ࡱ > !a bjbj / ( ] ? Z Z + p
, P y2 y2 y2 2 2 2 3 6 2 z : ?? ^ ? 1@ 1@ E =I IJ ) ˹ y2 z E @ [E @ z z ], ], 1@ 1@ R Ó Ó Ó z : ], 1@ 2 R 1@ Ó z Ó Ó l | q0 W2 " 1@ Pvt 2 g H [ [ $ ^ [ y2 J [ 4 Ó h q J J J J J J z z z z [ J J J J J J J J J Z c* :
Conditional Cash Transfers in Latin America:
Problems and Opportunities
Julia Johannsen ( HYPERLINK "mailto:jjohannsen@iadb.org" jjohannsen@iadb.org)*
Luis Tejerina ( HYPERLINK "mailto:luist@iadb.org" luist@iadb.org)*
Amanda Glassman ( HYPERLINK "mailto:amandag@iadb.org" amandag@iadb.org)*
*Inter-American Development Bank (Social Protection and Health Division SCL/SPH)
Contents
TOC \o "1-3" \h \z \u HYPERLINK \l "_Toc245867603" 1. Introduction PAGEREF _Toc245867603 \h 2
HYPERLINK \l "_Toc245867604" 2. Basic Principals and Achievements of CCT Programs PAGEREF _Toc245867604 \h 3
HYPERLINK \l "_Toc245867605" 2.1. The Dual Policy Strategy of Consumption Smoothing and Long-Term Human Capital Investment PAGEREF _Toc245867605 \h 3
HYPERLINK \l "_Toc245867606" 2.2. Target Population PAGEREF _Toc245867606 \h 5
HYPERLINK \l "_Toc245867607" 2.3. Relevance and Impacts in Latin American Countries PAGEREF _Toc245867607 \h 6
HYPERLINK \l "_Toc245867608" 2.4. Positive Externalities of CCT Programs in Public Policy Administration PAGEREF _Toc245867608 \h 10
HYPERLINK \l "_Toc245867610" 3. Persisting challenges PAGEREF _Toc245867610 \h 13
HYPERLINK \l "_Toc245867611" 3.1. CCT Programs in Urban Areas PAGEREF _Toc245867611 \h 13
HYPERLINK \l "_Toc245867612" 3.2. CCT Programs Capacity to Respond in Crisis Situations PAGEREF _Toc245867612 \h 17
HYPERLINK \l "_Toc245867613" 3.3. Design of Targeting Instruments PAGEREF _Toc245867613 \h 21
HYPERLINK \l "_Toc245867614" 3.4. The Missing and Mistaken Exit Strategy PAGEREF _Toc245867614 \h 23
HYPERLINK \l "_Toc245867615" 3.5. Beyond the First CCT Wave: Developments of the Next CCT Generation? PAGEREF _Toc245867615 \h 26
HYPERLINK \l "_Toc245867616" 4. Conclusions PAGEREF _Toc245867616 \h 29
HYPERLINK \l "_Toc245867617" 5. References PAGEREF _Toc245867617 \h 31
HYPERLINK \l "_Toc245867618" 6. Annex PAGEREF _Toc245867618 \h Error! Bookmark not defined.
Abbreviations and Acronyms
CCT conditional cash transfers
PMT proxy means test
Introduction
Conditional cash transfer (CCT) programs began in the mid 1990s in Latin America, with Bolsa Escola in 1995 in Brazil, Chile Solidario in 1996, and Progresa in 1997 in Mexico. Today, CCT programs exist in most countries throughout the region. Exceptions include Nicaragua, Venezuela, and some Caribbean countries. REF _Ref234596832 \* MERGEFORMAT Appendix 1 lists the programs and their implementation years. The largest programs in terms of beneficiary numbers are those in Argentina, Brazil, and Mexico, which together benefit more than 16 million poor families (Bouillon and Tejerina 2007).
Based on the experience gained by the Inter-American Development Bank with CCT programs over the last decade in Latin America, this paper reviews some challenges and achievements of CCT programs in the region from an operational point of view. In addition, the paper summarizes recent areas of attention in CCT design and implementation, including urban expansions and response to crises, as well as the possible future development of CCT programs.
The second chapter explores the conceptual principals and some main achievements of CCT programs. It discusses the dual policy strategy pursued with conditional transfers, the intended target population, and the achieved impacts and relevance of the programs in terms of coverage and size. The chapter ends with the positive externalities of CCT programs in public policy administration beyond their immediate objectives.
The third chapter is dedicated to diverse persisting challenges in CCT program design and operation. These include the design adaptations to urban areas and a conceptual discussion about the capacity to respond to crisis situations such as the food price increases in 2008 and the current global economic crisis. The chapter then discusses challenges for designing targeting, exit strategies, and an outlook on potential future developments of CCT programs. The forth chapter briefly concludes.
Basic Principals and Achievements of CCT Programs
The Dual Policy Strategy of Consumption Smoothing and Long-Term Human Capital Investment
Many targeted antipoverty and emergency relief programs aim at both an immediate consumption effect and an investment effect related to the creation or conservation of human capital (or physical infrastructure). They achieve a varying balance between these two objectives. If the objective is to break the intergenerational poverty cycle through investment in the next generations health and education, this dual policy strategy (whether achieved by a single or several closely coordinated programs) is considered a key factor for success (Levy 2006). Indeed, the poor value transfers that are postponed into the future in the form of investment components considerably less than direct and immediate consumption smoothing, if they are asked to make a choice (Maloney 2001). In other words, on theoretical grounds, there is a critical level of direct transfers for current consumption that a household in extreme poverty needs to satisfy before being able to change its spending and other behaviors related to future social investment in terms of use of preventive healthcare, nutrition, and education. The design of CCT programs is based on this dual policy strategy.
Conceptually, CCT programs address market failures that are considered among the causes of low investment in human capital on behalf of the poor. Transfer schemes designed under this concept directly intervene on the demand side and aim at changing social accountability relationships between beneficiaries, service providers, and governments. By addressing demand-side barriers (such as constraints in the access to credit and insurance as well as high discount rates of investment in human capital), they combine short-term transfers for income support with incentives for long-run investments in human capital by conditioning the payments on changes in behavior that favor more adequate food consumption, higher school attendance, and use of preventive healthcare services. However, while CCT programs do not directly attempt to intervene on the supply side, they indirectly address market failures such as disincentives for coordination and information asymmetries because they not only produce detailed information on gaps in the supply of health and education services but also often create effective political pressure given that their program impacts depend heavily on complementary supply-side reforms. The following graph exemplifies how the potential effect of CCT programs can be conceptualized in the context of assumed market failures on the demand and supply side.
Figure SEQ Graph \* ARABIC 1: Potential Effects of CCT Programs on Assumed Market Failures in Demand and Supply [figure needs editing]
CCT = conditional cast transfer.
Source: Authors elaboration (using Rubinstein 2003, Borghans and Golsteyn 2001, and Maloney 2001).
Chronic poverty reduction can only be an indirect and long-term goal of CCT programs. Human capital accumulation is thought to contribute to interrupting the intergenerational transmission of chronic poverty because, as a consequence of program participation, beneficiary children are expected to be healthier and better educated than without the program. As a result, this future generation is intended to be better prepared to access the labor market at higher productivity rates and raise healthier and better educated children themselves. Based on this rationale, CCT programs cannot be expected to reduce chronic poverty immediately. Their direct task can only be to prepare the ground for this development and their ultimate impact strongly depends on how consequently and precisely the program design, identification of the beneficiary population, and alignment with necessary complementary policies respond to these overall objectives in a given country context.
Target Population
Consistent with their objective of strengthening human capital, CCT programs target specific subgroups within populations in structural poverty. Following the conceptual rationale just laid out, CCT programs are key instruments for alleviating extreme and chronic, not transitory poverty. They hold promise for addressing the intergenerational transmission of chronic poverty by explicitly targeting certain population groups. Consequently, in most countries in the region, CCT programs target households in extreme poverty. However, within this target population, project design has led to a subset of extremely poor households to be eligible for the program based on their demographic profile. For example, the CCT programs in Colombia, El Salvador, and Peru select beneficiaries only among household with children below a certain age limit. For other (usually the older) programs, all households are eligible for the benefit.
Another factor that has limited the outreach of programs to the poorest households is that implementation has been limited to the poorest municipalities within a country based on poverty mapping. For this reason, some programs (especially the most recent, as in El Salvador and Paraguay) have excluded households through geographical targeting that may be in extreme poverty but not in the preselected areas.
Relevance and Impacts in Latin American Countries
The potential program impact on proximate outcomes depends on the relevance in terms of coverage among the poor and relative magnitude and structure of the transfer in each country, among other factors including program design and supply-side conditions. REF _Ref234585649 \* MERGEFORMAT Table 1 reports coverage figures for some of the regions programs.
Table SEQ Table \* ARABIC 1: Coverage and Poverty Outreach of Some CCT Programs in Latin America
CountryProgram NameBeneficiaries (households)Beneficiaries (individuals)Poverty (%)Beneficiaries/
Poor (%)El SalvadorComunidades Solidarias (formerly Red Solidaria)80,000380,80047.512ArgentinaPlan Familias454,0002,161,04021.027Costa RicaSuperemonos58,000276,08019.034ChileChile Solidario221,0001,051,96013.747PeruJuntos420,0001,999,20044.517ParaguayTekopora100,000476,00060.513PanamaRed de Oportunidades55,000261,80030.827HondurasPrograma de Asignacin Familiar (PRAF)170,000809,20071.517ColombiaFamilias en Accin1,700,0008,092,00046.639Dominican Rep.Solidaridad400,0001,904,00044.546MexicoOportunidades5,000,00023,800,00031.772BrazilBolsa Familia11,000,00052,360,00033.384EcuadorBono de Desarrollo Humano1,200,0005,712,00043.0101Note: Among the programs not included in the table are: Nicaragua's Red de Proteccin Social (program concluded), Guatemala's Mi Familia Progresa and Jamaica's PATH. Beneficiary numbers only indicate estimations as of end of 2008 (for Paraguay, they refer to the projected target population in 2009); the calculation of individual beneficiary numbers is based on an average household size of 4.76.
For reasons of comparison, poverty figures refer to ECLAC poverty rates from the Social Panorama of Latin America (2008) and do not necessarily correspond to the country's own poverty rates nor the programs' defined target populations.
Source: Authors elaboration based on program documents and poverty rates from ECLAC (2008).
Clearly, programs that cover only a small fraction of a countrys poor population (such as most programs in Central America, see Figure 2) can only have a limited impact versus programs that cover the entire extremely poor population, as in Brazil and Mexico.
Figure SEQ Graph \* ARABIC 2: Estimated CCT Coverage and Poverty Rates in Central America and Jamaica (about 2008)
Note: Poverty levels do not reflect the actual poverty definition the programs used to identify their potential target population but are meant to reflect the relative size of programs according to a comparable poverty measure.
Source: Authors elaboration based on program documentation and ECLAC (2008).
Following the dual policy strategy of an immediate consumption and an investment effect, the actual relevance of the transfer relative to household expenditure or consumption levels varies. Comparisons across countries and years show an immediate consumption effect of the monetary transfer between approximately 10% and 30% in terms of the ratio of the transfer to household expenditures ( REF _Ref234590501 \* MERGEFORMAT Table 2).
Table SEQ Table \* ARABIC 2: Relevance of Some CCT Programs in Latin America in Magnitude of the Transferred Amount
CountryProgramRatio of Transfer to Household ExpendituresSourceBrazilBolsa Familia8%bColombiaFamilias en Accin30%a17% (2002)b13% (2006)bHondurasRed de Proteccin Social10%a11% (2002)bJamaicaPATH20%aMexicoOportunidades25%a20% (1999)bNicaraguaRed de Proteccin Social30% (2002)bSources: a = Handa and Davis (2006), b = Fiszbein et al. (2009).
Evaluation results show that CCTs can be effective tools for reducing poverty through demand-side incentives for an increasing accumulation of human capital. CCT programs are among the most evaluated social programs in the world. Impact evaluations have demonstrated CCTs results in improved child nutrition and household food spending; greater use of (preventive and general) healthcare; reduced child labor; and increased school attendance and enrollment, especially in rural areas ( REF _Ref234676325 \h \* MERGEFORMAT Appendix 2).
However, demand-side versus supply-side effects in the context of CCT implementations are not sufficiently analyzed. Given the limited analytical evidence on the actual causes of inequities in health and education as well as the endogenous relationship between income poverty and health or education outcomes (influencing each other in both directions), there are actually doubts that demand-side barriers are the only or even the most important cause for the observed low rates of social services use by the poor (Handa and Davis 2006). Only a tiny minority of the rich evidence of impact evaluations rigorously distinguishes between demand-side effects through the conditioned transfer payment and the potential effect of complementary supply-side reforms in terms of extended coverage and/or quality that CCT programs indirectly seek to foster and that ideally are implemented in parallel with or prior to the demand-side incentives (section REF _Ref234594200 \r \h \* MERGEFORMAT 2.1).
In cases where supply-side interventions (including those that are usually independent of any CCT program, such as health insurance schemes) took place in the investigated program regions, a contamination of demand-side evaluation results is probable. While a wide range of supply-side interventions have proven to be successful in reducing inequities in health and education outcomes on their own, so have the demand-side incentives of CCT programs. One example of impact evaluations that aim at separating the potential transfer impacts from those of complementary supply-side reforms (by strictly separated treatment groups for supply, demand, and a combination of both) is the attempt of the intermediate evaluation of the Honduran Programa de Asignacion Familia (PRAF) program (IFPRI 2003). Given the nearly nonexistent implementation of the health supply incentives (and only partial implementation in education), however, the evaluation was only able to demonstrate the isolated impacts of the demand-side transfers.
Based on the foregoing considerations, it can be concluded that both demand- and supply side-barriers are important causes of the inequities observed in health and education, and that CCT programs have proven to successfully intervene on the former while potentially having some indirect effects on stimulating complementary interventions on the supply side (section REF _Ref234594200 \r \h \* MERGEFORMAT 2.1). However, it is worrying to observe the lack of careful ex-ante diagnostics to identify the actual importance and nature of the demand- versus supply-side barriers to human capital in a specific country context prior to the design of concrete policy interventions, including CCT programs.
Careful and comparative cost-effectiveness studies of CCT programs are lacking. One relevant criterion for the above-mentioned ex-ante analysis of barriers to human capital and adequate policy responses refers to the cost-effectiveness of interventions. There is, however, only limited evidence on the comparison of the actual cost-effectiveness and cost-benefit ratios of the intended outcome between different types of demand- and supply-side interventions as well as within the CCT family across programs. This fact further limits the possibilities of taking evidence-based decisions between policy alternatives for human capital accumulation in health and education.
Positive Externalities of CCT Programs in Public Policy Administration
CCT programs play an important role in applying new social policy theories and program administration practices. In practice, many processes and instruments associated with CCT programs (including impact evaluations, targeting by proxy means test, and poverty maps) are, in fact, not intrinsic parts of nor initially motivated by these programs. However, the widespread use of these tools and the development of the associated technical capacity was in many cases a consequence of the implementation of a CCT program. CCT programs have the advantage that they represent large-scale and highly visible interventions. In many countries, the CCT is the flagship program of the administration in the social area (Brazil, El Salvador, Guatemala, and Mexico). In addition, by providing benefits in cash to poor familiesa sometimes controversial change from social programs that are predominantly based on in-kind deliverythe emphasis on the need for evaluation and accountability is increased. High visibility of a program creates a demand for transparency in the processes which, in turn, generates the need for objective instruments to improve the effectiveness of the programs.
CCT programs can have positive externalities in terms of administrative reforms within and across sectors. CCT programs often act as the main channel through which strategic reforms as well as transparency and accountability standards are stimulated (Levy 2006). CCT programs require strong leadership and political commitment in order to manage the complex alignment of executing actors across levels of public administration and civil society actors and the necessary coordination with the principal line ministries responsible for the supply of complementary social services. Apart from technical and operational factors, important political aspects determine the long-term success or failure of CCT programs and any other complex antipoverty intervention ( REF _Ref234212126 \h \* MERGEFORMAT Table 3).
Table SEQ Table \* ARABIC 3: Administrative Success Factors of CCT Programs
Source: Levy (2006).
The political success factors experienced by CCT programs in Latin America include credibility, financial transparency, a close relationship with congress, and their not being used for election purposes and campaigns. If the government or certain interest groups (such as service deliverers, intermediaries, or local distributers) use a program for political or economic purposes (e.g., as occurred in the school feeding program Vaso de Leche in Peru) the intended poverty-reduction impact as well as the financial efficiency and the necessary political sustainability of the program over time are likely to erode. Although much can still be improved in this regard, CCT programs represent the best practice among a large variety of development and social programs, and led to stimulus effects for administrative reforms in many countries.
CCT programs catalyze impact evaluations in social policy. While the practice of impact evaluation is not new to the social sectors and examples can be found as far back as the 1970s, the impact evaluation of Oportunidades in Mexico (called PROGRESA at the time) can be considered as the principal detonator of an expanded use of impact evaluations in Latin America. Figure 3 shows a timeline of impact evaluations by the publication date of the evaluation document.
Figure SEQ Graph \* ARABIC 3: Number of Impact Evaluations Published per Year in Latin America and the Caribbean
Source: Authors calculations based on Bouillon and Tejerina (2007).
The figure shows at least a clear relationship (if not causality) between the appearance of the Oportunidades evaluation and the growing number of impact evaluations implemented in social programs in the region. This process has been accompanied by many countries generation of local capacity for impact evaluation, as well as the practice of evaluating social programs outside the CCT realm. In many countries (including Colombia, Ecuador, and Mexico), the groups that were involved in the initial evaluations continue to be involved in the evaluation of social policy in general.
CCT programs have been key catalysts for the use of targeting instruments in social policy. In terms of targeting, CCTs have been instrumental in creating targeting tools that have been used by other programs. While in some countries, including Chile and Colombia, institutionalized targeting systems existed prior to the appearance of a CCT program, in many countries (Ecuador, El Salvador, and Paraguay, for example), the targeting mechanism created mainly for the CCT program is now being used by other programs, as shown in REF _Ref234383930 \h \* MERGEFORMAT Table 4. In other countries (such as the Dominican Republic, Jamaica, and Mexico), the CCT or another social program (such as Perus health program for the poor, Seguro Integral de Salud [SIS]) were the first in the sector to introduce objective targeting techniques such as proxy means testing to select individual households as beneficiaries, although the targeting instrument has not yet been widely adopted by other programs. REF _Ref234383930 \h \* MERGEFORMAT Table 4 summarizes some of the described cases about other social programs use of CCT targeting mechanisms.
Table SEQ Table \* ARABIC 4: Use of Targeting Systems beyond CCT Programs
CountryIndexNational SystemNo. of Programs Using CCT Tool for TargetingColombiaSISBENYes31ChileFicha de Proteccin SocialYes15EcuadorSELBENYes3El SalvadorCCT specificNo3ParaguayCCT specificNo2PeruSISFOHNo (in progress)0MexicoCCT specificNo0CCT = conditional cash transfer; SELBEN = Sistema de Identificacin y Seleccin de Beneficiarios de Programas Sociales; SISBEN = Sistema de Identificacin de Potenciales Beneficiarios de Programas Sociales; SISFOH = Sistema de Focalizacin de Hogares - System for Identifying/ Selecting/ Targeting Beneficiaries of Social Programs.
Source: Authors.CCT programs have helped rationalize fragmented social policy. This has been another way in which CCT programs have provided positive externalities and increased efficiency in Latin America and the Caribbean. CCT programs in many cases were created with the explicit purpose of merging existing programsgenerally duplicative in-kind programs of verified low effectivenessand reducing the operational complications involved. This was the case in Brazil, Ecuador (Bono Solidario), Jamaica, Mexico (the tortilla subsidy), and others.
Persisting challenges
CCT Programs in Urban Areas
Demand for social protection and CCT programs in urban areas is growing. Latin America is experiencing an increasing urbanization of povertyshares of the total poor living in urban areas having increased from 50% in 1961 to 78% in 2007 (World Bank 2009). This implies that, although poverty rates in rural areas tend to be higher in relative terms, absolute numbers of the urban population living in poverty are high and growing. The increasing effort in Latin America to expand CCT programs to urban areas is further motivated by the specific vulnerability of the urban poor to the recent economic shocks (section REF _Ref234585738 \r \* MERGEFORMAT 3.2),. Most programs in Latin America still mainly operate in rural areas (except for Argentinas Plan Familias). Only Colombia and Mexico have expanded their originally rural CCT programs into cities.
Urban CCT programs require a different design. The summary of impact evaluations of CCT programs in REF _Ref234676325 \h \* MERGEFORMAT Appendix 2 includes the urban areas of Colombia and Mexico. The achievements in nearly all of the desired sector outcomes are considerably lower in the programs that have been expanded to urban areas based on the same operational design as the rural program. Given this experience, countries that are currently considering implementing CCT programs or program components in urban areas (Honduras, El Salvador, and Colombias and Mexicos next phase) aim at developing a specifically urban program design. The required design adjustments include targeting tools and mechanisms, the amount of the transfers, and the type of conditionalities and logistical arrangements (such as opening hours of service providers) that help families fulfill their co-responsibility. Table 5 summarizes the urban expansion plans and limited related experience made by CCT programs in the region.
Table SEQ Table \* ARABIC 5: Urban Expansion of CCT Programs in Latin America
CountryYear of Urban ExpansionType of Design Adaptation
Lessons LearnedColombia2007Adapted urban design from the beginning based on pilot experiences, including
(i) modified amounts and differentiated structure of payments (increasing scholarships for secondary education) varying by city to take into account higher opportunity costs in urban areas;
(ii) payment via banks and debit cards instead of cash (including financial literacy activities); and
(iii) use of adjusted geographic targeting to identify poorest neighborhoods, including geographic aggregations by household targeting tool and additional adaptation of rankings by local authoritiesLower program impacts in previously covered small urban areas (municipal capitals) led to design pilot tests and evaluations accompanying the expansion to large citiesEl SalvadorplannedHondurasplannedMexico200420042009: same (rural) designLower program outreach and impact based on rural design ( REF _Ref234676325 \h \* MERGEFORMAT Appendix 2)
2009: pilot test of urban innovations including
(i) modified structure of scholarships to take into account higher opportunity costs in urban areas;
(ii) modified nutrition strategy to take into account presence of both micronutrient deficiencies and obesity (sprinkles plus nutrition education instead of papilla);
(iii) modified health package and attention strategy taking into account higher opportunity costs of seeking care and differing epidemiological profile in urban areas;
(iv) payment via banks and debit cards/savings accounts instead of cash;
(v) new education talks format (hiring specialized nongovernment organizations instead of using health clinic staff);
(vi) use of multidimensional rather than unidimensional (monetary) household targeting Experimental evaluations launched for education and health innovations implemented since 2009Source: Authors.
Specific targeting instruments and procedures are needed for urban areas. Evidence from Mexico indicates a need for specific instruments and procedures, including purely urban household targeting tools. Urban income poverty is very dynamic. Results from a panel study in urban areas of Mexico indicate that only 7% of people defined as extremely poor in 2002 were still extremely poor in 2007, and that there is a great deal of movement into and out of extreme poverty (Rascon and Rubalcava 2008). This dynamism has implications for the use of a monetary proxy means test to measure structural or chronic poverty at the household level in urban areas.
If geographic targeting is to be used, a much more complicated exercise should be expected compared to what is needed in rural areas. As urban poverty may be scattered in small concentrations around cities and may even be right next to high-income neighborhoods, urban poverty maps require much more geographic disaggregation to identify small areas of concentrated poverty. In addition, the identified poverty pockets do not usually correspond to an administrative division of geographic units in cities, which may require a redefinition of geographic urban areas to achieve an effective and operationally feasible targeting result. The case of Mexico further shows that the definition of urban clusters, in which beneficiaries are enrolled at registration booths based on demand in order to avoid a census-type beneficiary identification, may not have the expected effect. Close to half of the potential beneficiaries did not know about the program and/or did not approach the registration point to apply.
The structure and amount of the transfers provided by an urban program need to be adjusted. In urban areas, the opportunity cost in terms of labor market income and direct costs of schooling are likely to be high (compared with rural areas). To effectively function as compensation for school attendance, the amount and structure over time of a transfer designed to act as an incentive to stay in school and avoid early labor market participation should correspond to the age-related opportunity costs of urban school children. Recent simulations show that the amount needed to fully compensate for the labor market income of 16-year-olds would mean quadrupling the amounts currently paid at this age in Mexico, for example (Azevedo et al. 2009). While providing valuable input for theoretic program design, simulation exercises do not tell the whole story as observed in practice.
Specific social dynamics in poor urban neighborhoods require additional attention, particularly to youth at risk. An important percentage of school-age children and adolescents do not attend school or work. Their drop-out from school thus cannot be explained by the need to enter the labor market to complement household income. In Mexico, as many as 51% of the total urban 1218 year-olds (poor and non-poor) do not attend school or work (based on the National Household Income and Expenditure Survey (Encuesta Nacional de Ingresos y Gastos de los Hogares [ENIGH] data from 2006). Among the poor, school attendance is about 7 percentage points higher for Oportunidades beneficiaries than for nonbeneficiaries, suggesting a positive program impact on school enrollment (Azevedo et al. 2009). In El Salvador, as many as 13.6% of adolescents 1117 years in poor families comprise a similar nonworking, nonstudying urban youth population (Tejerina and Johannsen 2009). Reasons for not attending school include that the child does not like school, does not learn, or that school is too far away (de Janvry and Sadoulet 2006). In urban areas, however, the phenomenon is most likely also related to situations where children are prone to crime, neglect, or other risky behaviors (including drug addiction). The figures suggest that neither the school system nor the labor market is able to accommodate these children.
Under these conditions, while urban program design should consider a higher and age-differentiated transfer amount in favor of secondary education when compared to rural areas, the limited experience of urban CCT programs also indicates that an urban intervention should be integrated and combined with supply-side reforms to enhance service quality, as well as complementary social interventions and subsequent labor market policies for the urban youth.
CCT Programs Capacity to Respond in Crisis Situations
Since 2008, Latin Americas CCT programs have faced repeated and severe crisis challenges. Challenges have included the food price increase in early 2008 and the regionwide economic and social consequences of the global financial crisis since late 2008. Inter-American Development Bank simulations for the Andean region (Bolivia, Colombia, Ecuador, and Peru) during the 2008 food price increase concluded that the increase in extreme poverty in the four countries was higher than the estimated increase in terms of total poverty. In other words, a considerable proportion of households that were already poor prior to the food price increase probably entered the ranks of the extremely poor. These households outnumber the initially non-poor households that became (transitorily) poor after the food price increase (Cuesta and Jaramillo, 2009). The simulations indicate that Bolivia has been particularly affected; it does not have a CCT program targeted to the extremely poor.
CCT programs are principally considered to be able to mitigate the impact of aggregate shocks on the structurally poor. In Mexico, for example, the Oportunidades program protected poor rural families from further impoverishment in the aftermath of the Tequila crisis 1994/95 (Tejerina 2009). Conceptually, via the compensatory support to household incomes, the government purchases the complementary effort of the households that is needed for an effective investment in human capital and other productive assets.
In response to crisis, CCT programs are able to use their program instruments to protect the poor. REF _Ref234215217 \h \* MERGEFORMAT Table 6 summarizes the main instruments available to CCT programs that might be modified in the face of a sustained economic crisis.
Table SEQ Table \* ARABIC 6: CCT Instruments to Protect the Poor During Economic Downturns
Program InstrumentOptionsIssues to Consider1. CoverageExpand coverage of already eligible and identified households (compare poverty and coverage figures in REF _Ref234585649 \* MERGEFORMAT Table 1)Where undercoverage exists and eligible beneficiaries are already identified, resources can be reallocated from other uses to expand coverage
Undercoverage is generated by factors beyond budget limitations, including absence of banking facilities, suitable supply conditions, etc.2. Amount of cash transferMaintain transfer amount in real terms to prevent inflation-related erosion in purchasing power
Increase or decrease (temporarily) to compensate for estimated reductions in other sources of household income or to respond to severe budget constraints in a persisting crisisDifficult to revise transfer amounts downward (good communication strategy needed)3. Frequency of transfersIncrease the frequency of payments (if not already at least twice monthly)
Implies more frequent verification of compliance with conditions and a greater number of transactions, resulting in higher administrative costs
Cost-effectiveness of such a modification should be estimated ex ante and monitored4. Definition of eligible beneficiary populationConsider the expansion of the eligibility criteria for beneficiaries, resulting in new beneficiary populations either due to wider eligibility criteria in terms of household characteristics or because of a recertification in already covered regions (including those identified as extreme poor prior to the crisis)Eligibility for a new, temporary subsidy may include families without children
In practice, it is difficult to distinguish between a new and old extreme poor household when using proxy means tests to select beneficiaries
Consider (multidimensional) targeting strategies based on program objectives rather than income proxy means tests
5. Conditionalities/ co-responsibilitiesWaive conditionalities where supply conditions are not in place in order to expand program to eligible households in new areas
Announce intention to introduce conditionalities and verify compliance in the future
Work toward strengthening supply in the futureStudies from Ecuador and Honduras suggest that the announcement of compliance verification alone can generate behavior change toward the desired increased use of basic social services
Program can be introduced alongside supply-side strengtheningSource: Johannsen, Glassman, and Tejerina (2009)
The instrument to be adjusted and the scope of its adjustment depend on the specific country context. The context will include existing antipoverty programs, status of an existing CCT program, and microeconomic effects of crisis. A first priority could be to extend programs to the precrisis structural poor, which would be desirable independent of the nature and scope of the an economic downturn. A second priority might be to increase the amount paid (and payment frequency, if not already twice monthly). Programs with lower basic benefits have larger scope for increasing transfers, while in other settings, scaling up benefits might have undesirable effects on other policy objectives such as the number of beneficiary households entering the labor market. No matter which program instrument is used, it is critical to consider costs, timing, and reversibility of actions taken, as CCT programs are increasingly treated as entitlements. In addition, the weighting of these considerations may vary considerably according to the duration of a crisis. The only countries that have adjusted their CCT program in response to an economic crisis are Jamaica and Mexico. Mexicos Oportunidades increased its transfer amount by M$120 (about $10) per household in 2008 to compensate its beneficiaries for losses in purchasing power during the economic downturn due to the global financial crisis since 2008. In response to the food price inflation in 2008, Jamaica approved an increase in the individual cash subsidies from J$530 to J$650 and an expansion of the total number of Program for Advancement through Health and Education (PATH) beneficiaries from 245,000 to 360,000 beginning in June 2008.
If earlier recessions indicate the future impact of the current downturn in Latin America, urban populations will be more affected than rural populations. CCT programs are more likely to be in place in rural areas, which would generally call for an expansion of these programs to the previously uncovered, extremely poor in urban areas. Where the implementation of a new urban CCT program might be recommendable and feasible under an ongoing crisis context, the program would have to be adapted to the very different barriers to human capital accumulation in urban settings, as described in section REF _Ref234594132 \r \* MERGEFORMAT 3.1, and additional complementary policy options for the recently affected new poor might have to be considered.
The specific targeting strategies and instruments have implications for the ability of a CCT program to mitigate the effects of crisis on certain groups of poor people. Most programs target a subset of extreme poor households that are defined as eligible for the program based on their demographic profile. In general, a program with more restricted eligibility criteria may not reach households most affected by a recession but will in any case protect the highly vulnerable. In addition, CCT programs face specific targeting challenges to identify the new or transitorily poor under crisis situations. Even regularly up-dated proxy means tests (PMTs) will have difficulties identifying nonchronic, transitorily poor who have recently fallen below the poverty threshold. This is because PMTs rely on nonmonetary indicators to predict expenditures. The best predictor variables often include indicator categories such as education and, especially, housing characteristics or selected household assets, which imply certain minimum time lags until changes in a households endowment become measurable. Regularly up-dated PMTs are, therefore, considered to be good targeting tools for extreme, chronic monetary poverty (section 3.3), but not for identifying quickly changing, less severe poverty levels around the poverty line. PMTs will, therefore, not usually identify the new poor, who may still live in a decent house but now lack monetary income to put food on the table, buy medicine, or pay school fees. This limitation is particularly relevant for policy responses to crisis. Therefore, under a relatively short-term scenario of response to economic downturns, other policy and targeting instruments, such as (mostly urban) temporary on-the-job training, might be better solutions and necessary alternatives to effectively reach out to people who have recently and probably transitorily become poor.
Design of Targeting Instruments
Household targeting instruments used in CCT programs are most importantly distinguished by their conceptual and methodological differences. These lead to a distinction between monetary PMT, on the one hand, and composite indexes for alternative, relative poverty concepts and multidimensional deprivation, on the other hand.
Proxy means tests. PMTs provide an operationally feasible alternative to the exact measurement of household income or expenditures by long questionnaires based on the expenditure modules of national household surveys. PMTs either estimate household expenditures or the probability of being poor according to expenditure (or income) criteria by a limited number of proxy indicators, each with its weight, which are ideally selected so they are easy to ask, respond to, and verify. PMTs require simple and significantly shorter questionnaires while still achieving acceptable levels of accuracy, provided they are well designed and periodically updated. They are a common targeting tool implemented by CCT programs in Latin America (e.g., Argentina, Chile, and El Salvador), and at different levels of technical design and using different estimation and classification techniques.
Asset and wealth indexes. They are an alternative to monetary prediction tools in the form of PMT. Most CCT programs use composite asset and wealth indexes based on relative poverty concepts and using methodologies such as principal component analysis. Examples can be found in most CCT programs as well as other development policies (such as microfinance initiatives), including the current household targeting system (Sistema de Focalizacin de HogaresSISFOH) in Peru, the beneficiary identification system (Sistema de Identificacin de Potenciales BeneficiariosSISBEN) in Colombia, and the beneficiary identification and selection system (Sistema de Identificacin y Seleccin de BeneficiariosSELBEN) in Ecuador. Most of them are based on either principal component or factor analysis.
Multidimensional indexes. A further step away from monetary poverty and toward integral deprivation concepts is made by the recent advances in the development of truly multidimensional indexes that attempt to measure and predict poverty from a development concept that considers the monetary income dimension only as means to another, higher end and as part of a multidimensional holistic framework, such as the capability approach first presented by Amartya Sen. Partly because of the difficulties identifying country-specific and agreed criteria for the definitions of multidimensional poverty (and related reference criteria for the evaluation of targeting performance), multidimensional indexes are not yet being implemented by CCT programs. An exception is Mexico, where the National Council for Evaluation of Social Development Policy (CONEVAL) defined a first multidimensional poverty model and is currently pilot testing a related targeting instrument.
Conceptual and methodological inconsistencies persist in many targeting instruments. REF _Ref234739247 \h \* MERGEFORMAT Appendix 3 provides an overview of targeting methodologies at the household level used in selected CCT programs in Latin America. It shows not only the diversity of underlying concepts and approaches across countries and programs but also the methodological inconsistency within specific instruments in some cases. These inconsistencies occur because intended multidimensional poverty concepts lead to the choice of relative poverty indexes based on methods such as the principal component analysis, but are evaluated against monetary reference criteria and accuracy measures, such as inclusion and exclusion errors, with respect to the national (monetary) poverty line. Examples include Colombia, Ecuador, Mexico, Paraguay, and Peru. A more consistent and transparent way of targeting and evaluating a CCTs success would be to
either directly predict expenditure poverty by regression techniques on the basis of true PMTs and compare the resulting poverty classification of households against the poverty line (for a detailed methodological pathway, see Johannsen forthcoming),
or consequently apply alternative, integrated concepts of well-being and vulnerability based on conceptual consensus criteria for multidimensional indexes, and refrain from using the monetary poverty line for evaluating the tools targeting performance.
Most CCT targeting instruments lack appropriate tool updates. Even countries with well-developed survey capacity and information systems will find that their data become out of date and of limited use after a few years and even much earlier in times of crisis and/or rapid economic change. Ideally, the algorithms (including variable composition and weights) and the questionnaires of household targeting tools will be re-administered every few months or every year to keep pace with changing household circumstances, a feat that few countries can manage. The regular update of PMTs, which attempt to predict monetary household expenditures, is particularly recommendable because the level of household consumption is expected to change more quickly across time than long-term wealth or multidimensional poverty indicators, and the best-suited indicator set to predict changing consumption levels might have to be redefined or adjusted.
The Missing and Mistaken Exit Strategy
Current CCT programs in Latin America are not well prepared to release their beneficiaries. Experience with truly results-oriented exit strategies is scarce. Chile Solidario might classify as an example, although it also involves a fixed duration. The program releases its beneficiary households based on an explicitly phased program design consisting of a 2-year intervention and a subsequent 3-year exit phase with frozen payment amounts designed to encourage the graduation to other programs. Chile Solidario, however, was explicitly designed as a bridge program intended to connect families to a variety of social programs and services, rather than being a classical CCT program that seeks to build human capital among the youngest household members. The objectives and any related results-based exit strategy are, therefore, not strictly comparable to other CCT programs.
In most other cases, budgetary limits and government cycles simply lead to exit rules related to a certain maximum time period to be spent in the program or to natural age-related graduation when the maximum age limit for program eligibility is reached (as in Colombia, for example). Table 7 summarizes some of the exit rules or practices.
Table SEQ Table \* ARABIC 7: Exit Rules of Selected CCT Programs
CountryCCT ProgramType of Practiced Exit Rules
a=none
b=duration in years
c=graduation by age
d=results-based strategySequenced Exit
a=none
b=decreasing amounts over time
Comments
BrazilBolsa FamiliacHouseholds recertified every 2 years (poverty status) until age maximum for program participation of 17 years is reached (and depending on school attendance) aChileChile Solidariob+dThe intensive phase of psychosocial intervention and conditionalities lasts 24 months; the exit phase, another 3 yearsbAfter the principal participation phase of 2 years, the exit payment is fixed at the last level received before entering the exit phase The program is explicitly designed as a bridge to facilitate the access to other social programs. ColombiaFamilias en AccioncInitial contract for 4 years, prolongation (without recertification) until age maximum for program participation is reachedaEl SalvadorRed SolidariabFormally after 3 years (but without recertification or actual releases) aProgram started without predefined exit strategy and has not reached the point of dealing with program exitsMexicoOportunidadescRecertification of poverty every 3 years in urban areas and every 6 years in rural areasbDifferentiated exit schemes: households no longer certified as poor but still having children in eligible age groups receive only nutrition transfer for a period PeruJuntosbAfter 4 years (without recertification) aProgram currently plans to design an exit strategySource: Authors.
By the time the first beneficiary generation is realeased, programs experience political and academic pressure to reconsider exit rules. Particularly programs that operate with fixed participation periods may be expected to recertify participating beneficiary households to confirm continuing eligibility based on their poverty status, and to revise the rules for graduation (as is currently the case in Peru and may soon happen in El Salvador).
Consider a beneficiary family with an infant and a primary school child. If the family is released from the program after only 4 years, the infant may not have entered primary education and the older sibling would have just reached the age where the risks of parallel child labor or dropping out of school increase due to the increased opportunity costs of studying. In this situation, the programs impact might be seriously affected as the family continues to face financial constraints to accessing basic education and healthcare. Thus, fixing the years of program participation is the least recommended exit practice for CCT programs.
Most programs lack a results-based exit strategy that does justice to the actual program objectives. Although age-based graduation comes closer to this ideal than a fixed year limit, most programs lack a results-based exit strategy. Colombia provides an example for a de facto age-related graduation practice. Formally, program participation is limited to a fixed number of years, but participation is automatically prolonged as long as the households adolescents are below the maximum age for eligibility, conditional on the ongoing commitment of any new government.
Reference criteria are needed for a results-based exit strategy. The question of what are the reference criteria has led to intense and sensitive discussions in some Latin American programs due to misunderstandings or differing opinions about what the programs should achieve. As discussed in section REF _Ref234594200 \r \* MERGEFORMAT 2.1, CCT programs cannot be expected to reduce chronic poverty immediately and by themselves. Their direct objective can only be to prepare the ground for this by providing demand-side incentives for school attendance and the use of preventive healthcare and other human capital-enhancing services and practices for poor children. Increasing income opportunities for the current adult generation requires different policies and programs.
These considerations have consequences for the conceptual design and expectations related to program exit strategies. Properly defined program objectives and corresponding results-based exit strategies might liberate CCT programs from the burden of designing exit strategies from poverty, maybe even for all beneficiary households, or of being implicitly assigned additional objectives that relate to complementary labor market, agricultural, or microenterprise policies and thus confusing exit strategies from CCT programs with entry strategies to the job market. In contrast, results-oriented CCT exit strategies based on the education cycle (e.g., in terms of finishing a certain secondary school grade) would respond to the program objectives for which CCT programs are designed.
Given such exit strategies, and depending on the complementary supply-side reforms needed for the quality of health and education services, a CCT program would have fulfilled its task and released the graduates. They, and the rest of the countrys young generation, could be taken over by different policies. The main obstacle to these exit strategies is the related budget implications, especially in countries with high poverty rates that would require supporting 30% or more of the population for approximately 10 years and across government cycles. Careful analyses of the relative importance of demand-side versus supply-side barriers to health and education of the poor, as well as of the cost-effectiveness of alternative policy options, are particularly important in such countries.
Beyond the First CCT Wave: Developments of the Next CCT Generation?
The capacity of CCT programs to successfully increase service demand and to stimulate complementary supply-side reforms that enhance the quality of social services to the poor will remain challenging. Irrespective of the relative importance of quality-enhancing supply-side reforms versus monetary demand-side support (section REF _Ref234594160 \r \h \* MERGEFORMAT 2.3), there is consensus in the literature that, without these complementary sector policies, CCT programs will not be able to produce measurable impacts in long-term accumulation of human capital that lay the ground for sustainable poverty reduction. In fact, the quality of the supply side has considerable operational and impact implications for outcomes. Recent evidence from Mexico suggests a close link between the quality of education (especially in marginalized areas, indigenous primary schools, and schools that serve several grades in one classroom) and limited education outcomes of CCT beneficiaries (Mancera et al. 2008). The same applies to the health sectorevidence from Mexico and Nicaragua suggests that additional supply-side incentives within the health sector would increase CCT health impacts (Gutierrez et al. 2008).
In the context of service quality, the educational and knowledge efforts undertaken by the CCT programs merit more attention, given their limited effectiveness (i.e., talks by clinic staff on health issues or food preparation, in which they have no special expertise).
CCT programs are no blueprint or single solution for poverty reduction, nor do they replace an integrated social protection strategy. Complementary social and economic (growth and labor) policies are required and continue to be a major issue of concern to provide real and sustainable development opportunities for the young living in chronic poverty. Although the provision of such opportunities is not the ultimate task of CCT programs nor should it be, CCT programs will not be able to produce measurable long-term impacts without complementary policies in place that address the need for access to employment and increased labor productivity, a necessary condition for a successful interruption of the intergenerational poverty cycle (Levy 2006, Ibarrarn 2009). The experience in Mexico shows that, although beneficiary children are more educated and healthier than their parents, particularly after longer periods of participation in the program, the structure and dynamics of the labor market and macroeconomic circumstances dictate the extent of labor market insertion and productivity experienced by recent program graduates (Behrman et al. 2008, Levy 2008, Rodriguez-Oreggia and Freije 2008).
CCT programs in many but not all countries are already integrated in or are the cornerstone of a broader social protection network. These networks aim at coordinating sector policies that seek to jointly achieve improvements of the health, nutrition, and education status (and eventually social security) of the extremely poor. Examples include Colombias Red Juntos that aims at increasing coverage and impact by coordinating poverty-oriented programs, and Perus ongoing effort in the scope of the Crecer strategy. In other cases, a strategic integration with other sector programs and overall structural sector reforms has not yet been initiated. As a result, duplicated and fragmented social programs lead to inefficiencies and foregone development impacts.
Technology changes and related financial inclusion partnerships may be incorporated in future CCT programs. These two potentially related areas have started to challenge classical CCT design. The demand for measures to reduce operational costs and improve monitoring processes is increasinge.g., the use of information and communication technology to transfer payments and verify co-responsibilities. Examples include the use of debit cards to transfer payments in Colombias and Mexicos urban programs and Brazils Bolsa Familia, as well as plans in the Dominican Republic to take advantage of telebanking practices to transfer educational messages (for example, about deadlines for health check-ups related to conditional compliance) and announcements on payments and balances. Also, the effective use of sectoral information systems (on school attendance, for example) in coordination with the verification of co-responsibilities falls under this category.
Once telecommunication and/or bank cards are considered as an alternative to cash payment, CCT programs can serve as vehicles for the complementary promotion of financial inclusion policies such as basic financial literacy interventions and micro-savings accounts. Colombia, Mexico, and Peru provide examples of CCT programs that operate their payments through savings accounts. The microfinance community is making considerable progress with analyses and experiments using account-based and other technology-based CCT programs with their large coverage and outreach to the poor as potentially attractive vehicles for the wide-spread introduction of savings-mobilization schemes linked to bank accounts. Since 2008, the Proyecto Capital, cofinanced by the Ford Foundation, is planning to scale up the first pilot programs. The CCT programs in Argentina, Brazil, Ecuador, Jamaica, Paraguay, and other countries are potential candidates (Zimmermann and Moury 2009).
It is the CCT communitys turn to respond to these fast initiatives and carefully analyze the related benefits and potential challenges from the viewpoint of the CCT-specific program objectives.
Conclusions
The more than decade-long experience in Latin America and the Caribbean with CCT programs has led to a wealth of evidence on best practices and potential problems in their design and implementation. The culture of evidence-based policy design and transparency inherent in most CCT programs helped produce this information and evaluation results.
Given the many scientific publications based on summarizing and discussing the results of formal impact evaluations, the present paper adopts a different approach and focuses on some of the less discussed potentials and persisting challenges in the design and implementation based on the operational experience with many programs in Latin America. The following paragraphs summarize some of the conclusions that can be drawn from the topics covered in the previous sections.
Different design for different groups. Experience has shown the importance of adjusting the design of CCT programs to the heterogeneous living conditions of the poor in different regions in a given country. This is particularly the case when designing urban CCT programs. Adjustments need to be made to targeting mechanisms, amounts and structure of transfers, incentive structure (e.g., regarding secondary versus primary education), as well as design and logistics of conditionalities (e.g., adapting the health control timing), to name just a few.
Clarity of concepts and constant updates for targeting instruments. To understand how a program reaches out to its intended beneficiary population, the target population must be clearly defined in terms of poverty concepts and reference criteria for targeting success. This should be a basic characteristic of careful program design and part of the logical framework; it is also indispensable for selecting a targeting instrument. A clearly defined beneficiary population will lead to a clearly defined targeting mechanism, which is in turn an important precondition for evaluating a programs targeting performance and success in reaching the intended groups. Regular up-dates of the targeting instruments will ensure the consistency of the tools with the definition of the intended beneficiary population.
Carefully designed exit strategies consistent with CCT program objectives and directly related to human capital objectives. The paper highlights a lack of strategic exit rules designed in accordance with CCT programs objectives of human capital accumulation. While the overall indirect goals related to sustainable reduction of chronic poverty will only be seen gradually and in the long term, it is important that the program has a clearly defined strategy of how to release beneficiary households based on the direct program objectives. The exit steps should be based on the education cycle rather than on fixed year limits or the demographic development of the household.
CCT programs as catalysts of policy reform within and across sectors. CCT programs can be channels that stimulate strategic social policy reforms (including the rationalization of redundant programs) as well as administrative reforms related to the design and wide-spread use of targeting tools, and to monitoring and evaluation systems. If designed from an early stage, the installed capacity can be shared by other programs within and across sectors.
A remaining challenge: how to successfully increase service demand and stimulate complementary policies to enhance the quality of social services to the poor. Demand stimulation caused by CCT programs highlights the gaps in the delivery of social services. This has been shown to be an extremely important side-effect, as it can trigger sector efforts to improve the coverage and quality of social services in health and education, which could be understood as the governments co-responsibilities to fulfill.
Appendix SEQ Annex \* ARABIC 1: CCT Programs in Latin America and the Caribbean and their First Year of Implementation
First Year of ImplementationCountryProgram1995199619971998200020012002200320042005200620072008ArgentinaPlan FamiliasxBrasilBolsa FamiliaxBolsa Alimentao, PETI, Bolsa EscolaxChileChile SolidarioXEcuadorBono de Desarrollo Humano xxMexicoProgresa/ OportunidadesxHondurasPrograma de Asignacin FamiliarxNicaraguaRed de Proteccin SocialxColombiaFamilias en AccinxJamaicaProgramme of Advancement through Health and Education (PATH)xEl SalvadorRed SolidariaxParaguayTekopor xPeruJuntosxDominican RepublicSolidaridad xPanamaRed de OportunidadesxTrinidad and TobagoConditional Cash Transfer ProgramxUruguay Plan de EquidadxCosta RicaAvancemosxBoliviaJuancito Pinto (Education Transfer)xBono Juana Azurduy (Nutrition CCT, planned)GuatemalaMi Familia ProgresaxSource: Authors elaboration
Appendix SEQ Annex \* ARABIC 2: Summary of Impact Evaluations of Some CCT Programs in Latin America
Education NutritionChild LaborHealthCountrySchool AttendanceSchool Enrollment/ AchievementUsage/Access
Brazil+3% n.m.(Bolsa Alimentacao)
+9% food expenditures
+9% dietary diversity
+6% per car capita caloric availability
But less weight gain (PETI)
10% overall child labor
4% point targeted child labor.n.m.ChilePositiven.m.n.m.Positive in usage, none in accessColombia+1.9% rural (8 to 11 year-olds)
+5.1% urban (12 to 17 year-olds)
+7.2% rural (12 to 17 year-olds)n.m.
9% rural chronic undernutrition (height for age), no effect for urban areas
6% rural global undernutrition (weight for age), no effect for urban areas or older children
+2.8 months urban duration of breastfeeding, no effect for rural children5.5% rural labor force participation (1017 years old boys), no effect in urban areas
3.4% rural labor force participation (1017 year old girls), 2.7% in urban areasNo effect on DPT vaccinations
Mixed effect on weight height controls (from 11% to 44%)
No effect on probability of anemiaEcuador+10% enrollmentn.m.-17% n.m.Honduras+4.3% to 4.6% (1 extra day per month).
2.4 to 7% drop out during school year.
4.6% repetition rates grade 1 to 4+17% enrollment (512 year-olds from 2000 to 2001)
No enrollment impacts for average population.
NoneNone+7% to +10% children receiving DTP vaccinations on time.
No effect on diarrhea in children (03 years)
No impact on iron deficiency anemia
Increased fertility
+17% to +20% weight height control in last 30 days
No impact on measles, tetanus immunization in pregnant mothersJamaica+3.15 No effect in achievementn.m.None+38% preventive visits to healthcare practitioners (06 year-olds). No effect on the elderlyMexico+0.69 to +1.31 rural years of schooling completed (913 year-olds)+30.4% to +63.6% rural students with adequate progress (912 year-olds)
+ urban enrollment rates for some age groups (e.g. 10% for boys at age 6)
+7 to 15% urban school progress for all ages (excluding girls age 15 to 18)39.3% rural stunting (26 month-olds females), 19.4% for males-15% to -25% rural probability (1215 year-olds) +12% rural consultations (05 year-olds, last 6 months), +16% in urban areas
22% rural hospitalizations (50+ year-olds, last 12 months), 4% in urban areas
15% and 6% rural maternal and infant mortality (municipalities with > 35% of population)
26.1% rural anemia in girls (36 year-olds), no effect in boys
54% urban anemia prevalence (623 months-olds)
+61% rural births with adequate prenatal care, +32% in urban areasNicaragua+20% (+33% among the poorest)+7.3% passing/retention (+5.5% among the poorest) with higher impact in grades 1 and 4+23.2% per capita annual food expenditures (compared with 14% for the control group).
5.5% stunting (05 year-olds)
6.2% underweight children 5.6%+13.1% children participating in growth monitoring (visit to healthcare clinic and weighted of 03 year-olds).
No significant effect in 1223 month-olds with complete and adequate immunizations (but simple difference effect of 32.6%).Costa Rica+8.7% (1316 year-olds)
+4.8% students passing the school year 2001n.m.
n.m.Nonen.m.DPT = diphtheria, pertussis, and tetanus; n.m. = not measured
Sources: Tejerina (IDB, forthcoming), based on, for Brazil, Cardoso and Souza (2003); for Chile, Galasso (2006); for Colombia, Econometra Consultores (2004); for Ecuador, Leon and Younger (2006); for Honduras, IIFPRI (2003); for Nicaragua, Maluccio and Flores (2005); for Costa Rica, Duryea and Morrison (2004); for Mexico, Rodriguez and Levy (2004); and for Ecuador, Schady and Araujo (2006).
Appendix SEQ Annex \* ARABIC 3: Diversity and Inconsistency in Targeting Methods Used in CCT Programs
CountryNameGeographic ScopeUnderlying Poverty Concept of Index Method of Indicator SelectionEstimation MethodResulting Index Score/ Amount Reference Criteria for Targeting PerformancePoor/Nonpoor Classification (definition of cut-off value)Accuracy MeasuresAdditional Robustness or Validation CriteriaChileFPS National (by occupation levels and
sex)Vulnerability (risk to fall in monetary poverty)Income generating capacity: Manual + minimizing linear combination of inclusion and exclusion errorIncome generating capacity: linear regressions
Subsidy income: Pension institute data
Needs index: based on family members and
their dependency2072 points and up
(income generating capacity + subsidy income) / needs index Ranking into vulnerability deciles based on national survey distribution EcuadorSELBENNationalLiving standard, Income generating capacityManual (trial and error)PCA index scores 0 to 100Monetary poverty lines (household consumption)Poverty headcount (quintiles 1 + 2)Inclusion + exclusion error based on cut-off (also based on distance of > 2 distribution deciles) PeruSISFOHUrban/ rural separatelyHousehold welfare index (declared as nonmonetary)Manual +
correlation with poverty based on Sommers statistics (Madueno, 2007)PCA index scores 0 to 100, divided in 7 SISFOH segmentsMonetary household expenditures (segment 1+2
of index compared to expenditure quintile 1)Minimization of sum of inclusion error and 2x exclusion error, separately for each macro-region Inclusin/ exclusion error (over total population), total errorConfidence interval for cut-off value = not available; FPS = ficha de proteccion social; PCA = principal component analysis; SELBEN = System of Identifying and Selecting Beneficiaries (Sistema de Identificacin y Seleccin de Beneficiarios); SISFOH = Household Targeting System (Sistema de Focalizacin de Hogares).
Sources:
References
Adato, M. and J. Hoddinott. 2007. HYPERLINK "http://ideas.repec.org/p/fpr/2020br/bb17.html" Conditional cash transfer programs: A "magic bullet" for reducing poverty?. HYPERLINK "http://ideas.repec.org/s/fpr/2020br.html" 2020 vision briefs BB17 Special Edition, International Food Policy Research Institute (IFPRI).
Azevedo, V., C. Bouillon, and P. Yez-Pagans. 2009. Simulating Changes to Mexicos Oportunidades in Urban Areas. IDB Working Paper #680. Washington, DC: Inter-American Development Bank (IADB).
Behrman, J. R., L. Fernald, P. Gertler, L. M. Neufeld and S. Parker. 2008. Longer-Run Effects after Almost a Decade on Child Development, Education and Nutrition of Exposure of Infants and Toddlers to Oportunidades in 1998-9. Evaluacin Externa del Programa de Desarrollo Humano Oportunidades 20072008. Mexico City: Instituto Nacional de Salud Pblica.
Borghans, L., and B. Golsteyn. 2005. Imagination, Time Discounting and Human Capital Investment Decisions. Unpublished. http://www.roa.unimaas.nl/cv/borghans/pdfp/Time%20 Discounting.pdf.
Bouillon. C.P. and L. Tejerina. 2007. Do We Know What Works? A Systematic Review of Impact Evaluations of Social Programs in Latin America and the Caribbean. Inter-American Development Bank.
Cardoso, E. and A. P. Souza. 2003. The Impact of Flow Transfers on Child Labor and School Attendance in Brazil. Working Paper 04-WO 7. Department of Economics, Vanderbilt University.
Cuesta, J., and F. Jaramillo. 2009. Taxonomy of Causes, Impacts and Policy Responses to the Food Price Crisis in the Andean Region. IDB working paper #674. Washington DC: IADB
de Janvry, A. and E. Sadoulet. 2006. Making Conditional Cash Transfers More Efficient: Designing for Maximum Effect of the Conditionality. World Bank Economic Review 20 pp. 129.
Econometra Consultores. 2004. Evaluations of the Immediate Consequences of Empleo en Accion. Bogota: Departamento Nacional de Planeacion.
Duryea, S. and A. Morrison. 2004. The Effect of Conditional Cash Transfers on School Performance and Child Labor: Evidence from an Ex-post Impact Evaluation in Costa Rica. Research Department Working Paper 505. Washington DC: Inter-American Development Bank.
Economic Commission for Latin America and the Caribbean (ECLAC). 2008. Social Panorama of Latin America. Briefing Paper. 5. Santiago.
Filmer, D., and L. Prichett. 1999. The Effect of Household Wealth on Educational Attainment: Evidence from 35 Countries. Population and Development Review 25(1) pp. 85120.
. 2001. Estimating Wealth Effects Without Expenditure Dataor Tears: An Application to Educational Enrollments in States of India. Demography 38 (1) pp. 115132.
Fiszbein, A., N. Schady, F. Ferreira, M. Grosh, and N. Kelleher. 2009. Conditional Cash TransfersReducing Present and Future Poverty. Washington DC: World Bank.
Galasso, E. 2006. With Their Effort and One Opportunity: Alleviating Extreme Poverty in Chile. Development Research Group. Unpublished. Washington DC: World Bank
Grosh, M., C. del Ninno, E. Tesliuc, and A. Ouerghi .2008. The Design and Implementation of Effective Safety NetsFor Protection and Promotion. Washington DC: World Bank.
Gutirrez, J. P., J. Leroy, R. Lpez, et al. 2008. Evaluacin de la Calidad de los Servicios de Atencin a la Salud Asignados a la Poblacin Beneficiaria de Oportunidades. Mexico City: Instituto Nacional de Salud Pblica
Handa, S, and B. Davis. 2006. The Experience of Conditional Cash Transfers in Latin America and the Caribbean. Development Policy Review 24 (5) pp. 51336.
Hewett, P. and M. Montgomery. 2001. Poverty and Public Services in Developing Countries. Policy Research Division Working Paper no. 154. New York: Population Council.
Ibarrarn, P. 2009. Labor Markets and CCTs: Implications for Poverty Reduction. Presentation at the Fourth IZA/World Bank Conference Employment and Development. Bonn. 4 May 4.
International Food Policy Research Institute (IFPRI). 2003. Proyecto PRAF/BID Fase II: Impacto Intermedio. Washington, DC: IFPRI and IADB.
Johannsen, J. 2009. Operational Assessment of Monetary Poverty by Proxy Means Tests. Development Economics and Policy Series No. 65. Hamburg: Peter Lang.
Johannsen, J., A. Glassman, and L. Tejerina. 2009. Conditional Cash Transfer (CCT) Programs in the Context of the Financial Crisis in LAC. Unpublished. Washington DC: Inter-American Development Bank.
Leon, M. and S. Younger. 2007. Transfer Payments, Mothers Income and Child Health in Ecuador. Journal of Development Studies 43 (6) pp. 1126-1143.
Levine, R. 2007. Case Studies in Global Health: Millions Saved. Boston: Jones and Bartlett Publishers.
Levy, S. 2008. Good Intentions, Bad OutcomesSocial Policy, Informality, and Economic Growth in Mexico. Washington DC: Brookings Institution.
. 2006. Progress against Poverty: Sustaining Mexicos Progresa-Oportunidades Program. Washington DC: Brookings Institution.
Madueo, M. 2007. Revisin del algoritmo de focalizacin individual SISFOH. Unpublished. Lima: USAID, PRAES.
Maloney, W. 2001. Evaluating Emergency Programs: Intertemporal and Financing Considerations. Washington DC. World Bank.
Maluccio, J. and R. Flores. 2005. Impact Evaluation of a Conditional Cash Transfer Program: The Nicaraguan Red de Proteccion Social. Research Report No. 141. Washington DC: International Food Policy Research Institute.
Mancera C., C., L. Serna H., A. Priede S.. 2008. Modalidad educativa y organizacin multigrado como factores asociados a las brechas de aprendizaje de los becarios de Oportunidades de primaria y secundaria en 2007. Evaluacin Externa del Programa de Desarrollo Humano Oportunidades 20072008. Mexico City: Instituto Nacional de Salud Pblica.
Ministerio de Planificacin y Cooperacin (MIDELPLAN). 2004. Conceptos Fundamentales, Sistemas de Proteccin Social, Chile Solidario. Santiago de Chile: MIDEPLAN.
Montgomery, M.R., M. Gragnolati, K. Burke, et al. 2000. Measuring Living Standards with Proxy Variables. Demography 37 pp. 15574.
Rascon, E., and L. Rubalcava. 2008. Diagnstico General de las Barreras a la Acumulacin del Capital Humano en Zonas urbanas en Mxico (Perfil de la Pobreza Urbana). Consultant Report. Washington DC: Inter-American Development Bank.
Levy, S. and E. Rodriguez. 2004. Economic crisis, political transition and poverty policy reform: Mexicos Progresa-Oportunidades Program. Washington DC: Inter-American Development Bank.
Rodriguez-Oreggia, E. and S. Freije. 2008. Evaluacin Laboral de Jvenes Beneficiarios del Programa Oportunidades en el Largo Plazo. Evaluacin Externa del Programa de Desarrollo Humano Oportunidades 20072008. Mexico City: Instituto Nacional de Salud Pblica.
Rubinstein, A. 2003. Economics and Psychology? The case of Hyperbolic Discounting. International Economic Review 44(4) pp. 120716.
Sahn, D., and D. Stifel. 2000. Poverty Comparison Over Time and Across Countries in Africa. World Development 28 (12) pp. 212355.
Schady, N. and M. C. Araujo. 2006. Cash Transfers, Conditions, School Enrollment, and Child Work: Evidence from a Randomized Experiment in Ecuador. Policy Research Working Paper 3930. Washington DC: The World Bank.
Tejerina, L. 2009. Perfil de Riesgos Enfrentados por Familias de Bajos Ingresos en Areas Urbanas de Mxico. Unpublished. Washington DC: Inter-American Development Bank.
Tejerina, L., and J. Johannsen. 2009. Proteccin Social en El Salvador. Policy paper. Washington DC: Inter-American Development Bank.
World Bank. 2009. World Development Indicators. Washington DC: World Bank.
Zeller, M., S. Manohar, C. Henry, and C. Lapenu. 2006. An Operational Method for Assessing the Poverty Outreach Performance of Development Policies and Projects: Results of Case Studies in Africa, Asia, and Latin America. Washington DC: International Food Policy Research Institute.
Zimmermann, J., and Y. Moury. 2009. Savings-Linked Conditional Cash TransfersA New Policy Approach to Global Poverty Reduction. Global Asset Project Policy Brief. Washington DC: New America Foundation.
Maloney (2001) estimates the poors discount rate of investment components to be as high as 30%, which implies high utility costs for the poor if program resources are spent on investment in human or physical capital rather than in immediate consumption.
The discussion so far ignores the question of whether and which supply-side or demand-side barriers are more important determinants of the observed inequities in health and education (sections REF _Ref234594160 \r \h \* MERGEFORMAT 2.3 and REF _Ref234808237 \r \h \* MERGEFORMAT 3.5), and which policy options are more cost-effective at reducing the barriers. For now, it is assumed that both demand- and supply-side barriers are important and that CCT programs mostly intervene on the former while potentially having some indirect effects on the latter.
Complete coverage of the extremely poor is understood as a theoretically perfect outreach ignoring inevitable targeting errors.
Including interventions such as the scale-up of health technologies, strengthening of health systems, and the use of health education and policy change that showed to achieve impressive reductions in disease and disability, even in the poorest countries (Levine, 2007)
Demand-side transfers refer to the utilization of health services, in general, as well as of preventive care (such as growth controls and vaccinations), in addition to education impacts in terms of school inscription, attendance and reduction in drop outs (IFPRI 2003).
One relatively well-studied aspect of cost-effectiveness refers to the share of operating costs spent on program administration. Administrative costs of Latin American CCT programs amount to 8.2% on average, ranging from 4.1% in Ecuador in 2005 to 13% in Jamaica in 2004/05 (Grosh et al. 2008). As a rule of thumb, well-executed CCT programs show administrative costs of 8%15% (ibid), indicating neither an underdeveloped administration nor inefficient cost inflation due to leakages, losses, or design failures. It is, however, important to consider the year a particular program started as the share spent on setting up targeting, registration, payment processes, and monitoring systems might vary considerably across time within a program, with higher investment and processing costs during the first 3 years (Adato and Hoddinott, 2007.
In Brazil, the current program Bolsa Familia was created to replace former CCT programs that tackled specific problems such as child employment (the Programme for the Eradication of Child LabourPETI), food insecurity (Bolsa Alimentacao), and low school attendance (Bolsa Escola).
The results of simulating changes to Mexicos Oportunidades in urban areas indicate that eliminating or reducing school subsidies for primary education and increasing the transfer for older students is a cost-effective way to raise overall school enrollment in urban areas. Increasing school attendance of 16-year-olds to 80% or more, however, would indeed require quadrupling scholarships.
Anecdotal evidence from El Salvador and Mexico relates the phenomenon of neglect to high emigration rates and the social consequences of children left behind by their emigrating parents and raised by other family members.
The government of Bolivia is designing and preparing the CCT program Bono Juana Azurduy, to be implemented during 2009.
For a further discussion of specific adjustments in different contexts of crisis duration and CCT status, see Johannsen, Glassman, and Tejerina (2009).
Even at its increased value, the subsidy is still modest compared to other CCT programs, at about $9 per month for each beneficiary (representing about 10% of the per capita expenditures in poor beneficiary households) and an average monthly household grant of $27.
Poverty simulations for the Andean region during the 2008 food price increase concluded that impacts were considerably higher on urban than on rural poverty in some but not all cases, particularly Bolivia (Cuesta and Jaramillo 2009).
Chiles social protection index, existing since 2007, includes a PMT within its index of vulnerability to monetary poverty.
Examples in the literature include the asset index by Filmer and Prichett ( ADDIN EN.CITE Filmer199970707017Filmer, DeonPrichett, LantThe Effect of Household Wealth on Educational Attainment: Evidence from 35 CountriesPopulation and Development Review85-120251asset indexDHSpovertyindicatorsasset index199919801999 and ADDIN EN.CITE Filmer200171717117Filmer, DeonPrichett, LantEstimating wealth effects without expenditure data - or tears: An application to educational enrollments in states of IndiaDemography115-132381asset indexpovertyindicatorsIndia2001http://www.jstor.org/view/00703370/sp030001/03x0010j/02001); the multidimensional Consultative Group to Assist the Poor (CGAP) poverty index by Zeller et al. ( ADDIN EN.CITE Zeller200699999917Zeller, ManfredSharma, ManoharHenry, CarlaLapenu, CecileAn operational method for assessing the poverty outreach performance of development policies and projects: Results of case studies in Africa, Asia, and Latin AmericaWorld DevelopmentWorld Development446-464343poverty assessmenttargetingevaluationAfricaAsiaLatin America20062006/3http://www.sciencedirect.com/science/article/B6VC6-4J6WG18-2/2/fb31bcf41eb9e01d28568489d3b1fa902006); and the wealth indexes by Montgomery et al. ( ADDIN EN.CITE Hewett200127127117Montgomery, Mark R.Gragnolati, MicheleBurke, Kathleen Paredes, EdmundoMeasuring Living Standards with Proxy VariablesDemography155-1743720002000), Sahn and Stifel ( ADDIN EN.CITE Sahn200093939317Sahn, David E.Stifel, David C.Poverty comparison over time and across countries in AfricaWorld DevelopmentWorld Development2123-21552812welfare indexasset indexmulti-dimensional povertyAfricafactor analysis20002000), and Hewett and Montgomery ( ADDIN EN.CITE Hewett200137837837827Hewett, Paul C.Montgomery, Mark R.Poverty and public services in developing-country cities154asset indexfactor analysis2001Population Council Working Paper2001).
This caveat applies in the context that PMTs (and, even more so, the other targeting tools described) have been introduced as instruments that are particularly well suited to identify the chronically poor, given that even regularly updated PMTs will have difficulties in identifying the nonchronic, transitory poor who have recently fallen below the poverty threshold.
For more information on Chile Solidario, refer to MIDELPLAN (2004)
A strategy to increase supply-side coverage and quality could be designed on the basis of performance-based incentives (payment for performance schemes).
PAGE 2
PAGE 3
PAGE 2
PAGE 36
PAGE 31
Oportunidades baseline
.........
- . H I ] ^ ijğĉ|l|^|Q j ho h0 6U ho h[OW 0JK 6mH
sH
j ho h[OW 6U j ho h[OW 6U ho h[OW 6mH
sH
ho h[OW 0JK 6&j ho h[OW 6UmHsH j ho h[OW 6UmHsH ho h[OW 6ho hI$ 6CJ aJ ho h0 6CJ aJ ho hI$ ;ho h0 ;ho hAG ;hN . I J K L 3 4 5 6 7 8 9
x
q 5 4
!
gd 2 gdGM ? gd[OW ? gdb5 > gdAG / 0 3 tZD +ho h$Ya @ CJ OJ QJ aJ mH nH u2j ho h$Ya >*B*UmH nH ph u ho h$Ya mH nH uho h$Ya 0JK mH nH u$j ho h$Ya 0JK UmH nH u j ho h 2 Uho h[ ho h 2 ho hI$ ho h[OW ho h[OW 6ho h0 0JK 6j ho h0 6U j ho h0 6U ho h0 6
!
"
$
%
V
W
X
r
s
t
u
v
w
x
y
z
{aM{&j ho h$Ya UmH nH u 2j ho h$Ya >*B*UmH nH ph u ho h$Ya mH nH uho h$Ya 0JK mH nH u+ho h$Ya @ CJ OJ QJ aJ mH nH u$j ho h$Ya 0JK UmH nH u ho h><