PHILIPPINE SOCIAL PROTECTION NOTE October 2015 NO. 8 “Sa Pantawid, Malapit nang Makatawid!” (With Pantawid, we are closer to getting out of poverty!) An Update of the Philippine Conditional Cash Transfer’s Implementation Performance Pablo Ariel Acosta and Rashiel Velarde1 This is the second benefit incidence analysis of the Philippines’ conditional cash transfer program conducted by a World Bank team that uses standard measures to assess the implementation performance of the Pantawid Pamilyang Pilipino Program. We show that despite the program’s rapid expansion since it was piloted in 2007, it maintains good targeting accuracy, progressivity, and cost- efficiency in delivering assistance to the poor. We also show that Pantawid Pamilya helps reduce short- term poverty by closing the income gap of beneficiary households by up to 7.4 percentage points, helping them afford their basic needs. Among beneficiaries, the program reduces food poverty and total poverty by up to 6.7 percentage points. To maintain the program’s relevance, adequacy of assistance, sustained impact on beneficiary welfare, and adjustments in the benefit level and program coverage at a sustainable level are recommended. I. Introduction The Pantawid Pamilyang Pilipino Program (Pantawid Pamilya) is the Philippines’ national conditional cash transfer (CCT) program. It was conceptualized at a time when the Philippines’ economic growth had been sustained for a decade and yet not showing results on poverty reduction. The program provides cash grants to poor households that have limited capacity to invest in their children’s future and well-being. This Philippine CCT program was patterned after the pioneers of conditional cash transfer programs in Latin America, the Bolsa Familia of Brazil and Oportunidades of Mexico (now called Programa de Inclusión Social or PROSPERA), which received a lot of acclaim in effectively helping poor households. The Philippine Social Protection Note Pantawid Pamilya plays a central role in the Government’s overall strategy to address poverty and series aims to disseminate experiences, vulnerability among Filipinos. The program’s objective is two-fold. In the short-term, the Pantawid good practices, and key findings from Pamilya provides income support to poor households helping them afford their basic needs. In the the Philippines on the topics related to long-term, by conditioning the cash support to desired health and education behavior, it incentivizes social protection. It also aims to broaden the dialogue on social protection and stimulate public engagement in moving forward the policy agenda. The views expressed here are those of the authors 1 This note was prepared Pablo Ariel Acosta (Senior Economist) and Rashiel Velarde (Economist) as part of the and do not necessarily reflect those of Social Protection and Labor Programmatic Analytical and Advisory Activities (SPL PAAA), under the guidance of the World Bank. Aleksandra Posarac (Program Leader and Task Team Leader for the SPL PAAA) and Jehan Arulpragasam (Program Manager for SPL). The authors would like to thank the peer reviewers, Maddalena Honorati and Margaret Grosh, and other colleagues from the Social Protection and Poverty Global Practices for their invaluable comments. www.worldbank.org 1 “Sa Pantawid, Malapit nang Makatawid!” (With Pantawid, we are closer to getting out of poverty!) investment in the well-being of children so they can be more productive citizens in the future and break away from the cycle poverty. With the Department of Social Welfare and Development (DSWD) as lead agency, Pantawid Pamilya was piloted in the Philippines in 2007 with 6,000 household beneficiaries. It was launched formally in 2008 with coverage of 320,000 households. Since then, the program has expanded to 4.4 million households in 2015, or 100% of all poor households with children as identified by the national household targeting system for poverty reduction – Listahanan. Covering about 21 percent of the population, the Pantawid Pamilya is currently the third largest CCT program in the world, next to Brazil, covering 29 percent of its population, and Mexico, covering 27 percent (Figure 1). Figure 1. Pantawid Pamilya program coverage, 2007-2015 Sources: DSWD Pantawid Pamilyang Pilipino Program Implementation Status Reports (various issues); World Bank Group ASPIRE. Since the program started in 2007, significant resources have been invested in Pantawid Pamilya. The program started with a meager budget of PHP 4 million (USD 0.1 million2) to cover 6,000 pilot households in 2007, which represented a near-zero share of the country’s GDP and national budget (Figure 2). By 2014, the program grew to PHP 62.2 billion (USD1.5 billion), reaching 4.2 million household beneficiaries and representing 0.5% of the country’s GDP and 2.8% of the Government’s national budget. At this size, the program is mostly financed out of the Government’s own funds, with less than a quarter (22 percent) coming from external financing of development partners. II. Overview of Pantawid Pamilya program design Program eligibility combines the household’s poverty status and its demographic profile. As with other CCT programs, eligibility for the Pantawid Pamilya requires that a household has either a pregnant mother or has at least one child, and agree to comply with the program conditions. But first, a household has to be identified as “poor” as determined by the national household targeting system (Listahanan), after it has undergone the process of community validation. Poor households with children that are potential beneficiaries of Pantawid Pamilya are sourced from the Listahanan,3 the Government’s national household targeting system. Listahanan was created in 2007 to initially identify poor households that could benefit from the Government’s CCT program, the Pantawid Pamilya. Since its establishment, it has grown to become a nationwide household-based 2 PHP 43 = USD 1 for simplicity. 3 Formerly known as the National Household Targeting System for Poverty Reduction (NHTS-PR). For more information on the design and features of the NHTS-PR - Listahanan, refer to Fernandez (2012). 2 PHILIPPINE SOCIAL PROTECTION NOTE targeting system that is used to identify beneficiaries of almost all government programs targeted at poor and vulnerable population. Listahanan is a proxy means test – (PMT–) based targeting system. Using a standard set of household information that is easy to collect, measure, and verify, Listahanan estimates the income level of the household and compares it against the Government’s official income poverty threshold to determine the household’s poverty status. By collecting the same set of information across all households and gauging them against the same metric, all households are assessed in a uniform and objective manner, leaving subjectivity and political influence aside. Prior to Listahanan, various public social assistance programs applied their own beneficiary selection mechanisms, in many cases decided by subjective means and influenced by partisan politics at the local level. As a result, resources allocated to programs leaked to unintended beneficiaries – mostly to those who had political connections and were not necessarily poor – while the poor became marginal recipients of government assistance.4 Listahanan currently holds 11 million households in its database (out of 20 million households nationwide), of which 5.2 million households were classified as poor, among which 4.4 million had children 0-14 and or pregnant women and were gradually enrolled in Pantawid Pamilya as its beneficiaries. Other government programs have also been using the Listahanan to identify and locate their target beneficiaries.5 Department of Social Welfare and Development is currently completing the update of Listahanan. It is planned to collect information from 15.3 million households or 75 percent of all households in the Philippines. Beneficiary households are entitled to receive health and education cash grants, depending on the composition of their household and their actual compliance with program conditions. The Pantawid Pamilya provides uniform benefit levels for education and health across the country. However, different conditions apply to different members of a beneficiary household (Table 1). Education grants are given to each school-aged child in the household but only up to a maximum of 3 children per beneficiary household. Each child between 3 and 14 years old who is enrolled in a kindergarten/pre-school/elementary school is entitled to a grant of PHP 300 (USD 6.98) per month if she/he attends at least 85 percent of the school days in a month. Beginning January 2014, education grants were extended to poor households with children aged 15-18 years old and enrolled in secondary education. Each child between 15 to 18 years old who is enrolled in secondary school is entitled to a grant of PHP 500 (USD 11.63) per month if she/he attends at least 85 percent of the school days in a month. Education grants are only given for 10 months every year because there are two vacation months from school (usually March and April). The PHP 500 (USD 11.63) monthly health grant is an all-or-nothing grant. It is received by the beneficiary household only if it complies with all applicable health conditions as summarized in Table 1. Health grants are provided for twelve months every year. Based on this schedule, beneficiary households can receive up to PHP 21,000 per year, depending on their household composition and compliance with applicable program conditions. In 2013, prior to the program expansion to 15-18 year old children, beneficiary households could receive up to PHP 15,000 per year. Table 2 illustrates the grant schedule for beneficiary households with varying compositions. With the exception of the first release of the cash grant, succeeding payments of the grant to beneficiary households are based on actual compliance with the conditionality. Table 1. Summary of program conditions Education conditions Health conditions 1. Children 3-5 years old enroll in pre-school 1. Children below 5 years old go for monthly visit to health station to avail of or day care facilities and maintain school age-appropriate health checks and services as prescribed by DOH. attendance of at least 85% of school days 2. Children 6-14 years old take deworming pills twice a year in school. per month 3. Pregnant women go for trimestral consultations during pregnancy. 2. Children 6-18* years old enroll in elementary or high school and maintain 4. Pregnant women birth delivery attended by a skilled health worker. school attendance of at least 85% of 5. Grantee and/or spouse attend/s monthly Family Development Sessions school days per month (FDS). Source: Pantawid Pamilya Operations Manual (as of September 2014). *Extension of program conditions and benefits to 15-18 year old school children began in January 2014. 4 Manasan and Cuenca (2007); Fernandez and Velarde (2012). 5 Other national programs that used Listahanan to identify and locate beneficiaries include Universal Health Coverage component of the National Health Insurance Program (Philhealth), Student Grant in Aid Program of the Commission on Higher Education, Socialized Housing Project of the National Housing Authority, Targeted Actions to Reduce Poverty and Generate Economic Transformation (TARGET) of the Department of Agriculture, Social Pension of DSWD, Sustainable Livelihood Program of DSWD, and National Community Driven Development Program of DSWD. As of March 2015, total of 44 national government agencies use Listahanan to target 25 programs aimed at poor and vulnerable Filipinos. www.worldbank.org 3 “Sa Pantawid, Malapit nang Makatawid!” (With Pantawid, we are closer to getting out of poverty!) Table 2. Sample Pantawid Pamilya Grant Entitlements Grant entitlements for Type of Beneficiary Household Applicable Program Conditions* March and May 2015 Household A: Health: Health grant = Php500 Parents with a 4-year old child not enrolled in (i) Monthly FDS; and day care (ii) Growth and weight monitoring for the MAXIMUM ENTITLEMENT: FOR MARCH 4-year old child in the health center. 2015 = PHP 500 FOR MAY 2015 = PHP 500 Household B: Health: Health grant = PHP 500 Widow with a 10-month old infant, a (i) Monthly FDS; 5-year old child, and a 13-year old child in (ii) Growth and weight monitoring plus age- elementary school appropriate inoculations for the 10-month old infant in the health center; and (iii) Deworming for the 13-year old child. Education: Education grant = PHP300 for the 5-year old (i) If 5-year old child is enrolled in Day Care or child + PHP300 for the 14-year old child pre-school, 85% attendance in March; and (ii) 85% attendance in March for the 13-year MAXIMUM ENTITLEMENT: FOR MARCH old child. 2015 = PHP 1,100 FOR MAY 2015 = PHP 500 (no school) Household C: Health: Health grant = PHP 500 Grandparent looking after a 14-year old child (i) Monthly FDS. in high school and a 17-year old child in high Education: Education grant = PHP 500 for the 14-year school old child + PHP 500 for the 17-year old child (i) 85% attendance in March for the 14-year old child; and (ii) 85% attendance in March for the 17-year MAXIMUM ENTITLEMENT: FOR MARCH old child. 2015 = PHP 1,500 FOR MAY 2015 = PHP 500 (no school) *Applicable age-appropriate health conditions are as prescribed by DOH. Compliance monitoring is conducted in coordination with partner government agencies. Pantawid Pamilya is implemented in partnership with other government agencies that comprise the program’s National Advisory Committee (NAC). These agencies include the Department of Education (DepEd), Department of Health (DOH), National Nutrition Council (NNC), Department of Interior and Local Government (DILG), and National Anti-Poverty Commission (NAPC) among others.6 These agencies are regularly consulted on matters related to the implementation of Pantawid Pamilya. In addition, the DepEd and DOH have special roles in monitoring compliance of beneficiaries with the program’s education and health conditions. Compliance with conditions is systematically recorded through the program’s Compliance Verification System (CVS). CVS paper forms are filled out by school and health personnel based on their records. The accomplished forms are collected by Pantawid Pamilya ground coordinators, called Municipal Links, and sent to DSWD regional offices to be recorded in the computer-based CVS. Records that appear in the system are used to compute grants that will be paid to each household beneficiary and to generate the payroll. Based on latest program data, compliance to program conditions is high. Over years of implementation, compliance with program conditions has improved. As of September-October period in 2014, according to the program Compliance Verification System, beneficiary households are almost universally compliant with Pantawid conditions (Table 3). This, in turn, means that beneficiaries should be receiving their maximum entitlement from the program. 6 Joint Memorandum Circular No. 4 series of 2013. 4 PHILIPPINE SOCIAL PROTECTION NOTE Table 3. Compliance with program conditions Program Conditions % of Compliant HH beneficiaries (As of March (March- (March- (Sept-Oct (Jun-July 2010) April 2012) April 2014) 2014) 2015) Education 94.19 96.67 98.81 96.49 94.79 Attendance in Day Care Center/Pre- School for Children 3-5 years old 96.23 97.66 96.13 88.75 Attendance in Primary and Secondary Schools for Children 6-14 years old 97.11 99.02 97.34 95.61 Health 80.04 96.86 97.14 97.20 97.87 Check-up/immunization for Pregnant and Children 0-5 years old 95.48 96.05 96.31 95.45 Health (deworming) for Children 6-14 years old in Elementary Level 98.29 99.90 99.60 99.03 Attendance of FDS by Parents 84.28 96.15 95.08 94.86 95.07 Source: DSWD Pantawid Pamilyang Piliino Program Implementation Status Reports (various issues). Grants are delivered through various channels with the objective of minimizing the cost to the beneficiary. Initially, program grants were delivered to beneficiaries only through the Land Bank of the Philippines (LBP), a government depository bank. Where possible, grantees7 are given LBP debit cards to withdraw their grants through ATM machines. In areas where ATM machines are less accessible to beneficiaries, off-site cash payments are scheduled by LBP. In an effort to continuously improve the delivery of Pantawid Pamilya grants, DSWD thru LBP is engaging other service providers as payment conduits: Globe GCash Remit (a mobile cash delivery service), local rural banks, local cooperatives, and so on. Program grants are subject to a deduction, suspension, termination, or delays based on rules known to beneficiaries. Non-compliance to program conditions automatically means deduction of a specific grant amount. If a household beneficiary fails to comply with the condition for an elementary school-aged child, PHP 300 is automatically deducted for every month the child misses to comply with the 85 percent required attendance. If any one of the applicable health conditions is not met, PHP 500 is automatically deducted for each month of non-compliance. Proven misuse of program grants due to pawning of the LBP card or use as collateral are grounds for suspension of the grants (on the second offense) or termination and delisting from the program (on the third offense). Beneficiaries can also be delisted from the program if they were proven to have deliberately provided false/fraudulent information at the time of the data collection for Listahanan or enrollment into the Pantawid Pamilya. Inadvertent errors in records or outdated household information – such as misspelled names of grantees shown in the LBP cards, transfer from one school to another by a child beneficiary – may also delay the receipt of the grant if such mistakes or changes are not immediately corrected or updated in the beneficiary household records of Pantawid Pamilya. Pantawid Pamilya remains an efficient means of delivering assistance to poor households. Based on the most recent budget data for the program, it only costs the government an additional PHP 11.20 to deliver PHP 100 of cash grant to program beneficiaries. This includes the cost of setting- up and maintaining the targeting system Listahanan that is used to select potential beneficiaries of the Pantawid Pamilya spread over the seven years.8 In addition to the targeting system, program administrative costs also include staff and personnel costs, monitoring and evaluation, bank fees, and cost of equipment.9 From 18 percent in 2009 when the full cost of targeting and information systems figured significantly in the program budget, the cost of program administration has settled at about 8-10 percent annually (Figure 3). 7 The Grantee is usually the mother of the family. 8 The cost-benefit ratio for Pantawid Pamilya was previously compared with another large scale program of the government in 2009, the rice subsidy program. The latter was found to cost the government PHP 68.40 to deliver a PHP 100 direct assistance to its beneficiaries (Fernandez and Velarde, 2012). 9 Calculation of program administrative costs is extremely challenging and as such these estimates should be taken as a first approximation. Several indirect costs should also be considered (e.g. staff from different agencies and local government units in areas where the program is implemented), as well as other investment costs (e.g. of the national targeting system set up and updating, to be spread out across time and different user programs). www.worldbank.org 5 “Sa Pantawid, Malapit nang Makatawid!” (With Pantawid, we are closer to getting out of poverty!) III. Program Performance: Results from the recent national household surveys This section presents the results of a benefit incidence analysis for the Philippine CCT program using recent national household survey data. Benefit incidence analysis (BIA) is a standard analytical tool to assess performance of cash transfer programs using national household surveys (see Box 1). It evaluates the targeting performance and progressivity of a program by looking at how beneficiaries and benefits (e.g., cash grants) are distributed between poor and non-poor households or across income or consumption groups. In addition, BIA assesses poverty impacts of a program by comparing standard poverty and inequality indicators (e.g. changes in poverty status, poverty gap, and income distribution) with and without a program, assuming all other components of a household’s income or spending patterns remain unchanged. It is important to keep this assumption in mind because it implies that results of a BIA do not account for possible changes in behavior due to the program intervention (e.g., potential reduction in labor income for families receiving a Program).10 Such behavioral changes are typically covered in specialized treatment/control impact evaluation studies. A previous BIA exercise for Pantawid Pamilya was done in 2012 using data for 2009 when the program was only at 700,000 household beneficiaries.11 Given how the program has grown since then, which has now figured at 2.8 percent of the national budget and six-folds in coverage, an update of this exercise is needed to see how the program is performing in an expanded context. The results of the CCT benefit incidence analysis presented below are based on the 2013 Annual Poverty Indicators Survey (APIS) of the Philippine Statistical Authority (PSA). Where needed, the analysis is complemented by available administrative data on the program budget and implementation or by data from other nationally representative household surveys such as PSA’s Family Income and Expenditure Survey (FIES) 2009 and 2012. Box 1. What is benefit incidence analysis? Benefit incidence analysis (BIA) is a standard approach for examining who and how much benefits from public expenditures on various programs, in particular in social protection, health and education. In other words, it is a method of computing the distribution of public expenditure across different groups of the population, such as poor and non-poor, men and women, children and elderly, and so on. BIA does this by combining the unit cost of providing the service with information on the use or receipt of the service. In some countries, BIAs for social assistance programs had been effective tools to justify reforms that eliminate ineffective programs and replace them with better-targeted ones. Early applications of BIA in government-wide reforms include those of education and water and sanitation systems in Colombia and health systems in Malaysia and Ghana in the 1990s. Based on empirical evidence from BIAs, Indonesia in 2005 initiated the dialogue to reform ineffective general subsidies such as petroleum and reallocated funds to health, education, and a new cash transfer program (Indrawati, 2005). Just like any other research methodology, BIA has its own advantages and limitations. Among the advantages of BIA is that it does not require specialized and usually expensive data collection. BIA piggy-backs on existing national household surveys for data needed to undertake the analysis, which at the minimum requires the following information: (i) indicator of household participation to the program, and (ii) amount of assistance received from the program. If these two pieces of information are available from national household surveys, BIA can be a very straightforward exercise, which is the second most appealing trait of BIA. By comparing certain indicators with and without the program – or post-transfer versus pre-transfer scenarios – BIA can estimate program coverage of various segments of the national population as well as key questions for policy makers such as the program’s impact on overall poverty and income distribution. Findings from a BIA thus apply to the entire country unlike in many highly-specialized surveys. Unlike rigorous impact evaluation that assesses a program’s impact on specific development outcomes (i.e., education completion, child stunting, and wasting) due to intended behavioral changes induced by a particular program, BIA does not account for behavioral changes. It assumes that, except the provision of cash assistance to beneficiaries, everything else remained unchanged in the living conditions of both beneficiaries and non-beneficiaries.* In the case of the Philippines, this is the second BIA study for the Pantawid Pamilya conducted by the World Bank. The first BIA was done in 2012**. DSWD conducted its own round of the Pantawid BIA in 2013. The latest available national household surveys that could be used for this second BIA are the Annual Poverty and Indicators Survey (APIS) 2013 and the linked datasets of Labor Force Survey (LFS) and Family Income and Expenditure Survey (FIES) in 2012 produced by the Philippine Statistical Authority. These household datasets collected information on: (i) participation of households in the Pantawid Pamilya; and (ii) the amount of cash grant they received from the program for the reference year. Both surveys captured a nationally representative sample of the CCT beneficiaries. The FIES 2012 captured an unweighted sample of 5,864 household beneficiaries of the Pantawid Pamilya or about 15% percent of the total survey sample, which represented about 2.7 million of the 3.0 million nationwide household beneficiaries of the program in 2012. The APIS 2013 captured 1,845 Pantawid Pamilya household beneficiaries in the survey or about 17% of the national survey sample, which represented 3.4 million of the 3.9 million household beneficiaries in 2013. For this BIA, results presented are mostly from the APIS 2013. Where needed, it will be complemented with data from the FIES 2012 or available administrative data. * For more information on benefit incidence analysis, see Demery (2000). ** Fernandez and Velarde (2012). 10 As discussed later on in Box 2, Pantawid Pamilya does not disincentivize work, with adults in beneficiary households working as much as their non- beneficiary counterparts. 11 Fernandez and Velarde (2012). 6 PHILIPPINE SOCIAL PROTECTION NOTE 4 This section is drawn from World Bank. 2013. Philippines Conditional Cash Transfer Program Impact Evaluation 2012. Washington, D.C.: World Bank. • Program Coverage In 2013, Pantawid Pamilya covered a significant fraction of the targeted population. As already explained above, in 2013, the CCT program targeted poor households with children 0-14. Among poor households with children 0-14, the coverage rate was 58%, while among the food poor households with children 0-14, the coverage rate was 65% (Figure 4). Figure 4. Program coverage in 2013 (% of the population within a referenced group) Figure 5. Profile of the under-covered eligible population in 2013 Source: WBG staff calculations based on APIS 2013 using World Bank Social Protection ADEPT software. Despite its rapid expansion, the program maintains good levels of targeting accuracy especially when compared to CCT programs of other countries. Of the program’s total beneficiaries, 65 percent are income poor and 35 percent non-poor. This is an excellent result by international standards. 35 percent of all beneficiaries were food poor (Figure 6). In terms of income quintiles, 53 percent of beneficiaries belonged to the poorest 20 percent of the population (Quintile 1 or Q1) and the next 29 percent belonged to the second lowest income group (Q2). In other words, 82 percent were from the bottom 40 percent of the population, which in the Philippines comprises poor households as well as households vulnerable to poverty, e.g. likely to experience episodes of poverty during the year. These results are somewhat lower relative to those in the first BIA (based on FIES 2009), which found that 76 percent of all beneficiaries were poor, and 75 percent and 17 percent of beneficiaries were from Q1 and Q2, respectively. It is important to note that program coverage in 2009 was only about one-sixth of that in 2013. The targeting accuracy was expected to decline somewhat over time, given that the data collection for the first Listahanan took place in 2009 and 2010 when the household poverty status was more accurate than 3 years after the survey data was collected (2013). Poverty is a dynamic phenomenon and many households in the Philippines move in and out of poverty during the year. The issue of essentially static massive data bases and dynamic changes in household income is faced by other countries with similar programs as well. That is why the Philippines Government has stipulated that Listahanan should be updated every four years – to keep the data accurate, on the one hand, but also to allow for some stability in the program participation, which is an important concern for poor households. It should be kept in mind, that in the world-wide context, Pantawid Pamilya’s targeting performs better than that of many other CCT programs. Among countries with large CCT coverage, only Brazil’s Bolsa Familia has a more accurate targeting performance than the Philippines’ Pantawid Pamilya (Figure 6). In addition, the targeting performance of Pantawid Pamilya should also be compared with the targeting performance of various programs before Listahanan was introduced. Figure 6. Distribution of Beneficiaries (% of beneficiaries in income group) % Beneficiaries by income poverty status, 2013 % Beneficiaries belonging to Bottom 20% Source: Own calculations based on FIES 2009 and APIS 2013 using World Bank Social Protection ADEPT software. World Bank ASPIRE database for other countries. www.worldbank.org 7 “Sa Pantawid, Malapit nang Makatawid!” (With Pantawid, we are closer to getting out of poverty!) • Adequacy of Benefits On average, Pantawid Pamilya beneficiaries received a monthly grant of PHP 117 ($2.72) per person in 2013. This meant that for an average beneficiary household with 6 members, a Pantawid Pamilya household received PHP 703.50 ($16.36) per month or about PHP 8,442 ($196.33) for the full year of 2013. This corresponds to only 7 percent of beneficiary households’ income in 2013 (11.6 percent for households in Q1), and 56 percent of the maximum program entitlement of PHP 15,000 per year. This may be for various reasons. One is that, from the national surveys, we observe that Pantawid Pamilya households on average only have 2 eligible children between 3-14 years old. This means that many of them can only receive a maximum of PHP 12,000 per year – PHP 6,000 for health (PHP 500 for 12 months) and PHP 6,000 for education (PHP 300 for 10 months for an average of 2 eligible children). Another reason may be delays in payment, which could arise from unreported changes in household information such as transfer of residence or school of eligible children. Throughout 2014, some 6.7 million transactions related to beneficiary information updates were received by program implementers. A great majority (80 percent) of these transactions were requests to update schools where children’s attendance should be monitored. Pantawid Pamilya is highly progressive, with the greatest share of benefits going to the poorest households. The distribution of program benefits mirrors how well the Pantawid Pamilya was able to reach poor households in the bottom 20 or 40 percent. As in 2009, the biggest share of program benefits went to the poorest households (53.7 percent to Q1) and the share drops to nil in higher income groups (Figure 7). The Lorenz curve of the Pantawid Pamilya in Figure 7 is telling that the program is highly progressive, that is, the poorest receiving a higher share of program benefits than their actual share in the national income distribution. Beneficiaries spend their Pantawid Pamilya cash grants on basic needs. In the 2012 Family Income and Expenditure Survey of the Philippine Statistical Authority, beneficiary households were asked how they spent the grants they received from the Pantawid Pamilya program. Their self-reported responses indicated that half of the grants were spent on food and a quarter went to school-related expenses (Figure 8). Virtually none of the grants were used for recreation or adult goods like alcohol. Figure 7. Progressivity of the Pantawid Figure 8. How Do Pantawid Beneficiaries Spend their Pamilya in 2013 Cash Grants? Medical Housing 7% 1% Debt Invest 0% Education Clothing 2% Alcohol 0% 25% 11% Savings Recreation 0% 2% Other 3% Other use 3% Food 49% Source: Own calculations based on APIS 2013 using World Bank Social Protection ADEPT software, and FIES 2012 Over the years, the generosity of the Pantawid Pamilya program has declined and lagged in comparison to other CCTs of comparable size. The program was designed so that program generosity is on par with that of Brazil, Mexico, and Colombia. Based on simulations using available 2006 national survey data at program design stage, the benefit level was at 23 percent of potential households’ income. Over the years, the Government has prioritized covering all poor households with children over updating benefits to maintain a reasonable Program budget. However, because the nominal value of the grant has not been adjusted since it was piloted in 2007, at an annual inflation rate of 4.0 percent from 2007 to 2015, the maximum entitlement of PHP 15,000 (USD 348.84) per household in 2007 is valued only at about PHP 10,600 (USD 247) in 2015, equivalent to a cumulative reduction in real value of 42 percent. In comparison with other countries with large CCTs, at 11.6 percent the Pantawid Pamilya is now among the least generous, as measured by the share of benefits in the income of the poorest 20 percent (Figure 9).12 12 Based on grants reported as received by beneficiaries in the household surveys. Source: World Bank ASPIRE database, 2014. 8 PHILIPPINE SOCIAL PROTECTION NOTE • Program Impacts Pantawid Pamilya has been subject to rigorous impact evaluations that show that the program is doing well in extending aid to poor families and influencing behavior change. 13 Among the key findings of the series of impact evaluation studies are that the program successfully promoted safer facility-based birth deliveries, improved children’s access to health care services, improved usage of subsidized health care benefits, encouraged children’s enrolment and attendance in school, and did not show disincentive effects on adults’ labor participation. Many of these will have effects on the long-term welfare of beneficiary households. The impact evaluation studies have been important to introduce evidence informed changes in the Program, such as the extension to cover children in secondary education. Box 2 summarizes the main results from the impact evaluation studies. Consistently, Pantawid Pamilya is also showing improvements in the overall welfare of beneficiary households in the short-term. One of the main objectives of the Pantawid Pamilya is to provide short-term income support to poor beneficiaries. On average, in APIS 2013, poor beneficiary households’ income was 35 percent short of meeting the minimum income level needed to afford their basic needs. The program is helping beneficiaries decrease this gap. Assuming there are no other changes in the earning or spending patterns of beneficiary households, the annual cash grant of PHP 1,407 per person reduces the income gap of Pantawid Pamilya beneficiaries by more than a quarter of its absolute level, or down by 7.4 percentage points (Figure 10). In the case of poor households closer to the poverty threshold, the cash transfer has enabled them to temporarily cross the line of poverty, enabling them to fully close the gap and support the basic needs of their family. Among beneficiaries of the Pantawid Pamilya, this resulted in the reduction of total poverty from 64.5 percent to 58 percent, or by 6.5 percentage points. Among the food poor beneficiaries the Pantawid Pamilya has decreased food poverty rate from 35.2 percent before the transfer to 28.5 percent after transfer, or by 6.7 percentage points (Figure 9). This implies that 1.3 million Filipinos were lifted out of poverty through Pantawid Pamilya. The previous BIA based on 2009 FIES, similarly estimated that the Pantawid Pamilya effect on poverty reduction among the program recipient was 6.2 percentage points. The estimated effect on food poverty reduction among beneficiaries was 13.3 percentage points, coming from a high rate of 61.4 percent.14 At the national level, Pantawid Pamilya had a positive effect on the poverty gap and the incidence of food and overall poverty. The program reduced poverty gap from 9.1 percent to 8.2 percent. This translates to a reduction of PHP 61 in the nominal poverty gap for every PHP 100 spent on the Pantawid Pamilya. The program is also estimated to have reduced both total poverty and food poverty by 1.4 percentage points each: from an estimated pre-Pantawid rates of 26.4 percent for total poverty and 12.5 percent for food poverty (Figure 10). These estimates are in between those found in two other large CCT programs in the world – Mexico’s Oportunidades was estimated to have reduced total poverty by 1.8 percentage points while Brazil’s Bolsa Familia was estimated at 0.5 percentage point.15 Moreover, the Pantawid Pamilya contributed to reducing national income inequality as measured by Gini coefficient by 0.5 point: to 48.7 in 2013 with Pantawid Pamilya from 49.2 if without the program. Figure 9. Generosity of the Pantawid Pamilya Figure 10. Poverty rates with and without the (% of per capita income of households in Bottom 20%) Pantawid Pamilya program With Pantawid Pamilya Reduction 6.5 7.4 6.7 58.0 1.4 27.4 28.5 1.4 0.9 25.0 8.2 11.1 Income gap (% to Food Poverty rate Poverty rate Poverty gap (% to Food Poverty rate Poverty rate poverty line) (% of benes) (% of benes) poverty line) (% of popn) (% of popn) Source: Own calculations based on APIS 2013 using World Bank Social Source: Own calculations based on APIS 2013 using World Bank Social Protection ADEPT software. World Bank ASPIRE database for other countries. Protection ADEPT software. 13 For details, see Orbeta, et al (2014). 14 Velarde and Fernandez (2011). 15 Fiszbein and Schady (2009). www.worldbank.org 9 “Sa Pantawid, Malapit nang Makatawid!” (With Pantawid, we are closer to getting out of poverty!) Box 2. Impact evaluation findings for the Pantawid Pamilya* The Pantawid Pamilya has been subject to a rigorous impact evaluation to assess results in terms of the increase in human capital accumu- lation among beneficiaries. The key findings of the impact evaluation reports are summarized below. An impact evaluation study of the Pantawid Pamilya program was performed using a recommended randomized control trial (RCT) design during the first implementation phase (known as Set 1 which took place between June 2008 and April 2009) of the program, since the program’s scale-up plan was not yet in place. Among the most salient results found were the following: • Higher rates of school enrollment among children 3-11 years of age in the beneficiary households (by 10 percentage points for 3-5 year olds and by 4.5 percentage points for 6-11 year olds), compared to poor households who did not receive the program. • However, the program had no significant impact on increasing enrollment among older children aged 12-17 years old, not a sur- prise since at the time the program was not explicitly designed to improve schooling of children above age 14. • The program has helped improve the long-term nutritional status of younger children (6-36 months old) by reducing stunting by 10 percent compared to barangays that did not receive the program, • The program has also encouraged poor women to use maternal and child health services such as antenatal care, postnatal care, regular growth monitoring, and receipt of Vitamin A and deworming pills. In addition, it has helped increase healthcare-seeking behaviors among beneficiaries when their children become ill. • The program changed the spending patterns of poor households, with beneficiary households spending more on health and education than poor households who had not received the program. • The study also found that beneficiary households spent less on adult goods such as alcohol and that the program may have con- tributed to increased savings among beneficiary households. Since the program experienced a faster expansion, the second round of impact evaluation study performed in 2013 could not use a RCT design due to the absence of a randomly assigned counterfactual group. Instead, it employed a regression discontinuity design (RDD) which compares households that are very close to a pre-identified cutoff (e.g., those just below and those just above the poverty threshold). It was based on a sample of households from 30 municipalities which were introduced into the program between 2008 and 2011, and as- sessed in 2013. The key findings from this more recent study were the following: • The Program successfully keeps older children in school. Gross enrollment rate for high school children (12-15 years old) is higher for Pantawid children living near the poverty threshold. However, no impact is found in enrollment of younger children. • Pantawid children (10-14 years old) work seven days less in a month compared to non-Pantawid children. • The program encourages women to try modern family planning methods at least once. It also promotes facility-based deliveries and access to professional postnatal care for 7 in 10 live births among Pantawid mothers, compared to 5.5 in 10 births among non-beneficiary mothers. • It improves children’s access to some key health care services (Vitamin A, iron supplementation, deworming pills), as well as regular weight monitoring in health centers. However, no discernible impact was found among children below six years old on wasting, underweight, and stunting. • The Program stimulates households’ investments in education. Results show that Pantawid households spent PHP 206 more per school-aged child per year at the threshold compared to non-beneficiary households. • Pantawid Pamilya does not disincentives work. Findings reveal that adults in beneficiary households work as much as their non-beneficiary counterparts. Also, Pantawid households do not spend more on vice goods such as gambling, tobacco and alco- hol. • The Program seems to have improved parents’ perception of their situation and of their children’s future. Pantawid parents have higher hopes of their children finishing college compared to non-Pantawid parents. * For more information on the impact evaluation on the results, refer to DSWD (2014) and DSWD and World Bank (2014). 10 PHILIPPINE SOCIAL PROTECTION NOTE IV. Summary and Recommendations After 7 years of implementation, Pantawid Pamilya remains focused on achieving its long-term goal of assisting poor households invest in the education and health of their family. The program has grown to become a flagship social protection intervention in the Philippines, facing continuous expansion and constantly evolving to form the base of an effective safety net system. Having a legitimate and functioning targeting system has allowed the program to maintain its targeting accuracy and progressivity. To date, the Pantawid Pamilya is considered as one of the best targeted CCTs of its size in the world, with more than half of beneficiaries and program benefits accruing to the bottom 20 percent and more than 80 percent of beneficiaries and program benefits accumulating to the bottom 40. The program’s impact evaluation studies show that the assistance is resulting in improved education and health outcomes among beneficiaries. Independent national surveys show that it is also achieving its short-term objective of poverty mitigation by helping beneficiaries move closer to the poverty line. The program is estimated to reduce the income gap of beneficiaries by more than a quarter of its nominal level, equivalent to a reduction of 7.4 percentage points. For households close to the poverty line, the cash grant has enabled them to afford their basic needs and temporarily cross the line of poverty or move out of extreme poverty (food poverty). Total poverty and food poverty among beneficiaries is estimated to have been reduced by up to 6.7 percentage points. Due to its strong poverty focus, the Pantawid Pamilya also helped reduce short-term poverty and food poverty at the national scale by up to 1.4 percentage points. This level of impact is on par with the two other large CCT programs in the world – Mexico’s Oportunidades, which was estimated to have reduced total poverty by 1.8 percentage points and Brazil’s Bolsa Familia, which was lower at 0.5 percentage points. By augmenting the poor’s income, the Pantawid Pamilya also contributed to reducing national income inequality by 0.5 point. To keep the program’s good performance, regular updating of beneficiary database, periodic adjustment of benefits to keep up with inflation, and minimizing payment delays are key. Just like any other program, the Pantawid Pamilya is not error-proof. Regular checks are needed on the program’s targeting to minimize errors of capturing ineligible households and maximizing coverage of eligible households. These periodic checks can be done using available national household surveys. But a more important aspect is to consider ways for a more frequent updating of welfare indicators (ideally, once a year or once in two years) instead of a 4 to 5 years cycle of massive data collection. The program’s benefit level, as shown in this note, also needs to be regularly checked. While the program is large on the aggregate, it provides minimal income support to poor families and its value can easily be eroded by increasing prices. Ensuring that the program provides adequate support to poor families is important in achieving the expected impacts of the program, especially as it is considered as the Philippines’ flagship social assistance program. These adjustments in the program’s coverage and benefit level come with huge fiscal costs. So it is important to keep these adjustments at a sustainable level. At the implementation level, continuing proactive monitoring to minimize payment delays also help ensure that beneficiaries enjoy their maximum benefits from the program.  References Demery, Lionel (2000). “Benefit Incidence: A Practitioner’s Guide.” Poverty and Social Development Group, The World Bank (unpublished mimeo). DSWD (2014). “Keeping Children Healthy and in School: Evaluating the Pantawid Pamilya Using Regression Discontinuity Design.” DSWD and World Bank (2014). “Philippines Conditional Cash Transfer Program Impact Evaluation 2012.” Report Number 75533-PH. Fernandez, Luisa (2012). “Design and Implementation Features of the National Household Targeting System in the Philippines.” World Bank Philippine Social Protection Policy Note No. 5. Manila, Philippines. Fernandez, Luisa and Rashiel Velarde (2012). ”Who Benefits from Social Assistance in the Philippines: Evidence from the Latest National Household Surveys.” World Bank Philippine Social Protection Policy Note No. 4. Manila, Philippines. Fizsbein, Ariel and Norbert Schady (2009). “Conditional Cash Transfers: Reducing Present and Future Poverty”. World Bank: Washington, DC. Manasan, Rosario and Janet Cuenca (2007). “Who Benefits from the Food-for-School Program and Tindahan Natin Program: Lessons in Targeting.” Philippine Institute for Development Studies Discussion Paper Series No. 2007-10. Makati, Philippines. Orbeta, Aniceto, Arnelyn May Abdon, Michael del Mundo, Melba Tutor, Marian Theresia Valera, and Damaris yarcia (2014). “Keeping children healthy and in school: Evaluating the Pantawid Pamilya Using Regression Discontinuity Design. Second Wave Impact Evaluation Results.” Downloadable http://www.dswd.gov.ph/download/pantawid_pamilya_impact_evaluation/Pantawid%20Pamilya%20Impact%20Evaluation%202014%20 Report%20Final.pdf. Velarde, Rashiel and Luisa Fernandez (2011). “Welfare and Distributional Impacts of the Pantawid Pamilyang Pilipino Program.” World Bank Philippine Social Protection Policy Note No. 3. Manila, Philippines. www.worldbank.org 11