to scrutinize each part of the system. At each stage of the expenditure management cycle, one can go a long way toward building governance considerations into the technical assessment by asking the question of the extent to which the systems and procedures are transparent, and whether they provide for genuine accountability, predictability through application of the rule of law, and appropriate participation. It is especially important in this largely qualitative exercise to recognize the institutional reality that informal rather than formal rules may be the determining influence on peoples behavior and the effectiveness of the system. This recognition, in turn, leads directly to facing the key issue of incentive, and how to gradually provide stronger incentives for improved behavior, and disincentives for corruption and inefficiency. (Beyond this general principle, the extent to which institutional factors are taken into account in the guidelines and assessments of public expenditure management is discussed later).

COVERAGE OF THE INSTRUMENTS

Figure 1 summarizes the coverage of the main instruments (the detailed map of all 94 components of public expenditure management is in annex 2). Even a cursory look at figure 1 shows that there is both substantial specialization in the various instruments, in the sense of a particular focus on certain issues or aspects of public expenditure management as well as significant overlap in their coverage. Recalling that the high, medium, and low (black, gray, and white) coverage of each of the 15 major components in the map has been measured according to the frequency of coverage of the subcomponents, the main observations are as follows:

The scope and range of PERs vary widely but generally focus on traditional areas of public expenditure policy and analysis-both at the macro level and in key sectors such as health and education. Many PERs provide substantial coverage of intergovernmental fiscal relations, government structure, budget coverage, fiscal framework and expenditure programming budget preparation, and treasury systems, cash management, and expenditure monitoring.

CFAAs provide substantial coverage of budget coverage, budget preparation, treasury, cash management, and expenditure monitoring, internal control and internal audit, accounting, reporting, and records management, and external audit.

Consistent with its focused objectives, CPAR coverage of nonprocure-ment aspects of public expenditure management is limited to occasional mention of relevant issues of budget preparation and execution, particularly internal control and audit.

Fiscal ROSCs cover mainly the legal framework for expenditure management, relations between government and nongovernment entities, budget coverage, fiscal framework and expenditure programming, internal control and audit, and external audit (mainly from a fiscal transparency perspective).

EC audits focus on downstream issues such as treasury systems, cash management, and expenditure monitoring, procurement, and accounting and reporting.

None of the instruments provides more than limited coverage of administrative and financial management capacity (personnel quality, capacity, incentives, and management).

In countries where the World Bank is preparing or providing an adjustment loan, the functional equivalent of a PER is integral to the dialogue under the loan-but a formal PER is normally not done. Because expenditure reviews conducted in the context of adjustment lending tend to give greater weight to nuts-and-bolts problems of expenditure management than do formal PERs, their coverage of budget execution and financial control issues is greater, as is the overall scope of the instrument.

With the exception of PERs, the assessment instruments provide the most extensive coverage of budget execution, accounting, audit, and control. But within these areas there are substantial variations in coverage and approaches among PERs, CFAAs, Fiscal ROSCs, and EC audits-while seeming overlaps in coverage often mean little, and sometimes reflect different objectives and approaches. Moreover, figure 1 does not reflect differences in the quality and depth of analysis.

SIMILARITIES

Just as the depth and breadth of coverage vary, so do the procedures for carrying out assessments-including the intensity of discussions with recipient countries. Still, all the instruments share common objectives and features. They all examine aspects of public expenditure and financial accountability systems,

diagnose strengths, weaknesses, and risks, and recommend action plans for improving functions and procedures, with monitorable benchmarks or indicators identified to gauge and guide such efforts.

The recommendations of assessment reports have become more important in recent years, and in most cases feed (directly or indirectly) into the development agencys country assistance strategy or its equivalent. These recommendations often also influence lending and other donor support, affecting decisions on the type and amount of assistance and sometimes resulting in specific projects or technical assistance to support improvements in public financial management.

For example, EC audits of the use of EC program aid provide useful information about the quality of underlying public expenditure management systems and procedures for control, procurement, and the like-though, for reasons noted above, the European Commission is moving to a new approach that sharpens its assessment work by combining compliance testing (a modified form of audit) and performance measurement. Although the IMFs Fiscal ROSCs are in principle voluntary, acceptance of their analysis and recommendations increases the probability that the governments assessed will receive IMF support and influences the content of the stabilization programs the reports are designed to support. Moreover, multilateral development banks and other development agencies pay serious attention to the findings of Fiscal ROSCs when designing their aid programs.

Similarly, a PER can be a self-standing diagnostic study, but-whether as a formal product or not-is considered essential for a World Bank adjustment loan. As noted, in the absence of a recent formal PER, its functional equivalent is carried out during the preparation of an adjustment loan. Rarely does a Bank adjustment loan or credit not include a substantial component for improving public expenditure programming and management. Similarly, CFAAs and CPARs are required in all countries where the Bank is considering adjustment support. It would be disingenuous to pretend that the Bank would not (rightly) insist on implementation of some CFAA or CPAR recommendations as part of the reforms supported by an adjustment loan. Moreover, many bilateral agencies have always paid attention to Bank PERs, and appear just as willing to rely on the findings of CFAAs and CPARs.

The assessment instruments also share several operational approaches and processes. First, they all emphasize the importance of ensuring that recipient governments own and participate in the analysis. PERs, CFAAs, and CPARs encourage government participation throughout the process,