INTERNATIONALBANK FOR WORLD BANK R E T C N O E N STRUCTION PM AND DEVELO February 2003 No.18 A regular series of notes highlighting recent lessons emerging from the operational and analytical program of the World Bank`s Latin America and Caribbean Region COLOMBIA: REFORMING THE SOCIAL SAFETY NET Laura B. Rawlings The Economic Crisis and finance a response to the crisis. As part of this effort, In 1999, Colombia experienced its most severe recession in Colombia's social safety net was reviewed and a number of 70 years, fueled by unsustainable growth in government steps taken, including: (i) rapid crafting of a short-term spending beginning in the early 1990s and external macro- emergency safety net investment program called the Social economic shocks in the late 1990s. Compounding the eco- Support Network (Red de Apoyo Social, RAS); (ii) a nomic downturn, Colombia's internal conflict intensified, poverty assessment and a social safety net assessment; and hundreds of thousands of people were displaced, and vio- (iii) implementation of two social sector adjustment lence and insecurity increased. The social consequences in- operations focused on medium term reforms of the social cluded a doubling of the historical rate of unemployment in safety net, health and education systems. the late 1990s, a continued rise in inequality , and a dra- matic increase in poverty that reversed a decade of progress The Colombia Social Safety Net Assessment included: (i) a in poverty reduction. risk and vulnerability assessment based on the analysis of new and existing household survey data as well as a rapid Historically, Colombia's "safety nets" were economic qualitative study and (ii) an institutional analysis of growth and a now unsustainable expansion of social ser- Colombia's new and existing federal social assistance vices, notably in health, education, and pensions. Social as- programs. With the Colombia Poverty Assessment, this sistance (SA) was not included in the dramatic social sector analytical work provided an empirical basis and a reforms of the 1990s, implying that the social safety net did participatory process for outlining priorities for reforming not benefit from increased spending, decentralization and the social protection system. the strategic prioritization afforded to other social sector ar- eas. SA remains under-financed compared to national de- Vulnerable Groups and Risk Management mands and international norms (Box 1). Furthermore, the SA programs that could have provided a safety net during Despite a modest economic recovery, poverty, inequality the crisis were hampered by structural constraints, including and key social indicators continue to deteriorate, suggesting poor poverty targeting in certain programs, institutional in- that the shock to income and human capital is longer-term flexibility, and unfocused mandates. than the macroeconomic effects of the recession. Quantitative and qualitative analysis confirm that certain The Government of Colombia asked the World Bank and groups remained highly vulnerable and in need of attention Inter-American Development Bank (IDB) to help craft (see Table A, back page): Box 1 - International Comparisons Colombia's level of social assistance spending at 0.6 percent of GDP is very low compared to countries at a similar stage of development. Each of nine Latin America and Caribbean region countries included in a recent social protection expenditure review devotes a higher share of GDP to social assistance, including Argentina (0.9 percent), Mexico (1.1 percent), Peru (1.4 percent), Uruguay (3.4 percent) and Venezuela (1 percent). Source: Dulitsky, Gragnaloti, and Lindert 2001 1 · Children and adolescents are the main age-specific vul- Empleo en Acción community works program provides tem- nerable groups who should benefit from a reformed porary employment to poor, unemployed, low-skilled work- safety net. Preschool children are the age group most ers in labor-intensive social and economic investment prone to poverty: 70 and 90 percent in urban and rural projects such as school expansion and road repair. A condi- areas, respectively, are under the poverty line (Colombia tional cash subsidy program for poor families, Familias en Poverty Report, 2001). Roughly half of all Colombian Acción, aims to protect investments in children's health, children have not been vaccinated against basic commu- education, and nutrition by providing cash to families con- nicable diseases, a decrease in coverage of 10 percent in ditional on keeping children in school and providing them 5 years (Profamilia 1995, 2000). Adolescents are par- with basic preventive health care. Jovenes en Acción is a ticularly vulnerable to unemployment. Male adolescents training­apprenticeship program for young adults that uses represent a disproportionate share of perpetrators and competitive bidding among private firms and public entities. victims of crime and violence. Teen pregnancy has in- Colombia is committed to administering these programs for creased. four years (2001­04) and reviewing their results to assess whether they should continue as part of a broader safety net · People displaced by Colombia's internal conflict are an- strategy. other key vulnerable group, lacking access to assets and social services, notably education and housing. People in These programs are a step toward establishing a conflict areas also need attention. countercyclical strategy that is targeted to vulnerable groups affected by the crisis--notably the unemployed and · As a cross cutting challenge, the labor market plays a the young--with income support and cash transfers. The critical role in determining vulnerability. Facilitating ac- next steps are to evaluate the new RAS programs and cess to employment ­ particularly formal sector employ- integrate them into a coherent, reformed social safety net ment ­ is an effective, but elusive, poverty-reduction that will include a countercyclical component to address strategy. future crises with greater flexibility and effectiveness. If the RAS programs are not assessed as part of a broader The qualitative study of risk and vulnerability revealed a safety net reform, they risk becoming another institutional sophisticated knowledge of the complex challenges facing layer within the diffuse collection of social assistance Colombian society by focus group participants. Participants programs. uniformly assessed the crisis as multidimensional, involving political, economic, and social factors. They attributed their Institutional Analysis of Existing Programs rising vulnerability to the economic recession and to larger structural problems, such as the drop in international prices for A key challenge facing the new administration is to make coffee and other commodities, and the way the break-up of the social safety net an adaptable agent of social risk man- local drug-cartels and corresponding money laundering agement in crisis and non-crisis periods, drawing on social activities (mainly in construction) had contributed to rising assistance and social insurance programs. Although the unemployment. new RAS safety net programs have provided social assis- tance for the main types of risks faced by vulnerable Focus groups highlighted the lack of social safety nets and groups, the overall system is in need of reform. exclusion from benefits which especially affects children and adolescents. Migration, typically from rural to urban areas, Whereas funding for health and education rose from ap- was also a key issue, as it increases competition for scarce jobs proximately 4 percent of GDP in the early 1990s to over 8 and overburdens social services in receptor municipalities. percent by 1996, central government budgeted expenditures Economic insecurity--which emerged as paramount for all on social assistance fluctuated around 1 percent of GDP groups--was perceived as long-lasting by many groups. during the first part of the 1990s, and fell to close to 0.6 percent by 2000. Compounding this problem, social assis- Both the quantitative and qualitative work reveal that house- tance is the most procyclical component of social sector holds use three strategies to manage income risks presented by spending: a retrospective public social expenditure review the recent economic crisis: (i) mobilizing available labor; (ii) reveals that for each peso fall in GDP, social assistance reducing and diversifying consumption; and (iii) using avail- spending fell by nine pesos (CRECE 2001). The new RAS able physical assets, notably housing. These strategies can programs financed by international borrowing equivalent to have negative consequences, such as reduced consumption, 0.3 percent of GDP through 2004 was an important first child labor, and illegal activities including prostitution, gang step in establishing a counter-cyclical safety net in Colom- membership, drug trafficking, and extortion. bia and raising social assistance spending. However, these programs took nearly two years to be designed and launched, remain temporary in nature, and would require Colombia's New Social Assistance Strategy sustainable financing to be maintained past 2004. Colombia's new Social Support Network (Red de Apoyo So- cial, RAS) comprises three principal programs promoting Colombia's most important social assistance agencies, The human development, employment, and job training. The Colombian Family Welfare Institute (Instituto Colombiano 2 de Biensestar Familiar, ICBF) and the Social SolidarityNet- reason to create new social assistance programs. Rather, the work (Red de Solidaridad Social, RSS) lack a strategic assessment presents options for better structuring the system of safety net focus. Programs are fragmented and there is du- existing programs, and for addressing the critical relationship plication. Too few resources are available to fill gaps in the with social insurance institutions. safety net for key vulnerable groups (e.g., pre-school chil- dren, the elderly poor and the internally displaced). The Increased central allocations for social protection to 1 percent ICBF and RSS management of social public expenditures is of Gross Domestic Product (GDP) could be accomplished by complicated by partial and inconsistent social sector budget- reallocating resources from less effective and unfocused pro- ary data. There is no accepted definition for social spending grams for the unemployed in the formal sector to programs or social assistance in the national income accounts or in the with effective targeting of vulnerable groups. Also, a budget. Budgetary data are partial, and different sources re- countercyclical strategy that accumulates funds during periods port markedly different levels of spending on specific pro- of growth and draws them down during times of recession grams; most data are available only on an institutional basis would improve the social safety net's effectiveness in meeting and not by program (Alesina 2000). This lack of transpar- the needs of the transient vulnerable during crisis periods. ency has complicated effective management of social sector spending, particularly for SA programs. Impact of the Assessment Coverage and targeting are also a problem. ICBF estimates that only 13 percent of children in the poorest two deciles The main impact of the Colombia Social Safety Net Assess- are reached by their two main early childhood development ment was substantial reform of Colombia's social safety net. programs; coverage estimates derived from the 1997 Combined with the Bank and IDB lending operations and Encuesta de Calidad de Vida national household survey are the Poverty Assessment, it helped establish the three RAS even lower. Poor poverty targeting has led to a safety net emergency safety net programs for vulnerable people hit by "for the poor, but not the poorest." Existing social programs the recent recession (a conditional cash transfer program, a are primarily designed for formal sector workers and not for workfare program, a youth training program). The analysis the poorest who lack formal affiliation in the labor market. supported the Government's developing legislation for a Redistributive social assistance programs are limited. counter-cyclical fund to finance future safety net operations. Reforms in the proxy-means testing system were promoted, Finally, basic monitoring data on the numbers of program and successfully used to target the poor for selected social beneficiaries and budgets are often either lacking or unreli- assistance and insurance programs. Finally, regular monitor- able, complicating simple estimates of unit costs and re- ing and evaluation of social programs is now part of a re- views of program efficiency. Impact evaluation results are newed focus on results-based management of the system. almost nonexistent. About the Author Policy Recommendations Laura Rawlings is a Senior Monitoring & Evaluation Special- ist with the Social Protection Unit of the Human Development The Colombia Social Safety Net Assessment put forward a Department in the Bank's Latin America and the Caribbean series of recommendations, developed through extensive Region. analysis and consultation with government and civil society. A reformed, better funded social risk management system References would improve Colombia's ability to address present vul- nerabilities and the negative consequences of future shocks. Alesina, Alberto. 2000. "Institutional Reforms in Colombia." The existing system needs technical reforms, including up- Harvard University. dating the SISBEN proxy-means testing system, evaluating CRECE (Centro de Estudios Regionales Cafeteros y Empresariales). the impact of the main social protection programs, and more 2001. "Diseño del Marco Normativo e Institucional y del Esquema transparent national accounting. Also, the quality and cov- de Toma de Decisiones de la Red de Protección Social." Manizales, erage of key programs should be improved for better early Colombia. childhood development, support for the elderly poor, and as- sistance for displaced persons. Additional analysis of social Dulitzky, Daniel, Michele Gragnaloti, and Kathy Lindert. 2001. insurance programs (pensions, health and unemployment) "Social Protection Expenditure Review." August 2001 draft, World was also proposed. Bank. Because the current fiscal constraints make it unrealistic to Fedesarrollo, Coyuntura Social, #21. November 1999, p. 85. think of fresh, sustainable sources of funding for social as- Profamilia. 1995. "Encuesta Nacional de Demografía y Salud sistance, the proposed reforms are focused on reallocation 1995." Bogotá, Colombia. and cost recovery. If the RAS programs are included, Colom- bia now has programs for addressing the principal risks faced Vélez, Carlos Eduardo, Vivian Foster, Mauricio Santamaría, Natalia by each of the main age groups, and high risk groups in spe- Millán, and Bénédicte de la Brière. 2001. "Colombia Poverty Re- cial circumstances, so the report found no compelling port." Washington, D.C.: World Bank 3 Table A. Social Risks and Social Protection Programs in Colombia, 2000 Group/Risk Indicators No. Affected Percentage Social Protection Programs Age 0-6 Population = 4,873,000 Stunted development (E)ICBF school feeding; formal day-care 657,000 13.5% ->Severe chronic malnutrition (age 0-5) informal day-care (N)RAS Familias en Acción conditional cash transfer (health and nutrition grant) Limited early childhood development - Not attending: 3,562,000 Not attending :54% (SI) ISS health insurance (contributory and >Preschool or daycare attendance (age 0-7) subsidized regimes) Age 7-11 Population = 5,171,000 Low human capital development Not enrolled: 707,000 Enrolled: 83.6% (N) RAS Familias en Acción conditional cash ->Net primary school enrollment transfer (education grant) (SI) ISS health insurance (contributory and ->Overage (average years behind) (ages 7-11) 0.20 years subsidized regimes) Age 12-18 Population = 6,059,000 Low human capital development (E) PACES voucher program, but being phased Not enrolled: 1,901,000 Enrolled: 62.8% ->Secondary enrollment (ages 12-17) out (N) RAS Familias en Acción conditional cash ->Overage (average years behind ages 12-17) 0.82 years transfer (education grant) (SI) ISS health insurance (contributory and Poor health subsidized regimes) 420,000 19% ->Female teenage pregnancy (ages 15-19) Low income ->Inactivity (ages 12-17) 680,000 13.3% ->Unemployment (ages 12-18) 497,000 30.5% ->Informal sector (7 major cities, ages 15Ñ19) 244,000 79.0% Age 19-64 Population = 22,261,000 (N) RAS Jovenes en Acción youth training (N) Low income RAS Empleo en Acción workfare ->Unemployment 2,705,000 15.8% (SI) ISS health insurance (contributory and ->Informal sector (7 major cities, ages 20-29) 2,999,000 57.9% subsidized regimes) Age 65+ Population = 2,452,000 Low income (E) RSS Revivir ->No pensions 2,111,000 81.8% (SI) Pension system for formal sector employees ->Informal sector (7 major cities, age 60+) 234,000 82.7% General Population Population = 42.3 million Poor health -> Access to health insurance 19,000,000 un-insured 47% uninsured (E) Cajas de Compensación development funds (E) Utilities subsidies Low-quality housing (E) INURBE/rural housing programs ->No piped water 5,836,000 14.3% (E) RSS program for displaced ->No electricity 1,975,000 4.84% (SI) ISS healthInsurance (contributory and Violence and displacement subsidized regimes) ->Displacement 400,000 -1.9 million 1-5% Note: This Table covers the principal federal social protection programs. Unless otherwise noted, estimates were carried out by Departamento Nacional de Planeación's DDS and SES divisions using Encuesta Nacional de Hogares data from September 2000. **(E): Existing Social Assistance; (N): New Social Assistance; (SI): Social Insurance 4