Budget Process: Considerations for Updating the Budget
Enforcement Act (19-JUL-01, GAO-01-991T).
This testimony discusses issues related to the budget process
established by the Budget Enforcement Act (BEA), which will
expire in fiscal year 2002. Because the goal of achieving zero
deficit has been accomplished, the focus of the nation's budget
process needs to change. Now, instead of aiming for a zero
deficit, the goal shifts to how to allocate surpluses among debt
reduction, spending increases, and tax cuts. Also, the budget
process should be designed to avoid what has been described as
the year-end "train wreck". A year-end "train wreck" is the
result of failing to reach agreement--or at least a compromise
acceptable to all parties--earlier in the year. Although it is
possible that reaching agreement on some broad parameters early
on could facilitate a smoother process, it is not clear that such
an agreement will always prevent gridlock--it may just come
earlier. Two ideas that have been proposed to avert the year-end
disruption caused by an inability to reach agreement on funding
the government include joint budget resolutions and biennial
budgeting. In discussing alternatives for improving the budget
process, there is a broad consensus among observers and budget
analysts that the spending constraints established by the BEA is
necessary even with the advent of actual and projected surpluses.
Such constraints include (1) extending the discretionary spending
caps, (2) extending the pay-as-you-go mechanism, and (3) creating
a trigger device or set of rules specifically designed to deal
with the uncertainty of budget projections.
-------------------------Indexing Terms-------------------------
REPORTNUM: GAO-01-991T
ACCNO: A01419
TITLE: Budget Process: Considerations for Updating the Budget
Enforcement Act
DATE: 07/19/2001
SUBJECT: Future budget projections
Budget deficit
Budget surplus
Fiscal policies
Macroeconomic analysis
******************************************************************
** This file contains an ASCII representation of the text of a **
** GAO Testimony. **
** **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced. Tables are included, but **
** may not resemble those in the printed version. **
** **
** Please see the PDF (Portable Document Format) file, when **
** available, for a complete electronic file of the printed **
** document's contents. **
** **
******************************************************************
GAO-01-991T
Before the Committee on the Budget, House of Representatives
United States General Accounting Office
GAO For Release on Delivery Expected at 10 a. m. Thursday, July 19, 2001
BUDGET PROCESS Considerations for Updating the Budget Enforcement Act
Statement of Susan J. Irving Director, Federal Budget Analysis
GAO- 01- 991T
Page 1 GAO- 01- 991T
Mr. Chairman, Mr. Spratt, Members of the Committee: It is a pleasure to join
you as you think about how to extend and adapt the Budget Enforcement
regime. The discretionary spending limits and pay- asyou- go (PAYGO)
mechanism established by the Budget Enforcement Act (BEA) will expire in
fiscal year 2002. 1 Perhaps this timing is appropriate: although most of us
would argue that some controls are necessary even in a time of surplus, the
details will be different in a time of surplus than a time of deficit.
Among the issues your staff asked me to cover is whether- and if so how- the
budget process can be designed to help avoid what has been described as the
year- end ?train wreck.? Later in this statement I will talk about some of
the particular ideas that have been proposed in this area. First, however,
I?d like to talk a bit about what a process can and cannot do. A process can
surface important issues; it can seek to focus the debate on the important
choices. But it is not a substitute for substantive debate- no process can
force agreement where one does not exist.
We ask a great deal of our budget process. We use it to determine aggregate
fiscal policy and to allocate resources across different claims. We use it
to drive program management. In the context of the Government Performance
and Results Act of 1993, we turn to the budget to tell us something about
the cost of obtaining a given level of results. Asking the process to take
on the job of avoiding a ?train wreck? may be more than can reasonably be
expected.
A year- end ?train wreck? is the result of failing to reach agreement- or at
least a compromise acceptable to all parties- earlier in the year. Although
it is possible that reaching agreement on some broad parameters early on
could facilitate a smoother process, it is not clear that such an agreement
will always prevent gridlock- it may just come earlier. The details of
implementing broad agreements are often the subject of heated debate.
1 Although the overall discretionary spending caps expire in 2002, the
Highway and Mass Transit outlay caps established under Transportation Equity
Act for the 21st Century (TEA- 21) continue through 2003, and the
conservation caps established as part of the fiscal year 2001 Interior
Appropriations Act were set through 2006. In addition, the sequestration
procedure applies through 2006 to eliminate any projected net costs stemming
from PAYGO legislation enacted through fiscal year 2002.
Page 2 GAO- 01- 991T
BEA, when first developed and later when it was extended, was a process
established to enforce a previously reached substantive agreement. As we
move from seeking to reduce the deficit to debating how much of the surplus
should be used, agreement on a broad fiscal policy posture might help. How
much of the surplus should be used to meet demands for tax cuts and/ or
spending increases and how much for debt reduction? The Congress and the
President seem to have reached a tacit agreement that the Social Security
surplus should be used for debt reduction. While this agreement sets the
outside parameters for the budget debate, it does not settle either the
distribution between tax cuts or spending increases or the allocation of
either one. It is already evident that, by itself, this broad framework is
unlikely to make for smooth sailing.
While an orderly process may be important, and avoiding a ?train wreck?
desirable, I believe there are other important issues to consider in
designing the budget process. As I have testified before, the budget
represents the decisions made about a large number of often conflicting
objectives that citizens want the government to address. We should not be
surprised that it generates controversy. As BEA expires and you move from
fighting current deficits to prudent management of surpluses, you face a
wealth of options and choices. I appreciate the invitation to talk about
some of these today. Some of these points are discussed more fully in our
recent BEA compliance report 2 that we prepared at your request, Mr.
Chairman.
In the past, we have suggested four broad principles or criteria for a
budget process. 3 A process should
provide information about the long- term impact of decisions, both macro-
linking fiscal policy to the long- term economic outlook- and micro-
providing recognition of the long- term spending implications of government
commitments;
provide information and be structured to focus on important macro
tradeoffs- e. g., between investment and consumption;
2 Budget Issues: Budget Enforcement Compliance Report (GAO- 01- 777, June
15, 2001). 3 For a fuller discussion of these criteria, see Budget Process:
Evolution and Challenges (GAO/ T- AIMD- 96- 129, July 11, 1996), Budget
Process: History and Future Directions (GAO/ T- AIMD- 95- 214, July 13,
1995), and Budget Process: Comments on H. R. 853 (GAO/ T- AIMD- 99- 188, May
12, 1999). Principles for a
Budget Process
Page 3 GAO- 01- 991T
provide information necessary to make informed trade- offs between
missions (or national needs) and between the different policy tools of
government (such as tax provisions, grants, and credit programs); and
be enforceable, provide for control and accountability, and be
transparent, using clear, consistent definitions.
The lack of adherence to the original BEA spending constraints in recent
years, the nearing expiration of BEA, and the projection of continued and
large surpluses in the coming years suggest that now may be an opportune
time to think about the direction and purpose of our nation?s fiscal policy.
In a time of actual and projected surpluses, the goal of zero deficit no
longer applies. Rather, discussion shifts toward how to allocate surpluses
among debt reduction, spending increases, and tax cuts. Only then can limits
on subcategories of spending be set. Will the entire social security surplus
be ?saved?? What about the Medicare Part A surplus? In our work on other
countries that also have faced the challenge of setting fiscal policy in
times of surplus, we found that as part of a broad fiscal policy framework
some countries adopted fiscal targets such as debt- to- gross domestic
product (GDP) ratios to serve as guides for decision- making.
Complicating the discussion on formulating fiscal policy in a time of
surplus is the fact that the long- term picture is not so good. Despite
current projections that show surpluses continuing over the 10- year budget
window, our long- term budget simulations show a resumption of significant
deficits emerging after the anticipated demographic tidal wave of population
aging hits. These demographic trends serve to emphasize the importance of
the first principle cited above- the need to bring a longterm perspective to
bear on budget debates. Keeping in mind these principles and concerns, a
number of alternatives appear promising.
There is a broad consensus among observers and analysts who focus on the
budget both that BEA has constrained spending and that continuation of some
restraint is necessary even with the advent of actual and projected
surpluses. Discussions on the future of the budget process have primarily
focused on revamping the current budget process rather than establishing a
new one from scratch.
Where discussion has moved beyond a general call for continued restraint to
specific control devices, the ones most frequently discussed are (1)
extending the discretionary spending caps, (2) extending the PAYGO
mechanism, and (3) creating a trigger device or a set of rules specifically
Alternatives for
Improving the Budget Process
Page 4 GAO- 01- 991T
designed to deal with the uncertainty of budget projections. A new budget
process framework could encompass any or all of these instruments.
BEA distinguished between spending controlled by the appropriations
process-? discretionary spending?- and that which flowed directly from
authorizing legislation provisions of law-? direct spending,? sometimes
called ?mandatory.? Caps were placed on discretionary spending- and the
Congress? compliance with the caps was relatively easy to measure because
discretionary spending totals flow directly from legislative actions (i. e.,
appropriations laws). There is broad consensus that, although the caps have
been adjusted, they have served to constrain appropriations. This consensus
combined with the belief that some restraints should be continued has led
many to propose that some form of cap structure be continued as a way of
limiting discretionary appropriations. However, the actions in the last 2
years have also led many to note that caps can only work if they are
realistic; while caps may be seen as tighter than some would like, they are
not likely to bind if they are seen as totally unreasonable given current
conditions.
Further, some have proposed that any extension of BEA- type caps be limited
to caps on budget authority. Outlays are controlled by and flow from budget
authority- although at different rates depending on the nature of the
programs. Some argue that the existence of both budget authority and outlay
caps has encouraged provisions such as ?delayed
obligations? to be adopted not for programmatic reasons but as a way of
juggling the two caps. The existence of two caps may also skew authority
from rapid spendout to slower spendout programs, thus pushing more outlays
to the future and creating problems in complying with outlay caps in later
years. Extending only the budget authority cap would eliminate the incentive
for such actions and focus decisions on that which the Congress is intended
to control- budget authority, which itself controls outlays. This would be
consistent with the original design of BEA.
Eliminating the outlay cap would raise several issues- chief among them
being how to address the control of transportation programs for which no
budget authority cap currently exists, and the use of advance appropriations
to skirt budget authority caps. However, agreements about these issues could
be reached- this is not a case where implementation difficulties need derail
an idea. For example, the fiscal year 2002 budget proposes a revision to the
scorekeeping rule on advance appropriations so that generally they would be
scored in the year of enactment. If the Budget Committees and CBO agree,
this change could eliminate the practice of Extending Caps on
Discretionary Spending
Page 5 GAO- 01- 991T
using advance appropriations to skirt the caps. The obvious advantage to
focusing decisions on budget authority rather than outlays is that the
Congress would not spend its time trying to control that which by design is
the result of its budget authority decisions- the timing of outlays.
There are other issues in the design of any new caps. For example, for how
long should caps be established? What categories should be established
within or in lieu of an overall cap? While the original BEA envisioned three
categories (Defense, International Affairs, and Domestic), over time
categories were combined and new categories were created. At one time or
another caps for Nondefense, Violent Crime Reduction, Highways, Mass
Transit, and Conservation spending existed- many with different expiration
dates. Should these caps be ceilings, or should they- as is the case for
Highways and Conservation- provide for ?guaranteed?
levels of funding? The selection of categories- and the design of the
applicable caps- is not trivial. Categories define the range of what is
permissible. By design they limit trade- offs and so constrain both the
Congress and the President.
Because caps are phrased in specific dollar amounts, it is important to
address the question of when and for what reasons the caps should be
adjusted. This is critical for making the caps realistic. For example,
without some provision for emergencies, no caps can be successful. At the
same time, there appears to be some connection between how realistic the
caps are and how flexible the definition of emergency is. As discussed in
last year?s compliance report, the amount and range of spending considered
?emergency? has grown in recent years. There have been a number of
approaches suggested to balance the need to respond to emergencies and the
desire to avoid making the ?emergency? label an easy way to raise caps. In
the budget resolution for fiscal year 2001 [H. Con. Res. 290], the Congress
said it would limit emergencies to items meeting five criteria: (1)
necessary, essential, or vital (not merely useful or beneficial), (2)
sudden, quickly coming into being, and not building up over time, (3) an
urgent, pressing, and compelling need requiring immediate action, (4)
unforeseen, unpredictable, and unanticipated, and (5) not permanent,
temporary in nature. The resolution further required any proposal for
emergency spending that did not meet all the criteria to be accompanied by a
statement of justification explaining why the requirement should be accorded
emergency status. The fact that this provision was ignored during debates on
fiscal year 2001 appropriations bills emphasizes that no procedural hurdle
can succeed without the will of the Congress. Others have proposed providing
for more emergency spending- either in the form of a reserve or in a greater
appropriation for
Page 6 GAO- 01- 991T
the Federal Emergency Management Agency (FEMA)- under any caps. If such an
approach were to be taken, the amounts for either the reserve or the FEMA
disaster relief account would need to be included when determining the level
of the caps. Some have proposed using a 5- or 10year rolling average of
disaster/ emergency spending as the appropriate reserve amount. Adjustments
to the caps would be limited to spending over and above that reserve or
appropriated level for extraordinary circumstances. Alternatively, with
additional up- front appropriations or a reserve, emergency spending
adjustments could be disallowed. 4
Even with this kind of provision only the commitment of the Congress and the
President can make any limit on cap adjustments for emergencies work. States
have used this reserve concept for emergencies, and their experiences
indicate that criteria for using emergency reserve funds may be useful in
controlling emergency spending. 5 Agreements over the use of the reserve
would also need to be achieved at the federal level.
This discussion is not exhaustive. Other issues would come up in extending
BEA. Previously, we have reported on two issues- the scoring of operating
leases and the expansion of user fees as offsets to discretionary spending;
because I think they need to be considered, let me touch on them briefly.
We have previously reported that existing scoring rules favor leasing when
compared to the cost of various other methods of acquiring assets. 6
Currently, for asset purchases, budget authority for the entire acquisition
cost must be recorded in the budget up front, in the year that the asset
acquisition is approved. In contrast, the scorekeeping rules for operating
leases often require that only the current year?s lease costs be recognized
and recorded in the budget. This makes the operating lease appear less
costly from an annual budgetary perspective, and uses up less budget
authority under the cap. Alternative scorekeeping rules could recognize
4 The administration?s fiscal year 2002 budget submission included a
proposal to set aside a reserve for emergency needs in the annual budget and
appropriations process, arguing that this would limit the need for emergency
supplementals to extremely rare events.
5 Budgeting for Emergencies: State Practices and Federal Implications (GAO/
AIMD- 99- 250, Sept. 30, 1999). 6 Budget Issues: Budget Scorekeeping for
Acquisition of Federal Buildings (GAO/ T- AIMD- 94- 189, Sept. 20, 1994).
Miscellaneous
Discretionary Challenges: Leases and User Fees
Page 7 GAO- 01- 991T
that many operating leases are used for long- term needs and should be
treated on the same basis as purchases. This would entail scoring up front
the present value of lease payments covering the same period used to analyze
ownership options. The caps could be adjusted appropriately to accommodate
this change.
Many believe that one unfortunate side effect of the structure of the BEA
has been an incentive to create revenues that can be categorized as ?user
fees? and so offset discretionary spending- rather than be counted on the
PAYGO scorecard. The 1967 President?s Commission on Budget Concepts
recommended that receipts from activities that were essentially governmental
in nature, including regulation and general taxation, be reported as
receipts, and that receipts from business- type activities ?offset
to the expenditures to which they relate.? However, these distinctions have
been blurred in practice. Ambiguous classifications combined with budget
rules that make certain designs most advantageous has led to a situation in
which there is pressure to treat fees from the public as offsets to
appropriations under BEA caps, regardless of whether the underlying federal
activity is business or governmental in nature. Consideration should be
given to whether it is possible to come up with and apply consistent
standards- especially if the discretionary caps are to be redesigned. The
administration has stated that it plans to monitor and review the
classification of user fees and other types of collections.
The PAYGO requirement prevented legislation that lowered revenue, created
new mandatory programs, or otherwise prevented direct spending from
increasing the deficit unless offset by other legislative actions. As long
as the unified budget was in deficit, the provisions of PAYGO- and its
application- were clear. The shift to surplus raised questions about whether
the prohibition on increasing the deficit also applied to reducing the
surplus. Although the Congress and the executive branch have both concluded
that PAYGO does apply in such a situation, any extension should eliminate
potential ambiguity in the future.
This year, the administration has proposed- albeit implicitly- special
treatment for a tax cut. The budget states that the President?s tax plan and
Medicare reforms are fully financed by the surplus and that any other
spending or tax legislation would need to be offset by reductions in
spending or increases in receipts. It is possible that in a time of budget
surplus, the Congress might wish to modify PAYGO to permit increased direct
spending or lower revenues as long as debt held by the public is planned to
be reduced by some set percentage or dollar amount. Such a Extending and
Refining PAYGO
Page 8 GAO- 01- 991T
provision might prevent PAYGO from becoming as unrealistic as overly tight
caps on discretionary spending. However, the design of such a provision
would be important- how would a debt reduction requirement be specified? How
would it be measured? What should be the relationship between the amount of
debt reduction required and the amount of surplus reduction (i. e., tax cut
or direct spending increase) permitted? What, if any, relationship should
there be between this calculation and the discretionary caps?
While PAYGO constrained the creation or legislative expansion of direct
spending programs and tax cuts, it accepted the existing provisions of law
as given. It was not designed to trigger- and it did not trigger- any
examination of ?the base.? Cost increases in existing mandatory programs are
exempt from control under PAYGO and could be ignored. However, constraining
changes that increase the cost of entitlements and mandatories is not
enough. Our long- term budget simulations show that as more and more of the
baby boom generation retires, spending for Social Security, Medicare, and
Medicaid will demand correspondingly larger shares of federal revenues. The
growth in these programs will increasingly restrict budgetary flexibility.
Even if the Social Security surpluses are saved and used for debt reduction,
unified deficits are projected to emerge in about two decades, and by 2030
Social Security, Medicare, and Medicaid would require more than three-
fourths of federal revenues. 7
Previously we suggested some sort of ?lookback? procedure to prompt a
reexamination of ?the base.? Under such a process, the Congress could
specify spending targets for PAYGO programs for several years. The President
could be required to report in his budget whether these targets either had
been exceeded in the prior year or were likely to be exceeded in the current
or budget years. He could then be required to recommend whether any or all
of this overage should be recouped- and if so, to propose a way to do so.
The Congress could be required to act on the President?s proposal.
While the current budget process contains a similar point of order against
worsening the financial condition of the Social Security trust funds, 8 it
7 Long- Term Budget Issues: Moving From Balancing the Budget to Balancing
Fiscal Risk (GAO- 01- 385T, Feb. 6, 2001). 8 2 U. S. C. 632 (i), and
Medicare Reform: Issues Associated With General Revenue Financing (GAO/ T-
AIMD- 00- 126, Mar. 27, 2000).
Page 9 GAO- 01- 991T
would be possible to link ?tripwires? or triggers to measures related to
overall budgetary flexibility or to specific program measures. For example,
if the Congress were concerned about declining budgetary flexibility, it
could design a tripwire tied to the share of the budget devoted to mandatory
spending or to the share devoted to a major program.
Other variations of this type of tripwire approach have been suggested. The
1999 Breaux- Frist proposal (S. 1895) for structural and substantive changes
to Medicare financing contained a new concept for measuring
?programmatic insolvency? and required congressional approval of additional
financing if that point was reached. Other specified actions could be
coupled with reaching a tripwire, such as requiring the Congress or the
President to propose alternatives to address reforms or, by using the
congressional budget process, requiring the Congress to deal with
unanticipated cost growth beyond a specified tripwire by establishing a
point of order against a budget resolution with a spending path exceeding
the specified amount. One example of a threshold might be the percentage of
GDP devoted to Medicare. The President would be brought into the process as
it progressed because changes to deal with the cost growth would require
enactment of a law.
In previous reports we have argued that the nation?s economic future depends
in large part upon today?s budget and investment decisions. 9 In fact, in
recent years there has been increased recognition of the long- term costs of
Social Security and Medicare. 10
While these are the largest and most important long- term commitments- and
the ones that drive the long- term outlook- they are not the only ones in
the budget. Even those programs too small to drive the long- term outlook
affect future budgetary flexibility. For the Congress, the President, and
the public to make informed decisions about these other programs, it is
important to understand their long- term cost implications.
9 See Budget Process: Evolution and Challenges (GAO/ T- AIMD- 96- 129, July
11, 1996) and The Deficit and the Economy: An Update of Long- Term
Simulations (GAO/ AIMD/ OCE- 95- 119, Apr. 26, 1995), among others. 10
Budget of the United States Government, Fiscal Year 2002, OMB, Apr. 9, 2001;
The Budget and Economic Outlook: Fiscal Years 2002- 2011, CBO, Jan. 2001;
GAO- 01- 385T, Feb. 6, 2001; and Medicare: Higher Expected Spending and Call
for New Benefit Underscore Need for Meaningful Reform (GAO- 01- 539T, Mar.
22, 2001). Improving the
Recognition of LongTerm Commitments
Page 10 GAO- 01- 991T
While the budget was not designed to and does not provide complete
information on long- term cost implications stemming from some of the
government?s commitments when they are made, progress can be made on this
front. The enactment of the Federal Credit Reform Act in 1990 represented a
step toward improving both the recognition of long- term costs and the
ability to compare different policy tools. With this law, the Congress and
the executive branch changed budgeting for loan and loan guarantee programs.
Prior to the Credit Reform Act, loan guarantees looked ?free? in the budget.
Direct loans looked like grant programs because the budget ignored loan
repayments. The shift to accrual budgeting for subsidy costs permitted
comparison of the costs of credit programs both to each other and to
spending programs in the budget.
Information should be more easily available to the Congress and the
President about the long- term cost implications both of existing programs
and new proposals. In 1997 we reported that the current cash- based budget
generally provides incomplete information on the costs of federal insurance
programs. 11 The ultimate costs to the federal government may not be
apparent up front because of time lags between the extension of the
insurance, the receipt of premiums, and the payment of claims. While there
are significant estimation and implementation challenges, accrualbased
budgeting has the potential to improve budgetary information and incentives
for these programs by providing more accurate and timely recognition of the
government?s costs and improving the information and incentives for managing
insurance costs. This concept was proposed in the Comprehensive Budget
Process and Reform Act of 1999 (H. R. 853), which would have shifted
budgetary treatment of federal insurance programs from a cash basis to an
accrual basis.
There are other commitments for which the cash- and obligation- based budget
does not adequately represent the extent of the federal government?s
commitment. These include employee pension programs, retiree health
programs, and environmental cleanup costs. While there are various
analytical and implementation challenges to including these costs in budget
totals, more could be done to provide information on the longterm cost
implications of these programs to the Congress, the President, and the
interested public. At the request of this Committee, we are continuing to
address this issue.
11 Budget Issues: Budgeting for Federal Insurance Programs (GAO/ AIMD- 97-
16, Sept. 30, 1997).
Page 11 GAO- 01- 991T
As the budgeting horizon expands, so does the certainty of error. Few
forecasters would suggest that 10- year projections are anything but that-
projections of what the world would look like if it continued on a line from
today. And long- term simulations are useful to provide insight as to
direction and order of magnitude of certain trends- not as forecasts.
Nevertheless, budgeting requires forecasts and projections. Baseline
projections are necessary for measuring and comparing proposed changes.
Former Congressional Budget Office (CBO) Director Rudy Penner suggested that
5- year and 10- year projections are useful for and should be used for
different purposes: 5- year projections for indicating the overall fiscal
health of the nation, and 10- year projections for scorekeeping and
preventing gaming of the timing of costs.
No 10- year projection is likely to be entirely correct; the question
confronting fiscal policymakers is how to deal with the risk that a
projection is materially wrong. This year some commentators and Members of
the Congress have suggested dealing with this risk by using triggers.
Triggers were part of both Gramm- Rudman- Hollings (GRH) and BEA. The GRH
triggers were tied to deficit results and generally regarded as a failure-
they were evaded or, when deficits continued to exceed the targets, the
targets were changed. BEA triggers have been tied to congressional action
rather than to deficit results; sequesters have rarely been triggered- and
those were very small. This year the discussion of triggers has been tied
specifically to the tax debate and to whether the size of the tax cut in
future years should be linked to budget results in those years. There could
be several variations on this trigger: actual surplus results, actual
revenue results (this with the intent of avoiding a situation in which
spending increases can derail a tax cut), and actual debt results. There is
little consensus on the effectiveness of any triggers.
Although the debate about triggers has been tied to the tax debate in 2001,
there is no inherent reason to limit the discussion to taxes. Some might
wish to consider triggers that would cause decisionmakers to make proposals
to address fiscal results that exceed some specific target, such as debt or
spending as a share of GDP.
Former CBO Director Robert Reischauer suggested another way of dealing with
the fact that forecasts/ projections become less certain the further they go
into the future. Under his proposal, a declining percentage of any projected
surplus would be available- either for tax cuts or for spending increases.
Specifically, 80 percent of the surplus would be available to legislators in
years 1 and 2, 70 percent in years 3 and 4, 60 percent in years 5 and 6,
until reaching the 40- percent level in years 9 and 10. The Dealing With the
Uncertainty of Projections
Page 12 GAO- 01- 991T
consequence of not adhering to these limits would be an across- the- board
sequester. When a new Congress convenes, it would be given a new budget
allowance to spend based on a new set of surplus projections.
Others have suggested that mechanisms such as a joint budget resolution and/
or an automatic continuing resolution could avert the year- end disruption
caused by an inability to reach agreement on funding the government.
Biennial budgeting is also sometimes suggested as a better way to budget and
to provide agencies more certainty in funding over 2 years. Let me turn now
to these ideas.
Since agreement on overall budget targets can set the context for a
productive budget debate, some have suggested that requiring the President?s
signature on budget resolutions would facilitate the debate within such a
framework. Proposals to replace the Concurrent Resolution with a Joint
Resolution should be considered in the light of what the budget resolution
represents. Prior to the 1974 act only the President had a budget- that is,
a comprehensive statement of the level of revenues and spending and the
allocation of that spending across ?national needs? or federal mission
areas. Requiring the President to sign the budget resolution means it would
not be a statement of congressional priorities. Would such a change reduce
the Congress? ability to develop its own budget and so represent a shift of
power from the Congress to the President? Whose hand would it strengthen? If
it is really to reduce later disagreement, would it merely take much longer
to get a budget resolution than it does today? It could be argued that under
BEA the President and the Congress have- at times- reached politically
binding agreements without a joint budget resolution.
The periodic experience of government ?shutdowns?- or partial shutdowns when
appropriations bills have not been enacted has led to proposals for an
automatic continuing resolution. The automatic continuing resolution,
however, is an idea for which the details are critically important.
Depending on the detailed structure of such a continuing resolution, the
incentive for policymakers- some in the Congress and the President- to
negotiate seriously and reach agreement may be lessened. What about someone
for whom the ?default position? specified in the automatic continuing
resolution is preferable than the apparent likely outcome? If the goal of
the automatic continuing resolution is to provide a little more time for
resolving issues, it could be designed to permit the incurrence of
obligations to avoid a funding gap, but not the outlay of funds to liquidate
the new obligations. This would Other Ideas Proposed
to Smooth the Process
Page 13 GAO- 01- 991T
allow agencies to continue operations for a period while the Congress
completes appropriations actions.
Finally, you asked me to discuss proposals for biennial budgeting. Some have
suggested that changing the appropriations cycle from annual to biennial
could (1) provide more focused time for congressional oversight of programs,
(2) shift the allocation of agency officials? time from the preparation of
budgets to improved financial management and analysis of program
effectiveness, and (3) enhance agencies? abilities to manage their
operations by providing more certainty in funding over 2 years. Given the
regularity with which proposals for biennial budgeting are made, I believe
that at least some will consider the upcoming necessity to decide whether to
extend BEA as an opportunity to again propose biennial budgeting.
Whether a biennial cycle offers the benefits sought will depend heavily on
the ability of the Congress and the President to reach agreement on how to
respond to uncertainties inherent in a longer forecasting period, for there
will always be uncertainties. How often will the Congress and the President
feel the need to reopen the budget and/ or change funding levels?
Budgeting always involves forecasting, which in itself is uncertain, and the
longer the period of the forecast, the greater the uncertainty. Our work has
shown that increased difficulty in forecasting was one of the primary
reasons states gave for shifting from biennial to annual cycles. 12 The
budget is highly sensitive to the economy. Economic changes during a
biennium would most likely prompt the Congress to revisit its decisions and
reopen budget agreements. Among the issues that would need to be worked out
if the Congress moves to a biennial budget cycle are how to update the CBO
forecast and baseline against which legislative action is scored and how to
deal with unexpected events. The baseline is important because CBO scores
legislation based on the economic assumptions in effect at the time of the
budget resolution. Even under an annual system there are years when this
practice presents problems: in 1990 the economic slowdown was evident during
the year, but consistent practice meant that bills reported in compliance
with reconciliation instructions were scored based on the assumptions in the
budget resolution rather than updated assumptions. If budget resolutions
were biennial, this problem of
12 Since the mid- 1960s, 18 states have changed their budget cycles: 11 from
biennial to annual, 2 from annual to mixed, 4 from annual to biennial, and1
from biennial to annual and back to biennial.
Page 14 GAO- 01- 991T
outdated assumptions would be greater- some sort of update in the ?offyear?
likely would be necessary. In any consideration of a biennial budget, it is
important to recognize that even with annual budgets, the Congress already
has provided agencies with multiyear funding to permit improved planning and
management. As you know, it is not necessary to change the frequency of
decisions in order to change the length of time funds are available. Nearly
two- thirds of the budget is for mandatory programs and entitlements on
which decisions are not made annually. Even the remaining portion that is on
an annual appropriations cycle is not composed entirely of 1- year
appropriations that expire on September 30 of each year. The Congress
routinely provides multiyear or no- year appropriations when it seems to
make sense to do so. Thus, to the extent that biennial budgeting is proposed
as a way to ease a budget execution problem, the Congress has shown itself
willing and able to meet that need under the current annual cycle.
If BEA is extended in conjunction with biennial budgeting, a whole host of
technical issues needs to be considered. Would biennial budgeting change the
timing of the BEA- required sequestration report? How would sequestrations
be applied to the 2 years in the biennium and when would they occur? For
example, if annual caps are continued and are exceeded in the second year of
the biennium, when would the Presidential Order causing the sequestration be
issued? Would the sequestration affect both years of the biennium? Would
forecasts and baselines be updated during the biennium? These are just of
few of the many questions that would need to be resolved.
Regardless of the potential benefits, the decision on biennial budgeting
will depend on how the Congress chooses to exercise its constitutional
authority over appropriations and its oversight functions. We have long
advocated regular and rigorous congressional oversight of federal programs.
Annual enacted appropriations have long been a basic means of exerting and
enforcing congressional policy. Oversight has often been conducted in the
context of agency requests for funds. A 2- year appropriation cycle would
change- and could lessen- congressional influence over program and spending
matters since the process would afford fewer scheduled opportunities to
affect agency programs and budget.
Biennial budgeting would bring neither the end of congressional control nor
the guarantee of improved oversight. It would require a change in the nature
of that control. If the Congress decides to proceed with a change to
Page 15 GAO- 01- 991T
a biennial budget cycle- including a biennial appropriations cycle- careful
thought will need to be given to implementation issues.
To affect decision- making, the fiscal goals sought through a budget process
must be accepted as legitimate. For many years the goal of ?zero
deficit?- or the norm of budget balance- was accepted as the right goal for
the budget process. In the absence of the zero deficit goal, policymakers
need an overall framework upon which a process and any targets can be based.
Goals may be framed in terms of debt reduction or surpluses to be saved. In
any case, compliance with budget process rules, in both form and spirit, is
more likely if end goals, interim targets, and enforcement boundaries are
both accepted and realistic.
Enforcement is more successful when it is tied to actions controlled by the
Congress and the President. Both the BEA spending caps and the PAYGO
enforcement rules were designed to hold the Congress and the President
accountable for the costs of the laws enacted each session- not for costs
that could be attributed to economic changes or other factors.
Today, the Congress and the President face a different budgetary situation
than in the past few decades. The current budget challenge is not to achieve
a balanced unified budget. Rather, budgeting today is done in the context of
projections for continued and growing surpluses followed over the longer
term by demography- driven deficits. What process will enable policymakers
to deal with the near term without ignoring the long term? At the same time,
the challenges for any budget process are the same: What process will enable
policymakers to make informed decisions about both fiscal policy and the
allocation of resources within the budget?
Extending the current BEA without setting realistic caps and addressing
existing mandatory programs is unlikely to be successful for the long term.
The original BEA employed limited actions in aiming for a balanced budget.
It left untouched those programs- direct spending and tax legislation-
already in existence. Going forward with new challenges, we believe that a
new process that prompts the Congress to exercise more foresight in dealing
with long- term issues is needed. The budget process appropriate for the
early 21st century will have to exist as part of a broader framework for
thinking about near- and long- term fiscal goals. Conclusion
Page 16 GAO- 01- 991T
Mr. Chairman, this concludes my statement. I would be happy to answer any
questions that you or the Members of the Committee may have.
For future contacts regarding this testimony, please call me at (202)
5129142 or Christine Bonham at (202) 512- 9576. Melinda Bowman also made key
contributions to this testimony. Contacts and
Acknowledgments
(450061)
*** End of document. ***