Informing the sponsoring agencies and their development partners about fiduciary risk.

RISK

The notion of risk is inherently different and more ambiguous in public sector activity, including development aid, than in private enterprise. In commercial lending the biggest risk is that loans will not be repaid. But the risk of nonpayment is inapplicable to development grants and very remote for soft loans-especially for international financial institutions that do not allow loan rescheduling. Risk is generally defined as exposure to the chance of injury or loss (Merriam-Webster 1997), and fiduciary as a person or entity in a position of trust, obliged to act in the interest of another (Walker 1980). How such interest is defined thus becomes the central question.

A distinction is often made between fiduciary risk and development risk. Fiduciary risk refers to the possibility that funds provided will be misused or stolen. But in the context of development aid even a narrow conception of fiduciary risk must also include the possibility that actual expenditures will diverge from authorized expenditures (as reflected in the borrowing countrys budget), whether because of misappropriation or misallocation. This is the definition of risk used in the World Banks Country Financial Accountability Assessments (World Bank 2002b). To avoid fiduciary risk, the borrowing countrys budget must be reasonably comprehensive, its fiscal situation should be on a path to sustainability, and the borrower and the provider of funds must agree on how expenditures should be allocated. (All three conditions are also standard requirements of adjustment lending.)

This definition of risk also has a strong governance underpinning. In representative governments no funds can be mobilized from citizens or spent except through official and public sanction by their elected representatives. Thus unauthorized discrepancies between budgeted and actual spending place a cloud over the legitimacy of the entire public expenditure apparatus.

The above construction of fiduciary risk does not include an obligation to ensure value for money or any other efficiency objective. But notions of efficiency and effectiveness are integral to development assistance. Hence development risk adds to the risk of misappropriation and misallocation the risk that resources will be wasted as a result of inefficient institutions and organizational practices. This broader definition of risk is used, for example, by the U.K. DFID (though the agency still refers to it as fiduciary risk).

Assessing performance-let alone quantifying it-is exceedingly complex and fraught with potential misinterpretations and misapplications. But DFIDs implementation criteria for its value for money requirement are generic enough not to raise the risk of micromanagement or misplaced con-creteness. (For more details on the DFID approach, see annex 1.)

Thus the fiduciary objective is met by reasonable assurance that aid money will not be stolen or used for guns instead of butter. The development objective requires, in addition, some indication that money will be spent efficiently and effectively. These objectives share a kinship with the three classic levels of public expenditure management: overall expenditure control, strategic allocation, and operational management (Campos and Pradhan 1996).

Except in severely dysfunctional systems-where expenditure control is paramount-these three levels are interrelated. Improvements at one level facilitate, and are not sustainable without, improvements at the other two. For example, imposing a hard ceiling on sector spending during budget preparation without linking that ceiling to sector policies and programs will likely result in underfunding of economically valuable activities, because such activities tend to offer less potential for rent seeking. Conversely, efforts to strengthen the links between policies and budgets will fail if fiscal discipline is weak. Similarly, operational management cannot be improved without fiscal discipline and sound resource allocation-for which good management is crucial.

Over time, fiduciary and development risks tend to merge, and lasting improvements in expenditure management and accountability also should increase operational effectiveness because they reinforce the links between policies and budgets and strengthen fiscal discipline. Moreover, if reducing poverty through faster growth and pro-poor measures is assumed to be the main objective of recipient governments (which requires some judgment about the quality of governance), the fiduciary and development obligations of development agencies are to both the providers and recipients of aid funds. Correspondingly, ownership becomes a two-sided concept, and partnership becomes the hallmark of effective financial accountability assessments.

Still, a tension often exists between assessments of fiduciary risk and long-term development objectives and requirements. In principle, these are two sides of the same coin. Recipient governments should have a strong interest in precise assessments of the risks associated with their public expenditure management systems, as measured in terms of achieving aggregate fiscal discipline, efficient resource allocation, and efficient and effec-

tive service delivery. Indeed, such assessments can guide the design and implementation of public expenditure management reforms.

But in practice the process through which such information is collected and analyzed is extremely important and can influence government perceptions and incentives. If a development agency appears to be setting the agenda and collecting and using information that a country perceives could be used against its interests-for example, in setting difficult conditions for a structural adjustment loan or technical assistance operation-it is much harder to internalize the diagnostic process in a country, ensure its ownership by the government, and promote transparent provision and exchange of information. This point has important implications for the structure and management of assessment tools, and these issues are discussed later in this study.

ACCOUNTABILITY

Though accountability is at the core of good governance, the concept is inherently relative, requiring a specification of accountability to whom and for what. As noted, development agencies must be accountable to aid providers and share with aid recipients responsibility for achieving common objectives. Thus, there must be a way to systematically assess such achievements.

There is a panoply of possible results, from the immediate output of a specific activity to the broader outcomes of an overall program. In well-defined projects close to their final users (such as urban transport projects) the link between outputs and outcomes is clear and immediate enough to permit the use of output indicators as a proxy for outcomes. In adjustment lending and budget support this is not the case: to be defined meaningfully the results must be defined broadly, and should include both outcomes and process indicators. There is an accountability tradeoff: accountability can be tight or broad, but not both (Schiavo-Campo 1999). Individuals can be held strictly accountable for narrow and specific results, but only loosely for broad results-achievement of which partly depends on factors beyond their control.

Moreover, accountability has two elements: answerability and consequences. A systematic dialogue on the uses and results of expenditure can be valuable even in the absence of direct implications (provided the results are defined appropriately). As a rule, however, such a dialogue should be