III Social Safety Nets

Ritha Khemani, Sanjeev Gupta, Calvin McDonald, Louis Dicks-Mireaux, and Marijn Verhoeven
Published Date:
January 2000
  • ShareShare
Show Summary Details

In countries where the authorities could foresee that reform measures would have a sizable adverse social impact, the policy mix and sequencing have aimed to take this impact into account within a sustainable macroeconomic framework. For instance, IMF-supported programs have aimed to phase out subsidies for food and other items gradually, rather than at once (e.g., Indonesia, 1998; and Senegal, 1994-95). The adverse impact, however, cannot be totally eliminated even with an appropriate policy mix and sequencing. For instance, a change in relative prices that hurts the poor–such as a devaluation that could adversely effect the urban poor through increasing prices of imported products–may be at the heart of a reform program. A tension may emerge, therefore, between stabilization and social protection objectives.

Social safety nets are a means of easing this tension. IMF staff–drawing on the work of other institutions–has increasingly sought to incorporate social safely nets into adjustment programs. The IMF’s Executive Board, following its discussion of social safety nets in 1993, endorsed this approach.12 More recently, ESAF-supported programs have sought larger budget allocations for social safety nets. There has been an increase in the use of structural benchmarks and performance criteria aimed at securing social protection objectives (Box 3.1).

This review indicates that most IMF-supported programs have incorporated social safety nets, although there is scope for further improvement in their quality and implementation.

  • The social safety nets in IMF-supported programs have included new temporary arrangements–such as temporary subsidies and public works programs–as well as existing social protection instruments adapted to the needs of target groups–such as pensions and other permanent social security programs (see Chu and Gupta, 1998). Measures designed to foster financial stability, such as the adoption of deposit guarantees (limited or general) and other financial restructuring measures to maximize asset recoveries, to redistribute losses, and to sustain credit to the small and medium-sized industry segments, also have protected small depositors and vulnerable groups.
  • Over time, permanent social protection arrangements (e.g., pensions, unemployment insurance) also have been established in the context of reform programs.
  • Family-based safety nets have cushioned income losses during adjustment periods in many countries. For example, in Indonesia, in the wake of the recent crisis, about a fourth of families received informal transfers (Frankenberg, Thomas, and Beegle, 1999). Such informal arrangements have been generally well targeted.13 Thus the design of public social safety nets has sought to avoid duplicating the system of voluntary, private transfers.

The timely implementation of social safety nets has been hampered–all too frequently–by a lack of existing social policy instruments. Often these instruments can be speedily adapted to the needs of the new target groups. But sometimes the adaptation is difficult. Countries often have not had the will to reform costly existing social protection mechanisms or to shift social protection priorities. Those segments of the population suffering from the adverse effects of reforms may prove to be different from those protected by the permanent arrangements. In addition, the lack of data and administrative and financial constraints have hampered implementation and monitoring.

In most cases, IMF staff has relied largely on the World Bank, and regional development banks to some extent, to take the lead in the design of social safety nets for IMF-supported programs.

  • For example, in Indonesia, a targeted rice subsidy and community-based public works programs, designed by the World Bank, were incorporated in the 1998 IMF-supported adjustment program. This was also the case for the public works programs financed by the Asian Development Bank and the World Bank in Thailand in 1998. In Brazil in 1999, the Inter-American Development Bank and the World Bank cofinanced a special adjustment loan for a social protection project that was integrated into the IMF-supported program.
  • In some cases, the IMF Fiscal Affairs Department has provided limited technical assistance on social safety nets (e.g., Ecuador and Belarus in 1999). And in 1999, the African Department recruited two social policy specialists with backgrounds in sociology to assist the department in incorporating appropriate social safety nets in countries’ adjustment programs. To the extent possible, IMF missions have built on the work of the World Bank, regional development banks, other UN agencies, and NGOs.14 In many cases, program-related missions have also focused on social issues.

Box 3.1.Strengthening Social Safety Nets in ESAF-Supported Programs, 1994-98

Policy Framework Papers (PFPs) and Memoranda of Economic Policies (MEPs) for 44 countries that had ESAF-supported programs during 1994–98 were reviewed to ascertain what kind of measures had been incorporated in programs to mitigate the adverse social effects of structural adjustment policies. These measures were classified into two groups:

  • Budgetary allocations (either unspecified general allocations or quantitative commitments); and
  • Measures to strengthen social protection through more specific targeting, better monitoring of the affected population groups, widened coverage of safety net measures, or related reforms.

A significant group of countries have incorporated allocations for social safety nets in their ESAF-supported programs, and the use of structural benchmarks and performance criteria for achieving social protection goals has grown.

  • Half the countries had commitments in their PFPs for allocations for financing social safely nets during the program period (on average 3–3 ½ years), with a third of the 44 countries setting these targets in quantitative terms. During the review period, countries also included measures for strengthening the design and coverage of social safety nets; some established targets for the number of vulnerable people to be shielded by social safety nets (Armenia, Guyana, the Kyrgyz Republic, Mozambique, and Nicaragua).
  • About three-fourths of countries announced spending on social safety nets in their MEPs during the program period, with 60 percent of the 44 countries specifying quantitative allocations. In recent years, however, the use of quantitative targets with respect to social safety nets has declined. A similar trend is noticeable in the listing of structural measures for strengthening social safety nets. Although it is difficult to pinpoint the reasons for this trend, there are several possibilities. Priorities in programs could have shifted; the monitoring of social safety nets could have become more difficult because of data and other constraints; and in a number of countries, structural reforms with the greatest adverse social impact could have been felt earlier in the program periods. On a country-by-country basis, significantly fewer countries (about one-half) that sought allocations of expenditures on social safety nets in PFPs made the same commitments in their MEPs. Nine countries specified targets for the number of vulnerable to be shielded by social safety nets in MEPs (Albania, Armenia, the Republic of Congo, Haiti, the Kyrgyz. Republic, the former Yugoslav Republic of Macedonia, Mozambique, Nicaragua, and Yemen).

Greater use of structural benchmarks and performance criteria for social safety nets in ESAF countries is a recent phenomenon. Structural benchmarks have included, for example, improving the transparency of energy subsidies, replacing generalized subsidies with targeted subsidies, and establishing labor retrenchment funds. Programs for six countries incorporated benchmarks (Azerbaijan in 1996, Cameroon in 1998, the Kyrgyz Republic in 1994, 1995, and 1998, the former Yugoslav Republic of Macedonia in 1997 and 1998, Pakistan in 1995-98, and Yemen in 1997-98). Of these, two countries have included performance criteria for achieving social protection objectives, for example, by strengthening the revenue position and by reforming the benefit structure of the employment and the pension funds (the Kyrgyz Republic in 1994. 1997, and 1998. and the former Yugoslav Republic of Macedonia in 1998).

Design Issues

The design of social safety nets has been influenced, among other things, by the availability of existing social policy instruments. For example, transition economies had a broad range of social instruments that were poorly targeted (e.g., Moldova and Ukraine). A wide range of benefits covered the bulk of the population, including the nonpoor. Thus the principal aim of IMF-supported programs in these countries has been to make spending–for instance, food subsidies–better targeted, rather than to create new instruments. In contrast, low-income developing countries had few and limited social policy instruments; the effort there has been to create arrangements that could reach affected groups, such as transitory subsidies for the urban poor in the CFA franc zone countries in the aftermath of the 1994 devaluation. In general, establishing cost-effective social safety nets would have been easier had well-targeted social policy instruments been already in place before the onset of crises and economic reforms.15

The specific adverse effects on and characteristics of target groups have determined the types of social safety net instruments.

  • A sharp fall in output, reinforced by a large increase in prices of important staples, can result in a significant real income loss for those poor households who are net consumers of food. This occurred in Indonesia in 1998 and in transition economies in the initial stages of transformation. In these circumstances, income transfers or targeted food subsidies became critical.
  • When the prices of essential goods rose in countries where the elderly constituted a high proportion of the population, low-income pensioners needed to be helped through an adequate minimum allowance (e.g., the Russian Federation and Ukraine).
  • When there were regional pockets of unemployment, special programs to supplement incomes have been implemented (e.g., community-based public works programs in Indonesia and Senegal).

Nevertheless, the selection of target groups has raised fairness issues. To ensure political support for reform, social safety nets have been extended to politically vocal middle-income groups: for example, a subsidy for premium gasoline (in Indonesia, initially in 1998), severance payments for departing civil servants (Ghana and Lao People’s Democratic Republic) and for public enterprise employees (Argentina and Bolivia), and food subsidies (Jordan). Unemployment benefits–largely for formal sector beneficiaries—were strengthened in the presence of a large informal sector, where the majority of the poor may be residing (Brazil in 1998).

To identify target groups, IMF staff has generally relied on the national authorities and the World Bank, which have provided a measure of the poverty line and household expenditure survey data on household characteristics.16 The latter, however, are often conducted infrequently, with results that are not always timely or comparable across time.17 Because of these constraints, poverty lines and poverty profiles have typically been unavailable for the year when reform programs were put in place (e.g., Brazil and Thailand).

The weak administrative capacity in many countries has hampered the targeting of benefits, particularly on the basis of incomes. This has meant a greater reliance on programs that have self-targeting features, such as

  • public works with below-market wages (e.g., Indonesia, Malawi, Thailand, and Senegal);
  • subsidies on commodities consumed by the poor (e.g., inferior rice in Indonesia); and
  • shielding of groups that are easily identified as poor (pensioners, the unemployed, single mothers, and children).

Means testing based on wage income has been used in some countries–for the housing subsidy program in Ukraine, for example–but this carries the risk of mistargeting, especially in countries with a large informal sector, and could create disincentives for the supply of labor.

Financial constraints have limited the scope of social safety nets. The need to redress macroeconomic imbalances has typically precluded increasing total public spending; thus, a reallocation within the existing budgetary envelope to be tier-targeted programs has been necessary (e.g., Brazil).

  • In some cases, such as Venezuela, eliminating subsidies has freed resources for more targeted social protection programs.
  • Significant budgetary savings have often been achieved by reforming social safety nets, such as by replacing generalized subsidies with targeted ones (e.g., rice subsidies in Indonesia).
  • External donors, including the World Bank, have played a role in some cases, in particular by funding severance payments for departing civil servants and by providing food aid (e.g., Senegal).
  • In transition countries, the large decline in output has increased the demand for social benefits while reducing the availability of financing (Box 3.2).

Box 3.2.Social Safety Nets: Issues in Transition Economies

Key Issues

Transition economies have been trying to reform expansive, but increasingly ineffective, social protection arrangements, including subsidies, pensions, unemployment benefits, and family allowances. Declining taxes and social contributions have severely weakened the ability of many transition economies to provide the needed benefits.

  • The declining social contributions have transformed the earnings-related pensions and other social benefits into virtually flat minimum benefits.
  • Offsetting tax obligations of enterprises against obligations of the government has limited the ability of many governments to pay cash benefits. Moreover, these obligations of the government do not necessarily represent spending of high social value.

In many transition economies, the benefits have yet to be fully reformed.

  • In Moldova, the Russian Federation, and Ukraine, pension systems have allowed workers to collect benefits at a relatively young age, and workers in certain occupations have been eligible for pensions even earlier. Pensions have been based not only on the number of years of contributions, but also on years spent studying or taking care of a young child.
  • A large part of social contributions collected for assisting the unemployed continue to be wasted on low-priority programs and benefits administration; only a small share of the unemployed actually receive assistance in any form.
  • Traditional extensive and generous privileges for politically influential groups (judges, parliamentarians, internal security personnel) prevent the targeting of limited resources to the genuinely needy.

IMF Advice

Thus, the emphasis of IMF staff advice has been on improving compliance with tax laws, simplifying the rate structure, and stopping the collection of taxes and social contributions in kind (e.g., Azerbaijan, Moldova, and Ukraine). Some progress in simplifying the social contribution rate structure has been made, but significant change in other areas has been elusive.

In these and other similar cases, IMF staff have called for raising the pension age, eliminating privileged and early pensions and untargeted benefits, targeting social assistance and subsidies, increasing the size and coverage of unemployment benefits, and making social benefits more transparent. Pension reforms are under consideration in many transition countries, but the progress in implementing far-reaching reforms remains slow. For example, despite the emphasis given to pension reform, including the use of performance criteria, in IMF-supported programs in the Kyrgyz Republic, relatively little progress was achieved; a concerted and more comprehensive and resource-intensive World Bank adjustment operation appears to have had more success in pension reform since 1998.

When the adverse impact exceeded expectations, program targets have been changed to accommodate larger budgetary outlays for social safety nets. Indonesia, Korea, and Thailand raised spending on social protection programs to 5.2 percent of GDP in 1998/99, 2 percent of GDP in 1999, and 2 percent of GDP in 1998/99, respectively, from between ½ percent and 1 percent of GDP in each country before the crisis.

In general, and beyond the context of immediate program requirements, it would be desirable to identify the need for social policy instruments and advise government authorities to seek necessary assistance from the World Bank and others, in the course of IMF surveillance. Such efforts could, among other things, speed the establishment of cost-effective safety nets if difficulties arise and reform measures need to be undertaken.

Labor Market Implications

Labor market incentives have been a key concern in the design of unemployment benefits. The challenge has been to strike an appropriate balance between social protection and disincentive effects, although, certainly, this balance may differ among countries, depending on social preferences, norms, and other factors. In 1998, because of their concerns about possible disincentive effects, the Korean authorities initially hesitated to broaden the coverage of unemployment benefits. A relatively high wage for participants of public works programs can undermine their effectiveness as a safety net by attracting already employed workers.18

Establishing social safety nets, however, can help promote a fundamental labor market reform. Before the financial crisis in 1997, lifelong employment in large enterprises had been an important aspect of social protection in Korea. But it constrained the ability of enterprises to restructure in the face of changing economic conditions. Broadening the coverage of unemployment benefits to 70 percent of the labor force in early 1999 from around 30 percent supported labor market reform aimed at promoting labor market flexibility by providing income transfers to those switching jobs.


The staff monitoring of social safety nets has been infrequent. A review of 12 countries19 with ESAF arrangements during 1994-98 indicates that, although over three-fourths of Policy Framework Papers (PFPs) and Memoranda of Economic Policies (MEPs) reported on the performance of social safety nets under IMF-supported programs, such monitoring typically occurred only once or twice. The infrequent monitoring of social safely nets may have reflected weak national capacity to monitor the implementation of social policies and their impact on poverty.

The staff’s assessment and reporting of the effectiveness of social safety nets have been uneven. In only a third of the 12 countries reviewed did staff papers assess the coverage and incidence of social safety nets during the five-year period (e.g., with respect to temporary food subsidies and to fertilizer provision to smallholders in Malawi). For a larger number of countries (more than three-fourths), staff papers reported on the improvements in the benefit structure and financing of social protection mechanisms. In contrast, staff papers on IMF-supported programs for the countries affected by the recent crisis in Asia (Indonesia, Korea, and Thailand) reported extensively on social safety net developments.20

    Other Resources Citing This Publication