Where Does the Public Sector End and the Private Sector Begin?
Author: Mr. Ian Lienert

Contributor Notes

Author’s E-Mail Address: ilienert@imf.org

The boundary between the public and private sectors can be defined on the basis of ownership of institutional units. Nonmarket government-owned entities and corporations that are owned or controlled by government units belong to the public sector. “Economic ownership” is more important than majority ownership. Joint ventures, public-private partnerships, and social insurance funds (including for public employees) can be unambiguously allocated to the public or private sector on the basis of international public sector accounting standards. Boundary problems within the public sector are just as acute as those between the public and private sectors, mainly because of ambiguities in distinguishing “market” from “nonmarket” activities.


The boundary between the public and private sectors can be defined on the basis of ownership of institutional units. Nonmarket government-owned entities and corporations that are owned or controlled by government units belong to the public sector. “Economic ownership” is more important than majority ownership. Joint ventures, public-private partnerships, and social insurance funds (including for public employees) can be unambiguously allocated to the public or private sector on the basis of international public sector accounting standards. Boundary problems within the public sector are just as acute as those between the public and private sectors, mainly because of ambiguities in distinguishing “market” from “nonmarket” activities.

I. Introduction

Effective fiscal policy requires a clear definition of the public sector. In recent years, the use of private sector management techniques in the public sector, as well as private sector provision of public goods, has blossomed. Governments in many countries participate in public-private partnerships (PPPs), outsource or contract out to private sector suppliers, or use publicly-funded vouchers for supplying public goods (Blöndal, 2005, IMF, 2006). Certain functions and monopolistic-type activities, which were previously thought to be reserved for public sector provision, are now either furnished by the private sector, though funded publicly, or have been transferred to the private sector, through partial or full disengagement of the government from state-owned enterprises. In addition, private sector financial accounting standards have influenced changes in public sector accounting methodologies, with increasing use of accrual accounting for preparing government financial statements (Chan, 2002; OECD, 2003). More generally, the private sector is being looked at as the “model” for good budget and public financial management practice.

With increasing use by governments of market mechanisms and private sector provision of public services, has the borderline between the public and private sectors become blurred? Is it still possible to define the public and private sectors in black and white terms? If not, is there now a third “gray zone”, which is “semi-private” and “semi-public”, and whose activities belong neither to the traditional public sector, nor to the profit-driven private sector? If yes, could law be used to demarcate such a gray zone, clarify its conceptual basis, and offer a standard for defining the gray zone consistently across all countries?

This paper discusses these questions. It is argued that, to the extent that a gray zone exists, it is possible to allocate component parts to the public and private sectors. In Part II, conceptual issues—especially two key defining characteristics of the public and private sector, namely ownership and control—are discussed. The extent to which laws could be helpful for defining more clearly the public and private sectors is also investigated. Part III elaborates on a conceptual framework for defining the public sector. This is based on the IMF’s Government Financial Statistics Manual (GFSM2001) (see IMF, 2001). The application of this framework to joint ventures, leasing, PPPs and social security funds involving private and public companies or entities is examined in Section IV. Within the public sector, the difficulties of demarcating the boundary between public corporations and “general government”, as defined in GFSM2001, is discussed in Section V, prior to the concluding remarks of Section VI.

II. Conceptual Issues for Defining the Public and Private Sectors

A. Can the Public Sector Be Classified According to its Functions?

One possible approach for defining the public sector would be to identify the functions that the public sector is called to perform. In the 2001 version of the IMF’s Government Financial Statistics (GFS) Manual (GFSM2001) and in Eurostat’s European System of Accounts, ESA95 (Eurostat, 2007), the classification of functions of government (COFOG) is used as an internationally-agreed statistical grouping of government expenses.

The problem with using a functional approach is that most spending functions performed by government are also performed by the private sector. For example, the provision of services in health, education, social protection, and environmental protection—four COFOG categories—is shared between the public and private sectors. The extent to which the private or public sector supplies services in these areas varies considerably across countries, depending on political and cultural factors, and consensus on the role of government in society. There are, however, a few exceptions where the public sector has exclusive or near-exclusive functional responsibility, notably for national defense and foreign affairs. However, even for the provision of national defense, some countries may rely on partner countries, multilateral organizations or even the private sector.2

It is, therefore, impossible to delineate the public and private sectors solely on functional grounds, since nearly all functions undertaken by government are also performed by the private sector. It is nonetheless useful to classify government expenses by function, mainly because it allows the “study of the effectiveness of government programs in those areas” (paragraph 6.90, IMF, 2001). Should COFOG categories be applied to the private sector, they would allow comparisons of the effectiveness of private programs in the same functional areas. Such comparisons are useful for determining whether a government should continue to provide services in a particular functional area. In countries with performance-based budget systems, comparative studies and benchmarking are useful for considering changes in fiscal policies, including determining whether the government is the most efficient provider of that function or whether it should withdraw from providing that function.

B. The Concept of Ownership

The concept of ownership is crucial for defining the public and private sectors. Fundamentally, ownership relates to the possession of property, where the owner exercises rights and control over the property. “Property” may be land, real estate, intellectual property, and other assets, either financial or nonfinancial. Ownership implies responsibility for actions regarding the property, whose rights are usually protected under law.

The private sector could be defined as those entities of the economy that are owned by the private sector. Many private sector entities—notably privately-owned enterprises incorporated under law (corporations)—are geared to making profits. As a converse of the private enterprise sector, enterprises that are owned by the government would be part of the public sector. Privately-owned nonprofit organizations and households (appropriately defined) would also need to be included in private sector.

If ownership were the sole criterion for allocating between the two sectors, a four-way division of the economy would result. In Table 1, the first two boxes represent the private sector and the second two, the public sector (shaded).

Table 1.

Conceptual Splitting of Private and Public Sectors

article image

C. The Concept of Control

If there are unclear or differing notions of control, especially of corporations, then the borderline between the private and public sectors would be blurred. For example, an enterprise that is owned mainly be the private sector could have its internal policies set by government, if the latter is a minority shareholding. Economic ownership, in contrast to legal ownership, would need to be distinguished. It is, therefore, necessary to elaborate on understandings of “control”.

Statisticians and accountants have differing notions of government control:

  • Statisticians. A government exercises control of a corporation if the government has the ability to determine the general corporate policy (for a fuller discussion, see Section 4.30 of the System of National Accounts (SNA)—see UN 1993). This concept relates essentially to policy control.

  • Accountants. According to the International Public Sector Accounting Standards Board (IPSASB) of the International Federation of Accountants (IFAC), a government controls a corporation if it has the power to govern its financial and operating policies so as to benefit from its activities (see IFAC, 2008a, and Box 1). This definition focuses particularly on financial control.

Statisticians have recognized limitations of the SNA’s definition of control (Pitzer and Dupuis, 2005). In 2005, it was agreed that the SNA be further elaborated along the lines of the definition of the IPSASB (TFHPSA, 2005). The successive steps for determining control are shown in Figure 1.

Figure 1.
Figure 1.

Establishing Control of Another Entity for Financial Reporting Purposes

Citation: IMF Working Papers 2009, 122; 10.5089/9781451872699.001.A001

Source: p. 191, IFAC, 2008a.

The notion of control by government is not necessarily determined by the share held by the government. If the government’s shareholding in an enterprise exceeds 50 percent, it would be logical to classify it as a public enterprise and include it in the public sector. However, an enterprise in which the government has a minority shareholding but which is controlled, for example, by the government’s representation on its board of directors, or through government instructions requiring that it implement specific policy decisions, would be under government control. Such an enterprise would also be included in the public enterprise sector. To clarify further the different aspects associated with the notion of “control,” the IPSASB has identified various “power” and “benefit” criteria (Box 1). According to IFAC, government control of the entity exists when at least one of the power conditions and one of benefit conditions is exercised. This definition allows the public and private enterprises—and, hence, the public and private sectors—to be clearly defined.

When Does Government Control Exist?

The following guidance is provided by the international accounting profession for defining government control for financial reporting purposes. Control depends on the nature of the relationship between two entities. Two elements are important for the definition of control: the power element and the benefit element.

Power conditions—the power to govern the financial and operating policies of another entity:

  • Majority voting interest. The entity has, directly or indirectly through controlled entities, ownership of a majority voting interest in the other entity.

  • Power to appoint or remove members. The entity has the power to appoint or remove a majority of the members of the board of directors (or equivalent).

  • Majority vote. The entity has the power to cast, or regulate the casting of, a majority of the votes at general meetings or meetings of the board of directors, of the entity.

  • Other power indicators—including those relating to entities’ veto powers of budgets, decisions, and the legal mandates of other entities.

Benefit conditions—the ability of the controlling entity to benefit from the activities of the other entity. These include:

  • Power to dissolve. The entity has the power to dissolve the other entity and obtain a significant level of the residual economic benefits or bear significant obligations, e.g., an entity has responsibility for the residual liabilities of another entity.

  • Control over asset distributions or liabilities. The entity has the power to extract distributions of assets from the other entity, and/or may be liable for certain obligations of the other entity.

  • Government can direct the entity to cooperate with it in achieving its objectives.

  • Other benefit indicators—including holding direct or indirect title to the net assets/equity of the other entity, should also be examined.

For further details, see pp. 185–190 of IFAC, 2008a.

D. Could the Boundary Be Defined by Law?

If the demarcation between the public and private sectors is made by defining control carefully, as proposed above, what role, if any, could be assigned to law in defining, or clarifying the definition of, the public and private sectors? Could a law be adopted to resolve definitional issues in a decisive way? If adopted, would such a law be subject to varying interpretations, still leaving the door open to ambiguities and a perpetuation of a blurred frontier between the public and private sectors?

There are arguments in favor of a legal approach. Laws usually define terms and concepts. When statutes are written, these are made as clear as possible, particularly when there are thorough legal review processes prior to the promulgation of a new law, which help to clarify the law’s intentions. When the main criterion distinguishing the public and private sectors is that of control of an institutional unit by another unit—either public or private—a national law could define the meaning of an “institutional unit” and elaborate on the notion of control. It is particularly for corporations where the frontier between the public and private sectors can be blurred—where economic, but not legal, control exists.

Special legislation could be adopted to create a specific public enterprise. As with general legislation (see next paragraph), laws for specific State-owned companies usually include provisions that indicate the government’s legal rights to appoint directors and exercise control of the enterprise.

State Enterprise Laws provide a generic framework for all public enterprises. These often specify the prerogatives of the State. In some cases, all public enterprises covered by the law are listed.3 State Enterprise Laws may spell out the conditions under which control by the State can be exercised, even if shareholding is less than 50 percent. For example, in Senegal this is done by including the shareholding of any intermediate public entity on a pro rata basis. Also, a financial controller of the Ministry of Finance (MoF) is assigned to the enterprise to ensure that any instructions are followed up (Box 2). A similar provision is in Cambodia’s General Statute on Public Enterprises, 1996, where the enterprise’s controlling ministry is stated to assure the economic control of the enterprise;4 the financial controller is authorized by law to “oppose measures that are against the objectives of the enterprise”.

In contrast, Djibouti’s 1998 Public Enterprise Law defines public enterprises as companies in which the State or other legal entities defined in public law hold more than 50 percent of the shareholding. Although the law states that public enterprises are attached to a ministry that defines the general sector policies, the law makes clear that the enterprises’ policies are decided by the governing board, not the State’s representative. Thus, in Djibouti, the legal definition of public enterprises—and, hence, the public sector—appears to be unambiguous. However, the law also states that the prerogatives of the State are specified in the law’s implementing decree. Thus, for understanding whether the State exercises economic control of public enterprises where it holds less than 50 percent equity, it is necessary to review the dispositions of the implementing decree and how the legal framework is applied in practice.

Senegal: Provisions of the 1990 Law Relating to Public Sector Enterprises

Coverage of “para-public” sector: Included: (1) public companies of an industrial and commercial character (these are regulated by government decree); (2) nationalized companies (these are created by special laws); and (3) companies with majority public shareholding (regulated by private Company Law). Excluded: Administrative public establishments.

Governance. The Governing Board of the enterprise is composed of up to 12 members; the State nominates two board members for their professional expertise; the number of State-appointed Board members is limited to three (“in order to avoid abuse”). Voting is by majority.

Majority ownership. Public participation in the shareholding is calculated as: (1) if the direct public ownership is at least 50 percent, any public share of a parent organization is ignored from calculations of total State shareholding; (2) if the direct public ownership is less than 50 percent, then the public ownership of any additional controlling entity is added, on a pro rata basis, to determine the total State shareholding.

State representatives. Government appointees are not allowed to directly or indirectly intervene in company operations. However, exceptionally, the President of the Republic of Senegal can override these requirements.5

Control. Para-public enterprises are dispensed of all ex ante controls by the MoF. Nonetheless, a MoF controller oversees financial management, especially procurement operations, of each company. He also expresses views on the company’s planned investments. The controller is obliged to prepare periodic reports for the President of the Republic of Senegal, the MoF, the president of the Commission for Verifying and Auditing the Accounts of the Public Enterprise, the Inspection Générale des Finances (a high-level inspectorate under the MoF), and the company’s Chief Executive Officer. The controller must ensure that any Presidential directives to the enterprise’s managers are applied.

Source: The law is available on http://www.courdescomptes.sn/textes/loi9007.htm.

III. A Framework for Defining More Precisely the Public Sector

At a conceptual level, the GFSM2001 framework provides an adequate framework for distinguishing the public and private sectors6 In GFSM2001, the public sector is defined as the sum of units comprising “government” and public enterprises. This section examines “government” more carefully and elaborates on the various components of public sector, especially “general government”. Specific boundary issues, where interpretation of GFSM2001 is needed, are discussed more fully in section IV.

A. Economic Sectors and Institutional Units

In the SNA and GFSM2001, the resident economy is divided into five mutually exclusive sectors: households; the nonfinancial corporations sector; the financial corporations sector; the general government sector (entities that fulfill the general functions7 of government as their primary activity); and nonprofit institutions serving households sector.

Institutional units in each sector have similar objectives, and these objectives are, in turn, different from those of units in other sectors. In the GFSM2001, an institutional unit shows the following characteristics:

  • It is capable in its own right of owning goods and assets; it is able to exchange its ownership of goods and undertake asset transactions with other institutional units.

  • It is able to take economic decisions and engage in economic activities for which it is held directly responsible and accountable by law.

  • It is able to incur liabilities on its own behalf, to take on other obligations or future commitments, and to enter into contracts.

  • A complete set of accounts, including a balance sheet of assets, liabilities, and net worth, exists for an institutional unit. If such balance sheet does not exist, it should be possible, from both an economic and legal viewpoint, to compile a complete set of accounts if these were to be required.

There are two main types of institutional units: (1) persons or groups of persons in the form of households; and (2) legal or social entities whose existence is recognized by law or society independently of the persons or other entities that may own or control them.

B. The Public Sector: Public Corporations and General Government

In GFSM2001, the public sector is comprised, at the broadest level, of public corporations and “general government” (Figure 2). Since public corporations may be either financial or nonfinancial, the public sector therefore encompasses three broad economic sectors: government, financial and nonfinancial.

Figure 2.
Figure 2.

The Public Sector

Citation: IMF Working Papers 2009, 122; 10.5089/9781451872699.001.A001

Source: Figure 2.2, IMF, 2001.

Public Corporations

A corporation is a legal entity created to produce goods and services for the market. Corporations are collectively owned by shareholders who have the authority to appoint directors responsible for its general management. A corporation may be a source of profit or other financial gain to its owners. The profit motive is often the principal reason for establishing a corporation, although this may not necessarily be the case.

Public enterprises are resident8 public corporations that are owned and/or controlled by government units. The division between financial and nonfinancial public corporations is inconsequential for this paper (a country’s central bank and State-owned banks are clearly part of the public sector).

General Government

“General government” produces goods and services on a nonmarket basis, while public corporations are created for the purpose of producing goods and services for the market. The notion of “economically significant prices” is used for distinguishing between market production and nonmarket production (Box 3). The general government sector has two parts: (1) all general government units; and (2) all nonmarket nonprofit institutions (NPIs) that are both controlled and mainly financed by government units.

Market Output and Economically Significant Prices

Market output consists of goods and services that are sold at economically significant prices or intended for sale or otherwise disposed of on the market.

Nonmarket output consists of goods and services that are supplied free or at prices that are not economically significant to other institutional units or the community as a whole.

Economically significant prices are prices that have a significant influence on the amounts the producers are willing to supply and on the amounts purchasers wish to buy. An economically significant price does not have to be so high that all costs of production are covered.

A price that is not economically significant is one that is not quantitatively significant from the point of view of either supply or demand. Noneconomically significant prices may be nonzero. This is the case, for example, for prices charged to raise some revenue or achieve some reduction in the excess demand that occurs when services are provided completely free, but do not completely eliminate excess demand.

Source: IMF, 2001 pp. 10–11.

General government units are institutional units through which government authorities carry out their functions. They have legislative, judicial, or executive authority over other institutional units. General government units are distinguished from public corporations (see Box 4). All general government units are part of the general government sector and they assume responsibility for:

Distinguishing Between a Public Corporation and General Government

Three criteria are used for making the distinction:

  1. If a unit sells most or all of its output and its prices are economically significant, then the unit is a public corporation.

  2. If none of the unit’s prices are economically significant, then the unit is a part of general government.

  3. If only some of the prices are economically significant or if the unit sells only some of its output, it is first possible to identify whether a quasi-corporation9 exists within the unit? If yes, then the organizational component that sells its output at economically significant prices and has a complete set of accounts should be treated as a quasi-corporation. The remaining components would be treated as a general government unit. If it is not possible to identify a quasi-corporation, then the components selling their output at economically significant prices remain an integral part of the general government unit, and the proceeds of their sales are part of the unit’s revenue.

Source: IMF, 2001, paragraph 2.35.
  • providing goods and services to the community as a whole or to individual households on a nonmarket basis;

  • making transfer payments to redistribute income and wealth; and

  • financing their activities, directly or indirectly, mainly by means of taxes and other compulsory transfers from units in other sectors.

Public sector NPIs are legal or social entities created for the purpose of producing goods and services, but which cannot be a source of income, profit, or other financial gain for the units that established, control, or finance them. If a NPI engages in market production, then the NPI must either retain any surplus earned from its productive activities to support its future operations or distribute it to institutional units other than the units that established, control, or finance it. Market NPIs are classified as public corporations sector if they are owned or controlled by government units. Nonmarket NPIs are part of the general government sector if general government units control and mainly finance them.

In GFSM2001, statistical recording is for three levels of government: central; state, provincial, or regional; and local. The central government typically dominates “general government” as central governments provide collective services for the country as a whole, such as foreign affairs, national defense, and national economic management. In some countries, the decentralization of some government functions to lower levels is extensive.

Government entities that are financed fully by budgets adopted by national legislatures make up budgetary central government. A single unit usually encompasses the fundamental activities of the national executive, legislative and judiciary powers. No component is a separate institutional unit if it does not have the authority to own assets, incur liabilities, or engage in transactions in its own right. Budgetary central government holds the authority to exercise control over other, less important, government institutional units. These include autonomous institutional units and extrabudgetary units (those whose transactions are not included in the central government budget). The borderline between “nonmarket” and “market” is far from distinct in many autonomous government agencies, which are authorized to own assets, incur liabilities, or engage in transactions in their own right, e.g., public hospitals, universities, research institutes. The problem of defining subsectors within the public sector, rather than a public-private boundary definitional issue, is discussed below.

Social security administrations constitute general government units of a particular kind. A social security administration must satisfy the general requirements of an institutional unit, i.e., it must be separately organized from the other activities of government units, hold its assets and liabilities separately, and engage in financial transactions on its own account. There are two possible methods of GFS presentation: either by classifying social security administrations according to the level of government that operates them and combining them with other general government units at that level, or by combining all social security administrations into a separate subsector of general government. Employer social insurance funds need to be examined carefully to decide on their attribution to the public or private sector (discussed in Subsection IV.D below).

IV. Specific Boundary Problems Related to the Public and Private Sectors

This section examines the application of the above framework to specific arrangements between institutional units of the private and public sectors—joint ventures, leasing, PPPs, and employer social insurance.

A. Joint Ventures Involving Private Companies and Public Entities

The IFAC has defined and published financial reporting standards for three types of joint ventures (Box 5).

Types of Joint Ventures Between Public and Private Entities

  • Jointly controlled operations. Each venturer uses its own property, plant, equipment, and inventories, incurs its own expenses and liabilities and raises it own finance, which represent its own obligations. A new corporation is not formed. The joint venture agreement usually provides a means by which the revenue from sales or the provision of a joint product and expenses incurred are shared among the venturers, e.g., a State-owned aircraft manufacturer and private sector parts suppliers. The joint venture is not required to prepare its own financial statements; each venturer usually prepares management accounts concerning its interests in the joint venture.

  • Jointly controlled assets. This is similar to above, except that one or more of the venturer’s assets may be jointly controlled or owned for the purpose of obtaining benefits for the venturers. Again, a new corporation (institutional unit) is not established. Nor is there a requirement to prepare financial statements for the joint venture itself. Rather, each venturer recognizes in its own financial statements the share of jointly controlled assets and jointly shared revenues.

  • Jointly controlled entities. A key difference in this case is that the joint venture involves the establishment of a corporation or a partnership in which each venturer has an interest. The jointly controlled entity controls its own assets, incurs its own liabilities or expenses and earns its own revenues. It may enter into contracts in its own name, including raising finance for the joint venture’s activity. It maintains its own accounting records and reports it financial statements, as for similar legal entities. Each venturer is encouraged to recognize its interest in such joint venturers.

Source: IFAC, 2008b.

For joint ventures in which operations and/or assets are controlled jointly by public and private partners, but where a new institutional unit is not created, only minimal demarcation issues arise. For the first two types of joint ventures of Box 5, most of the operations and assets would be attributed to the units that undertake and own them. When sharing of assets and revenues takes place, these can be allocated to the public or private sector on the basis of the accounting conventions for sharing and reporting to the management of each entity. The public sector entity would also recognize its share of any liabilities and expenses incurred jointly in respect of its activities or its jointly controlled asset.

The third type of joint venture shown above has serious implications for the private/public sector frontier: when a joint venture involves the establishment of a new institutional unit, a single set of consolidated accounting statements needs to be prepared by the new entity.

However, for reporting to its shareholders, participating private companies can report its involvement in joint ventures, based on its shares in the financial statements of the new entity, which include the proportionate shares of revenues, expenses, assets, liabilities, and cash flows. To allocate the operations and balance sheet items of jointly controlled entities to the public or private sectors, the IFAC recommends using the proportionate method, as it reflects the substance and economic reality of a venturer’s interest in a jointly controlled entity. The equity method, although allowed, is not preferred (see Sections 35 to 46 of IPSAS 8 for details, IFAC, 2008b).

B. Public Sector Leasing from the Private Sector

A government can lease a privately-owned asset in two ways. First, it can take an operating lease (see Box 6 for a definition), under which payments by the government for services provided by the private operator would be recorded as a government operating expense. Any fees, lease payments, or other payments by private operators of concessions entered into with the government would be recorded as government operating revenue. The asset would be recorded on the private partner’s balance sheet. The second method is for the government to enter a financial lease arrangement with the private partner, under which the government acquires an asset. In this case, both the asset and the lease liability to the private sector would be recorded on the government balance sheet. Subsequent depreciation and any interest would be recorded in the government’s operating statement (assuming that the government has moved to an accrual basis for accounting).

Operating Leases and Financial Leases

Operating leasing is a productive activity that involves renting fixed assets for periods of time less than the expected service lives of the assets. It is a form of production in which the lessor provides a service to the lessee in exchange for lease payments. The lessor is typically responsible for the maintenance and repair of the equipment as part of the service provided to the lessee. When the government is the lessee, for accounting purposes, the lease payments are treated as a government expense.

Financial leasing is an arrangement for financing acquisitions of fixed assets. A contract is concluded between a lessor and a lessee whereby the lessor owns a fixed asset and puts it at the disposal of the lessee, and the lessee contracts to pay rentals that permit the lessor to recover all or almost all of its costs, including interest. As a result, several risks of ownership pass from the lessor to the lessee. To capture the economic reality, under accrual accounting, a change of ownership from the lessor to the lessee is deemed to take place, even though the leased good legally remains the property of the lessor until the termination of the lease, at which time the legal ownership is usually transferred to the lessee. In GFSM2001, which is based on accrual accounting, the private lessor is treated as having sold the asset to the government lessee and financed the sale with a loan.

Source: IMF, 2001, Appendix 2.

Financial leases are sometimes used by governments to obtain major capital equipment such as airplanes (PPPs—discussed in next subsection—also resemble financial leases in several respects). However, operating leasing arrangements are much more common. In the United States, for example, over 10 percent of all office space of federal agencies is leased, under operating leasing arrangements. The private sector is the owner10 of the office building from which the federal government rents space. There is no ambiguity concerning the public-private sector ownership of the underlying asset, even though such operating leases are quite costly compared with the alternatives of construction or lease-purchase.11

More generally, when the private sector is the owner of the asset from which the government obtains benefits and the government controls the operation under an operating lease, there is a strong risk that some financial commitments of the government are excluded in the annual budget and accounts. Operating leasing contracts (under cash accounting) give rise to a perception that they are low- or zero-cost arrangements, since the annual charges are recognized and paid over time, rather than up-front.

C. Public-Private Partnerships

PPPs have become a popular method of financing large-scale infrastructure projects (Box 7). In an increasing number of countries (see IMF, 2006, for a selective review), PPP contracts have been drawn up in economic sectors (e.g., transportation—roads, railways, tunnels), social sectors (e.g., education and health), and general public administration (e.g., prisons).

Public-Private Partnerships—What Are They?

PPPs often refer to arrangements where the private sector supplies infrastructure assets and services. In traditional acquisition of government assets or provision of services, the government provides financing, and retains all risks and responsibilities related to the infrastructure project. Under PPPs, the financing, risks and responsibilities are shared between the public and private sectors. Detailed contracts specify PPP arrangements.

Most PPPs include the following characteristics:

  • Private execution and financing of public investment.

  • The private sector is the legal owner of the infrastructure asset.

  • The service provision from the infrastructure asset is provided by the private sector.

  • Some risks of the infrastructure project are transferred to the government.

When contemplating infrastructure projects and their related services, governments face a trade-off between quality and efficiency. The theory of ownership and contracting can provide an analytical justification for PPPs (see IMF, 2006, for a review). The main argument in favor of PPPs relates to their potential efficiency gains. It is perceived that the government can benefit from private sector management and obtain better value for money than by embarking on an infrastructure project by itself, using traditional public ownership and procurement methods (Grimsey and Lewis, 2004; OECD, 2008). However, successful PPPs require efficiency gains to be large enough to cover the additional private sector borrowing and transaction costs.

PPPs are often quite complex and the risks associated with PPP projects are numerous (Box 8). The degree of private sector involvement and risk-sharing in PPPs varies: from service contracts or management contracts (low private sector involvement), design-build-finance-operate-maintain arrangements (intermediate private involvement), and build-own-operate-(transfer) arrangements (high private involvement) (see Chart A of IFAC 2008c). Risks for the government are particularly high for PPPs that are poorly-planned or executed.

Risks in PPP Projects

Seven PPP project-related risks can be identified, some of which overlap:

  • Construction Risk: related to design problems, building cost overruns and project delays.

  • Financial Risk: related to project financing problems, for instance due to interest and exchange rate variability.

  • Availability Risk: related to the continuity of quality service provision, which in turn depends on the availability of a well-maintained PPP asset.

  • Demand Risk: related to the ongoing demand for the project’s services.

  • Political Risk: where government action could impair the private partner’s earnings potential.

  • Residual Value Risk: uncertainty regarding the market price of the infrastructure asset at the end of the contract period (when private asset ownership reverts the public sector).

  • Force majeure: risks beyond the control of PPP partners (e.g., natural disasters).

Source: IMF, 2006.

In many PPPs, capital investment is financed by a private sector entity, which is the legal owner of the asset, at least for a predetermined time period. If the PPP takes the form of an operating lease, the PPP asset would be recorded on the private partner’s balance sheet. If the PPP is in the nature of a financial lease, the government’s acquires an asset. Both the asset and the lease liability to the private sector would be recorded on the government balance sheet. Most PPPs are set up as operating leases, not as financial leases.

For infrastructure projects under typical PPPs, the assets are often built and legally owned by the private operator. However, the government, rather than the private owner, bears many of the risks associated with ownership. This raises the question as to whether the government, although not the legal owner, is the de facto economic owner. If yes, then should the associated asset be recorded on the government’s balance sheet, in full or in part? If so, when: once the PPP project becomes operational? Or at the end of the contractual period, when the private sector’s depreciated asset is transferred to the public sector? These questions raise complex accounting issues, which are in the course of being resolved.

In 2004, Eurostat issued a decision classifying the assets of PPP projects as public or private, based on which risks are borne by the private sector. According to Eurostat, the assets of PPP projects are classified as private assets and recorded in the balance sheets of the private partner if the private partner bears: (i) the construction risk; and (ii) either the availability or the demand risk (Eurostat, 2004). Since the private sector typically bears the construction and availability risks, under the Eurostat guideline, there is a strong incentive for governments to design PPPs to meet the two criteria, resulting in PPPs not being recorded in governments’ financial statements. This is despite a strong risk (for some projects) that the government is the economic owner of the asset: once the PPP project becomes operational, the government may begin bearing some of the nonconstruction risks listed in Box 8. This is likely for PPP projects in which the government guarantees the private partner against the risk of default.

A more general question is: should risks be used as the key criterion for PPP accounting purposes? To begin addressing these issues, the IPSASB of IFAC issued a consultation paper in March 2008. Once finalized after August 2008, an internationally-agreed standard for treating specific PPP asset ownership and other accounting issues will be established. A summary of the paper and the key proposals put forward by IPSASB are shown in Box 9.

The IPSASB paper focuses on accounting issues related to PPPs that combine a construction element with an operations concession element—termed “service concession arrangements” (SCAs). The accounting for SCAs is particularly difficult because these arrangements involve a relatively even sharing between the government and the private constructor-operator of the various risks, responsibilities, benefits, control of the underlying infrastructure, and delivery of the associated services. The IPSASB aims to clarify which party should report the underlying infrastructure as an asset in their financial statements.

The key proposal of IPSASB is that the government should include the underlying asset in its financial statements if it is considered to control the property. The criteria for determining “control” are crucial. Two criteria are proposed for public sector control, notably if the public sector entity:

  • Controls or regulates what services the operator must provide with the underlying property, to whom it must provide them, and the price ranges or rates that can be charged for services; and

  • Controls—through ownership, beneficial entitlement or otherwise—the residual interest in the property at the end of the arrangement.

IPSASB Proposals for Accounting for PPPs

The IPSASB consultation paper:

  • Provides an overview of the various PPP arrangements, of which SCAs are a subset. Lack of specific guidance for SCAs has caused divergence in how the property in SCAs is reported, even occasionally resulting in the property not being reported by either the public sector entity or the private sector entity.

  • Notes that governments have used SCAs to fulfill their infrastructure needs while not recognizing the property and related financing in their financial statements.

  • Addresses whether the public sector entity should report the underlying property as an asset. The main issues relates to the timing of the recognition and measurement of the underlying property, as well as any associated liabilities.

  • Examines three other issues related to SCAs: (1) the recognition of revenue from resource inflows; (2) guarantees and commitments made by the public sector entity; and (3) financial statement note disclosures.

Proposals of the IPSASB consultation paper. The government should:

  • Report the property underlying an SCA as an asset in its financial statements if it is considered to control the property.

  • Measure the government’s asset and any related liability (obligations to provide compensation to the operator), based on the fair value of the property, except in cases where scheduled payments made by the government can be separated into a construction element and a service element.

  • Recognize the SCA’s contractually determined inflows of resources to be received by a government from an operator as government revenue as they are earned over the life of the SCA, beginning when the underlying property is fully operational.

Source: IFAC, 2008c.

In making its proposals for accounting for SCAs, the IPSASB considered alternatives, such as basing financial reporting of the asset on: (1) risks—as recommended by Eurostat in 2004; (2) the access to the (future economic) benefits and exposure to the associated risks—as used by the United Kingdom,12 for example; or (3) unbundling of the components—as suggested by Hemming and others.13 The third option would require quantification of the risks, rewards and benefits, and a splitting of the value of assets and liabilities to the private and public sector partners. While conceptually appealing compared with the binary approach of attributing assets to only one of the two sectors, this suggestion would be extremely difficult to apply in practice, especially since significant value judgments would be required to allocate a portion of the PPP asset to the private sector and another part to the public sector.

Irrespective of how the accounting issues are resolved, given the risk-sharing nature of PPP projects and the long-term commitments of governments under PPP contracts, it is appropriate to assess and quantify the risks borne by the government and to disclose all associated risks, especially those arising from PPPs and government guarantees. In this context, Fiscal Risk Statements are helpful for discussing the future medium-term financial implications of contingent liabilities and any other potential costs of PPP projects to the government (see also Boxes 6 and 7 of the IMF’s Manual on Fiscal Transparency, IMF, 2007). Such statements should identify the principal financial hazards and quantify them if possible. A fiscal risk statement can be prepared and presented to parliament in a document accompanying the draft annual budget law, thereby enhancing the transparency of government financial operations.

D. Social Insurance Schemes

According to the GFSM2001, social protection can be organized as social assistance schemes, social security schemes, or employer social insurance schemes. Social assistance schemes are organized and operated by government units; there is no earmarking of revenue for social assistance benefits, which are treated as an expense in general government accounts. Social security schemes in GFSM2001 are social insurance schemes that are imposed, controlled and financed by government. By definition they are organized and operated only by government units. In many countries, entire institutional units are devoted to the operation of a social security scheme. Such units—social security funds—are special types of government units. For a social security fund to exist, it must be separately organized from other government units, hold its assets and liabilities separately, and engage in financial transactions on its own account. For social assistance and social security schemes, the payment of some benefits, e.g., invalidity benefits or retirement benefits, may be spread over several accounting periods—a liability for future payments has been incurred. Under accrual accounting, the present value of those benefits should be recorded on the balance sheet of the government. For this paper, the key point regarding these two types of schemes is that they unambiguously belong to the public sector. In contrast, the allocation of employer social insurance schemes to the public or private sectors is less clear, in part because there are several types of schemes, notably funded and unfunded, and the private sector may be contracted to manage the scheme.

A social insurance scheme is funded if there are identified reserves or accounts assigned for the payment of benefits. Funded employer schemes can be operated by: (1) insurance enterprises; (2) autonomous pension funds; or (3) non-autonomous pension funds. Unfunded social insurance schemes are operated by the employer without assigning specific accounts or creating special reserves for the payment of benefits. Instead, the benefits are paid from the employer’s general resources.

If a public sector unit organizes an employer social insurance scheme but contracts with an insurance enterprise for its operation, then the general government unit or public corporation that is the employer will pay the required social contributions to the insurance enterprise on behalf of its employees. All other transactions of the social insurance scheme will be the responsibility of the insurance enterprise. If the insurance enterprise is a private corporation, then its transactions, such as the payment of benefits, administrative expenses, investment of assets, and incurrence and liquidation of retirement and other liabilities, will be recorded in its financial statement—the accounts of the public sector will not be affected. If, on the other hand, the insurance enterprise is a public corporation, then the accounts of the public sector would be affected by the operations of the insurance scheme (although there would be no impact on the accounts of the general government sector).

Thus, when governments, as an employer, set up social insurance schemes for its employees, and a privately-owned insurance enterprise manages the government’s scheme, all of the assets and liabilities would be recorded in the accounts of the private financial corporation. Only the government’s expenses would be recorded in its operating and cash flow statements. The same would apply if an autonomous pension fund were to operate the pension scheme for civil servants (and/or military). In contrast, if a non-autonomous (i.e., a government owned) pension fund operated the scheme, all assets, liabilities, revenues, and expenses would be recorded on the government unit’s own accounts.

V. The Boundary Between the Public Sector and General Government

Applying the principles and definitions shown above in Boxes 3 and 4 to distinguish “general government” from “public corporations” proves to be difficult in practice. The main difficulty is in distinguishing “nonmarket” and “market” producers for each government-controlled entity.

A. Extrabudgetary Funds

In addition to social insurance funds (discussed above), countries establish other extrabudgetary government entities to carry out specific functions (e.g., road construction; health services, higher education). These institutional units typically have their own off-budget revenues and make expenditures that are not necessarily included in the government’s annual budget appropriations (although these units’ resources are often supplemented by transfers from the national budget). Managers of extrabudgetary funds often have discretion over the volume and composition of their spending.

Extrabudgetary entities can be classified to be in the general government sector or in the public corporations sector. Clearly extrabudgetary entities that they are nonmarket producers and controlled by another government unit belong to general government, even when legally established as separate entities and somewhat independent from central government ministries. Governments may choose to use such nonmarket producers, rather than government ministries, to carry out certain government policies, since they can be detached, objective and, in to large degree, free from political pressures, e.g., for research and development; setting of standards in health, education, safety, and the environment.

If an extrabudgetary entity is producing market output, it fulfills the GFSM2001 criterion to be a market establishment. Market establishments that have: (1) complete accounting records about it production activities, but not (2) a complete balance sheet, nor (3) the ability to engage in financial transactions in its own name, are deemed to belong to “general government”. Market establishments that also satisfy the last two qualities are deemed to be quasi-corporations and, as such, would belong to the public corporation sector. For example, government-owned universities or public hospitals that are market producers—i.e., charge fees based on the costs of producing the services or that are sufficiently high to have a significant impact on the demand for services provided—would be classified as institutional units of the general government sector, not the nonfinancial corporations sector, when it has accounting records of its output and costs of production, but does not satisfy criteria (2) and (3) above. In contrast, a public road authority that receives road tolls in its own name, affects the demand for usage of the road by the size of its fees, and has full accounting records including a balance sheet, would be classified in the public nonfinancial corporations sector (even if unincorporated).

In practice, the distinction between “nonmarket” and “market” is difficult to apply. For example, a university that is controlled by the government would be “nonmarket” if it does not charge fees or charges only nominal fees that do not significantly affect demand for its services. Even a modest fee would affect demand coming from students of low-income families. A first question relate to whether it is possible to determine the threshold level of tuition fees, below which the entity would be classified as “nonmarket”. Even if a market establishment is identified, a second set of questions pertain to whether or not the entity can act in its own name and has a complete balance sheet, i.e., would be regarded as a quasi-corporation, not part of general government.

To apply the “market” versus “nonmarket” categorization for government entities, Eurostat has established a rule for use in European Union countries (Box 10). In other countries, where reports are prepared for compiling the “general government” statistics (e.g., those published in IMF’s GFS Yearbook), agencies may be categorized according to their legal status, as this is an easier option. The extent of misreporting from this source is not known, although it is generally thought to be low.

The European Union’s 50 Percent Rule

In the ESA95, the concept of economically significant prices is implemented on the basis of whether or not production costs covered by sales covers more than 50 percent.

In distinguishing market and nonmarket producers by means of the 50 percent criterion, “sales” and “production costs” are defined as follows:

  • “Sales” cover the sales excluding taxes on products but including all payments made by general government that are linked to the volume or value of output. Payments to cover an overall deficit are excluded.

  • “Production costs” are the sum of intermediate consumption, compensation of employees, consumption of fixed capital and other taxes on production. Other subsidies on production are not deducted. To ensure consistency of the concepts of sales and production costs when applying the 50 percent criterion, the production costs exclude all costs made for own-account capital formation.

The 50 percent criterion is applied by looking over a range of years. The criterion is applied when it holds for several years or holds for the present year and is expected to hold for the near future.

Source: Section 5.2, p. 14, Eurostat, 2002.

B. Autonomous Government Agencies

Various types of autonomous agencies have emerged in recent years as instruments of new public management. The classification of many of these agencies within the “general government sector” can be problematic. In many cases, such agencies are expected to behave in a “commercial” manner. For example, they are authorized to raise their own revenues by charging for the services they provide. In some OECD countries, budgetary appropriations are voted by parliament on a net basis—ministries’ gross expenditures minus the operating revenues of the ministries or units they control. By allowing agencies to retain the revenues they raise—and to spend them—agencies have an incentive to maximize their income, which also forces users to pay for government-provided services.

In the United Kingdom, for example, public management reforms led to the creation of over 130 executive agencies, established to carry out functions of government within a policy and resources framework set by the parent department. Executive agencies are required to publish their corporate and business plans, to prepare annual reports and financial accounts, and present these to parliament after audit by the National Audit Office. Few executive agencies have their own legal identity. Although many executive agencies appear to fulfill several of the criteria for institutional units (described in section III.A above), such agencies are usually not directly responsible and accountable by law for their outcomes, since the parent minister/ministry establish agency policies. They are therefore appropriately classified as part of “budgetary central government.”

A different situation arises for institutional units whose functions and policies are at a greater distance from ministerial control. Again taking the United Kingdom as an example, a nondepartment public body (NDPB) is an entity that is not a government department or part of one, and which operates independently of ministers, even though they are ultimately responsible to a minister. In contrast to executive agencies, non-advisory NDPBs have their own legal identity and their own governing boards, although appointments to their boards are made by ministers (or officials on behalf of ministers).

When OECD attempted to classify the diverse arrangements for “autonomous” agencies in member countries, it identified several distinguishing features, including legal status, governance structures, staffing rules, and funding (Table 2).

Table 2.

Different Types of Autonomous Public Bodies

article image
Source: pp. 13–19, OECD, 2002.

“Autonomous” agencies also differ according to the extent to which the government exercises policy control over their operations—via general guidelines, representation on governing boards, and reporting requirements to “parent” ministries and to parliament. To clarify these issues for the various types of agencies in New Zealand, an agencies (“Crown Entities”) law was adopted in 2004 (Box 11). Although this law provides a clear framework for categorizing agencies according to the extent of policy control by government Ministers (or parent ministries), it is not possible in this law to identify which agencies charge “economically significantly” prices for their outputs and which do not. For example, a privately-owned school partly funded by government would, in the GFSM2001 framework, be classified as a private NPI (if it is deemed to be operating on a nonmarket basis) or a private quasi-corporation (if it charges “economically significant” prices). In either case, the ownership criterion would require it to be part of the private sector. In contrast, a public-owned school would be classified as a public NPI (and hence “general government”) or a public quasi-enterprise (and hence the public corporations sector) in the GFSM2001. Thus, for properly allocating schools to the private or public sectors—and, more generally, all Crown Entities covered under the scope of the law—it would be necessary to distinguish the entities according to the ownership and control criteria. This case provides another example of the limitations of law in attributing institutional units to the public or private sectors—as defined in the GFSM2001 framework.

New Zealand: Categories of Government Agencies

The Crown Entities Act, 2004 provides the legal basis for the following categories of agencies:

  • Government agent. These are entities whose service delivery is very closely intertwined with government policy and undertake non-independent executive functions.

  • School Boards of Trustees and Tertiary Education Institutions. These are a special category of “government agent” that have their own governance and accountability regimes, and are not necessarily government-owned (e.g., private schools that are funded primarily by government).

  • Autonomous government entities. These entities are required to give effect to, or have regard for, the policies of the government, but do not need to be subject to a high degree of ministerial control.

  • Independent government entities. For decision-making, these entities operate with independence from government ministers. They are not required to give effect to, or have regard for, the policies of the government.

  • Government-owned companies. These are companies registered under the Crown Entities Act, rather than the general Companies Act, because the company’s objectives are not exclusively commercial. Examples include Radio/Television New Zealand and New Zealand Venture Capital Fund Limited.

The regulations accompanying the 2004 law specifies the agencies’ financial reporting requirements, including any exceptions e.g., with respect to reporting bank accounts; borrowing; government guarantees; and paying surpluses to the State.

Sources: p. 145, Gill, 2002; New Zealand Treasury, 2006.

VI. Conclusions

The boundary between the public and private sectors, at a conceptual level, can be defined unambiguously. After ruling out the impossibility of defining the public sector on the basis of the functions it undertakes, this paper emphasizes the concept of ownership as the basis for delineating the two sectors. The notion of an “institutional unit” is also fundamental.

The concept of ownership is critical in defining the institutional units that belong to the public or private sectors. International accounting standards give precedence to economic ownership over legal ownership. Economic ownership is exercised by a controlling entity when it has the power to govern the financial and operating policies of the other entity. Two conditions are essential for economic control: first, at least one “power” condition, such as the majority voting interest or the power to appoint or remove governing board members, is required; second, at least one “benefit” condition, such as the power to dissolve the entity or to control asset distributions, is also needed.

Laws can elaborate on the notion of control. However, law does not provide an internationally-acceptable definition of the two sectors. Unlike the accounting profession, which has established international norms useful for delineating the two sectors, no international law body has attempted to prepare international legal guidelines for defining the public and private sectors. For corporations that are perceived to be on the borderline between the public and private sectors, a review of public enterprise laws and their implementing decrees is needed for understanding how the legal framework clarifies or blurs the delineation of the two sectors. If, for example, the law does not explicitly proscribe political intervention into the operations of a public corporation in which the government has minority shareholding, it is still possible for the government to exercise economic (or de facto) control of the enterprise, even when de jure control is by the private sector.

Institutional units engaged in economic activities, take economic decisions in their own name, own goods and assets, incur liabilities, and are able to compile a complete set of accounts. In GFMS2001, the central government of a country is one such institutional unit.

The public sector is comprised of “general government”—central and local governments, and all extrabudgetary funds controlled by them—and all public corporations. The private sector comprises households, private corporations, and privately-owned nonprofit organizations. The GFSM2001 provides an adequate conceptual framework for demarcating the public sector, which is comprised of all nonmarket government-owned entities and corporations (including quasi-corporations—corporations that are not legally incorporated) that are owned or controlled by government units.

When it comes to applying the GFSM2001 framework, guidelines are needed. In this context, the IPSASB of the International Federation of Accountants has issued various guidelines useful for demarcating the two sectors and measuring them in a mutually exclusive manner. For joint ventures in which the public and private sector partners establish a jointly-controlled entity, accounting standards permit the operations and balance sheets of the entity to be allocated partly to the public sector and partly to the private sector. For PPPs, accounting norms propose standards for the most difficult-to-categorize PPPs—those that involve a construction element (usually the private partner constructs an asset on behalf of the public sector) and a concession element (under which the public sector leases assets from the private partner). As in public-private financial leasing arrangements, if the public sector is considered to control the property (even if the private sector is the legal owner) the public sector should report the underlying property in its financial statements. Thus, accounting norms enable the transactions, assets and liabilities of institutional units involved in PPPs to be reported clearly, in the accounts of either the private sector entity or in the public sector.

Social insurance funds can belong to either the public or private sectors. One borderline issue relates to civil service pension schemes. When a privately-owned insurance enterprise manages a funded pension scheme of government employees (to which the government contributes), then all of the assets and liabilities would be on the balance sheet of the private company. In many countries, however, government employee pension schemes are not autonomous, not (fully) funded, and are recorded entirely in the government’s accounts.

It could be argued that the risks for the government of some PPPs are extensive and that, even if the accounting standards appear to delineate the two sectors adequately in the present, there is still considerable likelihood that a semi-private/semi-public leviathan will emerge from the grayness of accounting backwaters and record its “rightful” place in the government’s expense account. This fear for the future is understandable and justified. The asymmetrical risk-sharing involved in many PPP projects—and the accounting standards for allocating these risks in the present—may be reversed in the future. This is case especially for PPPs with risks for the government that are not adequately captured in the accounts of either the public or private sectors. A government guarantee is one particular contingent liability that may induce large budgetary payments in the future that is not captured in today’s government accounts.

The existence of such risks to the government’s future budget and accounts is acknowledged. However, when the risk eventuates, there will be a transfer of ownership. This will induce a transfer of the recording of assets and, especially, of liabilities. For example, if the private sector partner of a PPP defaults on a loan from its commercial bank and the government assumes the guaranteed loan, then the net worth in the government’s balance sheet would deteriorate by the amount of the guaranteed loan. This would enlarge the balance sheet of the public sector, but it would not alter the definition of the two sectors.

Given that governments may lose assets or gain liabilities in the future, it is desirable for the governments to monitor, disclose and manage all fiscal risks, so as to minimize undesirable future pressure on its budget and financial statements. In this context, guidelines of good practice have been published14. These include the regular publication, at the time the annual budget is presented to parliament, of a Statement on Fiscal Risks.

This paper argues that boundary problems within the public sector are particularly acute. This is because of ambiguities in distinguishing “market” from “nonmarket” activities. However, as with the delineation of the frontier between the public and private sectors, this problem also can be overcome by establishing rules such as those of ESA95.

Establishing, then applying, accounting rules is a pragmatic way to unravel the complex realities induced by joint public-private ownership and de facto government control. What appears at first sight to be an undefined gray zone between the public and private sectors, can in fact be split into component parts and measured without ambiguity, admittedly by rules that, at the detailed level, contain a degree of arbitrariness.


  • Blöndal, Jón R., 2005, Market-type Mechanisms and the Provision of Public Services, OECD Journal on Budgeting, Vol. 5, No. 1, pages 79-106 (Paris: OECD).

    • Search Google Scholar
    • Export Citation
  • Chan, James, 2002, “Models of Public Budgeting and Accounting Reform”, OECD Journal on Budgeting, Vol. 2, Supplement 1 (Paris: OECD, 2002).

    • Search Google Scholar
    • Export Citation
  • Eurostat, 2002, ESA95 (European System of Accounts 95) Manual on Government Deficit and Debt (Luxembourg: Office for Official Publications of the European Communities).

    • Search Google Scholar
    • Export Citation
  • Eurostat, 2004, Treatment of Public-Private Partnerships, News Release 18/2004, February 11.

  • Eurostat, 2007, Manual on Sources and Methods for the Compilation of COFOG Statistics: Classification of the Functions of Government (COFOG), Eurostat Methodologies and Working Papers (Luxembourg: Office for Official Publications of the European Communities).

    • Search Google Scholar
    • Export Citation
  • GAO, 2005, Federal Real Property: Reliance on Costly Leasing to Meet New Space Needs Is an Ongoing Problem, Statement of Mark L. Goldstein, Director, Physical Infrastructure Issues, United States Government Accountability Office (GAO), GAO-06-136T, October http://www.gao.gov/docsearch/abstract.php?rptno=GAO-06-136T

    • Search Google Scholar
    • Export Citation
  • Gill, Derek, 2002, “New Zealand”, Chapter 7 of OECD, 2002, op. cit.

  • Grimsey, Darrin and Mervyn K. Lewis, 2004, Public Private Partnerships: The Worldwide Revolution in Infrastructure Provision and Project Finance (Cheltenham: Edward Elgar)

    • Search Google Scholar
    • Export Citation
  • IFAC, 2008a, 2008 Handbook of International Public Sector Accounting Pronouncements - Volume I, International Public Sector Accounting Standards Board, (New York: International Federation of Accountants) http://www.ifac.org/Members/Pubs-Downloading.tmpl?PubID=120915713024448&File=Public_Sector/2008_ IPSASB_Handbook_Combined_Vol_I.pdf&Category=Public%20Sector%20Accounting

    • Search Google Scholar
    • Export Citation
  • IFAC, 2008b Interests in Joint Ventures, International Public Sector Accounting Standard (IPSAS) 8, International Public Sector Accounting Standards Board (New York: International Federation of Accountants) http://www.ifac.org/Members/Pubs-Downloading.tmpl?PubID=120915713024448&File=Public_Sector/2008_IPSASB_Handbook_Combined_Vol_I.pdf&Category=Public%20Sector%20Accounting

    • Search Google Scholar
    • Export Citation
  • IFAC, 2008c, Accounting and Financial Reporting for Service Concession Arrangements, Consultation Paper, IPSAS Board, http://www.ifac.org/Guidance/EXD-Details.php?EDID=0109 (New York: International Federation of Accountants)

    • Search Google Scholar
    • Export Citation
  • IMF, 1986, Manual on Government Finance Statistics (Washington: International Monetary Fund).

  • IMF, 2001, Government Finance Statistics Manual (Washington: International Monetary Fund).

  • IMF, 2006, Public-Private Partnerships, Government Guarantees and Fiscal Risk, staff team lead by Richard Hemming (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • IMF, 2007, Manual on Fiscal Transparency http://www.imf.org/external/np/fad/trans/index.htm (Washington: International Monetary Fund).

  • IMF, 2008, Fiscal Risks – Sources, Disclosure, and Management, available on http://www.imf.org/external/pp/longres.aspx?id=4265 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • New Zealand Treasury, 2006, Financial Powers Provisions of the Crown Entities Act 2004, Wellington, http://www.treasury.govt.nz/crownentities/#legislation

    • Search Google Scholar
    • Export Citation
  • OECD, 2002, Distributed Public Governance: Agencies, Authorities and Other Public Bodies (Paris: OECD)

  • OECD, 2003, Accrual Accounting and Budgeting: Key Issues and Recent Developments Issues in Accrual Budgeting, OECD Journal of Budgeting, Vol. 3, No. 1, (Paris: OECD).

    • Search Google Scholar
    • Export Citation
  • OECD, 2008, Public-Private Partnerships: In Pursuit of Risk Sharing and Value for Money (Paris: OECD)

  • Pitzer, John, 2004, “Government/Public Sectors/Private Sector—Delineation Issues”, paper for the meeting of the Task Force on Harmonization of Public Sector Accounting, February 6-11, http://www.imf.org/external/np/sta/tfhpsa/2005/09/092705.htm

    • Search Google Scholar
    • Export Citation
  • Pitzer, John, and Jean-Pierre Dupuis, 2006, “The General Government and Public Sectors”, paper presented to the fifth meeting of the Task Force on Harmonization of Public Sector Accounting, March 8-10, http://www.imf.org/external/np/sta/tfhpsa/2006/03/pdf/govern.pdf.

    • Search Google Scholar
    • Export Citation
  • TFHPSA (Task Force on the Harmonization of Public Sector Accounting), 2005, “Government/Public Sector/Private Sector Delineation Issues”, SNA/M1.05/9.1, Advisory Expert Group for Update of the 1993 SNA, http://www.imf.org/external/np/sta/tfhpsa/2005/09/092705.htm

    • Search Google Scholar
    • Export Citation
  • UN (United Nations), 1993, System of National Accounts (New York: United Nations).


This paper was first presented at the 16th International Symposium of the Review Politiques et Management Public, held in Florence, Italy, March 2007. For this revised version, the author wishes to thank particularly V. Marie (STA), R. Allen, M. Lazare (FAD) and other IMF colleagues who provided useful comments on earlier versions of the paper. Remaining errors are the author’s.


For example, in Haiti, whose army was abolished in 1994, the defense function is provided in part by the national police and in part by UN forces. Concerning the private sector, defense logistics is increasingly being contracted out to private companies which, in some cases, have assumed responsibility for protection of sensitive buildings or important persons.


See, for example, Djibouti’s 1998 law on National Companies, Public-Private Companies, and Commercial Public Establishments, whose annex lists 15 public enterprises, mostly utility, transport- and food-related companies, and also a development bank. Source: a website with public enterprise laws for various francophone countries http://droit.francophonie.org/df-web/publication.do?publicationId=2998.


See article 20 of Loi portant statut général des entreprises publiques, available on website of footnote 3,


Article 13 of Law 90-07 states: Interdiction est faite aux administrateurs représentant l’État de prendre ou de conserver un intérêt direct ou indirect dans une opération effectuée par l’entreprise pour son compte ou par un organisme dans lequel celle-ci aurait une participation financière. Toutefois, à titre exceptionnel, une décision expresse du Président de la République peut déroger aux dispositions du présent alinéa.


Although the GFS Manual is not accompanied by an implementation manual that would elaborate further on conceptual issues, there are nonetheless a number of notes on specific issues, including on the coverage and sectorization of the public sector; compilation and consolidation; and reclassification of GFSM 1986 data. See http://www.imf.org/external/pubs/ft/gfs/manual/comp.htm.


Paragraph 2.1 of GFSM2001 specifies the generic economic functions of a government, namely to: (1) provide goods and services to the community on a nonmarket basis; and (2) redistribute income and wealth.


This paper does not discuss differences in treatment of nonresident public enterprises between GFS and IPSAS. Foreign subsidiaries of public corporations are not part of the public sector in SNA, whereas they are included in the IPSASs (see Pitzer, 2004).


A quasi-corporation is defined as entities that are not incorporated or otherwise legally established, but which function as if they were corporations. In the GFS system, they are institutional units separate from the units to which they legally belong.


Exceptions occur if a federal agency rents space from a government business enterprise that owns rental buildings. In such cases, the leasing arrangement would resemble a public-public partnership.


See GAO, 2005, which cites the Patent and Trademark Office’s operating lease for long-term space needs costing $48 million more than construction and $38 million more than lease-purchase. Under cash accounting, operating leases look cheaper than the ownership option, which would be recorded fully upfront in the budget. This has resulted in an over-reliance on operating leases for federal agencies’ long-term space needs. GAO also notes that many assets have alarmingly deteriorated; restoration and repairs would require billions of dollars.


See pp. 17–20 of IFAC, 2008c for a discussion of the key elements of accounting standards for reporting “private finance initiatives” (the U.K.’s term of PPPs that are of the nature of a SCA).


See pp. 27–28 of IMF, 2006.


See pp. 32–25 of IMF, 2008.