ANNEX 3 Country Desk Reviews: Methodology and Summary Findings

International Monetary Fund. Independent Evaluation Office
Published Date:
April 2007
  • ShareShare
Show Summary Details

This annex presents evidence gathered during desk review work on 14 SSA countries with PRGF-supported programs. It also outlines the methodology used by the evaluation team to collect and analyze the information. The desk review work was designed to complement the results emerging from the quantitative analyses and surveys, which covered a broader sample of 29 SSA countries with PRGF-supported programs. It focused on reviews of PRGF program documents, supplemented at a later stage by staff interviews and six country visits (Table A3.1). The evidence emerging from the desk reviews was important in establishing working hypotheses for the evaluation and in testing emerging conclusions.

Table A3.1.Desk Review Country Sample




Country Policy and

Institutional Assessment

Quintile, 2004


(In millions)
Real GDP Per Capita,


(Constant 2000 U.S. dollars)
Burkina FasoXXU113.2246.5
Cape VerdeU10.51277.8
Central African RepublicXI54.0231.1
Congo, Democratic Republic ofXU557.686.8
Côte d’IvoireI518.2575.8
Gambia, TheI41.5322.7
São Tomé and PríncipeI40.2350.7
Sierra LeoneU45.5206.5
Sources: World Bank, World Development Indicators database; and World Bank (2004).

Methodological Background and Sources of Information

The criteria for selecting the 14 countries (out of the broader sample of 29 SSA countries with PRGF-supported programs) for in-depth desk review included economic and institutional performance, representativity, donor presence, and modalities for aid delivery (Table A3.1). Ten of the 14 countries had had long program engagement with the IMF, without serious interruption (Burkina Faso, the Democratic Republic of the Congo, Ethiopia, Ghana, Mozambique, Rwanda, Senegal, Tanzania, Uganda, and Zambia); the other 4 had experienced serious recent program interruptions (Cameroon, the Central African Republic, Guinea-Bissau, and Malawi).

To ensure consistency of treatment across country cases, a common set of templates was used to gather evidence from PRGF program documents. Individual templates covered the following subjects, with focus on identifying program objectives, use of program instruments, and the evolution of program implementation over time: (1) aid forecasting; (2) fiscal expenditure; (3) current account absorption; (4) stability considerations (inflation, domestic financing); (5) wage bill ceilings; (6) priority expenditures; and (7) domestic resource mobilization.

The evaluation team reviewed documents that are also available, in most cases, to the broader public on the IMF’s external website1—such as PRGF-supported program documents, Article IV surveillance reports, and Selected Issues papers. The evaluation team also had access to internal Fund documents—such as mission briefing papers and comments made during the internal review process. Reviewers focused on PRGF-supported programs, including of recent vintage, in order to examine the extent to which staff assessments, objectives, and program design itself have changed during program implementation. Sample program periods varied by country, while some reviews included two fully-fledged PRGF-supported programs.

Summary of Findings

The findings are organized along the lines of the main report: aid context, stance of macroeconomic policies, and social impact. These findings complement Chapter 2 of the main report.

Aid issues

Program documents revealed similarities in aid discussions with countries. The early PRGF-supported programs cautioned against indefinite aid dependence (Burkina Faso, Malawi, Mozambique, and Tanzania) linking it often to the need to improve domestic resource mobilization. Prudence in program aid was based on discussions with donors (Malawi, Mozambique, and Rwanda). There were general references to the need for higher aid flows to enable countries to achieve poverty and development goals (the Democratic Republic of the Congo, Ethiopia, Mozambique, and Senegal). Aid issues are discussed in Chapter 2, section on “Analysis of Aid.”

Aid predictability and its potential implications were frequent program themes (the Democratic Republic of the Congo, Ghana, Guinea-Bissau, Malawi, Mozambique, Senegal, Tanzania, and Uganda). However, links to aid forecasting were rare (Ghana, Malawi, and Mozambique), the tendency of discussions focusing on aid predictability or volatility for the current year of the program.

There was little transparency in how programs fore-casted aid. There was generally very little information on the methodology, key assumptions, and discount factors used to forecast aid. Explicit references to past aid forecasting errors figured in only half of the cases reviewed (Burkina Faso, the Democratic Republic of the Congo, Ghana, Malawi, Mozambique, Tanzania, and Zambia). Discussions of how these translated into current forecasts were not explicit or remained at a general level.

Current account issues

Current account absorption issues were addressed in connection with international reserves positions and Dutch disease. These issues are discussed in Chapter 2, section on “Accommodation of Aid.”

Discussions of the treatment of international reserves were prevalent in PRGF-supported programs (except for CFA franc zone countries). Document reviews showed that for many countries in the sample, programs had, at some point in time, targeted higher net international reserves (NIRs) in order to reduce vulnerabilities to external shocks—including terms of trade and aid volatility—but with differences in emphasis. For cases with low NIR positions (Cameroon, the Democratic Republic of the Congo, Ethiopia, Ghana, Malawi, and Zambia) programs underlined the need to raise international reserves. For countries where NIRs were at an appropriate level, the focus was on maintaining reserves at such levels (Tanzania and Uganda). In a few cases, reserve accumulation was considered to have been excessive (as in Rwanda) and programs dwelt on the issues of excessive reserves accumulation and insufficient aid absorption.

Dutch disease was not a major concern—although the exchange rate and issues of competitiveness were common themes in program discussions. Table A3.2 shows that there was some early program concern regarding exchange rate appreciation and possible Dutch disease, which gave way to concerns about aid underutilization in Tanzania; while in Rwanda Duch disease concerns persisted, albeit with reduced implications for the programmed absorption of aid. Program discussions on exchange rate issues relied on indicators of competitiveness and real exchange rates (Ghana, Mozambique, and Zambia), with competitiveness sometimes framed in terms of enhancing productivity, efficiency, and growth through structural reforms and infrastructure investment (Ethiopia and Zambia).

Table A3.2.Spending and Absorption Issues1
CaseMicroeconomic issuesMacroeconomic issuesDutch Disease
Burkina Faso (2003)General absorptive capacity concerns.Rather liberal stance throughout program.Not an issue.
Cameroon (2000)Weak spending capacity limited HIPC-related spending. FAD also expressed concerns over capacity to absorb large spending increase in investment at program request.Program aimed at consolidating fiscal adjustment achieved in previous program.No overvaluation of real effective exchange rate. Acknowledged that Cameroon maintained large competitiveness gains that resulted from the 1994 devaluation of the CFA franc.
Central African Republic (1998)Only micro issues are mentioned.Weak administrative capacity of government is named as risk to program but is not explicitly related to spending limits.
Democratic Republic of the Congo (2002)No specific discussion of limitations to spending aid.Focus on stabilization.
Ethiopia (2001)Program aims at limiting inflation to low single digits while rebuilding international reserves.Although authorities argued for weaker exchange rate, in face of increasing aid flows, staff noted that case was not compelling.
Ghana (1999, 2003)Absorption issues not a significant concern.Fiscal consolidation and containing domestic debt.No reference.
Guinea-Bissau (2000)Weak administrative capacity of government mentioned but not explicitly linked to spending limits.
Malawi (2000)Program allowed higher expenditure if foreign financing is available (PDR showed concerns over capacity to implement an expenditure increase in PRSP priority areas).Program aimed at fiscal consolidation throughout entire program period.No explicit concerns over appreciation pressures.
Mozambique (1999, 2004)Absorptive capacity limitations called for saving exceptional or peak aid flows (1999) and posed challenges for achieving Millennium Development Goals (2004).Overall program context in 1999 and 2004 one of maintaining macroeconomic stability and fiscal consolidation. By fourth review of 2004, broad program context highlighted better-than-programmed fiscal situation.No overvaluation of exchange rate (according to various measures and export volumes).
Rwanda (2002)Continuous concerns about expenditure transparency and allocation of resources toward military spending led to program that was inflexible regarding spending of unanticipated resources without prior discussion with IMF.Dutch disease concerns discussed in 2002 program; also in fourth review (2005), along with underabsorption concerns.
Senegal (1998, 2003)Capacity constraints in finance and spending ministries, in spite of ambitious spending plans.Fiscal consolidation program objectives.Not a concern at prevailing aid level, analysis needed of potential Dutch disease effects of higher aid (2003 program).
Tanzania (2000, 2003)Budget system inefficient—i.e., unable to absorb all aid resources available. Need to enhance fiscal transparency.Initial Dutch disease concerns expressed in 2000 program, but no longer a concern by 2003 program.
Uganda (1997, 2002)Limited expenditure efficiency—capacity and governance issues in social spending, notably universal primary education.Program objectives maintain low inflation (5 percent) and comfortable level of international reserves.During first two years of program, aid inflows (and high level of remittances) led to currency appreciation (1997). Liquidity injected into economy by donor-funded poverty reduction spending posed threat to price stability (2002).
Zambia (2004)Need to strengthen budgetary processes and public expenditure management.Centerpiece of policy framework is strong, front-loaded fiscal adjustment to halt unsustainable rise of domestic debt and interest payments, and increase poverty-reducing spending.Appreciation pressures not yet a concern.

Fiscal issues

The document review focused on issues of domestic financing of the fiscal deficit, domestic resource mobilization, the public sector wage bill, and fiscal governance. These issues are discussed in Chapter 2, sections on “Accommodation of Aid” and “Key Features Agenda.”

Domestic financing was a key program parameter, linked to macroeconomic stability and private sector crowding in or crowding out. Most PRGFs limited domestic financing of the fiscal deficit. The size of the fiscal deficit or domestic financing was typically used as a performance criterion (Ethiopia, Ghana, Guinea-Bissau, Malawi, Mozambique, Rwanda, Senegal, Tanzania, Uganda, and Zambia). PRGF documents often justified this on (1) limited capacity to borrow domestically without significant negative impact on macro stability and growth—crowding out private sector investment and other spending (Cameroon, the Central African Republic, the Democratic Republic of the Congo, Ethiopia, Ghana, Guinea-Bissau, Malawi, Mozambique, Rwanda, Tanzania, Uganda, and Zambia); (2) domestic demand pressures (Ethiopia, Guinea-Bissau, and Mozambique); and (3) need to reduce domestic debt and large debt-service burdens (Ghana, Malawi, Rwanda, and Senegal).

Revenue mobilization was a frequent program theme in PRGFs. Many programs had tax revenue targets, mostly in the form of indicative targets or benchmarks (Burkina Faso, Cameroon, the Central African Republic, Ghana, Guinea-Bissau, Mozambique, Senegal, Tanzania, and Uganda). However, discussions of the rationale for greater tax revenue mobilization have evolved over the years from the early “aid dependency” reduction motive (Burkina Faso, Malawi, and Mozambique) to creating fiscal space for priority expenditures (the Central African Republic, Ghana, Mozambique, and Uganda) and building adequate capacity for government operations (Tanzania and Uganda) in recent years.

Wage bill targets were common in PRGFs, stemming from fiscal concerns as well as macroeconomic stability considerations. Wage bill conditionality has featured widely—5 of the 14 cases reviewed had performance criteria (PCs) at some point in time, 8 had indicative targets or benchmarks, and Malawi had both in various program reviews (Table A3.4). In some cases, repeated slippages led to strengthened conditionality (from indicative targets to PCs in Malawi), while in others with good performance, targets were downgraded (from PC to benchmark in the Central African Republic). In two cases, the wage bill target was eliminated altogether (Mozambique in 2006 in the context of better-than-expected fiscal performance, and Tanzania in 2003 with the focus having shifted to civil service pay reform). In terms of rationale, documents revealed that program targets on the wage bill stemmed from macroeconomic stability concerns, in most cases with reference to large wage bill increases in the immediate past (Ethiopia, Guinea-Bissau, Ghana, Malawi, Mozambique, and Zambia). Additional motivation included the need to free up fiscal space for other expenditures, including poverty-reducing expenditure (PRE) (the Central African Republic and Mozambique). Wage bill ceilings were also linked frequently to discussions of civil service reforms (the Democratic Republic of the Congo, Ghana, Mozambique, Senegal, Tanzania, and Uganda). The latter was especially important in Mozambique and Tanzania, in connection with the aforementioned elimination of the wage bill targets.

Table A3.3.Evidence on Adjusters1
CaseAid Shortfall FinancingRationaleAid Windfall Spending or

Burkina Faso (1999, 2003)Limited financing.No explicit rationale.Full spending was replaced by full saving in 2001. Full saving was replaced by limited spending on social sectors in 2005.No explicit rationale.
Cameroon (2000)Domestic financing for 50 percent of shortfall.No explicit rationale.Reduce domestic financing for full amount of excess.For crowding-in.
Central African Republic (1998)Limited financing.No explicit rationale.Equivalent amount deducted from government borrowing. Adjusters in 2004 and 2006 Emergency Post-Conflict Assistance allowed use for priority spending or reduction of debt—no proportions specified.No explicit rationale.
Democratic Republic of the Congo (2002)No financing until third review, which stated that 50 percent of any foreign financing shortfall could be financed. By fifth review, full financing was allowed.No explicit rationale.Excess foreign financing to be used to finance poverty reduction expenditure. Subsequent reviews added need to use excess external assistance to reduce net banking system credit to the government.Government’s ambition to reach HIPC completion point was a factor in targeting pro-poor spending. Subsequent focus on reducing banking system credit to government was to ensure success of stabilization effort.
Ethiopia (2001)50 percent financing up to $20 million.Restrain demand pressures.Full saving for any amounts exceeding those programmed. By fourth review limited use for poverty reduction expenditures.Build reserves—which staff noted were precarious, given needs and shocks.
Ghana (1999, 2003)Full financing (1999), from third review, limited financing. Limited financing continued in 2003 program but from third review, no financing allowed.No explicit rationale.Equivalent amount deducted from limit on government borrowing. From third review of 2003 PRGF, full use.No explicit rationale.
Guinea-Bissau (2000)Financing of 50 percent.No explicit rationale.50 percent can be used for priority spending on social and infrastructure areas.Pressing nature of social needs.
Malawi (2000)Initially a maximum of $50 million financing but reduced to zero at the time of Emergency Assistance (2002) and first review (2003).Need to reduce domestic debt to lower interest rates. Strengthened over time in response to repeated slippages.Initially a maximum of $50 million could be used but raised to unlimited.Need to reduce domestic debt.
Mozambique (2004)Initially no domestic financing. By fourth review, partial financing.Maintain pace of fiscal consolidation and create room for private sector. Context of change in adjusters (fourth review) was one of better than expected fiscal performance, with aid decline no longer perceived a risk to the program.Initially partial use (on capital expenditures) and absorption. By fourth review, full use (on priority spending) and absorption.Justified initially on high yearly volatility of aid.
Rwanda (2002)Initially no domestic financing, then changed to limited financing in 2003.No explicit rationale.Full saving.Concern that resources would be diverted to military spending.
Senegal (1998, 2003)Limited financing to CFAF 20 billion.Level of adjustment had to be consistent with regional protocol on monetary policy and fixed exchange rate.No use of excess funds allowed.Level of adjustment had to be consistent with regional protocol on monetary policy and fixed exchange rate.
Tanzania (2000, 2003)Initially limited (to $60 million). By third review (2000), full financing allowed and retained in following program.Initially to safeguard international reserves—relaxed as reserves increased to give government more flexibility in making financing and spending decisions.Initially no use of excess foreign financing allowed. By fourth review (2000), full use allowed which continued in 2003 program.Initially to build international reserves, but use of excess resources later left to government discretion.
Uganda (1997, 2002)Full financing allowed (with the exception of the second review in the first PRGF).Enable government to meet commitments, notably those of Poverty Action Fund (PAF). Net credit to government ceiling would be lowered for any unspent PAF commitments.Full saving for any excess, throughout programs.Enable country to meet debt payments, especially arrears.
Zambia (2004)Partial financing (initially $14 million increased to $20 million)Full saving of windfalls, except to reduce domestic debt.
Table A3.4.Wage Bill Ceilings1
CountryInstrument2RationaleConsideration of Impact on

Priority Sector in Design
Adjustment in Context of

Program Review
Burkina Faso (2003)Indicative target.Contain medium-term pressures on expenditures.
Cameroon (2000)No formal conditionality. But program underlined importance of containing wage escalation.Ensure targeted noninterest expenditure and aimed at containing large wage increase at beginning of the program.Stronger program wording reflecting repeated fiscal slippages.
Central African Republic (1998)PC. In addition, civil service positions (including military and security forces) were not to grow (PA).Ceiling is part of an effort to ensure that adequate resources are available for social spending and critical infrastructure investment.Ceiling allowed for recruitment of 880 new personnel in priority sectors of education and health.PC was turned into a benchmark for the second annual program, with actual wage and salaries in 1998 sectors programmed.
Democratic Republic of the Congo (2002)Ceiling on wage arrears for civil service (kept at zero).Raise morale in civil service.Not explicit—but implication on efficiency in public sector and service delivery.
Ethiopia (2001)Indicative targets.Limit size of wage bill.Program concerns with wage bill eased as issues of macroeconomic management took hold.
Ghana (1999, 2003)In 1999, no target. In 2003, a PA was used in second review, and a PC was used from third review onward. In addition, two structural PCs were introduced relating to civil service reform.Past increases in wage bill that contributed to noncompletion of fifth review of 1999 program.
Guinea-Bissau (2000)Performance indicator.Ceiling is part of fiscal consolidation, reflecting demobilization of troops.Ceiling allowed for an increase in number of civil servants.
Malawi (2000)Benchmark (first review).Need tight stance in order to restore fiscal discipline. Also aimed at containing large wage increase at beginning of program.In 2005, wage bill for priority sector protected by ceiling adjuster (linked to additional aid for health SWAp).Stronger form of conditionality toward end of program in response to repeated fiscal slippages.
Mozambique (1999, 2004)In 1999 no target, in 2004 indicative target. Target abandoned in fourth review.In 2004, in the context of fiscal consolidation and past large wage increases and need of public sector reform (ghost workers). Target abandoned in fourth review (2006) in the context of better than expected fiscal position.Target set with explicit reference to greater employment in health and education.
Rwanda (2002)None.
Senegal (2003)PC.Contain impact on expenditure.Not explicit. But program anticipated that improvements in wage reform would have positive impact on social service delivery.No change. Monitoring included monthly reporting to Fund on changes in wage bill.
Tanzania (2000, 2003)Indicative targets.Contain expenditure on wages, rationalize wage bill.Compensation and wage incentives identified as key for public service delivery.
Uganda (2002)No wage ceilings.
Zambia (2004)Benchmark.Limit wage increases of recent past.Program modified in the course of first review—in coordination with additional donor assistance—to allow for additional hiring in priority sectors.

Fiscal governance and transparency were important pillars of PRGFs. Discussions of public expenditure management and financial accountability (PEFA) issues centered around fiscal governance and transparency issues, including budgetary frameworks, budget execution, monitoring and reporting, and financial management and information systems. The use of structural conditionality in PEFA was extensive (as in Cameroon, the Central African Republic, the Democratic Republic of the Congo, Ghana, Malawi, Mozambique, Rwanda, Tanzania, Uganda, and Zambia), covering expenditure execution, monitoring and control (including on commitments), coverage and timing of budget reporting, information systems (including on public sector payrolls), and in some instances more specific areas—public procurement, auditing, code of ethics in civil service. The program focus on PEFA has been complemented by extensive technical assistance from the Fund, notably in public expenditure management and financial accountability (Burkina Faso, the Democratic Republic of the Congo, Ghana, Mozambique, Rwanda, Senegal, Tanzania, Uganda, and Zambia), including budget preparation and execution, expenditure monitoring and control, and information systems (including for tracking PRE).2

Social impact

With respect to priority PRE, the focus of PRGF-supported programs was generally on tracking activities, and less so on program adjusters or conditionality. Documents reviewed showed that direct program targets on priority expenditures (PCs in Rwanda and Uganda, indicative targets in Ghana and Malawi) were infrequent (Table A3.5). But programs did track priority expenditures, with tables dedicated to this in staff reports. In some instances, documents described in general terms recent developments with priority expenditures and government plans going forward (e.g., Mozambique and Zambia). As discussed in Chapter 2, the section on “Key Features Agenda,” program adjusters for incremental aid were linked to priority expenditures in 8 of the 14 cases reviewed. But, except in a few instances (the Democratic Republic of the Congo, Guinea-Bissau, and Uganda) where the pressing nature of social needs and protecting government commitments were noted, there was little explicit rationale for linking adjusters to priority expenditures.

Table A3.5.Priority Poverty-Reducing Expenditures 1
CountryInstrumentAid Shortfall Adjuster: Link to

Priority Expenditure
Aid Windfalls Adjuster: Link to Priority

Burkina Faso (2003)No conditionality.No link to priority expenditureAdjuster allowing limited spending of windfalls only on poverty reduction and special programs as defined by HIPC/PRSP process.
Cameroon (2000)No conditionality.
Central African Republic (1998)No conditionality.No link to priority expenditure.No link to priority expenditure in 1998 program. The adjusters on windfalls in 2004 and 2006 Emergency Post-Conflict Assistance allow for priority public spending or reduction of domestic arrears or reduction of domestic and/or external debt—but no proportions specified.
Democratic Republic of the Congo (2002)Ceiling on net credit to government raised to meet programmed financing of poverty reduction.Excess to be used for poverty reduction spending.
Ethiopia (2001)No link to priority expendituresUp to $50 million of excess foreign financing (including HIPC relief) would be targeted at poverty reduction, and a similar amount on “special programs” (fourth review).
Ghana (1999, 2003)Indicative target.
Guinea-Bissau (2000)No conditionality.No link to priority expenditure.Given pressing nature of social needs, adjusters were to partially allow for increased directed spending with 50 percent of resources to be spent on social and infrastructure projects identified in collaboration with World Bank.
Malawi (2000, 2005)Indicative target on pro-poor expenditure, first review, 2003.No adjusters on indicative targets on pro-poor expenditure.No adjusters on indicative targets on pro-poor expenditure.
Mozambique (2004)No program targets on PRSP expenditures, but tracking of developments and government plans.In fourth review, partial financing of shortfalls. No link to priority expenditure.Limited accommodation for additional capital outlays financed by budgetary grants. By fourth review, full accommodation to be used in priority expenditures identified in budget.
Rwanda (2002)Performance criteria on broadly defined “priority spending” (mainly social and infrastructure) and “exceptional expenditures” (mainly post-genocide-related expenditures).
Senegal (2003)Indicative targets on programmed spending of HIPC debt relief—but sectors of focus not specified.Ceiling on net cumulative change on credit to government to be raised for aid shortfalls on HIPC-related (i.e., priority) expenditure from programmed levels.Ceiling to be lowered from higher HIPC-related (i.e., priority) expenditure than programmed levels.
Tanzania (2003)No conditionality.No link to priority expenditure.No adjuster on excess financing—use left to government’s direction.
Uganda (2002)A performance criterion on minimum expenditures under Poverty Action Fund (including universal primary education). An adjuster indicated that any amounts falling below those programmed would lead to lowering of the ceiling on net government credit.No link to priority expenditure.Ceiling on net credit to the government was to be lowered (raised) by shortfall (excess) expenditure on areas in Poverty Action Fund—universal education, primary healthcare, access to clean water, and so on.
Zambia (2004)No program target.

Wage bill ceilings were often set without consideration of the impact on expenditures in priority areas. In only a few cases (the Central African Republic, Guinea-Bissau, and Mozambique) did documents acknowledge explicitly that program design took priority sectors into account while setting wage bill ceilings (and not throughout the evaluation period but only more recently, as in the case of Mozambique). Only in the case of Malawi were adjusters included to allow additional aid to be used to increase wages in priority areas. In Zambia, the PRGF was adjusted in the context of the program review to accommodate additional employment in priority sectors, when the wage bill ceiling proved binding.

PSIA results were frequently reported but rarely informed PRGF programs. Since the creation of the PSIA group in FAD in 2004, the Fund has conducted nine assessments (Table A3.6); six were focused on subsidies (electricity, petroleum, agriculture, and fertilizers) and the rest on other macroeconomic areas (taxation, devaluation, and external shocks). The results from PSIAs were generally presented in program documents (except in Mali and Malawi), but were rarely part of appraisals (except for Burkina Faso and Djibouti). Program documents indicated no specific countervailing measures linked to the PSIAs, in some cases because the recommendations were not adopted (Malawi and Uganda). In two instances, programs noted that the resulting fiscal space would be used by the authorities to increase priority expenditures (Ghana and Mali).

Table A3.6.Poverty and Social Impact Analysis Conducted by Fiscal Affairs Department1
Country (PSIA

Completion Date)
Sector or TopicDiscussion in PRGF DocumentsCountervailing Measures
Burkina Faso2 (2006)Electricity tariff reform.PSIA recommendation on electricity tariffs reflected in sixth review of PRGF in 2006 (recommendation was to raise tariffs because of marginal impact on the poor).No explicit countervailing measures in PRGF for increase in electricity tariffs. PSIA report had argued that few poor households were connected to electricity grid.
Djibouti3 (2005)Devaluation.PSIA finding that devaluation would be disruptive because of import dependence, featured prominently in the Staff Report for the 2005 Article IV consultation and staff monitored program. Board discussions also alluded to PSIA findings.No devaluation suggested by the staff-monitored program. Concerned over competitiveness, the staff-monitored program suggested lowering government wages—taking into account poorest households.
Ghana4 (2005)Petroleum pricing.PSIA was done before a number of petroleum pricing reforms were undertaken in February 2005, notably implementation of a new automatic price adjustment mechanism (see Staff Report for the 2005 Article IV consultation).The “fiscal space” created, inter alia, by removal of petroleum price subsidies was to be spent on health and education and infrastructure in rural areas (Memorandum of Economic and Financial Policies in Staff Report for 2005 Article IV consultation).
Madagascar5 (2006)Rice subsidies
Malawi6 (2006)Fertilizer subsidies.PSIA pricing reforms not explicitly reflected in August 2006 PRGF review. reforms put off by drought and food crisis.PSIA report had no policy impact on fertilizer subsidy, and so no mitigation in PRGF required.
Mali7 (2006)Petroleum pricing.The fourth review of PRGF (June 2006) mentions that “external” studies were crucial in determining petroleum pricing mechanism.No special measures for mitigation considered in PRGF—but authorities indicated that the resulting fiscal space was to be used to develop infrastructure and transport networks to address poverty.
Mali8 (2005)Impact of external shocks and macro responses on poverty.No explicit reference to PSIA exercise in subsequent staff reports, but general reference to strategies for poverty reduction (see fourth review, June 2006).
Senegal9 (2005)reform of groundnut marketing.Groundnut sector reform was an ongoing process before PSIA. But groundnut parastatal was privatized after PSIA (had failed before), although there was little change in edible oil pricing policies (private company still a protected monopoly). This was discussed in third and fourth reviews (December 2005).No countervailing measures in PRGF (PSIA measures not implemented).
Uganda10 (2005)Value-added tax (VAT).PSIA analysis used in staff report of May 2005 to suggest two alternative means of raising revenue with minimum negative impact on poor: change VAT rate (from 17 percent to 18 percent) and increased excise taxes (on petroleum).No countervailing measures in PRGF as tax changes were not adopted by government.

Other issues

Bank-Fund collaboration was most frequently noted on PEFA and financial sector work. Program documents reported frequently, but not always, on the division of labor between the Bank and the Fund, specifying the lead institution as well as areas requiring joint work (Burkina Faso, Ghana, Mozambique, Rwanda, Senegal, Tanzania, and Zambia). In general, PRGF programs put macroeconomic issues within the IMF’s core areas of responsibility and sectoral and social issues within those of the World Bank. As noted above, the IMF and the World Bank shared responsibilities for PEFA and financial sector work. More specific delineations of inputs into the collaborative effort were sometimes indicated. For example, in some programs the IMF would focus on financial management information systems or medium-term expenditure frameworks and the Bank on other areas of PEFA (Mozambique and Zambia). With respect to alternative scenarios, and with the exception of Ethiopia, program documents were not clear on the role collaboration with the Bank played.

Table A3.7.Public Expenditure Management and Financial Accountability1
CaseStructural Conditionality2Technical Assistance
Burkina Faso (1999, 2003)Computerized monitoring of investment expenditure execution (IT).Strengthening budget preparation and expenditure control.
Specific codes for identifying social expenditure and expenditure financed under the HIPC Initiative (IT).Strengthening system to track poverty-reducing public expenditures.
Cameroon (2000)Render operational the interim system for public procurement (PC).Review of public expenditure management.
Issue quarterly reports on budgetary execution (B).
Central African Republic (1998)Complete validation process for domestic debt (B).No technical assistance (TA) related to public expenditure management and financial accountability (PEFA).
Democratic Republic of the Congo (2002)Introduce code of ethics for civil service (PC)Expenditure management.
Ethiopia (2001)No PEM-related conditionality.No PEFA-related TA.
Ghana (2003)Publish past month’s fiscal report (PA).Five instances of TA on public expenditure management.
Payroll information system (PC).
Monthly fiscal report (B).
Guinea-Bissau (2000)No structural conditionality specified.Strengthening fiscal controls; assessing budget management and tax system.
Malawi (2000)Effective implementation of expenditure monitoring and control (PA).Budget management, expenditure control, and expenditure management.
Monthly reports on commitment levels (PC).
Launch of Ministry of Finance unit to monitor parastatal spending (PC).
Commitment controls; reports on proverty-reducing expenditure (PRE); anti-corruption; parastatal borrowing (B).
Mozambique (2004)Quarterly budget reporting (PA).Seven instances of TA on public expenditure management.
Implement integrated financial management system (B).
Rwanda (2002)Incorporate any extrabudgetary and off-budget projects and transactions into the budget to the extent appropriate (PC).Budget execution; expenditure management; tax policy; assessment of tracking of poverty-reducing expenditure.
Senegal (2003)Adopt WAEMU expenditure management directives (PA).Capacity to track PPE.
Undertake pilot on monthly treasury accounts (PC).
Auditing of treasury accounts (B).
Tanzania (2003)Identify budget codes for PRE (PC).Public expenditure management and fiscal decentralization.
Quarterly reports from spending agencies (B).
Uganda (2002)Submit plan for implementation of report on public administration budgeting to cabinet (PC).Local government budgeting; budgeting and commitment control; public expenditure management.
Zambia (2004)Approval of PEFA program (PA).Six instances of TA on public expenditure management.
Publication of quarterly budget execution plans; introduction of financial information system (PC).

Private sector development and its contribution to economic development and growth were frequent themes in PRGF-supported programs. It was discussed in relation to removing obstacles to private sector growth by improving the business climate, including the regulatory and judicial environment, and basic infrastructure. But programs left specific work to the World Bank. PRGFs rarely included structural conditionality in these areas and the IMF did not provide technical assistance. The main channel through which the Fund addressed private sector development issues in program design was in the context of crowding-out considerations when setting fiscal targets (as discussed above), and in a few instances through structural conditionality in the financial sector (Mozambique, Tanand Zambia). The latter especially related to the regulatory and supervisory infrastructure—including for microfinance. The IMF has also provided significant technical assistance for financial sector issues, including through Financial Sector Assessment Programs (as in Ghana, Rwanda, Tanzania, Uganda, and Zambia).

Recent evaluations of the effectiveness of Fund technical assistance in the PEFA area indicate a mixed picture, mirroring the performance of IMF-supported programs (IMF, 2004a and 2005i). Countries further ahead in the reform process (e.g., Cameroon, Uganda, Tanzania, and Rwanda) showed greater progress in the PEFA area than those where the reform pace had been slower (e.g., Côte d’Ivoire, the Central African Republic, Malawi, and Zambia). On the effectiveness of technical assistance delivery in PEFA, a recent IEO evaluation (IEO, 2005b) noted that longer-term, resident technical assistance was more effective than shorter-term interventions, because of greater access to expertise and training possibilities.

    Other Resources Citing This Publication