An synopsis of this essay
appears in the June 1991 issue of the Region.
The views expressed in this annual report are solely
those of the authors; they are not intended to represent a formal position
of the Federal Reserve System.
The federal budget mess just won't go away. Despite the
discipline of Gramm-Rudman-Hollings (GRH), the government is running ever
larger budget deficits, making poor decisions and spending an inordinate
amount of time on the process. In 1990 the budget deficit reached $220
billion. This year, after the torturous passage of a package of expenditure
cuts and tax increases, it's projected to exceed $300 billion. Yet the
original GRH deficit target for fiscal 1991 was zero. Voters are concerned;
Congress is concerned; the administration is concerned. Although there
is widespread agreement that something is seriously wrong with the budget
process, there is less agreement about what should be done about it. In
response to this concern, many proposals for reform have been suggested.
Yet they miss the mark: they fail to address the problems inherent in
the budget process and are not based on sound economic principles.
We provide a conceptual framework for budget policy. Based on
this framework, we propose that the federal government change accounting
practices, institute rules on debt issue and impose enforcement
mechanisms. Our proposal will produce budgets that are balanced
over time in an appropriate rather than an arbitrary sense. Our
proposal also will help inform the decision makers and the public
about their policy options and the financial consequences of those
options. Of course, our proposal will not cure all the ills in the
budget process. Hard choices will still have to be made.
But first a bit of history. Since the late 1960s, the federal
government has consistently run large deficits. These large deficits
have led to a dramatic increase in the federal debt as a percentage
of GNP in the 1980s. Interest payments to service this debt have
absorbed an increasing proportion of our national product.
The debilitating consequences of a growing federal debt are well-
known. Large interest payments leave us with less to pay for education,
highways, national defense and a variety of useful causes. A growing
federal debt tends to raise interest rates and to increase pressure
on the Federal Reserve System to follow inflationary policies. This
litany of ills may be familiar; nonetheless, it is alarming.
To make things worse, even though total non-defense expenditures
have been allowed to grow at double-digit rates, too little money
has been allocated for capital projects, such as highways, airports
and sewer systems. Federal capital spending in fiscal 1991 is projected
to be roughly 2.2 percent of GNP, down from 4.4 percent of GNP in
the early 1960s. Construction expenditures, which are an important
component of capital spending, have declined over the postwar period.
Surely we can do better than to leave our children decayed highways,
crumbled bridges and antiquated sewer systems, with little or no
ability to repair or replace any of them because of the enormous
tax bills coming due for services consumed before they were ever
We are not the first to offer solutions: The Gramm-Rudman-Hollings
Act(GRH) and the reforms of last fall were attempts to respond to
these problems. While these attempts to reach a balanced budget
are laudable, we think our proposal is better. Specifically, we
propose that the federal government adopt the following:
- Record transactions on an accrual basis and maintain separate operating
and capital budgets,
- Require that the combined operating budget in the current and subsequent
fiscal years be balanced, and establish overall limits on capital
- Institute enforcement mechanisms based on performance to ensure
that the rules are being followed,
- Set up rainy day funds to meet contingencies.
We would, of course, not be averse to an escape clause suspending
the rules in the event of a war or national emergency. But we think
that such suspension should require a supermajority of votes in
Congress and the assent of the president.
Why do we need these or any rules at all? Why not rely on policymakers
to make good decisions? The problem is that the policymaking process
is fundamentally biased against the future. Without rules to constrain
policy decisions, we will continue to have bad outcomes. So the
question is not whether to have rules, it is which rules to have.
The current rules do not address the bias, nor are they based on
sound economic principles, In the rest of the essay, we explain
the bias in policymaking, show that the current set of rules is
inadequate and argue that our rules will yield good outcomes.
The Need for Policy Rules
Our political System encourages elected officials to adopt policies
that are biased against the future. They are biased, because they
are not what voters ideally would like. Voters tend to have a long-term
view on policy issues. They care about the outcomes of policy decisions
not only over their own lifetimes but over the lifetimes of their
descendants. Elected officials, however, tend to adopt a short-term
view on policy issues.
But Why should this be so? If Voters could keep themselves fully
informed, closely monitor decisions and understand the effects of
alternative policies, surely they would force their elected officials
to act in the public's interest. Voters would boot out officials
who acted badly or develop institutional arrangements that set up
proper incentives. But voters have neither the ability nor the incentive
to get the information necessary to do these things.
It is especially difficult for voters to decide whether policy-
makers are making wise choices on decisions that will yield benefits
or costs several years into the future. So even though voters care
about the future, they weight the current results heavily. They
rationally reward policymakers who make decisions yielding large
current benefits even if some of those decisions impose future costs.
In order to get reelected, policymakers have incentives to make
decisions that are systematically biased against the future. Thus,
it is not especially helpful to argue that if incumbent policymakers
were replaced, wiser policies would be chosen. The problem is that
the system encourages policymakers to adopt the short-term view.
This bias results in decisions weighted toward current rather
then future consumption. In particular, these decisions mean too
much deficit financing, too little capital investment and proclivity
to put off until tomorrow projects which should be undertaken today.
The immediate benefit to elected officials is easily seen. Deficit
financing shifts the tax burden for current consumption into the
future. This allows more consumption now but less in the future
as the taxes are paid. Less capital investment frees up resources
and allows more consumption now, but since fewer resources have
been invested, there is a smaller supply of goods available for
future consumption. Elected officials can also make more resources
available for current consumption by putting off projects such as
maintenance of the infrastructure or closure of insolvent thrifts.
But the problem is that such actions leave fewer resources available
in the future as the real costs of the necessary, but delayed, expenditures
This policymaking bias against the future is not unique to the
government, however. Corporate stockholders face many of the same
problems as voters in deciding whether corporate managers are acting
wisely. Corporate managers act on behalf of stockholders yet they
have different interests. Stockholders have incomplete knowledge
about the managers' decisions and the consequences of the managers'
actions. These are features of the so-called AGENCY PROBLEM.
Furthermore, no single stockholder has much incentive to monitor
management's actions. Stockholdings are typically dispersed among
many individuals. Each stockholder has an incentive to let other
stockholders monitor the firm. As a result, typically there is less
monitoring of firms' actions than would occur if the stockholders
could act jointly. This is known as the FREE RIDER PROBLEM.
Voters also face the agency and free rider problems. It is difficult
to know the consequences of policymakers' actions and each voter
has an incentive to free ride on other voters. The agency and free
rider problems can be mitigated but not entirely solved. If stockholders
cannot entirely solve these problems, certainly voters cannot be
expected to solve them completely either, since compared to corporate
decisions, government policy decisions are aimed at more diverse
objectives and the responsibility for them is more diffuse.
How stockholders attempt to manage the agency and free rider problems
suggest ways voters might deal with them, though. Publicly traded
corporations, for example, are required to adopt standard accounting
practices (known as Generally Accepted Accounting Principles) to
inhibit inaccurate presentation of information. Corporate charters
also limit management's discretion. We think these ideas can and
should be used to design better policy procedures for the federal
government. Corporations also adopt a variety of other practices
to align managers' and stockholders' interests, including incentive
contracts for managers, hostile takeovers and so on. It is not apparent
how, or even whether, these other practices can be transferred to
the government, so here we will stick to practices that seem readily
The agency and free rider problems point to a need for rules to
limit policymakers' discretion. To understand why we are proposing
new rules it is important to understand what's wrong with the old
The Problems With the Old Rules
The Gramm-Rudman-Hollings Process
Prior to the fall 1990 reforms, the budget process was designed
to work roughly as follows: Early in January the president would
submit to Congress a budget for the fiscal year beginning October
1. By the time Congress would adjourn in the summer it would pass
bills and a budget resolution specifying amounts to be appropriated
to discretionary spending programs and changes to be made to rules
for entitlement programs and taxes. In mid-August the Office of
Management and Budget (OMB) would determine whether the projected
deficit for the upcoming fiscal year exceeded the GRH target by
more than $10 billion. If it did, automatic spending cuts would
be made according to a statutory formula to ensure that the projected
deficit met the target. Then at some point Congress would raise
the debt ceiling so that there was authority to issue additional
debt to finance the projected deficit.
The GRH procedure was supposed to lead to balanced budgets and
to eliminate the government's deficit financing bias by the use
of rules. The procedure did not come close to achieving its goal
and the original GRH targets had to be revised several times. In
addition, the process led to other problems.
Rather than curbing the government's bias to excessively discount
the future, the GRH process fed it. By focusing on the projected
budget deficit in the approaching fiscal year, the process encouraged
deficit financing, discouraged capital spending and made it more
attractive to delay necessary expenditures.
Paradoxically, the GRH process made deficit financing easier by
encouraging the substitution of gimmicks for real actions. The gimmicks
included the exchanging of assets, time-shifting of payments, movement
of expenditures off the budget and use of unrealistic economic and
technical assumptions. Selling government assets, such as loans,
for cash resulted in budget savings, even though all that occurred
was the exchange of one asset for another one of comparable value.
Time-shifting created one-time fictitious savings by taking payments
scheduled for the approaching fiscal year and moving them into the
current fiscal year where they were simply spilt milk, or into the
following fiscal year where they were not yet subject to a sequester.
For example, the government achieved savings in an approaching fiscal
year by taking payments such as employees' pay and farm subsidies
due out in late September and delaying them until early October
of the following fiscal year. Then without retribution, the government
was able to switch the payments back once the new fiscal year was
entered. Movement of expenditures off-budget, such as was done for
the U.S. Postal Service and the Resolution Trust Corporation (RTC),
produced budget savings by a stroke of the pen without doing anything
real. Finally, projected deficits were reduced by use of optimistic
economic and technical assumptions that overestimated tax revenue,
underestimated interest expense and understated the costs of existing
Just how extensively were these gimmicks used? Robert Reischauer,
director of the Congressional Budget Office (CBO), notes that the
amount of permanent deficit reduction enacted in the GRH period
averaged a bit less than in the pre-GRH period of same duration:
"What is different about the two periods is the reliance on one-time
savings that became a feature of the GRH period. ...In the pre-GRH
period, these gimmicks occurred so infrequently that CBO did not
keep systematic track of them. In the GRH era, fully half of the
apparent deficit reduction has been achieved by such devices" (Reischauer,
1990). Reischauer also notes that under GRH the amount of budget
savings in the president's budget attributable to overly optimistic
economic and technical assumptions more than tripled.
The GRH process, combined with existing accounting practices,
also made capital spending highly susceptible to the knife. Existing
accounting practices treat spending on capital projects, such as
bridges, no differently than spending on current consumption, such
as legal counseling. Yet the two are fundamentally different. A
capital project provides services in the current year as well as
in the future, while current consumption provides services only
in the year in which they were purchased. As a result of this difference,
spending on a capital project can be preferable to spending on current
consumption yet can return less in benefits in the current year
since it also provides benefits in future years. When GRH forced
budget cuts in an approaching year, the existing accounting practice
led government officials to believe that a dollar cut from capital
spending would cause less immediate loss in benefits than would
a dollar cut from current consumption. To understand better why
this is so, consider the following example.
Suppose the government has decided to build a bridge at a cost
of $30 million which is expected to provide services over 30 years
worth $6.4 million per year. Suppose the interest rate is 10 percent
so that the present-value of these services is $60 million. With
the present value of the benefits at double the cost, the bridge
obviously should be built. But under the government's accounting
system, dropping the project saves $30 million in costs and sacrifices
only $6.4 million in benefits. In contrast, a $30 million cut in
spending on a less attractive current consumption item, for which
benefits only matched costs, would sacrifice $30 million in benefits.
Given the short-term focus of GRH, accounting exercises such as
these might well explain why capital spending was so susceptible
to the knife.
In addition to capital spending, GRH made it unattractive to spend
money on other projects having short-term costs and long-term benefits.
For example, the government avoided proper maintenance of nuclear
armaments plants to save money in the short term. Because of that
neglect, the cost of keeping these plants operating has escalated
and the additional cost is now estimated by the CBO to be roughly
$160 billion spread over the next 40 years. Another painful example
is the savings and loan crisis. If the government had dealt with
the problem when it surfaced in 1986, it would have meant closing
some insolvent thrifts at an estimated cost of $10 billion to $15
billion. But the government balked, at least in part because it
didn't want to put more pressure on itself to abide by the GRH targets.
So spending to remedy the problem was delayed, the problem mushroomed,
and the cost to the government as estimated by the CBO and the General
Accounting Office rose to over $150 billion in present-value terms.
The GRH process not only generated poor results, it also contributed
to delays and confusion. Without a conceptual framework, policy-makers
were forced to debate each minor budget variation as if it were
a new theme.
The Fall 1990 Reforms
The government was well aware that the budget process was in need
of repair and direction. The administration suggested numerous reforms,
including instituting a line-item veto and changing the accounting
procedures for loans and guarantees. The budget committees in both
chambers of Congress held hearings on reform, several bills proposing
reforms were introduced and the budget package that finally was
passed last fall contains important reforms of the process.
It's too early to judge whether the fall reforms will lead to
improvements in the process, but it's not too early to argue they
don't go far enough. These reforms still fail to provide government
officials with a conceptual framework for making decisions, and
the definition of budget deficits and the targets for them remain
The reforms make it more difficult to exceed the deficit targets,
but they also make it easier to raise the targets (see below the
details on the reforms). Under the new process, there are constraints
placed not only on the size of the total budget deficit as before,
but also separately on entitlements, revenues and three types of
discretionary spending. In addition, a series of three sequesters
can be prompted by spending overruns in the prior fiscal year. These
reforms give Congress less maneuvering room to exceed budget deficit
targets. But at the same time, the reforms allow the deficit targets
to be raised for economic and technical reasons, emergencies and
increased spending on deposit insurance activities. Since the targets
are arbitrary, the government is highly likely to use these loopholes
to continue its deficit spending ways.
The reforms also fail to resolve problems associated with the
current accounting system. Under the fall reforms it is still possible
to achieve budget "savings" by shifting payments forward into future
fiscal years, although the reforms remove much of the incentive
to shift them back to the current fiscal year. That's because past
spending overruns could prompt a sequester and thus are no longer
treated as spilt milk. However, the reforms do nothing to address
the bias against capital spending. Capital expenditures are still
treated no differently from current expenditures.
We have argued that rules are needed to address the policy bias
problem, but we have seen that bad rules can create further problems.
While the fall 1990 reforms improve the original GRH rules, a lot
more should be done. That is why we offer our proposal.
The fall reforms of the budget process provide a transition to
a more balanced federal budget. But even if they take us to that
destination, problems will remain. Without a logical basis for their
targets, policy- makers will find ways to violate them. Without
a sound accounting system, they will continue to bias their decisions.
And without a conceptual framework, they will continue to debate
every budget nuance as if it were a new problem.
The Case for Our Rules
Our budget proposal is a set of reforms intended to reduce both
the policy-making bias and the confusion associated with current
procedures, and it's guided by four basic economic principles.
First, the budget should be balanced in a present-value sense
without use of the inflation tax. This principle is based on an
accounting identity and on a stated goal of macroeconomic policy.
The identity says that what goes out from the government must come
in, and it implies that the present value of government expenditures
cannot exceed the present value of government receipts. Since it
would be inefficient for the government to take in more than was
needed, it follows that the present value of government expenditures
should equal the present value of government receipts. In this equality
receipts can include proceeds from the inflation tax, that is, the
depreciation caused by inflation in the value of government nominal
liabilities. However, based on statements in Humphrey-Hawkins legislation
and congressional testimony of high officials in the Federal Reserve,
we take price stability to be a goal of policy. With zero inflation
as a goal, the first principle follows.
Second, benefits should outweigh costs. This follows from the
theory of economic policy-making which requires that government
officials weigh alternative programs in terms of their economic
benefits and costs to society. Since the services of programs occur
over time, the government must measure benefits and costs in expected,
present-value terms. Those benefits and costs are dated to occur
when resources are transferred in and out of the private sector.
Thus, when the government hires workers, the economic cost occurs
when the workers enter the public sector and not when the government
gets around to mailing their checks.
The third economic principle is that users pay. That beneficiaries
of government programs should pay is partly a fairness argument.
It also has the virtue of making it more likely that the benefits
of public programs exceed their costs, since those costs cannot
be pushed off on non-involved parties. This principle suggests that
borrowing to finance current consumption is unacceptable because
that method of financing pushes the costs off to future generations
who do not benefit from the consumption. In contrast, it also suggests
that borrowing to finance capital spending is acceptable since future
generations will benefit from the services of that capital. Obviously
this principle cannot be applied across the board. By definition,
income redistribution programs, such as welfare, cannot be financed
by recipients. But some other programs, such as the national parks
and the highway system, would fall squarely under the user-pays
principle. And most other programs would fall under this principle
in a general way. Current services should be paid for by the current
generation. And transfers to the poor of one generation, which are
designed to even the income distribution, should be paid for by
the wealthy of the same generation.
Our fourth economic principle, tax smoothing, is an implication
of studies of the tax structure. The implication follows as long as
the deadweight loss, the distortions caused by the tax and the resources
burned up in collecting it, rises disproportionately with the tax rate.
That is, the deadweight loss more than doubles when the tax rate doubles.
Tax smoothing means that when the government has commitments to spend
in the future, it should begin taxing for them today. This is true whether
those commitments are contractual, such as underfunded pensions, or
non- contractual but fairly certain to occur, such as wars or natural
disasters. What this means in practice is that it's more efficient to
raise taxes a little bit now and keep them there than it is to wait
and raise them a lot when the spending takes place.
We believe that these four simple principles suggest reforms of
the budget process which help deal with the policy bias problem.
We also believe they can provide guidance on many current budgetary
Our Reforms in More Detail
Our proposal for reform is hardly radical. It is composed of modest
changes in accounting procedures, rules on debt issue and enforcement
mechanisms. Most of the changes are either incorporated into budget
practices of corporations and state and local governments or included
in other proposals for federal budget reform.
The accounting changes we propose are that expenditures and receipts
be recorded on an accrual basis and that separate accounts be maintained
for operating and capital items. These accounting changes follow
directly from our cost-benefit timing and user-pays principles.
Our cost-benefit timing principle requires that expenditures and
receipts be recorded when the activity giving rise to them occurs;
that is, they should be recorded on an accrual basis. Our user-pays
principle suggests that it is not appropriate to borrow for operating
expenses but it may be appropriate to borrow for capital. Therefore,
it follows that separate accounts should be maintained for operating
and capital items.
These accounting changes allow the financial effects of alternative
policy actions to be more accurately represented. This facilitates
official decision making and also makes it easier for voters to
monitor officials' actions. Our proposals for accounting changes
are not original. They have been proposed by the General Accounting
Office (GAO) and they have been included in a bill introduced by
Sen. Herbert Kohl of Wisconsin. Most firms and state governments,
as well as the Federal Reserve, maintain separate operating and
capital accounts. The federal budget is reported on an accrual basis
in the National Income Accounts, and the budget, calculated as the
GAO and we recommend, is produced by the OMB in a timely manner.
Thus, all that is new here is that we are proposing using this existing
budget information as the basis for policy deliberations and rules.
The rule changes we propose limit the amount of debt the government
can issue on its operating and capital accounts. The rules follow
from our principles and from our attempts to reduce the policy bias.
Although they involve only minor changes to existing rules, they
provide explicit policy targets.
We propose to limit the debt that can be issued on the operating
budget by requiring that the combined estimated and projected budget
balance be zero in the current and subsequent fiscal years. Since
the accounts would be maintained on an accrual basis, the proposal
allows operating debt to be issued temporarily when there is a mistiming
of payments and receipts. It also could be issued temporarily when
unforeseen spending increases or revenue losses occur. However,
by including the current year's deficit in the calculation, the
government would have to implement policies to eliminate debt caused
by mistakes in budget projections. This proposal is similar to current
GRH procedures and suggested balance-the-budget amendments. What
is new in our proposal is that the budget being balanced is the
operating budget and that adherence to the rule leads straightforwardly
to present-value balance of the entire budget, without inflation,
as our first principle requires.
But why a two-year rule rather than a five-year rule or a month-by-
month rule? Given the nature of the policy bias against the future,
a rule requiring a balanced budget over a fairly short time frame
is desirable. Otherwise, policy-makers can continue to run deficits
while claiming they will be offset by surpluses at some distant
time. However, neither spending nor tax revenues can be forecasted
very accurately, and unforeseen events do occur. Thus, if the time
frame is too short, policy-makers will continually be forced to
make changes to expenditure programs or tax rates. In our view,
a two-year rule is a reasonable compromise.
We propose to limit the debt that can be issued on the capital
budget by requiring Congress to pass a bill annually authorizing
debt issue up to a specified ceiling. While this is much like current
procedures, our ceiling applies only to debt issued to finance capital
spending. This means that the ceiling would be an independent control
on capital spending. It would not be redundant, required as it is
now, to accommodate the operating deficits Congress has planned.
Since all capital spending would be financed by debt issue, setting
a ceiling on debt would be equivalent to setting a ceiling on federal
capital spending. This would let policy- makers better decide on
a desirable mix of private and public capital. More capital spending
would be desirable as long as the benefits of a project were at
least as large as its costs. We view a ceiling on total capital
spending as desirable so that Congress as a whole can effectively
force constituencies for capital spending to compete with each other.
This reduces the incentives to spend excessively on capital equipment.
We propose to enforce the rules using approaches similar to current
practices. The rule on operating debt would be enforced with a sequester.
The sequester could be applied in a disaggregated way, as it is
under current procedures. The sequester would be triggered whenever
the combined operating deficit in the current and succeeding year
exceeded some small amount--say $10 billion to match the trigger
amount under GRH. The sequester would require cuts in spending or
increases in revenue to achieve combined budget balance. Thus, if
there were an unforeseen deficit of $20 billion in the current fiscal
year, the government would have to adopt policies leading to a $20
billion surplus in the succeeding year. We would limit the amount
of deficit reduction in a sequester to 0.5 percent of GNP, which
is roughly the amount of reduction that experts testified could
be implemented without causing major economic disruptions.
The rule on capital debt would be self-enforcing. The Treasury
simply would not be authorized to issue debt above the legislated
Our proposals so far are derived from our economic principles
of present-value balance, cost-benefit comparison and user-pays,
but seem in conflict with our tax-smoothing principle. The reason
is that government spending and revenues fluctuate due to causes
that cannot be perfectly anticipated. Wars and recessions are as
likely to occur in the future as they have in the past, but it's
hard to know when. Therefore, meeting the two-year balanced budget
rule would require sharp changes in tax rates when these contingencies
occur. To avoid these kinds of changes in tax rates, we propose
that rainy day funds be set up to meet contingencies. These rainy
day funds would be set apart from the operating budget. Inflows
of cash into these funds would be counted as outlays for the operating
budget and outflows from these funds would be counted as receipts.
By drawing down the funds in bad times and building them up in good
times, tax rates would not have to be adjusted in conflict with
our tax-smoothing principle. Since we recognize the temptation to
raid these funds in good times, we suggest that a supermajority
in Congress be required to use these funds.
What Our Reforms Will Accomplish
Our reforms are intended to lessen the government's bias to overly
discount the future and to remove some of the confusion that surrounds
current budgetary practices. We argue that they lessen the bias
by making deficit financing more difficult, capital spending more
attractive and procrastinating more costly. We argue that they reduce
the confusion by providing a framework based on economic principles.
How do we make deficit spending more difficult? Reporting the
accounts on an accrual basis takes away the budget "savings" options
of selling off assets for cash and delaying payments to government
employees or program beneficiaries. Accrual accounting records when
activities take place and not when exchanges or payments are made.
Including an explicit makeup for past errors in the enforcement
mechanism reduces the incentive to use overly optimistic economic
and technical assumptions. Mistakes require painful adjustments
in the upcoming year. And, as we argue later, requiring present-value
balance makes the movement of items to off-budget status less advantageous.
However, the most important contribution the proposal makes to controlling
deficits is that it provides a definition of budget deficit and
specifications of targets which are guided by economic principles
and, thus, have some logical basis.
One problem with the GRH targets is that they were designed to
lead to budget balance for an arbitrary definition of the budget.
There is no economic principle that suggests the deficit should
be zero when capital transactions are included in the definition
of balance. Moreover, under the GRH deficit definition there are
different targets depending on whether Social Security or the RTC
Our definition and targets are not as arbitrary as those of GRH,
and that should make it harder to raise or disregard the deficit
targets. They are guided by the present-value principle. The definition
makes clear that capital transactions are excluded from the zero
deficit target. Having a clearer idea of the reason for the targets
should make it easier to stay the course.
Our proposal also makes capital spending more attractive by putting
it on a more equal footing with current spending. A dollar cut from
capital spending would have a comparable effect in the current fiscal
year to a dollar cut from current spending. That is because in our
proposal the operating cost of a capital asset is spread out over
the life of the asset. While the purchase price of the asset is
reported in the capital budget, only the annual depreciation and
interest financing expense are reported on the operating budget.
Thus, a dollar cut from capital spending cuts current spending by
the amount of depreciation and interest. Our method spreads the
cost of capital equipment over the years it provides services, while
the current method charges it all to the current year. Our yearly
charge is essentially what it would cost the government if it rented
the capital from a private party.
The main difference between our method and current practices is
how it treats dollars saved on capital spending in the current year.
If the government decided not to purchase capital equipment, our
method would show that the savings in current expenses would be
only depreciation and interest. According to current practices the
savings would be the cost of the capital purchase, which is much
larger. As a result, current practices make cuts in capital spending
look more attractive to policy-makers than they really are.
As is usual under standard accounting principles, we would require
that the government's assets be carried on its books at the lesser
of cost or market value. Some assets of the government have an ascertainable
market value such as the assets acquired from failed savings and
loans. Thus, for such assets the government would have an incentive
to provide appropriate maintenance. If the government did not maintain
such assets appropriately, their market value would fall, thereby
resulting in a larger depreciation charge and adversely affecting
the government's operating budget. Even for assets without a readily
ascertainable market value, such as nuclear armaments plants, standard
accounting practices provide better incentive for maintenance than
Our method also requires quick action to balance the budget when
circumstances change. In this sense, under our proposal the federal
government would be forced to act like state and local governments
now do. Under our proposal, difficult choices could not be simply
passed on to future Congresses and administrations.
To illustrate how our proposal and the economic principles on
which it is based could work to reduce the confusion surrounding
current budgetary issues, we examine the treatment of trust funds,
the RTC, loans and guarantees and future commitments.
The controversy over trust funds, such as Social Security, is
whether they should be on-budget or off-budget. If on-budget, their
balances would be included in deficit calculations and targets.
If off-budget, they would not.
Our present-value balance principle gives some guidance on this
issue. To move a program off-budget means that the program should
have an independent budget. It should neither rely on revenue from
the general budget nor should its earmarked revenue be accessible
to other programs in the budget. If it's not independent, then it's
not truly a trust fund and it's not truly off-budget. The question
of whether Social Security should be off-budget is then a question
of whether its budget should be independent of the general budget.
If the answer is yes, then by the present-value principle the Social
Security budget and the general budget should be balanced independently
in a present-value sense. If the answer is no, then just the sum
of the two budgets should be balanced in present value. Within this
framework, policy-makers must first decide whether they want Social
Security to have an independent budget, and if they do, they will
find their choices to be quite limited on the financing of committed
Social Security benefits. For instance, experts believe that given
current benefit schedules and tax rates the Social Security system
is balanced in present-value terms. Thus, by the present-value and
tax-smoothing principles, policy-makers would not be allowed to
lower Social Security tax rates unless they also lowered the benefits.
We should also point out that our analysis of the policy bias problem
suggests that trust fund accounting can be a useful disciplinary device.
Because voters lack the information required to monitor the actions
of policymakers, it's hard to monitor whether policy-makers are following
the cost-benefit principle. This monitoring difficulty is particularly
acute when expenditures are financed out of general tax revenues. Beneficiaries
have every reason to argue that the benefits accruing to them are large,
whether they value the services a lot or a little. Dedicated programs
with independent revenue sources that have strong safeguards against
raiding the treasury can be useful in solving the monitoring problem.
From this perspective, trust funds are not merely an accounting device;
rather, they serve an important economic function.
We also recognize that trust funds can be abused. Given the bias
in policy-making, policy-makers have an incentive to postpone costs
and accelerate benefits. For example, policy-makers have an incentive
to run a deficit or a smaller surplus than is desirable on the Social
Security system. The result is that future benefits must be reduced
or future taxes raised if the system is to be independently balanced.
One crude way to limit abuses of this kind is to require that trust
funds not run a deficit.
The issue on RTC spending is how to split it up into on-budget
and off-budget. The RTC handles the assets and liabilities of failed
thrifts. Since RTC spending relies on general revenues, our reasoning
on trust funds suggests all of it belongs on-budget. The drive to
move some of it off- budget was mainly a result of the current procedure's
failure to distinguish capital spending from operating expenses.
What typically occurs is that the RTC takes over a failed thrift
with assets valued at, say, $700 million and insured deposits of,
say, $1 billion. The $1 billion must be paid off immediately, while
the $700 million in assets is sold gradually over a number of years.
By current methods the $1 billion is treated as a current expenditure.
Then, when the assets are sold over time, the sale receipts are
treated as revenue. The pay-off to depositors is funded by debt
issue and the debt is in effect reduced when the assets are sold.
The interest is also treated as an expenditure. The current procedure
clearly overstates the deficit in the current year, since it assigns
no value to the assets the government acquires. The drive to move
RTC spending off-budget was a clumsy attempt to correct this problem.
Using our procedures, only the capital loss and interest expense
would show up on the operating budget. That budget would not be
affected by the timing of asset sales. When the RTC initially takes
over the failed thrift, it would be considered a capital purchase
of $1 billion financed by debt. However, since the assets were worth
only $700 million, there would be an immediate write-off of $300
million charged to depreciation. As with other capital purchases,
there also would be an associated interest expense. Future asset
sales would affect the operating budget only to the extent that
actual sale values differed from the capital budget's assumed market
values. Thus, using our procedures, RTC spending would be treated
no differently from other capital spending.
Capital budgeting also would clarify the treatment of government
loans and guarantees. These items involve subsidies that are realized
when private parties fail to maintain payments on loans. Past budget
practices have treated the government loss of loan revenue or payment
on a loan guarantee as a budget deficit increase at the time they
occur. Thus, it appears to policy-makers as a good way to give out
subsidies now and pay for them much later. Under our proposal, loans
and guarantees would be included in the capital budget as assets
and liabilities, respectively. The subsidies on loans and guarantees
would show up on the operating budget at the time the loans and
guarantees were granted. Under our proposal policy-makers would
be confronted immediately with the costs of the subsidies. We should
point out that the way loans and guarantees affect the operating
budget under our proposal is similar to how they will affect the
GRH budget following last year's reforms.
Finally, our tax-smoothing principle suggests that revenue should
be collected today for future commitments. Currently the money is
not collected until after the event occurs. Some of the commitments
are contractual or explicit, such as pensions, and for these the
government might make advance payments into something like an escrow
account. Other commitments are not explicit but are fairly certain
to occur in the future, such as wars and natural disasters. For
these we have proposed that the government make advance payments
into a rainy day account. The purpose of the escrow and rainy day
accounts is to provide present-value budget balance without having
to change tax rates. Assuming that the government's capital expenditures
rise at the same rate as national income, the effect of the accounts
is to lower the government's debt-to-income ratio over time until
the commitments are realized and then to allow them to rise at that
time. Over long periods of time, the debt-to-income ratio would
The purpose of using our procedures to examine these issues is
to show their practical value. We believe they can considerably
reduce the confusion surrounding current budget practices.
Objections to Our Reforms
Since aspects of our proposal have been tossed around for some
time, we can anticipate two important objections to it. One objection
is that it does not accommodate countercyclical policy, and the
other is that our proposal would encourage policy-makers to move
everything over to the capital budget. Although these objections
have some validity, we believe they are not decisive. We believe
the constraint on countercyclical policy is not very costly, and
we believe safeguards can be put in place to limit misclassification
Consider first the loss in flexibility to conduct fiscal policy.
Our proposal allows some limited countercyclical policy, but it
does not allow the government to suspend the rules in case of a
recession. The availability of a rainy day fund would, in any case,
allow for some countercyclical fiscal policy. But when an unforeseen
shortfall does occur, it could be accommodated in the current year,
provided it is made up for in the succeeding year. The government
would also be able to increase capital spending in a recession when
interest rates were low, because such spending then would generate
less interest expense.
Nonetheless, our proposal is more rigid on countercyclical policy-
making than current procedures. We do not think this rigidity is
very costly because we are unaware of any evidence that discretionary
countercyclical budget policy works. Most studies show that lags
in responding to recessions cause any stimulative effects of fiscal
policy to occur too late, well into the ensuing recovery. Furthermore,
the rigidity could be beneficial. If the decline in the growth rate
of real output goes on for a long period, perhaps there is a secular
as well as a cyclical element. To the extent that an output decline
signals a long-term reduction in output growth, the government should
Consider next the objection that policy-makers will want to move
everything to the capital budget. To a great extent that's true,
but we argue that now everything, in effect, is treated as a capital
expense. The government can borrow to finance any expenditure. So
by strictly defining what is a capital expenditure, as states have
done and as the GAO proposes for the federal government, many expenditures
can be kept off the capital budget.
Even with strict definitions, though, there will be problems with
misclassifications or understating of depreciation on capital items.
Some expenditures that provide benefits in future years would be
classified as current. In fact, we would favor including most human
resource programs, such as those for education, crime control and
health, on the operating budget. We do not deny that a better educated,
better protected and healthier population will make people better
off in the future. We also do not deny that requiring these expenditures
to be paid in full in the current year could lead to underfunding.
It is just our judgment that the underfunding bias would be no greater
than the policy-makers' bias to overspend on current consumption.
Thus, we judge that by strictly limiting the capital budget to long-lived
physical and nominal assets, a small cost in terms of underfunding
of some expenditures would be more than offset: there would be a
smaller bias toward overspending on current consumption, which now
is facilitated by abuse of the debt-issue option.
The government would also try to understate depreciation, as states
and corporations have been known to do. It could classify some current
consumption items as capital items and assign them value, even though,
in a sense, they are fully depreciated in the current year and have
no value. Or it could just overstate the value of some of its physical
or nominal assets. This is where watchdogs such as the CBO and GAO
would have to be on the alert. The logic of our proposal requires
that depreciation be accurately recorded on the operating budget.
If it were not, the budget situation could be seriously misrepresented.
Understatement of asset depreciation led to the unrecognized deterioration
in the financial condition of many state and local governments and
various financial institutions.
How do we get from the current system to our proposed system?
Some of our reforms--accrual accounting and separating the capital
and operating budgets--could and should be adopted for fiscal 1992.
All that is required is that policy-makers look at a different set
of books. However, an immediate move to a balanced budget would
require enormous and disruptive increases in taxes or reductions
in spending. We believe the government should move to a balanced
operating budget over a three-to-five-year period. Over this period,
the goal of monetary policy should be to reduce the inflation rate
gradually to zero.
Such transition periods have been abused in the past, but we think
our enforcement mechanism provides a way to limit future abuses.
Specifically, we propose imposing annual limits on the operating
budget deficit during the transition period. These limits would
be enforced with a sequester. If the limits are exceeded within
a given year, then the sequester would require cuts in spending
or increases in revenues in the following year. These proposals,
combined with the fall 1990 reforms enacted by Congress, would go
a long way to reducing the deficit to zero over roughly a five-year
It's also possible to frontload the pain of spending reductions
and tax increases to a greater extent than is now mandated under
the fall 1990 reforms. One major problem with the GRH process was
that large deficit reductions were supposed to occur toward the
end of the targets, and this problem persists though to a lesser
extent with the fall 1990 reforms. When the real pain of deficit
reduction is postponed, however, the temptation to revise the targets
often becomes irresistible. The only credible way around this problem
is to ensure that substantial deficit reduction occurs in the early
years of the transition period.
Our Rules Are No Panacea
We would like to conclude by claiming it would be all smooth sailing
if only our proposal were accepted. But of course we know that's
not true. No change in the process can make the difficult choices
confronting policy- makers easy. They still would have to decide
whose ox to gore by cutting spending or increasing taxes. But we
think policy-makers would make better decisions if they understood
what they were up against and what the consequences of their actions
would be. No change in budget process is going to solve the policy
bias problem or keep the government out of financial difficulty.
Better budget processes than the federal government now employs
have not stopped these problems with corporations or state governments.
We nevertheless strongly believe our proposal can lessen the magnitude
of the problems.
Will our proposal work, or is more drastic measure such as a constitutional
amendment necessary? We believe that the situation is not yet so
dire as to warrant such an extreme action. Concern over the budget
is widespread enough in the nation and among policy-makers that
we feel the problems described here can be addressed legislatively.
The budget mess will not be completely cleaned up even if all
our reforms are adopted. Hard choices will still have to be made.
But we will no longer have the choice of inflicting costs upon future
generations for programs that benefit us.
Main Features of the Fall 1990 Budget Process Reforms
- Five-Year Budgeting. Budget resolutions and necessary reconciliation
bills must project spending, revenues and deficits for five years.
- Discretionary Spending Caps. Appropriations bills must stay within
separate caps for defense, foreign aid and domestic discretionary
spending for fiscal 1991-93; for fiscal 1994-95, the law sets overall
discretionary spending caps.
- Enforcement. A complicated set of sequesters ensures that spending
stays within the caps in the bill. Essentially, these sequesters
apply if the Office of Management and Budget determines that spending
will exceed the caps. The sequesters also "look back" to offset
spending increases or revenue cuts in the prior fiscal year.
- Pay-As-You-Go Entitlements and Revenues. Bills containing increases
in entitlement or other mandatory spending or reducing revenues
must be offset by entitlement cuts or revenue increases.
- Social Security and Deposit Insurance. Social Security receipts
and expenditures will no longer be included in budget calculations.
Increased spending for deposit insurance activitieschiefly
the savings and loan salvage operationwill not be allowed
to trigger sequesters.
- Emergencies. If requested to do so by the president, Congress could
enact emergency appropriations, entitlement increases or revenue
cuts without triggering sequesters.
- War and Recession. A declaration of war would still cancel the
sequester process. Congress could still vote to cancel the sequester
process in the event of a projected recession or measured economic
growth below 1 percent for two consecutive quarters.
On Political Economy:
A classic in the field is...
Downs, Anthony. 1957. Economic Theory of Democracy.
New York: Harper Press.
Other excellent readings are...
Buchanan, James and Tullock, Gordon. 1962. The Calculus
of Consent. Ann Arbor: University of Michigan Press.
Olson, Mancur. 1971. The Logic of Collective Action.
Cambridge, Massachusetts: Harvard University Press.
On Agency Theories of the Firm:
Alchian, Armen A. and Demsetz, Harold. 1972. Production,
information costs, and economic organization. American
Economic Review 62, pp. 777-795.
Jensen, Michael C. and Meckling, William H. 1976. Theory
of the firm: Managerial behavior, agency costs and ownership
structure. Journal of Financial Economics 3,
On Recent Budget Policy:
Miller, Preston. 1989. Gramm-Rudman-Hollings'
hold on budget policy: Losing its grip? Federal Reserve
Bank of Minneapolis Quarterly Review 13, pp. 11-21
Reischauer, Robert. 1990. Taxes and spending under Gramm-Rudman-Hollings. National Tax Journal 43, pp. 223-232.
Schick, Allen. 1990. The Capacity to Budget.
Washington, D.C.: TheUrban Institute Press
On Last Fall's Reforms:
Congressional Quarterly staff. 1990. Budget-reconciliation
bill. Congressional Quarterly, December 1, pp.
On Proposed Reforms:
Bowsher, Charles A. 1988. Budget reform for the federal government.
Statement before the Committee on Governmental Affairs, U.S.
Senate, June 7.
Budget reform proposals. 1989. Joint hearings before the
Committee on Governmental Affairs and the Committee on the
Budget, U.S. Senate, October 18, 26.
We wish to thank Rudolph Penner, Alice Rivlin, Mark Sniderman, Eugene
Steuerle, and John Sturrock for useful comments on an earlier draft.