VI Key Feature 5: More Selective Structural Conditionality
- Benedict Clements, Gabriela Inchauste, Nita Thacker, Thomas Dorsey, Shamsuddin Tareq, Emanuele Baldacci, Sanjeev Gupta, and Mark Plant
- Published Date:
- September 2002
This key feature sets out the expectation that conditionality in PRGF-supported programs should focus on the IMF’s core areas of expertise and limit conditionality to key measures for the program.
Limit Conditionality to Key Measures
Considerable progress has been made in streamlining structural conditionality on the IMF’s core areas of expertise. The mandate in this area is that measures specified in the PRGF-supported programs should cover only those areas where the IMF has primary responsibility (and in these areas conditionality should be used parsimoniously). The only exception to this rule would be where a structural measure has such a direct, critical macroeconomic impact that the PRGF-supported program would be derailed unless the measure was implemented.31 Substantial progress has been made in this regard.
This streamlining of conditionality reflects the changing roles and modalities of collaboration in the PRSP framework between the IMF and the World Bank. In July 2001, the IMF and World Bank Executive Boards endorsed a strategy to strengthen collaboration between the two institutions on conditionality. Based on principles agreed to in 1998, the key features in the strengthened framework are clarity about responsibility, early and effective consultation, and separate accountability. The Boards also agreed that, to further clarify the delineation of responsibilities, it would be useful to adopt the practice of identifying one institution as the “lead agency” responsible for designing and monitoring conditionality in each policy area. In the future, Board documents should transparently and systematically set out the staff views of the lead agency on various issues and conditionality as an input to Board discussions, including more effective use of Bank inputs to Article IV consultations and IMF inputs in the context of the Bank’s country assistance strategies, economic and sector work, and programmatic lending.32
Overall, PRGF-supported programs have shown a strong shift toward more streamlined conditionality.33 The average number of structural conditions (performance criteria and prior actions) has declined from more than eight in the most recent ESAF-supported annual programs approved before December 1999 to six for programs approved since July 2000 (including new PRGF-supported programs and continuation of programs approved before 2000). When the scope of measures is broadened to include structural benchmarks, the average number of structural measures has declined from 16.9 in the last ESAF-supported annual programs approved before December 1999 to 11.8 for programs approved since July 2000 (Table 5). Moreover, implementation of the guidelines has been increasingly consistent: the average number of structural measures has declined from 13 for programs approved in 2000 to fewer than 11 for those approved in 2001. This reduction in the number of structural measures has been uneven across programs, though, as indicated by a relatively large, although declining, standard deviation (falling from 8.1 for programs approved in 2000 to 5.2 for those approved in 2001).
|ESAF Arrangements||PRGF Arrangements|
|Total conditions and benchmarks||16.9||11.8|
|Average core conditions and benchmarks||7.0||6.9|
|Share of measures in core||41.0||58.0|
|Average shared conditions and benchmarks||3.7||2.3|
|Share of measures in shared areas||22.0||20.0|
|Average non-core conditions and benchmarks||6.2||2.5|
|Share of measures in non-core areas||34.6||21.9|
Streamlining is evident across forms of conditionality and other program measures (Figure 9). The reduction in conditionality was particularly significant for prior actions and benchmarks, whose number on average was cut by 47 percent to three and six conditions per program, respectively. The average number of performance criteria, already moderate under ESAF-supported programs, has declined further to fewer than three measures per program. In some programs, streamlining benefited from efforts to limit conditions that specify multiple steps to achieve the same objective, such as the introduction of a VAT or the design and implementation of civil service reform.
Figure 9.Streamlining Conditionality in PRGF-Supported Programs
Sources: National authorities, country policy intentions documents; and IMF staff estimates.
Confine Conditionality to Measures in the IMF’s Domain
Conditionality has been increasingly focused on the IMF’s core areas of expertise (Box 4).34 In PRGF-supported programs approved since July 2000, 58 percent of structural measures focused on the IMF’s core areas, compared with 41 percent in earlier programs. The proportion of measures in core areas has been even larger for new PRGF-supported programs approved since July 2000 (68 percent), compared with transformed PRGF-supported programs (52 percent). The number of conditions in core areas has continued largely unchanged in absolute terms, while those in areas shared with the World Bank and in areas outside the IMF’s core responsibilities have fallen by roughly one-third and three-fifths, respectively (Table 5).
The concentration of measures under PRGF-supported programs shows a clear shift to measures related to fiscal policy (Figure 10). Measures in the areas of fiscal management and fiscal transparency rose from 15 percent of structural measures in the ESAF subsample to 31 percent in the PRGF subsample (see Section VII). Tax policy and administration measures roughly maintained the same share in both the subsamples (16 to 18 percent). The concentration of structural conditionality in the fiscal area is consistent with the renewed emphasis on poverty reduction in the PRGF and the related need to maintain revenue and ensure that fiscal resources are spent efficiently for appropriate purposes.
Figure 10.Composition of Structural conditionality in PRGF-Supported Programs
Sources: National authorities, country policy intentions documents; and IMF staff estimates.
For areas outside the IMF’s domain, conditionality in most recent programs has generally been applied when such measures were critical to the country’s fiscal and/or external targets (Box 5). In such cases most staff reports provided justification for their inclusion in the program. Justification was clear in 60 percent of all PRGF-supported program reports and in 73 percent of reports for new PRGF-supported programs approved since July 1, 2000. For example, the importance of ensuring a healthy banking sector, and thus containing the fiscal cost of restructuring, was particularly clearly presented in the case of Vietnam. Financial sector reforms have also taken center stage in Mozambique’s PRGF-supported programs, and a clear case has been made in the relevant staff report. In countries where governance is a critical issue, such as Kenya and Cameroon, programs have also focused on fiscal transparency as well as governance-related issues not in the IMF’s core area of expertise.
Box 4.Structural conditionality and Areas of IMF Expertise
|Tax policy||Financial sector reforms||Public enterprise reforms|
|Fiscal transparency||Trade policy||Privatization|
|Fiscal management||Private sector promotion||Marketing and pricing reforms|
|Monetary policy||Civil service restructuring|
|Foreign exchange regime||Social safety nets|
|Exchange rate policy||Monitoring poverty reduction|
|Macroeconomic data||Sectoral policies|
|Tax and customs administration|
The streamlining of structural conditionality is being complemented with a deepened collaboration with the World Bank. With the IMF focusing its conditionality on its core areas of expertise, conditionality in noncore areas has increasingly been picked up by the World Bank. In more than two-thirds of such cases where the IMF retained conditionality in non-core areas (mainly because these were critical to reach the macroeconomic objectives of the program), this has been supplemented by the World Bank, either through its own conditionality in its lending programs, or by providing policy advice in cases where an appropriate lending program was absent. In the subsample of ESAF-supported programs, this was true in only about half of all cases. In contrast, in areas where the IMF and the Bank share responsibility, efforts have been directed more toward avoiding duplication. In these areas, the share of IMF conditionality that has been supplemented by World Bank conditionality/advice has remained at around 50 percent.35 These shares may increase as more countries are covered by the World Bank’s Poverty Reduction Support Credit, simplifying coordination across instruments.
External comments on streamlining structural conditionality broadly concurred with the factual assessment set out above, but views were mixed as to the implications and advisability of this shift. Some donors and NGOs expressed the view that streamlining structural conditionality is an important element in strengthening program ownership, but also thought there was a risk of serious gaps in program monitoring between the Bank and the IMF. Other donors and NGOs (including some of the same cited immediately above) also expressed a concern that if structural conditionality is added by the Bank to the same extent that it is dropped by the IMF, there will be no net change in conditionality. Still other NGOs thought that notwithstanding streamlining to date, IMF conditionality was still too extensive.36
Box 5.Structural Conditionality Boxes—What Story Do They Tell?
Starting in May 2001, staff reports for PRGF-supported program requests and reviews have routinely included a box on structural conditionality. These boxes follow a standard format specifying the status of topics covered by conditionality previously in IMF-supported programs, the structural conditionality in the current program, areas now covered by World Bank conditionality or conditionality of other institutions, and areas that are no longer covered in Bank or IMF conditionality.
One clear trend seen from these boxes is the phasing out of structural conditionality related to privatization. Whereas privatization measures used to constitute a significant part of structural conditionality in many PRGF-supported arrangements, recent programs include no such conditionality, except where it had macroeconomic importance (e.g., Ghana and Tajikistan). In most cases the structural agenda for privatization has been taken over by the World Bank.
The streamlining of structural conditionality also affects other areas. In the programs with Burkina Faso and Mali, IMF conditionality in the energy and telecommunications sectors has been phased out, as has conditionality relating to drafting of procurement and investment codes in the program with The Gambia. In all of these cases, conditionality has been taken over by the World Bank. The streamlining efforts have also extended to the area of the financial system (an area with shared responsibilities between the IMF and the World Bank); in most of these cases (e.g., Mali, Mongolia, and Niger) the World Bank has also taken over the conditionality from the IMF.
These efforts at streamlining structural conditionality have been pursued flexibly. In cases where measures have been deemed to be essential on macroeconomic grounds, conditionality has been retained. For example, the programs with Azerbaijan, Cameroon, and Ghana include conditionality in the energy sector. In the cases of Azerbaijan and Ghana, domestic arrears buildup or domestic debt buildup and inappropriate pricing policies are posing serious risks to macroeconomic stability, whereas in the case of Cameroon, the conditionality—related to the formulation of a reform strategy for the oil sector—is viewed as essential for enhancing the economy’s growth prospects. Other cases where sector-specific problems have important macroeconomic repercussions are Cambodia (logging sector), Mali (cotton), and Senegal (groundnuts).