There is little doubt that government regulation in America could be improved. It often doesn't do much good; it always costs a lot. The Organization for Economic Cooperation and Development, for example, recently reported that our environment is still deteriorating despite 25 years of cleanup effort, although total cleanup costs (both in the aggregate and as a percentage of GDP) are higher than in any other country, in part because our cleanup efforts are less cost-effective (The Economist, October 21, 1995: 32-33).

The interesting question is not whether, but how government regulation could be improved. The last Congress proposed to fix regulation by requiring all new rules costing more than $25 million to be subjected to a rigorous benefit/cost analysis. Would this have fixed regulation?

There is, in fact, nothing new about regulatory review or even mandated benefit-cost analysis. The United States has had mandatory benefit-cost analysis for twenty years (Thompson & Jones, 1980). President Clinton's Executive Order #12,866, which requires formal review of all new "significant," substantive federal rules and makes the Office of Information and Regulatory Affairs in the Executive Office of Management and Budget responsible for reviewing proposed regulations, merely reaffirms this long-standing commitment.


It is also a fact that Executive Order #12,866 is something of a toothless wonder -- a triumph of process over purpose. For example, Jim Tozzi (1995), of the Institute for Regulatory Policy, looked at 510 new rules proposed by Environmental Protection Agency between April and September of 1995. He couldn't find a single rule that was rejected because it failed to meet the cost-benefit test or because its costs were deemed excessive. Bob Hahn (1996) arrived at a similar conclusion as the result of a careful look at 92 regulatory analyses carried out since 1990.

Because regulatory review has not led to better regulation -- more benefits and less cost, some pundits have concluded that what we really need is a regulatory budget [Crandall, 1979; DeMuth, Shackson, Stork, & Wright, 1979; Thompson & Jones, 1981; Harrison & Portney, 1982; Litan & Nordhaus, 1983; Crews, 1996]. This notion has been endorsed by think tanks like the Heritage Foundation and the Competitive Enterprise Institute and bills have begun to surface in Congress that would put America on a regulatory budget. The idea was part of the Contract with America and is the main element of HR-994, the Regulatory Accountability Act.


The basic justification for regulatory budgeting is quite simple: government regulation is the economic equivalent of government spending and ought, therefore, to be treated the same by legislators. As Robert Litan and William Nordhaus (1983: 4) explained:

From an economic viewpoint, federal regulations are akin to federal expenditure programs. Both require that resources be devoted to the pursuit of objectives the nation collectively deems to be important. The only difference is that, in the case of federal expenditures, the resources are first collected through taxation and then spent directly by the government. In the case of regulation, the government orders individuals or firms in the private sector to make such expenditures -- a kind of balanced budget expenditure program.

From this economic perspective, the need for a centralized process for coordinating regulation to parallel that of expenditures becomes readily apparent. How can such a mechanism be established? Conceptually the solution is straightforward. [T]he Congress could establish a 'regulatory budget' to coordinate and manage the diverse regulatory activities of the federal government. Indeed, it might be desirable to integrate all federal activities -- expenditures, tax expenditures, credit allocation programs, and regulatory programs -- into a single 'super-budget.'

Of course, one could treat expenditure and regulation alike by treating them both like regulation. But most people accept that the existing budget process leads to better decisions, in a way that regulatory review does not (GAO, 1996) -- not optimal, or even good decisions, just better ones. Indeed, there is considerable evidence to show that the budget process, despite its flaws, constrains the growth of spending and allocates resources to where they are most needed.

According to Aaron Wildavsky, the budget process works because it brings spending interests into conflict, thereby forcing elected officials to acknowledge scarcity. This guarantees that budget examiners will have allies among their political masters, making it possible for them to compel spenders and their sponsors to justify their proposals. The adversarial nature of the process also encourages political evaluation of benefits and costs and mutual adjustment on the part of participants (Wildavsky & Jones, 1994). Without political allies, all the onus of cutting would fall on budget examiners, who have no real constituencies of their own. In which case, they would be as impotent as their counterparts across the hall in the Office of Information and Regulatory Affairs.


While there have been calls to put the federal government on a regulatory budget, proponents have never really explained what a regulatory budget would look like or how it work in any detail or shown how we could get from here to there. To assess the feasibility, let alone the desirability, of a regulatory or super-budget, one must first understand how it would work. That is the purpose of this essay, starting with the existing expenditure budget process.

The Constitution stipulates that to spend money executive branch departments and agencies must first be given the authority to do so by an appropriation "made by law." The Constitution also requires the federal government to give an accounting of its finances "from time to time." To comply with these requirements, to promote economy and efficiency in government, and to assert congressional prerogatives, budgeting in the federal government comprehends three legislative processes -- authorization, appropriation, and budget enforcement -- and is sustained by a comprehensive system of administrative controls and financial accounts.

Under a super-budget, regulatory rule making would have to conform to these processes as well.1 In that case, it might take the following shape:

1. Authorization: in addition to setting out the purpose and techniques of regulation, as is now the case, authorizing statutes would also set limits on the financial burden that a regulatory agency could impose on individuals and organizations in the private sector, state governments, and subordinate jurisdictions. For example, future amendments to the Clean Air Act could authorize the EPA to require industry to reduce SOx emissions by an additional 8 million tons per year at an annualized cost of up to $8 billion. Authorizing legislation could provide permanent authority, multi-year authority (must be renewed on a periodic basis), or annual authority (must be renewed annually).

2. Appropriation:

a. The agency would forward the rules by which it proposed to carry out the purpose of its authorization, together with an estimate of the costs of those rules and their consequences, to program analysts in OMB. For example, EPA might propose that coal-fired electric-power generating plants convert to low-sulfur natural gas, reducing SOx emissions by an additional 6.5 million tons per year at an incremental annualized cost of $8 billion.

b. Under the direction of the President, OMB would review all proposed rules, their justifications, and accompanying analyses, and make choices between alternative regulatory proposals. These choices would have to be consistent with Budget Enforcement Act mandated subceilings and totals (see below). For example, the OMB might reject EPA's natural gas rule in favor of another regulatory program/activity designed to reduce SOx emissions, a rule designed to promote water quality, or perhaps even a spending program in another area altogether.

c. All the proposed rules embraced by the OMB would be included in the President's Budget and submitted to Congress. The rule changes included the President's Budget would be individually scheduled and their costs assigned to one or more budget functions and departments. The logic of a "super-budget" would seem to call for the creation of a new object-of-expenditure class -- regulatory compliance spending -- to parallel the existing classes -- salaries, benefits, insurance, grants/subsidies, capital assets, etc. However, compliance spending would be assigned to existing functions and departments. For example, the estimated regulatory compliance costs of reducing SOx emissions would be assigned to Subfunction 305, Pollution Control and Abatements, and to the EPA.

d. Congress would pass appropriations bills authorizing regulatory agencies to encumber funds, to enforce existing rules requiring individuals and organizations to incur compliance expenses, and to issue notices of proposed rulemakings that would cause individuals and organizations to incur compliance expenses. Regulatory budget authority could be provided on an annual, multi-year, or no year basis. Litan and Nordhaus (1983: 145) proposed that "the regulatory budget for a given year should be stated in terms of the incremental spending mandated by the rule over the lifetime of the rule." Thus, the appropriation for the hypothetical natural-gas rule, designed to reduce SOx emissions by 130 million tons over the program's twenty-year life, would be, say, $160 billion, rather than $8 billion for one year. As Litan and Nordhaus correctly point out (146), this is the way Congress has traditionally appropriated funds for large-scale, multi-year projects (e.g., the construction of a dam or aircraft carrier). Estimating errors were then corrected in subsequent years with supplemental appropriations.

There are, however, two problems with this approach. In the first place, it doesn't account for the timing of expenses. To account for the timing of expenses, lifetime spending should be discounted to present values. For example, an $8 billion annuity paid out over twenty years has a present value of only $100 billion -- that number is much closer to the real economic cost of the hypothetical decision to require power plants to burn natural gas than $160 billion. Discounting future spending to present values is now standard practice in budgeting for federal credit guarantee programs, insurance programs, and Corps of Engineers construction and reclamation projects. It should be universal where authorization and appropriation of new federal initiatives are concerned.

In the second place, the Litan and Nordhaus proposal ignores the need to estimate annual outlay rates and enact them into law along with appropriation of multi-year and no-year budget authority. Annual outlay estimates are needed to ensure that the budget is executed as enacted and also to comply with the requirements of the Budget Enforcement Act, which focuses on annual outlay totals rather than new obligational authority.

3. Budget Enforcement: under the Budget Enforcement Act (Doyle & McCaffery, 1991, 1992, 1993), the successor to Gramm-Rudman-Hollings, Congress is committed to enacting appropriation bills that respect legislated annual ceilings on total outlays and negotiated subceilings. If a regulatory budget is to have more than symbolic value, regulatory compliance spending must be subject to the same disciple as other federal spending. The process must make it clear to all the participants that regulatory spending requires sacrifice. Hence, Congress must be forced to cut spending by an offsetting amount from another function to spend more on regulatory compliance in any one year. This means that the Budget Enforcement Act must be stretched to accommodate regulatory compliance spending and that regulatory compliance expenses must be budgeted annually.

These are difficult, but not insuperable problems. Skilled budget examiners regularly solve these kinds of problems.


Accounting and control raise more serious problems. When Litan and Nordhaus made their super-budget proposal, they dismissed a system of regulatory accounts and voucher controls as unnecessary and infeasible (144-158). They were certainly wrong on the first count. Lacking a system of controls, a regulatory budget would be little more than a metaphor -- useful, perhaps, for highlighting some of the defects in the existing regulatory process,2 but without bite. Controls are needed to insure that budgets are executed as enacted. Litan and Nordhaus were probably wrong on the second as well. The federal government's experience with credit reform, changes in FASB rules, and improvements in information technology make it possible to envision a workable system of regulatory accounts, although bringing one into conformity with the federal government's Byzantine encumbrance accounting system is no simple matter.

Congressional appropriations authorize government agencies to encumber funds. An encumbrance occurs when the agency enters into a contract that will cause an outlay of funds from the U.S. Treasury. An agency's encumbrance authority is typically divided into scores of separate accounts, hundreds of management codes, and, perhaps, thousands of accounting lines. This system is repeated at each level of the organization, from the top to the bottom. Hence, each time someone within an agency proposes to incur an obligation, the encumbrance must be matched to the appropriate account, code, and line, and this process must be repeated all they way up or down the organization to verify that the spending office, bureau, and agency/department have the authority to enter into the contract. Encumbrance accounting has no direct counterpart in the private sector, although firms are required to disclose material contingent liabilities. Encumbrance accounting, as noted earlier, was designed to insure that funds are expended only for purposes authorized by Congress and that neither lower level employees nor higher level officials exceed the obligational authority granted them.

Obligations are followed by expenditures and disbursements. Disbursements are measured when payment is made -- when a check drawn on the Treasury is issued or cash disbursed. The federal accounting system also matches expenditures and disbursements to the appropriate budget accounts, codes, and lines and on a quarterly basis reports the unencumbered obligational authority, unliquidated encumbrances, and expended appropriation (disbursements plus accounts payable) for each and every budget line and organizational level.

Government agencies also generate expenses. Expenses measure the cost of the assets consumed in the production of goods or services, which is what business accounting concerns itself with. Agencies also spend money to produce things like national defense, the capacity to meet our overseas commitments, or to cleanup hazardous waste sites, airsheds, or waterways. However, government agencies do not now generally measure their expenses, although designing an accounting system that tracks expenses as well as encumbrances and expenditures is fairly easy (Anthony & Young, 1994: 429). The Air Force, for example, suspends expenditures in working capital accounts until a component unit consumes the asset that actuated the expenditure. It should be just as easy to reverse engineer this system,3 although it is not really necessary to do so to bring regulatory accounts into conformity with encumbrance accounting. Under a regulatory budget, revenues would automatically reflect compliance costs. This is the procedure followed in accounting for tax expenditures (special provisions in the tax laws that reduce revenues): the revenue line reflects the tax that would have been paid in the absence of the tax expenditure and the expenditure item reflects the transfer. Like tax expenditures, regulatory compliance expenses do not lead to a disbursement of funds from the Treasury and do therefore not affect the size of the deficit or federal borrowing.4 That being the case, there is no reason why regulatory agencies should not report the expenses incurred by firms in complying with federal regulations, in lieu of expenditures. In that case, expenses would liquidate encumbrances. Their timing would be governed directly by the Budget Enforcement Act and indirectly by appropriations to agencies' regulatory accounts.

The key to bringing regulatory accounting into conformity with encumbrance accounting lies in the recognition that regulating is directly analogous to contracting. Regulation typically involves issuing a license or permit to an applicant, a firm or subordinate jurisdiction, which lays out the terms of compliance agreed to by both regulator and regulated. For example, the Federal Water Pollution Control Act requires firms discharging wastewater into navigable streams and their tributaries to obtain a permit from the Environmental Protection Agency dictating the treatment that is required, the quality of water discharged, and plans to reach those limits. Furthermore, under the Federal Accounting Standards Board's SFAS No. 5 and the Government Accounting Standards Board's Rule No. 18, the costs of complying with federal regulatory permits and licenses must be recognized as a contingent liability when the permit or license is issued and expensed when the liability is liquidated (Hawkins & Goff, 1994; Hogan, 1996). One would merely have to require the contingent liability to be reported on the regulatee's balance sheet to be included in the permit to provide a measure of regulatory encumbrances, in conformance with appropriations. Reported expenses could then be used to measure their liquidation, in conformance with the Budget Enforcement Act .

An agency's regulatory account would be divided into codes by act or title and lines for each program and for each level of the organization. Hence, each time someone within a regulatory agency proposed to issue a permit, the encumbrance would be be matched to the appropriate code and line, and this process would be repeated up and down the organization to verify that the permitting office, bureau, and agency had the authority to do so -- thereby insuring that the regulatory budget was executed as enacted.


Unfortunately for advocates of a regulatory budget there are two practical problems with the approach outlined here -- one minor, the other more serious. The minor problem is that some regulatory programs do not involve issuance of a license. Where that is the case, the administrative similarities between spending and regulation are attenuated and the problem of regulatory accounting is correspondingly exacerbated. Conformity with encumbrance accounting would require redesign of regulatory programs and not merely the provision of additional information. This problem can be described as minor, however, because the number of regulatory programs involved are few and relatively insignificant in terms of the burden they impose upon the economy.

The more difficult problem is that SFAS No. 5 may not provide adequate guidance for regulatory reporting. Comparing estimated regulatory liabilities for all U.S. firms with the regulatory contingencies reported on the 10Ks and Annual Reports of publicly traded companies, Ennis Hawkins and Don Goff (1994) found that liabilities were understated by nearly half.

In Canada, which like the United States mandates cost-benefit analysis of all new regulations, a public-private partnership involving the Canadian Manufacturer's Association, the Treasury Board Secretariat, Industry Canada, and Consulting and Audit Canada has developed an activity-based analytical tool (BIT ©CMA Quality Software) designed to help government, business, or other interested parties measure the expense of complying with regulation and other government actions. This activity-based analytical tool can also be used by firms as a "should cost model" to estimate the total cost of alternative regulatory compliance strategies. The Canadian government is committed to using this tool to measure regulatory compliance costs and, by comparing after-the-fact cost measures with its before-the-fact cost estimates, using it to revise and refine parameter values, thereby increasing the accuracy of should-cost estimates.


Fifteen years ago, when Robert Litan and William Nordhaus (1983: 157) evaluated the regulatory budget concept, they concluded:

In our view, none of the problems -- with regard to the scope of the regulatory, the nature of the dollar constraints, the manner of executive and congressional involvement in the process, and sanctions -- is individually great enough to dissuade us of the desirability of a regulatory budget. But taken together they are sufficiently severe to make it unwise to move toward a budget plan in one leap.

That was then; this is now. In the mean time, we have learned a lot. We have a far better idea of what a regulatory budget would look like and how one would work. It is now reasonable to conclude that regulatory budgeting is feasible.

Whether regulatory budgeting is desirable is another matter. I have reservations about the existing federal budgetary process and system of encumbrance accounting. Nevertheless, the time seems ripe to consider the idea. Thanks to improvements in information technology, the administrative costs of implementing a regulatory budget have dramatically decreased in recent years. Moreover, Congress seems to be concerned with reforming the regulatory process. Regulatory budgeting makes as much sense as any of the other procedural reforms currently under consideration -- and more than most.

The next logical step is probably the development of a model regulatory accounting system and draft legislation to implement it. If people are going to give serious consideration to the regulatory budget concept, it must be given tangible form. Besides, better regulatory cost measures are needed whether a regulatory budget is adopted or not. At a minimum, better cost measures are needed to improve cost estimates. If the numbers used in debates over regulatory proposals were more reliable, they might be taken seriously.



Crandall, R. Is Government Regulation Crippling Business? Saturday Review. 20 [January 1979], 34-40.

Crews, C. Wayne, Jr. Promise and Peril: Implementing a Regulatory Budget. Unpublished Study, Competitive Enterprise Institute, Washington DC, 1996.

DeMuth, C., Shackson, R.H., Stork, E.O., and Wright, A.W. The Regulatory Budget as a Management Tool for Reforming Regulation. Paper commissioned by the Joint Economic Committee, processed May 29, 1979.

GAO Regulatory Reform: Implementation of the Regulatory Review Executive Order #12,866. General Accounting Office, Report T-GGD-96-185, Washington DC, September 26, 1996.

Hahn, Robert W. Regulatory Reform: What Do the Government's Numbers Tell US? Working Paper, American Enterprise Institute for Public Policy, Washington DC, 1996.

Harrison, David and Portney, Paul R. Regulatory Reform in the Large and in the Small, in L. Graymer and Fred Thompson (eds.) Reforming Social Regulation. Los Angeles: Sage, 1982: 219-246.

Hawkins, Ennis and Goff, Don The Environmental Timebomb: Disclose, Recognize or Disregard? Paper No. 94-07A, College of Business Administration, Sam Houston State University, Huntsville, Texas 77341, 1994.

Hogan, William Environmental Regulation: The "Cost" of Compliance, Municipal Finance Journal. 17/1 [Spring 1996], 29-36.

Litan, Robert E. and Nordhaus, William D. Reforming Federal Regulation. New Haven, CN: Yale University Press, 1983.

Merrow, E., Chapel, S., and Worthing, C. A Review of Cost Estimation in New Technologies. R-2481-DOE. Santa Monica, CA: RAND, 1979.

Quarles, Ross How Much Does It Really Cost to Be Green? The Need for an Environmental Cost Classification System. Paper No. 95-19A, Department of Accounting, Sam Houston State University, Huntsville, Texas 77341, 1995.

Quirk, James P. and Terasawa, K. Sample Selection and Cost Underestimation Bias in Pioneer Projects, Land Economics. 62/2 [May 1986], 192-200.

Thompson, Fred Mission-Driven, Results-Oriented Budgeting: Financial Administration and the New Public Management, Public Budgeting & Finance. 14/3 (Fall 1994) 90-105.

Thompson, Fred and Jones, L.R. Reforming Regulatory Decision Making: The Regulatory Budget, Sloan Management Review.  22/2 [1981], 53-61.

Thompson, Fred and Jones, L.R. Fighting Regulation: The Regulatory Review, California Management Review. 23/2 (Winter 1980), 5-19.

Tozzi, James Ensuring Accountability for Developing Well-Founded Federal Regulations: An Initial 'Report Card' on Compliance with Key Directives of the Regulatory Executive Order [E.O. 12866]. The Institute for Regulatory Policy, Federal Focus, Inc., Washington DC, April 1995.

Wildavsky, Aaron, and Jones, L. R. Budgetary Control in a Decentralized System, Public Budgeting and Finance. 14/1 (Winter 1994), 7-22.

Wood, L.D., Laws, E.P., and Breen, B. Restraining the Regulators: Legal Perspectives on a Regulatory Budget for Federal Agencies, Harvard Journal on Legislation. 18/1 [1981], 1-33.