3. IMF Operational Involvement in Social Protection
- International Monetary Fund. Independent Evaluation Office
- Published Date:
- October 2017
A. Bilateral Surveillance
32. The IMF addressed an extensive range of social protection issues in Article IV consultations with a number of countries during the evaluation period. To measure coverage, this evaluation examined Board assessments in Article IV Summings Up for advice related to social protection.42 Based on Summings Up examined for all Article IV consultations concluded in 2006, 2008, 2010, 2013, and 2015, this evaluation identified seven themes of IMF advice that were closely related to social protection;43 (i) reforming the pension/social security system; (ii) reforming unemployment benefits/minimum wage schemes; (iii) improving the targeting of social benefits/transfers; (iv) protecting vulnerable groups or limiting the social cost of reforms/policies/shocks; (v) protecting or creating fiscal space, i.e., increasing budgetary allocations, for social spending; (vi) strengthening the social safety net/social transfers/provision of social services; and (vii) pursuing active labor market policies. Examples of each theme are provided in Annex 2. Many Summings Up contained advice on more than one theme.44
33. Coverage of social protection issues was high. On average, around 60 percent of all Article IV Summings Up in the five years examined contained social protection advice in one or more of the seven themes. There were only four countries where no social protection advice appeared in any Article IV Board assessment during the entire 2006–15 period (Afghanistan, Argentina, Kosovo, and Somalia).45 Coverage of social protection issues was highest in advanced economies and lowest—but rising—in LICs (Figure 4A.) Coverage varied across regions but was relatively more frequent in countries covered by the Western Hemisphere Department (WHD), the Middle East and Central Asia Department (MCD), and the European Department (EUR) (Figure 4B).
Figure 4.Incidence of Social Protection Advice in Article IV Consultations
34. The topics of IMF advice on social protection varied across country income groups and regions. Most advanced economies have comprehensive social protection systems centered on social insurance schemes, such as contributory pensions and unemployment insurance, whereas most LICs do not.46 Not surprisingly, therefore, the IMF’s coverage of social protection in advanced economies focused mainly on modifying aspects of the existing social protection system (e.g., reforms to the pension and/or social security system, reforms to unemployment insurance, and greater use of active labor market policies), while coverage in LICs emphasized the need to protect vulnerable groups against the impact of reforms or external shocks and to develop or expand social protection schemes (Figure 5). Coverage in EMEs was more evenly distributed across key identified topics. As for coverage across regions, pension and social security reforms were the dominant theme in EUR countries, while protecting vulnerable groups and expanding social spending were the dominant themes in AFR, APD, and MCD countries.
Figure 5.Types of Social Protection Advice in Article IV Summings up
35. IMF country teams provided advice on social protection policies largely with the aim of addressing fiscal, labor-market, and other core macroeconomic concerns, but also to reduce income inequality:
▶ Fiscal consolidation and long-term fiscal sustainability: Concerns about large fiscal deficits and growing public debt in many advanced economies (and some EMEs) prompted the IMF to press for fiscal consolidation in the short and medium term; in pushing for fiscal retrenchment, the IMF often stressed entitlement reform as a key component (e.g., Bulgaria, United States). Long-term fiscal sustainability was the principal factor motivating the IMF’s work on reforming public pension schemes in advanced economies and EMEs (e.g., Cyprus, Dominican Republic, El Salvador, Ireland, Japan). As noted in Heller (2017), many of these countries are rapidly aging and will have a high elderly dependency burden by 2050, raising the prospect that they could face social stresses as their public sectors are exposed to significant future explicit or implicit pension liabilities borne by a shrinking active labor force.
▶ Expenditure efficiency: In many EMEs and LICs, the IMF recommended targeted social protection policies to improve expenditure efficiency while protecting the vulnerable. For example, a common strand of advice in these countries was to reduce or eliminate generalized food and/or energy price subsidies in favor of “better targeted” social safety net programs. While such price subsidies are not usually categorized as social protection policies, in many countries they have a social protection element. As discussed in Feltenstein (2017), during the evaluation period the IMF recommended food or fuel price subsidy reforms in up to a quarter of all Article IV Summings Up across the membership, with the highest incidence of such advice observed in MCD countries (e.g., Egypt, Jordan, Morocco, Tunisia) (Figure 6A).47 The same Summings Up also contained advice to introduce measures to limit the negative impact of price subsidy reform on vulnerable groups in about 50 percent of the cases, and advice to use the fiscal savings from price subsidy reform for social spending in up to 23 percent of these cases (Figure 6B).48
▶ Unemployment, employment, and labor force participation: The impact of the global financial crisis shifted the IMF’s attention toward active labor market policies and unemployment benefit reforms to address protracted, elevated levels of unemployment in many advanced economies and some EMEs. In these countries, the IMF advised reducing employment protection and expanding social protection (“flexicurity”) to lessen labor market duality (e.g., Korea); lowering social security contribution rates (reducing the labor tax wedge) to stimulate employment demand (e.g., Bulgaria); supporting job search and skills acquisition (e.g., Latvia); or changing pension benefit provisions that could contribute to workers dropping out of the labor market or opting to limit their participation (e.g., Japan).
▶ Domestic saving: The IMF encouraged some advanced and emerging market economies to strengthen social safety nets to reduce domestic saving. This motivation—which featured also in multilateral surveillance, notably External Sector Reports—was based on the argument that “weaker safety nets tend to distort saving rates upwards” resulting in excess current account surpluses that contribute to global imbalances (IMF, 2015b). This argument was usually applied to countries in Asia (e.g., China, Korea, Malaysia, and Singapore).
▶ Income inequality: The IMF encouraged some countries to strengthen social safety nets to reduce income inequality (e.g., Bolivia, Ethiopia, Korea). This motivation largely drew on IMF research on inequality and growth, notably Berg and Ostry (2011), and the policy implications suggested therein for improving income distribution, including social protection policies such as social assistance spending and active labor market policies.
Figure 6.Food and/or Fuel Subsidy Reform Advice in Article IV Consultations
36. IMF staff have provided increasing background and analysis on social protection issues in Selected Issues Papers as part of Article IV consultations. On average, over one-fifth of Article IV consultations featuring social-protection advice in the Summing Up in 2006, 2008, 2010, 2013, and 2015 had one or more SIPs focusing on social protection issues. The share of Article IV consultations with one or more social-protection-related SIPs in the total number of consultations with accompanying SIPs rose from an average of 23 percent in 2006–11 to an average of 45 percent in 2012–15.49 In terms of papers, the share of social-protection-related SIPs in the total number of SIPs rose from an average of 9 percent in 2006–11 to an average of 15 percent in 2012–15 (Figure 7A). Most of these SIPs were prepared for advanced and emerging market economies (Figure 7B); some countries had more than one social-protection-focused SIP in the same Article IV consultation. By contrast, the share of SIPs discussing social protection issues in LICs was relatively small.50
Figure 7.Selected Issues Papers Discussing Social Protection Issues
37. In many advanced and emerging market economies in particular, these SIPs were a way for the IMF to contribute its analysis to an ongoing policy debate in the country. For example, pension reform loomed large in Article IV discussions during the evaluation period; most advanced economies had one SIP on the topic every few years, and over one-third of EMEs had one or more SIPs on pension reform during the past decade. Labor market reforms were another area where the IMF contributed to the policy debate, particularly in Europe but also increasingly in other advanced and emerging market economies facing high unemployment in the wake of the global crisis. In a number of EMEs, SIPs discussed social protection policies in the context of improving expenditure efficiency and, after 2010, in the context of achieving “inclusive growth.” In LICs, SIPs on social protection were usually part of the discussion on poverty reduction as well as inclusive growth.
38. In line with its mandate, the IMF’s Article IV analysis of social protection issues generally had a macroeconomic rather than a social focus. For example, Heller (2017) notes that while IMF staff focused on macro-critical issues associated with existing pension systems, such as fiscal sustainability and the short- term expenditure burden, they typically did not address social issues such as the extent of pension coverage in the population or the adequacy of the pension replacement rate. The social sustainability of the pension system (i.e., the extent to which social protection objectives were accomplished and perceived as satisfactory by citizens) was rarely analyzed. Interviews for this evaluation revealed that often these social issues related to fundamental decisions that were considered by staff and authorities alike to lie outside the IMF’s mandate.
39. Staff’s analysis sometimes lacked much depth as it moved beyond “core macroeconomic” concerns. IMF staff did quite detailed analytical work on a range of issues, as noted above, but coverage became more formulaic as staff tried to link social protection with some of the newer areas of IMF emphasis, such as addressing income inequality. For example, this evaluation found little useful analysis to support generic exhortations to increase social protection to promote “inclusive growth.” A review of SIPs on inclusive growth for EME Article IV consultations during 2006–15 found very few that substantially addressed the link between growth and poverty or inequality in a specific country, let alone how social protection fit into the picture.51 Several papers simply referred to Berg and Ostry’s (2011) research on the link between inequality on growth spells without explaining its relevance to the country in question; some relied on rather mechanical aggregate benchmarking against other countries done by others, without providing their own analysis. Staff explained to the IEO that data limitations, specifically the availability of comprehensive household survey data, restricted their ability to conduct a thorough customized analysis in many countries. The first batch of “inequality pilot” Article IVs contained a deeper analysis of inequality and poverty outcomes, but very few (e.g., Bolivia) included a serious attempt to use the analysis to build beyond the rather generic recommendations for social protection offered during the consultations.52
40. Staff’s analysis sometimes lacked sufficient country-specific knowledge. For example, the evaluation found that IMF staff, in advising some advanced economies (e.g., Korea and Singapore) to reinforce/expand social welfare programs, could have done more to incorporate societal preferences on the size and scope of the social protection system. Similarly, IMF advice to countries to “improve the social safety net” in order to lower the domestic saving rate was not found, in most cases, to be supported by much explanation or analysis as to what kind of social safety net measures should be improved, how they would affect the domestic saving rate, and by approximately how much. Instead, such advice seemed to be simply drawn from a comparison with the “social spending norm” calculated in the IMF’s external balance assessments.53 In LICs, outside experts interviewed by the IEO observed that IMF staff tended to “mechanically” recommend conditional cash transfers without analyzing if other types of transfers might be more appropriate for the particular country or situation.
41. Country authorities interviewed for this evaluation expressed mixed views about the IMF’s effectiveness in the surveillance context. Many country authorities interviewed for this evaluation did not recall having any significant dialogue with the IMF on social protection. Those who did sometimes noted that when the IMF’s recommendations on social protection were generic or based on a mechanical application of an empirical methodology or benchmarking, they were often seen as not being particularly relevant in their country circumstances and not sufficiently respectful of their country’s social and cultural characteristics and values. In fact, some country authorities indicated that IMF projections, e.g., about long-run fiscal sustainability, that were insufficiently tailored to country circumstances could do more harm than good (by “scaring the public” in the words of one Executive Director). On the other hand, when IMF country teams delved more deeply into the issues (including meeting with the appropriate country experts on such matters) and presented critical analysis in a sensitive way, pointing out inequities, misallocations, and best practices, their work was considered to have value-added in informing the policy debate in the country. That said, however, not many country authorities interviewed were familiar with staff analyses contained in SIPs, and it is not clear that these analytical contributions were followed through with much broader outreach.
42. The IMF approved over 170 arrangements during 2006–15 (Figure 8). In 2008, Iceland became the first advanced economy in more than three decades to receive IMF financial support. As the crisis spread in Europe, the IMF extended financial support to four euro area members. Among EME member countries, almost 40 percent had IMF-supported programs approved during the evaluation period, most of them to help cope with the effects of the global recession. Just over half of all low-income member countries had IMF-supported arrangements approved during the evaluation period. Most of these countries had multiple arrangements, and most of the arrangements were concessional, funded by the Poverty Reduction and Growth Trust (PRGT).
Figure 8.IMF Arrangements, by Year Approved
Note: Includes arrangements under: Extended Credit Facility (ECF); Extended Fund Facility (EFF); Exogenous Shocks Facility (ESF); Flexible Credit Line (FCL); Precautionary Credit Line (PCL; replaced by the PLL in 2011); Precautionary and Liquidity Line (PLL); Poverty Reduction and Growth Facility (PRGF; replaced by the ECF in 2010); Policy Support Instrument (PSI); Stand-By Arrangement (SBA); Standby Credit Facility (SCF). Blended arrangements were counted as one.
43. On average, around 10 percent of IMF-supported arrangements approved during 2006–15 included structural conditionality explicitly to strengthen or better target social protection (Figure 9). This number is based on a review of the IMF Monitoring of Fund Arrangements (MONA) database for prior actions, structural performance criteria, or structural benchmarks classified under “other social sector reforms (e.g., social safety nets, health, and education).”54,55 These measures typically sought to improve the targeting of social protection programs and/or increase their coverage to protect the vulnerable during fiscal adjustment. They were found in advanced economy, EME, and LIC programs. Examples include structural benchmarks to implement a new (better targeted) social protection system (Armenia: 2009 SBA; Cyprus: 2013 EFF); increase the coverage of the public conditional cash transfer system (Dominican Republic: 2009 SBA; Paraguay: 2006 SBA); and design a strategy to strengthen the social safety net (Latvia: 2008 SBA; Pakistan: 2008 SBA).
Figure 9.IMF Arrangements with Conditionally Aimed at Strengthening Social Protection
Source: IEO analysis, drawing on IMF MONA database and program documents.
44. An increasing share of LIC programs included social and other priority spending floors, though the objective of these tar gets typically went much beyond supporting social protection.
▶ In the earlier part of the evaluation period, half of the 36 PRGF-supported arrangements approved during 2006–09 included an explicit floor on nationally-defined poverty-reducing (or “pro-poor”) social expenditures, monitored as a quarterly indicative target. These spending floors were motivated by countries’ commitments to attain the 2015 UN Millennium Development Goals (MDGs), although social protection was not explicitly part of the MDGs.56
▶ The share of arrangements with social and other priority spending floors rose to 93 percent of the 57 arrangements approved during 2010–15 under the post-2009 LIC facilities.57 However, only 19 percent of these arrangements (10 arrangements in 7 countries) contained indicative targets defined to focus primarily and specifically on social protection.58 The definitions of the indicative targets in the rest of the arrangements either did not include expenditures on social protection or were too broad or insufficiently specific to determine if social protection expenditures were a meaningful component.
45. A few EME arrangements also included an indicative target (floor) on social spending (see Figure 9). In most cases, the floor applied to spending on specific social protection programs, reflecting the intention to protect these programs (and the vulnerable populations they served) in the face of budgetary retrenchment (e.g., Armenia: 2014 EFF; Jamaica: 2013 EFF; Pakistan: 2013 EFF). There were no social spending floors in the advanced economy programs. Some advanced economy and EME programs explicitly incorporated social safety net expenditures in the macroeconomic framework without specifying formal conditionality (e.g., Iceland: 2008 SBA; Romania: 2009 SBA; El Salvador: 2010 SBA).
46. In advanced economy programs, the IMF mainly played a supportive rather than lead role in efforts to minimize the social costs of adjustment. IMF press releases for the programs in Greece, Ireland, Portugal, and Cyprus emphasized the need to protect vulnerable groups and/or strengthen social safety nets. In Iceland and Ireland, the two cases where program ownership was highest, the authorities themselves played a key role in designing a fiscal adjustment plan consistent with their social objectives including the maintenance of universal benefits. The IMF was generally supportive of the authorities’ plans, although it endorsed Iceland’s Nordic welfare state model but urged Ireland to move away from universal benefits towards targeted support to the most vulnerable groups.59 In the other countries, IMF staff addressed macro-critical social protection issues (e.g., pension reform in Greece) in coordination with the European Commission (EC). Only in Cyprus was the IMF actively involved through structural conditionality and related TA to reform the social protection system (with assistance from World Bank experts).
47. In EME programs where the IMF played an active role on social protection, it did so in collaboration with partner institutions. In all the EME program country cases studied for this evaluation, specific social safety net measures were designed with the assistance of the World Bank or other institution(s). The IMF’s main contribution was to embed the measures in its program design and to insert structural benchmarks in the programs if necessary to keep the reforms on track (e.g., in the Dominican Republic’s 2009 SBA). In interviews for this evaluation, World Bank staff were uniformly appreciative of the IMF’s support and rated highly their cooperation with IMF staff.
48. In LIC programs, the IMF sometimes played a significant advocacy role in favor of strengthened social safety nets. In most cases, the World Bank and/or other institutions were already involved in assisting the country authorities in setting up or expanding the social safety net, including designing and implementing means-testing mechanisms. Interviews with Fund and Bank staff usually revealed close cooperation in this area, especially in the field. Fund staff were aware of and generally supported the Bank’s work in this area although they did not always report on its details in staff reports. IMF staff with experience in African LICs noted that finance ministers were sometimes unenthusiastic about cash transfers and skeptical that targeted schemes could work.60 Importantly, while the World Bank and/or other development partners may have worked with the relevant ministry on the practicalities of improving the social safety net, the IMF had a direct line to the more powerful finance ministry. Staff of the World Bank and other institutions (such as UNICEF and the ILO) noted that recommendations from the IMF tended to carry more weight with those in power within the government.
49. Concretely, the IMF’s role in many LIC programs was to try to find fiscal space in the macroeconomic framework for social protection expenditures. In identifying expenditures to cut to make room for enhancing the social safety net, IMF staff often zeroed in on energy price subsidies.61 Thus, a frequent theme of IMF advice—which was in line with the World Bank’s approach—was to reduce and eventually abolish energy price subsidies to create fiscal space for a well-targeted social safety net.62 Where staff did manage to carve out fiscal space for social protection, the amounts were typically modest—well under 1 percent of GDP, for example. In one case (Mongolia), the IMF-supported program included conditionality to shift from universal to targeted social transfers.
50. How effective was the IMF’s involvement in social protection in the program context? This evaluation did not undertake an impact assessment of IMF interventions on the welfare of vulnerable groups, which would have required a much larger study. Instead, it focused more narrowly on the authorities’ perceptions of the usefulness of the Fund’s advice on social protection and the extent to which the advice was implemented. On this basis, among advanced economies, the IMF’s involvement was more effective in Cyprus, where a guaranteed minimum income scheme was introduced to support vulnerable groups affected by the crisis (a structural benchmark under the 2013 EFF-supported program) than in Portugal and Greece. According to Portuguese authorities interviewed for this evaluation, IMF staff were primarily focused on fiscal consolidation during the program and they evinced little interest in the social impact of reform measures, preferring to leave such matters to the European institutions. Former Greek authorities criticized the IMF for lacking sufficient understanding of the country’s social protection system (see Heller, 2017).
51. In EMEs, the IMF’s involvement helped to strengthen social protection in some cases but not in others. Country authorities were pleased when the IMF supported their proposals to scale up social protection during an adjustment program (e.g., in the Dominican Republic and El Salvador). However, the IMF’s efforts to protect the vulnerable from adverse effects of program measures were not always well-received by authorities and conditionality to that effect was not always met. In some countries (e.g., Latvia), the authorities pushed back against the IMF’s advice to raise social spending, arguing that the existing social safety net was adequate and that a slower fiscal adjustment would only prolong the pain. In others (e.g., Tunisia), implementation capacity constraints and administrative delays led to the indicative target on social spending being missed repeatedly. Country authorities interviewed for this evaluation said they were surprised to see the IMF taking such an active interest in protecting social spending in Fund-supported programs and some felt that the indicative targets placed an unwelcome additional constraint on their ability to meet what were usually already ambitious fiscal balance targets.
52. In LICs, the IMF’s advocacy of social protection was generally highly appreciated by its development partners. Staff of other institutions including the World Bank, UNICEF, and the ILO indicated to this evaluation that the IMF’s involvement in social protection gave greater visibility to the issue in the country and more importantly, helped to secure budgetary funding for their proposed reforms (see Klugman and others, 2017). While some commentators criticized the IMF for paying lip service to social protection—in the form of the now ubiquitous phrase “… while protecting the most vulnerable in society”—many development partners and CSOs interviewed for this evaluation said that this was sufficient and appropriate as a contribution from the Fund. Many of them, in fact, suggested that the IMF should stop there and not go on to recommend “well-targeted” policies. On the other hand, some IMF staff told the IEO that they were not entirely comfortable advocating what seemed to them a “black box” of social protection measures.
53. However, the IMF’s efforts to promote better targeting of social protection in LICs had mixed results on the ground. IMF staff often underestimated the time and complexities involved in developing and implementing means-tested benefits and the administrative and political challenges inherent in excluding population groups from benefits. Staff did not always follow up consistently on the country’s progress in developing a better targeted social safety net, which some mission chiefs saw to be outside their domain. Sometimes the degree of attention to social protection in an IMF-supported program would shift over time, as other issues rose in urgency. The view from LIC authorities who spoke to the IEO was that their countries usually implemented social safety net programs independently of the IMF, with the assistance of the World Bank, UNICEF, and/or other development partners. In some cases, replacing universal benefits with targeted schemes met with not just political but also cultural resistance and even when implemented, did not have lasting effects (e.g., Mongolia).
54. Social and other priority spending floors in IMF-supported LIC programs were generally not very useful for safeguarding social protection expenditures. As noted earlier, while over 90 percent of arrangements approved during 2010–15 under the post-2009 LIC facilities included social and other priority spending targets, these included a broad range of spending items and less than 20 percent of these spending targets focused primarily and specifically on social protection.63 Previous IEO evaluations also saw room for improvement in focusing these spending targets.64 Staff interviewed for this evaluation were well aware of the shortcomings of this indicator and the box-ticking nature of the monitoring exercise. In one case (Mozambique), staff simply stopped monitoring the indicative target, explaining that it was basically of no use in protecting critical social spending. But there were better experiences too, e.g., in Honduras, where staff supported a new indicator more narrowly focused on social protection that was proposed by the authorities in the 2010 SBA/SCF-supported program, and included an adjustor to allocate a portion of any excess tax revenue over projected amounts to such spending.
55. Did IMF-supported programs inadvertently “do harm” to social protection? This evaluation did not assess whether social protection increased or decreased as a result of IMF-supported programs due to conceptual and attribution problems.65 However, the evaluation did find that the IMF, in pressing for social-protection-reducing reforms (such as pension reform or energy subsidy reform) in a program context, often sought to mitigate potential adverse impacts on vulnerable groups.
▶ Pension reforms: As shown in Figure 10, conditionality pertaining to reforms of the pension and/or social security system was a feature of many programs. These reforms were usually motivated by short- and medium-term budgetary pressures in the context of fiscal adjustment, and benefit cuts essentially involved reducing protection for certain population groups. In many cases, staff addressed distributional concerns by proposing that reforms protect the pensions received by the bottom group of pensioners and/or limit the pensions received by the top group or “privileged pensioners” although the authorities did not always agree with staff (Heller, 2017). Program measures involving pension cuts were usually very contentious—Greece being one of the most challenging cases for the IMF—and, in several European countries, ended up being reversed by the Constitutional Court because they were judged to violate the acquired rights of pensioners.66
▶ Energy subsidy reforms: While reforms of generalized energy price subsidies were often promoted by the IMF to finance more efficient social protection (i.e., expenditure reallocation), in many LIC and EME programs they were needed first and foremost for expenditure rationalization (see Figure 10). These energy subsidy cuts were often viewed by governments as weakening social protection.67 This evaluation found, based on case study evidence, that even when expenditure rationalization was the primary reason for reducing energy price subsidies, IMF staff also paid attention to the social implications and suggested targeted mitigating measures for the poor.68 In most cases in practice, country authorities chose to retain certain subsidies such as those on fuel products commonly used by the poor, lifeline electricity tariffs for low-usage consumers, public transportation vouchers, etc., and IMF staff, recognizing their pragmatic benefits, did not object (Feltenstein, 2017).69
Figure 10.IMF Arrangements with Conditionality on Pension Reform and Energy Subsidy Reform
Source: IEO analysis, drawing on IMF MONA database.
C. Technical Assistance
56. During the evaluation period, FAD provided TA related to social protection issues to over 60 countries.70,71 Some countries received multiple TA missions. The provision of TA by the IMF reflects both demand (for assistance by country authorities) and supply (of resources in FAD).72 Although historically IMF TA was concentrated in EMEs and LICs and rarely needed in advanced economies, during the evaluation period, the IMF provided TA in these areas to some crisis-struck advanced economies (Figure 11). About 9 percent of TA reports addressing social protection issues were for advanced economies, with the rest roughly equally divided between EMEs and LICs. On average, around 45 percent of TA reports addressing social protection issues were prepared in the context of an ongoing program.
Figure 11.IMF TA Reports Addressing Social Protection
57. The IMF was by no means the central player providing TA on social protection. The World Bank and other development partners—UN agencies (particularly the ILO and UNICEF), regional development banks such as the Inter-American Development Bank (IDB) and the Asian Development Bank (ADB), and bilateral aid agencies, among others—were much more actively involved in LICs and EMEs, with specific agendas for poverty reduction, sectoral policies, and the extension of social protection. These institutions provided technical support on social protection programs, including design and implementation details, as well as monitoring and evaluation. Most advanced economies and many EMEs had sufficient domestic capacity and/or access to other resources (e.g., private consultants) to analyze social protection issues without IMF support.
58. IMF TA missions addressed social protection policies both directly and indirectly. The direct focus included assistance on reforming the pension system or the social safety net. In addition, TA missions whose primary purpose was to advise on options for expenditure rationalization, food or energy price subsidy reform, or tax reform sometimes touched on social protection policies as well. Both types of TA focused primarily on the fiscal dimension. They typically looked at distributional implications but did not dig deep into specific design aspects of social protection programs or questions such as the adequacy of program coverage or benefits. In that respect, there was little overlap with the World Banks work.
► Expenditure rationalization (usually in the context of an ongoing program) was the main avenue through which TA addressed social protection policies in advanced and emerging market economies (see Figure 11). TA reports on expenditure rationalization were essentially mini versions of World Bank Public Expenditure Reviews. They typically provided a brief background, outlined the main issues, and discussed some reform options for improving expenditure efficiency, including in social protection categories such as pension/social security and social assistance. Equity implications of reform options were usually considered as well, but not with a full poverty and social impact analysis (PSIA) of each option.
► Energy subsidy reform was the main avenue through which TA addressed social protection policies in LICs, in both a program and non-program context. These TA reports usually (but not always) included an analysis of the distributional impact of existing energy subsidies and suggested some possible measures for mitigating the impact of subsidy removal on vulnerable groups. This evaluation found good use of PSIA by FAD TA missions on energy subsidy reform in several countries (e.g., Bangladesh, Bolivia, Burkina Faso, Ghana, Honduras, Mozambique, and Senegal).
► Pension reform was the main topic in over one-fifth of social-protection-related TA in EMEs and LICs. Most of these TA reports included pension outlay projections. About half of them addressed distributional and allocation issues. Similarly, about half of them discussed alternative pension reform options. TA reports on pension-related issues further from the social protection realm, such as pension administration, were not included in this count.73
► Social safety net reform was often addressed (though not in detail) by TA missions on expenditure consolidation or energy subsidy reform. Dedicated TA missions on social safety net reform were relatively unusual.74
► In a small number of cases, at the request of the authorities, TA missions on tax reform explicitly addressed social protection concerns, e.g., introducing a negative income tax (Korea, Tailand) or reforming personal income tax credits to assist needy households (Iceland).
59. IMF TA was always appreciated by the authorities that requested the assistance.75 Interviews with country officials and FAD staff indicated that the authorities were appreciative of the TA missions’ efforts to bring social protection issues into the policy debate and/or contribute to capacity building, even if the mission’s policy recommendations were not implemented. As noted in IEO (2014b) and according to FAD staff interviewed for this evaluation, “there is always excess demand” for IMF TA—the IMF does not have the resources of the World Bank for PSIA and energy subsidy reform. A former Greek Finance Minister reflected that notwithstanding “the very good technical assistance from the IMF,” the Fund lacked the capacity of the World Bank to provide the needed “serious support for institutional reform” as part of the 2010 SBA and 2012 EFF-supported program in Greece.76
60. FAD has found ways to address the “excess demand” for TA within its existing resource envelope. TA support was effectively provided through participation by FAD economists/experts in area department missions;77 through mobility by FAD economists to area departments; by using World Bank and outside experts—notably, in Cyprus and Portugal, where the World Bank did not take the lead on social protection TA; and through “knowledge expansion,” i.e., internal and external training (including a massive open online course on energy subsidy reform), and tools, materials, and other resources available on the IMF website.