6 Monitoring and Improving Governance Subsystems

International Monetary Fund
Published Date:
May 2006
  • ShareShare
Show Summary Details

Governance being multidimensional, both reform and monitoring to support reform are best tackled through specific subsystems. The framework laid out in the previous chapter distinguished among three broad subsystems: cross-cutting public financial management and administrative control agencies that underpin bureaucratic capabilities; front-line service provision and regulatory agencies; and checks and balances institutions. This chapter deepens the focus on developing-country governance by considering for each of these subsystems some options for monitoring and improving performance—and links these to different approaches to scaling up aid. (The developed-country, multilateral, and global dimensions of the governance challenge are considered in depth in chapter 3, 4, and chapter 7, respectively.) Inevitably, countries reforming governance differ from one another in the pace at which these different subsystems improve. A final section of the chapter examines some of the dilemmas this poses for the country and for its development partners.

Monitoring and Improving Bureaucratic Capability

An effective bureaucracy facilitates the scaling-up of aid. The bureaucracy formulates detailed policies that translate the goals of society and its political leaders into programs of action. It manages the implementation of these policies. And it reports on progress.

Helping to build bureaucratic capability has long been a focus of development assistance. Before 2000 it was viewed as principally technocratic, with a gradual accumulation of lessons and advice on good practice. Even as these lessons crystallized, the profile of the work remained low, because its focus on building country systems was at odds with the dominant approaches to providing aid and technical support through self-standing projects, hermetically sealed off from often dysfunctional public sectors. But with the new approach to aid placing increasing emphasis on mutual accountability, the profile of efforts to build bureaucratic capability has risen dramatically.

Better public finance and administration in developing countries is essential for the new approach, introduced in December 1999 with the Poverty Reduction Strategy (PRS) process. The PRS builds on a hard-learned lesson of development experience—that externally imposed conditionality generally fails to achieve its intended results (World Bank and IMF 2005: 1, 10). The national budget and the public bureaucracy that prepares and implements the budget and is accountable to its citizens are critical vehicles for ensuring country ownership and leadership (World Bank and IMF 2005: 12, 15, 19).

Budget support is the natural way to transfer resources to support a country’s PRS objectives without undermining country ownership through excessive external oversight. But where governments focus less on poverty reduction and participation, are less constrained to be accountable to their citizens, and have less capacity, the combination of a PRS process and budget support does not offer a ready way of resolving the tension between country ownership and donor fiduciary obligations.

Countries are making progress in developing a long-term holistic vision for poverty reduction and translating that vision into a coherent, medium-term, sequenced strategy. But most have a long way to go. The PRS Review of 2005, reporting survey data covering 59 countries, concluded that only 7 had well-developed strategic programs (World Bank and IMF 2005; and World Bank 2005). The majority of the remaining countries had activity under way—though not yet advanced to the point that long-term visions could serve as a reference point for policy makers.

This section considers some aspects of country progress in moving from general assertions of development goals to the specific articulation, costing, and implementation of strategies for poverty reduction. It first reports on progress in monitoring the quality of public expenditure management and highlights some patterns across countries revealed by monitoring. It also highlights emerging lessons about how to strengthen public expenditure management systems in different country settings. The section then reports on efforts to monitor public administration, drawing on experience to offer practical guidance on how to improve administrative capability, both for developing countries on a path of improving governance and for their development partners seeking to monitor progress.

Monitoring and Improving Public Financial Management

Public financial management is particularly relevant to the new aid architecture. It is key for getting results on the ground and for assuring donors that aid resources are being used prudently. Setting the stage for the analysis here is a framework based on the 2005 report of the multiagency Public Expenditure and Financial Accountability (PEFA) partnership program (figure 6.1). That report synthesized the results of more than a half-dozen years of work by PEFA partners to develop a common platform for assessing the quality of public finance systems, including those in aid-recipient countries. The framework depicts four facets of the budget cycle:

  • Policy-based budgeting—the formulating process for translating public policies, including policies that emerge from a PRS process, into specific budgeted expenditures
  • Arrangements for predictability, control, and stewardship in the use of public funds (for example, payroll and procurement systems)
  • Systems of accounting and recordkeeping to provide information for proper management and accountability
  • External audit and other mechanisms that ensure external scrutiny of the operations of the executive (for example, by parliament)

FIGURE 6.1Public financial management: a performance monitoring framework


Comprehensiveness of budget coverage and transparency of fiscal and budget information cut across these four facets. The framework also identifies credibility—that the budget is realistic and implemented as intended—as a key intermediate outcome, a result of the operation of the whole cycle.

There are many ways of measuring the quality of a country’s public financial management system. As box 6.1 highlights, the International Monetary Fund (IMF) has developed some useful tools.1 This section focuses on two measures at two different levels—an overall measure of the quality of public expenditure management, and measures of specific expenditure management subsystems.

An overall assessment. The results of Country Policy and Institutional Assessment (CPIA) criterion 13 (see box 5.2)—abbreviated here as CPIA-budget—can be used to assess overall patterns in the quality of budget management systems across countries. As a country improves its budget management system, its CPIA-budget score moves from weakest (1) to strongest (6). The scale is built from four dimensions of budget management, which broadly correspond to the facets of the PEFA performance management framework in figure 6.1.2 A CPIA-budget score at or above 4 is consistent with the “good enough governance” pattern described in chapter 5.

As of 2004 only 10 of 66 low-income aid-recipient countries had the “good enough” (though imperfect) budget system implied by a CPIA-budget score of 4 (figure 6.2). These higher-performing countries are Azerbaijan, Benin, Burkina Faso, India, Indonesia, Mali, Serbia and Montenegro, Sri Lanka, Tanzania, and Uganda. Almost half the countries scored at or below 3. Of the 10 better-performing countries, Azerbaijan, Mali, and Tanzania raised their CPIA-budget scores by at least one point between 2001 and 2004.

FIGURE 6.2Low-income aid recipient countries with CPIA 13 (quality of budgetary and financial management) scores, 2004

Source: World Bank CPIA Database.

Note: CPIA 13 measures the quality of budgetary and financial management. The 66 IDA countries are divided into groups based on a CPIA 13 score.

BOX 6.1Two IMF tools to support fiscal management and transparency

The Code of Good Practices on Fiscal Transparency was developed in response to concerns that a lack of comprehensive information on fiscal activity made it difficult to properly assess the objectives of fiscal policy. Greater fiscal transparency was also believed to be linked to improved governance and fiscal outcomes more generally. The code contains 37 good practices that are organized according to four main principles of fiscal transparency: clarity of roles and responsibilities; public availability of information; open budget preparation, execution, and reporting; and assurances of integrity. These practices, when observed, are critical not only for holding leaders accountable, but also for preventing any mishandling of finances during budget execution.

The IMF regularly undertakes assessments of fiscal transparency called fiscal Reports on Observance of Standards and Codes (ROSCs) in its member countries. Participation in an ROSC is voluntary and the authorities retain the right not to publish the final report, although most have agreed to publish fiscal ROSCs.a As of the end of 2005, fiscal ROSCs have been completed for 80 countries, and 76 of these have been published. ROSC participation is distributed unevenly across regions, with most countries in Europe and the continental Western Hemisphere having completed ROSCs, while a much smaller share of countries in Africa, the Middle East, and Asia have agreed to participate. A number of countries, especially in Europe, have been working on improving fiscal transparency and have opted to undertake one or more ROSC updates to reflect this progress. In addition, a growing number of countries are undertaking full reassessments. Both reassessments and updates are published on the IMF Web site with the original ROSC.

Source: International Monetary Fund, Fiscal Affairs Department, Fiscal Transparency Unit.a. All the published reports are available on the IMF ROSC Web site at

Disaggregated public financial management indicators. The heavily indebted poor countries (HIPC) debt reduction initiative spurred a sustained effort to develop actionable indicators of budget quality. A first set of 16 indicators was developed jointly by the World Bank and the IMF and applied in 2001 in 23 HIPC through a joint assessment with recipient-country governments, with a follow-up assessment in 2004.

Building directly on the HIPC tracking process, seven donors (the World Bank, the IMF, the European Commission, the U.K. Department for International Development (DFID), France, Norway, and Switzerland) plus the Strategic Partnership with Africa embarked on a joint PEFA program to support “integrated and harmonized approaches to assessment and reform in the field of public expenditure, procurement, and financial accountability.”3 In 2005 PEFA issued its public financial management performance measurement framework, including 28 high-level monitoring indicators. PEFA participants have committed to harmonize their assessments of the quality of the public management systems of aid-recipient countries around the PEFA framework.

The HIPC tracking indicators score each question on an A-C scale, with detailed descriptions of how to score each question and an explicit benchmark of “good enough” performance for each question.4Table 6.1 aggregates the HIPC tracking results for 2004 for the 16 indicators into five categories aligned with the public financial management (PFM) framework laid out in figure 6.1.

TABLE 6.1Quality of budget management systems in 25 heavily indebted poor countries, 2004
CPIA-budget ratingCountryPolicy-based budgeting (1 measurea: A-C ranking)Budget comprehensiveness (4 measuresb: # met)Budget credibility (2 measuresc: # met)Budget Execution (5 measuresd: # met)External budget scrutiny (2 measurese: # met)
Best-performing group (both CPIA and HIPC)MaliA2241
Burkina FasoA2231
Middling group 1GuyanaB3222
Middling group 2CameroonC1131
Sierra LeoneC2121
Weaker-performing groupSão Tomé and PríncipeC2011
Congo, Dem. Rep. ofC2000
Gambia, TheB2000
Source: IDA and IMF (2005). For details of the individual HIPC indicators, see

The measure is HIPC indicator 7.

The measures are HIPC indicators 1, 2, 4, and 5.

The measures are HIPC indicators 3 and 8.

The measures are HIPC indicators 9–13.

The measures are HIPC indicators 14–15.

Source: IDA and IMF (2005). For details of the individual HIPC indicators, see

The measure is HIPC indicator 7.

The measures are HIPC indicators 1, 2, 4, and 5.

The measures are HIPC indicators 3 and 8.

The measures are HIPC indicators 9–13.

The measures are HIPC indicators 14–15.

Control of procurement and payroll was not part of the 2001 HIPC tracking indicators. In practice, procurement and payroll—plus cash transfers—make up the overwhelming majority of public spending, so strong controls in these areas are vital for good financial management. Recent advances in monitoring the quality of procurement highlight some emerging lessons (box 6.2).

Consider first the cross-country patterns for policy-based budgeting. Done well, policy-based budgeting can sharpen the focus on longer-term priorities, enable phasing in shifts in priority expenditures over time, and potentially reconcile capital costs and their recurrent cost implications (if capital and recurrent budgets are integrated). The HIPC tracking indicator reported in the third column of table 6.1, labeled “policy-based budgeting,” focuses on medium-term projections. A score of A signals that medium-term projections exist and are integrated into the budget formulation cycle; a score of B that they exist but are not integrated; a score of C that they exist for only a few sectors or not at all. In 2004 only 7 of the 25 countries tracked had integrated medium-term projections into their budget cycles, but 13 of the remaining countries made projections (but did not integrate them into the cycle).

Now consider the cross-country patterns for budget implementation in the fourth, fifth, and sixth columns in table 6.1:

  • The fourth column reports on measures of whether the budget is comprehensive, with no significant extrabudgetary funds (including unfunded contingent liabilities), and with donor funds also reported on budget. Only Bolivia, Chad, Ethiopia, and Guyana can be said to have comprehensive budgets in the sense that they met at least three of the four benchmarks. Seven countries met no more than one benchmark.
  • The fifth column reports on budget credibility, as measured by the closeness of actual expenditure out-turns (both aggregate and sectorally) compared with the original approved budget, and limits on the extent of arrears. Six countries have fully credible budgets (meet both indicators), but 12 countries met neither of the credibility benchmarks.
  • The sixth column reports on whether countries have a well functioning expenditure execution system, including an internal audit mechanism, and other in-budget-year controls. Only 8 of the 25 countries met three or more of the five budget execution indicators tracked in the HIPC process—and 9 countries met only one or none.

BOX 6.2Recent advances in monitoring the quality of procurement

More governments and advisers have recognized that the mechanisms for the government to purchase goods, works, and services and the effectiveness of these practices influence the financial well-being of nations, the ability of citizens to access public services, and the competitiveness of domestic firms. Spending on procurement is at the core of discretionary government spending; even minor improvements in efficiency can yield substantial cost savings.

With greater visibility has come a profusion of efforts to monitor the quality of public procurement systems and enhance their functioning. Consider the Philippines, where in 2005 the government initiated work to measure public procurement performance in 10 of its largest agencies. The work complemented other procurement monitoring efforts in the country, which included the application of the Organisation for Economic Co-operation and Development-Development Assistance Committee (OECD-DAC) Baseline Indicator Set for Procurement (BIS) tool; the observance of procurement proceedings by civil society representatives; and the publishing of information on the award of procurement contracts and other relevant statistics on the government’s e-bulletin Web site.

A robust approach to monitoring procurement and linking monitoring to improved performance is beginning to emerge. The simplest form is physical observation of procurement practices and outcomes.

A second form of procurement monitoring focuses on transparency: the publication of procurement opportunities and outcomes. In many countries electronic procurement systems have tremendously increased the visibility of public contracting and allowed government and nongovernment bodies alike the opportunity to review the distribution of contract awards as well as the price the government pays for its goods, works, and services.

A third form of procurement monitoring is assessing performance of public procurement systems using defined performance indicators. Work on developing tools suitable for monitoring public procurement has been undertaken jointly by donors and partner countries over a two-year time period. A procurement-specific indicator has been included in the PEFA Performance Indicators, and an entire tool, the BIS, has been developed as part of the OECD-DAC Working Party on Improving Aid Effectiveness.

The BIS has been applied in more than 10 countries in the first six months after it was finalized. The recommended action here is for the BIS to be used as a regular monitoring tool.

Source: World Bank.Note: The BIS tool is available at

Overall, seven countries (Benin, Burkina Faso, Ethiopia, Honduras, Mali, Senegal, and Tanzania) can be said to implement their budgets reasonably effectively, in the sense that they met the benchmarks for half or more of the criteria in each of columns 4-6. Another seven countries (Bolivia, Democratic Republic of Congo, The Gambia, Guinea-Bissau, Mozambique, Uganda, and Zambia) met fewer than half of the benchmarks in at least two of the three categories, and so appear to have significant weaknesses in budget implementation.

The seventh column of table 6.1 reports on the quality of budget reporting and external scrutiny. Adequate accounts are a precondition for effective scrutiny. As of 2004, 14 countries met one of the two benchmarks—closing annual accounts within two months of the end of the fiscal year. Formal oversight of the budget is the responsibility of parliament, based on independent audits of the accounts, and is measured by the second benchmark. But not one of the HIPC-monitored countries submitted audited reports to its legislature within 6 months of the end of the fiscal year—and only 7 countries submitted an audit within the benchmark of 12 months.

The results confirm that the quality of budget management systems of the 25 HIPC-tracked countries remains uneven. Only Burkina Faso, Mali, and Tanzania score in the top half of possible (absolute) scores in all five categories. Ethiopia, Ghana, Guyana, Honduras, and Rwanda avoid the bottom rung in all four categories. The remaining 17 countries had budget systems with at least one deep flaw.

This unevenness raises concern. The budget process is like a chain in the sense that it is only as strong as its weakest link. Even well-formulated budgets add modest value if there is little relation between the budget on paper and the way money is actually spent. And the impact of a well-prioritized and well-executed formal budget is undercut if much of the public spending is off budget.

Strengthening public financial management. Why is performance on public financial management so uneven? In some countries poor performance may be a consequence of clientelism, extended civil conflict, and the evasion of formal rules and external scrutiny. Serious improvement is unlikely without changes in a country’s political dynamics. This is more likely for one-third or so of the low-income aid-recipient countries that have consistently been stuck in the fourth and fifth quintiles of all governance performance measures, with no improvement over the past five years, most of them with capabilities undermined by conflict.

But many countries have shown the capacity for quite rapid improvement in their public financial management systems:

  • The CPIA and HIPC tracking assessments reveal that many countries strengthened their budget systems in just three years—some by significantly more than any plausible margin of error. Of the 66 International Development Association (IDA) recipient countries included here, 19 improved their CPIA-budget score between 2001 and 2004–7 of them by one or more points. A comparison between the 2001 and 2004 HIPC tracking assessments of 25 countries identifies 6 (Cameroon, Ghana, Mali, Niger, Senegal, and Tanzania) that improved their scores in a net of at least three categories.
  • Even in a brief three-year time span, there were some countries that made substantial improvements for each of the five budget subsystems. For budget execution Senegal went from meeting none of the benchmarks in 2001 to meeting three in 2004, Ghana from one to three. Cameroon improved its score on both “external scrutiny” benchmarks (though in 2004 it still took more than two months to close its annual accounts). Guinea’s score on policy-based budgeting went from C to A. And Bolivia and Guyana increased by two the number of “budget comprehensiveness” benchmarks met.
  • As the sustained improvements in Ghana, Mali, Senegal, and Tanzania suggest, countries with stronger starting capacity (measured, say, by having more benchmarks met in 2001) may be better able to achieve rapid gains in the short run than countries with weaker starting points (see Dorotinsky, Kisunko, and Pradhan 2005). But Niger—which improved its ranking in a net of five categories5—suggests that significant gains also are possible where the starting point is weak.

These patterns suggest that heightened attention to budget management and strong incentives for better performance can result in quite rapid gains. For countries determined to improve their public financial management systems, achieving a “good enough” standard within, say, a 5- to 10-year period may be feasible. How budget reform is designed and implemented will be key.

The HIPC tracking results suggest that a few countries—those with committed developmental leadership, plus a preexisting baseline of capacity—appear able to adopt and rapidly implement a comprehensive program of budget reform, to the point that country systems can provide a robust platform for ensuring effectiveness in the use of resources. But where capacity is weaker, there is a need to set realistic goals for what is achievable and implement them in a realistic sequence.

A first lesson for strengthening public financial management systems is that in most countries, the approach should be incremental. The reforms proposed for specific budget management subsystems have sometimes been very ambitious. Recent reviews by both the IMF and the World Bank have examined the experience with medium-term expenditure frameworks (MTEFs). The IMF review captures the shared conclusion, namely that “developing comprehensive medium-term expenditure frameworks can be effective when circumstances and capacities permit. Otherwise, it can be a great consumer of time and resources and might distract attention from the immediate needs for improving the annual budget and budget execution processes.”

The IMF review also offers some useful guidance in noting that

… the MTEF, as a feasible means of improving budgeting, requires the following: reliable macroeconomic projections, linked to fiscal targets in a stable economic environment; a satisfactory budget classification and accurate and timely accounting; technical capacity … and disciplined policy decision-making, [including] budgetary discipline … and political discipline for fiscal management. Before introducing an MTEF, one should raise a question: is the country ready for such an exercise in the sense of having adequate support for the above preconditions? When this support was not adequate in a number of African countries, the MTEF was introduced prematurely, and is turning out to be merely a paper exercise. (IMF 2005)

Efforts to install computerized financial management information systems (FMIS) also are often overambitious and invariably encounter significant delays. Reviews by both the World Bank and the IMF of efforts to install FMIS in African countries concluded that large and therefore more spectacular projects are often preferred because they can be easily communicated as evidence of political action, but they are more volatile and subject to greater likelihood of failure than are smaller, more focused interventions (IMF 2005; Heidenhof and others 2002).

A useful guide to sequencing public financial management reform in low-capacity settings is suggested by DFID. The new “platform approach” for Cambodia involves a cumulative sequence of budget reforms that focuses each round on achieving specific budget functionalities, building on these functionalities in the subsequent round (DFID 2005). The sequence that emerges is almost the reverse of that often associated with PRS implementation (figure 6.3). Efforts to implement PRSs have focused on their costing and translation into medium-term budget frameworks and on strengthening countries’ statistical capacities to monitor results.6 By contrast, the Cambodian platform sequence focuses first on the basics: budget credibility, then predictability and control in budget execution. (Achieving these basics depends also on achieving predictability in the year-to-year flows of aid, a serious problem as noted in part 1.) Only after these first two platforms are locked in will they move on to medium-term budget planning—and only once that is in place will they foster public management reforms to support a results culture throughout the public bureaucracy. Country leadership has been an important feature of Cambodia’s public financial management program, ensuring that the design of reforms reflects domestic priorities, rather than those of donors.

FIGURE 6.3Cambodia: A platform approach to budget management reforms

Source: DFID 2005.

A second emerging lesson for budget management is to complement the technocratic reforms with greater transparency. Although the PRS approach highlights inclusiveness, its implicit route to effectiveness tends to be technocratic: design a robust poverty-reducing budget, execute it effectively, monitor results, and recalibrate policy and budgeting on the basis of what is learned. The lesson emerging from experience is that, in developing countries with weaker capacity, this process may be better viewed as a long-run trajectory than as a feasible path to better results in the short to medium term. That explains interest in a more demand-side approach, complementing the technocratic route: along with participatory priority-setting in PRSs, foster transparency in budget management—and emphasize the potential of public information to improve the developmental discourse among citizens, their governments, and development partners.

Monitoring and Improving Administrative Quality

Getting development results depends on much more than good financial management. For any organization, public or private, delivery depends on the quality not only of the financial side of its balance sheet, but also of its real side—the quality of its people, and how effectively they are deployed and led.

As the framework in chapter 5 highlighted, public administration comprises both downstream service provision and regulatory agencies (schools and ministries of education, customs agencies, roads authorities, and the like) and upstream cross-cutting control agencies within the bureaucracy (pay, human resource, and performance management control agencies, for example). Public administration reforms generally combine a focus on improving upstream systems—to have a broad impact across multiple systems—with targeted efforts to improve the performance of specific, high priority agencies. This section focuses on upstream reforms; later sections consider sector-specific approaches.

In the 1980s and early 1990s a first generation of administrative reform focused principally on scaling back the bloated apparatus of government. In the late 1990s attention shifted toward improving administrative capability. Some consensus has been generated on the characteristics of an effective public administration. As the CPIA subcategories used to score the “quality of public administration” suggest, the standard prescription typically includes the following:

  • Well-functioning mechanisms for policy coordination, which ensure policy consistency across departmental boundaries and facilitate clear decisions on policy and spending priorities. To be effective, these coordinating mechanisms need to be at the apex of government, supported by top political leadership.
  • Well-designed administrative structures for individual line ministries and semi-autonomous executive agencies, with little duplication of responsibility, and with clear lines of authority—plus streamlined business processes and a focus on results.
  • Human resource management underpinned by the principle of meritocracy—including for recruitment, promotion, and major disciplinary actions. This includes insulation from undue political or personal interests, as well as practices that reward good performance (for example, through career advancement and financial rewards) and penalize poor performance.
  • Pay and benefits adequate to attract and retain competent staff, including at senior and technical levels.
  • Establishment and wage bill control sufficiently robust to ensure that the public sector wage bill is sustainable under overall fiscal constraints.

Monitoring administrative capability. As with CPIA-budget, the 1-6 scale of CPIAadmin describes the gradations for a country to move through as it works to improve the quality of its public administration.

The track record of efforts to close the gaps between the desired and actual quality of public administration is (to put it gently) uneven in both developed and developing countries. A landmark review of public administrative reform in 10 OECD countries—including such noted public management reformers as Australia, New Zealand, Sweden, the United States, and the United Kingdom—concluded:

Reform-watching in public management can be a sobering pastime. The gaps between rhetoric and actions … are frequently so wide as to provoke skepticism. The pace of underlying, embedded achievement tends to be much slower than the helter-skelter cascade of new announcements and initiatives. Incremental analysis and partisan mutual adjustment seem to have been very frequent features of public management reform, even if more-than-incremental changes were frequently hoped for. (Pollitt and Bouckaert 2000: 184, 188-89)

The CPIA-admin scores provide a snapshot of administrative system performance and reform for the 66 IDA countries. The ability to measure is less well developed for administrative quality than for budget management. No disaggregated actionable measures paralleling the HIPC tracking and PEFA indicators are available—although an initiative to fill the gap is at an early stage of piloting (box 6.3). The 2004 CPIA-admin results and a composite Kaufmann-Kraay (KK)-style measure of administrative quality produce a correlation coefficient of only 0.56—a reminder of the large margins of error all in governance measures.7

Public administrative systems are weaker than their budget management counterparts (figure 6.4). Of the 66 IDA countries, only 2 score 4 or higher on CPIA-admin (versus 10 on CPIA-budget), and only 17 score 3.5 (versus 24 on CPIA-budget). Trends in CPIA-admin suggest that, though change generally comes slowly, committed countries can achieve quite rapid improvement in their systems of public administration: Between 2001 and 2004 Armenia, Azerbaijan, Cameroon, Georgia, and Vietnam lifted their CPIA-admin scores by one or more points, more than any plausible margin of error.

FIGURE 6.4Low-income aid recipient countries with CPIA 15 (quality of public administration) scores, 2004

Source: World Bank CPIA Database.

Note: CPIA 15 measures the quality of public administration.

BOX 6.3Actionable indicators on public administrative quality

World Bank-supported operational work in Albania, FYR Macedonia, and Romania has yielded some actionable indicators to monitor the extent to which the immediate objectives of civil service management are being furthered. The table below identifies specific subobjectives for civil service management, and indicators to monitor each subobjective.

Civil service management actionable indicators
Merit-based civil service (CS) management
Competition in recruitment and selectionPercentage of CS vacancies filled through advertised, competitive procedures
Turnover unrelated to changes in political leadershipQuarterly CS turnover rates plotted against changes in political leadership
Effective performance evaluation practicesPercentage of CS staff for whom annual performance evaluations were completed

Percentage of CS performance evaluations falling in each rating category
Attracting and retaining qualified staff
Competitive remunerationAverage CS total remuneration as a percentage of average economic sector wages
Ratios of average CS to private sector total remuneration by title
Vertical decompressionRatio of average Secretary General total remuneration to average Junior Officer total remuneration
Attracting qualified staffAverage number of qualified (long-listed) candidates per advertised CS opening
Continuously weeding out poor performing staffPercentage of civil servants receiving the lowest performance rating in two successive years who have left the CS within the following year
Fiscally sustainable wage bill
Budget-financed wage bill is fiscally sustainableActual budget-financed overall wage bill as a percentage of GDP

Albania was first to begin using these indicators (in early 2000). Three examples illustrate their impact on reform implementation. First, reformers documented a significant increase in requests from ministers for exemptions from the competitive recruitment procedures mandated by the CS Law, and used the data to successfully make a case for imposing regulations that would make it more difficult to justify such exemptions. Second, a survey of public and private sector salaries was used to develop a new CS salary structure, which would ensure consistency in the competitiveness of CS salaries across types of CS positions. Third, evidence on a rising incidence of qualified applicants per advertised CS position in Albania has helped to convince doubters about the efficacy of Albania’s competitive recruitment and selection procedures.

Source: World Bank.

A comparison of the results among low-income aid-recipient countries for CPIA-admin and CPIA-budget—and the relation between each and the corruption and policy-quality outcomes—again suggests some unevenness across governance subsystems. While the overall correlation between CPIA-budget and CPIA-admin is quite high (0.73), the quality of budget management and of public administration can vary greatly from one country to another.

The correlation between budget systems and control over corruption is low at 0.46. This result is not as surprising as it may appear at first—corruption is an outcome of the quality of national governance systems as a whole, not simply budget management (chapter 5) and can be unrelated to public expenditure management. Even so, the result highlights an acute dilemma for approaches to aid that give special prominence to improving budget systems to monitor the use of donor resources. The strong focus on strengthening budget management may help in underpinning good resource allocation and related policies—but not prove a panacea in the fight against corruption. Greater clarity is needed in the global dialogue on governance, corruption, and development impact as to what is achievable—and how it can realistically be achieved.

Strengthening administrative capability. Building effective public administrative systems in developing countries is difficult. A 1999 review of 102 World Bank operations to support civil service reform (CSR) between 1987 and 1997 found that only 33 percent of closed CSR interventions and 38 percent of ongoing efforts achieved satisfactory out-comes.8 Useful lessons are emerging as to both the reasons for the disappointing track record of efforts to improve administrative systems, and constructive options for proceeding.9

Much of the administrative reform agenda aims to improve processes, and process reforms tend to be soft, with progress difficult to observe or measure. Even when these reforms work, their impact is evident only over the long term. From the start, though, they threaten the authority of established interests throughout the bureaucracy. Resistance to reform within the bureaucracy—either overt, or through halfhearted implementation—is therefore likely to be endemic.

Then there is the political logic of reform. Political leaders need to balance a technocratic view of good reform practice with the political imperatives of building and sustaining alliances with powerful patrons, avoiding conflict with powerful social groups, and maintaining electoral support. Such a calculus is not favorable for serious administrative reform: the upfront political costs are substantial, and the time horizon long before benefits are evident in the form of improved public performance. It is, however, much more favorable for more cynical politicians with a short time horizon to promise bold reforms to clean up government and get government working, in full knowledge that the seriousness—or otherwise—of the reform effort will be invisible to the average citizen.

Administrative reforms therefore need to seek a good fit—one that aligns the agenda with a country’s political realities on the ground. Rapid, comprehensive administrative reform is appropriate only in those rare cases where there is a strong enough baseline of capacity for sustained administrative reform—plus political leadership with the commitment, mandate, and time horizon needed to see the effort through.

Latvia and Tanzania are two countries where the environment for administrative reform was propitious. Between 2000 and 2003, Latvia promulgated an ambitious, and generally well-regarded agenda of administrative reform including a new civil service law which guaranteed meritocratic recruitment, and introduced performance appraisal; a new control framework for the large number of semiautonomous state agencies; and a new framework for coordinating policy making and administrative reform from the Prime Minister’s Office. It also made ongoing efforts to reform the public sector salary structure. In Tanzania, the momentum of administrative reform built incrementally, sustaining a consensus as the program unfolded. An eight-year (1991–9) civil service program first brought employment and the wage bill under control, and then clarified the appropriate roles—and rightsized—across a wide range of government ministries, departments, and agencies. In 2000 a new phase—an ambitious 11-year program—began. The program incorporates both a phased approach to pay reform, and a performance improvement model that gives individual agencies incentives to clarify their role and mission, develop strategic plans (including well-defined results and a well-publicized service delivery charter), and identify and address capacity development needs.

Even in these favorable environments, implementation has been quite challenging. In ical interests have made both pay reform and agency restructuring an uphill struggle. Tanzania, too, has had to scale back the ambition of some of its more far-reaching initiatives. In countries with less favorable environments, the agenda of administrative reform needs to be more modest. However, as box 6.4 illustrates for Albania, even in these more difficult settings, carefully designed incremental reforms can achieve quite significant results.

Improving the Governance of Service Provision—Some Targeted Approaches

In most countries top-down reforms of cross-cutting public financial management and administrative systems will take a long time before they help improve service provision. So it is natural to complement them with approaches that work closer to the service provision front-line. This section will highlight five service-centric approaches to improving governance and service provision:

  • Using public expenditure reviews to highlight sectoral spending priorities as part of an integrated dialogue on strengthening country systems
  • Engaging via sectorwide programs
  • Using information to improve accountability at the service-provision front-line
  • Decentralizing to shift responsibility for service provision closer to the front-line
  • Adopting community-based approaches to local infrastructure investments

The discussion focuses first on approaches that are relatively more helpful in institutionally stronger settings, working its way down to the difficult challenges posed by countries where governance is weak. Some of the approaches presented are relevant regardless of whether a country’s governance is strong or weak.

Identifying sectoral spending priorities. Cross-cutting public management systems aim to ensure that scarce public resources are targeted toward activities with high social returns, and are deployed efficiently. But past allocation decisions may not work out as intended, and new opportunities continually arise. chapter 1, for example, highlights the potential for new, highly productive public investments in key infrastructure areas. The development returns can thus be high from reviewing public expenditures to identify specific expenditure with high potential returns, and ongoing, low-return expenditures that could usefully be redirected toward high-return uses.10 Where this process works well, the fiscal space opened up for new investment or productive current expenditure can be large:

  • Chile invested on average 5 percent of GDP in infrastructure during the second half of the 1990s without resorting to significant borrowing, primarily through reallocation of expenditure, increased efficiency, and the use of public-private partnerships. One consequence was that the country’s credit rating increased, enlarging its unused borrowing capacity, and giving it greater fiscal flexibility for potential future use.
  • Thailand initiated in 2005 a large five-year public investment program of 2.5-5 percent of GDP annually to upgrade and improve infrastructure, addressing widely recognized bottlenecks, including mass transit in Bangkok and the country’s inter-provincial highways. Credit rating agencies have assessed the investment program to be an important driver of growth over the medium term—assessments that were based on the country’s earlier fiscal credibility and enabled it to finance these investments via borrowing. However, both the IMF and rating agencies have noted that effective management of the investment program will be needed to ensure continued access to markets.
  • The United Kingdom routinely incorporates spending reviews into its budget preparation process. Despite this, a 2003 independent review of public sector efficiency identified over US$15 billion of ongoing spending which was not being efficiently used, and was directly “cash releasing” and so available for reallocation.

BOX 6.4Albania—administrative reform in an unpropitious environment

In the wake of communism, skepticism was pervasive among Albanians about the value of state authority and collective action. Politics was fiercely competitive, factionalized, and patronage based. There was no appetite or capacity for far-reaching administrative reform. Even so Albania’s administrative system made important gains between 1998 and 2005.

The gains came through the skillful exploitation, by both domestic reformers and their international champions, of a window of opportunity that opened between 1998 and 2002: the appointment (by the political leaders of an electorally victorious political party) of a reformist prime minister willing to champion an administrative reform agenda. Backstopped by strong conditionality from the World Bank, the agenda was carefully calibrated to be feasible in a setting with limited commitment to reform. Albania’s administrative reform focused narrowly on introducing meritocracy, plus market-competitive pay, for the country’s top 1,300 civil servants. Targeting only this top tier is not enough for systemwide improvements, but it can yield important gains in the quality of policy making and in the management of public resources. It can also establish a precedent of new ways of doing business, with the scope of application broadening over time.

In 2002 the reformist prime minister was replaced, and momentum shifted away from reform and toward Albanian politics as usual. Yet the reforms, which had been widely publicized and enjoyed both the support of donors and broad approval among Albania’s citizens, had crowded in a powerful constituency for their continuation—the senior civil servants. The arrangements for a meritocracy have largely been sustained. Indeed in 2005 parliament intervened directly to reject legislation that would have reduced the ability of the Department of Public Administration to enforce the pro-meritocracy 1999 civil service law.

Source: World Bank.

For all of the potential benefits of arrangements to review and adjust earlier decisions on resource allocation, putting them in place is difficult—for at least three reasons. First, as the previous section of this chapter has detailed, in many low-income countries even the basics of cross-cutting budget and administrative systems do not work well. Second, even where the systems work well, they might not drill down in sufficient detail to distinguish effectively between low- and high-return activities: the knowledge needed to assess development returns can be highly specialized, and reside within sectors, not in budget central agencies. Third, many high-return investments cut across sectoral boundaries—as illustrated in chapter 2 by the high benefits for childhood health of upgrading wood-burning stoves or dirt floors.

While the returns are thus high from strengthening budget systems so they can prioritize more effectively, especially in low-income and weaker governance settings the need to use resources well is too urgent to be dependent solely on systemic improvements. The case is compelling for complementing efforts at system improvements with more targeted efforts—within individual sectors and across sectors—to identify high-return investment opportunities, as well as opportunities for freeing up resources locked into low-return activities. This is an activity for which development partners can provide targeted assistance. The Public Expenditure Reviews facilitated by the World Bank, already an established part of the landscape of development dialogue, offer a ready-made vehicle.

Sectorwide programs. Partial approaches that focus on improving governance and service provision in one sector have the potential to achieve many goals simultaneously. These include getting quick wins in a high-priority area; providing a mechanism for concentrating limited country capacity; creating a focal point for harmonizing multiple, overlapping donor programs around a coherent agenda; providing a clear focal point for results-based monitoring and evaluation; and serving as catalysts for broader change in country systems. Financing mechanisms can run from sector-specific budget support (donors pool all their funding and channel it through the budget using country procedures but carefully monitoring flows to the preferred sector) to approaches that partly “enclave” the use of funds. To realize their potential, though, two issues need to be confronted.

First, even at the level of an individual sector, the challenges of aid harmonization and alignment remain formidable. Donors must be willing to subsume their particular priorities under the umbrella of a country-led sectorwide program and to harmonize their procedures. This is proving difficult, even in Tanzania, a global leader in incorporating aid into country-led strategies and systems. In that country’s sectorwide program (SWAP) in education, for example, donors provide support through basket funding, but have not all harmonized their reporting requirements. As of October 2004, there were an estimated 110 projects still on the books, with an average size of only $906,000 (Economic and Social Research Foundation 2005).

Second, focusing public management reforms on a single sector risks making systemic reform more difficult later on. Road funds, for example, aim to strengthen accountability by harnessing the incentives of users, who have a stake in the efficient and honest use of resources, including stakeholders from the road transport industry, chambers of commerce, and farmers organizations. These have an extra incentive to provide oversight because some of the revenues used for road investment and maintenance comes from earmarked vehicle licenses and fuel levies.11 As another example, sectorwide programs in education sometimes (notably in Francophone Africa) have included a move toward community schools, with increased parental oversight. This has been shown to improve both school management and educational outcomes.12 In addition, locating responsibility for the contracting of teachers with communities can also help reduce unit costs: “even when offering average salaries as low as half the civil service teacher wage, countries have found more qualified applicants than they can hire” (World Bank 2005: 82–85).

Despite their advantages, such sectorally focused initiatives can be at variance with standard approaches to strengthening cross-cutting bureaucratic controls. Ring-fencing through road funds can undercut the ability to make choices among competing uses of resources, fragment the systems of budgetary control, and complicate efforts to achieve broader improvements in the financial management system. Community contracting of teachers risks undercutting efforts to introduce transparent meritocratic practices of recruitment and promotion, and can also create new opportunities for informality and patronage. Each of these criticisms presumes that broader systemic reforms are directly feasible. But in many settings this is unlikely to be the case: the challenge is to achieve gains in an imperfect world, where the best can be the enemy of the good. Further, partial reforms also have the potential to nudge along incremental change in broader systems: A well-managed road fund could spur more far-reaching public financial management reforms. Community teachers might create an opening for more flexible approaches to civil service reform. Engaging citizens in public sector governance within individual sectors can be a valuable spur to civic engagement more broadly.

Using information to improve accountability at the service provision front-line. Open information on the performance of public agencies can engage citizens in a continuum of ways. At one end is political accountability: citizens can use information on the quality of service provision as part of their decision regarding the reelection of incumbents, at national or local levels. At the other end is the use of performance information by citizens directly involved in the governance of service provision facilities, for example through community-driven approaches discussed a little later. Two intermediate examples illustrate further the potential of empowerment through information.

The first example highlights how detailed, public information can enable citizens to make informed judgments regarding the performance of politicians, policy makers, and providers—and to respond with support, or pressure for change. Frustrated by years of inaction on public services which increasingly were unable to keep up with Bangalore’s dynamism and population pressure, in 1994 a group of citizens introduced the idea of a user survey-based “report card” on public services. Initially, the impact was modest. Nonetheless, the sponsors persisted, establishing a nongovernmental organization (NGO), the Bangalore Public Affairs Center, to institutionalize the effort, building coalitions with other NGOs and repeating the report card survey in 1999 and 2003. Figure 6.5 highlights the extraordinary turnaround in perceptions of the quality of service delivery. The Public Affairs Center describes how this was achieved:

The first and second report cards had put the city’s public agencies under the scanner. The adverse publicity they received, according to many observers, acted as a trigger for corrective action. Inter-agency comparisons seem to have acted as a proxy for competition. Citizen activism and dialogues with the agencies also increased during this early period. These developments prepared the ground for a positive response from the Government. A good example is the political leadership and vision displayed by the Chief Minister S. M. Krishna in the past four years. He provided the framework within which a set of able administrators could set in motion a series of actions and reforms in the agencies. Many civil society groups and the media have stimulated and supported this momentum. Sustaining this movement is the challenge for the future. (See Paul 2002: 71)

FIGURE 6.5Perceptions of service delivery performance in nine Bangalore agencies, 1994–2003

Source: Samuel Paul, Public Affairs Centre, Bangalore, presentation at 6th Global Forum on “Reinventing Governments.”

Note: BMP = Bangalore Municipal Corporation; BESCOM = electricity; BWSSB = water supply; BSNL = Telecom Department; BDA=Land Development Authority; BMTC = Metropolitan Transport Corporation; RTO = Motor Vehicle Licensing

The service provision scorecard approach pioneered in Bangalore has been widely implemented—in countries ranging from Brazil to the Philippines, Ukraine, and Turkey.

The second example involves more hands-on citizen monitoring of official mechanisms and reports of how public resources are used. A few instances are summarized below:

  • In Rajasthan, India, building on the passage of a Right to Information Act in the state in 2001, the Movement for the Rights of Peasants and Workers (MKSS) organized public hearings in rural areas at which figures from the records of licensed distributors of subsidized food rations were compared with figures from the ration books of recipients. Social audits were also carried out of hospitals during which data from medical records were compared with patients’ actual experience. In both cases, large discrepancies between the two sets of figures were revealed. This led to further investigation, which in turn disclosed evidence of corruption, embezzlement, and maladministration.13
  • A Philippines civil society organization (CSO), the Ateneo University Group, set up a citizen monitoring effort, together with government agencies responsible for textbook distribution and highway infrastructure, to make delivery more effective. The project determined that 21 percent of textbooks were not actually delivered to schools designated to receive them, creating losses of more than US$3 million, which the Department of Education promised to rectify. The template developed for this project has been used by many other CSOs.
  • In Tanzania, the Rural Initiatives and Relief Agency helped 10 local communities track government program expenditures for health and education. The pilot projects appear to have helped ensure that commitments to deliver funds were indeed followed through. The expenditure tracking tool has been made available to CSOs in other rural areas of the country.

The latter two instances both were funded by the Partnership for Transparency—an international NGO (supported by Sweden, the United Nations Development Programme [UNDP], and the World Bank) that provides micro-grants to CSOs engaged in fighting corruption. Independent evaluations have shown the large majority of these projects to be successful. The maximum grant size provided by the Partnership for Transparency is US$25,000—underscoring that empowerment through information can be a low-cost, high-return strategy for improving governance.

Decentralization has increasingly been seen as a response to governance dysfunction. It has a dual role in a national governance system. First, democratic subnational governments can offer an important check and balance against central executive power. Second, subnational governments potentially have advantages in the provision of some public services.

Decentralization often is driven by politics. Sierra Leone embarked on decentralization as a way of building simultaneously intergovernmental institutions, local government capacity, and bottom-up accountability. When that country’s civil war ended in 2002, the government initiated a process of national consultation on decentralization. In February 2004, it enacted a progressive Local Government Act, establishing 19 local councils, which, over the period of 2004–8, will take over a large set of responsibilities and resources related to primary education, primary health, agriculture, feeder roads, water, and sanitation. The World Bank has supported fiscal decentralization technically and has helped open political space for development-oriented local politicians to emerge and establish track records by promoting community-based approaches to local infrastructure investment.

Does decentralization help reduce poverty? For this to happen, two sets of accountabilities need to work well. The first comprises downward accountability to local residents. As the 2004 World Development Report (WDR) on improving service provision to the poor put it:

Where decentralization is driven by a desire to move services administratively closer to the people … the assumption is that [it] works by enhancing citizens’ voice in a way that leads to improved services…. Voters make more use of information about local public goods in their voting decisions because such information is easier to come by and outcomes are more directly affected by local government actions. And political agents have greater credibility because of proximity to the community and reputations developed through social interaction over an extended period. But on both theoretical and empirical grounds this could go either way. The crucial question is always whether decentralization increases accountability relative to its alternatives. If local governments are no more vulnerable to capture than the center is, decentralization is likely to improve both efficiency and equity. (World Bank, World Development Report (WDR) 2004: 90)

The second set of accountabilities comprises the allocation of responsibilities between central and local governments. These include the assignment of responsibilities for service provision (clarifying which services are assigned to local authorities, which are assigned to national authorities, and which involve complementary responsibilities for both local and central authorities); the allocation of fiscal resources (including some tax base for local authorities) and fiscal accountability; and regulatory, fiduciary, and other forms of central oversight over local activity.

Clarifying these responsibilities in ways that give each tier of government an incentive to perform its role efficiently is a complex task and deeply political. Decisions over the decentralization “rules of the game” involve a zero-sum contest between national and local politicians and bureaucracies over who controls resources and influence. The interplay between technical complexity and political jockeying can sometimes create difficulties. A comparative review of experience in six East Asian countries (Cambodia, China, Indonesia, Philippines, Thailand, and Vietnam) concluded that

The result [has been] a kind of “institutional limbo”…. Whether by design or as a result of slippages in the implementation process, intergovernmental structures have substantial internal inconsistency. The functions of different levels of government overlap. Bottom-up accountability of locally elected bodies is dampened by top-down methods for appointing key officials. And the discretion given to local authorities in spending unconditional fiscal transfers is effectively curtailed by central government control over human resources. (White and Smoke 2005: 7)

As the 2004 WDR concluded:

Subnational authorities can be efficient providers and regulators of local services under the right institutional incentives and with clarity about who does what—and with what. But greater autonomy can also increase opportunistic behavior and create moral hazard, resulting in costs that diminish accountability and the benefits of decentralization. Good design, sound management, and constant adaptation by both central and subnational authorities are needed to make decentralization work. (World Bank, WDR 2004: 185)

Community-based approaches to local infrastructure investments. In recent years, community-based approaches to local investments have been pursued aggressively under the rubric of community-driven development (CDD)—an approach that “gives control over planning decisions and investment support to community groups and local governments.”14 It seeks to synthesize two types of development interventions which historically have been considered separately from one another—decentralization, as described above, and social investment funds. The latter have been used extensively by donors to transfer resources to poor communities in a participatory way. Between 1999 and 2005, the World Bank alone channeled over US$10 billion to poor communities. According to most reviews, these CDD operations have helped to get services to citizens more cost-effectively and equitably, and have supported participation and accountability. Nonetheless, fierce debate surrounds CDD. Underlying this debate are contrasting views regarding the likely interplay, in weaker governance settings, between bottom-up approaches, and efforts to strengthen national governance systems.

Certainly, the risks are large. As with many donor-funded initiatives, early generation social funds bypassed the public administration with the usual costs associated with parallel implementation (see box 6.5). But in addition, such programs offer a sometimes irresistible opportunity to political leaders. In Peru, for example, between 1994 and 2000, over US$900 million was allocated to the Peruvian Social Fund, FONCODES. The poverty benefits were significant: 80 percent of the resources went to the poorest 40 percent of municipalities. Increasingly, however, it became apparent that FONCODES was being used as a source of patronage and popularity by the country’s populist president, Alberto Fujimori. Disbursements increased in the months directly preceding elections, and while poorer areas were more likely to get funding, those poorer areas that were “swing voters” were favored in resource allocation (Schady 2000).

BOX 6.5Why stand-alone investment projects can be bad for governance

Over the past half-century, stand-alone investment projects have been the dominant response of external donors to the dilemma of ensuring accountability in weaker-governance settings. Projects implemented by autonomous units have a useful role, especially for large infrastructure initiatives. But from a governance perspective, the turn to wholly parallel, projectized arrangements is a conclusion of despair. Such projects substitute external for local accountabilities, thereby perpetuating weaknesses in national governance systems. They typically insulate themselves from the day-to-day business (and rules) of the public sectors in the countries in which they operate: they establish independent project implementation units; set up their own procedures; offer salaries higher than those available in the civil service; and attract away the best talent, demoralizing those who remain. Reducing the prevalence of separate project implementation units is therefore one of the aims of the Paris Declaration (chapter 3).

Practitioners of CDD have worked to address these risks by designing and implementing programs as part of a broader strategy of governance improvement—combining scaled-up participatory resource transfers to communities and longer-run institutional reform, by working closely with line ministries and local governments to help build their capabilities and interactions with community groups. Advocates argue that, especially in weaker governance settings, this hybrid approach can be a powerful way of supporting decentralization. Efforts along these lines are under way in many countries, ranging from Afghanistan to Albania, Brazil, Indonesia, the Kyrgyz Republic, Tanzania, and Zambia. Indonesia offers one example of how this integration can proceed (see box 6.6), but in some other countries programs have failed to evolve much beyond parallel mechanisms. Instead, by seeking to break out of the comfort zone provided by parallel projects, they have brought to the surface the many rivalries and unresolved tensions that characterize countries stuck in a syndrome of weak governance.

BOX 6.6Linking community-based resource transfers and decentralization in Indonesia

Indonesia’s Kecamatan Development Program (KDP) gives communities planning and decision-making power over development resources. The KDP was begun in 1998, in the aftermath of a major financial crisis and political turmoil. Over three phases, close to US$1 billion has passed through the program, which encompasses 28,000 villages—almost 40 percent of Indonesia’s total. The first phase funded more than 50,000 infrastructure and economic activities, benefiting some 35 million poor people.

The main initial motivation for the KDP was that traditional methods for disbursing funds through line ministries had failed. The KDP proved able to provide quick, high-volume disbursements of development funds down to the local level. These are channeled outside the usual government disbursement mechanisms, allowing financing to flow directly to kecamatan localities and village-level bank accounts controlled by communities. Direct financing resolves decision-making bottlenecks caused by central efforts to plan and control activities. KDP disbursement takes an average of two weeks between the time when a village places a request and when funds arrive in the village account. Field studies and audits show that projects deliver a broader range of services at lower-than-normal costs, with greater community involvement, with corruption reduced most effectively by a combination of external audits plus citizen participation.

Since 1998, Indonesia has progressively systematized its formal system of decentralization. Consequently, the second and third phases of KDP have emphasized greater oversight from district parliaments, government monitoring, links with sectoral agencies such as education and health, district matching grants, and local involvement in drafting formal decentralization regulations on village autonomy. The KDP platform has also provided lessons which are being incorporated into local governance reforms to support greater transparency and participation in district policies related to information disclosure, procurement, budget planning, and allocation, leading to higher pro-poor expenditures.

Source: World Bank documents; Wong and Guggenheim 2005.

Surfacing the reality of the difficulty and unpredictability of change in weaker governance settings need not, however, be a bad thing. The challenge for CDD practitioners is to learn more about how to improve the odds: What approaches make success in incrementally fostering sustainable institutional change more likely, and in which country settings? When might the net benefits of a CDD intervention be positive, even with no success in catalyzing institutional change? Demand-driven and incremental institutional reforms such as CDD tend to be judged against a standard of perfection. Unsurprisingly, they fall short. What is needed is some agreement as to what incremental improvement would look like—and a monitoring approach that systematically tracks and assesses incremental, demand-driven institutional change.

Monitoring and Improving National Checks and Balances Institutions

Strong checks and balances institutions are key to a well-functioning national governance system. Some of these checks and balances are global (including global financial and other markets) and are considered elsewhere in this report. The focus here is on national checks and balances institutions. Developmental leadership or a dynamic political movement can sometimes substitute for weak national checks and balances, at least for a period. But over the longer run, well-functioning checks and balances institutions are key to sustainability. They help keep the executive arm of government focused on the public purpose. They are vital for fighting corruption, for ensuring that state actors at all levels use public resources efficiently and effectively, and for helping to ensure that citizens perceive state institutions to be legitimate.

Figure 6.6 disaggregates checks and balances into a constellation, arranged in terms of their “distance” from the executive authority they oversee. The relationship of these institutions with one another is only loosely hierarchical. Depending on a country’s constitution, the judiciary may or may not be a constraint on legislative authority. Citizens may ultimately elect governments but on a day-to-day basis their role is more participatory than hierarchical. We can distinguish three broad groups:

  • An “outer constellation” of civic voice—the rules (for example on freedom of information) and actors (such as the media) that ensure the open operation of civil society—and the transparent flow of information and data that enables citizens to play an informed role in public discourse. (Though not an explicit focus in this report, the discipline provided by competitive markets is an important buttress of this outer constellation.)
  • A “middle constellation” of impartial dispute resolution—in particular the justice system
  • An “inner constellation” of direct over-sight—subnational governments, autonomous oversight agencies, and the legislature.

FIGURE 6.6A constellation of checks and balances institutions

Source: Authors.

The next three subsections consider each of these in turn, focusing on approaches to monitor the quality of the relevant checks and balances institutions and highlighting how some can be strengthened.

The Outer Constellation—Transparency and Voice

Citizen engagement, underpinned by access to high quality information, forms the outermost, and possibly the most important, element of a national system of checks and balances. Figure 6.7 depicts the “virtuous circle of transparency” in a way that highlights the links between the provision of information and state responsiveness.

FIGURE 6.7The virtuous circle of transparency: from disclosure to responsiveness

Source: Authors.

Information reveals the actions of policy makers; this facilitates evaluation and monitoring, activism rises, and with it the level of public debate. Policy becomes more contestable and citizens are motivated by the possibility of holding the government accountable. Communication with the government becomes a two-way flow, generating further demands for more reliable information. The virtuous circle is completed as government practices become more open and more responsive to citizens.

Strengthening the virtuous circle. Several factors are needed for this virtuous circle to work well. First is the production and dissemination of good quality information. Recent initiatives by international agencies, including the IMF and the World Bank, have begun to put in place a framework of internationally accepted norms for the collection and publication of economic and social data. Created in 1996, the Special Data Dissemination Standard (SDDS) is a voluntary standard whose subscribers18—countries with market access or seeking it—commit to meeting internationally accepted levels of data coverage, frequency, and timeliness. SDDS subscribers are required to maintain a Web site that contains the actual data. For countries that do not have market access, the General Data Dissemination System provides a detailed framework that promotes the use of internationally accepted methodological principles, the adoption of rigorous compilation practices, and ways in which the professionalism of national statistical agencies can be enhanced.

These norms can serve as benchmarks for statistical capacity building. There is, as yet, no agreed single measure of statistical capacity, but work carried out as part of the 14th replenishment of IDA’s resources (IDA 14) provides a basis for monitoring changes in the performance of statistical systems. Data are compiled annually on three key dimensions of capacity: statistical practice, data collection, and indicator availability. The indicators are combined to generate overall indicators for each dimension and to produce a single overall indicator. This measure paints a worrying picture of statistical capacity (figure 6.8). As the low average scores in figure 6.8 for “practice” and “collection” signal, many IDA countries lack the ability to provide basic statistics on a regular basis, resulting in a vicious circle—limited, poor-quality data reduce demand for data and lower interest in supporting data collection. Attempts to bridge the availability gap include proxy data or simulations. Capacity has been increasing slowly, if at all, in most poor countries. Especially in the poorest, the impact of projects to strengthen statistical capacity has often been disappointing. Investments are usually not sustained, often because of the piecemeal, short-term nature of projects. The international community has responded to these weaknesses with the Marrakech Action Plan for Statistics (MAPS). The objective of MAPS is to assist all developing countries to either implement or prepare a longer-term national statistical development plan by the end of 2006.

FIGURE 6.8Measuring country statistical capacity: IBRD, IDA, and IDA-Africa, 1999–2005

Source: Country Statistical Information Database ( See IDA (2004) for methodology.

A second factor needed for the virtuous circle of transparency to work well is disclosure—the critical step that turns information into a potent tool for civic accountability. All governments routinely disclose reams of information, including selective information aimed at shaping public opinion. Most democratic societies have some basic standards of disclosure—publication of judicial decisions, or the records of parliamentary debates, for example. More recently, however, global changes in politics, technology, and values have converged to provide a powerful impetus to efforts to strengthen the transparency of governance systems.

This global sea change is reflected in the growing number of countries that have adopted Freedom of Information Laws15—over 50 as of the end of 2004, with efforts under way in an additional 30. The trend is spreading worldwide: in Asia, nearly a dozen countries have either adopted laws or are on the brink of doing so. In South and Central America and the Caribbean, half a dozen countries have adopted laws and nearly a dozen more are currently considering them. South Africa enacted a wide-reaching law in 2001 and many countries in southern and central Africa, mostly members of the Commonwealth, are following that country’s lead.

A third factor is an independent media. Independent media are a crucial pillar of good governance, and a critical link in the accountability chain between the government and the governed. Investigative journalists increase the likelihood of detection of corruption, and punitive action, thus fostering good governance. Mass media also function as a channel of citizen voice, influencing government policies and actions to be more relevant and responsive to citizen preferences.16 As box 6.7 underscores, a vibrant and good-quality media can be a potent development asset.

A fourth factor is an engaged civil society. Perseverance of civil society is crucial as a way of ensuring that greater transparency translates into a change in the internal cultures of public institutions. Development practice has responded to the new focus on civil society—illustrated by the participatory nature of the PRS process. Box 6.8 outlines for Rwanda and Vietnam how the PRS helped crowd civil society more systematically into the policy discourse. This marked a departure from the earlier practice of donors and international financial institutions (IFIs) of focusing narrowly on the executive, and has sought to engage more directly citizens and their elected representatives. A recent review suggests progress on this front (table 6.2) but shows that engagement is well developed in only a minority of countries.

TABLE 6.2Participation in the PRS, 2005
Little actionAction under wayWell developed
PARLIAMENTARY INVOLVEMENTWell developedBhutan, Djibouti, Lao PDR, TajikistanBosnia and Herzegovina, Burkina Faso, Cambodia, Guinea, Honduras, Madagascar, Mali, Mauritania, Moldova, Mozambique, Timor-LesteGhana, Rwanda, Uganda
Action under wayAzerbaijan, Benin, Rep. of Congo, PakistanArmenia, Burundi, Cape Verde, Ethiopia, Kyrgyz Republic, Liberia, Mongolia, Niger, Serbia and Montenegro, Yemen, ZambiaTanzania, Vietnam
Little actionCentral African Republic, Dem. Rep. of Congo, Côte d’Ivoire, Dominica, Guinea-Bissau, Nepal, São Tomé and Principe, Sri Lanka, SudanAlbania, Bangladesh, Bolivia, Chad, The Gambia, Georgia, Guyana, Haiti, Kenya, Lesotho, Malawi, Nicaragua, Senegal, Sierra LeoneCameroon
Source: World Bank 2005b.
Source: World Bank 2005b.

BOX 6.7How media access can influence development outcomes

A variety of studies have documented the link between better-informed citizens and better-performing governments. Besley and Burgess (2002) show that state governments in India were far more responsive to food crises in those states that had high newspaper circulation than in those that did not. Adsera, Boix, and Payne (2003) find, similarly, that corruption is significantly lower in countries with high newspaper circulation. And Strömberg (2004) finds that households with radios during the U.S. Great Depression were much more likely to benefit from relief efforts than were households that lacked them.17

BOX 6.8The Poverty Reduction Strategy process in Rwanda and Vietnam

Rwanda’s PRS process has complemented and helped deepen dialogue initiated through the National Unity and Reconciliation Commission (NURC), set up to promote peace, tolerance, and respect following the 1994 genocide. There is easy public access to the PRS, including a summary in Kinyarwanda, and to fiscal data, as well as the emergence of some leading CSOs vocal on poverty issues. Participatory surveys and stakeholder seminars have been conducted by the Poverty Observatory, a strategic planning and monitoring directorate charged with monitoring PRS implementation. Efforts are under way to merge dialogue held by the NURC with that conducted by the Poverty Observatory. The development debate is being consolidated with stronger analytical underpinnings.

Vietnam produces a Socio-Economic Development Plan (SEDP) every five years. The SEDP typically has been prepared by central government agencies with little consultation outside the communist party. Subsequent to the finalization of the 2001-5 SEDP, the government embarked on developing a PRS—the Comprehensive Poverty Reduction and Growth Strategy (CPRGS)—together with local experts and researchers as well as international and local CSOs. The existence of parallel processes and strategies has caused some confusion on the reference point for policy makers, but has provoked unprecedented lively debate on policy directions in the National Assembly. In preparing the 2006–10 SEDP, the government has committed to emulate the participatory approach to planning that characterized the preparation of the CPRGS.

Monitoring transparency and voice. One broad and one specific set of indicators are used in this subsection to monitor transparency and voice (TV). The broad indicator captures the overall TV environment—including the human rights and political governance dimensions. The specific indicator focuses more directly on those aspects of transparency most directly relevant for achieving the Millennium Development Goals (MDGs).

Two broad indicators were considered for this report—the aggregate KK “voice and accountability” indicator, and a related indicator that focuses more narrowly on transparency. Table 6.3 reports the better-established “voice and accountability” indicator—noting also that at 0.88 the correlation between the voice and accountability and the transparency results is high.18 The indicator is estimated from 19 separate disaggregated sources—each of which focuses on a specific aspect of TVA. KK note that these include “a number of indicators measuring various aspects of the political process, civil liberties and political rights. These indicators measure the extent to which citizens are able to participate in the selection of governments. We also include indicators measuring the independence of the media, which serves an important role in monitoring those in authority and holding them accountable for their actions” (see Kaufmann, Kraay, and Mastruzzi 2005).

TABLE 6.3KK voice and accountability 2004, 66 low-income countries
In relevant third (with 95% certainty)In relevant third (with less than 95 % certainty
Top thirdBenin, Ghana, India, Lesotho, Mali, Mongolia, São Tomé and Principe, Senegal, Serbia and MontenegroIn top half (with 95 % certainty): Albania, Bolivia, Bosnia and Herzegovina, Comoros, Guyana, Honduras, Madagascar, Mozambique, Nicaragua, Niger, Papua New Guinea, Solomon Islands
Middle thirdArmenia, Bangladesh, The Gambia, Guinea-Bissau, Indonesia, Malawi, Moldova, Nigeria, Sierra Leone, UgandaCould be in bottom third: Azerbaijan, Cambodia, Congo, Djibouti, Nepal, YemenCould be in top third: Burkina, Faso, Georgia, Kenya, Sri Lanka, Tanzania, Zambia
Bottom thirdDem. Rep. of Congo, Côte d’Ivoire, Eritrea, Haiti, Lao PDR, Pakistan, Sudan, Uzbekistan, Vietnam, ZimbabweIn bottom half (with 95 % certainty): Angola, Bhutan, Burundi, Cameroon, Central African Republic, Chad, Ethiopia, Guinea, Kyrgyz Republic, Mauritania, Rwanda, Tajikistan, Togo

Table 6.3 reports the distribution of the sample of 66 IDA-eligible countries across three groups, distinguishing among countries according to whether one can be at least 95 percent confident, using a two-tailed test, that given measurement errors they indeed fall into the category in which they are located. As with all governance measures, the indicator provides some useful benchmarking, but only for a subset of countries is it possible to assert with confidence that their environment for voice and accountability is relatively strong or relatively weak.

If we locate the high- and middle-income countries in the KK VA sample using the same cut-off points as for the 66-country IDA-eligible sample, a considerable number falls below table 6.3’s top-third group. While all of the high-income OECD member countries are significantly above the top third cut-off point for the 66-country sample, six non-OECD high-income countries are located below this cutoff point. Of 77 middle-income countries, 30 rank below the top-third cut-off point, and 16 of these score low enough to be in the bottom-third of the 66-country sample.

The specific indicators are taken from the Global Integrity Index (GII). As box 6.9 describes, the GII is an example of “good practice” methodology for governance indicators. As explained earlier, as with all governance indicators, the estimates have some margin of error. But because each measure is specifically defined, it provides “actionable” information for governance reform. The specific GII indicators cover the range of the checks and balances constellation, though so far country coverage remains limited. Table 6.1 reports the scores for three specific GII subindicators, which measure facets of the environment for transparency and civic participation for 25 OECD, middle-, and low-income countries. Low-income countries lag, especially in the right of access to information.

The Middle Constellation—Justice and the Rule of Law

Justice sector reform and promoting the rule of law have emerged as key goals of development policy. The justice sector covers a vast array of institutions, issues, and functions. In the broadest terms, it can be defined as the institutions and processes by which laws are devised and enforced. It includes legal services and their providers (for instance, lawyers and paralegals), police, prosecutors, the judiciary, courts and their officials, other institutions that resolve disputes, and institutions that execute judgments. The justice sector fulfills two distinct, but complementary, sets of essential functions. It provides services to citizens, such as safety and security, and resolving disputes. And it can help to constrain the arbitrary and discretionary use of state power.

TABLE 6.4Global Integrity Index—transparency and civic participation (by group)
Civil society organizationsAccess to information lawFreedom of the media
OECD countries958791
Middle-income countries886074
Low-income countries824179
Source: The covered by the index, grouped in the three categories shown above, are Australia, Germany, Italy, Japan, Portugal, United States; Argentina, Brazil, Guatemala, Mexico, Namibia, Panama, Philippines, Russian Federation, South Africa, Turkey, Ukraine, Republica Bolivariana de Venezuela; and Ghana, India, Indonesia, Kenya, Nicaragua, Nigeria, Zimbabwe.Scoring: Each question within each category is scored on a 0-100 scale, using specific guidelines. The category score is the average of the scores for the individual question.
Source: The covered by the index, grouped in the three categories shown above, are Australia, Germany, Italy, Japan, Portugal, United States; Argentina, Brazil, Guatemala, Mexico, Namibia, Panama, Philippines, Russian Federation, South Africa, Turkey, Ukraine, Republica Bolivariana de Venezuela; and Ghana, India, Indonesia, Kenya, Nicaragua, Nigeria, Zimbabwe.Scoring: Each question within each category is scored on a 0-100 scale, using specific guidelines. The category score is the average of the scores for the individual question.

BOX 6.9The Global Integrity Index as a tool for governance monitoring

The GII, developed by the Center for Public Integrity, focuses on measurement of “the existence and effectiveness of mechanisms that prevent abuse of power and promote public integrity, and on the access that citizens have to their government.” The GII has a nested design—with answers to more than 290 detailed questions providing the basis for estimating a variety of indicators at different levels of aggregation. This enables users to move from the more aggregated indicators to the most disaggregated, and thereby identify strengths and weaknesses. Country-specific scoring is done by a diverse panel of in-country experts, each operating individually to avoid “contamination by consensus,” and with rigorous, “blind” peer review. So far, the index has been estimated only for 25 countries—6 OECD countries, 12 middle-income countries, and 7 IDA-eligible low-income countries. Global Integrity aims to increase the number of countries covered to over 100 by the end of 2006. It remains to be seen, though, whether the GII will receive the sustainable financing and broad legitimacy necessary for it to become a widely used part of the arsenal of governance indicators. The table below details the questions for 10 indicators that are most directly relevant to the dimensions of checks and balances highlighted for this report.

Some specific GII indicators

Indicators of transparency and civic participation

Civil society organizations—In law, do citizens have a right to form CSOs? Do they in practice? Can citizens organize into trade unions? In practice, do CSOs actively engage in public advocacy campaigns? Are civil society activists safe when working on corruption issues?

Access to information law—In law, do citizens have a right of access to information? In practice, is the right of access to information effective?

Freedom of the media—In law, is freedom of the media guaranteed? In law, is freedom of speech guaranteed? Are citizens able to form media entities? Is the media able to report on corruption? Are journalists safe when investigating corruption?

Indicators of justice and the rule of law

Judiciary—In law, is the independence of the judiciary guaranteed? Is the appointment process for high court judges effective? Can members of the judiciary be held accountable for their actions? Can citizens access the judicial system? In law, is there a program to protect witnesses in corruption cases? Are judges safe when adjudicating corruption cases?

Rule of law and access to justice—In practice, does the criminal justice process function according to the rule of law? In law, is there a general right of appeal? Are citizens protected from detention without trial? Are individual economic rights guaranteed?

Law enforcement—Is the law enforcement agency (that is, the police) effective? Can law enforcement officials be held accountable for their actions?

Indicators of direct oversight

Legislature—Can members of the legislature be held accountable [by the judiciary] for their actions? In law, are members of the legislature subject to prosecution? Are there regulations governing conflict of interest by members of the legislature? Can citizens access the asset disclosure records of members of the legislature? Can citizens access legislative processes and documents? Does the legislature have control of the budget? Can citizens access the national budgetary process?

National ombudsman—In law, is there a national ombudsman, public protector, or equivalent agency covering the entire public sector? Is the national ombudsman effective? Can citizens access the reports of the ombudsman?

Supreme audit institution—In law, is there a national supreme audit institution, auditor general, or equivalent agency covering the entire public sector? Is the supreme audit institution effective? Can citizens access reports of the supreme audit institution?

Anticorruption agency—In law, is there an agency (or group of agencies) with a legal mandate to address corruption? Is the main anticorruption agency effective? Can citizens access the main anticorruption agency?

Source: Center for Public Integrity 2004.

A well-functioning justice sector is expected to reflect certain basic qualities, most notably the rule of law. Indeed, the term “rule of law” is sometimes seen as synonymous with, or used as a proxy for, a well-functioning justice sector. Yet, as with the justice sector generally, there is no shortage of conceptions as to what the rule of law is said to entail.19 Both the rule of law and justice reform have been defined broadly with reference to their essential role in ensuring democracy and human rights—or narrowly with reference to their impact on predictability for business processes and investment climate. Differences in priority and definition will have a direct impact on which reform efforts are prioritized to improve the functioning of the justice sector and the rule of law and, in turn, what should be measured.

There already exist a number of broad and specific indicators on justice and rule of law issues. Most of them are not aimed specifically at justice and the rule of law, however, and only incorporate a section on it, as part of a broader focus or theme. Others, while focusing on justice and rule of law issues, focus on specific processes or institutions and do not seek an overall view of the state of the rule of law. The two sets of broad indicators that make global comparisons among countries comprise the KK “Rule of law” aggregate indicator, and the “Property rights and rule-based governance” CPIA criterion. As examined in chapter 5, the KK Rule of Law indicator aggregates data from multiple sources, namely: “… several indicators which measure the extent to which agents have confidence in and abide by the rules of society. These include perceptions of the incidence of crime, the effectiveness and predictability of the judiciary, and the enforceability of contracts …” The CPIA-rules indicator focuses primarily on the extent to which the legal system facilitates private economic activity, but also looks at broad outcomes (safety), specific outcomes and functions (provision of business licenses, contract enforcement), and formal characteristics of the system. The correlation between the two indicators is quite high, at 0.83. Combining the two indicators for 66 low-income countries yields 12 countries that are both in the top third of the KK rankings for the indicator, and have a CPIA-rules score of 3.5 or above (Armenia, Bhutan, Ghana, Honduras, India, Lesotho, Madagascar, Malawi, Mali, Senegal, Sri Lanka, and Tanzania).

Efforts are under way to develop more specific, actionable indicators. Three are noteworthy. The first two comprise the Doing Business and Investment Climate surveys. As discussed in chapters 1 and 5, both include indicators that can be used to monitor the performance of the justice system relevant to specific features of the business environment. (See this chapter’s annex for specific indicators relevant to the justice system.) The third comprises the three GII justice and the rule of law subindicators identified in box 6.9. Table 6.5 summarizes the subindicator scores for 25 OECD, middle-, and low-income countries. The results suggest that for many of them, improving the justice system is a large challenge. This area has been prominent in the development agenda only for a short period, and much remains to be learned as to what reforms work. Box 6.10 summarizes some emerging lessons.

TABLE 6.5The quality of some attributes of the justice system in 25 countries (by group)
JudiciaryRule of law and access to justiceLaw enforcement
OECD countries799390
Middle-income countries717463
Low-income countries587259
Source: See table 6.4 for a list of countries in each category.
Source: See table 6.4 for a list of countries in each category.

BOX 6.10Strengthening justice—three initial lessons

The World Bank’s work in justice reform has largely focused on formal institutions, beginning with the courts, and institutions supporting the market. More recently, an increased priority has been placed on access to justice, and attention has also focused on the link between formal and informal institutions, as well as the role that informal institutions play in helping to fulfill the roles of the justice sector. Three lessons emerge:

Lesson #1: Do not work on independence without simultaneously working on accountability. Early efforts to increase independence of courts tended to focus on how judges are selected and evaluated, and their capacity to deliberate and decide cases without undue influence from other branches of government or other pressure. Accountability of judges, particularly to the public, was sometimes not given as much emphasis as judicial independence. Yet citizens’ greatest complaints about courts tend to focus first on corruption and second on delay, neither of which are likely to improve substantially without greater judicial accountability.

Lesson #2: Reforms that seek to overhaul the way justice systems operate will endure only if they deliberately strengthen the management of the reform process. Profound institutional change requires professional change management at the planning and implementation stages. This is pointed out quite clearly in studies of the criminal justice reform processes in many countries of Latin America, in which attempts to replace written processes with oral hearings and trials suffered backsliding in the absence of improved management of the institutions involved.

Lesson #3: If one is looking to increase the amount or quality of justice for the average citizen, look at both formal and informal justice systems. Most World Bank projects, and many of the justice reform projects of major donors, focus on improving the administration of justice through formal institutions—courts, prosecutors’ offices, ministries of justice, and so on. Assistance has also been provided to legal aid institutions both formal and informal, and major donors have supported development of alternative dispute resolution mechanisms such as mediation and arbitration—though often as a recognized, and sometimes court-annexed, part of a formal proceeding. The role of informal, local justice systems—which in some countries govern as much as 95 percent of the population—has only recently begun to receive more attention.

The Inner Constellation—Direct Oversight

Direct oversight institutions in the first two inner rings of figure 6.6 include elected sub-national authorities, ombudsmen, supreme audit institutions (with independent authority to review national accounts, monitor the probity with which public resources are used, and report on their findings to parliament), anticorruption agencies (with independent authority to investigate and sometimes also prosecute accusations of corruption), and the national legislature, to which the executive generally is directly accountable. Though this section focuses principally on monitoring, box 6.11 illustrates for one direct oversight institution—the legislature—some of the challenges of improving performance.

The GII provides disaggregated measures of the quality of direct oversight. As table 6.6 summarizes, in most OECD and some of the middle-income countries restraints on the executive are rated as high; Zimbabwe stands out among the low-income countries considered, as having few effective executive constraints.

TABLE 6.6The quality of some direct oversight institutions in 25 countries (by group)
LegislatureNational ombudsmanSupreme audit institutionsAnticorruption agency
OECD countries79839885
Middle-income countries66819268
Low-income countries73737877
Source: See table 6.4 for a list of countries in each category.
Source: See table 6.4 for a list of countries in each category.

One of the most widely used aggregate indicators is the “executive constraint” measure of the POLITY data set.20 This measure refers to “… the extent of institutionalized constraints on the decision-making powers of the executive. Such limitations may be imposed by any ‘accountability group.’ “In Western democracies these are usually legislatures. Other kinds of accountability groups are the ruling party in a one-party state, councils of nobles or powerful advisors in monarchies, the military in coupprone polities, and in many states a strong, independent judiciary.

BOX 6.11Legislative oversight in Africa—a work in progress

A recent study of legislatures in four African countries—Benin, Ghana, Kenya, and Senegal—identified large differences in their effectiveness. The Kenyan parliament emerged as the most independently assertive; the Ghanaian and Beninese legislatures were described as semi-independent (and certainly more independent as of 2002 than 10-15 years earlier); but the Senegalese legislature was judged to be almost entirely subservient to the executive.

These variations in independence translated into variations in how parliamentarians allocated their time between policy-related and constituency-support activities, with the Kenyans most (and the Senegalese least) preoccupied with the former. But even in Kenya, there was only limited real engagement with the budgeting process (as distinct from other aspects of policy making), and even this engagement tended to focus narrowly on the implementation of spending commitments within the districts of individual members. Multiyear delays in the presentation of audits have led some parliamentarians to refer disparagingly to audit committees as the “post-mortem committees.”

Legislative strengthening is best seen as a complement to related governance improvements involving civil society. Civil society organizations are sources of technical expertise and can provide specialized legislative committees with information about the effects of public policies and policy alternatives:

Treating legislatures as self-contained entities that can be fixed by repairing internal mechanisms is unlikely to get very far. Rather, … it is more useful to think in terms of helping a society develop the capacity to enact laws that incorporate citizens’ interests … [this means] working with many people and groups outside the legislature. (Carothers 1999: 107, 186-87)

The internal workings of legislatures can nonetheless be important to give parliaments the ability to sustain their interventions, bring significant independent expertise to bear, and exercise effective leverage in their oversight activities. These might include creating a permanent, independent nonpartisan staff for parliament; making changes in internal rules to permit tougher scrutiny of key executives; establishing and strengthening specialized committees (including those focused on budget, education, health, roads, rural development, and cross-cutting themes—including poverty reduction); building links with CSOs and independent policy-advisory institutions; establishing well-paid research capabilities to serve parliament; and “putting their own house in order” to improve credibility, for example by establishing codes of conduct for members of parliaments, and by making campaign financing transparent, honest, and constrained.

Source: Barkan, Adamolekun, and Zhou 2004; World Bank Institute.

POLITY IV scores on a 7-point scale. Most OECD and many middle-income countries score in the high and medium-high categories, but many countries that confront difficult governance issues also score medium-high on the indicator. A country’s political and bureaucratic leadership can find itself constrained either as part of a well-functioning overall institutional environment or as part of an overall syndrome of state weakness. The final section of the chapter considers these issues further.

Sequencing Governance Reforms

This final section brings together some of the individual governance measures examined in this chapter to pose a complex question—how to engage countries with an uneven mix of governance strengths and weaknesses? This is a somewhat different problem from the question of how to engage with countries with severe all-round governance weaknesses, in part because the uneven mix may reflect turnaround cases rather than stable, clientelistic equilibria.

Trajectories of change. Table 6.7 applies the governance indicators used in earlier sections to identify 28 countries that rate well in the quality of either their bureaucracies or their checks and balances institutions. While 10 countries rate well in both areas, performance across the remaining countries is uneven. Ten countries (Rwanda and Vietnam, for example) have relatively capable public bureaucracies, but less strong checks and balances institutions. And the pattern is reversed in the other 8 countries (Albania and Lesotho, for instance) where relatively stronger indicators for checks and balances are not matched by correspondingly capable public bureaucracies.

TABLE 6.7State capacity and state accountability
Quality of checks and balances institutions
Bureaucratic capabilityMedium or LowHigher
Higher10 countries (Azerbaijan, Bhutan, Burkina Faso, Ethiopia, Indonesia, Pakistan, Rwanda, Tanzania, Uganda, Vietnam)10 countries (Armenia, Benin, Bolivia, Ghana, Honduras, India, Mali, Senegal, Serbia and Montenegro, Sri Lanka)
Medium or low38 countries8 countries (Albania, Guyana, Lesotho, Moldova, Mongolia, Nicaragua, Niger, Papua New Guinea)
Source: Collated by the authors.Notes: States with higher bureaucratic capability are those with CPIA-budget scores of 4 and above, or both CPIA-admin and CPIA-budget scores of 3.5 and above. States with higher quality of checks and balances institutions are those that score “high” on at least two of the voice and accountability, rule of law, and executive constraints broad checks and balances measures reviewed in earlier subsections of this chapter.
Source: Collated by the authors.Notes: States with higher bureaucratic capability are those with CPIA-budget scores of 4 and above, or both CPIA-admin and CPIA-budget scores of 3.5 and above. States with higher quality of checks and balances institutions are those that score “high” on at least two of the voice and accountability, rule of law, and executive constraints broad checks and balances measures reviewed in earlier subsections of this chapter.

Why might patterns such as those in table 6.7 be observed? figure 6.9 illustrates three possible trajectories for governance turnarounds. These might vary depending on both the initial political impetus within a country and the longer-term historical processes that can shape and constrain political and institutional reform.

FIGURE 6.9Governance turnarounds: three trajectories

Source: Authors.

In trajectory 1 a developmentally oriented political leader takes power in a hitherto clientelistic setting (as when President Rawlings took power in Ghana in the early 1980s, or President Museveni in Uganda in the mid-1980s). A common early focus of reform might be to liberalize the economy and strengthen the performance of the public sector. This can emphasize strengthening the capabilities of the public bureaucracy—public administration and financial management, and the service provision frontline. The strengthening of checks and balances institutions can initially be a low priority, though countries vary as to whether there is an initial weakening of checks and balances relative to the status quo (as in Ghana) or a modest improvement (as in Uganda). But once the reform process matures, the priority for government reform might usefully shift from strengthening bureaucratic authority to enhancing stability by increasing transparency, participation, and accountability of the state. This subsequent phase is, in practice, advanced in Ghana and more tentative in Uganda.

In trajectory 2 a turnaround is initiated by a move to political pluralism. Examples in Africa include democratic transitions over the past 15 years in countries as varied as Benin, Kenya, Malawi, Nigeria, and Zambia. Examples in Europe and Central Asia include Albania and Romania in the early 1990s. The initial political opening is only a first move in the direction of stronger checks and balances institutions. The dotted line signals a second phase of governance reform in which the momentum for greater accountability continues—and the reinvigorated legitimacy that comes from stronger participation and accountability provides a platform for ongoing improvements in bureaucratic capability. Whether and how this subsequent phase unfolds is, of course, an empirical matter.

In trajectory 3 turnaround starts from a state collapse. Sometimes external intervention helps to reintroduce the precondition for an effective state: a monopoly on the legitimate use of violence. This umbrella of security provides an opportunity for reestablishing both the bureaucracy and checks and balances institutions. Once a new base has been established, the process can continue in a balanced way, with momentum coming from the newly reestablished domestic institutions. This pattern is evident in countries ranging from Bosnia and Herzegovina to Mozambique.

These varying trajectories pose some dilemmas for the design and sequencing of governance reform:

  • Change that focuses first on improvements in bureaucratic quality has the potential for rapid gains in public sector performance. But without a subsequent effort to strengthen checks and balances institutions, it risks subsequent reversal—perhaps by a reversion to corrupt behavior by the political leadership, perhaps by a loss of legitimacy with citizens.
  • Change that begins with a political opening can generate a surge of confidence and improve the climate for private investment. But unless the gains are consolidated, the country risks becoming trapped in a cycle of what Thomas Carothers (2002) has called “feckless pluralism”—with short-lived governments repeatedly voted out of power, never having sufficient support and longevity to build the base of bureaucratic capability on which effectiveness and legitimacy will eventually depend.

These varying trajectories also pose dilemmas for a country’s development partners—both for scaling up aid and for ensuring the sustainability of development support.

Scaling up across different country settings. Consider first the dilemma uneven bureaucratic capability poses for efforts to scale up aid. As table 6.7 summarizes, perhaps about 20 low-income aid recipients currently have budget management and administrative systems reasonably capable of targeting spending on poverty reduction priorities—and of executing and monitoring spending in a comprehensive, credible, and transparent way. With a few exceptions, World Bank budget support via Poverty Reduction Support Credits has been targeted to these institutionally stronger countries, in the upper quintiles of the CPIA (see Gelb and Eifert 2005).

What might be the “mutual accountability” basis for scaling up aid in the remaining countries? Three possibilities are worthy of note.

First, even where current systems fall short, budget support might be scaled up for countries based on a clearly improving trend in the quality of their budget and administrative management systems. This is not simply because the additional resource transfers can be poverty reducing: a shift from project aid to budget support can also be seen as an investment in strengthening country systems. (See Gelb and Eifert 2005 for this argument.) As the principles of the Paris Declaration on Aid Effectiveness underscore, heavily fragmented project aid complicates and disrupts national systems, whereas budget support, combined with technical assistance, can facilitate the improvement of these systems, particularly if scaling up depends on continuing system improvements (see chapters 3 and 4). Tanzania illustrates this potential: it has shown rapid improvement in budget management systems since 2001 and has been a beneficiary of progressively scaled-up budget support over the period. The other examples highlighted in this chapter suggest that, for countries determined to improve their administrative budget systems, achieving a “good enough” standard within, say, 5-10 years may be feasible. Budget support might be initiated quite early in the cycle of improvement—and scaled up as long as the carefully monitored improvement continues to be evident.

Second, priority could be given to reforms that foster transparency—in budget management and more broadly. Transparency relies on public information as a source of pressure for better public sector performance—in a less technocratic way than is implied by top-down reforms of bureaucratic capability. To be sure, the route from transparency to performance is circuitous, and the timing of impact, unpredictable. So far, no study definitively pinpoints the relationship between transparency and performance. But many examples, including some in this report, highlight the potential—from the tracking of education expenditures in Uganda, to service delivery report cards in Bangalore and Brazil, to the impact of media prevalence across India’s states. Even with continuing weakness in bureaucratic capability, a case could thus be made for scaling up aid (including some component of budget support) to countries that clearly commit themselves to facilitating transparency in how public resources—and state power more broadly—are used.

The third possibility for countries is to target scaled-up aid more directly toward poverty-reducing services, which can be done in several ways. A key distinction here is between countries where bureaucratic capability may be on the upturn but is only at an early stage of improvement—and those where there is little sign of political commitment to improve governance and capacity. In the former group, the sectorwide approaches described earlier that focus on improving governance and service provision in part of the overall system are attractive. In the latter group, the focus might be on infrastructure and other service delivery investment projects—complete with project implementation units and related mechanisms that operate apart from country systems. Box 6.5 detailed some well-founded objections to these approaches. But where there is little political commitment to improve country systems and little sign that governments would have targeted pro-poor spending, these objections have less relevance.

Sustainability—bringing checks and balances onto the agenda. In the short-term, aid can thus straightforwardly be scaled up to countries with improving budget and administrative systems. But a longer-run challenge looms. While trajectories of improvement vary, and in the short run no one type of turnaround is superior to another, unless the gains in the bureaucracy and the checks and balances institutions eventually evolve in a balanced way, the risk is high that initial improvements in governance will not be sustained. Over the medium term, it may therefore become necessary to focus the governance dialogue on the complementary aspects of the bureaucratic and institutional agenda that are not spontaneously coming to the fore. How can these sensitive issues best be addressed?

A first consideration is timing. In some settings it may not be practical to press very early in a turnaround process for far-reaching reforms of checks and balances. In Uganda, for example, in the immediate aftermath of the Amin and latter-Obote years the state was in total collapse, and the ability of the new government to assert authority over the nation was limited. Under such circumstances, it is difficult to find fault with the readiness of donors to support government efforts to focus principally on strengthening bureaucratic capability and development policy—and to emphasize decentralization as a means of bringing government closer to the people. As is well known, far-reaching reductions in Ugandan poverty resulted from the early actions of government, and donor support. At the same time, it also seems clear that development partners can wait for too long—until it is too late to put the challenge of strengthening checks and balances squarely on the agenda. An example here is President Suharto’s Indonesia—where a failure to focus early enough on checks and balances was associated with rising corruption, financial crisis, and a difficult process of political succession which led to some significant reversal of the development gains of earlier decades. Overall, the track record of recent decades suggests that (perhaps partly as a consequence of the Cold War) in many countries development partners may have waited too long before putting checks and balances institutions higher on the agenda of development dialogue.

A second consideration is that our knowledge of how to get “from here to there” is less developed than our understanding of what well-functioning checks and balances institutions should look like. One exception to this proposition is the value of transparency, which is essential for the effective working of all checks and balances and which can and should be enhanced in almost all settings—at quite low cost. Donors and IFIs can play a direct role here, including by ensuring that all analytical work is made widely available, with translation into local languages. We know less as to when and how improvements in transparency translate into genuine gains in accountability and performance, but it does seem to be at least a necessary condition.

Given the limitations of current knowledge, perhaps all that can be offered at this stage is a modest process suggestion. Even—or perhaps especially—when it is still uncomfortable, governments and their development partners might usefully begin a dialogue on how to strengthen checks and balances institutions. The aim of this dialogue would be to agree on a phased sequence of steps for strengthening these institutions, perhaps emphasizing more those directly relevant to MDG outcomes—transparency, gender, the justice system, and local governance for example. Recipient countries would be accountable for proceeding with an agreed sequence. In return, they would enjoy more certainty over what is expected by the international community. Donors, in turn, having agreed on a way forward would be expected not to shift the goalposts after the fact.


1.For details of, and results from, the Code of Fiscal Transparency, see and Hameed (2005).
2.The correspondence between the four dimensions of CPIA-budget—(a) through (b)—and the PEFA PFM framework in figure 6.1 is as follows: (a) corresponds to policy-based budgeting and the formulating process; (b) corresponds to a combination of the comprehensiveness of budget coverage, credibility that the budget is realistic and implemented as intended, plus the budget execution arrangements for the exercise of predictability, control, and stewardship in the use of public funds; (c) corresponds to the systems of accounting and recordkeeping to provide the information needed for proper management, plus auditing mechanisms that ensure external scrutiny. Intergovernmental finance—the focus of (d)—is not directly incorporated in the PEFA framework. For the detailed scoring system used in the CPIA, see
3.See the PEFA Web site at
4.For some questions the benchmark was set at the score of B and for others at A. Further details, including the descriptions of how to score each question, are available at
5.Niger’s ranking improved in eight categories and declined in three. Five of the improvements were sufficient to achieve the benchmark (but all three declines were from benchmark level to below).
6.See, for example, the Africa Action Plan recently issued by the World Bank.
7.The composite measure comprises a subset of the data used for the KK government effectiveness aggregate indicator; it excludes responses on the quality of public service provision and on the credibility of government’s commitment to policies, and it excludes the CPIA-admin (because it is being used as a cross-check).
8.World Bank, Operations Evaluation Department (1999: ii-iii). The OED review highlighted four specific weaknesses in Bank-supported interventions: the poor quality of information on civil service reform performance, needed for monitoring and evaluation; the limited role afforded to strategic management and cultural change; the absence of checks and balances on arbitrary action; and a failure to appreciate key contextual contexts.
9.See the articles by Mike Stevens and Stefanie Teggemann; Kithinji Kiragu, Rwekaza Mukandala, and Denyse Morin; Poul Engberg-Pedersen and Brian Levy in Levy and Kpundeh (2004).
10.For the detailed analysis on which this subsection is based, see International Monetary Fund and the World Bank (for the Development Committee), “Fiscal Policy for Growth and Development: An Interim Report,” April 2005.
11.For a detailed discussion of road sector reform, see Heggie and Vickers (1998).
12.For a review of the role of community schools in Francophone Africa, and the relevant lessons from international experience, see Gershberg and Winkler (2004).
13.For information on the MKSS, visit or contact the organization at Press coverage of MKSS activities has been extensive and includes Deccan Herald (September 21, 2003) and Mail & Guardian Newspaper, South Africa (February 20, 2004).
15.This review of the role of Freedom of Information Laws is adapted from Bellver and Kaufmann (2005).
16.For a pioneering, in-depth analysis, see Islam (2002).
18.Consideration was given to using three new measures of transparency produced by Kaufmann and a co-author, but it was decided to stick with the better-known and more thoroughly scrutinized “voice and accountability” measure. The correlation between the aggregate voice and aggregate transparency indicators is 0.88. As for the two transparency subindicators, the correlation between “voice and accountability” and “political transparency” is 0.93. The correlation is much lower—0.41—with the measure of “economic and institutional transparency.” But the latter measure generates very large standard errors relative to the other KK indicator, raising questions as to the coherence of the underlying concept it is intended to measure. See Bellver and Kaufmann (2005).
19.The inception of the term “rule of law” reaches to the roots of Western political thought—for instance, in early Greek and Roman political writings—and also appears as a cornerstone in the genesis of various European legal and political systems. See, for example, Rigo and Gruss (1991: 5–8) for an excellent overview of the origins of the rule of law, in Greek and Roman thinking, as well as its emergence as central tenets in the British and French legal orders. For a more extensive discussion, covering both Western and non-Western conceptions and origins of the Rule of Law, see Hager (2000: 3–20).
20.The POLITY project, ( run from the University of Maryland, is the world’s most widely used data resource for monitoring regime change and studying the effects of regime authority. For details, see Marshall and Jaggers (2002: 23–24).

ANNEX Doing Business Indicators and Investment Climate Surveys—Some Useful Measures for Governance Monitoring

A: Measures of corruption (ICS)

Unofficial payments for firms to get things done (% of sales)

Average value of gifts or informal payments to public officials to “get things done” with regard to customs, taxes, licenses, regulations, services, and so on. The values shown indicate a percentage of annual sales.

Firms expected to give gifts in meetings with tax inspectors (%)

Percentage of firms for which a gift was expected in meeting with tax inspector.

Value of gift expected to secure government contract (% of contract)

Percentage of contract value expected as a gift to secure government contract.

Corruption a “major or severe” obstacle (% of firms)

Percentage of firms that say corruption is a major or severe obstacle to the operation and growth of their business.

B: Measures of transactions costs associated with red tape

(i) Doing Business indicators

Starting a business

The number of procedures, average time spent during each procedure, and official cost of each procedure involved in incorporating and registering a commercial or industrial firm.

Dealing with licenses

The number of procedures, average time spent during each procedure, and official cost of each procedure involved in obtaining necessary licenses and permits, completing required notifications and inspections, and obtaining utility connections (using construction of a warehouse as a benchmark).

Registering property

The number of procedures, average time spent during each procedure, and official cost of each procedure involved in registering property (using as a benchmark the case of an entrepreneur who wants to purchase land and buildings in the largest business city—already registered and free of title dispute).

Trading across borders

Number of documents, approvals, signatures, or stamps required, and the time and associated cost necessary to comply with all procedural requirements for exporting and importing a standardized cargo of goods.

(ii) Investment climate indicators

Senior management time spent dealing with requirements of regulations (%)

Average percentage of senior management’s time that is spent in a typical week dealing with requirements imposed by government regulations (such as taxes, customs, labor regulations, licensing, and registration), including dealings with officials, completing forms, and so on.

Time spent in meetings with tax officials (days)

Average time firms spend in meetings with tax officials (days).

Time to claim imports from customs (days)

Average number of days that it takes from the time goods arrive in their point of entry (for example port, airport) until the time they can be claimed from customs.

Customs and trade regulations a “major or severe” obstacle (% of firms)

Percentage of firms that say customs regulations present major or severe obstacles to the operation and growth of their business.

C: Measures of quality of provision of specific public services (ICS)

Delay in obtaining a connection (days) [electricity, water, telephone]

Average actual delay, in days, that firms experience when obtaining a connection, measured from the day the establishment applied to the day it received the service or approval.

Supply failures and outages (days) [electricity, water, telephone]

Average number of days per year the establishment experienced supply failures and outages from the public network.

Value lost to supply failures (% of sales) [electricity, water, telephone]

Total losses over the course of a year resulting from interruptions in electricity service, as a percentage of sales, including losses due to lost production time from the outage, time needed to reset machines, and production and sales lost due to processes being interrupted.

Supply weaknesses a “major or severe” obstacle (% of firms) [electricity, water, telephone]

Percentage of firms that say the shortcomings of the infrastructure present major or severe obstacles to the operation and growth of their business.

D: Measures of justice and the rule of law (ICS, except “Enforcing contracts”)

Enforcing contracts (DB)

The number of procedures involved from the moment a plaintiff files a lawsuit over a payment dispute until actual payment, and the associated time in calendar days, and cost, necessary to resolve the dispute.

Confidence in the judiciary system (%)

Percentage of firms that agree with the statement, “I am confident that the judicial system will enforce my contractual and property rights in business disputes.”

Dispute resolution time (weeks)

Average amount of time, in weeks, that it usually takes to resolve an overdue payment.

Legal system a “major or severe” obstacle (% of firms)

Percentage of firms that say the legal system presents major or severe obstacles to the operation and growth of their business.

Crime, theft, and disorder a “major or severe” obstacle (% of firms)

Percentage of firms that say crime, theft, and disorder present major or severe obstacles to the operation and growth of their business.

    Other Resources Citing This Publication