Experience with World Bank Conditionality Stefan Koeberle and Thaddeus Malesa 1. Summary. This note summarizes key trends in the application of conditionality in World Bank policy-based lending since fiscal year 198. Many of the findings indicate that over time the World Bank has adopted a number of lessons learned and incorporated them into business practices. The main conclusions of the analysis and questions going forward are: The volume of adjustment lending has stabilized at around 3% of World Bank lending in the past decade; The average number of conditions has decreased; Loans with fewer conditions are generally of higher quality; Single tranche adjustment operations are usually of higher quality; World Bank adjustment lending is selective, favoring better performing countries; The average number of conditions is lower in better performing countries; and The focus of conditionality has changed from short-term reforms, to longer-term and more complex issues; Given the experience with conditionality, should the World Bank continue to use it in its present form? And Programmatic lending emerges as a fairly robust approach to reconcile numerous tensions and concerns. 2. The volume of adjustment lending has stabilized at around 3% of World Bank lending in the past decade. Adjustment operations have accounted for -2% of all Bank lending operations (see Figure 1) in recent years, with a far smaller share in the early 198s. There has been a greater fluctuation in the share of lending volumes, with adjustment lending commitments representing around one third of the World Bank portfolio on average (see Figure 2). The number of adjustment operations approved to date has been almost evenly split between IBRD and IDA countries (see Figure 3), but the share of commitments to middle -income countries (IBRD) has been far larger (see Figure 4). This is also reflected in regional lending comparisons: the Africa region accounts for 34% of adjustment operations, but only 16% of commitments (see Figure ). 1
Figure 1 3 Number of World Bank Lending Operations, FY8-3 Investment Lending Adjustment Lending 2 Number of Operations 1 FY8 FY81 FY82 FY83 FY84 FY8 FY86 FY87 FY88 FY89 FY9 FY91 FY92 FY93 FY94 FY9 FY96 FY97 FY98 FY99 FY FY1 FY2 FY3 Figure 2 World Bank Lending Commitment in US$ billion, FY8-3 3 2 US$ 1 billion Investment Lending Adjustment Lending FY8 FY81 FY82 FY83 FY84 FY8 FY86 FY87 FY88 FY89 FY9 FY91 FY92 FY93 FY94 FY9 FY96 FY97 FY98 FY99 FY FY1 FY2 FY3 2
Figure 3 Number of World Bank Adjustment Operations, FY8-3 4 4 IDA IBRD 3 Number of Operations 3 2 1 FY8 FY81 FY82 FY83 FY84 FY8 FY86 FY87 FY88 FY89 FY9 FY91 FY92 FY93 FY94 FY9 FY96 FY97 FY98 FY99 FY FY1 FY2 FY3 Figure 4 16 14 World Bank Adjustment Lending Commitments in US$ billion, FY8-3 IDA IBRD 12 US$ billion 8 6 4 2 FY8 FY81 FY82 FY83 FY84 FY8 FY86 FY87 FY88 FY89 FY9 FY91 FY92 FY93 FY94 FY9 FY96 FY97 FY98 FY99 FY FY1 FY2 FY3 3
Figure Regional Share of Adjustment Lending by Number of Operations, FY8-3 MNA 6% LCR 24% AFR 34% Regional Share of Adjustment Lending by Volumes, FY8-3 LCR 34% AFR 16% EAP 1% ECA 22% SAR 6% EAP 8% MNA % ECA 24% SAR 6% 3. The average number of conditions has decreased. The average number of conditions 1 fell from above 3 in the late 198s to below 18 in FY3 (see Figure 6), with the exception of FY92-93 where the average number of conditions was over 4. If non-binding conditions or desired actions are included, 2 the total number has fallen from around 6 to 42 during the same time somewhat above previous years averages (34 in FY1 and 27 in FY 2). Furthermore, the average number of conditions has been decreasing across all types of adjustment lending 3 used by the Bank in recent years (see Figure 7). While the average number of conditions has decreased across all Bank adjustment loans, the number of conditions in IBRD and IDA countries has been converging (see Figure 8). 1 Legally binding conditions are prior actions preceding Board presentation, effectiveness conditions, and conditions for tranche release set out in the legal agreements signed by the borrowing government and the World Bank. 2 Non-binding conditions are other desired actions included in the loan documentation but not included in the legal agreement. Client government compliance with these actions is not tied to the release of funds. 3 The different types of adjustment lending are grouped as follows: SAL & SSAL Structural Adjustment Loans and Special Structural Adjustment Loans; SECAL Sectoral Adjustment Loans; PRSC & PSAL Poverty Reduction Support Credits and Programmatic Structural Adjustment Loans. 4
Number of Conditions per Loan/ Credit 7 6 4 3 2 Figure 6 Average Number of Conditions and Desired Actions per WB Adjustment Loan/Credit, FY8-3 Desired Actions Conditions FY8 FY81 FY82 FY83 FY84 FY8 FY86 FY87 FY88 FY89 FY9 FY91 FY92 FY93 FY94 FY9 FY96 FY97 FY98 FY99 FY FY1 FY2 FY3 Number of Conditions per Loan/ Credit Figure 7 Average Number of Conditions per WB Adjustment Loan/Credit by Different Types of Adjustment Lending, FY8-3 4 3 3 2 1 34 3 24 22 23 23 18 FY8-89 FY-2 24 FY9-99 FY3 14 SAL & SSAL SECAL PRSC & PSAL
4 4 3 3 2 1 Figure 8 Average Number of Legally-binding Conditions by Agreement Type, FY9-3 IBRD IDA 199 1996 1997 1998 1999 2 21 22 23 4. Loans with fewer conditions are generally of higher quality. In general, quality of adjustment operations, as judged by the independent Operations Evaluation Department, has markedly improved in the last decade from 69% of operations with satisfactory ratings exiting in FY92, to 81% of operations with satisfactory ratings exiting in FY3 (see Figure 9). A lower number of legally binding conditions is generally associated with higher quality (see Figure ). Figure 9 OED Quality-at-Exit Ratings of WB Adjustment Lending Operations, Exit FY9-3 Percent of satisfactory operations according to OED outcome ratings 9 8 7 6 4 3 2 86 81 84 82 83 81 76 7 69 72 74 72 63 64 FY9 FY91 FY92 FY93 FY94 FY9 FY96 FY97 FY98 FY99 FY FY1 FY2 FY3 6
Figure Percent of satisfactory operations according to OED outcome ratings Average OED Quality-at-Exit Ratings of WB Adjustment Lending Operations by Number of Conditions, Weighted by Operations, Exit FY8-3 9 8 7 6 4 3 2 66 61 82 9 FY8-89 FY9-94 FY9-98 FY99-3 7 77 78 1-1 16-44 4 or more Number of Conditions 6 6 77 67. Single tranche adjustment operations are usually of higher quality (see Figure 11). This was not affected by a significant increase in the share of single tranche adjustment operations in the overall portfolio (see Figure 12). Percent of satisfactory operations according to OED ratings Figure 11 OED Ratings of WB Adjustment Lending Operations by Number of Tranches, Exit FY9-3 % 9% 8% 7% 6% % 4% 3% 2% % % 88% 79% 69% 61% Single tranche Multiple tranche 37% 44% Outcome Sustainability Institutional Development Impact 7
Figure 12 Share of single tranche adjustment operations Single Tranche WB Adjustment Operations as a Share of Total WB Adjustment Operations, FY9-3 % 4% 4% 3% 3% % 2% 1% % % % 19% 31% 4% FY9-94 FY9-98 FY99-3 6. World Bank adjustment lending is selective, favoring better performing countries. The majority of adjustment lending commitments go to high and good performing country groups 4, according to the countries CPIA rating (see Figure 13). The CPIA Rating The CPIA rating is carried out once a year. It s main purpose is to assesses the quality of a country s current policy and institutional framework on the basis of observable policies, not on the amount of improvement since the previous yearly exercise, nor on intentions for future change, unless the latter are virtually in place. Quality means how conducive that framework is to fostering poverty reduction and sustaining growth and the effective use of development assistance. Bank staff follow specific guidelines and rate countries based on these guidelines and to benchmark countries in each region. Bank staff assess and rate 2 policy and institutional performance areas every year, which are then grouped and averaged into the four clusters: (a) economic management, (b) social inclusion and equity, (c) public sector management and institutions, and (d) structural reform policies. The CPIA rating (a numerical value between 1 and 6) is then discussed and agreed to by Regional management. Some aspects of the CPIA ratings are disclosed. CPIA ratings for 199 existed only for IDA countries, IBRD ratings for 199 were staff estimates. CPIA ratings for 199-1997 were rescaled from a 1- scale to 1-6 scale. 4 The unofficial groupings of high, good and poor performers were made using CPIA (Country Policy and Institutional Assessment) ratings over the period between fiscal years 1996 and 23. A country with a CPIA rating of 4. and above are assigned to the high performing group, those countries given a rating between 3. and 3.99 are assigned to the good performing group, and the countries with an overall CPIA rating 2.99 and below are included in the poor performing group. These groupings have been made exclusively for this exercise, and are not in line with any official World Bank classifications. 8
Figure 13 Average Annual Adjustment Lending Commitments by Country's CPIA Rating, in US$ billion US$ billion 9 8 7 6 4 3 2 1 Poor Performing Country Group Good Performing Country Group High Performing Country Group.2 2..7 1. 9. Avg. FY96-97 Avg. FY98-99 Avg. FY-3 2.7.4 4.7 1.4 7. The average number of conditions is lower in better performing countries. On average for FY98-3, adjustment loans in very high performing countries (CPIA 4. or more) contain an average of 23 legally binding conditions, whereas for very low performing countries (CPIA 2.49 or less) the average number of conditions is 47. Operations in countries with higher overall CPIA ratings have higher average outcome ratings. 8. The focus of conditionality has changed from short-term reforms, to longer-term and more complex issues. The focus of overall conditionality has also changed: from an initial focus on short-term macroeconomic and trade-related reforms, to longer-term and more complex social sector and public sector management issues (see Figures 14 and 1). The change in concentration of policy-based lending is also reflected in the sharp decline of supported public enterprise restructuring and privatization reforms in recent years (see Figure 16). 9
Figure 14 Share of Conditions by Sectors in % 4 4 3 3 2 1 Share of Conditions Associated with WB Adjustment Loans/Credits by Sectors, FY8-3 38 31 Industry & Trade 3 23 18 1 16 Agriculture & Infrastructure 4 11 16 2 Social Sector 18 1 12 FY8-89 FY9-98 29 Public Sector Management FY9-94 FY99-3 24 2 19 16 Finance Sectors Sectors are high-level, mutually exclusive grouping of economic activities based on the types of goods or services produced. Sectors are used to indicate which part of the economy is supported by the WB intervention. For our analysis we classified the sectors into categories: Industry & Trade - Industry & Trade Sector Group; Agriculture & Infrastructure - Agriculture, Fishing and Forestry Sector Group, Information and Communications Sector Group, Energy and Mining Sector Group, Transportation Sector Group and Water, Sanitation and Flood Protection Sector Group; Social Sector - Education Sector Group, Health and Other Social Services Sector Group, Compulsory Pension and Unemployment Insurance Sector and Health Insurance Sector; Public Sector Management - Central Government Administration Sector, Law and Justice Sector, Sub-national Government Administration Sector and General Public Administration Sector; and Finance - Banking Sector, Capital Markets Sector, Housing Finance and Real Estate Markets Sector, Non-compulsory Pensions, Insurance and contractual savings Sector, Payment Systems, Securities Clearance and Settlement Sector and General Finance Sector, Micro- and SME Finance Sector.
Share of Conditions by Themes in % 4 4 3 3 2 1 Figure 1 Share of Conditions Associated with WB Adjustment Loans/Credits by Themes, FY8-3 27 FY8-89 FY9-98 21 17 Trade & Economic Management FY9-94 FY99-3 21 11 6 3 Agriculture & Infrastructure 6 23 24 24 18 17 14 Social Sector & Environment 28 Public Sector Management 22 31 42 3 Finance & Private Sector Themes Themes are goals/objectives of WB activities.?themes are used to capture WB support to the Millennium Development Goals.?For our analysis we classified the themes into categories: Trade & Economic Management - Trade and Integration Theme Group and Economic Management Theme Group; Agriculture & Infrastructure - Infrastructure Services for Private Sector Development Theme, Access to Urban Services for the Poor Theme, Other Urban Development Theme, Rural Non-farm Income Generation Theme, Rural Policies and Institutions Theme, Rural Services and Infrastructure Theme and Other Rural Development Theme; Social Sector & Environment - Social Protection and Risk Management Theme Group, Social Development, Gender and Inclusion Theme Group, Human Development Theme Group and Environment and Natural Resources Management Theme Group; Public Sector Management - Public Sector Governance Theme Group, Rule of Law Theme Group and Municipal Governance and Institutional Building; and Finance & Private Sector - Corporate Governance Theme, Regulation and Competition Policy Theme, Small and Medium Enterprise Support Theme, Standards and Financial Reporting Theme, State Enterprise/Bank Restructuring and Privatization Theme, Other Financial and Private Sector Development Theme, Rural Markets Theme and Municipal Finance Theme 11
Figure 16 % 9% 8% Loans with State enterprise/bank Restructuring and Privatization Conditionality, FY8-3 Share of conditions Share of loans 96% 7% 6% % 4% 3% 2% % % conditions 26% 1% 4 conditions % 198 1981 1982 1983 1984 198 1986 1987 1988 1989 199 1991 1992 1993 1994 199 1996 1997 1998 1999 2 21 22 23 9. Given the experience with conditionality, should the World Bank continue to use it in its present form? While ex ante conditionality has often been criticized as corrosive and ineffective, leading to volatility of resource flows, straining the donor-recipient relationship and undermining the respective country s sovereignty, it is still a necessity. If for no other reason, donors have a fiduciary responsibility to exercise due diligence while spending their taxpayers money. However, alternative approaches such as ex post conditionality and outcome-based conditionality have been under consideration as possible ways forward. Yet such approaches are though to be fraught with practical difficulties including the identification of meaningful intermediate indicators on which to base disbursements, and the considerable time lag resulting from the collection and measurement of results indicators in often challenging circumstances.. Programmatic lending emerges as a fairly robust approach to reconcile numerous tensions and concerns. Many concerns remain with outcome-based adjustment lending such as the diminishing predictability of available aid resources for the client, as well as having a country s aid program tied to indicators that are not necessarily under their control. Programmatic lending on the other hand, emerges as a promising way to reconcile focus on performance with resource predictability by having support be reflected by a country s track record while supporting sustained engagement for a prudent mix of medium-term institutional and policy reforms. Not only is the lending product able to assist the implementation of complex reforms, but it also serves as a vehicle for policy dialogue that also typically involves transfer of advice and knowledge. 12