Chairman Goss, Chairman Dreier, and Members of the Subcommittees, thank you for inviting me to testify today on the budget process. In my testimony, I will discuss three topics:

• The evolving role of the Congressional budget process;
  
  
• Lessons from recent experience with budgeting; and,
  
  
• Criticisms of both the current process and some proposed reforms.

THE ROLE OF THE CONGRESS IN THE FEDERAL BUDGET PROCESS

For many decades, the role of the Congress in the federal budget process has included prescribing general rules and procedures for presenting the President's budget and for enacting spending and tax laws. Historically, the Congress has also assumed responsibility for controlling spending and the size of the budget deficit or surplus. Three decades ago, when discretionary spending accounted for nearly 70 percent of budget outlays, the lion's share of the responsibility for spending control fell to the Appropriations Committees.

Over time, however, both the composition of the budget and the nature of budget control have changed dramatically. In the current year, discretionary spending is expected to make up only 36 percent of total outlays, and by fiscal year 2000 the share directly controlled by the Appropriations Committees will fall to 30 percent. Growth in entitlement spending has outpaced both discretionary spending and revenues and, thus, led to unanticipated large deficits. Further, such growth occurs without Congressional action and is highly responsive to economic conditions and other factors that are difficult to predict. Although the budget process has had some success in controlling discretionary spending, controlling entitlement spending presents a tougher challenge. Recent reform efforts have recognized the inherent difficulties of controlling entitlement growth, but the problem persists.

How the Budget Process Developed

The general rules of the budget process are contained in two laws. The first is the Budget and Accounting Act of 1921. That law created the Bureau of the Budget (now the Office of Management and Budget) and the General Accounting Office. In addition, the Budget and Accounting Act required the President to present a set of budget recommendations annually to the Congress and has been amended several times since to include new requirements for Presidential budget submissions.

The current Congressional budget process was established with the Congressional Budget and Impoundment Control Act of 1974. That law was passed both in response to the frustration generated by the fragmented nature of Congressional budgeting and as a way to provide the Congress with greater ability to carry out its budgetary responsibilities. Part of the reason for the reform, too, was a sense that President Nixon had abused his ability to impound appropriated funds; the 1974 act severely restricted the President's ability to rescind such funds.

The act strengthened the Congressional role in the budget process by centralizing and augmenting the Congress's capacity to budget. Thus, the House and Senate Budget Committees were created to coordinate Congressional action on the budget, and the Congressional Budget Office (CBO) was established as a source of nonpartisan analysis and information on the budget and the economy. The budget resolution gives the Congress the ability to consider the broad outlines of the budget as a whole and establishes the guidelines for future Congressional action on the budget~the passage of bills governing taxing and spending.

Using the Budget Process Specifically to Reduce the Deficit

The large budget deficits in the 1980s ushered in a new era for the budget process. During this period, the primary focus of the process has been on establishing rules specifically intended to reduce the deficit. That role is manifest in two major laws-the Balanced Budget and Emergency Deficit Control Act of 1985 (popularly known as Gramm-Rudman-Hollings) and the Budget Enforcement Act of 1990 (BEA).

Gramm-Rudman-Hollings had a single goal: to reduce the size of the deficit to specified levels each year until expenditures were in balance with revenues. If annual deficit targets were not met by appropriate spending restraint or tax increases, automatic across-the-board spending cuts (or sequestrations) were to take effect. According to the targets specified in the legislation, the budget was to be balanced by fiscal year 1991. In the meanwhile, the first Gramm-Rudman-Hollings law was replaced in September 1987 by a successor, which delayed the target for budgetary balance to fiscal year 1993.

Yet the deficit did not come down. Instead of enacting the changes in laws governing revenues and spending necessary to reduce the deficit, the President and the Congress were able to avoid the intent of the act by changing the deficit targets or by basing budget estimates on overly optimistic forecasts. For example, the original deficit target for 1990, the last year the Gramm-Rudman-Hollings procedures were fully hi place, was $36 billion. The revised 1990 target, established in 1987, was $100 billion. The reported deficit for that year was $220 billion. The deficit for 1993 (the year in which the revised targets were to require a balanced budget) was $255 billion.

The BEA created a much different process than Gramm-Rudman-Hollings. It was enacted to enforce compliance with the deficit-reducing actions agreed to in the 1990 budget summit agreement. Its purpose was not to force the enactment of further reductions or to achieve some specific deficit level. Rather, it attempted to control spending and prevent policy actions that would cause the deficit to increase. Two separate procedures were set up to achieve that end. The first set caps on annual discretionary appropriations, which limited the level of budget authority and outlays through 1995. The second instituted the pay-as-you-go (PAYGO) process, which required that the cumulative effect of Congressional legislative actions controlling mandatory spending and revenues be deficit neutral. The Omnibus Budget Reconciliation Act of 1993 extended both the discretionary caps and PAYGO through 1998.

The Congress and the President have lived within the BEA's constraints. Although the deficit outlook deteriorated just after the 1990 budget agreement, that turn of events did not result from evading the strictures set up by the BEA. Rather, the deterioration of the economy and the open-ended nature of entitlement programs such as Medicare and Medicaid were largely responsible for increasing the deficit over the 1990 projections. Virtually none of the darkening deficit picture after 1990 resulted from policy actions controlled under the BEA. Furthermore, although the BEA did not include any procedure to force additional deficit reduction measures beyond the scope of the original agreement, such reductions have actually occurred.
 

SOME LESSONS FROM RECENT BUDGETING EXPERIENCE

Based on more than 20 years of experience with a variety of approaches to Congressional budgeting, some tentative conclusions are beginning to emerge. First, the Congressional Budget'Act created an infrastructure for budgeting that has proved to be of lasting value. The Budget Committees, the timetables and rules of procedure, and the information systems put in place by this legislation afford the Congress an expanded capacity to frame a budget in a systematic and informed manner, in addition to providing it with the means to pursue an explicit fiscal policy role. Although the current system is well short of perfect, the Congress now possesses the tools to achieve budget outcomes that are closer to its objectives than has been the case in the past. In short, the essential elements for effective budget control by the Congress are in place.

Second, although the budget process is clearly useful to the Congress, the process is limited in what it can accomplish alone. Budget institutions can provide policymakers with information that is relevant to the decisions they face; it can permit extensive participation in making those decisions; and it can constrain decisions to prevent some undesired outcomes. But it cannot guarantee a desirable outcome. The process alone cannot substitute for specific policy decisions required to produce a desired result. Budgeting has no magical powers: at best it can provide opportunities for the Congress to make informed decisions that will result in desired outcomes.

Third, in designing budget procedures, control mechanisms should focus on those things that can be controlled directly. That lesson comes from the contrasting experiences with Gramm-Rudman-Hollings and the Budget Enforcement Act. Gramm-Rudman-Hollings explicitly attempted to reduce the deficit without specifying how that was to be accomplished. The deficit is powerfully affected by employment, inflation, and other economic variables; it is also influenced by the decisions of individual states regarding the scope of benefits and rules of eligibility for many entitlement programs. In the case of health care, it is further affected by the behavior of prices, technology, and use of covered services. Thus, Gramm-Rudman-Hollings suffered from the absence of a direct link between the object of control and those variables the Congress can control.

The BEA avoided that shortcoming by putting controls on appropriations through the discretionary spending caps and on legislation affecting entitlements and taxes through the PAYGO provisions of the act. The relative success of the BEA stems in part from its confinement to controllable legislative actions. That is particularly apparent in the case of discretionary spending, which has been consistently declining as a percentage of gross national product.

Inevitably, the BEA's changed emphasis has come at a price. By focusing on legislative action through the PAYGO rules, the BEA does not force enactment of measures to reduce the deficit further. Under the BEA, entitlements can still expand because of changes in population, inflation, and utilization rates that occur under current law. The BEA firewall between discretionary and PAYGO controls also means that tax reduction cannot be paid for directly with cuts in discretionary spending. In addition, because the distinction between mandatory and discretionary spending is necessarily somewhat arbitrary, a single program can involve both direct and discretionary spending. That adds a whole new layer to the already complex process of drafting authorizing legislation that accomplishes what its sponsors intend for program activity and the budget. But as the "dismal science" tells us, every good thing involves trade-offs or, as the Rolling Stones point out, "You can't always get what you want."
 

CRITICISMS OF THE CURRENT PROCESS

Many observers are dissatisfied with the current process for a number of reasons, but most criticisms fall into the following broad categories:

• The budget process is too complex, time consuming, and rule-bound;
  
  
• The system controls mandatory spending less effectively than discretionary spending;
  
  
• Too much attention is paid to short-run budgetary effects of legislation and too little to overall economic effects and the long run;
  
  
• The President's role in the process is too inactive and inflexible; and
  
  
• Current budgetary accounting might bias policy outcomes.

The Budget Process Is Too Complex. Time Consuming, and Rule-Bound

Much of the complexity and time requirements of the process result from piling additional requirements onto the existing process in the pursuit of various budgetary goals. The budgetary changes that have occurred since 1974 either have expanded the use of procedures (such as reconciliation) that were not used much until the 1980s or those changes have created new procedures, reports, and rules on top of the old ones. Gramm-Rudman-Hollings added deficit targets, sequestration, and sequestration reports. The BEA gave us spending caps and PAYGO.

One complaint frequently made is that the Congress is required to act too many times on the same spending-when the budget resolution is adopted, when programs are authorized, and finally when funds are appropriated. Critics argue, for example, that if the budget was enacted less frequently (as would be the case under such proposals as biennial budgeting), more time would be freed up for legislating and oversight. But enacting the budget less frequently would decrease fiscal flexibility and could divert attention away from deficit reduction.

The process is also criticized as too rule-bound. Certainly, the PAYGO procedures, for example, have created more complex rules. As a result, a great deal of power is transferred to technical staff, and legislators who pursue changes in policies find their task much more difficult. Although the complexity of the process is certainly troubling, it may be a natural by-product of repeated attempts to bring the budget into balance. Fewer rules are desirable, but so are lower deficits. Given that many of the current rules try to curb the deficit, they may eventually outlive their usefulness. That time, however, might not come as soon as we would like, even if this year's efforts to balance the budget are successful. The pressures on the budget in the next century brought about by the government's commitment to an aging population could mean that it will be as difficult to maintain balance as it is to achieve it.

The Budget System Is Inadequate to Control Direct Spending

Mandatory spending has continued to increase as a share of the budget since 1990. The relative uncontrollability of mandatory spending has fostered numerous suggestions for bringing it under greater control, including sunset provisions for all programs, annual appropriations for mandatory programs, and an enforceable cap on mandatory spending. Social Security, still the largest of the mandatory spending categories, has its own process of evaluation and periodic review; right now, it is "off the table" for current budget negotiations.

The costs of some programs--notably Medicare and Medicaid--are estimated to grow much faster than their beneficiary populations and general inflation over the next decade. A mandatory cap might tie the growth of spending for individual programs to the increase in the eligible population and inflation. Such a cap could also use a sequestration to enforce compliance with the cap, and savings would be achieved if spending was held to the cap level.

The difficulty with both the annual appropriation and cap approach is that they are potentially inconsistent with the underlying, legislated commitment to provide benefits to the designated population. Some acceptable procedure is required to ration capped benefits among those declared by law as eligible to receive a full entitlement. Moreover, the Congress is limited in its ability to delegate authority to executive branch agencies to reduce unilaterally those payments made from mandatory spending programs.

Some critics also view the existing process as inadequate because it creates incentives to use the BEA's emergency designation, which has the effect of increasing the deficit. They contend that the emergency designation gives the Congress and the President an incentive to underfund certain accounts in the normal appropriation process in order to fit under the caps, only to increase the deficit when a supplemental appropriation is provided after an emergency is declared. They propose that emergencies be funded by budgeting for those events in advance of their occurrence under the existing caps. Others support the current process, citing the difficulty of predicting the timing of and amounts needed for emergencies.

Too Little Attention Is Paid to the Long Run and to Nonfederal Costs

Some recent studies, such as that of the Bipartisan Commission on Entitlement and Tax Reform, have decried the lack of attention to the long-term implications of budget decisions. According to that view, the five-year budget horizon creates incentives to be short-sighted and to push costs beyond the expiration date of the BEA. Accordingly, the Senate already has instituted a 10-year point of order in response to that concern. Others would go further: the Entitlement Commission suggested a 30-year window for budget decisions. Of course, a longer time horizon is essential to a complete assessment, but it significantly adds to the uncertainty surrounding budget estimates.

A second set of concerns has to do with the limited coverage of the budget. Federal policy often imposes costs on nonfederal entities. Mandating that state and local governments or private businesses provide certain services, for example, can substitute for direct federal spending. Nonfederal costs, however, do not show up in the federal budget. That outcome has led to proposals that state and local mandates and private-sector regulatory costs be moved formally into the budget. As an alternative, others have proposed that more information be made available on the total costs of such requirements before legislative action is taken and that additional procedural impediments be placed in the way of costly increases in those requirements. In fact, the Congress and the President took that approachinenacting the Unfunded Mandate Reform Act of 1995.

The President's Role in the Process Is Too Inactive and Too Inflexible

Some critics argue that the President needs to be involved more fully and continually in the budget process. That argument usually emphasizes that the President is the only elected official with a national constituency and that he or she needs to remain accountable for the budget throughout the process. Two reforms are proposed. The first would require the President's signature on budget resolutions, thus providing the opportunity for the President and the Congress to reach agreement at an earlier stage than currently. The second would grant increased authority to the President to veto individual budget items. Indeed, both Houses of Congress have enacted separate statutory versions of the line-item veto during this session.

Such an increase in Presidential involvement could restrict Congressional budgetary prerogatives. In the case of a joint budget resolution, requiring the President's signature could result in conflict and stalemate at an earlier stage of the process, making it more difficult to adopt a Congressional budget. One can easily imagine the added difficulties that would have been encountered this year if the budget resolution had had to meet the President's approval as well as that of both Houses of Congress. Conversely, we have yet to experience the endgame of this year's struggle; it is unlikely to be resolved smoothly. The line-item veto would increase the President's power to pursue his or her budget priorities at the expense of those of the Congress. Depending on the relative weights one assigns to Presidential and Congressional priorities, that might or might not be a positive development.

Current Budgetary Accounting Might Bias Policy Outcomes

A final criticism of the budget process is that current budgetary concepts--in particular the rules governing accounting in the budget--prevent budgetary costs from reflecting economic costs in a timely fashion. That criticism recognizes that policymakers respond to the estimated costs of various actions as they are scored in the budget. To the extent that costs are misstated, decisions can be biased away from the most efficient outcome.

Numerous proposals have been offered to address the possibility of a cost bias. Most would make greater use of an accrual, rather than a cash, basis of accounting in the budget. One frequently cited proposal would change the budgetary treatment of federal investments to report costs over the expected life of the investment rather than up front when the asset is acquired. To proponents, that form of capital budgeting would correct a bias that they believe results because the cost of an investment project is not shown in the budget as it is consumed through use over many years. Alternatively, smoothing out costs over an extended period might make things seem more inexpensive than they are if policymakers only consider the short run.

Another approach to the same concern about a potential bias against capital investment argues that the goal of a balanced budget should be targeted toward the operating component of the budget, not its capital component. In that view, a balanced budget rule would apply to government expenditures for current consumption and operations, but borrowing-subject to a specified limit-would be permitted to finance long-lived federal investments. Critics of that approach are concerned that relaxing budgetary discipline for spending classified as investment would create strong incentives to classify many activities as capital, regardless of whether they met suitable investment criteria or not.

Another set of proposals would expand the use of accrual accounting for federal pensions and insurance programs. Accounting for the cost of those programs when cash is paid, critics argue, does not accurately reflect the cost of an activity when costs are controllable. Opponents contend that this method of accounting would make the budget much too dependent on the uncertain estimates of events many years in the future.
 

CONCLUSIONS

In conclusion, I would suggest that-some developments to the contrary-the last two decades of Congressional budgeting have been fruitful. The Congress has created a budgetary infrastructure that enables it to address the most pressing fiscal issue of our time: how to balance the division of scarce resources between public and private uses and between present and future generations. No doubt the unchecked growth of entitlement spending poses a continuing threat to the desired balance. Moreover, the current budget process is clearly better suited to controlling discretionary spending than entitlements.

Nonetheless, as the current Congress has demonstrated in the new budget resolution, the budget infrastructure is flexible and can be used to outline the changes required to control mandatory spending. No one can predict the outcome of this year's efforts, but failure to achieve the objectives of the resolution is not likely to be the fault of the budget process. Enacting required legislation involves a consensus, not only among divergent views within the Congress but also between the Congress and the President. Such agreement is not always easy to achieve.

In sum, getting mandatory spending under control is urgent. Although potential improvements to the process are certainly welcome and should be pursued, the current system appears adequate to support policy decisions that would reduce the deficit. The hard fact remains that eliminating the deficit will ultimately require changes in the legislation that authorizes the spending in the first place.