Thank you for the opportunity to present testimony on accounting for capital assets and the pertinence of accounting to budgeting for capital assets. My views on this subject are admittedly colored by my service as a deputy to the elected Comptrollers of both New York State and New York City, and by my experience on both the state and local Governmental Accounting Standards Board and the Federal Accounting Standards Advisory Board. The biases that I developed during those years have not changed during my recent service as an adjunct professor of public administration.

In state and local government, there is a linkage between capital assets and both finance and budgeting. Finance, of course, is directly related to budgeting, because state and local budgets are financed in three basic ways: (a) taxes and user fees, (b) grants from higher levels of government, and (c) the issuance of long-term debt. Obviously, issuing long-term debt for a capital asset (or anything to which capital-like characteristics can be attached) reduces the need for current-period taxation and passes the burden on to taxpayers of a future period. The notion of generational equity tells us that future-period taxpayers ought not to be burdened with debt (and the taxation resulting from that debt) unless they receive an accompanying benefit.

I make this point in order to introduce a significant event that has influenced my views on governmental capital asset accounting, budgeting, and finance - namely, the financial crisis that New York City endured in the mid-1970's. One of the ways that New York City used to mask the impending fiscal crisis was to take advantage of laws keying the issuance of debt to objects or purposes for which the legislature established a "period of probable usefulness." As City expenditures increased relative to the tax base in the decade preceding the crisis, the City sought increasing amounts of state aid. It didn't take long before City and State officials agreed to attach "periods of probable usefulness" to all sorts of annually recurring non-capital activities that might be characterized as "capital" because they theoretically produced "future benefit." This enabled the City to finance those activities, not by taxes or state aid, but rather by issuing debt. And what were some of those activities?
    o expenditures for the annual costs of vocational education in the high schools - presumably because those expenditures produced benefits that lasted throughout the working lives of those who received vocational education.
    o expenditures for the annual cost of carting garbage to the City dump on Staten Island - presumably because the garbage created land (or perhaps even ski slopes!)
    o any administrative cost that could be linked - however remotely - to projects financed with debt.

Those who studied the City's financing practices (including several Federal agencies) strongly criticized the City. One of my jobs, when I was appointed as New York City's Deputy Comptroller in 1976, was to help develop the rules to govern future debt-issuance practices. We all agreed that the rules should be based on accepted accounting notions of capital assets - like brick and mortar projects and sanitation trucks. We even eliminated, from debt-financing, smaller items of tangible capital, like sedans and police cars. Based on previous liberties taken with the notion of "future benefit," I take an extremely dim view of any attempt to spread the cost of annually recurring operating expenditures over future periods, whether it be for budgetary, financing, or accounting purposes - even though one may argue that those expenditures produce a "future benefit."

Let me respond now to the specific matters that you asked me to cover. First, how do Federal, state, and local governments account for capital assets and what is included in capital assets?

In state and local governments, the term capital assets is generally used synonymously with "fixed assets." Fixed assets are defined as long-lived tangible assets, including buildings, equipment, improvements other than buildings, and land. Capital asset accounting depends on whether the assets are held for tax-financed general governmental purposes (such as education, highways, corrections, and mental health) or whether they are held for fee-financed business-type activities.

    o Business-type activities account for their capital resources in a manner similar to private enterprise. They therefore generally capitalize long-lived tangible assets on acquisition, and depreciate them over their estimated useful lives. (I use the word "generally" here because there is no absolute uniformity in governmental accounting.)

    o General governmental activities are accounted for on what I call a modified cash basis of accounting, which is close to budgetary accounting. Capital acquisitions are accounted for as expenditures. They are not capitalized and depreciated. For financial reporting purposes, the historical cost of fixed assets is reported, but there is no requirement to report the accumulated depreciation on those assets. Further, there is no requirement whatsoever to report the government's cumulative investment in infrastructure, such as streets, roads, street lighting, and bridges.

When state and local governments finance the acquisition of capital assets by incurring debt, the transactions are actually budgeted for twice: once when the asset is acquired (as part of a debt-financed capital budget) and again when the asset is ultimately paid for in the form tax-financed debt service. If the term of the debt is consistent with the life of the asset, you can make a good case that the debt service is a reasonable surrogate for depreciation. On the other hand, issuing "backloaded" debt or zero-coupon debt without a sinking fund causes the debt service to lose its value as a depreciation surrogate.

(I should mention at this point, that the Governmental Accounting Standards Board currently has an agenda project that is likely to affect future accounting for state and local capital assets. The Board recently proposed a dual method of accounting measurement and reporting, once on the existing modified accrual method and again on a full-accrual basis, with depreciation of capital assets shown as an expense. The proposal is controversial, and the GASB had not yet made a final decision on it.)

Capital asset accounting in the Federal government is governed by two standards recently issued by the Federal Accounting Standards Advisory Board (FASAB). The FASAB has developed financial accounting and reporting standards for several types of capital investment of concern to your Commission: (a) property, plant, and equipment (PP&E) and (b) so-called "stewardship investment."

    o PP&E. FASAB defined PP&E as tangible assets that (1) have an estimated useful life of 2 or more years, (2) are not intended for sale in the ordinary course of business, and (3) are intended to be used by the entity. The PP&E standard was influenced by the usefulness (or lack of usefulness) of depreciation in determining costs of services or products. There are four categories of PP&E, each of which are accounted for somewhat differently:

    - General PP&E, which is PP&E used to provide general government services or goods. These assets are capitalized on acquisition and depreciated over their estimated useful lives.
    - Mission PP&E, which is being re-defined by a proposed amendment so as to be limited to national defense weapons systems and support, such as aircraft, ships, tracked combat vehicles, and missiles. These assets (which constitute the greater part of Federal PP&E) are to be expensed on acquisition and not depreciated.
    - Heritage assets, which are assets that have historical significance, cultural or artistic importance, or significant architectural characteristics. Costs of constructing or renovating these assets are expensed on acquisition. However, if these assets are also used for governmental operations, reconstruction costs directly related to those operations are capitalized and expensed over the period that they provide benefits.
    - Stewardship land, which is expensed on acquisition.

    o Stewardship investments include (a) non-federal physical property and (b) human capital and research and development (R&D). Investments in non-federal physical property are grants made by the Federal government for properties to be acquired and owned by state and local governments. Investments in human capital are for educational and training programs that will benefit the public. Investments in R&D are for basic and applied research. The FASAB concluded that all these costs should be expensed when incurred, rather than capitalized and amortized over future periods. Because these expenses might produce future benefit, however, the FASAB felt it would be useful to report on them in a special report. The reporting would include showing the expense trend for a five-year period. In addition, the FASAB requires reporting of outputs and outcomes data for investments in human capital and R&D in order to justify their continued treatment as "investments."

I would like now to provide some personal views on the pertinence of accounting for capital assets to budgeting. My thinking here is driven partly by the notion that public sector accounting must (1) help managers and legislators to make decisions and (2) facilitate public accountability. Financial reporting must deal with financial accountability (which includes generational accountability), budgetary accountability, and program accountability. Just as costs can be different for different purposes, so might accounting be different for each of these three aspects.

    1. Relation of a Capital Budget to the Balanced Budget. I'm not aware of all the factors that led to the establishment of your Commission. I suspect that some of the factors relate to balanced budgets and the computation of surpluses and deficits. Hence, I'd like to talk a bit about "balanced budgets" from an accounting perspective.

    The general perception is that state and local governments have balanced budgets. Well, New York City almost went bankrupt even though it was required to have a balanced budget. Both New York City and State regularly achieve so-called "balanced budgets" through budgetary gimmicks, some of which have the effect of issuing long-term debt to finance current operating expenditures. As Don Axelrod says in his budget book: "A good deal of myth-making... surrounds this requirement. Actually, the constitutional mandate [of balanced budgets] is less of a fiscal barrier than meets the eye."

    Virtually all state and local government budgets are "balanced" on the cash basis of accounting. This contrasts with the accrual basis of accounting, where expenses are charged to the period in which they are incurred. The proposed constitutional amendment on the balanced Federal budget was also predicated on the cash basis of accounting. Using the cash basis means you can balance the budget simply by not paying bills, by not scoring current operating expenses (such as pensions and post-employment benefits earned by current employees) as expenses of the current period.

    But depreciation is an accrual accounting concept. From my perspective as an accountant, adopting depreciation as a means of scoring Federal expenses mixes cash- basis apples with accrual-basis oranges. Actually, as a result of Federal credit reform in the early 1990's, the Federal budget process already has some elements of accrual accounting. I favor accrual accounting, but there is a potential for mis-using it, as I shall point out in a moment. Further, if your Commission reforms the budget by scoring on the basis of depreciation, then I would hope the Federal government adopts accrual accounting for all its activities.

    2. Budgeting for Expenditures that Produce Future Benefit. You have been asked to report on the definition of "capital" for Federal budgeting purposes. Should "capital" be limited to physical capital or should it also embrace "soft" expenditures for human capital and basic research? My attitude here is covered by the anecdote at the beginning of my testimony regarding New York City's experience in issuing debt for recurring operating expenditures. For budgetary presentation purposes, it may be politically useful to separate, from all other expenditures, those recurring operating expenditures that might be considered "investments" because they produce future benefit. But for budgetary scoring purposes, I strongly advise against capitalizing them and amortizing them over future periods. Once started, I fear that the Federal budget will begin to look like New York City budgets of the early 1970's.

    For one thing, many recurring governmental operating expenditures (for functions like education, higher education, and health) produce some form of future benefit. But those expenditures occur year after year, and should be paid for by current taxes. For another thing (even though economists, accountants, and actuaries are in the business of estimating), it is virtually impossible to predict either the degree of future benefit the period of time over which future benefits are likely to be obtained from these expenditures. For example, I understand studies of the Head Start Program show that program benefits are no longer apparent after a child reaches the third or fourth grade. Serving in New York City 15-20 years ago, I saw little evidence that the Federally-funded job training programs were producing significant long-range benefits for those who received the training. In short, I believe the discipline that applies to estimating the useful lives of tangible fixed assets does not extend to these intangible "human capital" investments. Hence, the potential exists for abusing what appears to be an extension of accrual accounting. As a member of the FASAB, I voted not to capitalize "human capital" and R&D for accounting purposes, and voted to require reporting of outcomes measures for items classified as investments.

    The experience of the private sector accounting profession regarding costs that have similar "future benefit" characteristics may also be helpful in your deliberations. (As I said earlier, state and local governmental accounting standards make no provision for capitalizing these intangible investments.)
    o Although many companies capitalized and amortized research and development costs before 1975, the FASB has since then required those costs to be charged to expense as they are incurred. Pfizer's new wonder drug Viagra may indeed produce long-term future benefit to the company, but the cost of developing it is not reported as an asset. However, the potential investor in Pfizer can obtain significant data about Pfizer's R&D program by reading its annual report. Pfizer's latest report shows that its 1997 R&D expenses were nearly $2 billion, more than double its 1993 expenses, and up by 17% a year over the past 10 years.
    o Private sector rules on many other types of expenses that theoretically provide future benefit has produced financial reporting inconsistencies because of difficulties in estimating either the extent or the time period of those benefits. For example, the costs of developing new markets may be written off as expenses when incurred or over several years. America Online had historically written them off over two years, causing it to report higher profits than if it had expensed them immediately. AOL abandoned that practice 18 months ago by writing off $385 million of an asset called "deferred subscriber acquisition costs."
    o In a speech five years ago, Walter Scheutze, who was then chief accountant for the Securities and Exchange Commission (and a former FASB member), took issue with the private sector definition of assets as "probable future economic benefits." He suggested instead that assets be defined as "cash, contractual claims to cash or services, and things that can be sold separately for cash." Although directed at the private sector, I think his advice (which is an extension of the private sector R&D standard) is appropriate when assessing proposals to capitalize so-called "intangible or human capital" investments for governmental accounting or budgeting purposes.

    3. Budgeting for Government-Owned Tangible Capital Assets. I can see some logic in budgeting for the consumption (or depreciation) of government-owned tangible capital assets, rather than for the acquisition of those assets. However, I'd like to emphasize a few matters for your consideration, including the potential for achieving the same benefits through analysis of a separate capital section of the unified budget.
    o The private sector certainly accounts for tangible capital assets by spreading their costs over their estimated useful lives. This is done so as to measure financial performance of private enterprise by linking expenses to revenues earned in a period or to the time period itself.
    o State and local governments budget for the acquisition of tangible capital assets through capital budgets. As I said earlier, they do not depreciate them, except to the extent that they are debt-financed, and the annual debt service might be construed as a depreciation surrogate.

Here are some of the matters I think you ought to consider in deciding whether the Federal government should budget for depreciation.
    a. As an accounting standards-setter, I was guided by the notion that generational (or interperiod) equity plays a role in public sector accounting. It should also guide budgeting. To me, this means that (except for war, severe economic problems, etc.), taxpayers of a period should pay for the services they receive in a period. It also means that taxpayers of a period have a duty to at least replenish the capital assets they have consumed in a period. Depreciation (or its surrogate) helps to provide the measurements needed to assess whether generational equity has been achieved and whether there has been a net disinvestment in our capital plant. If depreciation is used, it might be better to calculate it on current cost, rather than historical cost.
    b. Although it is useful, depreciation in a very large government may not be necessary. This is because many tangible capital expenditures recur year after year. When I was in New York City, I often used to think that there was no point in issuing debt for new schools. All we needed to do was add up the number of schools (say 1,000), divide by the average life of a school (say, 50 years), replace the resulting number of schools (20) each year out of the tax budget, and issue debt only for the schools needed to meet the expanding population. I also tried to measure the financial maintenance of portions of our infrastructure by comparing each year's expenditures with depreciation, computed on the basis of replacement cost. The point of these anecdotes is that it may be possible to achieve the same result as depreciation by analytical devices that would accompany the budget submission.
    c. Budgeting for Defense Department capital asset expenditures causes special problems for a variety of reasons. First, I suspect that measuring the useful lives of those assets relates more to obsolescence than to physical life, which complicates the measurement of depreciation. Second, I suspect that the annual amount spent to maintain many of those assets is very large in proportion to the initial investment, making the maintenance cost more significant than the annual depreciation of the original cost of the assets. DOD told FASAB that depreciation of weapons systems served no useful managerial purposes, and FASAB decided not to require it.
    d. Tangible capital assets can be acquired either through direct purchase or through leasing. At the moment, I believe these methods create different budget scoring results. Outright purchase creates an immediate charge to the budget for the entire cost. Leasing spreads the cost out over term of the lease, creating a budgetary effect similar to that of depreciation. We lease buildings and we lease computer equipment. I don't think we lease tanks and aircraft, but I know of no reason why we can't. I don't believe that leasing should have a budgetary scoring advantage over direct purchase, and depreciation provides one method of equalizing the scoring.

So, on balance I think there are some good arguments for introducing depreciation into the budget process. The basic question I would raise is the method by which it ought to be introduced.

    4. Budgeting for Non-Government Owned Tangible Capital Assets. Another issue that you face is budgeting for expenditures that result in the ownership of tangible capital assets by entities other than the Federal government. I refer specifically to capital grants the Federal government makes to state and local governments.

    Based on recent GASB proposals, I believe many accountants would look on these outlays as current period expenses, rather than as a form of asset to be amortized over a future period. This is because the assets aren't owned by the grantor government. However, I think those accountants miss two key points. First, intergovernmental grants form a major part of the way the three levels of government finance programs. Under our federal form of government, many programs are mandated by one level of government, carried out at another, and financed by two or three. Who owns capital assets is not relevant to the way the Public gets things done. Second, a grantor government has the option of financing capital grants as a lump sum up front, or by providing annual grants in the form of debt service or some other form to the grantee. I believe the budgetary scoring effect of that option ought to be equalized.

    Therefore, for budgeting purposes, I would be inclined to treat capital outlays for tangible capital assets the same whether the Federal government owns the asset or whether some other government owns the asset acquired with Federal grants. If you choose to capitalize and depreciate Federally-owned capital assets, I see no reason why you cannot do the same with capital assets acquired by another government with Federal funds. This can be done with an amortization schedule based on the estimated lives of the assets.

That completes my formal testimony. I shall be pleased to answer your questions.