But over the past 25 years four developments have produced far-reaching changes in approaches to and perspectives on external assistance. First, it was recognized that all resources are fungible-especially convertible resources. Thus aid ostensibly earmarked for a well-prepared, well-supervised donor project could result in government resources being used to finance another project about which the donor knew nothing. (The notion of resource fungibility was first formulated by de Vries 1967 and subsequently elaborated by Schiavo-Campo and Singer 1970; Devarajan and Swaroop 2000 recently reviewed these arguments.)

Moreover, safeguards to ensure the integrity of donor resources provided to a project meant little in the absence of safeguards on the use of government resources effectively released by the aid. Thus the World Bank and other donors began to consider it necessary, when financing a large share of a countrys public investment, to appraise the entire investment portfolio rather than just aid-financed projects. The Banks resulting Public Investment Reviews were also useful to other development agencies in the context of their project-centered assistance. In addition, because Public Investment Reviews required identifiable public investment portfolios, they led to the provision of technical assistance in formulating public investment programs-which quickly became standard in most developing countries.2

Second, in line with the global shift away from central planning and mounting evidence of the importance of social factors for development, the distinction between development and nondevelopment spending lost meaning. It became accepted that new teachers and new textbooks are just as crucial for development as new schools, and thus that current spending and investment form (or should form) an integrated whole. Hence, the World Banks Public Investment Reviews expanded into Public Expenditure Reviews (PERs)-though this is often a misnomer because most such reviews cover only government spending, and usually only central government spending. (PERs are discussed later in this study along with other donor instruments used to assess public expenditure and financial accountability.)

Third, the dominant influence of economic and social policies on project effectiveness came to be more widely understood. After the early 1980s it was no longer possible to believe that project-centered aid could be effective for development if macroeconomic policies were badly flawed. Accordingly, untied lending conditional on policy reforms began to complement project assistance. This gradual shift to policy-based adjustment lending- originally permissible only to finance foreign exchange gaps, then from the

mid-1990s justifiable as pure budget support-made increasingly clear the need to periodically assess the recipient governments budget policies and budget management.

Yet for much of the 1980s most donors, still in the grip of Harrod-Domar thinking (which argues that growth is a function only of the amount of physical investment and has no relation to the quality of governance or public sector management), remained focused on aggregate spending, its compatibility with the macroeconomic framework, and its sectoral and economic allocations. Not until the early 1990s were the links between good public sector management and development sufficiently understood.3 It became increasingly accepted that effective development outcomes require not only sufficient resource transfers and sound macroeconomic and social policies, but also efficient spending. Efficient spending, in turn, depends on strong systems for budgeting, financial management, and accountability. The need for donors to get involved in budgeting, accounting, auditing, and control could no longer be ignored.4

Finally, the recognition from late 1996 of the destructive effects of corruption-official and private, and highlighted emphatically by the East Asian crisis of 1997-99-gave new urgency to donor agencies need to assure their constituencies that aid resources were not being diverted for private ends or misallocated to activities not conducive to promoting growth and reducing poverty. To that end the World Bank adopted an anti-corruption policy in 1997, the Asian Development Bank followed suit in 1998 (see http: www.adb.org), OECD countries negotiated an Anti-Bribery Convention in 1999, and other agencies placed corruption at the fore of their concerns. In addition, because lack of transparency had permitted deep-seated financial and governance problems to fester until they erupted in the 1997 East Asian crisis, the IMF developed Standards and Codes of Fiscal Transparency addressing similar issues (see IMF 2001). (IMF Fiscal ROSCs are examined in greater detail later in the study.)

This evolution in development policies and practices has raised the pressure on governments and donors to understand the public expenditure management and financial accountability environment in which all budget funds are used. Thus, donors have created instruments that share many objectives, processes, analytical methods, and development goals, but that are grounded in each agencys specific concerns. As noted, later sections of this study examine these instruments and assess their coverage in detail. But underlying all these instruments are basic notions of fiduciary risk and accountability and concepts of public expenditure management-the subject of the next section.

NOTES

1. Strengthening such instruments is one of the two main activities of the Public Expenditure and Financial Accountability (PEFA) program; the other is supporting country assessments and reforms. PEFA is a joint program of the World Bank, European Commission, International Monetary Fund (IMF), development agencies from France, Norway, Switzerland, and the United Kingdom, and the Strategic Partnership with Africa (SPA). The program was established in December 2001 and functions through a secretariat based in World Bank headquarters under the guidance of a steering committee representing all the partner institutions. PEFA was established out of concern that the instruments used by different development agencies to assess public expenditure, procurement, and financial accountability are insufficiently integrated, impose unnecessarily high transaction and compliance costs on recipient governments, and may be on a divergent course. Moreover, with increasing aid for budget support-where the fungibility of the assistance precludes a direct link between the aid and desired out-comes-these instruments have become more important, especially in terms of strengthening financial accountability.

2. First-generation public investment programs suffered from a number of weaknesses, were often formalistic documents produced as wish lists for consultative groups or other donor meetings, and sometimes even produced adverse budget outcomes over the long term. More recent second-generation public investment programs preserve the advantages of sound medium-term investment programming while avoiding the mistakes of earlier years (Schiavo-Campo and Tommasi 1999).

3. These links were first emphasized in World Bank (1989a) and subsequently formalized in a policy paper (World Bank 1992) and progress report (World Bank 1994). Other international organizations issued similar policy statements (such as Asian Development Bank 1995), as did bilateral donor agencies. The general consensus is that good governance revolves around four pillars: accountability, transparency, rule of law, and popular participation. The World Banks public sector and governance strategy was first discussed by its Board in 1991; a progress report was issued in 1994, and a revised strategy was issued in October 2000.

4. Budget support also avoids the suggestion of donor interference and micromanagement inherent in some forms of earmarked assistance. For example, the World Banks first loan, in May 1947, was for $250 million and