monitoring and improving governance subsystemssiteresources.worldbank.org/intglobalmonitoring... ·...

Governance being multidimensional,both reform and monitoring to sup-port reform are best tackled through

specific subsystems. The framework laid outin the previous chapter distinguished amongthree broad subsystems: cross-cutting publicfinancial management and administrativecontrol agencies that underpin bureaucraticcapabilities; front-line service provision andregulatory agencies; and checks and balancesinstitutions. This chapter deepens the focus ondeveloping-country governance by consideringfor each of these subsystems some options formonitoring and improving performance—andlinks these to different approaches to scaling upaid. (The developed-country, multilateral, andglobal dimensions of the governance challengeare considered in depth in chapters 3, 4, and 7,respectively.) Inevitably, countries reforminggovernance differ from one another in the paceat which these different subsystems improve. Afinal section of the chapter examines some ofthe dilemmas this poses for the country and forits development partners.

Monitoring and ImprovingBureaucratic CapabilityAn effective bureaucracy facilitates the scal-ing-up of aid. The bureaucracy formulatesdetailed policies that translate the goals ofsociety and its political leaders into programs

of action. It manages the implementation ofthese policies. And it reports on progress.

Helping to build bureaucratic capabilityhas long been a focus of development assis-tance. Before 2000 it was viewed as principallytechnocratic, with a gradual accumulation oflessons and advice on good practice. Even asthese lessons crystallized, the profile of thework remained low, because its focus onbuilding country systems was at odds with thedominant approaches to providing aid andtechnical support through self-standing pro-jects, hermetically sealed off from often dys-functional public sectors. But with the newapproach to aid placing increasing emphasison mutual accountability, the profile of effortsto build bureaucratic capability has risendramatically.

Better public finance and administration indeveloping countries is essential for the newapproach, introduced in December 1999 withthe Poverty Reduction Strategy (PRS) process.The PRS builds on a hard-learned lesson ofdevelopment experience—that externallyimposed conditionality generally fails toachieve its intended results (World Bank andIMF 2005: 1, 10). The national budget andthe public bureaucracy that prepares andimplements the budget and is accountable toits citizens are critical vehicles for ensuringcountry ownership and leadership (WorldBank and IMF 2005: 12, 15, 19).

G L O B A L M O N I T O R I N G R E P O R T 2 0 0 6 139

Monitoring and ImprovingGovernance Subsystems

Bureaucratic Capability, Front-Line Provision,Checks and Balances

6

Budget support is the natural way totransfer resources to support a country’s PRSobjectives without undermining countryownership through excessive external over-sight. But where governments focus less onpoverty reduction and participation, are lessconstrained to be accountable to their citi-zens, and have less capacity, the combinationof a PRS process and budget support doesnot offer a ready way of resolving the tensionbetween country ownership and donor fidu-ciary obligations.

Countries are making progress in develop-ing a long-term holistic vision for povertyreduction and translating that vision into acoherent, medium-term, sequenced strategy.But most have a long way to go. The PRSReview of 2005, reporting survey data cover-ing 59 countries, concluded that only 7 hadwell-developed strategic programs (WorldBank and IMF 2005; and World Bank 2005).The majority of the remaining countries hadactivity under way—though not yet advancedto the point that long-term visions could serveas a reference point for policy makers.

This section considers some aspects ofcountry progress in moving from generalassertions of development goals to the specificarticulation, costing, and implementation ofstrategies for poverty reduction. It firstreports on progress in monitoring the qualityof public expenditure management and high-lights some patterns across countries revealedby monitoring. It also highlights emerginglessons about how to strengthen publicexpenditure management systems in differentcountry settings. The section then reports onefforts to monitor public administration,drawing on experience to offer practical guid-ance on how to improve administrative capa-bility, both for developing countries on a pathof improving governance and for their devel-opment partners seeking to monitor progress.

Monitoring and Improving PublicFinancial Management

Public financial management is particularlyrelevant to the new aid architecture. It is keyfor getting results on the ground and for assur-ing donors that aid resources are being usedprudently. Setting the stage for the analysishere is a framework based on the 2005 reportof the multiagency Public Expenditure andFinancial Accountability (PEFA) partnershipprogram (figure 6.1). That report synthesizedthe results of more than a half-dozen years ofwork by PEFA partners to develop a commonplatform for assessing the quality of publicfinance systems, including those in aid-recipientcountries. The framework depicts four facets ofthe budget cycle:

� Policy-based budgeting—the formulatingprocess for translating public policies, in-cluding policies that emerge from a PRSprocess, into specific budgeted expenditures

� Arrangements for predictability, control,and stewardship in the use of public funds(for example, payroll and procurementsystems)

� Systems of accounting and recordkeepingto provide information for proper man-agement and accountability

C H A P T E R 6

140 G L O B A L M O N I T O R I N G R E P O R T 2 0 0 6

BudgetCredibility

ComprehensivenessTransparency

The budget cycle

Predictability& control in

budgetexecution

Externalscrutiny &

audit

Accounting,recording, &

reporting

Policy-based

budgeting

FIGURE 6.1 Public financial management: a performance monitoring framework

Source: www.pefa.org.

� External audit and other mechanismsthat ensure external scrutiny of the oper-ations of the executive (for example, byparliament)

Comprehensiveness of budget coverageand transparency of fiscal and budget infor-mation cut across these four facets. Theframework also identifies credibility—thatthe budget is realistic and implemented asintended—as a key intermediate outcome, aresult of the operation of the whole cycle.

There are many ways of measuring thequality of a country’s public financial man-agement system. As box 6.1 highlights, theInternational Monetary Fund (IMF) hasdeveloped some useful tools.1 This sectionfocuses on two measures at two different lev-els—an overall measure of the quality of pub-lic expenditure management, and measures ofspecific expenditure management subsystems.

An overall assessment. The results of Coun-try Policy and Institutional Assessment (CPIA)

criterion 13 (see box 5.2)—abbreviated hereas CPIA-budget—can be used to assess over-all patterns in the quality of budget manage-ment systems across countries. As a countryimproves its budget management system, itsCPIA-budget score moves from weakest (1) tostrongest (6). The scale is built from fourdimensions of budget management, whichbroadly correspond to the facets of the PEFAperformance management framework in fig-ure 6.1.2 A CPIA-budget score at or above 4is consistent with the “good enough gover-nance” 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 bya CPIA-budget score of 4 (figure 6.2). Thesehigher-performing countries are Azerbaijan,Benin, Burkina Faso, India, Indonesia, Mali,Serbia and Montenegro, Sri Lanka, Tanzania,and Uganda. Almost half the countries scoredat or below 3. Of the 10 better-performingcountries, Azerbaijan, Mali, and Tanzania

M O N I T O R I N G A N D I M P R O V I N G G O V E R N A N C E S U B S Y S T E M S


The Code of Good Practices on Fiscal Transparency was developed in response to concerns that alack of comprehensive information on fiscal activity made it difficult to properly assess the objec-tives of fiscal policy. Greater fiscal transparency was also believed to be linked to improved gover-nance and fiscal outcomes more generally. The code contains 37 good practices that are organizedaccording to four main principles of fiscal transparency: clarity of roles and responsibilities; publicavailability of information; open budget preparation, execution, and reporting; and assurances ofintegrity. These practices, when observed, are critical not only for holding leaders accountable, butalso for preventing any mishandling of finances during budget execution.

The IMF regularly undertakes assessments of fiscal transparency called fiscal Reports on Obser-vance of Standards and Codes (ROSCs) in its member countries. Participation in an ROSC is vol-untary and the authorities retain the right not to publish the final report, although most have agreedto publish fiscal ROSCs.a As of the end of 2005, fiscal ROSCs have been completed for 80 coun-tries, and 76 of these have been published. ROSC participation is distributed unevenly acrossregions, with most countries in Europe and the continental Western Hemisphere having completedROSCs, while a much smaller share of countries in Africa, the Middle East, and Asia have agreedto participate. A number of countries, especially in Europe, have been working on improving fiscaltransparency and have opted to undertake one or more ROSC updates to reflect this progress. Inaddition, a growing number of countries are undertaking full reassessments. Both reassessmentsand 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 http://www.imf.org/external/np/rosc/rosc.asp.

BOX 6.1 Two IMF tools to support fiscal management and transparency

raised their CPIA-budget scores by at leastone point between 2001 and 2004.

Disaggregated public financial manage-ment indicators. The heavily indebted poorcountries (HIPC) debt reduction initiativespurred a sustained effort to develop action-able indicators of budget quality. A first set of16 indicators was developed jointly by theWorld Bank and the IMF and applied in 2001in 23 HIPC through a joint assessment withrecipient-country governments, with a follow-up assessment in 2004.

Building directly on the HIPC trackingprocess, seven donors (the World Bank, theIMF, the European Commission, the U.K.Department for International Development(DFID), France, Norway, and Switzerland)plus the Strategic Partnership with Africaembarked on a joint PEFA program to sup-port “integrated and harmonized approachesto assessment and reform in the field of pub-lic expenditure, procurement, and financialaccountability.”3 In 2005 PEFA issued itspublic financial management performancemeasurement framework, including 28 high-level monitoring indicators. PEFA partici-pants have committed to harmonize their

assessments of the quality of the public man-agement systems of aid-recipient countriesaround the PEFA framework.

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

Control of procurement and payroll wasnot part of the 2001 HIPC tracking indicators.In practice, procurement and payroll—pluscash transfers—make up the overwhelmingmajority of public spending, so strong controlsin these areas are vital for good financial man-agement. Recent advances in monitoring thequality of procurement highlight some emerg-ing lessons (box 6.2).

Consider first the cross-country patternsfor policy-based budgeting. Done well, pol-icy-based budgeting can sharpen the focus onlonger-term priorities, enable phasing in shiftsin priority expenditures over time, and poten-tially reconcile capital costs and their recur-rent cost implications (if capital and recurrentbudgets are integrated). The HIPC trackingindicator reported in the third column oftable 6.1, labeled “policy-based budgeting,”focuses on medium-term projections. A scoreof A signals that medium-term projectionsexist and are integrated into the budget for-mulation cycle; a score of B that they exist butare not integrated; a score of C that they existfor only a few sectors or not at all. In 2004only 7 of the 25 countries tracked had inte-grated medium-term projections into theirbudget cycles, but 13 of the remaining coun-tries made projections (but did not integratethem into the cycle).

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

� The fourth column reports on measuresof whether the budget is comprehensive,with no significant extrabudgetary funds

C H A P T E R 6


Source: World Bank CPIA Database.Note: CPIA 13 measures the quality of budgetary and financial management. The 66 IDAcountries are divided into groups based on a CPIA 13 score.

0Average = 2.4

Low

16 countries

5

30

Number of countries scored

Average = 3Lower middle

Weaker Stronger

16 countries

Average = 3.5Middle

CPIA-budget score

24 countries

Average = 4.1High

10 countries10

15

20

25

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

(including unfunded contingent liabilities),and with donor funds also reported onbudget. Only Bolivia, Chad, Ethiopia, andGuyana can be said to have comprehensivebudgets in the sense that they met at leastthree of the four benchmarks. Seven coun-tries met no more than one benchmark.

� The fifth column reports on budget credi-bility, as measured by the closeness ofactual expenditure out-turns (both aggre-gate and sectorally) compared with theoriginal approved budget, and limits on

the extent of arrears. Six countries havefully credible budgets (meet both indica-tors), but 12 countries met neither of thecredibility benchmarks.

� The sixth column reports on whether coun-tries have a well functioning expenditureexecution system, including an internalaudit mechanism, and other in-budget-yearcontrols. Only 8 of the 25 countries metthree or more of the five budget executionindicators tracked in the HIPC process—and 9 countries met only one or none.



TABLE 6.1 Quality of budget management systems in 25 heavily indebted poor countries, 2004

Policy-based Budget Budget Budget External budgeting comprehensiveness credibility Execution budget scrutiny

(1 measurea: (4 measuresb: (2 measuresc: (5 measuresd: (2 measurese:CPIA-budget rating Country A–C ranking) # met) # met) # met) # met)

Best-performing Mali A 2 2 4 1group (both CPIA Tanzania A 2 2 4 1and HIPC) Burkina Faso A 2 2 3 1

Benin A 2 1 4 0Uganda A 1 0 3 2

Middling group 1 Guyana B 3 2 2 2Chad B 3 0 2 2Rwanda A 1 1 2 2Senegal B 2 2 3 0Ghana B 1 1 3 1Honduras B 2 1 2 1

Middling group 2 Cameroon C 1 1 3 1Ethiopia B 3 1 1 1Sierra Leone C 2 1 2 1Bolivia B 4 0 0 0Niger B 2 0 2 0Guinea B 2 0 2 1Malawi B 2 0 2 0Madagascar A 2 0 1 0Mozambique B 1 2 0 1Zambia B 0 0 1 2

Weaker-performing São Tomé and Principe C 2 0 1 1group Congo, Dem. Rep. of C 2 0 0 0

Gambia, The B 2 0 0 0Guinea-Bissau C 0 0 0 0

Source: IDA and IMF (2005). For details of the individual HIPC indicators, see www1.worldbank.org/publicsector/pe/FinalHIPCAAPGuidance2003-04.pdf.a. The measure is HIPC indicator 7.b. The measures are HIPC indicators 1, 2, 4, and 5.c. The measures are HIPC indicators 3 and 8.d. The measures are HIPC indicators 9–13.e. The measures are HIPC indicators 14–15.

Overall, seven countries (Benin, BurkinaFaso, Ethiopia, Honduras, Mali, Senegal, andTanzania) can be said to implement their bud-gets reasonably effectively, in the sense thatthey met the benchmarks for half or more ofthe criteria in each of columns 4–6. Anotherseven countries (Bolivia, Democratic Repub-lic of Congo, The Gambia, Guinea-Bissau,Mozambique, Uganda, and Zambia) metfewer than half of the benchmarks in at leasttwo of the three categories, and so appear tohave significant weaknesses in budget imple-mentation.

The seventh column of table 6.1 reports onthe quality of budget reporting and externalscrutiny. Adequate accounts are a precondi-tion for effective scrutiny. As of 2004, 14countries met one of the two benchmarks—closing annual accounts within two months ofthe end of the fiscal year. Formal oversight ofthe 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 countriessubmitted audited reports to its legislaturewithin 6 months of the end of the fiscal year—

C H A P T E R 6


More governments and advisers have recognized that the mechanisms for the government to pur-chase 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 domesticfirms. Spending on procurement is at the core of discretionary government spending; even minorimprovements in efficiency can yield substantial cost savings.

With greater visibility has come a profusion of efforts to monitor the quality of public procure-ment systems and enhance their functioning. Consider the Philippines, where in 2005 the govern-ment initiated work to measure public procurement performance in 10 of its largest agencies. Thework complemented other procurement monitoring efforts in the country, which included the appli-cation of the Organisation for Economic Co-operation and Development–Development AssistanceCommittee (OECD-DAC) Baseline Indicator Set for Procurement (BIS) tool; the observance of pro-curement proceedings by civil society representatives; and the publishing of information on the awardof 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 perfor-mance is beginning to emerge. The simplest form is physical observation of procurement practicesand outcomes.

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

A third form of procurement monitoring is assessing performance of public procurement sys-tems using defined performance indicators. Work on developing tools suitable for monitoring pub-lic procurement has been undertaken jointly by donors and partner countries over a two-year timeperiod. 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 onImproving 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 www.oecd.org/dataoecd/12/14/34336126.pdf.

BOX 6.2 Recent advances in monitoring the quality of procurement

and only 7 countries submitted an auditwithin the benchmark of 12 months.

The results confirm that the quality of bud-get management systems of the 25 HIPC-tracked countries remains uneven. OnlyBurkina Faso, Mali, and Tanzania score inthe top half of possible (absolute) scores in allfive categories. Ethiopia, Ghana, Guyana,Honduras, and Rwanda avoid the bottomrung in all four categories. The remaining 17countries had budget systems with at leastone deep flaw.

This unevenness raises concern. The bud-get process is like a chain in the sense that itis only as strong as its weakest link. Evenwell-formulated budgets add modest value ifthere is little relation between the budget onpaper and the way money is actually spent.And the impact of a well-prioritized and well-executed formal budget is undercut if muchof the public spending is off budget.

Strengthening public financial manage-ment. Why is performance on public financialmanagement so uneven? In some countriespoor performance may be a consequence ofclientelism, extended civil conflict, and theevasion of formal rules and external scrutiny.Serious improvement is unlikely withoutchanges in a country’s political dynamics.This is more likely for one-third or so of thelow-income aid-recipient countries that haveconsistently been stuck in the fourth and fifthquintiles of all governance performance mea-sures, with no improvement over the past fiveyears, most of them with capabilities under-mined by conflict.

But many countries have shown the capac-ity for quite rapid improvement in their pub-lic financial management systems:

� The CPIA and HIPC tracking assessmentsreveal that many countries strengthenedtheir budget systems in just three years—some by significantly more than any plausi-ble margin of error. Of the 66 InternationalDevelopment Association (IDA) recipientcountries included here, 19 improved theirCPIA-budget score between 2001 and2004—7 of them by one or more points. A

comparison between the 2001 and 2004HIPC tracking assessments of 25 countriesidentifies 6 (Cameroon, Ghana, Mali, Niger,Senegal, and Tanzania) that improved theirscores in a net of at least three categories.

� Even in a brief three-year time span, therewere some countries that made substantialimprovements for each of the five budgetsubsystems. For budget execution Senegalwent from meeting none of the benchmarksin 2001 to meeting three in 2004, Ghanafrom one to three. Cameroon improved itsscore on both “external scrutiny” bench-marks (though in 2004 it still took more thantwo months to close its annual accounts).Guinea’s score on policy-based budgetingwent from C to A. And Bolivia and Guyanaincreased by two the number of “budgetcomprehensiveness” benchmarks met.

� As the sustained improvements in Ghana,Mali, Senegal, and Tanzania suggest,countries with stronger starting capacity(measured, say, by having more bench-marks met in 2001) may be better able toachieve rapid gains in the short run thancountries with weaker starting points (seeDorotinsky, Kisunko, and Pradhan 2005).But Niger—which improved its ranking ina net of five categories5—suggests that sig-nificant gains also are possible where thestarting point is weak.

These patterns suggest that heightenedattention to budget management and strongincentives for better performance can result inquite rapid gains. For countries determined toimprove their public financial managementsystems, achieving a “good enough” standardwithin, say, a 5- to 10-year period may be fea-sible. How budget reform is designed andimplemented will be key.

The HIPC tracking results suggest that afew countries—those with committed devel-opmental leadership, plus a preexisting base-line of capacity—appear able to adopt andrapidly implement a comprehensive programof budget reform, to the point that countrysystems can provide a robust platform forensuring effectiveness in the use of resources.



But where capacity is weaker, there is a needto set realistic goals for what is achievableand implement them in a realistic sequence.

A first lesson for strengthening public finan-cial management systems is that in most coun-tries, the approach should be incremental. Thereforms proposed for specific budget manage-ment subsystems have sometimes been veryambitious. Recent reviews by both the IMFand the World Bank have examined the expe-rience with medium-term expenditure frame-works (MTEFs). The IMF review capturesthe shared conclusion, namely that “develop-ing comprehensive medium-term expenditureframeworks can be effective when circum-stances and capacities permit. Otherwise, itcan be a great consumer of time and resourcesand might distract attention from the imme-diate needs for improving the annual budgetand budget execution processes.”

The IMF review also offers some usefulguidance in noting that

. . . the MTEF, as a feasible means ofimproving budgeting, requires the fol-lowing: reliable macroeconomic projec-tions, linked to fiscal targets in a stableeconomic environment; a satisfactorybudget classification and accurate andtimely accounting; technical capacity . . .and disciplined policy decision-making,[including] budgetary discipline . . . andpolitical discipline for fiscal manage-ment. Before introducing an MTEF, oneshould raise a question: is the countryready for such an exercise in the sense ofhaving adequate support for the abovepreconditions? When this support wasnot adequate in a number of Africancountries, the MTEF was introducedprematurely, and is turning out to bemerely a paper exercise. (IMF 2005)

Efforts to install computerized financialmanagement information systems (FMIS)also are often overambitious and invariablyencounter significant delays. Reviews by boththe World Bank and the IMF of efforts toinstall FMIS in African countries concluded

that large and therefore more spectacularprojects are often preferred because they canbe easily communicated as evidence of polit-ical action, but they are more volatile andsubject to greater likelihood of failure thanare smaller, more focused interventions (IMF2005; Heidenhof and others 2002).

A useful guide to sequencing public finan-cial management reform in low-capacity set-tings is suggested by DFID. The new“platform approach” for Cambodia involvesa cumulative sequence of budget reforms thatfocuses each round on achieving specific bud-get functionalities, building on these func-tionalities in the subsequent round (DFID2005). The sequence that emerges is almostthe reverse of that often associated with PRSimplementation (figure 6.3). Efforts to imple-ment PRSs have focused on their costing andtranslation into medium-term budget frame-works and on strengthening countries’ statis-tical capacities to monitor results.6 Bycontrast, the Cambodian platform sequencefocuses first on the basics: budget credibility,then predictability and control in budgetexecution. (Achieving these basics dependsalso on achieving predictability in the year-to-year flows of aid, a serious problem asnoted in part 1.) Only after these first twoplatforms are locked in will they move on tomedium-term budget planning—and onlyonce that is in place will they foster publicmanagement reforms to support a resultsculture throughout the public bureaucracy.Country leadership has been an importantfeature of Cambodia’s public financial man-agement program, ensuring that the designof reforms reflects domestic priorities, ratherthan those of donors.

A second emerging lesson for budget man-agement is to complement the technocraticreforms with greater transparency. Althoughthe PRS approach highlights inclusiveness, itsimplicit route to effectiveness tends to betechnocratic: design a robust poverty-reduc-ing budget, execute it effectively, monitorresults, and recalibrate policy and budgetingon the basis of what is learned. The lessonemerging from experience is that, in develop-

C H A P T E R 6


ing countries with weaker capacity, thisprocess may be better viewed as a long-runtrajectory than as a feasible path to betterresults in the short to medium term. Thatexplains interest in a more demand-sideapproach, complementing the technocraticroute: along with participatory priority-set-ting in PRSs, foster transparency in budgetmanagement—and emphasize the potentialof public information to improve the devel-opmental discourse among citizens, their gov-ernments, and development partners.

Monitoring and ImprovingAdministrative Quality

Getting development results depends onmuch more than good financial management.

For any organization, public or private, deliv-ery depends on the quality not only of thefinancial side of its balance sheet, but also ofits real side—the quality of its people, andhow effectively they are deployed and led.

As the framework in chapter 5 highlighted,public administration comprises both down-stream service provision and regulatory agen-cies (schools and ministries of education,customs agencies, roads authorities, and thelike) and upstream cross-cutting controlagencies within the bureaucracy (pay, humanresource, and performance management con-trol agencies, for example). Public adminis-tration reforms generally combine a focus onimproving upstream systems—to have a broadimpact across multiple systems—with targetedefforts to improve the performance of specific,



FIGURE 6.3 Cambodia: A platform approach to budget management reforms

Enables more accountability for performance management

Enables focus on what is done with money

Enables a basis for accountability

Platform 1A credible budget delivering a reliable and predictable resource to budget managers

Broad Activities

• Integration of budget (recurrent & capital budgets)

• Strengthen macro and revenue forecasting

• Streamline spending processes

Platform 2Improved internal control to hold managers accountable

Broad Activities

• Redesign budgeting classification system • Initial design of FMIS for core business processes • Define internal audit function

Platform 3Improved linkage of priorities and service targets to budget planning and implementation

Broad Activities

• Redesign budget cycle (e.g. MTEF)

• Pilot program- based budgeting & budget analysis

• Further fiscal decentralization

Broad Activities

• Full design of FMIS

• Develop IT management strategy

• Initial design of asset register

Platform 4Integration of accountability and review processes for both finance and performance management

Source: DFID 2005.

high priority agencies. This section focuseson upstream reforms; later sections considersector-specific approaches.

In the 1980s and early 1990s a first gener-ation of administrative reform focused prin-cipally on scaling back the bloated apparatusof government. In the late 1990s attentionshifted toward improving administrativecapability. Some consensus has been gener-ated on the characteristics of an effective pub-lic administration. As the CPIA subcategoriesused to score the “quality of public adminis-tration” suggest, the standard prescriptiontypically includes the following:

� Well-functioning mechanisms for policycoordination, which ensure policy consis-tency across departmental boundaries andfacilitate clear decisions on policy andspending priorities. To be effective, thesecoordinating mechanisms need to be at theapex of government, supported by toppolitical leadership.

� Well-designed administrative structuresfor individual line ministries and semi-autonomous executive agencies, with littleduplication of responsibility, and withclear lines of authority—plus streamlinedbusiness processes and a focus on results.

� Human resource management under-pinned by the principle of meritocracy—including for recruitment, promotion, andmajor disciplinary actions. This includesinsulation from undue political or per-sonal interests, as well as practices thatreward good performance (for example,through career advancement and financialrewards) and penalize poor performance.

� Pay and benefits adequate to attract andretain competent staff, including at seniorand technical levels.

� Establishment and wage bill control suffi-ciently robust to ensure that the public sec-tor wage bill is sustainable under overallfiscal constraints.

Monitoring administrative capability. Aswith CPIA-budget, the 1–6 scale of CPIA-

admin describes the gradations for a countryto move through as it works to improve thequality of its public administration.

The track record of efforts to close the gapsbetween the desired and actual quality of pub-lic administration is (to put it gently) unevenin both developed and developing countries.A landmark review of public administrativereform in 10 OECD countries—includingsuch noted public management reformers asAustralia, New Zealand, Sweden, the UnitedStates, and the United Kingdom—concluded:

Reform-watching in public manage-ment can be a sobering pastime. Thegaps between rhetoric and actions . . .are frequently so wide as to provokeskepticism. The pace of underlying,embedded achievement tends to bemuch slower than the helter-skelter cas-cade of new announcements and initia-tives. Incremental analysis and partisanmutual adjustment seem to have beenvery frequent features of public man-agement reform, even if more-than-incremental changes were frequentlyhoped for. (Pollitt and Bouckaert 2000:184, 188–89)

The CPIA-admin scores provide a snap-shot of administrative system performanceand reform for the 66 IDA countries. Theability to measure is less well developed foradministrative quality than for budget man-agement. No disaggregated actionable mea-sures paralleling the HIPC tracking andPEFA indicators are available—although aninitiative to fill the gap is at an early stage ofpiloting (box 6.3). The 2004 CPIA-adminresults and a composite Kaufmann-Kraay(KK)-style measure of administrative qual-ity produce a correlation coefficient of only0.56—a reminder of the large margins oferror all in governance measures.7

Public administrative systems are weakerthan their budget management counterparts(figure 6.4). Of the 66 IDA countries, only 2score 4 or higher on CPIA-admin (versus 10

C H A P T E R 6




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

BOX 6.3 Actionable indicators on public administrative quality

Civil service management actionable indicators

Objective Indicator

Merit-based civil service (CS) managementCompetition in recruitment and selection Percentage of CS vacancies filled through

advertised, competitive proceduresTurnover unrelated to changes in political leadership Quarterly CS turnover rates plotted against changes

in political leadershipEffective performance evaluation practices Percentage of CS staff for whom annual

performance evaluations were completedPercentage of CS performance evaluations falling in each rating category

Attracting and retaining qualified staffCompetitive remuneration Average CS total remuneration as a percentage of

average economic sector wagesRatios of average CS to private sector total remuneration by title

Vertical decompression Ratio of average Secretary General total remuneration to average Junior Officer total remuneration

Attracting qualified staff Average number of qualified (long-listed) candidates per advertised CS opening

Continuously weeding out poor performing staff Percentage of civil servants receiving the lowest performance rating in two successive years who have left the CS within the following year

Fiscally sustainable wage billBudget-financed wage bill is fiscally sustainable Actual budget-financed overall wage bill as a

percentage of GDP

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

Source: World Bank.

on CPIA-budget), and only 17 score 3.5 (ver-sus 24 on CPIA-budget). Trends in CPIA-admin suggest that, though change generallycomes slowly, committed countries canachieve quite rapid improvement in their sys-tems of public administration: Between 2001and 2004 Armenia, Azerbaijan, Cameroon,Georgia, and Vietnam lifted their CPIA-admin scores by one or more points, morethan any plausible margin of error.

A comparison of the results among low-income aid-recipient countries for CPIA-adminand CPIA-budget—and the relation betweeneach and the corruption and policy-quality out-comes—again suggests some unevenness acrossgovernance subsystems. While the overall cor-relation between CPIA-budget and CPIA-admin is quite high (0.73), the quality of budgetmanagement and of public administration canvary greatly from one country to another.

The correlation between budget systems andcontrol over corruption is low at 0.46. Thisresult is not as surprising as it may appear atfirst—corruption is an outcome of the qualityof national governance systems as a whole, notsimply budget management (chapter 5) and canbe unrelated to public expenditure manage-

ment. Even so, the result highlights an acutedilemma for approaches to aid that give specialprominence to improving budget systems tomonitor the use of donor resources. The strongfocus on strengthening budget managementmay help in underpinning good resource allo-cation and related policies—but not prove apanacea in the fight against corruption. Greaterclarity is needed in the global dialogue on gov-ernance, corruption, and development impactas to what is achievable—and how it can real-istically be achieved.

Strengthening administrative capability.Building effective public administrative sys-tems in developing countries is difficult. A1999 review of 102 World Bank operations tosupport civil service reform (CSR) between1987 and 1997 found that only 33 percent ofclosed CSR interventions and 38 percent ofongoing efforts achieved satisfactory out-comes.8 Useful lessons are emerging as to boththe reasons for the disappointing track recordof efforts to improve administrative systems,and constructive options for proceeding.9

Much of the administrative reform agendaaims to improve processes, and process reformstend to be soft, with progress difficult toobserve or measure. Even when these reformswork, their impact is evident only over the longterm. From the start, though, they threaten theauthority of established interests throughoutthe bureaucracy. Resistance to reform withinthe bureaucracy—either overt, or through half-hearted implementation—is therefore likely tobe endemic.

Then there is the political logic of reform.Political leaders need to balance a techno-cratic view of good reform practice with thepolitical imperatives of building and sustain-ing alliances with powerful patrons, avoidingconflict with powerful social groups, andmaintaining electoral support. Such a calcu-lus is not favorable for serious administrativereform: the upfront political costs are sub-stantial, and the time horizon long beforebenefits are evident in the form of improvedpublic performance. It is, however, muchmore favorable for more cynical politicianswith a short time horizon to promise bold

C H A P T E R 6


2225

17

2

Average = 2.4Low

Average = 3.5Middle

Average = 3Lower middle

Average = 4High

Weaker Stronger

CPIA-admin score

0

5

30

Number of countries scored

10

15

20

25

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

Source: World Bank CPIA Database.Note: CPIA 15 measures the quality of public administration.

reforms to clean up government and get gov-ernment working, in full knowledge that theseriousness—or otherwise—of the reformeffort will be invisible to the average citizen.

Administrative reforms therefore need toseek a good fit—one that aligns the agendawith a country’s political realities on theground. Rapid, comprehensive administra-tive reform is appropriate only in those rarecases where there is a strong enough baselineof capacity for sustained administrativereform—plus political leadership with thecommitment, mandate, and time horizonneeded to see the effort through.

Latvia and Tanzania are two countrieswhere the environment for administrativereform was propitious. Between 2000 and2003, Latvia promulgated an ambitious, andgenerally well-regarded agenda of administra-tive reform including a new civil service lawwhich guaranteed meritocratic recruitment,and introduced performance appraisal; a newcontrol framework for the large number ofsemiautonomous state agencies; and a newframework for coordinating policy makingand administrative reform from the PrimeMinister’s Office. It also made ongoing effortsto reform the public sector salary structure. InTanzania, the momentum of administrativereform built incrementally, sustaining a con-sensus as the program unfolded. An eight-year(1991–9) civil service program first broughtemployment and the wage bill under control,and then clarified the appropriate roles—andrightsized—across a wide range of govern-ment ministries, departments, and agencies. In2000 a new phase—an ambitious 11-yearprogram—began. The program incorporatesboth a phased approach to pay reform, and aperformance improvement model that givesindividual agencies incentives to clarify theirrole and mission, develop strategic plans(including well-defined results and a well-pub-licized service delivery charter), and identifyand address capacity development needs.

Even in these favorable environments,implementation has been quite challenging. InLatvia, the passage of reformist legislation pro-ceeded straightforwardly, but entrenched polit-

ical interests have made both pay reform andagency restructuring an uphill struggle. Tanza-nia, too, has had to scale back the ambition ofsome of its more far-reaching initiatives. Incountries with less favorable environments, theagenda of administrative reform needs to bemore modest. However, as box 6.4 illustratesfor Albania, even in these more difficult set-tings, carefully designed incremental reformscan achieve quite significant results.

Improving the Governance of Service Provision—SomeTargeted ApproachesIn most countries top-down reforms of cross-cutting public financial management andadministrative systems will take a long timebefore they help improve service provision.So it is natural to complement them withapproaches that work closer to the serviceprovision front-line. This section will high-light five service-centric approaches toimproving governance and service provision:

� Using public expenditure reviews to high-light sectoral spending priorities as part ofan integrated dialogue on strengtheningcountry systems

� Engaging via sectorwide programs� Using information to improve account-

ability 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 approachesthat are relatively more helpful in institution-ally stronger settings, working its way downto the difficult challenges posed by countrieswhere governance is weak. Some of theapproaches presented are relevant regardlessof whether a country’s governance is strongor weak.

Identifying sectoral spending priorities.Cross-cutting public management systemsaim to ensure that scarce public resources aretargeted toward activities with high social



returns, and are deployed efficiently. But pastallocation decisions may not work out asintended, and new opportunities continuallyarise. Chapter 1, for example, highlights thepotential for new, highly productive publicinvestments in key infrastructure areas. Thedevelopment returns can thus be high fromreviewing public expenditures to identify spe-cific expenditure with high potential returns,and ongoing, low-return expenditures thatcould usefully be redirected toward high-return uses.10 Where this process works well,the fiscal space opened up for new investmentor productive current expenditure can be large:

� Chile invested on average 5 percent of GDPin infrastructure during the second half ofthe 1990s without resorting to significantborrowing, primarily through reallocationof expenditure, increased efficiency, and theuse of public-private partnerships. Oneconsequence was that the country’s credit

rating increased, enlarging its unused bor-rowing capacity, and giving it greater fiscalflexibility for potential future use.

� Thailand initiated in 2005 a large five-yearpublic investment program of 2.5–5 per-cent of GDP annually to upgrade andimprove infrastructure, addressing widelyrecognized bottlenecks, including masstransit in Bangkok and the country’s inter-provincial highways. Credit rating agen-cies have assessed the investment programto be an important driver of growth overthe medium term—assessments that werebased on the country’s earlier fiscal credi-bility and enabled it to finance these invest-ments via borrowing. However, both theIMF and rating agencies have noted thateffective management of the investmentprogram will be needed to ensure contin-ued access to markets.

� The United Kingdom routinely incorpo-rates spending reviews into its budget

C H A P T E R 6


In the wake of communism, skepticism was pervasive among Albanians about the value of stateauthority 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 admin-istrative system made important gains between 1998 and 2005.

The gains came through the skillful exploitation, by both domestic reformers and their interna-tional champions, of a window of opportunity that opened between 1998 and 2002: the appoint-ment (by the political leaders of an electorally victorious political party) of a reformist primeminister willing to champion an administrative reform agenda. Backstopped by strong condition-ality from the World Bank, the agenda was carefully calibrated to be feasible in a setting with lim-ited commitment to reform. Albania’s administrative reform focused narrowly on introducingmeritocracy, plus market-competitive pay, for the country’s top 1,300 civil servants. Targeting onlythis top tier is not enough for systemwide improvements, but it can yield important gains in thequality of policy making and in the management of public resources. It can also establish a prece-dent 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 reformand toward Albanian politics as usual. Yet the reforms, which had been widely publicized andenjoyed both the support of donors and broad approval among Albania’s citizens, had crowded ina powerful constituency for their continuation—the senior civil servants. The arrangements for ameritocracy have largely been sustained. Indeed in 2005 parliament intervened directly to reject leg-islation that would have reduced the ability of the Department of Public Administration to enforcethe pro-meritocracy 1999 civil service law.

Source: World Bank.

BOX 6.4 Albania—administrative reform in an unpropitious environment

preparation process. Despite this, a 2003independent review of public sector effi-ciency identified over US$15 billion ofongoing spending which was not being effi-ciently used, and was directly “cash releas-ing” and so available for reallocation.

For all of the potential benefits of arrange-ments to review and adjust earlier decisionson resource allocation, putting them in placeis difficult—for at least three reasons. First, asthe previous section of this chapter hasdetailed, in many low-income countries eventhe basics of cross-cutting budget and admin-istrative systems do not work well. Second,even where the systems work well, they mightnot drill down in sufficient detail to distin-guish effectively between low- and high-return activities: the knowledge needed toassess development returns can be highly spe-cialized, and reside within sectors, not in bud-get central agencies. Third, many high-returninvestments cut across sectoral boundaries—as illustrated in chapter 2 by the high benefitsfor childhood health of upgrading wood-burning stoves or dirt floors.

While the returns are thus high fromstrengthening budget systems so they can pri-oritize more effectively, especially in low-income and weaker governance settings theneed to use resources well is too urgent to bedependent solely on systemic improvements.The case is compelling for complementingefforts at system improvements with more tar-geted efforts—within individual sectors andacross sectors—to identify high-return invest-ment opportunities, as well as opportunitiesfor freeing up resources locked into low-return activities. This is an activity for whichdevelopment partners can provide targetedassistance. The Public Expenditure Reviewsfacilitated by the World Bank, already anestablished part of the landscape of develop-ment dialogue, offer a ready-made vehicle.

Sectorwide programs. Partial approachesthat focus on improving governance and ser-vice provision in one sector have the potentialto achieve many goals simultaneously. Theseinclude getting quick wins in a high-priority

area; providing a mechanism for concentrat-ing limited country capacity; creating a focalpoint for harmonizing multiple, overlappingdonor programs around a coherent agenda;providing a clear focal point for results-basedmonitoring and evaluation; and serving ascatalysts for broader change in country sys-tems. Financing mechanisms can run fromsector-specific budget support (donors poolall their funding and channel it through thebudget using country procedures but care-fully monitoring flows to the preferred sector)to approaches that partly “enclave” the use offunds. To realize their potential, though, twoissues need to be confronted.

First, even at the level of an individual sec-tor, the challenges of aid harmonization andalignment remain formidable. Donors mustbe willing to subsume their particular priori-ties under the umbrella of a country-led sec-torwide program and to harmonize theirprocedures. This is proving difficult, even inTanzania, a global leader in incorporating aidinto country-led strategies and systems. Inthat country’s sectorwide program (SWAP) ineducation, for example, donors provide sup-port through basket funding, but have not allharmonized their reporting requirements. Asof October 2004, there were an estimated110 projects still on the books, with an aver-age size of only $906,000 (Economic andSocial Research Foundation 2005).

Second, focusing public managementreforms on a single sector risks making sys-temic reform more difficult later on. Roadfunds, for example, aim to strengthenaccountability by harnessing the incentives ofusers, who have a stake in the efficient andhonest use of resources, including stakehold-ers from the road transport industry, cham-bers of commerce, and farmers organizations.These have an extra incentive to provideoversight because some of the revenues usedfor road investment and maintenance comesfrom earmarked vehicle licenses and fuellevies.11 As another example, sectorwide pro-grams in education sometimes (notably inFrancophone Africa) have included a movetoward community schools, with increased



parental oversight. This has been shown toimprove both school management and educa-tional outcomes.12 In addition, locating respon-sibility for the contracting of teachers withcommunities can also help reduce unit costs:“even when offering average salaries as low ashalf the civil service teacher wage, countrieshave found more qualified applicants than theycan hire” (World Bank 2005: 82–85).

Despite their advantages, such sectorallyfocused initiatives can be at variance withstandard approaches to strengthening cross-cutting bureaucratic controls. Ring-fencingthrough road funds can undercut the abilityto make choices among competing uses ofresources, fragment the systems of budgetarycontrol, and complicate efforts to achievebroader improvements in the financial man-agement system. Community contracting ofteachers risks undercutting efforts to introducetransparent meritocratic practices of recruit-ment and promotion, and can also create newopportunities for informality and patronage.Each of these criticisms presumes that broadersystemic reforms are directly feasible. But inmany settings this is unlikely to be the case: thechallenge is to achieve gains in an imperfectworld, where the best can be the enemy of thegood. Further, partial reforms also have thepotential to nudge along incremental change inbroader systems: A well-managed road fundcould spur more far-reaching public financialmanagement reforms. Community teachersmight create an opening for more flexibleapproaches to civil service reform. Engagingcitizens in public sector governance withinindividual sectors can be a valuable spur tocivic engagement more broadly.

Using information to improve accountabil-ity at the service provision front-line. Openinformation on the performance of publicagencies can engage citizens in a continuum ofways. At one end is political accountability:citizens can use information on the quality ofservice provision as part of their decisionregarding the reelection of incumbents, atnational or local levels. At the other end is theuse of performance information by citizensdirectly involved in the governance of service

provision facilities, for example through com-munity-driven approaches discussed a littlelater. Two intermediate examples illustrate fur-ther the potential of empowerment throughinformation.

The first example highlights how detailed,public information can enable citizens to makeinformed judgments regarding the performanceof politicians, policy makers, and providers—and to respond with support, or pressure forchange. Frustrated by years of inaction on pub-lic services which increasingly were unable tokeep up with Bangalore’s dynamism and pop-ulation pressure, in 1994 a group of citizensintroduced the idea of a user survey–based“report card” on public services. Initially, theimpact was modest. Nonetheless, the spon-sors persisted, establishing a nongovernmen-tal organization (NGO), the Bangalore PublicAffairs Center, to institutionalize the effort,building coalitions with other NGOs andrepeating the report card survey in 1999 and2003. Figure 6.5 highlights the extraordinaryturnaround in perceptions of the quality ofservice delivery. The Public Affairs Centerdescribes how this was achieved:

The first and second report cards had putthe city’s public agencies under the scan-ner. The adverse publicity they received,according to many observers, acted as atrigger for corrective action. Inter-agencycomparisons seem to have acted as aproxy for competition. Citizen activismand dialogues with the agencies alsoincreased during this early period. Thesedevelopments prepared the ground for apositive response from the Government.A good example is the political leader-ship and vision displayed by the ChiefMinister S. M. Krishna in the past fouryears. He provided the frameworkwithin which a set of able administratorscould set in motion a series of actionsand reforms in the agencies. Many civilsociety groups and the media have stim-ulated and supported this momentum.Sustaining this movement is the chal-lenge for the future. (See Paul 2002: 71)

C H A P T E R 6


The service provision scorecard approachpioneered in Bangalore has been widely imple-mented—in countries ranging from Brazil tothe Philippines, Ukraine, and Turkey.

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

� In Rajasthan, India, building on the pas-sage of a Right to Information Act in thestate in 2001, the Movement for the Rightsof Peasants and Workers (MKSS) orga-nized public hearings in rural areas atwhich figures from the records of licenseddistributors of subsidized food rationswere compared with figures from theration books of recipients. Social auditswere also carried out of hospitals duringwhich data from medical records werecompared with patients’ actual experience.In both cases, large discrepancies betweenthe two sets of figures were revealed. Thisled to further investigation, which in turndisclosed evidence of corruption, embez-zlement, and maladministration.13

� A Philippines civil society organization(CSO), the Ateneo University Group, setup a citizen monitoring effort, togetherwith government agencies responsible for

textbook distribution and highway infra-structure, to make delivery more effective.The project determined that 21 percent oftextbooks were not actually delivered toschools designated to receive them, creatinglosses of more than US$3 million, which theDepartment of Education promised to rec-tify. The template developed for this projecthas been used by many other CSOs.

� In Tanzania, the Rural Initiatives andRelief Agency helped 10 local communi-ties track government program expendi-tures for health and education. The pilotprojects appear to have helped ensure thatcommitments to deliver funds were indeedfollowed through. The expenditure track-ing tool has been made available to CSOsin other rural areas of the country.

The latter two instances both were fundedby the Partnership for Transparency—aninternational NGO (supported by Sweden,the United Nations Development Programme[UNDP], and the World Bank) that providesmicro-grants to CSOs engaged in fightingcorruption. Independent evaluations haveshown the large majority of these projects tobe successful. The maximum grant size pro-vided by the Partnership for Transparency isUS$25,000—underscoring that empowerment



5 6 4 9

25

114

4147 42

67

34 34

16

32 32

73

94

73

92

73 7885

96

77

0

20

40

60

80

100

BMP BESCOM BWSSB BSNL Government hospitals

Police BDA BMTC RTO

Agencies

Percent satisfied

1994 1999 2003

n/a n/a

FIGURE 6.5 Perceptions 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 = MetropolitanTransport Corporation; RTO = Motor Vehicle Licensing

through information can be a low-cost, high-return strategy for improving governance.

Decentralization has increasingly been seenas a response to governance dysfunction. It hasa dual role in a national governance system.First, democratic subnational governments canoffer an important check and balance againstcentral executive power. Second, subnationalgovernments potentially have advantages inthe provision of some public services.

Decentralization often is driven by politics.Sierra Leone embarked on decentralization asa way of building simultaneously intergovern-mental institutions, local government capac-ity, and bottom-up accountability. When thatcountry’s civil war ended in 2002, the govern-ment initiated a process of national consulta-tion on decentralization. In February 2004, itenacted a progressive Local Government Act,establishing 19 local councils, which, over theperiod of 2004–8, will take over a large set ofresponsibilities and resources related to pri-mary education, primary health, agriculture,feeder roads, water, and sanitation. The WorldBank has supported fiscal decentralizationtechnically and has helped open politicalspace for development-oriented local politi-cians to emerge and establish track records bypromoting community-based approaches tolocal infrastructure investment.

Does decentralization help reduce poverty?For this to happen, two sets of accountabilitiesneed to work well. The first comprises down-ward accountability to local residents. As the2004 World Development Report (WDR) onimproving service provision to the poor put it:

Where decentralization is driven by adesire to move services administrativelycloser to the people . . . the assumptionis that [it] works by enhancing citizens’voice in a way that leads to improvedservices. . . . Voters make more use ofinformation about local public goods intheir voting decisions because suchinformation is easier to come by andoutcomes are more directly affected bylocal government actions. And political

agents have greater credibility becauseof proximity to the community and rep-utations developed through social inter-action over an extended period. But onboth theoretical and empirical groundsthis could go either way. The crucialquestion is always whether decentral-ization increases accountability relativeto its alternatives. If local governmentsare no more vulnerable to capture thanthe center is, decentralization is likelyto improve both efficiency and equity.(World Bank, World DevelopmentReport (WDR) 2004: 90)

The second set of accountabilities comprisesthe allocation of responsibilities between cen-tral and local governments. These include theassignment of responsibilities for service provi-sion (clarifying which services are assigned tolocal authorities, which are assigned to nationalauthorities, and which involve complementaryresponsibilities for both local and centralauthorities); the allocation of fiscal resources(including some tax base for local authorities)and fiscal accountability; and regulatory, fidu-ciary, and other forms of central oversight overlocal activity.

Clarifying these responsibilities in ways thatgive each tier of government an incentive toperform its role efficiently is a complex task anddeeply political. Decisions over the decentral-ization “rules of the game” involve a zero-sumcontest between national and local politiciansand bureaucracies over who controls resourcesand influence. The interplay between technicalcomplexity and political jockeying can some-times create difficulties. A comparative reviewof experience in six East Asian countries (Cam-bodia, China, Indonesia, Philippines, Thailand,and Vietnam) concluded that

The result [has been] a kind of “insti-tutional limbo”. . . . Whether by designor as a result of slippages in the imple-mentation process, intergovernmentalstructures have substantial internalinconsistency. The functions of different

C H A P T E R 6


levels of government overlap. Bottom-up accountability of locally elected bod-ies is dampened by top-down methodsfor appointing key officials. And thediscretion given to local authorities inspending unconditional fiscal transfersis effectively curtailed by central gov-ernment control over human resources.(White and Smoke 2005: 7)

As the 2004 WDR concluded:

Subnational authorities can be efficientproviders and regulators of local ser-vices under the right institutional incen-tives and with clarity about who doeswhat—and with what. But greaterautonomy can also increase opportunis-tic behavior and create moral hazard,resulting in costs that diminish account-ability and the benefits of decentraliza-tion. Good design, sound management,and constant adaptation by both centraland subnational authorities are neededto make decentralization work. (WorldBank, WDR 2004: 185)

Community-based approaches to localinfrastructure investments. In recent years,community-based approaches to local invest-ments have been pursued aggressively underthe rubric of community-driven development

(CDD)—an approach that “gives control overplanning decisions and investment support tocommunity groups and local governments.”14

It seeks to synthesize two types of developmentinterventions which historically have been con-sidered separately from one another—decen-tralization, as described above, and socialinvestment funds. The latter have been usedextensively by donors to transfer resources topoor communities in a participatory way.Between 1999 and 2005, the World Bankalone channeled over US$10 billion to poorcommunities. According to most reviews,these CDD operations have helped to get ser-vices to citizens more cost-effectively and equi-tably, and have supported participation andaccountability. Nonetheless, fierce debate sur-rounds CDD. Underlying this debate are con-trasting views regarding the likely interplay, inweaker governance settings, between bottom-up approaches, and efforts to strengthennational governance systems.

Certainly, the risks are large. As with manydonor-funded initiatives, early generationsocial funds bypassed the public administra-tion with the usual costs associated with par-allel implementation (see box 6.5). But inaddition, such programs offer a sometimesirresistible opportunity to political leaders. InPeru, for example, between 1994 and 2000,over US$900 million was allocated to thePeruvian Social Fund, FONCODES. The



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

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

poverty benefits were significant: 80 percentof the resources went to the poorest 40 per-cent of municipalities. Increasingly, however,it became apparent that FONCODES wasbeing used as a source of patronage and pop-ularity by the country’s populist president,Alberto Fujimori. Disbursements increased inthe months directly preceding elections, andwhile poorer areas were more likely to getfunding, those poorer areas that were “swingvoters” were favored in resource allocation(Schady 2000).

Practitioners of CDD have worked toaddress these risks by designing and imple-menting programs as part of a broader strat-egy of governance improvement—combiningscaled-up participatory resource transfers tocommunities and longer-run institutional

C H A P T E R 6


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 amajor financial crisis and political turmoil. Over three phases, close to US$1 billion has passedthrough 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 35million poor people.

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

Since 1998, Indonesia has progressively systematized its formal system of decentralization. Con-sequently, the second and third phases of KDP have emphasized greater oversight from district par-liaments, government monitoring, links with sectoral agencies such as education and health, districtmatching grants, and local involvement in drafting formal decentralization regulations on villageautonomy. The KDP platform has also provided lessons which are being incorporated into localgovernance reforms to support greater transparency and participation in district policies related toinformation disclosure, procurement, budget planning, and allocation, leading to higher pro-poorexpenditures.

Source: World Bank documents; Wong and Guggenheim 2005.

BOX 6.6 Linking community-based resource transfers anddecentralization in Indonesia

reform, by working closely with line min-istries and local governments to help buildtheir capabilities and interactions with com-munity groups. Advocates argue that, espe-cially in weaker governance settings, thishybrid approach can be a powerful way ofsupporting decentralization. Efforts alongthese lines are under way in many countries,ranging from Afghanistan to Albania, Brazil,Indonesia, the Kyrgyz Republic, Tanzania,and Zambia. Indonesia offers one example ofhow this integration can proceed (see box6.6), but in some other countries programshave failed to evolve much beyond parallelmechanisms. Instead, by seeking to break outof the comfort zone provided by parallel pro-jects, they have brought to the surface themany rivalries and unresolved tensions that

characterize countries stuck in a syndrome ofweak governance.

Surfacing the reality of the difficulty andunpredictability of change in weaker gover-nance settings need not, however, be a badthing. The challenge for CDD practitioners is tolearn more about how to improve the odds:What approaches make success in incremen-tally fostering sustainable institutional changemore likely, and in which country settings?When might the net benefits of a CDD inter-vention be positive, even with no success in cat-alyzing institutional change? Demand-drivenand incremental institutional reforms such asCDD tend to be judged against a standard ofperfection. Unsurprisingly, they fall short.What is needed is some agreement as to whatincremental improvement would look like—and a monitoring approach that systematicallytracks and assesses incremental, demand-driven institutional change.

Monitoring and ImprovingNational Checks and Balances InstitutionsStrong checks and balances institutions arekey to a well-functioning national governancesystem. Some of these checks and balances areglobal (including global financial and othermarkets) and are considered elsewhere in thisreport. The focus here is on national checksand balances institutions. Developmentalleadership or a dynamic political movementcan sometimes substitute for weak nationalchecks and balances, at least for a period. Butover the longer run, well-functioning checksand balances institutions are key to sustain-ability. They help keep the executive arm ofgovernment focused on the public purpose.They are vital for fighting corruption, forensuring that state actors at all levels use pub-lic resources efficiently and effectively, and forhelping to ensure that citizens perceive stateinstitutions to be legitimate.

Figure 6.6 disaggregates checks and bal-ances into a constellation, arranged in terms oftheir “distance” from the executive authoritythey oversee. The relationship of these institu-

tions with one another is only loosely hierar-chical. Depending on a country’s constitution,the judiciary may or may not be a constraint onlegislative authority. Citizens may ultimatelyelect governments but on a day-to-day basistheir role is more participatory than hierarchi-cal. We can distinguish three broad groups:

� An “outer constellation” of civic voice—the rules (for example on freedom of infor-mation) and actors (such as the media)that ensure the open operation of civilsociety—and the transparent flow of infor-mation and data that enables citizens toplay an informed role in public discourse.(Though not an explicit focus in thisreport, the discipline provided by compet-itive markets is an important buttress ofthis outer constellation.)

� A “middle constellation” of impartial dis-pute resolution—in particular the justicesystem

� An “inner constellation” of direct over-sight—subnational governments, autono-mous oversight agencies, and the legislature.

The next three subsections consider eachof these in turn, focusing on approaches tomonitor the quality of the relevant checks andbalances institutions and highlighting howsome can be strengthened.



Executive

Civil society - Media

Judiciary

Legislature

Subnational governments andautonomous oversight agencies

FIGURE 6.6 A constellation of checks and balances institutions

Source: Authors.

The Outer Constellation—Transparencyand Voice

Citizen engagement, underpinned by accessto high quality information, forms the outer-most, and possibly the most important, ele-ment of a national system of checks andbalances. Figure 6.7 depicts the “virtuous cir-cle of transparency” in a way that highlightsthe links between the provision of informa-tion and state responsiveness.

Information reveals the actions of policymakers; this facilitates evaluation and moni-toring, activism rises, and with it the level ofpublic debate. Policy becomes more contestableand citizens are motivated by the possibility ofholding the government accountable. Commu-nication with the government becomes a two-way flow, generating further demands for morereliable information. The virtuous circle is com-pleted as government practices become moreopen and more responsive to citizens.

Strengthening the virtuous circle. Severalfactors are needed for this virtuous circle towork well. First is the production and dissem-ination of good quality information. Recentinitiatives by international agencies, includingthe IMF and the World Bank, have begun toput in place a framework of internationally

accepted norms for the collection and publi-cation of economic and social data. Created in1996, the Special Data Dissemination Stan-dard (SDDS) is a voluntary standard whosesubscribers18—countries with market accessor seeking it—commit to meeting internation-ally accepted levels of data coverage, fre-quency, and timeliness. SDDS subscribers arerequired to maintain a Web site that containsthe actual data. For countries that do not havemarket access, the General Data Dissemina-tion System provides a detailed frameworkthat promotes the use of internationallyaccepted methodological principles, the adop-tion of rigorous compilation practices, andways in which the professionalism of nationalstatistical agencies can be enhanced.

These norms can serve as benchmarks forstatistical capacity building. There is, as yet,no agreed single measure of statistical capac-ity, but work carried out as part of the 14threplenishment of IDA’s resources (IDA 14)provides a basis for monitoring changes in theperformance of statistical systems. Data arecompiled annually on three key dimensions ofcapacity: statistical practice, data collection,and indicator availability. The indicators arecombined to generate overall indicators foreach dimension and to produce a single over-all indicator. This measure paints a worryingpicture of statistical capacity (figure 6.8). Asthe low average scores in figure 6.8 for “prac-tice” and “collection” signal, many IDAcountries lack the ability to provide basic sta-tistics on a regular basis, resulting in a viciouscircle—limited, poor-quality data reducedemand for data and lower interest in sup-porting data collection. Attempts to bridge theavailability gap include proxy data or simula-tions. Capacity has been increasing slowly, ifat all, in most poor countries. Especially in thepoorest, the impact of projects to strengthenstatistical capacity has often been disappoint-ing. Investments are usually not sustained,often because of the piecemeal, short-termnature of projects. The international commu-nity has responded to these weaknesses withthe Marrakech Action Plan for Statistics(MAPS). The objective of MAPS is to assist all

C H A P T E R 6


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

Accountability

Availability ofinformation

Analysis ofoutcomes and

policy proposals

Public advocacy Public debate

Source: Authors.

developing countries to either implement orprepare a longer-term national statisticaldevelopment plan by the end of 2006.

A second factor needed for the virtuouscircle of transparency to work well is disclo-sure—the critical step that turns informationinto a potent tool for civic accountability. Allgovernments routinely disclose reams ofinformation, including selective informationaimed at shaping public opinion. Most demo-cratic societies have some basic standards ofdisclosure—publication of judicial decisions,or the records of parliamentary debates, forexample. More recently, however, globalchanges in politics, technology, and valueshave converged to provide a powerful impe-tus to efforts to strengthen the transparencyof governance systems.

This global sea change is reflected in thegrowing number of countries that haveadopted Freedom of Information Laws15—over 50 as of the end of 2004, with effortsunder way in an additional 30. The trend isspreading worldwide: in Asia, nearly a dozencountries have either adopted laws or are onthe brink of doing so. In South and CentralAmerica and the Caribbean, half a dozen

countries have adopted laws and nearly adozen more are currently considering them.South Africa enacted a wide-reaching law in2001 and many countries in southern andcentral Africa, mostly members of the Com-monwealth, are following that country’s lead.

A third factor is an independent media.Independent media are a crucial pillar of goodgovernance, and a critical link in the account-ability chain between the government and thegoverned. Investigative journalists increasethe likelihood of detection of corruption, andpunitive action, thus fostering good gover-nance. Mass media also function as a channelof citizen voice, influencing government poli-cies and actions to be more relevant andresponsive to citizen preferences.16 As box 6.7underscores, a vibrant and good-quality mediacan be a potent development asset.

A fourth factor is an engaged civil society.Perseverance of civil society is crucial as away of ensuring that greater transparencytranslates into a change in the internal cul-tures of public institutions. Developmentpractice has responded to the new focus oncivil society—illustrated by the participatorynature of the PRS process. Box 6.8 outlines



Source: Country Statistical Information Database (www.worldbank.org/data/countrydata/csid.html). See IDA (2004) for methodology.

0

10

20

30

40

50

60

70

80

90

100

IBRD IDA IBRD IDA IDA-AFRIDA-AFR IDA-AFRIBRD IDA

Practice Collection Availability

1999 Change between 1999–2005

Percent

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

for Rwanda and Vietnam how the PRShelped crowd civil society more systemati-cally into the policy discourse. This marked adeparture from the earlier practice of donorsand international financial institutions (IFIs)of focusing narrowly on the executive, andhas sought to engage more directly citizensand their elected representatives. A recentreview suggests progress on this front (table6.2) but shows that engagement is well devel-oped in only a minority of countries.

Monitoring transparency and voice. Onebroad and one specific set of indicators are usedin this subsection to monitor transparency and

voice (TV). The broad indicator captures theoverall TV environment—including the humanrights and political governance dimensions.The specific indicator focuses more directly onthose aspects of transparency most directly rel-evant for achieving the Millennium Develop-ment Goals (MDGs).

Two broad indicators were considered forthis report—the aggregate KK “voice andaccountability” indicator, and a related indica-tor that focuses more narrowly on transparency.Table 6.3 reports the better-established “voiceand accountability” indicator—noting also thatat 0.88 the correlation between the voice and

C H A P T E R 6


A variety of studies have documented the link between better-informed citizens and better-performinggovernments. Besley and Burgess (2002) show that state governments in India were far more respon-sive to food crises in those states that had high newspaper circulation than in those that did not.Adserá, Boix, and Payne (2003) find, similarly, that corruption is significantly lower in countries withhigh 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 thatlacked them.17

BOX 6.7 How media access can influence development outcomes

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

Vietnam produces a Socio-Economic Development Plan (SEDP) every five years. The SEDP typ-ically has been prepared by central government agencies with little consultation outside the com-munist party. Subsequent to the finalization of the 2001–5 SEDP, the government embarked ondeveloping a PRS—the Comprehensive Poverty Reduction and Growth Strategy (CPRGS)—together with local experts and researchers as well as international and local CSOs. The existenceof parallel processes and strategies has caused some confusion on the reference point for policy mak-ers, 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 participatoryapproach to planning that characterized the preparation of the CPRGS.

BOX 6.8 The Poverty Reduction Strategy process in Rwanda and Vietnam

accountability and the transparency results ishigh.18 The indicator is estimated from 19 sep-arate disaggregated sources—each of whichfocuses on a specific aspect of TVA. KK notethat these include “a number of indicatorsmeasuring various aspects of the politicalprocess, civil liberties and political rights.These indicators measure the extent to whichcitizens are able to participate in the selectionof governments. We also include indicatorsmeasuring the independence of the media,which serves an important role in monitoringthose in authority and holding them account-able for their actions” (see Kaufmann, Kraay,and Mastruzzi 2005).

Table 6.3 reports the distribution of the sam-ple of 66 IDA-eligible countries across threegroups, distinguishing among countries accord-ing to whether one can be at least 95 percentconfident, using a two-tailed test, that givenmeasurement errors they indeed fall into the cat-egory in which they are located. As with all gov-ernance measures, the indicator provides someuseful benchmarking, but only for a subset ofcountries is it possible to assert with confidence

that their environment for voice and account-ability is relatively strong or relatively weak.

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

The specific indicators are taken from theGlobal Integrity Index (GII). As box 6.9describes, the GII is an example of “goodpractice” methodology for governance indi-cators. As explained earlier, as with all gover-nance indicators, the estimates have somemargin of error. But because each measure isspecifically defined, it provides “actionable”information for governance reform. Thespecific GII indicators cover the range of the



TABLE 6.2 Participation in the PRS, 2005

CIVIL SOCIETY PARTICIPATION

Little action Action under way Well developed

PARL

IAM

ENTA

RY IN

VOLV

EMEN

T

Well developed Bhutan, Djibouti, Lao PDR, Tajikistan

Bosnia and Herzegovina,Burkina Faso, Cambodia,Guinea, Honduras,Madagascar, Mali, Mauritania, Moldova,Mozambique, Timor-Leste

Ghana, Rwanda, Uganda

Action under way Azerbaijan, Benin, Rep. of Congo, Pakistan

Armenia, Burundi, Cape Verde,Ethiopia, Kyrgyz Republic,Liberia, Mongolia, Niger, Serbiaand Montenegro, Yemen,Zambia

Tanzania, Vietnam

Little action Central African Republic, Dem. Rep. of Congo, Côted’Ivoire, Dominica, Guinea-Bissau, Nepal, São Tomé andPrincipe, Sri Lanka, Sudan

Albania, Bangladesh, Bolivia,Chad, The Gambia, Georgia,Guyana, Haiti, Kenya, Lesotho,Malawi, Nicaragua, Senegal,Sierra Leone

Cameroon

Source: World Bank 2005b.

checks and balances constellation, thoughso far country coverage remains limited.Table 6.4 reports the scores for three specificGII subindicators, which measure facets ofthe environment for transparency and civicparticipation for 25 OECD, middle-, andlow-income countries. Low-income coun-tries lag, especially in the right of access toinformation.

The Middle Constellation—Justice andthe Rule of Law

Justice sector reform and promoting the ruleof law have emerged as key goals of develop-

ment policy. The justice sector covers a vastarray of institutions, issues, and functions. Inthe broadest terms, it can be defined as theinstitutions and processes by which laws aredevised and enforced. It includes legal servicesand their providers (for instance, lawyers andparalegals), police, prosecutors, the judiciary,courts and their officials, other institutionsthat resolve disputes, and institutions that exe-cute judgments. The justice sector fulfills twodistinct, but complementary, sets of essentialfunctions. It provides services to citizens, suchas safety and security, and resolving disputes.And it can help to constrain the arbitrary anddiscretionary use of state power.

C H A P T E R 6


TABLE 6.3 KK voice and accountability 2004, 66 low-income countries

In relevant third (with 95% certainty) In relevant third (with less than 95 % certainty)

Top third Benin, Ghana, India, Lesotho, Mali,Mongolia, São Tomé and Principe,Senegal, Serbia and Montenegro

In top half (with 95 % certainty):Albania, Bolivia, Bosnia and Herzegovina, Comoros, Guyana, Honduras,Madagascar, Mozambique, Nicaragua, Niger, Papua New Guinea, SolomonIslands

Middle third Armenia, Bangladesh, The Gambia, Guinea-Bissau, Indonesia, Malawi,Moldova, Nigeria, Sierra Leone,Uganda

Could be in bottom third: Azerbaijan, Cambodia, Congo,Djibouti, Nepal, Yemen

Could be in top third: Burkina, Faso, Georgia, Kenya, Sri Lanka, Tanzania, Zambia

Bottom third Dem. Rep. of Congo, Côte d’Ivoire,Eritrea, Haiti, Lao PDR, Pakistan,Sudan, Uzbekistan, Vietnam,Zimbabwe

In bottom half (with 95 % certainty): Angola, Bhutan, Burundi, Cameroon, Central African Republic, Chad,Ethiopia, Guinea, Kyrgyz Republic, Mauritania, Rwanda, Tajikistan, Togo

Source: Kaufmann, Kraay, and Mastruzzi 2005.

TABLE 6.4 Global Integrity Index—transparency and civic participation (by group)

Civil society organizations Access to information law Freedom of the media

OECD countries 95 87 91Middle-income countries 88 60 74Low-income countries 82 41 79

Source: www.globalintegrity.org.Notes: 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, República Bolivariana deVenezuela; 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 scoresfor the individual question.



The GII, developed by the Center for Public Integrity, focuses on measurement of “the existence andeffectiveness of mechanisms that prevent abuse of power and promote public integrity, and on the accessthat citizens have to their government.” The GII has a nested design—with answers to more than 290detailed questions providing the basis for estimating a variety of indicators at different levels of aggre-gation. This enables users to move from the more aggregated indicators to the most disaggregated, andthereby 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 rigor-ous, “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 increasethe number of countries covered to over 100 by the end of 2006. It remains to be seen, though, whetherthe GII will receive the sustainable financing and broad legitimacy necessary for it to become a widelyused part of the arsenal of governance indicators. The table below details the questions for 10 indica-tors that are most directly relevant to the dimensions of checks and balances highlighted for this report.

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

Some specific GII indicators

Indicators of transparency and civic participationCivil society organizations—In law, do citizens have a right to form CSOs? Do they in practice? Can citizensorganize into trade unions? In practice, do CSOs actively engage in public advocacy campaigns? Are civil societyactivists 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 ofaccess 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 wheninvestigating corruption?

Indicators of justice and the rule of lawJudiciary—In law, is the independence of the judiciary guaranteed? Is the appointment process for high courtjudges effective? Can members of the judiciary be held accountable for their actions? Can citizens access thejudicial system? In law, is there a program to protect witnesses in corruption cases? Are judges safe whenadjudicating corruption cases?Rule of law and access to justice—In practice, does the criminal justice process function according to the rule oflaw? In law, is there a general right of appeal? Are citizens protected from detention without trial? Are individualeconomic rights guaranteed?Law enforcement—Is the law enforcement agency (that is, the police) effective? Can law enforcement officials beheld accountable for their actions?

Indicators of direct oversightLegislature—Can members of the legislature be held accountable [by the judiciary] for their actions? In law, aremembers of the legislature subject to prosecution? Are there regulations governing conflict of interest bymembers of the legislature? Can citizens access the asset disclosure records of members of the legislature? Cancitizens access legislative processes and documents? Does the legislature have control of the budget? Cancitizens access the national budgetary process?National ombudsman—In law, is there a national ombudsman, public protector, or equivalent agency coveringthe 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 equivalentagency covering the entire public sector? Is the supreme audit institution effective? Can citizens access reportsof the supreme audit institution?Anticorruption agency—In law, is there an agency (or group of agencies) with a legal mandate to addresscorruption? 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 expectedto reflect certain basic qualities, most notablythe rule of law. Indeed, the term “rule of law”is sometimes seen as synonymous with, or usedas a proxy for, a well-functioning justice sector.Yet, as with the justice sector generally, there isno shortage of conceptions as to what the ruleof law is said to entail.19 Both the rule of lawand justice reform have been defined broadlywith reference to their essential role in ensuringdemocracy and human rights—or narrowlywith reference to their impact on predictabilityfor business processes and investment climate.Differences in priority and definition will havea direct impact on which reform efforts are pri-oritized to improve the functioning of the jus-tice sector and the rule of law and, in turn, whatshould be measured.

There already exist a number of broad andspecific indicators on justice and rule of lawissues. Most of them are not aimed specificallyat justice and the rule of law, however, andonly incorporate a section on it, as part of abroader focus or theme. Others, while focus-ing on justice and rule of law issues, focus onspecific processes or institutions and do notseek an overall view of the state of the rule oflaw. The two sets of broad indicators thatmake global comparisons among countriescomprise the KK “Rule of law” aggregate indi-cator, and the “Property rights and rule-basedgovernance” CPIA criterion. As examined inchapter 5, the KK Rule of Law indicator aggre-gates data from multiple sources, namely: “. . .several indicators which measure the extentto which agents have confidence in and abideby the rules of society. These include percep-tions of the incidence of crime, the effective-

ness and predictability of the judiciary, and theenforceability of contracts . . .” The CPIA-rulesindicator focuses primarily on the extent towhich the legal system facilitates private economic activity, but also looks at broad out-comes (safety), specific outcomes and func-tions (provision of business licenses, contractenforcement), and formal characteristics of thesystem. The correlation between the two indi-cators is quite high, at 0.83. Combining thetwo indicators for 66 low-income countriesyields 12 countries that are both in the topthird of the KK rankings for the indicator, andhave 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 spe-cific, actionable indicators. Three are note-worthy. The first two comprise the DoingBusiness and Investment Climate surveys. Asdiscussed in chapters 1 and 5, both includeindicators that can be used to monitor theperformance of the justice system relevant tospecific features of the business environment.(See this chapter’s annex for specific indica-tors relevant to the justice system.) The thirdcomprises the three GII justice and the rule oflaw subindicators identified in box 6.9. Table6.5 summarizes the subindicator scores for25 OECD, middle-, and low-income coun-tries. The results suggest that for many ofthem, improving the justice system is a largechallenge. This area has been prominent inthe development agenda only for a shortperiod, and much remains to be learned as towhat reforms work. Box 6.10 summarizessome emerging lessons.

C H A P T E R 6


TABLE 6.5 The quality of some attributes of the justice system in 25 countries (by group)

Judiciary Rule of law and access to justice Law enforcement

OECD countries 79 93 90Middle-income countries 71 74 63Low-income countries 58 72 59

Source: www.globalintegrity.org.Note: See table 6.4 for a list of countries in each category.

The Inner Constellation—Direct Oversight

Direct oversight institutions in the first twoinner rings of figure 6.6 include elected sub-national authorities, ombudsmen, supremeaudit institutions (with independent author-ity to review national accounts, monitor theprobity with which public resources are used,and report on their findings to parliament),anticorruption agencies (with independentauthority to investigate and sometimes alsoprosecute accusations of corruption), and thenational legislature, to which the executivegenerally is directly accountable. Though thissection focuses principally on monitoring,box 6.11 illustrates for one direct oversightinstitution—the legislature—some of thechallenges of improving performance.

The GII provides disaggregated measuresof the quality of direct oversight. As table6.6 summarizes, in most OECD and some ofthe middle-income countries restraints onthe executive are rated as high; Zimbabwestands out among the low-income countriesconsidered, as having few effective executiveconstraints.

One of the most widely used aggregateindicators is the “executive constraint” mea-sure of the POLITY data set.20 This measurerefers to “. . . the extent of institutionalizedconstraints on the decision-making powersof the executive. Such limitations may beimposed by any ‘accountability group.’” InWestern democracies these are usually legis-latures. Other kinds of accountability groupsare the ruling party in a one-party state,



The World Bank’s work in justice reform has largely focused on formal institutions, beginning withthe courts, and institutions supporting the market. More recently, an increased priority has beenplaced on access to justice, and attention has also focused on the link between formal and informalinstitutions, as well as the role that informal institutions play in helping to fulfill the roles of the jus-tice 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 eval-uated, and their capacity to deliberate and decide cases without undue influence from other branchesof government or other pressure. Accountability of judges, particularly to the public, was some-times not given as much emphasis as judicial independence. Yet citizens’ greatest complaints aboutcourts tend to focus first on corruption and second on delay, neither of which are likely to improvesubstantially without greater judicial accountability.

Lesson #2: Reforms that seek to overhaul the way justice systems operate will endure only if theydeliberately strengthen the management of the reform process. Profound institutional changerequires professional change management at the planning and implementation stages. This ispointed out quite clearly in studies of the criminal justice reform processes in many countries ofLatin America, in which attempts to replace written processes with oral hearings and trials sufferedbacksliding 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 jus-tice reform projects of major donors, focus on improving the administration of justice through for-mal institutions—courts, prosecutors’ offices, ministries of justice, and so on. Assistance has alsobeen provided to legal aid institutions both formal and informal, and major donors have supporteddevelopment 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 roleof informal, local justice systems—which in some countries govern as much as 95 percent of thepopulation—has only recently begun to receive more attention.

BOX 6.10 Strengthening justice—three initial lessons

C H A P T E R 6


TABLE 6.6 The quality of some direct oversight institutions in 25 countries (by group)

Legislature National ombudsman Supreme audit institutions Anticorruption agency

OECD countries 79 83 98 85Middle-income countries 66 81 92 68Low-income countries 73 73 78 77

Source: www.globalintegrity.org.Note: See table 6.4 for a list of countries in each category.

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

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

Legislative strengthening is best seen as a complement to related governance improvementsinvolving civil society. Civil society organizations are sources of technical expertise and can providespecialized legislative committees with information about the effects of public policies and policyalternatives:

Treating legislatures as self-contained entities that can be fixed by repairing internal mecha-nisms is unlikely to get very far. Rather, . . . it is more useful to think in terms of helping asociety 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 abil-ity to sustain their interventions, bring significant independent expertise to bear, and exercise effec-tive leverage in their oversight activities. These might include creating a permanent, independentnonpartisan staff for parliament; making changes in internal rules to permit tougher scrutiny of keyexecutives; establishing and strengthening specialized committees (including those focused on bud-get, education, health, roads, rural development, and cross-cutting themes—including povertyreduction); building links with CSOs and independent policy-advisory institutions; establishingwell-paid research capabilities to serve parliament; and “putting their own house in order” toimprove credibility, for example by establishing codes of conduct for members of parliaments, andby making campaign financing transparent, honest, and constrained.

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

BOX 6.11 Legislative oversight in Africa—a work in progress

councils of nobles or powerful advisors inmonarchies, the military in coup-prone poli-ties, and in many states a strong, indepen-dent judiciary.

POLITY IV scores on a 7-point scale. MostOECD and many middle-income countriesscore in the high and medium-high categories,but many countries that confront difficult gov-ernance issues also score medium-high on theindicator. A country’s political and bureau-cratic leadership can find itself constrainedeither as part of a well-functioning overall insti-tutional environment or as part of an overallsyndrome of state weakness. The final sectionof the chapter considers these issues further.

Sequencing GovernanceReformsThis final section brings together some of theindividual governance measures examined inthis chapter to pose a complex question—how to engage countries with an uneven mixof governance strengths and weaknesses?This is a somewhat different problem fromthe question of how to engage with countrieswith severe all-round governance weak-nesses, in part because the uneven mix mayreflect turnaround cases rather than stable,clientelistic equilibria.

Trajectories of change. Table 6.7 applies thegovernance indicators used in earlier sectionsto identify 28 countries that rate well in thequality of either their bureaucracies or theirchecks and balances institutions. While 10countries rate well in both areas, performanceacross the remaining countries is uneven. Tencountries (Rwanda and Vietnam, for example)have relatively capable public bureaucracies,but less strong checks and balances institu-tions. And the pattern is reversed in the other8 countries (Albania and Lesotho, for instance)where relatively stronger indicators for checksand balances are not matched by correspond-ingly capable public bureaucracies.

Why might patterns such as those in table6.7 be observed? Figure 6.9 illustrates threepossible trajectories for governance turn-arounds. These might vary depending on boththe initial political impetus within a countryand the longer-term historical processes thatcan shape and constrain political and institu-tional reform.

In trajectory 1 a developmentally orientedpolitical leader takes power in a hithertoclientelistic setting (as when President Rawl-ings took power in Ghana in the early 1980s,or President Museveni in Uganda in the mid-1980s). A common early focus of reformmight be to liberalize the economy and



TABLE 6.7 State capacity and state accountability

Quality of checks and balances institutions

Bureaucratic capability Medium or Low Higher

Higher 10 countries (Azerbaijan, Bhutan, 10 countries (Armenia, Benin, Bolivia, Burkina Faso, Ethiopia, Indonesia, Ghana, Honduras, India, Mali, Senegal, Pakistan, Rwanda, Tanzania, Serbia and Montenegro, Sri Lanka)Uganda, Vietnam)

Medium or low 38 countries 8 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 of3.5 and above. States with higher quality of checks and balances institutions are those that score “high” on at least two of the voice and account-ability, rule of law, and executive constraints broad checks and balances measures reviewed in earlier subsections of this chapter.

strengthen the performance of the public sec-tor. This can emphasize strengthening thecapabilities of the public bureaucracy—pub-lic administration and financial management,and the service provision frontline. Thestrengthening of checks and balances institu-tions can initially be a low priority, thoughcountries vary as to whether there is an initialweakening of checks and balances relative tothe status quo (as in Ghana) or a modestimprovement (as in Uganda). But once thereform process matures, the priority for gov-ernment reform might usefully shift fromstrengthening bureaucratic authority to en-hancing stability by increasing transparency,participation, and accountability of the state.This subsequent phase is, in practice, advancedin Ghana and more tentative in Uganda.

In trajectory 2 a turnaround is initiated bya move to political pluralism. Examples inAfrica include democratic transitions over thepast 15 years in countries as varied as Benin,Kenya, Malawi, Nigeria, and Zambia. Exam-ples in Europe and Central Asia include Alba-nia and Romania in the early 1990s. Theinitial political opening is only a first move inthe direction of stronger checks and balancesinstitutions. The dotted line signals a second

phase of governance reform in which themomentum for greater accountability con-tinues—and the reinvigorated legitimacythat comes from stronger participation andaccountability provides a platform for ongo-ing improvements in bureaucratic capability.Whether and how this subsequent phaseunfolds is, of course, an empirical matter.

In trajectory 3 turnaround starts from astate collapse. Sometimes external interven-tion helps to reintroduce the precondition foran effective state: a monopoly on the legiti-mate use of violence. This umbrella of securityprovides an opportunity for reestablishingboth the bureaucracy and checks and balancesinstitutions. Once a new base has been estab-lished, the process can continue in a balancedway, with momentum coming from the newlyreestablished domestic institutions. This pat-tern is evident in countries ranging fromBosnia and Herzegovina to Mozambique.

These varying trajectories pose some dilem-mas for the design and sequencing of gover-nance reform:

� Change that focuses first on improve-ments in bureaucratic quality has thepotential for rapid gains in public sector

C H A P T E R 6


high

Trajectory I Trajectory II Trajectory III

lowlow high low high low high

Quality ofbureaucracy

Quality of checks and balancesinstitutions

Desirable follow-throughInitial turnaround

FIGURE 6.9 Governance turnarounds: three trajectories

Source: Authors.

performance. But without a subsequenteffort to strengthen checks and balancesinstitutions, it risks subsequent reversal—perhaps by a reversion to corrupt behav-ior by the political leadership, perhaps bya loss of legitimacy with citizens.

� Change that begins with a political open-ing can generate a surge of confidence andimprove the climate for private invest-ment. But unless the gains are consoli-dated, the country risks becoming trappedin a cycle of what Thomas Carothers(2002) has called “feckless pluralism”—with short-lived governments repeatedlyvoted out of power, never having sufficientsupport and longevity to build the base ofbureaucratic capability on which effective-ness and legitimacy will eventually depend.

These varying trajectories also pose dilem-mas for a country’s development partners—both for scaling up aid and for ensuring thesustainability of development support.

Scaling up across different country settings.Consider first the dilemma uneven bureau-cratic capability poses for efforts to scale upaid. As table 6.7 summarizes, perhaps about20 low-income aid recipients currently havebudget management and administrative sys-tems reasonably capable of targeting spendingon poverty reduction priorities—and of exe-cuting and monitoring spending in a compre-hensive, credible, and transparent way. Witha few exceptions, World Bank budget supportvia Poverty Reduction Support Credits hasbeen targeted to these institutionally strongercountries, 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 coun-tries? Three possibilities are worthy of note.

First, even where current systems fall short,budget support might be scaled up for coun-tries based on a clearly improving trend in thequality of their budget and administrativemanagement systems. This is not simplybecause the additional resource transfers canbe poverty reducing: a shift from project aidto 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 onAid Effectiveness underscore, heavily frag-mented project aid complicates and disruptsnational systems, whereas budget support,combined with technical assistance, can facil-itate the improvement of these systems, par-ticularly if scaling up depends on continuingsystem improvements (see chapters 3 and 4).Tanzania illustrates this potential: it hasshown rapid improvement in budget manage-ment systems since 2001 and has been a ben-eficiary of progressively scaled-up budgetsupport over the period. The other exampleshighlighted in this chapter suggest that, forcountries determined to improve their admin-istrative budget systems, achieving a “goodenough” standard within, say, 5–10 years maybe feasible. Budget support might be initiatedquite early in the cycle of improvement—andscaled up as long as the carefully monitoredimprovement continues to be evident.

Second, priority could be given to reformsthat foster transparency—in budget manage-ment and more broadly. Transparency relieson public information as a source of pressurefor better public sector performance—in aless technocratic way than is implied by top-down reforms of bureaucratic capability. Tobe sure, the route from transparency to per-formance is circuitous, and the timing ofimpact, unpredictable. So far, no study defin-itively pinpoints the relationship betweentransparency and performance. But manyexamples, including some in this report, high-light the potential—from the tracking of edu-cation expenditures in Uganda, to servicedelivery report cards in Bangalore and Brazil,to the impact of media prevalence acrossIndia’s states. Even with continuing weaknessin bureaucratic capability, a case could thusbe made for scaling up aid (including somecomponent of budget support) to countriesthat clearly commit themselves to facilitatingtransparency in how public resources—andstate power more broadly—are used.

The third possibility for countries is totarget scaled-up aid more directly toward



poverty-reducing services, which can be donein several ways. A key distinction here isbetween countries where bureaucratic capa-bility may be on the upturn but is only at anearly stage of improvement—and thosewhere there is little sign of political commit-ment to improve governance and capacity. Inthe former group, the sectorwide approachesdescribed earlier that focus on improvinggovernance and service provision in part ofthe overall system are attractive. In the lattergroup, the focus might be on infrastructureand other service delivery investment pro-jects—complete with project implementationunits and related mechanisms that operateapart from country systems. Box 6.5 detailedsome well-founded objections to theseapproaches. But where there is little politicalcommitment to improve country systems andlittle sign that governments would have tar-geted pro-poor spending, these objectionshave less relevance.

Sustainability—bringing checks and bal-ances onto the agenda. In the short-term, aidcan thus straightforwardly be scaled up tocountries with improving budget and adminis-trative systems. But a longer-run challengelooms. While trajectories of improvementvary, and in the short run no one type of turn-around is superior to another, unless the gainsin the bureaucracy and the checks and bal-ances institutions eventually evolve in a bal-anced way, the risk is high that initialimprovements in governance will not be sus-tained. Over the medium term, it may there-fore become necessary to focus the governancedialogue on the complementary aspects of thebureaucratic and institutional agenda that arenot spontaneously coming to the fore. Howcan these sensitive issues best be addressed?

A first consideration is timing. In some set-tings it may not be practical to press very earlyin a turnaround process for far-reachingreforms of checks and balances. In Uganda, forexample, in the immediate aftermath of theAmin and latter-Obote years the state was intotal collapse, and the ability of the new gov-ernment to assert authority over the nationwas limited. Under such circumstances, it is

difficult to find fault with the readiness ofdonors to support government efforts to focusprincipally on strengthening bureaucraticcapability and development policy—and toemphasize decentralization as a means ofbringing government closer to the people. As iswell known, far-reaching reductions in Ugan-dan poverty resulted from the early actions ofgovernment, and donor support. At the sametime, it also seems clear that development part-ners can wait for too long—until it is too lateto put the challenge of strengthening checksand balances squarely on the agenda. Anexample here is President Suharto’s Indone-sia—where a failure to focus early enough onchecks and balances was associated with risingcorruption, financial crisis, and a difficultprocess of political succession which led tosome significant reversal of the developmentgains of earlier decades. Overall, the trackrecord of recent decades suggests that (perhapspartly as a consequence of the Cold War) inmany countries development partners mayhave waited too long before putting checksand balances institutions higher on the agendaof development dialogue.

A second consideration is that our knowl-edge of how to get “from here to there” is lessdeveloped than our understanding of whatwell-functioning checks and balances institu-tions should look like. One exception to thisproposition is the value of transparency,which is essential for the effective working ofall checks and balances and which can andshould be enhanced in almost all settings—atquite low cost. Donors and IFIs can play adirect role here, including by ensuring that allanalytical work is made widely available,with translation into local languages. Weknow less as to when and how improvementsin transparency translate into genuine gainsin accountability and performance, but itdoes seem to be at least a necessary condition.

Given the limitations of current knowl-edge, perhaps all that can be offered at thisstage is a modest process suggestion. Even—or perhaps especially—when it is still uncom-fortable, governments and their developmentpartners might usefully begin a dialogue on

C H A P T E R 6


how to strengthen checks and balances insti-tutions. The aim of this dialogue would be toagree on a phased sequence of steps forstrengthening these institutions, perhapsemphasizing more those directly relevant toMDG outcomes—transparency, gender, thejustice system, and local governance forexample. Recipient countries would beaccountable for proceeding with an agreedsequence. In return, they would enjoy morecertainty over what is expected by the inter-national community. Donors, in turn, havingagreed on a way forward would be expectednot to shift the goalposts after the fact.

Notes1. For details of, and results from, the Code of

Fiscal Transparency, see http://www.imf.org/external/np/fad/trans/index.htm and Hameed (2005).

2. The correspondence between the fourdimensions of CPIA-budget—(a) through (b)—and the PEFA PFM framework in figure 6.1 is asfollows: (a) corresponds to policy-based budgetingand the formulating process; (b) corresponds to acombination of the comprehensiveness of budgetcoverage, credibility that the budget is realistic andimplemented as intended, plus the budget execu-tion arrangements for the exercise of predictabil-ity, control, and stewardship in the use of publicfunds; (c) corresponds to the systems of account-ing and recordkeeping to provide the informationneeded for proper management, plus auditingmechanisms that ensure external scrutiny. Inter-governmental finance—the focus of (d)—is notdirectly incorporated in the PEFA framework. Forthe detailed scoring system used in the CPIA, seehttp://siteresources.worldbank.org/IDA/Resources/CPIA2004questionnaire.pdf.

3. See the PEFA Web site at http://www.pefa.org/index2.htm.

4. For some questions the benchmark was setat the score of B and for others at A. Furtherdetails, including the descriptions of how to scoreeach question, are available at http://www.pefa.org/about_test.htm.

5. Niger’s ranking improved in eight categoriesand declined in three. Five of the improvementswere sufficient to achieve the benchmark (but allthree declines were from benchmark level to below).

6. See, for example, the Africa Action Planrecently issued by the World Bank.

7. The composite measure comprises a subsetof the data used for the KK government effective-ness aggregate indicator; it excludes responses onthe quality of public service provision and on thecredibility of government’s commitment to poli-cies, and it excludes the CPIA-admin (because it isbeing used as a cross-check).

8. World Bank, Operations Evaluation Depart-ment (1999: ii–iii). The OED review highlightedfour specific weaknesses in Bank-supported inter-ventions: the poor quality of information on civilservice reform performance, needed for monitoringand evaluation; the limited role afforded to strate-gic management and cultural change; the absence ofchecks and balances on arbitrary action; and a fail-ure to appreciate key contextual contexts.

9. See the articles by Mike Stevens and StefanieTeggemann; Kithinji Kiragu, Rwekaza Mukandala,and Denyse Morin; Poul Engberg-Pedersen andBrian Levy in Levy and Kpundeh (2004).

10. For the detailed analysis on which this sub-section is based, see International Monetary Fundand the World Bank (for the Development Com-mittee), “Fiscal Policy for Growth and Develop-ment: An Interim Report,” April 2005.

11. For a detailed discussion of road sectorreform, see Heggie and Vickers (1998).

12. For a review of the role of communityschools in Francophone Africa, and the relevantlessons from international experience, see Gersh-berg and Winkler (2004).

13. For information on the MKSS, visithttp://www.freedominfo.org/case/mkss/mkss.htmor contact the organization at [email protected]. Press coverage of MKSS activities hasbeen extensive and includes Deccan Herald (Sep-tember 21, 2003) and Mail & Guardian Newspa-per, South Africa (February 20, 2004).

14. See http://www.worldbank.org/cdd.15. This review of the role of Freedom of Infor-

mation Laws is adapted from Bellver and Kauf-mann (2005).

16. For a pioneering, in-depth analysis, seeIslam (2002).

17. Adserà, Boix, and Payne (2003); Besley andBurgess (2002); Strömberg (2004).

18. Consideration was given to using three newmeasures of transparency produced by Kaufmannand a co-author, but it was decided to stick withthe better-known and more thoroughly scrutinized“voice and accountability” measure. The correla-tion between the aggregate voice and aggregatetransparency indicators is 0.88. As for the two



C H A P T E R 6


transparency subindicators, the correlationbetween “voice and accountability” and “politicaltransparency” is 0.93. The correlation is muchlower—0.41—with the measure of “economic andinstitutional transparency.” But the latter measuregenerates very large standard errors relative to theother KK indicator, raising questions as to thecoherence of the underlying concept it is intendedto 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 politicalwritings—and also appears as a cornerstone in thegenesis of various European legal and political sys-

tems. See, for example, Rigo and Gruss (1991:5–8) for an excellent overview of the origins of therule of law, in Greek and Roman thinking, as wellas its emergence as central tenets in the British andFrench legal orders. For a more extensive discus-sion, covering both Western and non-Western con-ceptions and origins of the Rule of Law, see Hager(2000: 3–20).

20. The POLITY project, (www.cidcm.umd.edu/inscr/polity) run from the University of Maryland, isthe world’s most widely used data resource formonitoring regime change and studying the effectsof regime authority. For details, see Marshall andJaggers (2002: 23–24).

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

Dealing with licensesThe number of procedures, average time spent during each procedure, and official cost of each procedure involvedin obtaining necessary licenses and permits, completing required notifications and inspections, and obtaining utilityconnections (using construction of a warehouse as a benchmark).

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

Trading across bordersNumber of documents, approvals, signatures, or stamps required, and the time and associated cost necessary to com-ply with all procedural requirements for exporting and importing a standardized cargo of goods.

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



(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 imposedby government regulations (such as taxes, customs, labor regulations, licensing, and registration), including dealingswith 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 oftheir 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 estab-lishment 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, includ-ing losses due to lost production time from the outage, time needed to reset machines, and production and sales lostdue 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 operationand 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 actualpayment, 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 contrac-tual 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 theirbusiness.

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 growthof their business.

ANNEX (continued)

monitoring and improving governance subsystemssiteresources.worldbank.org/intglobalmonitoring... ·...

Documents