for procurement actions that would trigger government oversight, for example, promises to reduce the department's workload considerably. DoD expects to save billions of dollars by relying more on commercial products than on costly military specifications in purchasing goods and equipment. The department has also reexamined the process that governs procurement of major weapon systems. The review, which was headed by the Defense Acquisition Board and supported by the services' own acquisition management structures, was intended to reduce overhead and to ensure that the smallest number of people were involved and that coordination was minimized. The department has already begun to implement various streamlining measures the board recommended, such as using integrated product teams to reduce the number of oversight meetings and paperwork required to support the acquisition process.

Although such reforms could result in efficiencies and the need for fewer employees if they are successful, past efforts at procurement reform have not generated the major breakthroughs the department and the Congress have hoped for. Nearly every Administration in the past three decades has tried to

reform the acquisition process. Yet acquisition costs for weapons continue to increase beyond initial expectations. Reducing the size of the civilian workforce before policy reforms have proved their effectiveness could jeopardize their potential to be integrated into the acquisition process.

The department could cut back the acquisition workforce since the services are purchasing considerably fewer weapons than in the past. In 1990, for example, DoD bought 392 fixed-wing aircraft; this year the Administration has requested authorization to purchase only 73. The Navy is buying many fewer ships, and the Army is no longer building new tanks. Moreover, the services are developing fewer new systems to manage. In 1991, the Defense Acquisition Board oversaw 131 major programs compared with only 104 in 1997. DoD, however, plans to raise overall acquisition spending by 19 percent between 1997 and 2001, increasing the value of the procurement workload and placing a greater burden on a smaller workforce. During the past decade, reductions in the acquisition workforce have generally corresponded to reductions in procurement spending.

Contractors producing goods and services for the Department of Defense currently hold over $8 billion worth of government-owned industrial plant equipment (IPE) and other plant equipment (OPE). IPE includes metal presses and milling machines that are widely used in industry; OPE includes items such as commercial computers, filing cabinets, and desks. Believing that private ownership would be more economical, DoD officials have sought since the early 1970s to reduce the department's role in providing such assets.

This option would facilitate DoD's efforts to reduce its inventories of both types of equipment. It provides for legislation that would grant the General Services Administration (GSA) clear authority to negotiate the sale of equipment to the holding contractor in situations in which continued DoD ownership was not necessary but the contractor required the equipment for defense production. Moreover, in future contracts in which the contractor could demonstrate that it was in DoD's best interest to provide equipment, contractors would have to obtain the assets from DoD on an explicit rental or lease basis. Rental charges for DoD-owned equipment would encourage contractors to invest in their own equipment. (Under the current system, the department does not charge contractors rent for the use of IPE and OPE in defense production. Instead, DoD benefits to the extent that providing such assets lowers the prices of the goods and services that contractors provide.)

DoD's desire to reduce its role in providing industrial assets to defense contractors appears to be justified. Contractors complain that the costs of tracking such equipment in accordance with government standards sometimes outweighs the value of the

assets. Government auditors report that items are sometimes lost and that contractors hold on to unneeded or underused items rather than return them to DoD. The costs DoD incurs in providing industrial equipment are not fully reflected in the estimates of weapon system costs that are used in making program decisions. The benefits to DoD--the lower prices paid for the goods and services that contractors provide--are uncertain. Because contractors are generally not free to use DoD assets to produce goods and services for non-DoD customers, they may be discouraged from integrating defense and commercial production and may thus lose economies of scale. Moreover, contractors with access to assets supplied by the government may have little incentive to invest in more modern and efficient equipment.

The department's efforts to reduce inventories, however, have not been very successful. Today, the total value of industrial and other plant equipment appears to be slightly above the 1990 level. Improved reporting of assets may be partly responsible for that increase, but another factor is that GSA lacks clear authority to conduct negotiated sales of IPE and OPE to the contractors who hold those assets. (GSA already has the authority to conduct such negotiated sales of real property that DoD identifies as necessary for defense production but feels it does not have to own.) By providing that authority, this option would permit DoD to divest itself of IPE and OPE without disrupting the work of the contractors using those assets in defense production.

DoD has also found it difficult to enforce policies that limit the provision of DoD-owned equipment to contractors while still giving program managers the flexibility to provide equipment when it is in DoD's interest to do so. Program managers may have an

incentive to try to help contractors and reduce measured program costs by authorizing the use of DoD equipment whenever they can. The rental payments required under this option would eliminate that incentive by making the costs of DoD-provided equipment clearly visible. The Defense Logistics Agency or GSA could be responsible for setting rents at levels that would fully amortize the cost of the equipment and the overhead costs associated with its management (including the costs of carrying inventories). Faced with such rental prices, contractors would have a strong incentive to purchase their own equipment.

CBO estimates that savings under this option could amount to $2.1 billion between 1997 and 2002. That estimate reflects reduced purchases of new equipment for the use of contractors but does not include revenues from asset sales. Although sales of federal assets reduce the deficit in the short run, they do not count as savings under the provisions of the Budget Enforcement Act. Over the long run, part of the savings from fewer government purchases of OPE and IPE would be offset by higher prices for

defense goods. Some savings would remain, however, since contractors who provided their own capital would try to use it efficiently and would not have to bear the cost of monitoring and tracking government-owned assets.

The need to set explicit rental prices and enter into rental agreements with contractors is one disadvantage of this option. In practice, however, rental agreements would be needed only when it was in the government's interest to provide equipment and when renting was more attractive to the contractor than either a negotiated purchase from DoD or a purchase from a commercial source. Another potential disadvantage is that the government might not receive as high a price in a negotiated sale as it would if the asset was declared excess federal property, removed from the hands of the contractor, and sold to the highest bidder. Limiting GSA's authority to conduct such sales to a three- to five-year period might alleviate that concern while still permitting DoD to reduce existing inventories of industrial and other plant equipment.

The Department of Defense operates three distinct categories of activities that support or improve the morale and welfare of military personnel and the recreational opportunities available to them. Category A activities are those that it considers "mission sustaining" (such as gymnasiums, fitness centers, and libraries). Category B activities provide basic community support (such as arts and crafts, child care, and youth programs). Category C activities provide goods and services similar to those offered by private enterprises in civilian communities (such as golf, bowling, clubs for officers and for enlisted personnel, department stores managed by the military exchange system, fast-food outlets, and slot machines in overseas locations).

Each category relies on a different mix of funds appropriated by the Congress and nonappropriated funds (NAF) derived from user fees and sales receipts. Appropriated funds cover much of the cost of category A and B activities. Even in civilian communities, many of those types of recreational activities receive some public funding. In contrast, category C activities receive only limited support from appropriated funds and depend primarily on fees and receipts to cover their costs. The Congress's intent is that category C activities that are not in isolated locations will be largely self-supporting.

In some types of category C activities (such as telephones, exchanges, liquor stores, slot machines, recycling programs, and financial investments made with nonappropriated funds), total nonappropriated fund receipts exceed the operating costs that they must cover. For example, in 1994 those six activities generated net revenue of more than $400 million after depreciation and other NAF costs. DoD used part

of those earnings, together with appropriated funds, to help cover the costs of category A and B activities that have only a limited ability to generate their own revenue. DoD invested another portion of those earnings in buildings and equipment for category C activities. CBO estimates that the department also used approximately $200 million of the net revenue generated by profitable category C activities to cover losses in other category C activities. (That estimate includes allocated overhead costs and does not count as income the earnings transferred from one category C activity to another that is losing money. In contrast, DoD's accounting methods, which do not always fully allocate overhead costs and commonly count such transfers as income, do not show those losses.)

This option would prohibit the military services from using the earnings generated by one type of category C business activity to subsidize the losses or the overhead costs of another. Instead, DoD would be required either to use those earnings to support category A and B activities or to reinvest the money in the activities that generated the earnings. DoD would thus have to manage each category C activity so that the revenue the activity generated, together with any authorized support from appropriated funds, would cover its operating costs, including overhead and depreciation.

This option would allow the Congress to reduce the amount of appropriated funds for category A and B activities by about $200 million annually without affecting either the level of services or the fees charged for those activities. Those savings estimates assume that each type of category C business activity would be required to at least break even. An individ

ual golf course could incur a loss, but golf courses as a whole could not. Because some establishments within each business activity lose money, the total amount of nonappropriated funds generated by category C establishments that have positive net revenue is much greater than those estimated savings indicate. A more stringent approach, which would require each category C establishment that was not in an isolated location to operate at least on a break-even basis, would yield even greater savings.

Some restriction on the use of NAF earnings may be necessary if the Congress's efforts to limit the use of appropriated funds by category C activities are to have any impact on DoD's total allocation of resources. Under the current system, DoD's freedom to transfer nonappropriated funds to activities that experience losses means that it can operate activities at a loss for an indefinite period even without appropriated funds. DoD can overcome the limits on the level of appropriated funds used by officers' clubs, for example, by allocating the NAF earnings generated by the exchanges or slot machines to the clubs rather than to category B activities, which remain eligible for appropriated funds. There is some indirect evidence that this kind of substitution is taking place. According to DoD figures, the amount of nonappropriated funds spent in category A and B activities fell by almost 15 percent between 1992 and 1994 compared with only a 3 percent drop in category C.

Subsidies for businesslike activities--whether paid for out of appropriated or nonappropriated funds --may not be a wise use of DoD's resources. Outside of isolated areas, the inability of a business activity to break even probably indicates that service members have other, more attractive recreational opportunities. Requiring category C activities to cover their costs is one way to ensure that the value of the activity to today's service members warrants the cost of providing it. That rationale might underlie the Congress's decision to limit the use of appropriated funds in category C activities. Extending that limit to support from nonappropriated funds could be a logical next step. In today's austere fiscal climate, it is important that all funds for morale, welfare, and recreation, regardless of their source, be used to provide the greatest possible benefit to service members.

There are different interpretations, however, of what breaking even means, and one might argue that the break-even criterion in this option is too generous or too restrictive. On the one hand, including depreciation charges as a cost might seem overly restrictive; depreciation charges often reflect sunk costs (those already incurred) rather than the cost of providing current services. On the other hand, this option would allow DoD to operate individual facilities that did not cover their depreciation charges to the extent that other, similar facilities produced earnings after depreciation. It would also allow activities to count any support they received from appropriated funds as if that support was business revenue. Moreover, this option would not require DoD businesses to take account of the cost of land, the cost of capital apart from depreciation, or the tax revenue that local governments would lose because of DoD's federal status.

This option has some clear disadvantages. Restricting the services' ability to use nonappropriated funds to subsidize business activities would force the services to reduce costs, increase prices, or close some category C establishments (including golf courses, theaters, bowling alleys, clubs, and hotels). If prices were raised or establishments closed, then service members, retirees, and other eligible patrons would experience a loss. Many of the activities that might suffer, such as officers' clubs, represent traditional features of military life. Subsidies that encourage service members to seek recreational opportunities on-base may be warranted because they contribute to the sense of community within the military.

These disadvantages might not be insurmountable. If subsidies for a particular activity could be justified, perhaps on the grounds that the activity enhances the cohesion of military units, the Congress would be free to authorize appropriations or to move the activity from category C to category B. This option would limit only nonappropriated fund subsidies, which are not currently controlled by the Congress. Moreover, a recent study by the Logistics Management Institute provides evidence that many category C activities that do not currently earn a return could do so under more professional management. To date, DoD has been reluctant to embrace the kinds of far-reaching reforms--for example, mov