March 1, 2004

By  Richard Kogan

PDF of report

Related Report:
Administration’s Proposed Discretionary Spending Caps Represent Unsound and Inequitable Policy

View Related Analyses

If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.

Key Conclusions

1. Caps on appropriations can be very effective in enforcing a broad agreement to restrain deficits. They worked well from 1991 through 1998.

2. However, appropriations caps that entail cuts in discretionary programs are unlikely to be adhered to unless tax cuts and entitlement increases also are being held in check.

3. Caps that are too severe are not sustainable and may actually undermine fiscal discipline.

4. The Administration’s proposed caps fail the test of fiscal prudence for both reasons two and three.

In its budget, the Administration proposes establishing caps, or statutory limits, on the total cost of programs (known as “discretionary programs”[1]) whose funding is determined by annual appropriations bills.  These caps would cover a little less than two-fifths of the budget.  (The remaining three-fifths is made up of entitlements such as Social Security and Medicare and interest payments on the debt.)

The caps the Administration has proposed are very tight and would necessitate deep cuts in domestic discretionary programs.  Of particular note, the caps would not be accompanied by comparable fiscal restraint elsewhere in the budget.  They would not be part of a larger plan to reduce projected deficits.

To the contrary, the administration continues to propose the permanent extension of expensive tax cuts and also is calling for additional tax cuts.  Both OMB data and a new Congressional Budget Office analysis show that, taken as a whole, the Administration’s budget proposals would make projected deficits larger, not smaller, than they otherwise would be.  As a result, the savings from the cuts in discretionary programs that the proposed caps would require would effectively be used to finance a modest portion of the cost of the tax cuts, which primarily benefit the most well-off, rather than to rein in deficits in conjunction with tax and entitlement changes that produce further savings.  Caps that are established to finance tax cuts rather than reduce deficits and that are as harsh as the caps the Administration has proposed do not represent equitable or sound public policy.

What History Teaches Us About Discretionary Caps.  Experience shows that discretionary caps can work.  It is possible to set caps in advance and adhere to them for a number of years.

In 1990, caps were established that covered fiscal years 1991 through 1995.  In 1993, the caps were extended through 1998.  The caps covering fiscal years 1991-1997 were adhered to.  The 1998 cap was eventually raised, but only by a relatively modest amount.

The 1991-1998 caps were not imposed unilaterally.  They were negotiated at length between the executive branch and the congressional leadership and Appropriations Committees.  The 1990 and 1997 negotiations produced bipartisan agreement at the highest levels.  Moreover, the 1991 and 1993 caps did not stand alone; they were part of a set of balanced, multi-year, deficit-reduction policies that also raised taxes, especially on those who could most afford to pay more, and trimmed entitlements, especially entitlements such as Medicare (primarily through restraints on payments to health care providers).

In adhering to the discretionary caps enacted in 1990 and 1993, Congress exercised restraint regarding the funding levels that it provided for domestic discretionary programs.  Expenditures for discretionary programs grew slowly from 1990 through 1998, an average of 0.7 percent per year after adjusting for inflation and population growth.  Relative to the size of the economy, expenditures for domestic discretionary programs actually fell during these years, from 3.16 percent of the Gross Domestic Product in 1990 to 3.04 percent of GDP in 1998.  This small decline contributed modestly to deficit reduction.  In addition, from 1990 to 1998, with the Cold War having just ended and the former Soviet Union having collapsed, expenditures for defense and international affairs programs decreased substantially, from 5.6 percent of GDP to 3.3 percent, contributing significantly to deficit reduction.

CBO on Budget Process Reform

In 1993, CBO reviewed the history of the failed Gramm-Rudman-Hollings Act, enacted in 1985, as well as the successful start of the Budget Enforcement Act of 1990. CBO concluded that “efforts to reduce the deficit are most likely to be successful if the President and the Congress first agree on policy actions and then set up processes to enforce them: deficit reduction does not work well if the process changes precede the policy actions. … The congress and the President should avoid any temptation to substitute process for policy…"

Eleven years later, CBO’s view is unchanged. In its January 2004 report on the budget and the economy, CBO cautions: “[S]ome lawmakers and other observers assert that the current Congressional budget enforcement procedures are inadequate to control deficits. They argue that an additional framework such as the BEA is needed to strengthen fiscal discipline. However, experience under the BEA — and with the budget process in general — suggests that no procedures to control deficits or impose budgetary restraint will be effective in the absence of an overall political consensus to achieve those goals. Whether or not the BEA framework (or something like it) is renewed, political agreement on fiscal policy objectives is probably the largest single factor in ensuring that budget enforcement procedures and the budget process function smoothly.”

If the experience from 1991 through 1998 shows that reasonable discretionary caps can be effective in imposing fiscal restraint, however, experience also shows that if caps are too tight, they will not be adhered to and will prove ineffective.  When the Balanced Budget Act of 1997 was designed, a new set of caps running through 2002 was negotiated.  These caps were much more austere.  The levels of these caps were set with an eye to having the 1997 budget package achieve, under the economic and technical assumptions in use at the time, the goal of balancing the budget by 2002.  Had the caps set in 1997 been adhered to, expenditures for discretionary programs would have fallen significantly.  Domestic discretionary programs were supposed to decline by 0.5 percent of GDP over this period, five times the amount they fell from 1990 through 1998.  Expenditures for domestic discretionary expenditures were supposed to be cut 10 percent in real, per-capita terms by 2002.

In other words, the caps established in 1997 were much more severe than the caps established in 1990.  These caps proved too harsh, and they were not adhered to.  (See Table 1, on the next page.)

A related factor that contributed to the undermining of the caps set in 1997 was Congressional efforts to weaken fiscal discipline in other parts of the budget.  In its 2000, 2001, and 2002 budgets, Congress called for — and subsequently adopted — very large tax cuts, with none of the tax cuts being offset, despite the fact that the budget rules that the 1997 Balanced Budget Act extended also required tax cuts and entitlement increases to be paid for.[2]  Senate Majority Leader Trent Lott initially maintained that the emergence of budget surpluses justified enacting tax cuts without paying for them (in contravention of the Pay-As-You-Go rule) but that the discretionary caps must still be adhered to because “a deal is a deal.”  This one-sided view of political deals could not hold.  If the surplus meant that some bets were off, then all bets were off.  As a result, Congress first engaged in various gimmicks to breach the discretionary caps without formally admitting it.  Then in 2001 and 2002, Congress simply ignored the caps, enacting discretionary funding at levels that seemed appropriate and, after the fact, raising the caps to the levels actually provided.

This experience has led some knowledgeable observers to conclude that unrealistically tight caps are more damaging to fiscal responsibility than no caps at all.  If there are no caps, Congress may be able to set a politically acceptable discretionary spending ceiling for the coming year in its annual “budget resolution” and adhere to that ceiling as appropriations bills are designed and adopted through the year.  (This is what occurred for the 2003 and 2004 budgets.[3])  But if caps exist — even very tight ones — Congress may feel compelled to start the year acting as though the caps will be adhered to.  If these caps subsequently prove substantively or politically untenable and are breached, no cap remains.  Members of Congress may then come to believe the sky is the limit.  In such circumstances, the total level of funding may end up exceeding the level that would have been provided if the caps had originally been set at a higher but more realistic and sustainable level.

The Concord Coalition reached such a conclusion in September 2000, when the caps negotiated in 1997 were supposedly in force but were being ignored at will and no obvious alternative existed.  Concord wrote:[4]

It is time to acknowledge the obvious — the 1997 spending caps are dead.  In effect, there are no caps, which means that policymakers are now operating in an “anything goes” environment.  This is detrimental to both fiscal discipline and the credibility of the budget process. … Rules that are routinely violated are worse than no rules at all because they fail in their basic purpose to control spending and they breed contempt for the budget process by encouraging gamesmanship and chaos.

Table 1:
The Discretionary Caps Agreed to in 1990 Were Adhered to,
but the More Severe Caps Agreed to in 1997 Were Not

Change in expenditures for domestic discretionary programs

Actual change from 1990 through 1998 (reflects actual results, which were consistent with the caps)

Assumed change from 1998 through 2002 under the caps enacted in 1997,(a)

Administration’s proposed change from 2004 to 2009

As a share of GDP

-0.1% of GDP

-0.5% of GDP

-0.7% of GDP

Average annual growth rate, adjusting for inflation and population

+0.7% per year

-2.5% per year

-3.0% per year

 (a) Assumes the subdivisions between defense, international, and domestic programs set forth in the 1997 budget agreement.  Under that agreement, defense and domestic programs were assumed to be squeezed equally hard, although that was not required as a matter of statute.  In practice, funding for all types of discretionary programs — defense, international, and domestic — exceeded the plan by increasingly large amounts starting in fiscal 1999. 

How Tight Are The President’s Proposed Caps?  The caps that the Administration’s budget now proposes are even tighter than the 1997 caps, which were evaded or ignored from 1999 on.  Given the President’s budgeted levels for defense, international affairs, and homeland security, the caps would result in expenditures for domestic discretionary programs outside homeland security being reduced by 0.7 percent of GDP between 2004 and 2009.  This exceeds the 0.5 percent of GDP reduction that the 1997 budget legislation called for over the 1998-2002 period but that Congress could not sustain (see Table 1).  In fact, under the Administration’s proposal, expenditures for domestic discretionary programs outside homeland security would, by 2009, be cut 14 percent below the 2004 level in real per-capita terms, and would fall to their lowest level as a share of GDP since 1963.[5]

Adding to the problems with the Administration’s proposal, its proposed caps are being advanced as part of an unbalanced package of budget-process changes.  Discretionary programs would be squeezed substantially, and entitlement liberalizations would be prohibited.  (Any entitlement improvements would have to be financed with offsetting entitlement cuts.)  But new and unlimited tax cuts would be allowed.  Moreover, the President’s budget proposes to make permanent the 2001 and 2003 tax cuts, the cost of which over the next ten years would be many times the savings from squeezing domestic appropriations.  The Administration has proposed an agenda that essentially cuts domestic discretionary programs substantially as a way to defray a modest fraction of the cost of its tax cuts, rather than as part of a larger plan to shrink deficits.  History suggests such an approach is unlikely to be politically sustainable over time.

A recent analysis of the President’s multi-year plan for domestic discretionary programs shows how severe it is.  Under the Administration’s budget, nearly all domestic discretionary programs outside homeland security would be cut between 2005 and 2009, and many programs would be cut quite substantially. [6]  For example:

The President’s proposed caps thus suffer from substantive and political liabilities akin to those that doomed the 1997 caps.  Furthermore, the 1997 caps were negotiated on a bipartisan basis.  It is highly unlikely that the caps the Bush Administration is now proposing will be agreed to on such a basis.



The President has proposed very tight caps on appropriated programs that would likely do noticeable harm to a number of such programs and the people they assist.  Because the proposed caps are so tight, and because they are part of a budget that features tax cuts as well as one-sided budget rules under which tax cuts could continue being enacted without limit, history suggests the proposed caps would not be adhered to for long.  History also suggests that caps that are set too tight and consequently end up being bypassed are far less beneficial from a fiscal responsibility standpoint than caps set at reasonable and achievable levels — and may even be counter-productive.

What are reasonable and achievable targets?  Another lesson from the last two decades is that achievable discretionary spending targets are those that are set realistically and are part of an overall deficit-reduction plan that also raises revenues and reduces entitlement spending, so that there is a sense of shared sacrifice and policymakers and other interested parities view the plan as being fair and balanced.  CBO has written that “[t]he congress and the President should avoid any temptation to substitute process for policy” and that “[w]hether or not the BEA framework (or something like it) is renewed, political agreement on fiscal policy objectives is probably the largest single factor in ensuring that budget enforcement procedures and the budget process function smoothly.”

Caps and Emergencies

When the 1990 caps were negotiated, it was with the understanding that Iraq’s invasion of Kuwait could lead to a war and also with the understanding that the caps made no allowance for natural disasters such as hurricanes or earthquakes.  The plan therefore permitted the funding of emergencies outside the constraints of the caps.  The challenge was how to design this exception so it would not open a large loophole.

The solution was to have Congress and the President act as a check on each other.  Funding could not be treated as being outside the caps unless Congress designated that funding as an emergency by statute.  Even then, such a congressional designation was insufficient unless the President, independent of the statute, also designated the funding as an emergency (and he could designate part rather than all of the congressional amounts as an emergency).  The White House viewed the President’s independent ability to agree with some, all, or none of a congressional designation as critical to the checks and balances inherent in the system.

In analyzing the Budget Enforcement Act of 1990, CBO praised the flexibility allowed by emergency designations.  In 1993 it wrote:

Any budget process must be flexible enough to deal with unforeseen circumstances that require budgetary responses.  For example, it is often appropriate for the federal government to engage in counter-cyclical fiscal policy when the economy is in recession.  The federal government also responds to other emergencies, such as natural disasters or international crises, that cannot always be anticipated.

The budget process must recognize these realities; indeed, its continued survival depends on providing policymakers with the flexibility to deal with these unanticipated events.  The BEA assisted in this goal by establishing an explicit exception for discretionary appropriations, mandatory spending increases, or tax cuts that the legislation and the President designate as emergency requirements.  Despite some predictions to the contrary, the President and the Congress have resisted the urge to use this safety valve to evade the BEA’s strictures on a large scale.  Any process designed to enforce future deficit reduction agreements should maintain such exceptions.

This solution worked well through fiscal 1998.  Eventually it broke down, but the problem was not with the method by which emergencies were designated.  The problem was that after 1998 the caps were too tight, substantively and politically, and both Congress and the President sought to evade them.  At that point, the emergency designation was misused.  But if emergency designations had not been available, Congress and the President would have evaded the caps in other ways once they decided they wished to do so.  From 1999 on, the problem was the caps themselves, not the method for designating emergencies.

Congressional Budget Office, The Economic and Budget Outlook: Fiscal Years 1994-1998, January 1993, pp 89-90

Given CBO’s advice, attempts to set multi-year caps in isolation from a broader deficit-reduction package seem premature and ill-advised.  As long as tax increases and reductions in the largest entitlements are off the table, focusing caps and cuts on domestic appropriations bills will not achieve meaningful deficit reduction and is likely to prove ineffective or inequitable.

How Would The Caps Be Enforced?

The administration proposes that the caps be enforced through a mechanism called “sequestration,” another word for automatic across-the-board cuts.  If, at the conclusion of a session of Congress, OMB estimated that funding for discretionary programs for the fiscal year that has just started would exceed the applicable caps, the President would be required to order across-the-board funding cuts deep enough to bring appropriations into conformance.[7]

An alternative approach to enforcing discretionary caps is to establish them as a part of House and Senate rules and to forbid appropriations bills or congressional budget plans (called “budget resolutions”) from breaching those caps.  Such rules can be set aside by a vote of the House or Senate, but in the Senate, such “waivers” require a vote of 60 of the 100 Senators.

The possibility that discretionary caps established by congressional rules could be waived may make it seem that statutory enforcement of discretionary caps would be much stronger than enforcement by congressional rules.  Yet that may not be the case.  Suppose, for example, that caps were set in statute and enforced by automatic across-the-board cuts.  In that event:

If the President is not willing to go along with appropriations that exceed the cap, it also may not matter much whether the cap is in statute or has been set by a Congressional rule.  Either way, the President can veto the offending appropriations bills.  He does not need the threat or excuse of a pending across-the-board cut to do so.  It is simply a question of his willingness to exercise his authority to sign or veto legislation.

End Notes:

[1]   Annually appropriated programs are called “discretionary” programs because the Appropriations Committees have the legal discretion to decide on the level of funding these programs receive each year.  The administration’s proposal to establish discretionary caps is spelled out in Chapter 14 of OMB, Analytical Perspectives, February 2, 2004, page 215.

[2]   The first two of those tax cuts were vetoed by President Clinton, but the third — the largest of them all — was requested and signed by President Bush.  The Pay-As-You-Go Rule, supposedly in force throughout this period, was the companion to the discretionary caps.  In the original 1990 budget summit agreement and again in 1997, Congress and the President had agreed to use discretionary caps and the Pay-As-You-Go Rule to enforce the substantive budgetary decisions they had negotiated.  Under the Pay-As-You-Go Rule, tax cuts or entitlement increases were not allowed unless they were fully offset by entitlement cuts or tax increases.  When Congress enacted the 2001 tax cuts, that statutory Pay-As-You-Go Rule was set aside.

[3]   No statutory caps constrained those budgets, but Congress nevertheless adhered to discretionary levels requested by the President and approved by Congress in its budget resolutions.  (Congress engaged in some gimmicks to increase the agreed-upon 2002 and 2004 levels very slightly, but these hardly count in the scheme of things.) 

[4]   Concord Coalition, “Discretionary Spending Caps: What Next?”  Issue Brief, September 11, 2000

[5]   See Concentrating On The Wrong Target: Bush Cuts Would Reduce Domestic Discretionary Spending, As A Share of GDP, To Its Lowest Level in 46 Years, Isaac Shapiro and David Kamin, Center on Budget and Policy Priorities, February 27, 2004

[6]   Administration’s Budget Would Cut Heavily Into Many Areas of Domestic Discretionary Funding After 2005,   Center on Budget and Policy Priorities, February 27, 2004

[7]   This brief description of sequestration refers to funding, also called budget authority (i.e., it refers to the amounts appropriated by Congress in appropriations bills).  The administration proposal may also require enforcement of expenditure levels, independent of the enforcement of funding levels.  Since all expenditures derive from enacted funding (and it is both unconstitutional and illegal to expend amounts that have not been provided by Congress), control over the ultimate amount of spending for discretionary programs can be achieved by controlling the amount of funding.  Separate expenditure limits are neither necessary nor desirable because they may distort the funding decisions that Congress would otherwise make within a fixed funding cap.  To its credit, to date the administration has focused solely on funding levels when judging whether appropriations bills are within the President’s budget or exceed it.  However, Congress can use a few gimmicks to evade funding limits to some extent, such as increasing funding from the transportation trust funds or shifting forward-funded programs to an advance appropriations cycle.  These gimmicks are discussed at greater length in the appendix to The Omnibus Appropriations Bill: Are Appropriations For Domestic Programs Out of Control?  Center on Budget and Policy Priorities, February 1, 2004.  If funding caps are established, it would be advisable to address the gimmicks that can distort funding totals; the administration has proposed a limit on the use of advance appropriations.