Papers - DDCI Research

DDCI Research Other Research Conference Papers Briefing Papers Articles

Executive Summary

This study examines the Poverty Reduction Support Credit, the World Bank’s programmatic lending introduced in 2001.  According to the World Bank the PRSC is based on a more flexible approach to conditionality.  The key questions are: what has changed with the introduction of the PRSC; what are the continuities with previous World Bank approaches to conditionality.   The aim of the study is two-fold:

§      To contribute to the World Bank’s conditionality review due to be presented to the Development Committee at the autumn meetings 2005.  The particular aim is to test the World Bank’s claim that the PRSC allows more flexibility in terms of conditionality.

§      To contribute to the broader debate on the level and content of conditionality, in particular  economic policy conditions.  Positions taken recently by a number of major donors have started to break the consensus which concedes the lead role in setting conditions to the World Bank and IMF.

The methodology consisted of documentary research, examining 13 PRSC programmes across Africa, Asia and Latin America.  The focus of the Report is on conditionality.  Particular attention is paid to privatisation for two reasons: a) privatisation is particularly controversial and action by governments’ to implement these conditions has led to unrest in many countries [1] ; b) some donors, in particular DFID, have stated explicitly that they will no longer call for privatisation as a condition for their aid.

The findings show that:

§      the number of binding conditions imposed by the World Bank has decreased.

§      this decrease has been accompanied by a sharp upturn in the number of benchmark or  ‘soft’ conditions.  While the World Bank may argue that compliance with all benchmarks is not essential in order for PRSC funds to be released, progress on meeting the benchmarks is taken into account in judging whether there has been ‘satisfactory progress’. 

§      In terms of the content of conditions, there has been a significant increase in public finance management and good governance conditions which now form 30% of overall conditions.

§      The overall thrust of the conditions are similar to previous World Bank approaches: encouraging private sector investment; decentralising public service provision; private sector provision of basic infrastructure (water, energy, telecommunications); liberalisation of trade; privatisation of agricultural State Owned Enterprises; reform of Public Financial Management and accountability procedures.

§      Privatisation conditions were found in all but 2 PRSCs.

§      While PSIAs were carried out on most PRSCs, their function was to plan how to alleviate any negative effects of the policy rather than to open up debate as to whether the proposed policy was the best one.

Over all while the reduction in binding conditions is welcome, the study concludes that the World Bank continues to ‘buy reforms’ with money only released after a range of ‘prior actions’ have been met.   Further, the large number of benchmarks ring alarm bells – raising the fear that the World Bank wishes to continue to micro manage economies . Effectively there has been no streamlining (if all types of conditions are taken into account) and the application of conditionality continues to undermine government ownership and potentially to impact on aid predictability.   This concern is reinforced by the lack of clear evidence of a more flexible approach in terms of policy choices.  It is striking that PSIAs still focus mainly on identifying the need for safety nets rather than opening up debate on what are the appropriate policies.  Generally there seems to be little variation in the formulation of PRSC programmes.  On privatisation, in spite of the World Bank’s recognition that these are often in direct conflict with public opinion and indeed parliamentary motions,  these conditions continue to appear in most PRCS in our sample.  Finally the Report highlights the lack of transparency in how PRSCs are developed and adopted with both Parliaments and civil society excluded from examining a programme which will impact on their lives and may also mount as debt.  PRSCs are only released publicly when they have been agreed with the government.

1. Policy Conditions: PRSCs should cease to include privatisation and other economic policy conditions.
2. Criticality: World Bank should make clear the criteria on which critical conditions are based.  Prior actions must be properly explained and open to debate and challenge.
3. Consultation/transparency:  PRSCs should be made public in draft form.  Both Parliaments and civil society groups must have an opportunity to input into PRSCs before they are agreed.
4. Cross conditionality: there should be no cross conditionality with IMF or WTO programmes.
5. Donor Alignment: other multilateral, regional and bilateral donors should align with the PRSP and not the PRSC and PAFs should be drawn from the PRSP not the PRSC.
6. PSIAs: assessments carried out on key elements of PRSCs should factor in consideration of alternative approaches.

Contents

1.     Introduction
2.     Poverty Reduction Support Credits
3.     Conditionality and PRSCs
4.     Moving from the base-case lending scenario
5.     Criticality
6.     The focus of PRSC conditionality
7.     PRSC conditionality and privatisation
8.     Poverty and Social Impact Analysis
9.     PRSC conditionality and ownership
10.   Cross conditionality and explicit links between the PRSC and PRGF
11.   Donor Harmonisation
12.   Conclusions
13.   Recommendations

Annexe

Table 1: Prior Actions and Monitorable Actions Per Country
Table 2: Prior Actions and All Monitorable Actions Per Sector
Table 3: Prior Actions and Benchmarks by Reform Area Included
               In a 3-year PRSC Programme
Table 4: Privatisation and Public-Private Partnership Prior Actions and
              Benchmarks by Sector
Table 5: Increase in Total Number of Conditions from Earlier to Later PRSC

Documents Consulted

1.  Introduction

The World Bank is one of the major providers of finance to low income countries.  For almost two decades the conditions underpinning the Bank’s structural adjustment lending have been steeped in controversy.  Criticism of conditions has been two fold:  they restrict the decision making space for elected governments and undermine democracy; they are not poverty proofed and therefore can have a damaging effect on borrowing countries.

The World Bank introduced a new lending programme for low-income countries in 2001, the Poverty Reduction Support Credit (PRSC).  The PRSC is the World Bank’s programmatic lending to support low-income countries’ PRSPs.  According to the World Bank its programmatic lending marks a shift from the traditional approach to conditionality, allowing for greater flexibility and medium term support.  As the PRSC is a fairly new World Bank lending facility, very little civil society research has been carried out on how it operates and what it supports.  The purpose of this research is to provide a profile of the PRSC focussing in particular on conditionality and testing World Bank claims outlined above.  We see this work as both a contribution to the World Bank Review of Conditionality and, in the longer term, the basis for discussion with bilateral donors, the World Band and other interested bodies.

The Report is based on an examination of PRSCs agreed with 13 countries [2] - at the time of carrying out this research PRSCs had been agreed with 20 countries. The focus is on conditionality with a view to identifying both what has changed and the degree of continuity with the previous World Bank approach to conditionality.  The focus in particular is on privatisation for two reasons: a) privatisation is among the most controversial conditions laid down by the World Bank and has led to protests within countries; b) some major donors now no longer use specific policy conditions as a condition for aid.  ‘We will not make our aid conditional on specific policy decisions by partner governments, or attempt to impose policy choices on them (including in sensitive economic areas such as privatisation or trade liberalisation’. [3]   The EC has also indicated a shift in its position on economic policy conditions.  The EC will not necessarily cut budget support to a country if it is off track on its IMF programme in relation to e.g. privatisation conditions [4] .  

All tables referred to in the text can be found at the end of the document in the annexe.

2.  Poverty Reduction Support Credits (PRSCs)

Poverty Reduction Support Credits (PRSC) were introduced by the World Bank in 2001 to provide financial support for a government’s medium term development programme (usually the Poverty Reduction Strategy Paper (PRSP). PRSCs are provided as concessional loans or grants and are available to International Development Association (IDA) borrowers. Typically, three successive, single tranche (sometimes two tranche) PRSCs are bundled together into a medium-term programme in support of a country’s PRSP. Usually single tranche PRSCs are provided to those countries with what the World Bank judges to be a good track record of reform, whilst two tranche PRSCs are made available to those countries without one. The aim is to synchronise the release of the PRSC with a government’s budget and policy planning cycle and with the release of IMF Poverty Reduction and Growth Facility (PRGF) loans.

PRSCs are additional to other forms of lending such as investment projects and other policy-based lending.  However, it is the Bank’s intention to shift towards a greater use of PRSCs rather than investment loans as the main vehicle for providing finance and engaging in policy dialogue. Thus in some countries new investment projects are not being negotiated as old ones come to an end, and are being replaced with PRSCs.

PRSCs are considered preferable to investment loans because they can help to reduce transaction costs incurred by the recipient government. PRSCs are released directly to the Ministry of Finance and are disbursed through the budget.  Thus they avoid the creation of parallel structures to the government’s own, thereby facilitating improved aid and budgetary management by the recipient government.

3. Conditionality and PRSCs

Before a PRSC is sent to the World Bank Board for approval and resources are released the government must first complete a set of ‘prior actions’. [5] These are agreed between the World Bank and the government when the PRSC programme is formulated. Each individual PRSC in a series has prior actions attached to it. [6]

In addition to prior action conditions a PRSC programme includes trigger conditions. However, there is little discernable difference between the two. The only difference is that trigger conditions are provisional prior actions, [7] that is, they are actions that must be undertaken during the current PRSC to qualify for a subsequent PRSC. Their implementation signals the go-ahead to negotiate the subsequent PRSC. Typically the prior actions finally agreed as conditions for accessing the subsequent PRSC will be the same as the trigger conditions identified in the previous PRSC, although they might vary slightly to take account of developments in the economy or delays in reforms or in economic and sector work, and additional prior actions might be included: “in the context of programmatic operations, triggers are the expected prior actions for a subsequent loan, credit, or grant.” [8]

In the case that not all prior action conditions are met it does not mean that the government will receive no budget support from the Bank. Depending on the reason for non-implementation, the Bank can delay the subsequent PRSC, reduce the amount of the loan/grant, or revise the PRSC programme. This continues standard Bank practice in which certain conditions might be waived or bundles of conditions might be linked to separate loan tranches so that if not all conditions are met disbursement of some tranches can still be made.

In addition to the prior actions which are specifically tied to PRSC release, a PRSC programme includes ‘benchmark’ measures (also known as “milestones”) which are written into the policy matrix and are also expected to be undertaken prior to the PRSC loan/grant release. Benchmarks are included in addition to prior actions to “help keep the program on track during its implementation, in particular by creating a timetable and incentives for completing actions.” [9] Benchmarks/milestones are not directly tied to the release of the PRSC loan/grant and therefore the timing of their implementation is more flexible, and because they are ‘non-binding’ their implementation “is, by definition, regarded as a lower priority than the achievement of conditions and triggers.” [10] However, their implementation is monitored by Bank staff to determine whether a government is on track with its programme. Before a PRSC is released, in addition to fulfilling the prior actions and other measures, the Bank’s Board must be assured that there has been ‘satisfactory progress’ in implementing these ‘benchmark’ conditions: “The Bank makes the loan funds available to the borrower upon maintenance of an adequate macroeconomic policy framework, implementation of the overall program in a manner satisfactory to the Bank, and compliance with these critical program conditions.” [11] Whilst failure to implement one or two of these benchmarks is not likely to cause a delay in PRSC release, failure to implement several could signal that the programme is going off-track and could lead to a halt in disbursement. [12]

In most cases PRSC conditions are focussed on inputs, ie specific reforms to be implemented or actions to be taken. Although there is reference in most PRSC programme matrices to expected outcomes these are used for monitoring purposes only. Conditionality is not linked to these directly and therefore they are not used to determine the release of resources.

4.  Moving from the base-case lending scenario

Trigger conditions are also included in the Country Assistance Strategy attached to IDA lending scenarios. A trigger to move between the ‘base-case’ lending scenario and the ‘high case’ lending scenario might be the successful completion of all or some of a PRSC’s prior actions/triggers and benchmarks. For example, a condition for Nicaragua to move into the high-case lending scenario is that there is significant progress in completing the ENEL (the public power company) and ENITEL (the national telecom company) privatisation programmes or strengthening of property rights through approval of a Land Registry Law and an Indigenous Lands Demarcation Law. [13] The former condition is a benchmark in the PRSC1 programme. A trigger condition for Benin to move into the low case lending scenario includes unsatisfactory progress withprivatisationof Societe Nationale de Promotion Agricole (SONAPRA) (the National Cotton Company) and SBEE (the National Company for Water and Electricity), which are prior actions linked to PRSCs 2 and 3. [14] Linking completion of PRSC conditions to triggers for moving into different lending scenarios gives additional incentive to the government to implement the conditions: “The Government will have strong incentives to implement the PRSC program as IDA's possible future lending program will depend on satisfactory progress of the program” [15]

Alternatively, CAS trigger conditions might need to be fulfilled to bump up the government into a ‘high-case’ lending scenario where it is then able to access the next PRSC (or conversely if trigger conditions are not implemented then the government might fall into a low case scenario where it does not have access to a subsequent PRSC). For example, in the case of Vietnam, “Progress across all triggers will lead to the consideration of high-case lending for PRSC 4. Progress across most of them and no backtracking in any sector would keep Vietnam in the base-case lending scenario. Limited progress or serious backtracking in various sectors could move Vietnam to low-case lending, which in practice means that the PRSC 4 operation could be delayed until the reform process regains momentum.” [16]

Given that governments are likely to have both a preference for budget support rather than investment lending and a preference for more World Bank lending than less, the use of PRSC conditions as triggers for moving between lending scenarios and the tendency to include subsequent PRSCs in high-case lending scenarios gives greater incentive to governments to implement PRSC conditions.

5.  Criticality

Debates about the usefulness of conditionality have led to calls for the number of conditions to be kept to a minimum, both to facilitate government ownership of the reform agenda and to improve the contracting function of conditionality. [17] Whilst no limit has been imposed as to how many prior actions can be applied to a PRSC, the World Bank’s Interim Guidelines on PRSCs (which have now been subsumed into Operational Directive 8.60 and Best Practice Guidelines 8.60) suggest that binding conditions should be “imposed on those reforms which are regarded as “critical” for the success of the PRSP and the PRSC programme and are expected to have a “substantial impact on sustainable growth and poverty reduction”. Likewise, the Bank’s 2004 Adjustment Lending to Development Policy Lending: Update of World Bank Policy urges staff to “use only the highest priority actions as conditionality and triggers.” [18] This advice appears to have been reasonably well heeded. Koeberle (2005b) has found that the number of binding conditions per adjustment operation has dropped from a high of 46 in 1992 to 15 in 2004. On average, in the sample of PRSCs used for this study, a PRSC had 12 prior action conditions per tranche, although actual numbers varied between 7 for Mozambique and 23 for Senegal (see table 1). However, the fact that there is considerable variation in the number of prior actions imposed suggests that the staff is not clear as to what is and what is not a critical condition, and the Bank has failed to define this for the staff. It is problematic that the World Bank does not have detailed and clear guidelines for determining what is and what is not a ‘critical’ condition, and that staff are not required to justify their use of prior actions in the PRSC programme document, since it clouds transparency and prevents a proper assessment of the Bank’s justification of and application of conditionality. Although PRSC programme documents include a discussion of why particular sectors and reforms are supported by the PRSC, and may link prior actions and triggers to these, there is no discussion as to why or how the particular prior actions have been chosen above others.

To facilitate the minimal use of prior actions (and other binding conditions) the Bank has encouraged staff to make greater use of ‘benchmarks’ or ‘milestones’ to guide and monitor the reform process. In the sample used for this study the average number of benchmark conditions applied per tranche is 50, which is a little higher than the average of 42 per adjustment operation that Koeberle (2005a) calculates. However, again this masks a vast difference in the number of benchmarks applied from country to country, with Guyana’s first PRSC having as few as 16 whilst Rwanda's second PRSC programme provisionally has 82. The explanation for this extreme difference in the application of benchmark conditionality is not obvious and is not justified by the Bank in its programme documentation. What is apparent is that no criticality tests are applied to benchmark conditions.

Koeberle’s (2005b) analysis indicates that the use of benchmark conditionality is on the increase, particularly for programmatic loans. Whilst the sample used for this study is small it is interesting to note that for all but one of the countries from one PRSC to the next the number of benchmark conditions rises considerably (see Table 5).

Given the vagueness of benchmark conditionality, in the sense that a government does not know how many of the benchmarks need to be achieved to show reasonable progress in programme implementation and it is expected that “actions selected as milestones may not be achieved according to the initial timetable”, it is problematic that the Bank applies so many to its PRSC programmes. [19] The more that are applied the less clarity there is and the more subjective become disbursement decisions. The continued use of a large number of benchmarks suggests that the Bank is keen to continue micromanaging the reform process. For example, benchmarks such as “designate a manager for each program selected”, which was included in Benin’s programme seem unnecessary and could be counter-productive by undermining a government’s sense of programme ownership.

6.  The focus of PRSC conditionality

The Bank claims that PRSCs differ from its standard policy-based loans because they focus largely on social sectors and public sector management reforms, with a particular focus on the latter. This claim is more or less supported by analysis of the distribution of conditions across sectors, although PRSCs, like typical policy-based loans, also include a large variety of other sectoral reforms such as private sector development, agricultural sector reforms and civil service reforms (see tables 2 and 3). Whilst most conditions are applied to reforms in the areas of public sector management and health, more prior actions are imposed on private sector development, privatisation or reform of State Owned Enterprises (SOEs) than are imposed on education reforms.

The central focus of PRSC programmes on public sector management is facilitated by the fact that PRSCs are disbursed through the government’s budget. Previously, the development of government systems received limited support, because most donors tended to focus on their own aid delivery mechanisms. Another bonus from a World Bank perspective is that the provision of budget support offers the Bank the opportunity to engage the government on discussions about budget allocations and priorities, which means that the Bank has become even more deeply involved in a government’s economic decision-making: “The added value of the PRSC lies in its leverage to move forward the policy dialogue and policy actions, particularly in the area of strategic budgeting and planning, overall and sectorally.” [20]

Despite the Bank’s insistence that PRSC programmes have different priorities to the standard adjustment programme, there is an uncanny similarity in the focus of each PRSC programme (see tables 2 and 3), which raises the question of how much influence the Bank has on the design of the PRSP (and subsequently the priorities of the PRSC) and suggests that a standardised policy model is being advocated (if not imposed) by the Bank. Although the details of the reforms differ from country to country the overall thrust of reforms is more or less in the same direction: encouraging private sector investment; decentralising public service provision; private sector provision of basic infrastructure (water, energy, telecommunications); liberalisation of trade; privatisation of agricultural state owned enterprises; reform of Public Financial Management and accountability procedures.

7.  PRSC conditionality and privatisation

Most of the programmes reviewed in this study include one or more prior actions or benchmark conditions requiring privatisation of government enterprises or public-private partnerships. Only two out of the thirteen programmes –Tanzania and Uganda - did not include any such conditionality, whilst Burkina Faso’s programme included five conditions and Benin and Mozambique’s programmes both had four (see table 4). The sectors in which there is most privatisation conditionality are telecommunications, energy and water.

Most of the PRSPs – on which the PRSC should be based - do include references to privatisation and public-private partnerships in the sectors in which PRSC conditionality is imposed. [21] However, in several cases, there is no explicit mention of privatisation or public-private partnerships in the PRSP, and yet such reforms are included as conditions in the PRSC. For example:

§                Burkina Faso’s PRSP talks about reforming the role of the state and encouraging greater private enterprise. Reducing input costs of production – water, energy, telecommunications is a key strategy which involves privatisation. The PRSP does not mention privatisation of parts of the state cotton firm although this is a prior action condition in the PRSC.

§                In Ghana’s PRSP there are calls for the unbundling of the electricity supply system but there is no mention of private-public partnerships although this is a prior action for the PRSC.

§                In Mozambique’s PRSP promotion of private sector participation in the energy sector is identified as a main objective. Privatisation of telecommunications is not explicitly mentioned but restructuring is.  However, in the PRSC document it is required as a benchmark condition.

§                Senegal’s PRSP recognises that the supply of health care is not sufficient but it does not call for increased involvement of the private sector in the provision of health services although this is a benchmark in the PRSC.

§                Guyana’s PRSP calls for the creation of Guyana water with the merger of two water utilities and for the new company to be autonomous from the government. However, it does not explicitly call for private management of the new water company, which is a prior action in its PRSC1.

Given that in many countries privatisation and public-private partnerships are contentious issues, it is questionable why conditionality of this sort is being included in PRSCs, particularly in the water and electricity sectors. In the cases of Nicaragua, Mozambique, Benin and Burkina Faso the PRSC programme documents suggest that there may be opposition to planned privatisations from the public and/or parts of the government.

§                Nicaragua’s first PRSC programme includes benchmark conditionality on the first tranche to initiate discussion in the national assembly of a new water law which would allow private sector management of water companies, and in the second tranche to award management contracts through a bidding process. However, the PRSC programme document reveals that whilst discussion of a new Water Law “is a positive sign, public opinion is still adverse to private sector participation, since it is broadly critical of recent privatisation processes, especially that of the electricity sector.” [22]

§                In the case of Mozambique, the Bank notes that “Domestic support for reforms is particularly important at this stage in Mozambique’s development, to cushion against a potential backlash from utility and infrastructure privatization and to ensure policy continuity into the new Government (to be elected in December 2004), and so the PRSC will continue to draw conditions from the PAF [Performance Assessment Framework] matrix as revised from time to time through biennial joint reviews.” [23] Yet, benchmark conditions related to privatisation included in the PRSC matrix are not included in (i.e. are additional to) the PAF agreed with the government and other donors supporting the PRSP.

§                The PRSC programme document for Benin notes that, “The pace of implementation of the privatization program and establishment of the appropriate regulatory frameworks in privatized sectors has been slow and needs to be accelerated in order to reach the PRSP objectives.” [24] Progress has been slow in the cotton, electricity, telecommunications and port sectors. In particular, privatisation of the port management was contested by the unions. Renewed efforts towards privatisation coincide with the Bank’s PRSC programme which includes privatisation reforms as benchmark conditions, some of which are also trigger conditions to stay in the IDA base-case lending scenario.

§                The PRSC programme document for Burkina Faso notes that there is a risk that the government will not implement the privatisation programme in the run up to the presidential election in November 2005, which suggests that there is foreseen to be considerable domestic opposition to privatisation, and subsequently not strong government ownership.

It is worth noting that in two cases in the sample – Benin and Nicaragua – certain privatisation benchmark conditions were not only included in the PRSC but were also included as trigger conditions for CAS lending scenarios giving added incentive to (or putting increased pressure on) these governments to implement them (see above).

By benchmarking privatisation reforms in cases where there may be public opposition – for example in Benin, Mozambique and Nicaragua - rather than setting them as prior actions Bank staff appear to have heeded the Bank’s guidelines which state that staff should not “include [binding] conditions and triggers that might only stimulate political opposition before the government leadership can build its case and strengthen support for a particular reform.” [25] Whilst the Bank is apparently cautious not to trigger opposition to privatisation reforms by imposing binding conditions, nonetheless, the use of benchmark conditionality is still intended to steer the government in this direction.

Given that there is often strong opposition to privatisation, particularly outside the government, the World Bank should follow the lead set by the UK’s Department for International Development (DFID) not to include conditions related to privatisation in their lending programmes. [26]

8.  Poverty and Social Impact Analysis

Poverty and Social Impact Analysis (PSIA) is being undertaken in relation to most of the PRSCs in the sample (10 out of 13). However, only in four programmes (out of twelve requiring privatisation) was PSIA undertaken on the proposed privatisation or private-public partnerships reforms. These were for cotton privatisation in Benin; power sector reforms and price changes in Ghana; privatisation of tea companies in Rwanda; and private sector management of water services in Guyana. [27] In the case of Nicaragua, in the first and second tranches of PRSC1 benchmarks require the National Assembly to initiate discussion of a new water law that would allow signing of management contracts for water and sanitation services and bidding processes for the management contracts must be carried out, but it is not until PRSC2 – when the management of water facilities has already been handed over to the private sector - that PSIA is to be undertaken on the tariff regime at which point the tariff regime will be adjusted based on its findings.

In most cases PSIA is being conducted once a reform has been decided upon, rather than facilitating debate and decision-making between various reforms. Thus it is not helping to determine the choice of policy but is being used to ascertain whether and where safety-nets might be needed or to facilitate public support of reforms (such as in generating support for water privatisation in Nicaragua). PRSC programme documents do not elaborate why certain reforms are selected for PSIA and others are not. Whilst some PRSC programmes might briefly report the findings of PSIA few identify what precise measures will need to be taken to ensure that overall there is a net benefit to the poor (other than perhaps stating the more pro-poor spending will compensate for any negative effects), and none cross refer to where these policies are included in the current or future PRSC programmes.

9.  PRSC conditionality and ownership

The programme of reforms which the PRSC supports is supposed to be aligned with the Poverty Reduction Strategy Paper (PRSP) and is based on economic and sector work (ESW) carried out by the Bank. This analysis informs the Bank's advice to the country in its preparation of the PRSP, and the Bank's assessment in the Joint Staff Assessment (JSA). It also informs the development of reform actions in the PRSC from the underlying PRSP. Guided by the Country Policy and Institutional Assessment (CPIA), ESW focuses on areas where reform has been judged weakest. Weakness in reform efforts may signal lack of government agreement with policies or simply indicate a lack of government capacity.  It must be noted that the CPIA is under criticism with regard to the lack of transparency of its findings, the lack of neutrality of the assessments, and the one-size-fits all model that it promotes’ [28]

To the extent that the policy conditions attached to PRSCs are drawn from a government’s PRSP and are negotiated with the government then it could be argued that the government is willing to adopt the reforms and that the conditionality is owned by the government. However, in some cases the links between the PRSC programme and the PRSP have not been robust: “The link between PRSC measures and the PRSP impact indicators was at times tenuous…” [29] This is confirmed more generally by the World Bank’s Operations Evaluation Department (OED) review of the PRSP process, which found that “The Bank has effectively supported PRSP formulation, but alignment has not entailed major changes in Bank programs” and “overall, there is still little evidence that donors have coordinated and selected the majority of their programs in response to PRSPs.” [30] The Bank’s standard defence is that whilst a programme matrix should draw from the actions and targets laid out in a PRSP, “such strategy documents seldom include all of the relevant actions and outcomes appropriate for a particular reform program, and often they are insufficiently specific for a program supported by a development policy operation. Hence, the Program Matrix actions are typically more elaborate than those in the national strategy document.” [31] Efforts are being made in subsequent PRSPs to include more policy detail and monitoring criteria.

Whilst the links between the PRSC and the PRSP are clearly doubtful at times, even where the two do seem to be in alignment there is still considerable doubt cast on the degree of government ownership because of the extensive World Bank input into the formulation of PRSP priorities and policy design. Many analysts (see for example WDM 2005) argue that the Bank’s influence in PRSP formulation is so intrusive that it is not really possible to talk about domestic (government or national) ownership of PRSPs: “The homogeneity across PRSPs in widely differing countries, and the dearth of alternative policy approaches on these key economic issues, suggests that ownership of the economic policies in such countries is still a pipedream.” [32]

Since it is impossible to discern the true level of ownership from PRSP and PRSC documents, it may be useful to look at whether or not the reforms are implemented. In some cases not all prior actions are implemented which might indicate weakness in ownership, although it might also indicate a lack of capacity to implement reforms. Although most of the triggers/prior actions in the PRSCs in the sample were implemented there are a few examples of triggers being only partially implemented or met later than scheduled.

§                Burkina Faso had one unmet trigger/prior action in its PRSC4 that was later fulfilled, and two triggers/prior actions were only partially implemented in PRSC5.

§                In Ghana’s PRSC2 one trigger/prior action was unmet but this was expected to be completed later.

§                In Vietnam’s PRSC3 two triggers were only ‘partially’ implemented and a third was ‘unsatisfactorily’ implemented (all three related to the privatisation of a state owned commercial bank (SOCB).

Although it is not possible to determine why prior actions were not implemented, and no explanation is given in the PRSC programme documents, it seems likely that in the case of Vietnam it may have been because the government did not want to privatise the state owned commercial bank, that is, that it did not have ownership of this reform. This supposition is informed by the World Bank’s concern that the “Government’s decision to fully integrate with the world economy, introducing competition in sectors currently dominated by General

Corporations (GCs), is not unanimously shared by stakeholders. Similarly, there is a perception that Vietnam can “grow out” of its problems in the financial sector, without frontally addressing governance and supervision issues in SOCBs.” [33] Furthermore the programme document, whilst noting the government’s commitment to reform, encourages the government to take “a more decisive stance in the banking sector.” [34]

Because PRSCs are expected to be drawn from PRSP policy matrices there is no requirement for Bank staff to consult with stakeholders outside the government, which should have been undertaken during PRSP formulation. However, efforts have been made in some cases to meet with NGOs. In one case in the sample – Vietnam - the draft PRSC programme was shared with civil society organisations: “Documentation related to the proposed credit was shared with NGOs during its preparation. With the support and cooperation of the NGO Resource Centre, a meeting was held with civil society organizations (including both international and local NGOs) at an early stage of program design. The timeframe for project preparation was presented and the draft matrix of possible prior actions was distributed and discussed. There was a discussion on the strengths and weaknesses of recent Decree 79 on Grassroots Democracy and Decree 88 on local organizations. At the suggestion of an NGO working in the health sector, the team pursued the possibility of including actions which would accelerate implementation of the Government’s National Tobacco Control Policy and the international Framework Convention on Tobacco Control. Subsequent discussion with Government officials indicated that these actions could not be achieved within the timeframe of PRSC3.” [35] The Bank’s disclosure policy requires the Bank to disclose PRSC programme documents but there is no requirement to disclose drafts or discuss these outside the government.

10.  Cross conditionality and explicit links between the PRSC and PRGF

The Bank’s guidelines state that where possible PRGFs and PRSCs should be negotiated over the same time-frame and should be implemented in parallel. A prior condition for accessing a PRSC is that the country has an adequate macroeconomic environment, which is understood to mean that a government should be on track with its PRGF programme. In the case that a country does not have a programme then IMF staff are invited to assess the country’s macro position and outline any concerns in an assessment letter to Bank staff. [36]

Surprisingly only a little more than a third of the programmes in the study sample (5 out of 13) explicitly included macroeconomic conditionality, although it is a selectivity criteria. In each of these 5 countries the PRSC programme included at least one prior action related to the macroeconomy. Often the prior actions are formulated vaguely for example, “maintain a stable macroeconomic framework during the program period.” [37] This is a euphemism for compliance with a Poverty Reduction and Growth Facility (PRGF) programme, since Bank staff rely on IMF PRGF reviews to determine whether a country’s macroeconomic environment is satisfactory. It is not uncommon for Bank and Fund staff to join each other for their respective programme reviews.

The Bank also draws on the IMF’s analysis when determining macroeconomic reforms and the two institutions collaborate together on analytical work in several sectors including public finance management, any sectoral reforms that have budget implications (particularly privatisations), public sector reform, and financial sector reforms, which is the basis of their policy advice to governments.

For example, the IMF has collaborated with the Bank on Rwanda’s and Nicaragua’s PRSCs to help define macroeconomic criteria. However, the links between PRSC macroeconomic conditions and PRGF programmes are often not elaborated in detail. For example, Tanzania’s PRSC programme focuses on debt and taxation both areas relevant to the IMF, but although the programme document identifies where there has been engagement between the Bank and Fund, and when the Bank has drawn on the IMF’s analysis it does little more than state that the PRSC and PRGF programmes are “complementary and mutually reinforcing.” Likewise in Nicaragua’s programme the government is required to take specific actions on its allocation of pro-poor spending within the budget, and on debt and pension reform. Although Nicaragua’s programme document states that “the budgetary implications and implied actions in the policy matrix will be consistent with the spending constraints under the PRGF” and that “Numerous consultations with the IMF on the amount, timing and content of the operation have taken place”, there is no further elaboration of how the programmes fit together. Also, in Mozambique taxation reforms are included in the programme including a prior action that demands that “revenue-raising measures [are] deployed, including automatic fuel tariff adjustments and [a] withholding tax on GOM employee incomes”, and a benchmark which requires tax revenues to be 14.1% of GDP. Although the PRSC programme document states that “it is expected that the PRSC will be aligned with the IMF’s next PRGF arrangement” at the time the PRSC programme was written there was no PRGF programme.

A concern is that the World Bank’s programmes are more likely to be aligned to the IMF’s programmes rather than vice versa. The implication is that social sector spending and poverty spending more broadly are likely to be constrained by IMF fiscal conditions imposed on governments. Certainly the PRSC programme documents make no mention as to whether PRGF programmes have been adapted in the light of priorities identified in PRSC, but rather seem to suggest the opposite.

Whilst many welcome the greater coherence between IMF and Bank programmes that greater collaboration should achieve, some NGOs have expressed concern that governments will be increasingly boxed into IFI-advocated reforms, and there is some evidence of this.

§                In Guyana’s PRSC1 programme prior actions stipulate that the government must “agree with the IMF on the total amount to be spent on public sector wages in the budget for 2002” and “complete [an] overall review of tax system and administration with IMF assistance”. [38]

§                Senegal’s PRSC matrix includes benchmarks that require the government to “restrict [the] use of extraordinary [expenses] in accordance with the IMF program” and to “implement [the] recruiting sectoral strategy [in the Ministry of Justice] agreed with the IMF.” [39] The document states that “…the implementation of fiduciary and human resources reforms in the public sector, one of [the] priority areas selected under PRSC I and PRSC 11, enhances successful implementation of the PRGF.” [40]

During the IMF’s conditionality streamlining process concerns were raised about the use of cross-conditionality, and finally it was agreed that the World Bank and the IMF could duplicate conditionality in each others’ programmes “only when a policy measure was considered critical for the success of the programs supported by both institutions. In such cases, conditionality would be harmonized, using identical dates and benchmarks to the extent possible.” [41] In the two cases mentioned above, criticality of the reforms is not elaborated, although in Senegal’s case the fact that the condition is a benchmark rather than a prior action suggests that it is not considered critical by the Bank.  

Cross conditionality also arises in relation to other institutions or processes. Conditions stipulated as prior actions for the PRSC may also be included as conditions for reaching Heavily Indebted Poor Country (HIPC) completion point. For example, in Nicaragua, the passage of a law on civil service reform by the National Assembly was a PRSC trigger and a condition for reaching HIPC completion point. [42] And Ethiopia’s programme includes a prior action demanding “submission of [an] application for accession to [the] WTO”. [43] And Vietnam’s policy matrix for its third and fourth PRSCs include a benchmark and prior action condition which require the government to “conduct detailed preparatory work for WTO accession including tariff scenarios” and “issue regulations on import licenses in line with [the] WTO.” The Bank argues that “Compliance with WTO rules, in turn, would be a powerful mechanism to lock-in reforms, especially on “behind-the-border” issues.” [44]

11.  Donor Harmonisation

PRSC programmes and conditions are also negotiated right from the early stages with other donors who may co-finance the PRSC or align their programmes with it. Piggy-backing with the PRSC allows donors to participate in policy dialogues with the government in those areas that are of particular interest to them.

§                In Benin, “For 2004-05, the AfDB has prepared a program loan in close coordination with IDA…The content and conditionalities of this program loan are aligned with the PRSP and coordinated with PRSC triggers. For instance, on public procurement reform, the AfDB and IDA teams coordinated their technical advice and agreed on a joint set of conditionalities and triggers. Close coordination, in particular on public expenditure management reform including public procurement reform will continue (AfDB supported the PERAC reform agenda with a program loan that was also disbursed in 2001) to help ensure that AfDB’s program lending is consistent with the objectives of the PRSC.” [45] Likewise the EU’s budget support to Benin is governed by performance indicators for health and education, which “are either the same or fully consistent with those adopted in the PRSC.” [46]

§                In Senegal, “although the PRSC program has not been developed along one single matrix shared by all donors, such an approach was adopted for the fiduciary reforms where all donors have endorsed and organized their support along the implementation of the CFAA/CPAR [Country Financial Accountability Assessment and Country Procurement Assessment Report] actions plans…Throughout project preparation, coordination with the European Union, the Dutch and Canadian Governments has received special attention as they are also currently preparing budgetary support operation[s].  Not only will these operations complement the PRSC, but they also send a consistent message to the authorities regarding the priorities agreed by the donor community by supporting the same priority actions, notably in fiduciary reforms and health.” [47]

From these examples it appears that donors are aligning with the PRSC rather than the PRSP. The implication is that if a government fails to implement the PRSC conditionality then it puts other donors’ lending at risk. In Uganda “Much of the budget support from development partners is contingent upon the World Bank Board’s approval of the PRSC: about US$134 million a year from the United Kingdom, Netherlands, Ireland, Norway, Sweden, and Germany is explicitly linked to the PRSC.” [48] Alignment of donor economic and sector work and conditionality has advantages from a policy coherence and transaction cost perspective, but the big disadvantage is that alignment is occurring around the World Bank’s analysis and the conditions included in the PRSC programme, which reinforces the World Bank’s dominant influence on development analysis.

In other countries donors are aligning with Performance Assessment Frameworks (PAFs) which are based on the PRSP and are jointly determined by the government and its donors:

§                In Tanzania “A common Performance Assessment Framework (PAF)/policy matrix, including the set of prior actions for PRSC-2 and triggers for the PRSC-3, has been developed in close cooperation among the government, the PRBS donors, and the World Bank.” [49] Although the PAF and the PRSC are aligned, the Bank (and the other donors) has the option to include specific benchmarks in its PRSC programme that are not included in the PAF and to form an independent judgment from other donors as to whether significant progress has been made. [50]

§                In Mozambique “In April 2004, a Memorandum of Understanding (MoU) was negotiated between the Government and the G15 [Group of 15 donors] as a basis for harmonizing the G15 financing instruments, ensuring predictable flows, minimizing transactions costs, and ensuring consistency of policy dialogue. As is the case for other participating donors, within the framework of the MoU, the Bank retains flexibility to determine specific benchmarks and form an independent judgment as to whether the Government has made significant progress. Nonetheless, all parties to these documents are committed to minimizing differences to the greatest extent possible, in an effort to achieve agreed efficiency objectives and greater predictability in resource flows. As an example of the Bank’s commitment to harmonization, all seven triggers for PRSC2 are explicitly included in the restricted PAF matrix agreed between the Government and its partners in April 2004.” [51]

§                And in Ghana, “In the run up to the preparation of PRSC-2, development partners carried out two joint missions (June and September 2003) that were important first steps in aligning mission schedules, advancing agreements on policy actions and on the timing of future assessment reviews. These efforts led to a better alignment of programs for 2004, with an almost complete overlap between key policy actions agreed under the MDBS [Multi-donor Budget Support] 2004 and those envisioned under the proposed PRSC-2. The development partners also agreed with the Government of Ghana on a joint matrix of policy actions to be supported by PRSC-3 and MDBS-2005, aligning monitoring and evaluation frameworks.” (Ghana PRSC2 p17)

However, it is not clear whether PRSC matrices are aligned to PAFs or whether PAFs are aligned to the PRSC matrices. The quote from the Bank’s PRSC programme in Mozambique above suggests that it is the latter. This is problematic since it again implies that it is the PRSC around which donors are aligning rather than the PRSP.

While there has been some reduction in the level of conditionality and a shift in content, there is still significant continuity with the World Bank’s traditional approach to conditionality and the policy prescriptions demanded.  Continuity is clearly demonstrated in the high level of overall conditions, the lack of space for policy alternatives and the exclusion of Parliaments and civil society from inputting into the PRSC process. 

The reduction in binding conditions is obviously welcome but the high and apparently rising number of benchmarks raises serious questions.  Major concerns are that the World Bank still wishes to micro manage low-income countries and the lack of clarity over the status of benchmarks causes confusion, particularly in relation to aid predictability. 

The continuation of privatisation conditions in spite of the recognition of their political sensitivity coupled with the lack of evidence that privatisation leads to poverty reduction, suggests that the World Bank is still wedded to a particular policy approach.  It is a grave concern that PSIAs were carried out on relatively few of the privatisation conditions, which suggests that the Bank is not paying due attention to the potentially significant negative poverty impacts of privatisation.     

The lack of transparency on the PRSC process and the lack of a formal requirement to consult raise serious questions about the legitimacy of the PRSC. The World Bank’s argument that there is no need for consultation on the PRSC as it is based on the country’s participatory PRSP, does not stand up to scrutiny. The findings of the OED PRSP study shows that World Bank programmes have not changed substantially in response to the PRSP, undermine the World Bank’s case.  Following the example of Vietnam where the draft PRSC programme was discussed with civil society, World Bank management should make it a formal Bank policy to publicly release the PRSC in draft format before it is finally agreed with the government. As in the case of Vietnam, the PRSC document should detail the focus of discussions with civil society and whether or not their recommendations and suggestions were included.

A government’s desire to access the fast disbursing budget support that the PRSC provides is also used by the World Bank as an incentive to persuade the government to implement additional reforms (although these may be included in the PRSC) identified in the Country Assistance Strategy.  This moves the government up into a higher lending scenario. The incentive is achieved by including subsequent PRSCs in the high-case lending scenario which rewards government compliance with conditionality. 

Beyond the World Bank, the PRSC is part of the tight web of mutually reinforcing conditionality within which low-income countries are caught.  There is cross conditionality between the PRSC and the PRGF, between the PRSC and HIPC completion point and , casting the web even wider, between the PRSC and the WTO.  The web is tightened even further when other bilateral, regional and multilateral donors base their programmes on the PRSC matrix. While Debt and Development Coalition Ireland supports donor harmonisation, it is a concern that donors appear to be harmonising around the PRSC rather than the PRSP. Furthermore, it is problematic that in some cases the PRSC appears to be informing Performance Assessment Frameworks rather than vice versa.  In relation to Bank and IMF harmonisation, the IMF and World Bank programmes should be aligned with the poverty reduction goals of the PRSP

Finally, it was beyond the scope of this study to examine in detail other areas of PRSC conditionality.  Given the centrality of good governance and public finance management conditions to the PRSC, this is an area which needs detailed examination.  While all multilateral and bilateral funders give high priority to this area, there is surprisingly little agreement on how to determine minimum standards, and this gives rise to concerns that developing countries could be excluded from receiving funding, if the bar is set too high.

13.  Recommendations

In the lights of the findings we would make the following recommendations, to the World Bank and other multilateral, regional and bilateral funders:

1.     Policy Conditions: PRSCs should cease to include privatisation and other economic policy conditions.

2.     Criticality: World Bank should make clear the criteria on which critical conditions are based.  Prior actions must be properly explained and open to debate and challenge.

3.     Consultation/transparency:  PRSCs should be made public in draft form.  Both Parliaments and civil society groups must have an opportunity to input into PRSCs before they are agreed.

4.     Cross conditionality: there should be no cross conditionality with IMF or WTO programmes.

5.     Donor alignment: Other multilateral, regional and bilateral donors should align with the PRSP and not the PRSC and PAFs should be drawn from the PRSP not the PRSC.

6.     PSIAs: assessments carried out on key elements of PRSCs should factor in consideration of alternative approaches.

--------------------------------------------------------------------------

Print Version