A. Conclusions
72. Traditionally, the IMF’s role in social protection was limited and its approach fiscal-centric. Social issues were understood to lie outside the IMF’s core areas of responsibility. Formal guidance gave staff some latitude to decide when such issues were sufficiently important to warrant attention in surveillance or programs, and IMF involvement generally aimed to improve expenditure efficiency and/or to ensure medium- or long-term fiscal sustainability. Within this fiscal approach, staff addressed concerns for the poor and the vulnerable by recommending that social benefits be effectively targeted to those most in need. On more specialized issues, such as the design and implementation of social protection schemes, the IMF relied on the World Bank or other institutions with the relevant expertise, per the Board’s direction.
73. Starting in the 1990s but particularly over the past decade, the IMF has given greater attention to social protection from a widening perspective. This shift has responded to a range of challenges: dealing with the aftermath of the global financial crisis, addressing concerns from the impact of food and fuel price shocks, and recognizing the social and political threats to macroeconomic stability arising from strains on low-income groups and the most vulnerable. Thus, social protection has increasingly been seen as “macro-critical” going beyond purely fiscal concerns. This increased concern for social protection has been evident in IMF surveillance, programs, and TA, most notably in the number of SIPs in surveillance and related conditionality in programs. Social protection issues have often been on the table in policy advice and program design, while the IMF continued to rely on partner institutions for the detailed design and implementation of social protection policies.
74. While recognizing the increased overall attention to social protection issues, this evaluation found various levels of IMF involvement across countries and over time. In some cases, engagement was quite intensive and spanned different activities (surveillance, TA, and/or programs): country teams analyzed the possible adverse impact of policy measures, reforms, or shocks on vulnerable population groups; discussed possible policy options to mitigate such adverse impacts; pressed the authorities to enhance social protection; incorporated additional spending on social protection measures in the design of IMF-supported programs and/or specified program conditionality based on these measures; monitored progress in implementing social protection measures; and followed up when implementation fell short. In other cases, IMF engagement was much more limited, confined to emphasizing to the authorities the relevance of paying attention to the need for protection of vulnerable groups and increasing fiscal resources for related expenditures, but with little detailed analysis or follow-up.
75. This variation to some degree reflected an appropriate response to country-specific factors, but also idiosyncratic factors. The extent of existing social tensions and stresses on vulnerable groups, as well as fiscal problems arising from the need for more efficient public expenditures and fiscal risks arising from medium- to long-term demographic changes, played a role in determining if social protection was judged to be macro-critical in a particular country context, triggering IMF involvement in these issues. The availability of expertise from development partners and within the country itself also affected the extent of IMF involvement. However, idiosyncratic factors seem also to have played a part, particularly in the context of surveillance, as staff had different views as to what kind of work they were expected to do in this area and on the IMF’s role in social protection, and had varying degrees of expertise and interest in these issues. Thus, attention to social protection sometimes shifted over time with changes in country team members as well as country circumstances.
76. In the program context, the IMF almost always took account of social protection concerns but efforts to mitigate potential adverse effects of program measures on the most vulnerable had mixed success in implementation. The most effective implementation outcomes were in cases where the IMF lent its support to a package of measures that the authorities themselves had developed (often with the assistance of the World Bank or other partners) to shield the vulnerable; but this was clearly not a variable within the IMF’s control. Particularly challenging for the IMF were situations where existing social protection systems were not well developed to start with or where political resistance was high and/or implementation capacity was low. In some cases, program conditions to maintain or increase social protection expenditures were not well received by country authorities who viewed such requirements as an additional constraint on their already limited degrees of freedom to meet fiscal targets. In some cases, it took much longer than anticipated to build needed institutional capacity, e.g., for well-targeted transfer programs. The nationally defined social and other priority spending floors required in IMF-supported LIC programs were often too broad or insufficiently specific to be very useful in protecting critical spending for social protection.
77. In surveillance, as the IMF moved beyond the fiscal-centric approach to reflect broader considerations for addressing social protection, its policy advice was of varying depth. To be sure, the increasing number of SIPs on social protection some times contained excellent analysis. However, at other times, it seemed that attention to social protection devolved into a box-ticking exercise as staff tried to pay due attention to an increasingly broad range of policy issues. Staff often relied on a mechanical benchmarking of social expenditures (e.g., against OECD averages), or citing IMF cross-country research on income inequality and growth performance, without substantial analysis or explanation to make a solid link to policy recommendations on social protection. Many country authorities interviewed for this evaluation indicated that the IMF’s recommendations on social protection often did not reflect sufficient knowledge of local conditions or their specific cultural norms and societal preferences.
78. The IMF’s preference for targeting social protection to those most in need was broadly in line with the World Bank’s approach during the evaluation period, which made for generally good cooperation between the two institutions. The evaluation found an effective division of labor between the Fund and the Bank. The Bank had the lead in the design and implementation of social protection schemes, and the Fund provided valuable support by highlighting the need for such measures at Finance Ministries and helping authorities identify space in fiscal programs to ensure that fiscal sustainability was not compromised.
79. However, the targeting approach meshed less well with the rights-based approach to social protection espoused by UN agencies including the ILO. This difference in viewpoints—rooted in different mandates of and different legal frameworks applying to the respective institutions—posed a challenge to the IMF’s attempts at institutional-level collaboration with the ILO and UNICEF. To those promoting the rights-based approach, the IMF as an institution often came across as dismissive of the approach and uncooperative. At the end of the evaluation time-frame, as part of its undertakings on the SDGs, the World Bank joined the ILO in support of universal social protection, which could lead to complications for Bank-Fund collaboration going ahead depending on how this commitment is implemented at the Bank. Moreover, the IMF’s endorsement of the SDGs has raised questions about consistency with its continued support for targeted (means-tested) social protection schemes.
80. The IMF’s external communications tried to emphasize the Fund’s “human face,” but this created heightened expectations among external stakeholders regarding the IMF’s role in social protection that were sometimes disappointed. Against such expectations, some external commentators concluded that the IMF’s claims regarding its increasing attention to social protection often were not matched by the level or intensity of its efforts on the ground. This may have reduced recognition of the greater efforts the IMF has made to enhance its involvement in social protection over the past decade.
B. Recommendations
81. Looking forward, concerns about inequality, social and political stability, the impact of trade, immigration, and new technologies on vulnerable groups, and the consequences of aging populations seem likely to keep social protection issues high on the global policy agenda. This evaluation concludes with a number of recommendations to further enhance the IMF’s effectiveness in this area.
Recommendation 1: The IMF should establish a clear strategic framework to guide its involvement in social protection among multiple competing priorities.
82. The Executive Board should endorse a clear strategy setting the scope, objectives, and boundaries of IMF engagement in social protection. Board endorsement of a strategy for the IMF’s role in social protection would be an important signal to staff to guide priorities at a moment when many new institutional initiatives are being raised as potentially “macro-critical” and deserving attention. Without such clarity, staff risk losing focus in their country work and the IMF risks appearing inconsistent in its policy advice or insensitive to local conditions.
83. To support the discussion, staff should prepare a paper laying out the overarching issues for the Board’s consideration. The purpose would be to reach an institutional view endorsed by the Board setting out the scope, boundaries, and objectives of IMF engagement in social protection that would serve to guide staff’s operational work and dialogue with country authorities.
84. Key issues would include the following:
-
▶ The definition/scope of social protection relevant for the IMF. The first step would be to adopt a common understanding and usage of the term “social protection” within the IMF. This would eliminate the variety of terminology that now exists as well as assist in determining which policy areas fall under the rubric and which do not. In turn, this would provide clarity and transparency for internal and external discourse.
-
▶ The objective(s) of IMF involvement in social protection. That there are instances where the IMF should be involved in social protection is beyond doubt. However, while macro-criticality has been the operative criterion for determining IMF engagement on structural policy issues, it remains a somewhat nebulous standard that does not provide a clear working guide for when the IMF should or should not become involved in social protection. Staff need clearer strategic direction on how to identify those instances where social protection should be prioritized among multiple competing potentially macro-critical issues, and where it can be given less attention. A relatively narrow approach would be to focus on two well-established (macroeconomic) objectives of IMF involvement: to ensure that a country’s social protection policies are efficient and fiscally sustainable and to mitigate short-run adverse effects of IMF-supported adjustment programs on vulnerable groups that could erode public support for the programs. A broader approach would be to provide advice on policies to support vulnerable and low-income groups where concerns about social and political strains and inequality are judged to be material risks to economic growth and stability. A central challenge would be to assess when IMF involvement on this broader scale should be given priority, taking account of local conditions and preferences and limits on staff resources and expertise.
-
▶ The boundaries of IMF involvement in social protection. A central issue to address is the appropriate boundaries of IMF involvement in social protection, including to consider how the Fund’s work in this area fits within its legal framework. Social protection is a multi-dimensional issue, where the social and political dimensions are just as, if not more, important than the macroeconomic dimension. Questions to consider include: to what extent should IMF advice be guided or constrained by domestic social and political preferences? To what extent should the IMF embrace an overarching aspiration such as “universal social protection”? Should the IMF address social protection issues in the absence of an obvious or foreseeable fiscal/macroeconomic problem, for example, should staff be expected to advise countries on the size of the social welfare system or to assess the social sustainability of a country’s public pension system (including questions of inter- and intra-generational equity) in the absence of fiscal sustainability concerns? Clarifying the boundaries of IMF involvement in social protection would facilitate the setting of appropriate expectations—internally and externally—as to what the IMF will and will not be accountable for, and it would allow staff to better focus their efforts in this area.
Recommendation 2: When social protection is determined to be a macro-critical priority under the strategic framework, the IMF should provide tailored policy advice based on in-depth analysis of the particular country situation. The advice would draw on work by development partners or country authorities where available, but in its absence, the necessary analysis may need to be undertaken in-house.
85. Given the extent to which the nature and scope of social protection vary across the membership, the IMF’s advice on social protection issues must necessarily be based on country-specific analysis. For example, if reducing income inequality is identified as a macro-critical objective, the analysis should show, in the context of the country concerned, how the social protection system has contributed to inequality, how proposed social protection reforms might be expected to help, and whether there are more effective or less distortionary ways of achieving greater equity. Simply pointing out the risks to social stability of excessive inequality based on international experience is not sufficient to convince country authorities to embark on politically sensitive reforms. Similarly, listing examples from the rest of the world is of limited usefulness to countries seeking IMF advice on relevant reforms. A more coherent analysis tailored to country specifics is necessary if the IMF is to make a constructive contribution on social protection. And since reforms take time to put in place, such discussions should best take place in the surveillance context rather than in a program context in the midst of a crisis.
86. To the extent possible, advice on social protection should draw on work by development partners or by country authorities, but in-house expertise will sometimes be needed where social protection is judged to be a macro-critical priority. Institutions like the World Bank have much more experience and expertise on social protection and are often engaged on these issues in a given country context. Country authorities may have access to relevant data sets and will know well social, political, and legal constraints. However, there will be times when outside support will not be easily available on a timely basis, and the Fund will need to find the necessary resources in-house or to catalyze external partners to avoid the risk that macro-critical social protection issues are neglected in countries where they are judged to be a priority.
87. Resource implications of Recommendations 1 and 2 would need to be considered. Addressing social protection issues from a broader perspective and/or in greater depth in a significant number of members would require expertise and experience well beyond what is currently available (or can be easily repurposed) in the Fund. A relatively “resource-lite” approach to social protection by the Fund would work best with full commitment and a compatible approach by partner institutions like the World Bank or the availability of own resources in the country concerned, neither of which may be taken for granted.
Recommendation 3: The IMF needs to find more realistic and effective approaches to program design and conditionality to ensure that adverse impacts of program measures on the most vulnerable are mitigated.
88. The next Conditionality Review should consider how program design and conditionality can be more effectively applied to mitigate the impact of adjustment on vulnerable groups. The present evaluation—pointing to the mixed implementation record of social protection policy measures in the face of capacity constraints and political obstacles—suggests the need for the IMF to be more realistic in assessing implementation constraints. Building on the recent Board paper on “social safeguards” in LIC programs (IMF, 2017b), the Review should also examine IMF-supported programs in EMEs and advanced economies, covering the full range of social protection policies, and suggest how program design and conditionality can more effectively fulfill the purpose of protecting vulnerable groups. The proposed guidance note for staff on social safeguards put forward in IMF (2017b) should include good practices for addressing social protection concerns not just in LIC programs but in Fund-supported programs across the membership.
Recommendation 4: In external communications, the IMF should realistically explain its approach to social protection issues, and what it can and cannot do in social protection, given its mandate and limited resources and expertise.
89. The IMF would be better placed to receive greater recognition for its genuine efforts to tackle social protection issues if it sets realistic expectations. The IMF should be clear on how its approach to social protection issues fits with its mandate and available resources, and how its emphasis on targeting meshes with the rights-based approach. Clear external communications of the objectives, scope, and boundaries of the IMF’s involvement in social protection following the Board review recommended above will help to temper expectations of stakeholders and avoid reputational risk to the Fund.
Recommendation 5: The IMF should engage actively in inter-institutional cooperation on social protection to find ways to work constructively with development partners, particularly institutions with different mandates and policy priorities.
90. In an area such as social protection where the IMF is not a global leader and must rely heavily on other agencies for in-depth expertise, there is no alternative to the Fund cooperating, and being seen to cooperate, with others willingly and constructively. For the IMF to play a more effective role in social protection, at the least it needs to ensure continued close working relations with the World Bank, and it should also be willing to work constructively with other partners based on realistic assessments and agreement on common goals at the institutional level. The IMF should commit to regular attendance by senior-level staff at relevant inter-institutional meetings such as those of the SPIAC-B where efforts are being made to reconcile the targeting approach and the rights-based approach to reaching the goal of universal social protection. Such involvement would be helpful not just for good public relations but to keep up with developments in the field, to maintain open lines of communication with institutions that have different mandates and policy priorities, and to allow for the IMF’s particular perspective on social protection to be influential in the broader debate.