflected in the National Income and Product Accounts, take account of the declining value of capital items by requiring property and plant and equipment (but not land) to be "depreciated" or "amortized" over their "useful lives." The annual amounts of depreciation or amortization represent expenses that, along with salaries, supplies, rent, taxes, and other expense items, are deducted from annual revenue to determine profits each year.⁽⁶⁾ A number of different methods for depreciation and amortization are in use, ranging from the "straight-line" method (that computes the annual deduction simply by dividing the original capital investment by the years of useful life) to various forms of "accelerated depreciation" (that deduct more in the early years of an asset's useful life and less in later years). Businesses may also use depreciation methods for financial accounting purposes that are different from those they use to compute their income tax liability.

     Some state and local governments account for the declining value of their debt-financed capital assets by including in their annual budgets the annual debt service on the bonds they issued to finance the investments. Debt service includes interest and the annual amount of the principal of the bond that is paid off (similar to amortization of principal on a mortgage that individuals may take out to finance their homes) or put into a "sinking fund" that is eventually used to pay off the bonds when they mature. The amortization component of the debt service charge is analogous to depreciation, but with a time profile that is the opposite of accelerated depreciation--much larger deductions in the later years than in the earlier years.

     Two important distinctions are useful to keep in mind when considering trends in capital spending:

• The amount of capital spending in any given year is defined as 'investment." This is different from the total amount of the existing capital "stock" which is the cumulative total of all previous investment minus cumulative depreciation.

• Investment, in turn, is often measured in two ways: as either the "gross" amount of spending on capital items or the "net" investment, which is the gross figure minus annual depreciation. Many analysts concerned with the contribution of capital to economic output pay more attention to the net figures than the gross figures.

     Figure 1 depicts trends in net spending on physical assets alone, as a share of GDP, by the private sector, state and local governments, and the federal government. ⁽⁷⁾ Since World War II, the shares of such net spending in GDP have been reasonably stable--more so in the public sector than the private sector--with private investment substantially exceeding public investment. Meanwhile, within the government sector, since 1950 net investment at the state and local level has consistently outpaced federal spending. It is important to note, however, that about one-quarter of state and local infrastructure spending is financed by federal grants, and much of the rest has been subsidized by the federal tax exemption on municipal and state debt.

     One reason that public investment has been of special interest to economists, business, and policy makers is its impact on economic output. In particular, some economists have argued that the decline in the public investment-to-GDP ratio shown in Figure 1 has contributed significantly to the slowdown in long-term growth from the first half of the post-World War II era to the second.⁽⁸⁾ As shown in Figure 2, although it has picked up in recent years, over the past 25 years the annual rate of productivity growth in the United States, which determines the growth in average living standards, has been substantially below that of the 1948-73 period, which some have characterized as a "golden age." Figure 3 suggests that the slowdowns in public spending and productivity growth have occurred more or less around the same time.

     The claim that the first slowdown (in infrastructure spending) "caused" the other one (in productivity growth) has proven to be highly controversial, however. Among other things, various economists have claimed that the causation runs the other way: that is, (1) public capital spending has slowed because economic growth has slowed; (2) the public capital buildup in the 1950s and 1960s was largely associated with once-in-a-generation events--the construction of the interstate highway system and the construction of schools for the baby boom generation--that could not have been expected to be repeated after they were completed; and (3) the statistical estimates used to prove that the slower growth in capital spending caused the slowdown in productivity growth are highly sensitive to the time period examined.⁽⁹⁾ Moreover, as already shown in Figure 1, public sector investment in physical assets is considerably smaller in magnitude than private sector investment, which has also decreased relative to GDP during the same period in which public investment has declined. Both of these facts raise the question of whether and why public capital spending in particular should be singled out as being primarily responsible for the trends in productivity growth.

     There's no need to resolve the debate over what has caused the slowdown in measured productivity growth over the past 25 years to conclude that all types of capital (fixed assets, human capital, and intangibles), whether owned by the private or public sectors, remain important to economic growth. Economic theory has long pointed to that conclusion. The challenge for decision-makers in both the private and public sectors is to undertake those investments that realistically promise returns that exceed the cost of financing the investments; otherwise, scarce resources will be wasted.

     Some observers have attempted to draw policy implications for the United States by comparing the intensity of investment activity here (both public and private), as well as rates of return on investment, with similar figures for other industrialized countries. Such comparisons, however, do not provide a standard for judging the appropriateness of the amount of total investment in this country, and still less for judging the amount of investment spending by the federal government.

     In any event, several points should be made about those comparisons. First, though by conventional measurements the United States invests a smaller share of its GDP than other advanced countries do, that is not true (1) if investment is defined more broadly to include expenditures on education, research and development, consumer durables and defense capital, and (2) if the relative price of investment goods and other output is correctly calculated [Kirova and Lipsey, 1998]. Second, comparing investment/GDP ratios may not be as illuminating as comparing rates of return on capital. When this is done, the United States typically comes out on top of other countries. Third, significant differences in definitions and demographic conditions make comparisons of public investment particularly complicated across countries.⁽¹⁰⁾

     The executive order directs the commission to report specifically on capital budgeting practices used in the private sector, by state and local governments and in other countries, and then to explain the relevance of those practices for budget decisions made by the federal government.

     By definition, a budget is a constraint because it implies the existence of a finite amount of resources that can be allocated among alternative uses. But what is it that limits the amount of available money? The vastly different answers to this question for private firms, state and local governments, and the federal government help shed light on the extent to which capital budgeting practices followed elsewhere are suitable for the federal budget.

     Capital Budgeting in the Private Sector

     The American economy is populated by over twenty million businesses, large and small, which surely have different ways of budgeting capital expenditures. Nonetheless, certain conventions have become standardized through custom and repetition, as well as through formal professional practice. As a result, it is possible to describe a stylized process that many firms, typically larger publicly held corporations, use to analyze their capital spending options, to choose among them, and then to account for those choices. To help understand these conventions, it is useful to refer to three basic financial statements that are found in the annual reports of publicly held companies: the balance sheet, the income statement, and the statement of cash flows.

     The balance sheet provides a financial snapshot at a single point in time, usually at the end of a reporting year, of the firm's assets (on one side) and liabilities and net worth (on the other). The two sides add to the same total. Assets are "financed," as it were, by borrowing (liabilities) and shareholders' contributions (paid-in capital and retained earnings). Broadly speaking, three categories of assets are reported on the balance sheet: short-term assets (such as cash, marketable securities, receivables, and inventories), fixed assets (structures and equipment) minus any cumulative depreciation, and intangible assets minus any cumulative amortization. Using the nomenclature of this report, capital for private firms consists of fixed assets and, under some definitions, intangible assets as well.⁽¹¹⁾ It is worth noting that private sector accounting has been standardized in Generally Accepted Accounting Principles (GAAP), which are used to prepare financial statements.⁽¹²⁾ The Financial Accounting Standards Board, an independent body of experts, is responsible for seeing that the principles embodied in GAAP are maintained, updated, and applied in a fair and reasonable manner.⁽¹³⁾

     The income statement is an accounting of revenues and expenses over a certain time frame, typically a year, with the difference representing the firm's profit or loss. Because businesses exist to generate profits, spending decisions by private companies--including whether and how much to invest in capital projects--are judged predominantly by their likely impact on profitability. Investments in capital projects by definition are designed to deliver benefits over the long run, so capital spending does not appear on the income statement. Instead, the depreciation or amortization of existing capital recorded on the balance sheet shows up on the income statement as an expense that reduces reported profits.

     Where, then, might spending on capital show up? The typical place is on the statement of cash flows. This statement combines information on where a firm gets its money and where it spends it during the course of a year: on operating activities, interest on any outstanding debt, and the full cost of capital projects.

     How do firms decide how much capital spending to undertake, and of their many possible options, which projects to pursue? Here, again, practices surely vary. But certain facts and conventions are widely understood.

     First, most firms cannot spend without limits: they are constrained by their cash on hand, revenue likely to be realized in the short run, and how much additional cash they might be able to raise by selling existing assets, borrowing, or selling new equity.⁽¹⁴⁾ In turn, creditors and investors decide whether to provide funds, if they are requested, and on what terms based on the firm's ability to repay its debts (in the case of borrowings) and generate profits (in the case of equity sales). In short, firms in the private sector are subject to market discipline.

     Second, it is standard practice in private industry for firms to assess their capital projects by estimating their "net present value." Net present value (NPV) is calculated by projecting the future cash flows the investment is likely to generate (such as rentals from a building or cost savings from invesing in new equipment or machinery), "discounting" the future cash flows by the "time value of money," taking appropriate account of the risk of investment, and then subtracting the initial cost of the endeavor. Future cash flows are discounted because a dollar today is worth more than a dollar to be received in two, three, or several years hence (since the dollar today can be invested in a financial instrument and earn a rate of interest).

     According to standard practice, it makes economic sense to undertake a capital project only if its NPV is positive (the discounted returns are greater than the project's cost), and even then a firm may decide not to proceed.⁽¹⁵⁾ For example, if the discount rate is 10 percent, a project costing $1 million but projected to generate net revenues of $200,000 annually for ten years, would have a NPV of $229,000. But if annual net revenues are projected to be only $100,000 over the same time period, the project should not be pursued because its NPV is a negative $386,000 (which doesn't even cover the project's cost).

     Passing the NPV test, however, does not mean that a project will be authorized. A firm may have many potential projects that look promising when judged by their NPVs; however, it might not pursue all of them because it may have strategic objectives that cannot be readily quantified which limit the range of investments it can undertake. The firm may also be reluctant for other reasons to seek outside financing (preferring to undertake only those projects that can be financed with cash on hand), or to limit its borrowing or sale of equity.

     Third, regardless of which of these approaches (or others) private firms may employ to decide how much capital investment to undertake and which projects to pursue, all of them ultimately measure the probable success of the projects by a single metric--the likely effect on future financial performance. Moreover, the process of evaluating these undertakings is different from that of deciding whether to make certain expenditures for operating purposes (the expenses necessary to keep the business running on a day-to-day basis). These decisions do not require long-run projections of impacts or discounting into the future, although techniques such as calculating NPVs are often used to decide whether to terminate existing lines of activity. Accordingly, operating budgets are often prepared and overseen in the private sector through a process that is separate from the capital budget (although both processes are often linked by an overall management plan). ⁽¹⁶⁾

     Finally, a firm's decision to undertake one or more capital projects is not necessarily linked with a decision about how to finance those projects. Some firms, averse or unable to take on additional debt, may finance all, most, or part of their capital projects with cash on hand; others may borrow; and still others may sell equity. But just because capital spending may require a separate decision and budget, it need not be financed to any degree with additional debt.

     Capital Budgeting by State and Local Governments

     Just as there is no single capital budgeting practice prevalent in the private sector, the approach to capital budgets also varies among state and local governments. Nonetheless, some general tendencies are worth noting.⁽¹⁷⁾

     First, most state governments maintain a capital budget separate from the operating budget. However, states differ substantially in how they define capital, the degree to which capital is separate in the governor's proposed budget and in the legislature's budget, and the means by which they finance capital expenditures. ⁽¹⁸⁾

     Second, whether or not states budget capital spending separately from other expenditures, most states have long-range capital plans, ranging from three to ten years, with five years being the most frequent planning horizon. The spending figures in these plans tend not to be as detailed as the figures included in the annual budgets.

     Third, available survey evidence indicates that the states most satisfied with their capital budgeting process use some method of keeping their legislatures regularly informed about capital needs. Some state legislatures also have a separate committee charged with overseeing all or most capital projects and their financing.

     Fourth, unlike the private sector, where different capital projects can be judged by the common standard of impact on profitability, governments are responsible for a variety of functions, including police protection, health care, and education, whose benefits generally cannot be reduced to dollars and cents. This is a common situation shared by all levels of government. Nonetheless, governments must set priorities in deciding how to spend tax revenues and any borrowed funds.

     How do state governments set priorities in deciding on their capital expenditures? Although some do it project-by-project, or case-by-case, most states have formal mechanisms, either in statute or by practice, for setting priorities. Many states that take this approach set priorities on a functional basis, allocating expenditures for higher education, transportation, aiding local governments, or protecting natural resources. Others have statutes that give priorities to certain activities, such as health and safety.

     Fifth, contrary to popular belief, state governments do not always finance their capital projects by borrowing. To the contrary, states often dip into general revenues to pay for capital items, although the extent to which they are allowed or choose to do so varies. Other major sources of revenue for state capital spending include excise taxes (such as taxes on gasoline) or grants from the federal government. In addition, while debt service--interest and repayment of principal--typically shows up in state operating budgets, no state budget includes charges for depreciation.⁽¹⁹⁾ Many states impose user fees on intended beneficiaries of capital projects in order to help service the debt issued to finance them.

     Finally, most states have either constitutional or statutory limits (often with referendum requirements) on the amount of debt they may issue. State borrowing is also disciplined by the market. Rating agencies determine the ratings they give to a state's bonds, which strongly influence the interest rate at which those bonds can be marketed. These ratings are set in significant part by measuring the amount of state debt outstanding against the economic output generated in the state. Higher interest rates due to adverse ratings can force states to limit their borrowing.

     As a broad generalization, local governments follow procedures and conventions similar to those outlined for state governments.

     Current Budgeting by the Federal Government

     It may be surprising to some that throughout much of American history, the federal government had no central budget. Until the Budget and Accounting Act of 1921, which created the Bureau of the Budget, each individual agency submitted a budget to Congress. Since 1921, the Bureau of the Budget (now OMB) has coordinated the preparation and submission of a Presidential budget for the entire executive branch. The President is required to submit the budget for the coming fiscal year by the first Monday in February. This gives Congress eight months to enact the legislation that will continue the operation of most government operations and programs. If the necessary appropriations laws have not been enacted by October 1, temporary "continuing resolutions" usually provide funds until full-year appropriations are enacted.

     Although the Congress considers the President's budget proposals, it usually does not actually pass a law setting forth a budget (although, as discussed below, the "budget resolution" passed by Congress establishes a framework for later Congressional consideration of different pieces of the budget). Instead, it enacts thirteen separate appropriations bills for the approximately one-third of all federal spending that is deemed to be "discretionary." The thirteen appropriations bills are developed for full Congressional consideration by the same number of subcommittees of the Appropriations Committees of each chamber.

     The other two-thirds of the budget covers so-called "mandatory spending," which is mainly for entitlement programs such as Social Security, Medicare, Medicaid, and unemployment insurance. Mandatory spending continues at levels regulated by standing laws unless Congress enacts legislation to change them (for example, by changing a benefit formula). The same is true of tax receipts. Congress assigns responsibility for legislation governing mandatory spending and receipts to the authorizing (rather than appropriations) committees.

     Until the Congressional Budget Act of 1974, Congress had no procedures for coordinating legislation governing appropriations, mandatory spending, and revenues into an overall fiscal policy. Instead, a fiscal policy simply emerged as the sum of all of the enacted bills. The 1974 Act aimed at bringing more order to the budget process by creating separate budget committees in both the House and the Senate, and the Congressional Budget Office (the congressional counterpart to OMB), which provides information to Congress about the costs and effects of legislation. In addition, the Act requires Congress first to decide what the projected budget surplus or deficit should be and then to be guided by that decision in enacting spending and revenue bills.

     More specifically, the 1974 Act calls for Congress to adopt each year a "budget resolution" that sets a ceiling on total outlays and a floor on total receipts. The resolution, which is not presented to the President because it is technically not a law, also allocates "budget authority" and "outlays," by functional categories, to the appropriations committees (for discretionary spending) and the authorizing committees (for mandatory spending). The appropriations committees, in turn, further allocate budget authority among their thirteen subcommittees, which must report bills back to the full committee consistent with those allocations. The resolution may also direct authorizing committees to achieve a specified amount of savings by reducing mandatory spending or increasing receipts. Finally, the 1974 Act established parliamentary rules ("super-majority" voting requirements in the Senate) to stop bills that violate the budget resolution.

     The distinction between "budget authority" and "outlays" is fundamental to understanding the way budget decisions are actually made. Congress grants budget authority (BA), enabling agencies to incur obligations. Those obligations, in turn, require outlays (actual cash payments). Capital expenditures and operating expenses typically have very different "outlay rates." Capital projects are often completed over several years, so the outlays for them are spread out over some period of time. In contrast, the outlays for such things as salaries of government workers, repairs, and maintenance, along with payments under the various entitlement programs, typically coincide with the amount of BA for the same year.

     The Budget Enforcement Act of 1990 added further requirements to the budget process for fiscal years 1991-95. The BEA has been extended twice so that its requirements now apply (with amendments) through fiscal year 2002:

• The BEA divided all discretionary spending, of which capital spending is a part, into categories and imposed statutory limits or "caps" on each category (on both BA and outlays). The categories change from year to year, but currently consist of defense, non-defense, violent crime reduction, highways, and mass transit. The separate caps for defense and non-defense are replaced after fiscal year 1999 by a "discretionary spending" category, while the other categories remain intact. The violent crime reduction category expires after fiscal year 2000, leaving the discretionary, highways, and mass transit categories. Increases in taxes do not increase spending allowed by the caps (although the BEA rules allow discretionary spending to be offset by fees charged for goods and services when the fees are authorized in appropriations acts). The caps were intended to restrain the growth of spending, whether or not additional tax revenues for more spending could be found.

• The BEA contained "pay-as-you-go" (PAYGO) provisions to ensure that the cumulative impact of changes in legislation affecting mandatory spending or receipts do not increase the deficit. In other words, any increases in mandatory benefits must be financed either by cuts in other mandatory spending or by increased revenue. Because most capital expenditures are discretionary, the PAYGO rules seldom apply to capital spending.

     The federal budget contains several types of funds. The "general fund" is the broadest and includes income and some excise tax receipts. It also includes proceeds of general borrowing, on the revenue side of the budget; on the expense side, it includes national defense, interest on the federal debt, operating expenses of most federal agencies, and some capital expenditures (broadly defined) on R&D, education, and infrastructure and other physical capital spending. "Special funds" are earmarked for specific purposes; while they are not designated by law as "trust funds," they do not differ from them in substance.⁽²⁰⁾ Most special funds are financed by user fees. "Trust funds" also have dedicated uses, and are financed by user fees or taxes; when their surpluses are borrowed, the funds receive interest. A few of the best-known trust funds are those for Social Security, Medicare, and highways (although there are about 150 such trust funds in total).⁽²¹⁾

     Although each of the trust funds is technically distinct, they are reported on a combined basis in a "unified budget," a concept adopted in January 1968 (for the FY 1969 Budget). The unified budget provides the bottom-line impact of all federal spending and taxing on the economy by indicating--through the cash deficit or surplus--the impact on credit markets.

     The unified budget also consolidates both operating and capital expenditures, which means that the federal government does not have a separate budget for capital expenditures. The receipts and outlays shown in the unified budget are similar to a cash flow statement in the private sector, which also provides a comprehensive accounting of income and spending.

     There have been several efforts since World War II to address the question of whether budget procedures should be changed to provide for separate consideration of capital and operating expenditures.⁽²²⁾ For example, a capital budget was incorporated in the Taft-Radcliffe amendment to the Employment Act of 1945, which was passed by the Senate but rejected in the House. The 1949 Hoover Commission did not recommend a separate capital budget, but it did suggest that the government publish budget estimates for current operating expenditures and capital outlays separately under each major function or activity in the budget.

     There were periodic attempts in Congress during the subsequent two decades to adopt a capital budget, but these were often opposed by the executive branch and never resulted in legislation. The capital budget was firmly rejected in 1967 by the President's Commission on Budget Concepts, as it was in previous studies by the American Institute of Certified Public Accountants and the U.S. Chamber of Commerce. Interest in the idea returned in the 1980s with the apparent approval of Comptroller General Charles Bowsher and the suggestion by President Reagan in 1986 that the idea be studied. Interest in capital budgeting surfaced again during Congressional deliberations in 1995-96 over the proposed Balanced Budget Amendment (BBA) to the Constitution. Some of the proponents of the BBA wanted the amendment applied only to operating expenses of the federal government, excluding some defined capital that could be financed by government debt.

     The federal budget process today continues to budget operating and capital expenditures together.⁽²³⁾ During the course of its deliberations, the commission heard several explanations of why this is the case (although not all commissioners agree with each of them).

     First, for reasons already discussed, federal policy makers have not been able to agree on a single definition of capital or investment in the public sector. While a technical analysis that accompanies the budget (today it is known as Analytical Perspectives) has used a stable definition of investment for many years, the use of the term investment in the budget to describe policy proposals has changed with the political priorities of different administrations.⁽²⁴⁾ Given the changing priorities of the Congress and different administrations through time, it is not surprising that no single definition of public capital has emerged.

     Second, capital is one of a number of inputs (along with materials and labor) that the federal government uses to deliver its services (directly or through state and local levels of government) to the public. The public, in turn, judges the government not by the inputs it uses, but by the amount and perceived quality of the output it delivers. On this view, budget decisions should focus on the goals to be achieved (such as providing education or securing the national defense), and not on the mix between capital and other inputs judged necessary to achieve them.

     Third, although there is no necessary connection between capital spending and its financing--indeed, many states, localities, and other authorities have clearly defined capital budgets without financing all capital through borrowing--there have been fears that a "capital budget" would allow what is called capital to be debt-financed (in large part or in the entirety). Those who believe these concerns are justified also fear that adoption of a capital budget could create a strong temptation for policy makers to classify a wide range of expenditures as capital or investment (1) to avoid having to pay for them out of tax receipts or (2) to avoid having them subject to caps on discretionary spending. This is especially true for high visibility projects for which there are clear, short-term political benefits to elected officials in both branches of government who advocate them.

     The fears about excessive spending are of special concern: while it is true that the federal government cannot borrow without limit, federal borrowing is far less constrained by financial markets than is the case for borrowing by private firms and state and local governments. Investors understand that people and capital can easily move to other locales if state or local taxes are considered to be too high. This limits the ability of states and localities to borrow. Simply put, the added taxes that are required to service their debts could cause individuals or companies to move to other areas if they believe that the additional services are not worth the higher taxes.⁽²⁵⁾ By contrast, individuals and corporations in this country are far less likely to move to other countries in response to changes in taxes here. Furthermore, investors also understand that there is a buyer of last resort for federal debt--the Federal Reserve, which regularly adds to the money supply by buying Treasury securities.

     Capital Budgeting in Other Countries

     The national governments of very few other industrialized countries currently have a capital budget. At one time, Sweden, Denmark, and the Netherlands engaged in the practice, but all have since abandoned it. However, New Zealand and more recently the United Kingdom have adopted different versions of a capital budget for decision-making purposes.

     In 1988, New Zealand's national government introduced a capital budget for government-owned fixed assets. Spending on these items is separately budgeted and not shown on the government's operating budget, which is compiled under the accrual method of accounting. Depreciation of government capital is reflected on the operating statement, analogous to the way it would be accounted for in a private business in the income statement. Nonetheless, the full cost of capital assets must be appropriated in advance.⁽²⁶⁾

     In June 1998, the United Kingdom announced an even bolder capital budgeting initiative. Under this approach, the British government has established for a three-year period a budget for all physical investment and grants in support of capital spending. A two-part financing rule has been announced to accompany the budget: (1) the "golden rule" under which the government will borrow only to invest (and not to support current spending), averaged over the economic cycle; and (2) a limitation on borrowing to ensure that the public debt-to-national income ratio is stable over the economic cycle. The new system was adopted with the explicit intention of encouraging more spending on public capital, raising net public investment as a share of GDP from 0.75 percent to 1.5 percent [Brown, 1998, p. 6].

     It is too early to judge the results from either of these initiatives. Still, at least three features of the governmental systems in both countries are noteworthy. First, neither government counts expenditures on education and R&D--part of what we have labeled "national capital"--as capital for budgeting purposes. Second, the governments in both New Zealand and the United Kingdom operate within a parliamentary system under which the party controlling the executive branch also controls the majority in the Parliament. Accordingly, the proposed budget of the executive branch is expected to be adopted into law, unlike in this country. Third, agency heads in both New Zealand and the United Kingdom have greater authority to manage their operations, with incentive-based pay, than do their counterparts in the United States.

     A central question the commission has addressed is to what extent, if any, does the current federal budget process lead to less-than-ideal decision-making about capital spending? We answer this question in two parts: whether and to what extent the current process leads to a bias one way or another in (1) capital spending in the aggregate, and thus relative to other types of spending (the possible "macro" bias), and (2) the allocation of capital spending among different projects and activities, including maintenance of existing capital assets (possible "micro" biases).

     Is There a Macro Bias?

     The commission reviewed evidence and heard testimony suggesting that the current budget system has important biases in both directions with respect to capital spending--no matter how the term is defined. It is impossible to know which biases predominate, however, without first having an objective standard of what level of aggregate spending is optimal.

     It is difficult enough for a private firm to calculate its ideal level of capital spending, taking account of expected future profitability and the riskiness of the investments. But calculating an ideal level of capital spending for the government is far more complicated. Since the government is not a private firm, its activities cannot be judged by the profitability standard often used in the private sector. Instead, government has many different objectives that are not easily compared, such as influencing the distribution of resources among different geographic regions and income groups, ensuring national security, protecting the environment, and facilitating economic growth. In principle, it might be possible to calculate and even budget an ideal amount of capital spending for one of these purposes; but the commission has found nothing that provides a supportable and objective way of specifying an ideal level of all capital spending under any definition. For this reason, the commission does not believe that anyone can say authoritatively whether the existing budget process has a "macro" bias toward too much or too little total spending on capital.

     Even so, it may be interesting to know whether recent changes in budget conventions have caused capital spending totals to move either up or down without specifying whether such changes may be desirable. For example, what effect, if any, have the caps on discretionary spending that have been in place since fiscal year 1991 had on capital spending? In particular, have caps crowded out capital projects?

     To investigate this question, the commission examined multi-year averages for spending of different types as a share of GDP, both before and after 1990. Table 3 presents the results.

     The table shows essentially no difference in spending-to-GDP ratios in each of the four categories displayed, including overall discretionary spending, between the five years preceding the introduction of the caps and the succeeding years. It is true that the spending ratios for both periods are substantially below the levels in years before 1985, especially the 1970s; but with the exception of direct physical capital (whose spending as a share of GDP dropped in the 1970s), the declines in the spending ratios occurred in the 1980s during the Reagan Administration, before the caps were enacted.

     It is impossible to know what capital spending (or, for that matter, overall discretionary spending) would have been in the absence of the caps, so we cannot state with certainty that the caps had no constraining impact on capital spending. But Table 3 demonstrates that if the caps have suppressed capital spending they probably have done so to no greater extent than they have for discretionary spending in the aggregate.

     One feature of the current federal budget process--the general practice of having the full cost of all capital acquisitions appropriated by Congress before any portion of the acquisition can be made or the project started--has been alleged to act as a bias against public capital investment, specifically government-owned capital.⁽²⁷⁾ The commission believes, however, that full funding is important because it ensures that policy makers consider the total costs of an initiative before authorizing and appropriating the funds for it. Otherwise, policy makers would be tempted to fund only a portion of a capital project in the initial years, which means it would be too far along to stop later. We discuss below how failure to fully fund projects in the past has produced substantial waste.

     Nonetheless, it is possible that decision-makers defer some necessary, but large, capital projects because funding them requires authorized spending to "spike" in a given year. To the extent this occurs, aggregate public investment may fall short of some ideal figure.

     How serious a problem this actually turns out to be, however, depends to a significant degree on whether spending is more constrained in any year by the caps on budget authority or on outlays. As it turns out, the caps on budget authority (BA) seldom have constrained spending. Instead, in most years since the BEA was enacted, the outlay caps have been reached first. As already noted, capital projects also tend to have low outlay rates--that is, they spend out their budget authority over several years. When the outlay caps under the BEA are the binding constraint, the slower outlay rates for capital projects could induce Congress to spend more than it otherwise would on public capital. This is because operating expenses, including maintenance, tend to spend out quickly, and thus get scored as outlays in the forthcoming budget year.⁽ⁱ⁾ Of course, there are projects so large that even if the outlays are spread over several years, the annual outlay is still a "spike" and spending could be constrained if the outlay caps are binding.^(j)

     Efforts to get around budget spikes, meanwhile, produce distortions of their own. As just noted, agencies can be tempted to use "camel's nose under the tent" budget tactics that have led to inefficient outcomes. Another, potentially wasteful budget maneuver for avoiding spikes is for agencies (sometimes with Congressional blessing) to enter into short-term leases rather than to construct or purchase property at the outset--even when the life-cycle cost of the purchase would be lower than the cost of stringing together a series of short-term leases. Both of these "tricks" demonstrate that seemingly arcane scoring rules can have a real impact on budget decisions.

     Are There Micro Biases?

     Although it may not be possible to determine whether current budgeting procedures have caused a sub-optimal amount of total capital spending, there is much greater reason to believe that the current system generates biases at the micro level: that is, capital spending is allocated among capital projects and initiatives, including the maintenance of existing capital assets, in a less-than-ideal fashion.

     The Congressional Budget Office has reviewed the available studies of the measured economic returns from different activities, finding a very large variation--from programs that have produced estimated social returns well in excess of the cost of capital, to those that are producing almost no positive returns.⁽²⁸⁾   Significantly, the CBO cites evidence indicating that maintenance can pay social dividends well in excess of the returns realized on some large new projects [CBO].

     The commission recognizes that budgeting is not a mechanistic exercise solely in search of initiatives with the highest economic returns.⁽²⁹⁾   But in deciding how much attention to pay to efficiency and how much to distributional objectives, policy makers must work within a structured framework that (1) confronts them with the implications of the relevant tradeoffs and (2) provides maximum incentives for producing cost-effective decisions. Of particular interest to the commission is the need for federal decision-makers to take adequate account of the interests of American society over the long run. The commission has concluded, however, that in several respects, the current budget process impedes the ability of decision-makers to achieve these important objectives.

     To understand the basis for this conclusion, we first briefly review the key phases of the current federal budget cycle, and then discuss its shortcomings.

     Phases of the Current Budget Cycle

     The "budget process" of any organization is usefully understood as the combination of four important, separate functions: planning and analysis, which leads to budget recommendations; the making of budget decisions; accounting and reporting of the results; and evaluation of the outcomes of budget decisions and subsequent readjustment in programs, where appropriate. We have already described the legal process by which budget decisions are made. At the risk of some over-simplification, here are some key features that explain how the federal government carries out the other three functions.

     The process begins generally 18 months in advance of each fiscal year at the agency level, when individual departments and agencies develop internally the budget requests they will make to the President (initially through OMB) for that fiscal year. Until relatively recently, with few exceptions, agencies focused their budget plans only on a single year and generally paid little attention to their long-run plans. This changed to some extent with the enactment of the Government Performance and Results Act of 1993 (GPRA), which requires agencies to submit five-year strategic plans to OMB every three years. The first such plan was submitted in 1997, the next one is due in 2000.

     For the most part, the strategic plans are descriptive in nature and do not contain out-year spending/revenue projections. Nonetheless, the agencies separately provide to OMB their spending and revenue projections five years out under presidential policy. OMB uses these projections to present in the President's annual budget five-year projections of revenue, by major source, and outlays in aggregated form and at the function and program level (OMB's data base includes projections at the "account"level beyond the budget year, but these are not shown in the budget).

     The GPRA requires agencies to submit performance plans to OMB and the Congress each year. The Act also requires OMB to prepare a government-wide plan. These plans, the first of which was submitted with the President's budget for FY 1999, are supposed to lay out the agencies' goals in objective, quantifiable terms (such as the airplane accident rate for the Federal Aviation Administration) for that budget year.

     With respect to capital projects in particular, OMB's Capital Programming Guide requires agencies to analyze their life-cycle costs and benefits as part of any request for funding of planned projects. Once budget decisions are made, the results are reflected in annual reports issued by both OMB and CBO displaying the agencies' current and historical spending patterns.

     Agencies also prepare balance sheets that report their assets and liabilities. The Chief Financial Officers Act of 1990 required all cabinet departments, major independent agencies and the government as a whole to have audited financial statements. These financial statements are prepared in accordance with federal accounting standards developed by the Federal Accounting Standards Advisory Board (FASAB).⁽³⁰⁾ Of particular interest to this commission, these standards require the financial statements to disclose in footnote form estimates of deferred maintenance, effective with the statements for fiscal 1998. In his fiscal year 1999 budget, the President set a goal of having an unqualified opinion on the consolidated (government-wide) financial statements for that year. Furthermore, twenty of the twenty-four agencies under this Act are committed to obtaining unqualified opinions on their own statements in the same time frame [OMB and CFO Council, 1998].

     Various mechanisms are in place for evaluating the outcomes and ongoing progress of federal programs. The agencies typically have evaluation efforts under way. Congress periodically asks the General Accounting Office to prepare independent evaluations. Nonetheless, no ongoing systematic, government-wide evaluation process is in place, whether for capital spending (however defined) or other types of spending.

     Shortcomings of the Current Process

     As reflected in the foregoing summary, a number of significant improvements have been made in recent years in certain stages of the federal budget process. Even so, the commission has concluded that the existing process, at each of its various stages, still contains a number of important shortcomings. A broad theme that ties the various flaws together is that the federal government--both the executive and legislative branches considered together--is so heavily focused on each current budget year that too little attention is paid to longer-run matters. Furthermore, policy makers are not held sufficiently accountable for the longer-run implications of their current decisions. This shows up in part in wasteful spending on some capital projects, a shortchanging of maintenance of existing assets, and perhaps some missed opportunities (which are inherently difficult to measure, but nonetheless real).

• While the strategic plans and performance goals required of agencies under the GPRA are a major step forward, a number of important defects remain at the planning stage of the process:


--Updated long-term strategic plans are not required of the agencies annually, nor are they integrated with the five-year budget plans submitted by the agencies. Furthermore, because there is no systematic format for the strategic plans, they make it too easy for agencies simply to justify their missions rather than provide true forward-looking plans for achieving longer-term results-oriented objectives in a cost-effective manner.

--There is uneven progress among the agencies in stating goals and missions in performance plans, as required under the GPRA.

--Insufficient attention is paid to benefit-cost analyses of capital spending initiatives in particular, both before and after they are proposed. The analytical requirements set forth in OMB's Capital Programming Guide apply only to government-owned assets, and not to the broader types of assets that belong in any definition of national capital (at a minimum, infrastructure spending, R&D, and education and training). Many agencies and OMB lack the resources to design and conduct benefit-cost analyses, while Congress pays what the commission believes to be insufficient attention to such analyses in its oversight, authorizing, appropriations, and budget resolution activities. Furthermore, GAO and CBO do not have enough analytical resources to review the work of the agencies, and thus to assist Congress in assessing the merits of capital spending proposals.

• Because budget decision-making is inherently a political process, it is likely that a bias exists favoring projects with high local visibility and a concentrated impact on employment (such as roads, buildings, and waterways) and against those that are less visible and have a more diffuse impact on employment (such as computers for the Internal Revenue Service or the Social Security Administration). Although this problem can never be fully overcome, it can and should be mitigated by a commitment by both the executive and legislative branches (1) to sound analysis before approving new projects and (2) to supporting evaluations of the impact of those initiatives after they have been undertaken (see below).



• As discussed above, failure to fully fund capital projects in advance can lead to wasteful spending. For example, the General Accounting Office has found that incremental funding practices have led to substantial cost overruns, schedule slippages, and terminations in the Department of Energy's major acquisitions [GAO, 1996]. The associated waste in expenditures has been substantial: the four canceled projects since 1983--the Superconducting Supercollider being the prime example--cost $8 billion before they were terminated.



• Several aspects of the trust funds established to support certain types of capital spending--such as the construction of highways, airports, and water projects--are problematic. As a threshold matter, the existence of the capital-related trust funds themselves insulates the programs they support from the annual balancing of priorities across the government. At the same time, the commission recognizes that, in principle, the trust fund device may be justified (1) where the revenues going into them represent charges or fees on users of the services they support and (2) the earmarked fees and taxes are spent on the purposes for which the funds were created.

     The tendency toward surplus in some trust funds has become a problem under current scoring rules. Specifically, these rules treat revenues going into the trust funds on the mandatory side of the budget, but classify the spending out of the trust funds as discretionary spending and thus subject to caps. Congress and the administration took a major step toward rectifying the imbalance in the highway trust fund generated by this difference in scoring with the enactment of the Transportation Equity Act for the 21st Century in 1998. This legislation creates separate BEA caps for highway and mass transit spending, and it sets the caps equal to the receipts from motor fuels taxes collected the previous year.⁽³¹⁾ The commission does not endorse the specific spending formula in this act as a model for other trust funds; however, it does believe that the principle of tying spending out of the capital-related trust funds to the tax and fee revenue that flows into them, averaged over some reasonable time period, is a good one to follow.

     The current budget decision-making process also exerts biases against both routine and major maintenance, such as rehabilitation and remodeling (which represents a different type of capital expenditure). As already noted, the presence of the outlay caps feeds such a bias because the budget authority for both types of maintenance has associated with it a more rapid outlay rate than budget authority for new construction. In addition, there currently is no mechanism assuring that state and local governments receiving federal support for new capital projects adequately maintain those assets, once they have been constructed or acquired (nor do rating agencies generally allow maintenance to be bonded). This can defer maintenance, in turn leading to excessive funding for new assets when it may be more cost-effective to maintain existing assets.

     The shortchanging of maintenance is aggravated by the lack of accurate and timely information on the condition of federal and federally funded assets. Granted, recently adopted federal financial accounting standards require the audited financial statements of the agencies to be accompanied by footnotes disclosing the extent of deferred maintenance; yet footnote disclosure is not a substitute for a more complete and detailed report on the actual condition of federally owned assets. In addition, the federal government's financial statements do not contain information on the condition of assets at the state and local levels, some of which the federal government has funded.⁽³²⁾ Information about the current condition and even obsolescence of assets is critical if policy makers are to design effective maintenance and capital spending programs.

     The commission cannot stress too strongly the importance of having reliable estimates of deferred maintenance. Currently, there is no generally accepted method for agencies to use in estimating deferred maintenance. This is a significant shortcoming since sound policy making requires having accurate information of deferred maintenance in setting spending priorities and in deciding whether to purchase new assets or fix existing ones. This shortcoming has led the FASAB to propose an amendment to its current standards that would relax the audit requirement for the information reported on deferred maintenance. In conjunction with this change, OMB is planning to organize a task force to develop methods for making consistent, government-wide estimates of deferred maintenance, which should enable these estimates to be fully audited. Still, until better and more-consistent information about the condition of federally owned and financed assets is routinely made available, policy makers will be unable to make fully informed decisions about whether to fund new projects or put more money toward maintaining existing assets.

     Though efforts have been made to evaluate the effectiveness of government programs, we believe there is still little systematic retrospective analysis within either branch of the federal government to determine whether capital projects generated the benefits and came within the cost projections that were originally promised.

     In sum, we recognize that it is difficult to determine whether the existing budget process produces insufficient or excessive amounts of capital spending in the aggregate; however, there are several reasons for believing that aspects of the process contribute to a sub-optimal allocation of capital spending among various projects while shortchanging maintenance.

     The commission considered a range of proposals to address the problems that have just been identified. We believe the appropriate response is to make improvements in each of the component parts of the budget process. Many of the recommendations we outline below relate to improvements in information, but others also entail changes in the ways that budget decisions are actually considered and made.

     Better Planning and Analysis

     Long-range planning for all kinds of expenditures and operations of the federal government is essential (1) to ensure that services are delivered to the public in the most-effective manner and (2) to allow policy makers to judge how much and what kinds of capital are needed to provide public services.⁽³³⁾ Given the difficulty of terminating programs and initiatives once begun, the preparation and publication of long-run plans can help ensure that resources are wisely committed to new programs before they are launched, while facilitating ongoing readjustment in priorities when appropriate. The commission advances the following recommendations to help improve this process.

     Recommendation 1: Five-Year Strategic Plans

     Although the GPRA made major strides in requiring agencies to prepare five-year plans, we have pointed to a number of gaps in the existing planning process that should be filled.

     First, the five-year plans should be prepared annually (not just every three years) and should be integrated with the annual performance plans. Furthermore, the plans should be an integral part of the budget justifications sent to Congress.

     Second, the plans should be reconciled with the longer-run budget projections that the agencies already submit to OMB. In particular, the plans need to state results-oriented objectives--not just for the current budget year under current budget policy, but ideally with respect to future projected changes in policy.

     Third, the plans and annual budgets should be tied to the life-cycles of the agencies' capital assets. The following elements of capital planning are common in the private sector and among state and local governments, and should be standard practice for the federal government: a needs assessment for such additional capital assets; a realistic maintenance schedule, funded appropriately; and recognized replacement cycles.

     Fourth, OMB should develop standardized formats for the plans (in consultation with GAO and CBO) so that policy makers in both the executive and legislative branches can more easily compare the plans of one agency to another. Among other things, the plans should be less voluminous than many currently are, should record past successes in achieving defined results-oriented objectives, should identify shortcomings that need to be addressed, and should spot challenges that remain to be tackled. The plans should also identify major future outlays for physical assets (segregated in a separate "capital acquisition fund," as discussed below) in a level of detail that OMB should specify.

     Fifth, OMB should expand its efforts to evaluate the plans (together with benefit-cost analyses of major projects, as discussed below) and to consider them in connection with government-wide planning. Among other things, the plans should help identify programs and efforts that are no longer needed, programs that might be better carried out by other federal agencies or other levels of government, and new programs that may be truly necessary. The results of this exercise should be considered in the preparation of the President's annual budget.

     Sixth, in considering agency appropriation requests, the Congress should take account of the agencies' five-year plans and of OMB's annual evaluations of those plans, as reflected in the President's budget. Congressional authorization, appropriations, budget resolution, and oversight hearings should focus on these plans and evaluations. Congress should also study ways in which it might improve its own procedures to give more weight to the longer-run implications of its current year decisions and to issues with longer-run consequences. In undertaking this task, Congress might find it useful to take advantage of the wide range of institutional expertise available to it, including resources within the Congressional Budget Office, the General Accounting Office, and the Congressional Research Service.

     Recommendation 2: Benefit-Cost Assessments

     The benefits and costs (both expressed in monetary terms to the extent practical) of alternative options should be considered before decisions are made. This principle has been part of executive branch regulatory rulemaking (for "major" rules) for over two decades. It has recently been required of federal capital projects as well through OMB's Capital Programming Guide.

     The commission believes that several extensions beyond existing practice are warranted. First, the benefit-cost requirement should be extended beyond federally owned capital assets to the broader array of undertakings associated with a definition of national capital. To some extent, this is already done, although not in a systematic fashion. Most agencies fund evaluations of their programs. We are suggesting that the evaluation process become more systematic and institutionalized. Policy makers should not wait for sporadic economic studies of individual programs prepared by academic scholars to appear in the professional literature. Instead, there should be an ongoing effort within the government to analyze the benefits and costs of all major programs--whether or not related to capital expenditures--so that they can be adjusted, refashioned, or eliminated, as appropriate. As a practical matter, it may be useful to begin by requiring benefit-cost analyses only for "major" initiatives, such as those over a certain dollar threshold; later on, smaller capital projects and government programs could be analyzed in the same fashion.

     Second, more resources within the agencies, OMB, CBO, and GAO, should be devoted to carrying out this mission. Those resources should also support OMB in its effort to become a clearinghouse for "best practices" in evaluation techniques that