Strengths and Weaknesses of Capital Budget Accounting
Testimony for the President's Commission to Study Capital Budgeting
Representative John M. Spratt, Jr. and Richard Kogan(1)
January 30, 1998

    The federal budget shows income and outgo on a cash basis.(2) The primary "capital budget" alternative would amortize capital investment over its useful lifetime. In discussing this alternative, we assume:

    (1) The primary alternative would change the way that capital investment is "scored." Consumption would continue to be scored on a cash basis, but capital investment would be scored on an amortized basis. This alternative is what we mean by "capital budget accounting."

    (2) The capital budget alternative, if adopted, would change the "scoring" of capital expenditures, but there would still be only one federal budget. Capital investments would be scored when consumed rather than when purchased. All federal spending would be captured in the federal budget; but a new issue would be raised: in what year(s) should the capital spending be recorded?

    (3) Some conceive capital budget accounting as implying two budgets: a "capital budget" and an "operating budget." For this discussion, we conceive capital budget accounting as a change to the current budget, which would still leave one budget in place.(3)

    With this kind of capital budgeting, we expect that the budget surplus or deficit, at least over the near term following conversion, would be similar to the surplus or deficit under cash accounting, because the annual consumption (amortization) of existing capital would be about as great as the annual outlays for new capital goods.

    Some advocates of capital budget accounting have proposed that the cost of future capital investment would be amortized, but that capital investments purchased before the change in accounting would continue to be shown on a cash basis. As a result, in the immediate years after conversion, prior construction costs would not show up in the budget because the disbursements for such construction would have occurred in prior years, and new construction costs would show up only in small amounts because only the one-year amortization of such costs would be recorded. Hence, most of the highway budget -- and the budget of other capital investments -- would disappear, and the budget would suddenly seem to be improved by tens of billions of dollars. This opens the door to legerdemain. Our assumption -- that the federal surplus or deficit would be similar under either cash or capital accounting -- assumes a conversion to capital budget accounting in which the amortization, (i.e. the consumption or depreciation), of prior capital investment would be shown as a current federal cost.

    Because surpluses or deficits are likely to be much the same under either a cash budget or a correctly designed capital budget, capital budget accounting is unlikely to alter how Congress, the President, and the public react to surpluses or deficits. Therefore, the main effect of capital budget accounting may be to tip the scales in favor of capital goods and against current consumption. Suppose, for example, that Congress is forced to squeeze spending in order to meet a particular budget target. Under capital budget accounting, the incentive to cut capital investment would be much less than the incentive to cut consumption, because cutting capital investment would save only the first-year amortized cost of the capital goods.

    The prevailing rules of cash accounting have the advantage of expressing year-to-year changes in the surplus or deficit so as to measure whether the budget stimulates or restrains the economy. Cash accounting lends itself to use of the budget to support stabilization and counter-cyclical policies.

    Under capital budget accounting, the basic concept is different. Because the budget would show capital investment only when consumed (amortized over time), the budget would be easy prey to a "pay-as-you-go" philosophy. The view could easily develop that taxes should pay for current consumption, including consumption of capital. Since the benefits of capital investment accrue in the future, this view would hold that payment for capital goods should be deferred, based on the rationalization that future taxpayers will benefit from it and they should pay for it.(4)
    These are competing concepts of federal spending. The budget could be portrayed using both. But Congress has to work with one operative concept of budgeting, and it has to be used consistently over time. Having the budget measure consumption as it occurs (capital budget accounting) is a defensible principle, maybe as a good principle as using it to measure economic stimulus or restraint. But there are problems with the principle of capital budget accounting --

    The point has been made that amortizing capital investments might change the marginal incentives of budget makers, in the White House and in Congress, and give capital investment a scorekeeping or procedural advantage over other budget programs. Some proponents of capital budget accounting support it precisely for that reason; others oppose it for that reason.

    Here's why amortizing the cost of a capital investment is thought to give it an edge over purchases recorded under cash accounting concepts. Under cash accounting, the full cost of a purchase is shown up front, while under capital accounting only the amortized cost (the amount consumed in the first year) is reflected up front. Therefore, under capital accounting, whenever there is room under the budget-year spending caps, it is tempting to add capital investment because of the many jobs apparently created(5) at what appears to be very low cost, just the first-year amortization of the capital investment.

    Conversely, if tight caps force budget cuts, capital projects may be spared for the same reason. It is impossible to cut current "amortization allowances," derived from prior capital investment, being charged to the current budget. And cutting new capital investment gets very little scored savings. In short, by reducing the first-year scoring of capital investment, capital budget accounting appears to gives capital goods an advantage at the margin over consumption programs.

    We acknowledge a couple of caveats to the foregoing analysis.

    (1) The apparent budget advantage of capital investments may not occur in reality.

    (2) Even if capital budget accounting does give an advantage, or a further advantage, to capital investment, the result may not be appropriate.
  • It is not clear that the government overconsumes at the expense of capital investment. Some would argue that we are sinking too much into prisons, courthouses, stealthy bombers, and space stations relative to the nation's needs for education, nutrition, and child care.

  • Nor is it clear that there is an absolute shortage of public investment at the federal level. Even advocates of increased public investment, such as Professor Aschauer, have produced data showing that overinvestment by some states or communities can be correlated with slower-than-average economic growth.

  • We have noted above that infrastructure, education, job training, and scientific research and development all produce future benefits. In the private sector the rate of return on an employee training program or on research and development can be as high as the rate of return on a new building. But the inclusion of non-physical capital raises other issues:
  • Some would limit capital accounting to physical infrastructure as a way around the foregoing problem. But this limitation would create a new bias: tangible investments would be given a procedural advantage over intangible investments, including such investments as research and development; and this bias could make current decision making worse.
        There may be useful applications of capital accounting in agency decision-making. For example, the Department of Defense is the largest purchaser of capital goods in the government. By keeping a capital asset budget and calculating depreciation, the Defense Department would have a means of telling if its research, development, and procurement budgets provide adequately for replacement of its depreciating assets.

        Capital budgeting could also assist agencies in making trade-offs among competing uses of resources, particularly those with downstream benefits. But the planning tool best suited for this purpose is cost-benefit analysis. Cost-benefit analysis does not require amortization of capital costs. To the contrary, capital costs are treated as incurred when paid for, on a cash basis, and these up-front capital costs are then compared to the stream of benefits, which are priced and reduced to a present value by an appropriate discount rate. Capital budgeting is aimed at allocating capital costs over the useful life of the capital goods. This is of little use in cost-benefit analysis. Cost-benefit analysis allocates capital goods to the time period they are paid for, and then tries to determine if the value of future benefit stream exceeds the present value of the capital goods.


        On balance, it seems better to keep the federal budget on a cash basis, especially for purposes of congressional decision-making.

        However, experimenting with capital budget accounting at the agency level may aid budget making and planning. For example, by booking capital assets and calculating depreciation, the Defense Department would have one means of telling if its research, development, and procurement budgets provide adequately for replacement of its depreciating assets. Rather than applying capital budgeting to the whole of the federal budget, it might be tried experimentally in this one area of the budget where capital purchasing is concentrated: defense research, development, and procurement. But capital budgeting is of minor value when agencies use the most common analytical tool for weighing the merits of public investment: cost-benefit analysis. In any event, the Administration has devoted a good deal of resources to aggregate budget analyses of various types of investment. These are published in the annual Analytical Perspectives. It may be useful to enhance these analyses as a supplement to but not as a substitute for the cash-based system of budgeting.

        For purposes of getting a better look at the meaning of aggregate federal spending, surpluses, or deficits, the results that derive from capital budget accounting are so small that he disadvantages may outweigh the advantages. To review, these are our concerns:

        (1) The government would still require a budget expressed in budget authority -- scored up front, as it is now, and representing the full amount of the authority to incur and obligation -- because Congress requires this concept to appropriate and the Executive Branch uses it to manage agency spending and obligations of money.

        (2) To the extent that capital accounting lends itself to deferral of costs to future taxpayers, it is questionable and certainly not desirable in view of the impending retirement of baby boomers.

        (3) Capital budget accounting can be manipulated. The amortization schedules that would be used by the Executive Branch or the Congress are subject to considerable manipulation, at great risk to the integrity of the system.

        (4) The practical effect of capital accounting appears to be to be a scorekeeping advantage to capital investments, a bias that already exists in many cases.

        (5) The idea of capital accounting is to distinguish spending with only immediate benefits from spending with long-term benefits. But the distinction is problematic. No definition of capital captures that distinction, and many intangible investments, seldom treated as capital, generate substantial benefits into the future.

    1. Representative Spratt is Ranking Democratic Member of the House Budget Committee; Mr. Kogan is Director of Policy of the Democratic staff of the House Budget Committee. The authors do not intend that the views expressed in this testimony should be taken as representing the views of the Budget Committee, nor does each author necessarily subscribe to every point made in this testimony.

    2. "Cash" includes checks and "near checks" such as food stamps. It also includes IOUs: if, for example, the government pays a contractor by giving him a Treasury bond rather than a check, the transfer of funds (transfer of wealth) takes place at that time, not 30 years later when the bond is redeemed. Treating the issuance of unconditional, guaranteed IOUs as equivalent to cash leads to the result that federal borrowing, which brings in cash, does not reduce the deficit because it is offset dollar-for-dollar by the issuance of Treasury securities. The government cannot borrow its way out of debt.

    The principle of recognizing IOU's as equivalent to cash also means that the government records non-traditional IOU's (e.g. a non-marketable security, a "monetary credit," or a lease-purchase contract) as though it were a cash expenditure, financed by borrowing. Taken one step further, this principle also leads to so-called "credit scorekeeping," in which the cost of a federal loan is offset by the IOU (the loan asset) the government gets in return. In this case, the IOU issued by the borrower is worth less than face value because of a higher risk of default (and possibly because of interest rate differentials). The difference in value between the loan payment and the offsetting IOU constitutes a government subsidy or cost. From this perspective, credit scorekeeping is not an example of accrual accounting, but rather an example of up-front, cash accounting in which all IOUs are broadly construed to be a type of cash.

    3. The conception of a separate "capital budget" also facilitates the erroneous notion that it is appropriate -- perhaps necessary -- to borrow to finance the purchase of new capital.

    4. The rationale for capital budget accounting would be stated differently by an economist, but would have essentially the same meaning. That argument would go as follows: The amount of net national saving (private saving plus government surpluses or minus government deficits) constitutes the pool of money available for private-sector investment. In general, higher investment leads to faster increases in productivity, and from there to faster real economic growth. Thus, over time, the government can contribute to future prosperity by reducing deficits or increasing surpluses.

    If the underlying purpose is to increase net national saving, however, the federal budget should be measured in a way to facilitate that result. Consumption should be measured on a cash basis, but that federal investment which is equivalent to private-sector investment (e.g., transportation, education, scientific research, etc.) should be recognized as contributing to national saving. More precisely, a new federal capital investment of $10 billion requires that amount of federal borrowing -- thus decreasing national saving available for private-sector investment by $10 billion -- but it increases public investment by $10 billion. The increase and decrease net to zero. Therefore, the budget should not reflect capital investment expenditures as an outlay when they are made, because they do not decrease net national saving at that time. It follows that those capital expenditures should be reflected as an outlay later, when they are consumed (when they depreciate), because that is the time when the nation's capital stock shrinks.

    One consequence of this rationale, as pointed out by the General Accounting Office, is that we should amortize only that capital which is equivalent to private-sector investment in its effect on future productivity, e.g. roads, scientific research, education, job training. Under this concept, capital used primarily by the federal government (e.g. military hardware, federal courthouses, etc.) would not be amortized because it is not the equivalent of private-sector saving/investment.

    5. There are many circumstances in which an increase in federal spending, for example on a new construction project, will not increase the total number of jobs in the nation. For example, one would not expect an increase in total employment if the new spending is offset by a reduction in other spending or by a tax increase. Neither would one expect an increase in total employment if the Federal Reserve chooses to offset the increased fiscal stimulus with monetary contraction. But if all politics is local, then proponents of a particular project will see local job increases and may not care about offsetting decreases elsewhere.

    6. Because the budget authority and outlay caps are set years in advance by fallible humans, there is often a mismatch. More often than not, any given level of appropriations will be estimated to run up against the outlay cap while there is still room under the budget authority cap. In effect, then, if budget authority for capital investment continues to be recorded up front, in most years capital projects will have a relative advantage but in other years they won't.

    President's Commission to Study Capital Budgeting