IMF Policy Paper: Social Safeguards and Program Design in PRGT and PSI-Supported Programs

The Fund provides considerable support to low-income countries (LICs).


The Fund provides considerable support to low-income countries (LICs).

Introduction and Context

1. The Fund provides considerable support to low income countries (LICs). All Fund members benefit from Fund surveillance, access to financial resources through the General Resources Account (GRA), and access to technical assistance (TA). In addition, 70 Fund members, mostly LICs, also have access to concessional financing through the Poverty Reduction and Growth Trust (PRGT) (see Annex VIII). PRGT financing is made available on more favorable terms than are available under the GRA: specifically, it carries lower interest rates (currently set at zero percent) and longer maturities and grace periods (Box 1). PRGT-eligible members may make use of the Policy Support Instrument (PSI), a non-financial instrument designed to facilitate a close policy dialogue between the IMF and the member. Since the establishment of the current concessional financing architecture in 2010, 43 Fund members have benefited from 90 PRGT- and PSI-supported programs.1 The latter represented over 60 percent of Fund-supported programs during this period.

PRGT: Fund’s Support for LICs

The Fund has a set of tools specifically targeted at supporting LICs. This includes concessional financing using PRGT resources. The toolkit was significantly overhauled in 2009 to become more flexible and better tailored to meet the diverse needs of LICs and, since then, has continued to be fine-tuned.

PRGT-supported programs play an important role in helping low income countries (LICs). Most prominent, is the ability of PRGT-eligible members to borrow from the Fund on concessional terms. This support is provided in the context of an economic program aimed at maintaining or restoring stable and sustainable macroeconomic stability consistent with strong and durable poverty reduction and growth. Importantly, PRGT- and PSI-supported arrangements also play a role in catalyzing financial support from other donors. The types of arrangements include: 1/

  • Extended Credit Facility (ECF). The ECF provides medium and longer-term financial and policy support to meet protracted balance of payments needs. Borrowing under the ECF has a grace period of 5½ years and final maturity of 10 years. PRGT-eligible members may borrow up to 225 percent of quota on a cumulative basis, and 300 percent of quota in exceptional cases. Assistance is provided for an initial duration of three to four years, with a maximum duration of five years. A successor ECF arrangement may be approved as soon as an existing one ends (whether cancelled, expired, or terminated).

  • Standby Credit Facility (SCF). The SCF provides financial assistance to LICs with short-term balance of payments needs. It may also be used on a precautionary basis. The cumulative access limit is the same as under the ECF. The duration, however, is shorter. Its use is normally limited to two and half years out of any five years. However, if an SCF arrangement is used on a precautionary basis, where at the time of approval a member does not presently face a balance of payments need, then the time does not count toward the two and half years out of five limit. The SCF has a grace period of four years and a maturity of eight years.

  • Rapid Credit Facility (RCF). The RCF provides rapid concessional financial assistance with limited conditionality to LICs facing urgent balance of payments needs. It has lower access limits, normally 18.75 percent of quota per year (37.5 percent for certain types of exogenous shocks) and 75 percent of quota cumulatively. There are safeguards to prevent repeated use in situations where an SCF or ECF would be more appropriate. However, the RCF may be used repeatedly in some cases. This includes, for example, fragile states, which may use an RCF repeatedly as part of a strategy to build a track record for an upper credit tranche (UCT)-quality arrangement (i.e., whether authorities have the capacity and commitment to implement needed macroeconomic policies). The maturity of an RCF is the same as the ECF.

  • Policy Support Instrument (PSI). The PSI is a non-financial instrument for PRGT-eligible members that is designed to promote a close policy dialog between the IMF and the member. It supports members in maintaining or consolidating macroeconomic stability and debt sustainability, while deepening structural reforms. It also delivers clear signals to donors, creditors, and the general public about the strength of the country’s economic policies. A PSI may be approved for an initial duration of one to four years, and later extended up to a maximum of five years. Successive PSIs may be requested. It may also be used concurrently with an SCF or RCF, but not an ECF arrangement. It can also provide accelerated access to the SCF in case subsequent financial needs arise.

Interest rates for PRGT-supported loans are currently zero percent. The interest rate structure is typically reviewed every two-years and set for the following two-year period. The most recent review was in 2016 and set the rates for all Fund concessional loans under the PRGT to zero for the next two years. 2/ In 2015, the interest rate on the RCF was set permanently at zero percent.

PRGT resources are intended for the Fund’s lowest income and most vulnerable members. Currently, 70 members are PRGT-eligible. 3/ Eligibility is reviewed every two years, and the next review is scheduled for May 2017. The main criteria for assessing eligibility are that income per capita is below a threshold, which was US$1,215 in the last review (with higher threshold for small and micro-states), and the member lacks access to international financial markets.

The PRGT architecture has been periodically fine-tuned to meet the evolving needs of LICs. The architecture itself is reviewed every five years, with the next review scheduled for 2018. However, since the last review in 2013, some changes have made to better support LICs. Further reforms were approved in 2015 4/ to: (i) increase access limits and norms by 50 percent; (ii) better target concessional financing to the poorest and most vulnerable members, while also boosting access for better-positioned members through greater use of financing from the GRA; (iii) complement increased access under the RCF with a parallel increase of fast-disbursing support under the Rapid Financing Instrument (RFI) to assist all countries with urgent balance of payments needs; and (iv) to set the interest rate on RCF loans permanently at zero percent. IMF (2016) clarified that all Fund members have access to the GRA, and that PRGT-eligible members have additional access to concessional financing through the PRGT. 5/ Though, given the economic benefits, staff will continue to advise PRGT-eligible members to borrow from the PRGT up to the applicable limits before seeking GRA resources.

/1 The 2016 Handbook of IMF Facilities for Low Income Countries describes the facilities in greater detail.2/ See 2016 Review of the PRGT Interest Rate Structure.3/ See Eligibility to Use the Fund’s Resources for Concessional Financing, 2015.4/ See Financing for Development – Enhancing the Financial Safety Net for Developing Countries.5/ See Financing for Development – Enhancing the Financial Safety Net for Developing Countries—Further Considerations.

2. Poverty reduction is a core objective of PRGT and PSI-supported programs. A key goal of the Fund’s concessional architecture is to help eligible members achieve a stable and sustainable macroeconomic position consistent with strong and durable poverty reduction and growth (Box 2). In this regard, the 2009 PRGT reform called for program design in all PRGT facilities to support policies that safeguard social and other priority spending—and, whenever appropriate, to increase it. This was to be achieved by including a social and other priority spending indicative target in all PRGT- and PSI-supported programs wherever possible. In addition, if feasible and appropriate, any adverse effects of program measures on the most vulnerable should be mitigated by adopting countervailing measures (IMF 2002c, 2014a).

Poverty Reduction Strategy in PRGT and PSI-Supported Programs

Poverty reduction is a core objective of PRGT facilities architecture. The PRGT aims to assist LICs in achieving a stable and sustainable macroeconomic position consistent with strong and durable poverty reduction and growth. This box summarizes the evolution of practices with regards to poverty reduction in PRGT-supported programs.


The Fund has provided financial assistance on concessional terms to low-income countries (LICs) since the mid-1970s, first through a Trust Fund, and then through the Structural Adjustment Facility (SAF) and the Enhanced Structural Adjustment Facility (ESAF). The ESAF was replaced by the Poverty Reduction and Growth Facility (PRGF) in November 1999. The core aim of the PRGF is to arrive at policies that are more clearly focused on economic growth and poverty reduction and, as a result of better national ownership, more consistently implemented.

The development of the PRGF marked a milestone in the Fund’s struggle against poverty in the world’s poorest countries. In this process, the most important innovation was the introduction of a Poverty Reduction Strategy Paper (PRSP), which forms the basis of PRGF-supported programs. PRSPs were prepared by all LICs intending to borrow from the Fund or Bank or access debt relief under the Initiative for Heavily Indebted Poor Countries (HIPC) through a participatory process involving domestic stakeholders as well as development partners.

PRGF-supported programs emphasize more pro-poor and pro-growth budgets, and are thus expected to be featured by a reorientation of public spending towards the social sectors, basic infrastructure, or other activities that demonstrably benefit the poor; improved efficiency and targeting of spending in key sectors relevant to growth and poverty reduction; and tax reforms that simultaneously improve efficiency and equity. Subsequent reviews of PRSP and PRGF confirmed their effectiveness in boosting poverty-reducing spending. For example, staff review of the PRSP in 2002 indicated that budget allocations for pro-poor spending had increased in the full PRSP countries. 1/ A review of the PRGF in 2002 indicated that there was an increase in allocation of budgetary resources toward poverty-reducing spending, and fiscal frameworks were accommodating higher spending to support country-defined poverty reduction objectives. 2/

Current practice

In July 2009, the Executive Board agreed on reforms to tailor the Fund’s concessional lending facilities to the changing needs of LICs, and approved a new concessional financing framework. The new lending architecture includes three LIC facilities—the Extended Credit Facility (ECF) to provide flexible medium-term support; the Standby Credit Facility (SCF) to address short-term and precautionary needs; and the Rapid Credit Facility (RCF) to provide emergency fast-disbursing support. The 2009 IMF paper states that under all three facilities, social and other priority spending should be safeguarded—and, whenever appropriate, increased. This was to be achieved by including a social and other priority spending indicative target in all PRGT- and PSI-supported programs wherever possible. Meanwhile, a financing package of loan and subsidy resources to boost the concessional lending capacity was also approved.

In 2015, the World Bank decided to discontinue the PRSP requirement. Hence, the Fund reformed the PRSP approach and introduced the concept of Economic Development Document (EDD) as the new documentation requirement for members’ poverty reduction strategies, which remain the cornerstone for the Fund’s concessional financing. The EDD is prepared by the member and must describe its poverty reduction strategy. It may be either: (i) an existing national development plan or strategy document; or (ii) a newly prepared document. EDDs must meet the following minimum standards to ensure a clear strategic platform for policies supported under an ECF arrangement or PSI:

  • Include a strategy or plan describing how poverty reduction and growth objectives are to be achieved.

  • Articulate specific policies, including macroeconomic and financial policies, that would be pursued in the implementation of the strategy.

  • Specify the launch date and timeframe for implementation to ensure consistency with the strategy underlying IMF-supported programs.

  • In addition to these minimum standards, countries are strongly encouraged to prepare EDDs through a broad-based participatory process and conform to good practice guidelines. 3/

  • Note whether there was a participatory process and if there was one, should also note the nature of such process.

  • To complement the issuance of a new EDD, World Bank staff provides an assessment letter to the IMF Executive Board identifying strengths and risks of the member’s PRS and priority areas for attention during implementation. An EDD is required for ECF arrangements and PSIs. The PRS covered in the EDD is normally developed on a rolling five-year cycle.

/1 Review of the Poverty Reduction Strategy Paper (PRSP) Approach: Early Experience with Interim PRSPs and Full PRSPs, IMF (March 2002).2/ Review of the Poverty Reduction and Growth Facility: Issues and Options, IMF (March 2002).3/ Reform of the Fund’s Policy on Poverty Reduction Strategies in Fund Engagement with Low-income Countries—Proposals, IMF (July 2015).

3. In this paper, all measures aimed at safeguarding social spending and protecting the most vulnerable are referred to as “social safeguards.” Social safeguards include (i) minimum floors for social and other priority spending, typically established using indicative targets (IT),2 and (ii) specific reform measures designed to protect vulnerable groups, sometimes established as prior actions or structural benchmarks under Fund-supported programs. In the context of spending floors, social spending is generally defined to include spending on health, education and social safety nets.3 Regarding specific reform measures, these are, for example, measures that would seek to improve the targeting, tracking and monitoring of such spending and strengthen the social safety nets (e.g., increase social transfers to the poor). Vulnerable groups are defined in a country context and would include, for example, the poor, elderly, the youth and women. Fund policy does not call for the establishment of any other form of conditionality on social safeguards other than the social and other priority spending target, but such conditions can be established as needed and in accordance with the Guidelines on Conditionality.

4. This paper examines the use of social safeguards in PRGT- and PSI-supported programs. For purposes of brevity in this paper, from here on the paper refers to PRGT and PSI-supported programs as LIC programs, recognizing that the preponderance of PRGT-eligible countries are low-income countries (Annex VIII). That said, some middle-income small states also qualify to use PRGT resources and the PSI, and low-income countries also qualify for non-concessional financing using GRA resources: indeed, for some there is a presumption that PRGT financing will be blended with GRA resources. Thus, use of the term “LIC program" does not imply that PRGT facilities and the PSI are only, or exclusively used by low-income countries.

5. The paper seeks to identify good practices in the design and use of social safeguards.4 This covers the appropriate use of quantitative targets and structural benchmarks in program design, as well as the adequacy of the documentation of the safeguards in Board papers. The paper examines whether programmed measures were implemented (such as whether spending floors were met), but does not examine the broader question as to whether these social safeguards were successful in improving social outcomes. The latter is a major task beyond the scope of this paper and would require a substantially greater research effort, drawing also on specialist expertise from other institutions.5 Similarly, the paper does not look at the broader strategies adopted in LIC programs for reducing poverty. Since the focus is on LIC program design, the paper does not systematically cover the use of social safeguards in GRA-supported programs, nor does it cover Fund engagement on social safeguards through technical assistance or in Fund-surveillance.

6. The paper is organized as follows. The following section examines the fiscal context for the use of social safeguards in LIC programs. A subsequent section summarizes experience with social safeguards conditionality in LIC programs. Possible areas for improvement are then outlined followed by the concluding section. Annexes IV discuss good practice examples of social safeguard design, based on a review of program documents.

Fiscal and Social Spending Goals

7. This section discusses the context for considering social safeguards in LIC programs. It describes goals in recent LIC programs for fiscal consolidation and for public spending, including on health and education. These are important considerations in assessing the need for social safeguards to protect vulnerable groups.

8. LIC programs are broadly divided between those entailing fiscal expansion and those seeking fiscal consolidation. This assessment is based on stated goals for the fiscal balance and spending as described in the staff report for the program request.6 The review here covers 68 such staff reports issued since 2010. Overall, 50 percent of LIC programs involved fiscal consolidation, 43 percent involved fiscal expansion, and 7 percent were fiscally neutral (Figure 1). Staff reports offered various justifications for the fiscal expansion, but the most prevalent was to scale up infrastructure and social spending (Figure 2).

Figure 1.
Figure 1.

Fiscal Adjustment in LIC Programs 1/

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

1/Number of LIC program requests between January 2010 and May 2016.
Figure 2.
Figure 2.

Reasons for Fiscal Expansion in LIC Programs 1/

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

1/ Number of LIC program requests between January 2010 and May 2016.

A. WEO Forecasts: Fiscal Balance and Real Spending

9. Fiscal projections at the time of program approval corroborate the finding that LIC programs often do not feature fiscal consolidation. The median program aimed to achieve a consolidation (reduction in the deficit) of 1.9 percent of GDP phased over five years, while a quarter of programs encompassed a widening of the deficit (Figure 3). And, rather than cutting spending, a large majority of programs envisaged substantial increases in real expenditure, even in per capita terms. Across programs, the fiscal consolidation was to be achieved mainly through revenue mobilization efforts. The findings reflect the fact that programs have often served as a means for country authorities to signal sound policies and to mobilize external financing for additional priority spending.

Figure 3.
Figure 3.

Fiscal Projections at Approval of LIC Programs 1/

(By quartile)

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Note: Solid line depicts median projection, dashed lines show 25th/75th percentile.1/ Design of LIC programs approved during 2010-2016.

10. The typical fiscal policy design in LIC programs is markedly different from the emphasis on fiscal consolidation in many GRA-supported programs. This can be attributed in large part to the fact that a significant portion of Fund programs for LICs are associated with balance-of-payments needs linked to enhancing growth and reducing poverty rather than playing a crisis prevention and mitigation role (such as in GRA-supported programs). The average fiscal adjustment in LIC programs over five years, measured in percent of GDP, is twice as large in GRA-supported programs as in LIC programs (Figure 4). And the average GRA-supported program envisages a reduction in real per capita spending for the first three years after program approval, whereas the average LIC programs aims at a 15 percent increase over five years.

Figure 4.
Figure 4.

Average Fiscal Projections at Approval of LIC Programs 1/

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Sources: WEO; and MONA.1/ Design of LIC programs approved during 2010–2016.

B. Public Spending on Health and Education

11. Government spending on health and education has generally increased steadily since the mid-1990s across all countries. This result holds when measured as a share of overall spending or as a share of GDP (Figure 5).7 The data, which are from the World Bank, indicate that the increase in health and education spending was the highest in LICs. Health and education spending in LICs as a share of GDP has caught up with emerging economies.

Figure 5.
Figure 5.

Long-Term Trends in Public Spending on Health and Education

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Sources: World Bank; and IMF staff calculations.Note: Bars represent unweighted cross-country averages. Only countries with observations in both periods are included.1/ Constant USD, adjusted for Purchasing Power Parity.

12. In the short run, health spending is less pro-cyclical in LICs than in more advanced economies (Figure 6).8 This may reflect the fact that many LICs rely on donor support to finance health spending, which may reduce its sensitivity to fluctuations in domestic budget resources. Health spending in LICs is also less sensitive to revenue fluctuations than overall spending. Figure 6 shows that in LICs, average real expenditure growth is more than 8 percentage points higher in times of high revenue growth than in times of low revenue growth. By contrast, the average real growth rate of budget expenditure at the central government ministry of health is above 5 percent, irrespective of the pace of revenue growth.

Figure 6.
Figure 6.

Real Expenditure Growth in Times of High/Low Revenue Growth 1/

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Sources: WHO; WEO; and IMF staff calculations.1/Sample period: 1995-2015. Bars represent mean growth rates within income groups, conditional on whether revenue growth is above or below its country-specific median.

13. Regression analysis suggests that spending on health and education is protected in LIC programs. A 2013 study by Clements, Gupta, and Nozaki (2013) found that public spending on education was significantly higher in LICs implementing Fund-supported programs over the period 1985-2009. New staff analysis applying the same methodology to data for the period 1988-2014 confirms these findings (Table 1). Specifically, having a program increases education spending but the positive effect of programs on health spending is substantially smaller than the findings of the 2013 paper mentioned above and is no longer statistically significant. Nevertheless, the updated regression analysis yields no evidence that fiscal adjustment policies in LIC programs come at the expense of health and education spending.

Table 1.

Regression Results: Social Spending in LIC Programs

article image
Note: p-values in parentheses; significance levels: ***= 1%; **= 5%; *= 10%.Inverse Mills Ratios are included to correct for selection bias. For further details on the methodology, see Clements, Gupta, and Nozaki (2013).Other sources of data: World Economic Outlook, MONA.

Social Safeguards in PRGT and PSI-Supported Programs

14. This section reviews how social safeguards have been incorporated into LIC programs. The discussion focuses on experience with social and other priority spending targets and reform measures to protect vulnerable groups. In particular, it examines whether programs include social safeguards to protect vulnerable groups against fiscal adjustment or other structural measures.

15. The review focuses on the use of social safeguards in programs approved since 2010. The focus is on the period since 2010, given the introduction, as of 2009, of a Fund policy specifying that LIC programs should include an explicit program target to safeguard social and other priority spending wherever possible. The assessments draw on the results of a survey of Fund mission chiefs; a detailed review of Fund documents, especially those related to program requests; and case studies to identify good practices9 (see Annex VI for details).

A. Social and Other Priority Spending Indicative Targets

16. Social and other priority spending ITs have been incorporated into the preponderance of LIC programs. Consistent with the policy in place since 2009, 90 percent of LIC programs reviewed included an IT covering social and other priority spending. Of the 68 program request documents reviewed, only eight did not include such an IT. Of these, five cases explained that the social and other priority spending IT was excluded because of the need to first develop a better tracking system to monitor social spending;10 the other three did not offer any explanation. The targets were met in more than two-thirds of the reviews (Figure 7). Isolating reviews where the program design involved fiscal consolidation, the IT was met at a slightly higher rate (Figure 8). Hence, fiscal consolidation programs are not necessarily associated with more frequent non-implementation of the IT. Consistent with the logic of the social and other priority spending IT—to protect these items from bearing an undue burden of any fiscal adjustment—some programs took additional steps to ensure the fiscal deficit target was not met by cutting social and other priority spending. For example, the fiscal deficit target in some programs excluded social and other priority spending while others included an adjustor for social and other priority spending. Examples of this include Malawi (2010) and Grenada (2014).

Figure 7.
Figure 7.

Social and Priority Spending IT

(Percent share of total reviews for PRGT programs) 1/

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Source: IMF MONA database; and IMF staff estimates.1/ Includes PRGT and PSI-supported programs during 2010–2016.
Figure 8.
Figure 8.

Social and Priority Spending IT: Programs with Fiscal Consolidation

(Percent share of total reviews for PRGT programs) 1/

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Source: IMF MONA database and IMF staff estimates.1/ Includes PRGT and PSI-supported programs during 2010–2016

17. An examination of the cases of missed social and other priority spending ITs points to areas for improvement. It turns out that a handful of countries account for the majority of the cases where targets were missed (10 countries account for about 65 percent of misses). When missed, the targets were missed by an average margin of 20 percent. Common explanations included: (i) the need to cut spending; (ii) delays in project implementation;11 and (iii) external financing shortfalls. However, one-in-five LIC program documents did not provide any explanation of why targets were missed. A majority of program reviews in such cases did not discuss corrective measures to ensure future program targets would be met.

18. Another shortcoming is that social and other priority spending is often not effectively targeted on vulnerable groups. Country authorities take the lead in defining social and other priority spending in accordance with national poverty reduction and growth strategies (IMF, 2014b). This results in coverage that varies significantly from country to country, and often extends well beyond vulnerable groups. Often, spending floors are defined to cover total spending on health and education.12 In other cases, other priority spending includes a range of infrastructure projects, without specifying their contribution to poverty reduction. While social and other priority spending is, in principle, defined in program Technical Memoranda of Understanding (TMU), these often lack details on coverage and monitoring, and on reporting arrangements.

19. Most mission chiefs considered the social and other priority spending IT to be useful, but also recognized its shortcomings as an indicator of the quality and effectiveness of spending. Two-thirds of mission chiefs considered the IT to be useful. However, a significant majority also saw a need for a better targeted specification of the IT, focusing on the most critical sectors where the benefit could be substantial (Figure 9). As country authorities determine what constitutes social and other priority spending, seeking a more targeted specification of the IT would imply a need for closer dialogue between country authorities, Fund teams and other development partners at the program design stage. Two-thirds of the mission chiefs suggested seeking expertise from the World Bank and other development partners (DPs), given that Fund teams often lack the detailed knowledge needed for a meaningful design of the target.13 Country teams also noted that the effectiveness of these targets would improve if the quality of fiscal data could be strengthened, which would allow them to better monitor spending. About half of mission chiefs suggested including adjustors to the program target so that external financing shortfalls would no longer be a reason for missing the target.14 A few mission chiefs also mentioned that a more effective way to protect the poor may be through other policies, potentially subject to conditionality, to improve social safety nets.

Figure 9.
Figure 9.

Mission Chief Views on Social and Priority Spending Indicative Targets

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Source: Mission chief survey results. Reflects 15 responses.

B. Reforms to Protect Vulnerable Groups

20. Specific reform measures can also be utilized to help protect vulnerable groups. This can be important where economic reforms are expected to have an adverse distributional impact, where segments of the population are vulnerable, and where the social safety net system is underdeveloped. According to World Bank analysis (State of Social Safety Nets, 2015), this is common in low-income and lower-middle-income countries, where social safety nets cover only a quarter of the extreme poor, compared to 64 percent in upper-middle-income countries. The coverage gap is particularly acute in Sub-Saharan Africa and South Asia, where most of the global poor live. In these regions, only one-tenth and one-fifth of the poorest 20 percent are reached by social safety nets, respectively. The World Bank has stressed that better-coordinated systems can significantly increase the efficiency of social safety nets. They have recommended to: (i) invest in integrated systems necessitating multiple social protection programs to work together, (ii) adopt a social and beneficiary registry, and (iii) improve spending efficiency by strengthening institutional capacity, coordination, and program administration and evaluation.

21. Only a minority of LIC programs with fiscal consolidation or energy subsidy reform have included specific measures in the form of conditionality to protect vulnerable groups. Specifically, only about 15 percent of LIC programs with fiscal consolidation included such conditionality, mostly in the form of benchmarks on measures to strengthen social safety nets (Table 2).15 In addition, while a quarter of programs included conditionality on energy subsidy reforms, only 17 percent of these included an explicit program condition to mitigate its impact on the poor (Table 3). Most of these supporting conditionalities focused on improving the targeting, tracking, and monitoring of the social safety net.16 Measures to improve the safety net by increasing transfers, expanding coverage, or introducing new programs were rare.17 This rarity likely reflects the fact that many PRGT-eligible members have underdeveloped social safety net systems, limited financial resources, and implementation constraints. For comparison, GRA-supported programs more often included conditionality on the introduction of new social assistance schemes that increase transfers to groups affected by program adjustment measures (Box 3). For example, the Jordan (2012), Tunisia (2013), and Ukraine (2014) program requests included conditionality to introduce targeted assistance to protect the vulnerable from higher energy prices.

Table 2.

Fiscal Position and Social Safeguards in LIC Programs 1/

article image
Table 3.

Energy Subsidy and Social Safeguards in LIC Programs 1/

article image

Social Safeguards in GRA-Supported Programs

GRA-supported programs also include social safeguards. Indeed, reflecting in part more advanced social assistance programs and implementation capacity, GRA-supported programs often contain stronger social safeguards than LIC programs.

GRA-supported programs usually include a social and other priority spending IT if poverty reduction is considered to be a macro-critical program objective. Only one-quarter of GRA-supported programs have such an IT. 1/ Similar to LIC programs, the definition of these targets is determined by country authorities and are closely aligned with their medium-term development strategies. Compliance with program targets (about 62 percent) is also broadly similar to that observed in LIC programs. However, the reasons provided for missing the targets were less about weak implementation capacity and more due to country specific technical and administrative challenges. 2/ Similar to PRGT- and PSI-supported programs, the majority of program documents did not provide a clear discussion on corrective measures. However, the IT was better targeted as it mostly included funding to specific social assistance programs (such as conditional cash transfers, youth employment, or school feeding programs).


Social and Priority Spending Indicative Targets, 2013–2016

(Percent share of total reviews for GRA-supported programs) 1/

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Source: IMF MONA database and IMF staff estimates.1/ Includes 6 GRA-supported programs. Data based on test dates. Data downloaded January 2017.

GRA-supported programs rely more often on structural conditionality to strengthen social safety net policies. One quarter of GRA-supported programs compared to one-in-ten PRGT-supported programs had conditionality on social safety nets. Nearly half of GRA-supported programs with conditionality on energy subsidy reform were accompanied by conditionality on social safety nets in an effort to mitigate the impact of the removal of subsidies on the poor. A large majority of these social safety net measures (67 percent) focused on introducing new social assistance programs to protect the vulnerable from higher energy prices (such as Jordan (2012), Tunisia (2013), and Ukraine (2014) or using the savings from the energy subsidy reform to increase social spending (El Salvador (2010)). 3/

The clear majority of GRA-supported program documents discussed social safety net issues regardless of whether the program had conditionality in that area. Nearly ninety percent of GRA-supported program documents covered social safety net issues compared to only half in LIC programs. This underscores the room for LIC program documents to cover social safety net issues more consistently. The GRA-supported program document discussions of social safety nets also focus more on strengthening vocational and skill training to increase employment, 4/ implementing welfare-to-work exit strategies for vulnerable households, 5/ and creating a database that integrates social benefit programs across multiple government agencies with the aim of improving screening, monitoring and tracking of social benefit programs. 6/

About half of GRA-supported programs discuss support already given or pledged to country authorities on social safety net measures by various DPs. Similar to the experience in LIC programs, other DPs, the World Bank in particular, take the lead in assisting country authorities in this area. Hence, while any direct collaboration between the Fund and others is rarely discussed in program documents, support provided to country authorities by other DPs is covered. 7/

/1 See Annex VI and VII regarding which GRA-supported programs were included in the review. The following GRA program requests reviewed included a social and other priority spending target: Armenia (2014); Jamaica (2013), Pakistan (2013), Suriname (2016), and Tunisia (2013 and 2016).2/ Explanations provided for missing the targets included better targeting (Armenia 2014); and technical and administrative challenges in electronic payment mechanism and weak absorptive capacity or delays in expenditure for the poor (Tunisia 2013).3/ Other SSN conditionalities included issuing a study on policy options for targeted protection for vulnerable groups (Armenia (2014)) and establishing a databank on vulnerable households (Tunisia (2016).4/ Albania 2014, Armenia 2014, Georgia 2014 and Jamaica 2013.5/ Armenia 2014, Jamaica 2013 and Serbia 2015.6/ Armenia 2014 and Bosnia 2012.7/ For example, collaboration between country authorities and other DPs to strengthen the social safety net scheme is discussed in Iraq (2010), Romania (2013), Sri Lanka (2016), Suriname (2016).

22. The mission chief survey indicated that the limited use of specific reform measures to help protect vulnerable groups reflected a variety of factors. A key challenge cited by mission chiefs was the absence of a pre-existing social safety net framework (Figure 10). Where countries have safety nets (as is often the case with GRA-supported programs), these can be used at short notice to deliver support to vulnerable groups. However, when missing, it generally takes considerable time to design and implement new social safety net tools. This suggests the importance of collaborating with country authorities and other DPs early, such as during surveillance discussions, to allow sufficient time to identify affordable and effective social safety net tools to protect the poor. Most mission chiefs view weak capacity in the country as a key constraint to effectively implement social safety net programs. Respondents also indicated that lack of social indicators data hampers carrying out an impact assessment of specific social safety net measures. Technical assistance from the Fund or other DPs can improve the administrative and implementation capacity of the authorities and assist in improving data availability. The Fund’s Fiscal Affairs Department (FAD), often with joint missions with the World Bank, supports countries through the provision of TA missions on social safety nets that focus on measures to mitigate possible adverse effects of fiscal reforms (Boxes 4 and 6). More than half of mission chiefs also noted that limited financial resources to fund new social safety net programs was an issue. This highlights the importance of enhancing spending efficiency to help create fiscal space necessary to finance social safeguards measures in a fiscally sustainable manner. One in three respondents also cited a lack of Fund staff expertise on social safety nets as a constraint.

Figure 10.
Figure 10.

Mission Chief Views on Obstacles in Program Design

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

Source: Mission chief survey results. Reflects 28 respondents.

Fiscal Affairs Department and Social Protection 1/

The Fiscal Affairs Department (FAD) provides extensive support to member countries on the design of social protection systems. Much of this work began in the 1990s in the context of structural reforms in transition economies. This support focuses primarily on ensuring the fiscal sustainability of social protection spending and strengthening the capacity of safety nets to protect vulnerable households from unintended adverse economic consequences associated with macroeconomic and structural reforms, including fiscal reforms. Support is provided across a range of spending areas including public pension and health spending, energy subsidy reform, and social assistance spending.

Support is provided in the context of the various activities of the IMF; surveillance, use of Fund resources, and technical assistance, including capacity building. This approach is centered around: technical assistance missions, capacity building courses, supporting templates and tools, and analytical work.

Since 2004, FAD has provided extensive technical assistance (TA) to member countries to help ensure the sustainability of social protection spending and to strengthen social safety nets (Figure). TA missions have reached all regions, including Africa (30), Europe (28), Western Hemisphere (26), Middle East and Central Asia (19), and Asia and the Pacific region (9).

FAD has expanded its capacity building efforts to further promote knowledge sharing. This includes conferences, workshops, and online courses. In 2016, FAD’s Tokyo Fiscal Forum focused on Fiscal Policy for Long-term Growth and Sustainability in Aging Societies and a regional capacity building seminar in Sri Lanka on Enhancing Social Spending in Support of Inclusive Growth in Asia. In 2013, an annual workshop on “Reforming Fuel Subsidies” was launched in Kuwait and, in 2016, a course on “Energy Subsidy Reform” was launched in Central and Eastern Europe. In 2014, in collaboration with the ICD, FAD launched a massive open online course (MOOC) on “Energy Subsidy Reform.”


FAD Technical Assistance in Social Protection, FY04–FY17

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

User-friendly templates and tools have also been developed to support IMF country teams in their policy dialogue with members. These include tools for analyzing the distributional analysis of subsidy reforms, the fiscal sustainability of public pension systems, and long-term projections of public pension and health spending. In the area of energy subsidy reform, these toolkits together with supporting databases and other resource materials are available through FAD’s energy subsidy reform website.

FAD has undertaken substantial analytical work on social protection issues. This includes books on The Economics of Public Health Care Reform in Advanced and Emerging Economies’ (2012), Energy Subsidy Reform (2013), Equitable and Sustainable Pensions (2014) and Inequality and Fiscal Policy (2015).

/1 Prepared by FAD.

Operationalizing the IMF Work on Income and Gender Inclusion 1/

To support surveillance and program discussions, the IMF has significantly stepped up efforts to operationalize its work on inclusion over the last couple of years. A pilot initiative on inequality and gender started in 2015 and is currently under way. Concerning inequality, the analysis in the pilot studies mostly focuses on the distributional impact of macroeconomic policies and reforms. On gender, the focus is on labor force participation and the impact of policies. As part of the initiative, 9 pilot countries on inequality and 13 on gender have been completed.

To assess the distributional impact of macroeconomic policies and reforms, staff has developed an analytical framework, which has been applied in various pilot studies on inequality. Making use of a dynamic general equilibrium model customized to each country’s key macroeconomic features, the framework provides insights on the economic and distributional impact of reform packages and the transmission channels. It has been applied in studies conducted in the context of Article IV consultation discussions and provided inputs for the program design in the case of Malawi. Some examples include:

Ethiopia (2015 Article IV). The framework was used to assess the impact of structural reforms in the fiscal and financial sectors. Given that the analysis indicates that some of the policies could have regressive outcomes from a distributional standpoint, staff recommended adjusting the financial sector reform to increase financial service access and complementing the reform package with an expansion of the cash transfers for immediate support to the most vulnerable.

Honduras (2016 Article IV). Staff applied the framework to assess the distributional impact of the 2013 tax reform, which was part of the fiscal consolidation strategy to address large macroeconomic imbalances. The reform strategy also included an expansion of the cash transfer program aiming at protecting the most vulnerable from the potential negative effects of the reform. Based on the analysis, staff estimated that the reform package (tax reform and the expansion of the cash transfer program) had been progressive and helped boost growth while reducing inequality and poverty.

Malawi (2016 Article IV). In Malawi, the framework was used to quantify the impact of reforming agricultural subsidies. Based on the analysis, staff recommended targeted cash transfers to reduce the negative impact on the income of poor farmers in the short run and higher spending in agricultural research and development (R&D) to boost their productivity in the longer term.

Lessons from the pilot studies that applied the model-based framework were reported in a recently published IMF Staff Discussion Note (SDN/17/01). The note examines the distributional impact of macro-structural policies such as fiscal reforms, financial sector reforms, and reforms to the agricultural sector in low-income developing countries. The note provides important insights on the transmission channels, highlighting how the impact of macro-structural policies depends on both the reform design and on country-specific economic characteristics such as labor mobility, the level of informality, financial access and infrastructure gap. The note also explores accompanying measures that governments can adopt to offset potential adverse distributional consequences of reforms.

On gender, IMF staff analyses in the context of Article IV consultations and program discussions have examined potential barriers to labor participation in several low-income and developing countries. For example:

WAEMU (2016 Staff Report). Staff examined gender inequality and recommended that investment in infrastructure, human capital, and institutions should be coupled with efforts to close gender gaps in opportunities (e.g., education and health), remove legal barriers against women, and increase females’ access to finance.

Mali (2016 Article IV). Staff highlighted gaps in education and changing demographics as the main constraints to female labor force participation—they recommended expanding education spending and training programs for girls and women. Recommendations also included closing the gap between the demand for and the provision of contraception.

Niger (2017 New Extended Credit Facility Arrangement). The new arrangement for Niger seeks to maintain macroeconomic stability, reduce poverty, and make growth more inclusive. In this context, staff stressed the need for a strong reform agenda anchored, among other measures, on a comprehensive strategy for addressing gender issues.

To help enhance the role of fiscal policy in addressing gender inequality, the IMF has recently completed a study on gender budgeting efforts in 23 countries with significant initiatives. This work analyzes the use of fiscal policies to reduce women’s unpaid work burden, identifying tax structures that discourage female labor force participation. Main lessons suggest that gender budgeting efforts should focus on education, health, and infrastructure spending to reduce gender inequality in low-income developing countries. Details of this work are provided in a series of working papers and a publicly available toolkit.

1/ Prepared by SPR’s Developing Market Strategy Unit.

Mitigating Measures to Protect the Poor During Energy Subsidy Reform 1/

Measures to mitigate the impact of energy price increases on the poor are critical for building public support for subsidy reforms. Energy subsidies are inefficient, fiscally costly, and provide the most benefit to higher-income groups. They also crowd out pro-poor and growth-enhancing public spending on health, education, social protection, and infrastructure. However, eliminating energy subsidies can have a substantial adverse impact on the real incomes of the poor. Empirical studies have found that, on average, an increase of $0.25 per liter in fuel prices results in a 5.5 percent decline in real incomes of poor households.

The appropriate choice of measures to mitigate the adverse impact of energy subsidy reform will depend on the capacity of the current social safety net. The first step in identifying mitigating measures is therefore to assess the capacity to expand existing social programs in the short term. This requires an evaluation of the range of social programs available, their coverage of poor households, and their ability to minimize leakage of benefits to higher-income groups. Implementing or expanding targeted programs immediately prior to price reforms can help demonstrate the government’s commitment to protecting the poor. In combination with good targeting, setting compensation levels based on energy consumption by poor households will help generate fiscal savings since poor households typically consume substantially lower quantities of energy than higher-income groups.

Targeted cash or near-cash transfers are the preferred approach to compensation. Cash transfers give beneficiaries the flexibility to purchase the level and type of energy that best suits their needs and at a time and place of their choosing. They also remove the need for governments to be directly involved in the distribution of subsidized energy to households, which is often extremely costly and prone to abuse. The degree to which compensation should be targeted is a strategic decision that involves tradeoffs between fiscal savings, capacity to target, and the need to achieve broad acceptance of the reform.

When well targeted cash transfers are not feasible, other programs can be expanded while administrative capacity is developed. The focus should be on existing programs that can be expanded quickly, possibly with some improvements in targeting effectiveness, such as school meals, public works, reductions in education and health user fees, subsidized mass urban transport, and subsidies for consumption of water and electricity below a specified threshold. The challenge is to develop a package of measures that ensure adequate coverage and compensation while minimizing leakage of benefits to higher-income groups.

Appropriate phasing of price increases and sequencing them differently across energy products may also be desirable. The appropriate phasing and sequencing of price increases will depend on a range of factors, including the magnitude of the price increases required to eliminate subsidies, the fiscal position, the political and social context in which reforms are being undertaken, and the time needed to develop an effective communications strategy and social safety nets. As the safety net is strengthened, subsequent rounds of reform can include larger increases in prices for fuel products that are more important in the budgets of poor households (such as kerosene), and part of the budgetary savings can be used to finance targeted transfers to poor households.

1/ Prepared by FAD.

23. Coverage of social safeguard issues in program documents needs to be strengthened. Review of staff reports reveal that program documents do not consistently discuss how fiscal adjustment or other economic reforms may adversely impact poor and vulnerable groups, nor do they discuss the scope for measures to mitigate the impact on poor and vulnerable groups. Improving the coverage of social safeguards issues in program documents would be helpful and would provide a useful record on how such concerns were addressed in the policy discussions with national authorities. This applies equally to cases where the existing social safeguards are considered adequate, which may justify why mitigating measures were excluded. Only half of LIC program request documents discuss social safety net issues, compared to nearly 90 percent for GRA-supported programs.18 The most common discussions were related to expanding social protection programs and increasing social transfers (Figure 11). Program documents could also better cover the collaboration between the Fund and other DPs on social safeguards policies. This is important as other DPs, such as the World Bank, usually take the lead in advising country authorities on social protection issues. Program documents infrequently mentioned any direct collaboration with other DPs. One-quarter of LIC program requests discuss ongoing or planned support to the authorities by other DPs on social protection issues, compared to about half in GRA-supported programs.

Figure 11.
Figure 11.

Mitigating Measures in LIC Programs with Fiscal Consolidation 1/

Citation: Policy Papers 2017, 029; 10.5089/9781498346757.007.A001

1/ Number of LIC program requests between January 2010 and May 2016. Focuses on the discussion of measures.


This section discusses how vulnerable groups can be more effectively protected using different LIC program design approaches.

A. Social and Other Priority Spending Indicative Targets

24. The effectiveness of the social and other priority spending IT can be improved. Effectiveness, specifically, refers to protecting poor and vulnerable groups from bearing an undue burden of the economic adjustment costs. The key areas for improvement are in the design of the IT and the attention given to the topic in the policy dialogue with national authorities, as reported in country documents. The recommendations below identify good practices, which in most cases build on examples from actual program experience. The core principles behind the recommendations are flexibility in designing the IT—which is important to accommodate country-specific circumstances— and clarity in the discussion of how the IT achieves the desired social safeguard goal. These principles, moreover, will help ensure that the usage of social and other priority spending IT does not devolve into a mere box-ticking exercise.

25. The social and other priority spending IT should be designed to provide meaningful social safeguards tailored to country specific circumstances. The following are some recommended good practices:

  • Focus on safeguarding resources for vulnerable groups. Social and other priority spending can be a broad concept and may include many types of spending unrelated to vulnerable groups. Where public infrastructure spending is a development priority, it may provide few benefits to the poor and vulnerable in the short term. Thus, in cases where economic reforms are expected to have adverse distributional consequences, there is merit in defining spending floors to narrowly cover the needs of the most vulnerable. The current policy provides flexibility to country teams and authorities to focus the program target only on social spending or on social and other priority spending combined, depending on specific country needs.

  • Select social spending with the biggest impact. There may be scope in some countries to improve the targeting and quality of spending. The focus should be on targeting critical sectors where the benefit would be large and delivering targeted assistance to the vulnerable through social safety net programs. This may entail improving the targeting of existing schemes or introducing new social assistance schemes. For example, targeted cash or near-cash transfers are usually seen as the preferred approach to compensate for the adverse impact of energy subsidy removal on the poor (Box 6). Fiscal space to accommodate such increased spending can be created, for example, through domestic revenue mobilization or increasing spending efficiency (e.g., by eliminating inefficient spending such as on energy subsidies). The Fund already provides extensive TA and training in this area. Complementary program conditions, discussed below, could also be used to strengthen the social safety net. Social safeguard spending, moreover, could also have other economic benefits, including through relatively high fiscal multipliers.19 For example, the World Bank (2015) found that cash transfer programs have positive spillover effects on the local economy, with the total income multipliers ranging from $1.08 to $2.52 for each dollar transferred.

  • Leverage expertise from development partners. Close collaboration with and support from DPs and donors is important for designing the social and other priority spending IT. Different methods of impact and incidence analysis employed by other institutions such as the World Bank could be useful in informing what constitutes spending that could benefit the poor. Thus, staff should reach out early, ideally during surveillance discussions, to consult closely with donors and DPs such as the World Bank, United Nations agencies, and regional development banks (for example, the African Development Bank and Asian Development Bank).20 In addition, to address the concern raised by mission chiefs that Fund staff lacks the expertise in the design of social safeguards, mission chiefs could be provided more regular training opportunities in this area.

  • Avoid reliance on uncertain foreign financing. Where spending floors are established under Fund-supported programs, the government should have adequate control over its resources to meet these goals. This can be a problem where spending goals are designed to be met in part through external financing outside the control of the authorities. This argues for designing spending floors sufficiently tightly that they can be met, even in the event of shocks to foreign financing.

  • Periodically revisit the definition of social and other priority spending. In practice, the definition of these targets is rarely revised during the span of a program or even across multiple programs. In all cases, but especially when the target is habitually being missed, there would be merit in revisiting the definition of social and other priority spending more regularly. This would ensure that the definition is consistent with any changes in the economic environment (especially if, for example, the economy suffered from various unforeseen external shocks) and, importantly, still appropriate for achieving the underlying objectives of providing social safeguards. The benefit of revising the definition and design of the IT was also commonly noted by the mission chiefs (Figure 9). If the target seems to be well-designed, the focus should be on corrective measures to, for example, strengthen public financial management (PFM).

26. Social safeguard issues could be more thoroughly and consistently addressed in policy dialogue with country authorities. This dialogue would normally include the design and implementation of social safeguards, including in a program context. Documenting this dialogue in staff reports would be important to demonstrate the attention given to safeguarding vulnerable groups in PRGT-supported programs. Specific recommendations for policy dialogue include the following:

  • Elaborate on how the social and other priority spending IT supports social safeguards. While sometimes obvious, the link between the IT and social safeguards should be spelled out. This could include, for example, discussion of the key social spending programs that are covered by the IT. Regardless of whether the IT focuses on the “social” or “other priority” component of spending, the TMU should set the reporting requirement for disaggregated data to be provided by country authorities. This periodic reporting would allow for better monitoring and designing the IT to better target the poor. The provision of more detailed data should facilitate a more thorough discussion between country authorities, Fund teams and other DPs on the expected social benefit of certain priority projects. In addition, spending efficiency should be monitored through regular reporting on intermediate social objectives (e.g., school enrollment rate, access to health services, coverage rate of the poor by certain social programs). This regular reporting mechanism can usefully anchor the discussions on the spending categories included under the definition of social and other priority spending.

  • Describe the process for arriving at the definition of social and other priority spending. Program documents should note the institutional process used to identify critical social sectors and segments of the population to be covered under the IT, and identify any methodologies (e.g., impact analysis) used or drawn upon to guide what constitutes social spending. This would help to enhance transparency and accountability around the definition of the IT, and would provide a basis for more substantive discussions between Fund staff and country authorities on this aspect of program design. Staff reports requesting Fund arrangements could also explain how the budgetary resources will be made available in a timely and predictable manner. This could include any supporting TA from the Fund or other donors.

  • Discuss obstacles in design of the IT. A discussion of the challenges in designing the social and other priority spending IT—such as a lack of data, capacity, financial resources, or presence of well-functioning social safety net programs—will provide useful context. This will assist in identifying areas where either more TA and training needs to be provided to the authorities or areas where a complementary conditionality might ensure progress is achieved. If serious data gaps exist that hinder the tracking and monitoring of spending on social safeguards measures, including the impact assessment of programs and policies, there should be a discussion on how this will be addressed by technical assistance from the Fund or donors. Addressing these gaps is key to improving the analysis, design and impact of programs and policies. For example, in cases where tracking or monitoring of social safeguard expenditures is not feasible, the program documents could discuss measures to develop an adequate tracking system and to strengthen PFM. Progress in establishing such systems should be closely monitored and the program may include relevant structural measures, if appropriate.

  • Discuss slippages and corrective measures. When the IT is missed, program documents should elaborate on the reasons and provide concrete measures to avoid recurrence of the slippage. If the IT is missed again despite corrective measures taken, then consideration should be given to different corrective measures.

B. Reform Measures to Protect Vulnerable Groups

27. Given the limitations of the social and other priority spending IT, complementary social safety net measures could play a stronger role in policy advice. Good practices for strengthening the use of such complementary conditionality include (i) giving more explicit consideration to such options during the program design stage; and (ii) focusing more during surveillance on developing social safety nets, including in conjunction with DPs.

28. At the program design stage, consideration should be given to options for strengthening social safeguards. For example, as program conditionality on eliminating subsidies or tax hikes (notably indirect taxes) is likely to have a direct impact on the poor, it is important that program design clearly addresses mitigating measures (Box 6).21 This can be done by either introducing conditionality or by clearly explaining concrete measures underway, including specific timelines as to when they will be introduced and implemented. Ideally, the savings generated from the elimination of subsidies should be used at least in part to introduce targeted social assistance schemes or to increase transfers via established social programs. Also, appropriate phasing and sequencing of reform measures may be desirable when time is needed to put in place social safety net measures. A challenge in many LICs, however, is that such social assistance schemes may still be at an early stage of development and hence, not able to effectively deliver protection. In such cases, consideration could be given instead to determining TA and capacity building needs of country authorities and to ensure these identified gaps are met either by the Fund or with the support of other DPs. Program documents should clearly explain how capacity constraints that may impact the effective implementation of certain social safety nets are being tackled, even if this requires a more medium-term approach. In addition, in cases where the program got off-track or the country was subject to unforeseen economic shocks, there might be a need to reevaluate and revisit social safeguards needs and goals; hence, program documents should explain how such changes in conditions have been accounted for into the program design.

29. Ideally, Fund-supported programs should build on earlier discussions of social safeguard needs. Early engagement with experts and the authorities can give sufficient time to lay out the necessary groundwork to protect the poor in advance of adjustment policies. Areas for policy dialogue during Article IV discussions could include the following:

  • Adequacy of existing social policy instruments. Where existing instruments to protect vulnerable groups from economic shocks are inadequate, the Article IV discussions could identify scope to strengthen the institutional framework through TA or safety net reforms. Collaboration in these areas should extend to the World Bank and other DPs.

  • Fiscal sustainability. Early discussion of social safeguards can help ensure that they are both timely and affordable.22 Where fiscal space is limited, options include increasing domestic revenue mobilization and boosting spending efficiency.

  • Use of distributional impact analysis. The design of social safeguards can greatly benefit from exante technical analysis of the likely social and distributional impact of certain macroeconomic policies such as tax or fuel subsidy reforms (Boxes 4 and 5).23 As these types of analysis are time and resource intensive, they should ideally be conducted during surveillance discussions to provide timely input for program design (see Annex IV for good practice examples).

30. More broadly, close collaboration with development partners is warranted in program and surveillance discussions. If weaknesses in social safety net measures were identified during surveillance discussions, program conditionality can assist in addressing some of these; however, TA by the Fund or other DPs would need to support these efforts. Given resource constraints at the Fund, close engagement with development partners and donors at both the surveillance and the program design stage will be critical. The program document should include a discussion on how weaknesses identified will be addressed and how the Fund is collaborating with key stakeholders to provide assistance to the authorities.


31. Durable poverty reduction and economic growth is a core objective of LIC programs. Consistent with this, protecting the poor and vulnerable from bearing an undue burden of adjustment in LIC programs is also a priority. Social safeguards have an important role to play in LIC programs. This is especially true when these programs include adjustment measures that could adversely impact the poor and vulnerable.

32. LIC programs involving fiscal consolidation need to pay due regard to protecting the vulnerable from the burden of adjustment. Nearly half of programs build in fiscal expansions, and real per capita spending is forecast to rise on average by some 15 percent in the four years following the program. Moreover, empirical evidence continues to suggest that LIC programs are successful in protecting health and education spending. Nonetheless, many programs do contain fiscal consolidations and other measures (such as the removal of subsidies) that could hurt the poor. It is important to ensure that those LIC programs incorporate sufficiently strong social safeguards measures.

33. There is room for improvement in how LIC programs address social safeguards issues. Regarding the social and other priority spending IT, the definition of social and other priority spending can be designed to better target the poor. This would require careful discussions with country authorities, the World Bank and other DPs. Program documents should more thoroughly and consistently explain the challenges in the design of the target, collaboration with other DPs, any slippages, and corrective measures. The targets should also be revisited periodically. Meanwhile, complementary conditionality on reform measures to protect vulnerable groups have been used quite sparingly. Such conditionality, however, can often be the most effective way to protect the poor and vulnerable from economic policy reforms, especially as many LICs have fairly rudimentary social safety net systems. Thus, Article IV consultations could be used to take an in-depth look at social safety nets and discuss how an expansion of such programs can be financed in a fiscally prudent way, with the support of the World Bank or other DPs, and identify a roadmap for reform (including supporting TA).

34. The recommendations in this report can be implemented without changes in policies and procedures. Indeed, many are based on good practices drawn from existing experience. The underlying principle is to ensure that due consideration is given to social safeguards in the design of programs and in surveillance. This includes ensuring and, importantly, also documenting that the program is adequately protecting the poor and vulnerable from the burden of economic adjustment. Ultimately, this can best be achieved by country teams tailoring the program—and surveillance—to country-specific circumstances. Existing policies and procedures have sufficient flexibility to achieve this without any changes.

35. It is intended that this stocktaking will be followed by a guidance note to staff. The guidance note would clarify expectations on how to treat social safeguards measures in both Article IV consultations and program discussions for LIC programs. It could further help in strengthening and formalizing the institutional process for collaboration with the World Bank and other DPs on social safeguards. The guidance note will be issued to the Executive Board for information.

Issues for Discussion

  • Do Directors see a need to strengthen the design of the IT on social and other priority spending, including by improving its targeting on poverty reduction, by collaborating more closely with country authorities and other DPs in its design, and by revisiting and revising its definition more periodically?

  • Do Directors support the need to improve the incorporation of social safeguards issues in program design, including by relying more often on specific reform measures to strengthen the social safety net system to more effectively protect the poor from the impact of economic adjustment?

  • Do Directors support the need to improve the coverage of social safeguards issues in program documents, including the discussion on measures incorporated to protect the poor or the challenges faced in doing so and concrete measures proposed to tackle these constraints, improving the monitoring and reporting of social spending and intermediate objectives?

  • Do Directors support strengthening Fund engagement in social safeguards policies, through more structured and frequent coordination with country authorities and DPs during both Article IV consultation cycles and Fund-supported programs?

Annex I. Good Practice Examples in LIC Programs: Social and Other Priority Spending Indicative Target

Bangladesh (2012 ECF Request): Includes a detailed table of safety net program and respective ministries and expenditure codes. The expenditure under the safety net program is grouped into four: (i) cash transfer programs (e.g., old age allowance); (ii) food security programs (e.g., open market sales); (iii) micro-credit and miscellaneous funds (e.g., funds for climate change); (iv) development sector programs (e.g., school feeding program). The program makes efforts to avoid double-counting. The document also includes an informative discussion on social safeguard issues.

Honduras (2014 SBA-SCF Request): Its social safety net program is clearly defined and targeted. “Vida major - bono diez mil” is a cash transfer system covering a large share of families living in extreme poverty by providing cash benefits conditional on regular monitoring of child growth and use of health and education services. It is monitored quarterly on the basis of financial reports provided by the Ministry of Finance. Social safeguard issue is discussed frequently throughout the program document.

Kenya (2011 ECF Request): For the purposes of the program, priority social spending of the government is detailed and country-specific. It is defined as the sum of: cash transfers to orphans and vulnerable children; cash transfers to elderly; anti-retroviral treatment expenditures; free primary education expenditure; and free secondary education.

Sao Tome and Principe (2015 ECF Request): It includes a detailed description of pro-poor spending. Expenditures included under the pro-poor current spending are listed by functional classification and by budget codes in a matrix format. The pro-poor treasury-funded capital spending are defined as those deemed to have a direct impact on alleviating poverty in the following sectors: education, health, social safety nets, agriculture and fisheries, rural development, youth and sports, provision of portable water, and electrification. This definition of pro-poor spending was discussed and agreed with the World Bank and the IMF.

Annex II. Good Practice Examples in LIC Programs: Complementary Conditionality on Social Safeguards

Burkina Faso (2010 ECF Request)

The program includes conditionality on preparing a study on taxation of petroleum products and on a new pricing system. This is accompanied by a conditionality to produce quarterly reports on poverty-reducing expenditure with the goal to enhance monitoring of poverty reduction efforts. In addition, though not a program condition, the staff report mentions that the authorities commit to strengthening current social programs, through extended school lunch programs, support reduced-price sale of basic consumer goods in areas affected by food insecurity, reduced cost for maternal care for pregnant women, and promoting youth employment. The authorities have also initiated the preparation of a broader social protection system, in collaboration with the World Bank and other DPs. A nation-wide workshop on social protection was planned for later in the year.

Haiti (2015 ECF Request)

The program includes conditionality on the elimination of fuel subsidies and adoption of an automatic fuel price adjustment mechanism. Staff highlights the regressivity of fuel subsidies drawing upon World Bank analysis. The World Bank and IDB are providing key assistance in restructuring the energy sector. To mitigate the impact of fuel subsidy elimination, World Bank assistance is sought by authorities to design a social tariff for public transportation. The program also includes conditionality to streamline the spending categories covered under poverty reducing expenditure with the goal to strengthen targeting.

Madagascar (2015 SMP Request)

Program conditionality includes elimination of fuel price subsidy and cabinet approval on the national social protection policy to guide the design and implementation of social protection programs, including social safety nets for the poorest and most vulnerable households.

Yemen (2014 ECF Request)

Program conditionality includes subsidy reform (increasing domestic fuel prices) that would free up significant resources. The conditionality also specifically earmarks these savings to be used to increase Social Welfare Fund (SWF) allocations to the poor by 50 percent. Coverage of SWF is planned to be further improved with help of World Bank.

Annex III. Good Practice Examples in LIC Programs: Discussion of Social Safety Net Issues

Bangladesh (2012 ECF Request)

The Fund and World Bank staff have prepared a joint note detailing the type of social safety net measures in place and assessing the effectiveness in mitigating the impact of higher food and energy prices. The staff report provides a detailed stocktaking on existing social safety net programs accompanied by a discussion on poverty and inequality trends in the country. One of the key pillars of the program is energy subsidy reform. The program has conditionality to adopt automatic fuel price adjustment mechanism which is expected to free up significant resources. This newly created fiscal space is expected to be used to protect vulnerable households from rising fuel and food prices, building on World Bank TA and the work of other DPs.

Honduras (2014 SBA-SCF Request)

The program has a prior action to reinstate the automatic mechanism for the adjustment of fuel prices. This is accompanied by a comprehensive discussion on ongoing work by the authorities to strengthen the protection of vulnerable groups. The authorities plan to preserve the expansion of social programs experienced in recent years substantially increasing the share of social spending in overall primary spending. The government plans to use part of this spending on improving housing for low-income families under their recently established Vida Mejor program, which will consolidate existing social programs (notably the conditional cash-transfer program Bono 10 mil). The fiscal program envisages allocating part of the savings from reductions in the wage bill to Vida Mejor. Even though it is not a formal conditionality, IMF staff calls to complete a review of the “Bono 10 mil” and “Vida Mejor” social programs to ensure that they are adequately targeted to low-income families. The government is in the process of conducting a thorough review of their social programs, with the aim of improving their targeting. The authorities are also implementing a subsidy scheme for families with monthly electricity consumption below 75KW, which protect them from increases in electricity tariffs. In addition, the government plans to reform the pension and healthcare systems to increase coverage. With assistance from the World Bank and the IDB, the authorities will also revise the design of the safety net included in the draft legislation to improve its efficiency while keeping costs at fiscally-affordable levels.

Kyrgyz Republic (2015 ECF Request)

Staff and the authorities agreed that when reforming the subsidy programs mitigating the impact of these reforms on the vulnerable will be critical, and would require gradually increasing social spending over the course of the program. It details authorities’ plan to (i) streamline untargeted social spending by reviewing the existing programs; and (ii) to improve forecasting and procurement of social spending.

Mauritania (2010 ECF Request)

The staff report emphasizes the lack of coherent and comprehensive social protection and safety nets. It discusses that the authorities together with the World Bank and other donor assistance will work on strengthening and expanding, well targeted and coherent subsidy schemes and social safety nets programs. Authorities plan on launching, in collaboration with the UNICEF, a comprehensive study on existing social protection and safety nets schemes.

Mozambique (2011 Article IV and 2nd Review of PSI)

The staff report includes an extensive discussion on social protection policies even if the program does not include any conditionality in this area. Staff advised authorities to leverage fiscal space to design sustainable social protection policies. In this context, the mission emphasized that the removal of the fuel subsidy (not a formal program conditionality) should be accompanied by better targeted social protection schemes, which could be gradually rolled out, based on the joint costing pilot work conducted by UN (including the ILO), the Bank, the Fund, and bilateral partners. The exercise includes (i) a Bank-led review of existing social security programs and expenditure; (ii) a costing exercise led by the ILO, focusing on public works programs, conditional cash transfers, and a universal pension system; (iii) a Fund-led assessment of available fiscal space consistent with the macroeconomic framework; and (iv) a simulation of the impact of policy options on the poverty gap, led by UNICEF.

Annex IV. Good Practice Examples in LIC Surveillance Work: Discussion of Social Safety Net Issues

The Federal Democratic Republic of Ethiopia (2015 Article IV)

Ethiopia was a pilot case for the Fund’s project to operationalize its work on inequality, as the issues related to equitable growth were deemed macro-critical. Using a dynamic general equilibrium model tailored to Ethiopia, staff analyzed the macroeconomic and distributional impacts of economic and financial reforms. The discussion details how various policy measures would contribute to sectoral growth and promote structural transformation in the economy (e.g., in manufacturing and services and agriculture). The analysis then describes how this structural transformation would change poverty and inequality. Specific recommendations to sustain equitable growth included: (i) enhancing tax collection and lowering tax exemptions that would increase the fiscal space for poverty reduction; (ii) updating nominal income tax thresholds; (iii) continuing the pro-poor focus; and (iv) promoting financial inclusion via mobile banking and encouraging banks and microfinance institutions to increase their outreach.

Republic of Congo (Selected Issues Paper issued in conjunction with the 2015 Article IV)

Using benefit incidence analysis, staff estimates that prior to the 2008 adjustment in petroleum product prices, the fuel subsidies (except kerosene) mostly benefitted the wealthy and concluded that reducing or eliminating fuel subsidies is likely to improve the progressivity of public spending. However, to mitigate the impact of removing the subsidies on the poor, staff advised that the energy subsidy reform should be well-designed and take into consideration the impact on vulnerable households, including from interventions in the form of conditional cash transfers and subsidies for public transportation. Staff also recommends a rebalancing of expenditures in favor of social programs to mitigate the potential adverse impact of fiscal consolidation. The paper includes a detailed discussion on social safety net measures that the government at the time start working on such as introducing a cash transfer system, with the assistance of the World Bank and developing a Universal Health Coverage program.

Malawi (Selected Issues Paper issued in conjunction with the 2015 Article IV)

The Farm Input Subsidy Program (FISP) is one of the largest social expenditure items in Malawi, aimed at improving food security and reducing poverty. While the program was considered successful when it started in 2005, more recently, the program has become increasingly costly, without the accompanying improvements in maize production or poverty reduction. As a result, the authorities have been considering reforming the FISP program. FISP is also a component of the social and other priority spending indicative target under the ECF program. The study focuses on the macroeconomic and distributional impacts of reducing the subsidy rate for the FISP. If then discusses complimentary policies such as expanding the existing cash transfer program that could compensate the poor and foster equitable growth.

Senegal (Selected Issues Paper issued in conjunction with the 2014 Article IV)

Fund team conducted a review of the social safety net system and outlined the main strategies going forward to strengthen the system. The work has relied extensively on previous World Bank analysis. The review describes the dozen social safety net programs which were in place in 2011; discusses the average cost of these programs and explains why these programs are not sufficient to protect the poor. The main issues identified are the insufficient coverage of the poor; weaknesses in the targeting system; large variation in cost of beneficiary per program with some likely having only minor impact; lack of coordination across various programs; lack of a standardized monitoring of program implementation and heavy reliance on development partner financing. The note also describes the important steps the authorities have taken to strengthen the current system. Concrete recommendations are included such as strengthening the coordination between programs and moving towards a unified national social safety net system, designing a medium-term expenditure framework and integrating it into the budget process, developing central monitoring and evaluation mechanisms that informs strategic decision-making and allows program evaluation, improving targeting and further strengthening the institutional framework.

Annex V. Good Practice Examples in GRA-Supported Programs

Jamaica (2013 SBA Request): Besides including a floor on social spending to safeguard this spending category, the program document also discusses in detail authorities’ commitment to reduce the adverse impact of adjustment on the vulnerable by improving the existing social protection framework. The authorities intend to improve training and certification for labor market participants to tackle the high unemployment rate; enhance benefits and improve effectiveness of the targeting of beneficiaries under PATH (a conditional cash transfer programme); and implement welfare-to-work exit strategies for vulnerable households. The program also includes reforms for improving the efficiency and effectiveness of social spending, to help ensure that the available resources are better targeted to the poor (e.g., in health and education sectors), building on TA provided by FAD.

Jordan (2012 SBA Request): The fiscal adjustment program in Jordan highly relied on eliminating the costly and ill-targeted subsidies in the energy sector. An increase in electricity tariffs earlier in the year prior to the program had to be reversed because of heightened social tensions. The program tried to address these challenges by adding a conditionality to introduce targeted support to the poor should fuel prices increase drastically to accompany the conditionality to increase fuel prices. The conditionality on the energy sector reform was to be carried out in collaboration with the World Bank. Staff also stressed that even though there are many social assistance programs, better targeting was necessary, as these programs suffer from waste and inefficiencies where only 15 percent of beneficiaries would be classified as poor according to the national poverty line.

Romania (2013 SBA Request): While there is no conditionality either on social spending or social safety nets, the program document discusses in detail the authorities’ plan to mitigate the adverse impact of the energy subsidy reform. The revised budget allocates additional spending to mitigate the impact of gas and energy price liberalization through raising the Guaranteed Minimum Income (GMI) and family allowances as well as increasing the eligibility thresholds, consistent with the authorities’ social assistance reform program to consolidate benefits and improve means testing. At the same time the authorities lowered the VAT on bread temporarily and also extended the existing legislation by three years to provide a top-up to unemployment benefits for some SOE layoffs.

Ukraine (2014 SBA and 2015 EFF Requests): As part of a comprehensive reform agenda for the energy sector, program conditionality in both programs include measures to bring gas and heating prices to cost recovery based on international prices. Both program documents include an extensive discussion of mitigating measures of increasing energy prices on the poor. Increases in energy prices are to be accompanied by enhanced social assistance from the budget to cover most vulnerable. The 2014 SBA includes a prior action for the government to introduce a new social assistance scheme to protect vulnerable households not covered by the existing scheme. The 2015 EFF explains how existing social assistance programs are being streamlined to improve targeting. It also mentions that the energy subsidy benefit formula will be revised in consultation with IMF and World Bank.

Annex VI. Main Source of Inputs to the Review

article image

Annex VII. List of Program Documents Reviewed

article image
article image
article image
article image
article image
article image
article image

Annex VIII. List of Current PRGT-Eligible Countries

Afghanistan, I.S. of




Burkina Faso




Cape Verde

Central African Republic



Congo, Dem. Rep. of

Congo, Republic of

Côte d’Ivoire





Gambia, The










Kyrgyz Republic








Marshall Islands Mauritania

Micronesia, Fed. States of







Papua New Guinea



São Tomé and Príncipe


Sierra Leone

Solomon Islands


South Sudan

St. Lucia

St. Vincent and the Grenadines




Timor-Leste, Dem. Rep. of







Yemen, Republic of




  • Baldacci, Emanuele, Benedict Clements, Sanjeev Gupta, and Qiang Cui, 2008, “Social Spending, Human Capital, and Growth in Developing Countries,World Development, 36 (8), pp. 131741.

    • Search Google Scholar
    • Export Citation
  • Clements, Benedict, Sanjeev Gupta, and Masahiro Nozaki, 2013, “What Happens to Social Spending in IMF-Supported Programmes?Applied Economics, 45 (28), pp. 402233.

    • Search Google Scholar
    • Export Citation
  • Furceri, Davide and Aleksandra Zdzienicka, 2011, “The Effects of Social Spending on Economic Activity: Empirical Evidence from a Panel of OECD Countries,Munich Personal RePEc Archive (MPRA), Paper No. 30356.

    • Search Google Scholar
    • Export Citation
  • Gupta, Sanjeev, Louis Dicks-Mireaux, Ritha Khemani, Calvin McDonald, and Marijn Verhoeven, 2000, “Social Issues in IMF-Supported Programs,” IMF Occasional Paper No. 191 (Washington).

    • Search Google Scholar
    • Export Citation
  • Gupta, Sanjeev, Marijn Verhoeven, Erwin R. Tiongson, 2002, “The Effectiveness of Government Spending on Education and Health Care in Developing and Transition Economies,European Journal of Political Economy, 18, pp. 717737.

    • Search Google Scholar
    • Export Citation
  • Independent Evaluation Office of the IMF, 2014, “Revisiting the IEO Evaluations of the IMF’s Role in PRSPs and the PRGF (2004) and The IMF and Aid to Sub-Saharan Africa (2007).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2002a, “Review of the Poverty Reduction Strategy Paper (PRSP) Approach: Early Experience with Interim PRSPs and Full PRSPs” (Washington).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2002b, “Review of the Poverty Reduction and Growth Facility: Issues and Options” (Washington).

  • International Monetary Fund, 2002c, “Review of the Key Features of the Poverty Reduction and Growth Facility-Staff Analyses” (Washington).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2007, “IMF Executive Board Discusses Operational Implications of Aid Inflows for IMF Advice and Program Design in Low-Income Countries,” Public Information Notice (PIN) No. 07/83, July 19.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2014a, “Revised Operational Guidance to IMF Staff on the 2002 Conditionality Guidelines” (Washington).

  • International Monetary Fund, 2014b, “2014 Handbook of IMF Facilities for Low Income Countries” (Washington).

  • International Monetary Fund, 2015a, “Eligibility to Use the Fund’s Resources for Concessional Financing” (Washington).

  • International Monetary Fund, 2015b, “The Impact of IMF-Supported Programs on FDI in Low-Income Countries,” IMF Working paper 15/157 (Washington).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2015c, “Financing for Development—Enhancing the Financial Safety Net for Developing Countries” (Washington).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2015d, “Reform of the Fund’s Policy on Poverty Reduction Strategies in Fund Engagement with Low-Income Countries—Proposals” (Washington).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2016a, “Handbook of IMF Facilities for Low Income Countries” (Washington).

  • International Monetary Fund, 2016b, “Poverty Reduction and Growth Trust—Review of Interest Rate Structure” (Washington).

  • International Monetary Fund, 2016c, “Financing for Development—Enhancing the Financial Safety Net for Developing Countries—Further Considerations” (Washington).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2016d, “Review of the PRGT Interest Rate Structure” (Washington).

  • International Monetary Fund, 2016e, “Financing for Development—Enhancing the Financial Safety Net for Developing Countries—Further Considerations” (Washington).

    • Search Google Scholar
    • Export Citation
  • The World Bank, 2015, “The State of Social Safety Nets 2015” (Washington).


As of February 2017.


An IT is a quantitative condition in a program and is assessed in the context of overall program performance. It differs, however, from a quantitative performance criteria (QPCs). QPCs are established as a formal condition for the making of purchases or disbursements under a Fund arrangement and require a waiver from the IMF Executive Board if they are not met, whereas an IT does not.


Social safety nets, as used here, refers to “noncontributory measures designed to provide regular and predictable support to poor and vulnerable people.” (World Bank, 2015). Types of social safety nets include cash transfers, school feeding, in-kind transfers (such as targeted food assistance), public work programs, and fee waivers.


Annex IV provide a list of good practice examples.


As stated in the 2002 PRSP review (see IMF 2002a), there are methodological problems in isolating the impact of particular programs and also significant lags between spending and data and impact on poverty outcomes.


These staff reports were reviewed to assess whether during the first year of the program, the program had a fiscal expansion or consolidation. The assessment relied mostly on the description of the program in the text, and where this was unclear, whether the program was expansionary or contractionary was assessed by comparing the primary balance and the overall fiscal balance during the first year of the program to the year before.


The focus on health and education spending is due to the limited availability of data on other components of social spending.


Short-run fluctuations in other social expenditure items are less straightforward to compare, due to large variation in measurement.


See Annex I-V for good practice examples.


For example, the Benin (2010), Grenada (2010), Solomon Island (2010), Solomon Island (2011), and Tanzania (2010) program requests mention the authority’s efforts to develop a better expenditure tracking system to monitor social spending. This was included as a structural benchmark in the case of Tanzania.


Delays in project implementation is a common reason for missing the target especially when large infrastructure projects are included under the definition of social and ‘priority’ spending, as was the case in the Liberia (2012), Guinea (2012), and Mauritania (2010) programs.


For example, Burkina Faso (2010 and 2013), Central African Republic (2010), Chad (2014), Comoros (2015), Liberia (2012), Mozambique (2010, 2013, and 2015), and Senegal (2010 and 2015) included the entire health and education sector. Examples of expenditure items where the link to poverty reduction is unclear include Burkina Faso (2010), which included expenditure on animal resources; Central African Republic (2012), which included national solidarity and the family; and Mozambique (2010, 2013, and 2015), which included governance and judicial system.


Except for Sao Tome and Principe (2015), no staff report for a program request had a specific mention of collaboration with the World Bank or other development partners when designing the program target on social and other priority spending. However, collaboration may have taken place and simply not reported on explicitly in the program documents.


The definition of social and other priority spending should ideally only include domestically financed spending so that shortfalls in externally financed project spending do not cause the targets to be missed. Only half of PRGT and one-third of GRA-supported programs clearly limit the target to domestically financed spending.


There are 6 program requests out of the 68 that incorporated conditionality on social safeguards measures other than the social and other priority spending IT. Five of these were fiscal adjustment programs.


For example, structural benchmarks included in the program requests for Burkina Faso (2010) and Haiti (2010) were on publishing poverty reducing expenditure. Armenia (2010) and Tanzania (2010) structural benchmarks aimed at improving the targeting or tracking and monitoring of social spending. Yemen (2014) had a structural benchmark aimed at improving the targeting of spending. Madagascar (2015) had a structural benchmark on the approval of the national social protection policy. Haiti (2010) was a fiscal expansion program while the others were fiscal consolidation programs.


Examples include a structural benchmark in (i) Yemen (2014) to mitigate the impact of the domestic fuel price hike by using the freed-up fiscal space to increase the transfers to the poor; and (ii) Madagascar (2015) to approve a social protection policy to mitigate the impact of maintaining an automatic fuel price mechanism.


However, this increases to 60 percent for LIC programs with fiscal consolidation (Table 2).


Baldacci and others (2008) find that both education and health spending in developing countries support higher growth. Gupta, Verhoeven, and Tiongson (2002) also find evidence that increased public expenditure on education and health in developing and transition economies is associated with improvements in both access to and attainment in schools, and reduces mortality rates for infants and children. Furceri and Zdzienicka (2011), find that social spending in health and unemployment benefits in OECD countries seem to have the greatest impact on economic activity.


For example, the IMF-ILO collaboration on the Social Protection Floor Initiative (SPFI) that started in 2010 was piloted in several countries (including Vietnam, Mozambique, El Salvador, and Burkina Faso). This collaboration helped the Fund gain access to detailed assessments of the social protection mechanism in place and provided valuable input into discussions with the authorities on protecting social spending (see Annex III discussion on the Mozambique experience). The impact of the selected protection measures on reducing poverty was also assessed. The feasibility and sustainability of each of the identified options were then assessed within the existing and projected fiscal space assessment provided by Fund staff.


Ideally, a comprehensive analysis of the distributional impact of adjustment/reform packages would be undertaken to inform the design of such measures—but this can be resource intensive and would likely need to be carried out in collaboration with development partners.


See: Gupta, and others, 2000.


For example, the World Bank’s Poverty and Social Impact Analysis (PSIA) is an approach to assess the distributional and social impacts of policy reforms on various groups of the population. The 2014 IEO evaluation found limited Bank-Fund collaboration on PSIAs in the post-2007 period. This is after the clarification provided in the 2007 policy that stated that Fund staff was not responsible for conducting PSIAs, but to rely on the analysis of other development partners, especially the World Bank (IMF, 2007).


At the conclusion of the discussion, the Managing Director, as Chairman of the Board, summarizes the views of Executive Directors, and this summary is transmitted to the country’s authorities. An explanation of any qualifiers used in summings up can be found here:



Dependent variable data is from Clements, Gupta, and Nozaki (2013).


Dependent variable data is from World Development Indicators.


Focus is on social safeguards conditionalities other than the social and other priority spending IT.


Focus is on social safeguards conditionalities other than the social and other priority spending IT.


Includes 6 staff reports where a PRGT country blended with GRA resources.


The GRA comparison group is limited to non-PRGT eligible countries who fall into the lower-middle income. ($1,026 and $4,035) and upper-middle income ($4,036 and $12,475) groups defined by the World Bank.


Staff report list reflects documents downloaded as of end-May 2016.

Social Safeguards and Program Design in PRGT and PSI-Supported Programs
Author: International Monetary Fund