Public Disclosure Authorized 45618 Public Disclosure Authorized Conditional Gash Transfers in Indonesia: Program ILeEuarga Harapan and PNPM-Generasi Baseline Survey Report June 2008 Public Disclosure Authorized Public Disclosure Authorized ACKNOWLEDGEMENTS This report has been prepared by the World Bank Conditional Cash Transfer (CCT) Team. Susan Wong led the CCT Baseline Survey work as World Bank Task Team Leader, together with a core survey team consisting of: Junko Onishi, Gregorius Pattinasarany, Yulia Herawati, Scott Guggenheim, Menno Pradhan, Vivi Alatas, Kathy Macpherson, h a d Saleh (World Bank Indonesia), Ben Ollten and Jusuf Neggers (Harvard University, Boston). The team preparing the baseline survey report includes: Robert Sparrow (Instit~~te Social Studies, The Hague) as lead coordinator, Jossy P, of Moeis and Arie Damayanti (LPEM, University of Indonesia, Jakarta), and Yulia Herawati (World Balk Indonesia). The team was assisted by Hendro Tuhiman (World Bank, Indonesia) and Dhaniel Ilyas (LPEM, University of Indonesia, Jakarta). The University of Gadjah Madah (UGM), Center for Public Policy Studies, implemented the field survey. The team would like to express its appreciation to Pak Sukamdi, the Director, and his UGM team of some SO0 enumerators, data entry operators, and management staff for their excellent work. Financial support for the overall CCT program and this survey comes from the Government of Indonesia, the Decentralization Support Facility, the Royal Embassy of the Netherlands, and the World Bank. Significant support and contributions to this report were made by: Budi Hidayat, Vic Paqueo, and William Wallace (World Bank Indonesia). Peer Reviewers were: Deon Filmer, Margaret Grosh, and Emmanuel Skoufias (World Bank, DC). The Government of Indonesia through the Ministry of Planning (Bappenas), the Coordinating Ministry for Economy and Social Welfare (Menkokesra), and the Ministry of Home Affairs has provided tremendous support to the program and baseline survey over the past year. Special thanks to: Sujana Royat (Menltokesra), Bambang Widianto, Prasetijono Widjojo, Endah Muminingtyas, Pungky Sumadi, Vivi Yulaswati, and Woro S. Sulistyaningrum (Bappenas), Ayip Muflich, Elco Sri Haryanto, Bito Wikantosa, and Prabawa Eka Soesanta (Ministry of Home Affairs) for their generous support of the CCT program and more specifically, the baseline survey work. TABLE OF CONTENTS 1. INTRODUCTION...................................................................... 4 2 . PROGRAM DESIGN .................................................................. 6 3. BASELINE SURVEY METHODOLOGY ........................................ 14 . 4 EVALUATION DESIGN AND BASELINE ANALYSIS ..................... 21 5 . RESULTS OF THE BASELINE SURVEY ....................................... 26 . 6 CONCLUSIONS AND RECOMMENDATIONS ................................ 47 REFERENCES .............................................................................. 50 LIST OF TABLES ........................................................................ 51 LIST OF FIGURES ........................................................................ 53 1. INTRODUCTION In the second half of 2007, a large pilot project of two Conditional Cash Transfer (CCT) programs was initiated by the Goverrnnent of Indonesia. This pilot project, fitting in an overall national poverty reduction strategy, is motivated by concerns that Indonesia is still lagging behind in a number of health and education outcomes, and that poverty remains a reality for a large part of the Indonesia population. The main aim of the pilot is targeting the poor and simultaneously addressing short term poverty alleviation through cash transfers and improving access to health and education services, with the aim of long term poverty alleviation through increasing investment in human capital. Previous experience in other countries has found conditional cash transfers to be a successful policy instrument for achieving such objectives. The Indonesian CCT pilot project follows this example, but also introduces a novelty to the CCT approach. The Indonesian pilot consists of two independent programs. In addition to a household CCT program (Hopeful Families Program or Program Keluarga Harapan - PKH), comparable to those conducted in other countries, a community CCT program (PNPM Genevasi) targets block grants to communities, allowing these communities to formulate and i~nplementtheir own activities in order to meet the program targets. The pilot has been launched in seven provinces in Indonesia, targeting 500,000 households for the PIU3 transfers and 1,625 villages for participation in PNPM Generasi. An integral part of the program is the evaluation design. Participation in the CCT programs has been randomized over sub-districts as to be able to directly measure impact, following the exa~npleof CCT programs elsewhere in the world (most notable Latin America). In theory, randomization over sizable geographical units such as sub-districts solves the problenl of selection bias in the impact evaluation. The first stage of the evaluation process is the baseline analysis. To this end, a baseline survey was conducted in the treatment and control sub-districts, several months prior to implementatioll of the programs. The survey was conducted in 660 sub-districts, covering 26,326 households, 658 health centers, 1,861 junior high schools and 2,564 midwives, hom June to August 2007. This paper presents an analysis of the baseline survey, focusing in particular on the result of randomization in eliminating any systematic differences between treatment and coiltrol groups. Overall, we found the treatment and control groups to be balanced. Some minor discrepancies appear, in particular in child height, gross school participation rates, school travel costs and domestic work activities. However, these disparities are rare and do not show a clear pattern, indicating that there is no systematic bias in education or health status, hence colltrolling for pre-intervention differences should be sufficient to retrieve unbiased impact estimates. We also find no difference is health and education supply at village level, although we do find some significant differences between schools and health care providers, \vhich warrant consideration for future impact evaluation. The paper is structured as follows. The next section sets the context for the study, discussing the motivation, background and design of the two CCT programs. Section 3 provides details on the survey design, while the methodology for this baseline analysis is explained in section 4. The empirical results are presented in section 5. The results are presented separately for the CCT target indicators, key health and education outcomes, individual and household characteristics, village characteristics and social service delivery. Finally, section 6 concludes with some recommendations for the follow up surveys. 2. PROGRAM DESIGN 2.1 Motivation and background Motivatiorz Addressing lags in ltey human development indicators and poverty alleviation The key motivation driving the CCT pilot programs is to investigate whether these interventions are an effective means to bring about long term increase in human development indicators in combination with short term poverty alleviation, and hence, reducing scope for dynamic poverty traps. This is relevant for Indonesia sillce some indictors of the Millenium Development Goals in Indonesia have lagged behind, despite significant economic progress since the economic crisis of the late 1990s. For example, some areas of concern are: * Maternal health, which does not keep up with comparable countries in the Southeast Asian region. In Indonesia 72 percent of births are attended by skilled attendants, while 86 percent of births are attended in the region (UNICEF 2006). Indonesia's maternal mortality rate (307 deaths in 100,000 births) is three times that of Vietnam and six times that of China and Malaysia (World Bank, 2006). * Service coverage and child health outcome indicators, which are lower than countries in the region. The infant mortality rate in Indonesia (30 per 1,000 live births) is almost double that of Vietnam and three times that of Malaysia (World Bank, 2006). Malllut~itio~l are high and have risen in recent years: a quarter rates of children below the age of five are malnourished in Indonesia, with malnutrition rates stagnating in recent years despite reductions in poverty. The prevalence of underweight children under five in Indonesia between 1996 and 2004 is estimated to be 28 percent, compared to the regional rate of 15 percent during the same period (UNICEF, 2006). Immunization coverage is considerably lower than the regional average for 2004: in Indonesia 82 percent of one-year old children were immunized against BCG, 70 percent against polio, and 70 percent against measles, while in the region 92,87 and 83 percent were, respectively (UNICEF, 2006). Weak education outcomes. While there has been much progress in primary school enrollment, currently at 94 percent, transition rates from primary to secondary school are low, with net junior secondary school enrollment of 65 percent. On average, Indonesia has the capacity to provide junior secondary education to only 84 percent of the potential students in the 13 to 15 age group (World Bank, 2006). For all these indicators, there is a strong correlation with poverty, suggesting that a program that targets the poor, and provides the means to access basic health and education services could be an important component of a poverty strategy for Indonesia. In this context, conditional cash transfer can play a role, through income assistance for households and coininunities, conditioned on investment in health and education. Substitute for Unconditional Cash Transfers as social protection for the poor In August 2005, the Govermnent of Indonesia implemented an Unconditional Cash Transfer (UCT) program to help mitigate the negative effects of fuel price increase for poor households. Soille 19.2 million households received approximately US$120 in four installments over the course of one year, ending in September 2006. The total annual budget for the prograin is estimated to be close to US$2.4 billion. Later assessment of the program revealed that targeting was disappointing in that a significant portion of the transfers were allocated to the non-poor. In addition, the program was also criticized for creating dependency to govemment support (World Bank, 2007a). This led to calls for a different cash transfer program designed to effectively reach the poor and simultaneously address the many dimensions of human capital, supporting the idea for testing the implementation of a Conditiorlal Cash Transfer program in Indonesia. CCT programs have become a dominant social protection strategy in Latin America and the Caribbean (LAC) over the past decade. The rapid replication of CCT programs owes largely to the successes of the two largest CCT programs in the late 1990s: the Bolsa ' Fanzilia Program in Brazil and the Progressa Program in ~ e x i c o with annual budgets of $2.1 billion and $2.5 billion respectively (Handa and Davis, 2006). ' In March 2002, P~.ogi.essa changed name to Oporiunidades. Several features nlalce it distinct to earlier poverty alleviation programs. The traditional model of CCT prograins provides cash to households with the dual objectives of short- term poverty alleviation and investments in long-term human capital. CCT programs generally have two components: an education component and a health and nutrition component and so addressing various dimensions of human capital simultaneously. Unlike the unconditional cash transfer, with CCTs the benefit (a cash transfer) is made to a poor household conditional that the household performs certain obligations related to health and/or education aspects. For example, Progresa in Mexico links monetary educational grants to school attendance of children so that if a child misses more than 15 percent of school days in a month for unjustified reason, the family will not receive the grant that month. To receive monetary support for improved nutrition, families must complete a schedule of visits to the health care facilities. A distinct feature is the mechanism for delivering the benefits. Recognizing the potential of mothers to effectively and efficiently use resources in a manner that responds to the family's immediate needs, Progressa gives benefits exclusively to mothers (e.g. Skoufias and Parker, 2001). However, CCTs cannot operate in areas with supply-side constraints. In areas where distances to scl~ools and health facilities are large, transportation infrastn~ctureis deficient, and quality of services are poor, CCTs cannot effectively address the bottleneclcs to improve service usage. In such areas, it is not the lack of willingness on the part of the users but the supply-side deficits that limit poor households and communities to use basic health and education services. In such a setting, providing grants to communities that can be used to address specific supply-side problems within the community may work more effectively than transfers to households in order to increase health and education outcomes. Under such a community grant scheme, resources are provided to cominunities who will decide how best to use the block grant to reach several education and health targets. Indonesia will be the first country to test this type of CCT (World Bank, 2007a). TIze Z~zdorzesiarzCCTprogiunr In 2007, the Goveiu~lleiit of Indonesia launched a pilot test of a household and community CCT prograni. Household CCT: Program ICeluarga Harapan ( P M ) The Household-CCT program known as Prograrn Keluarga Harapan (PKH) provides cash transfer to households, similar to the traditional CCT program in Latin American and Caribbean countries. Eligible households must be classified as poor (Rumah Taizgga Sangnt Miskin) with children aged 6-15 years or less than 18 years but who have not completed basic education, or children aged 0-6 years or pregnantllactating mother. Cash transfers are made to households under the condition that certain health and education related obligations are met. The Ministry of Social Affairs is the implementing agency and the Post Office carries out the transfer of funds. The PKH household CCT program was implemented in 48 districts in seven provinces, targeting 500,000 household^.^ Selection of provinces is such that various types of areas are represented, e.g. higldmediurn/low poverty rate, urban/rural areas, coastal- areaslislands, accessibleldifficult-to-access areas. Beneficiary households are determined by combining geogaphic and household level targeting. Locations are first selected to meet several criteria: high incidence of poverty, high incidence of malnutrition, low transition rate of education from primary to secondary school, adequate supply of health and education facilities, and approval from the local government to participate in the CCT pilot project. The selection process of eligible households in sub-districts where PKH is implemented consists of two steps. An initial roster of beneficiaries was created using the Unconditional Cash Transfer beneficiaries list and then applying a proxy means test. It was further verified that only households in poverty would be selected for the program. To minimize any exclusion error, households not on the UCT list but deemed severely The provinces of West Java, East Java, West Sumatra, North Sulawesi, Gorontalo, East Nusatenggara, and DKI Jakalta. 9 Finally, households identified in the first step were verified poor were also ~onsidered.~ in terms of eligibility to the program, i.e. whether they have children aged 6-15 years or less than 18 years but who have not completed basic education, children aged 0-6 years or pregnantllactating mothers. Community CCT: PNPM Generasi Sehat dan Cerdas The Community CCT scheme PNPM Generasi Sehat dan Cerdas (PNPM Generasi) builds on the project infrastructure and capacities developed through the experiences of the Kecamatan Development Project (KDP). PNPM Generasi is implemented as part of the Government's new flagship poverty program, the National Community Empowerment Program or Program Nasional Penzberdayaan Masyaralcat (PNPM). The Ministry of Home Affairs implements both KDP and PNPM Generasi (World Bank, 2007a). PNPM Generasi differs from the household CCT in that the cash transfers are allocated to communities, not households. A condition for participation in the program was that communities have to commit to improve health and education conditions. While PNPM already allows for education and health activities, PNPM Gerzerasi places a stronger emphasis on such activities, emphasizing investments in certain lagging health and education outcomes. Communities will raise proposals to fund certain activities or investment (World Bank, 2007b). Examples of activities or investment include but are not limited to: . Paying transportation costs for midwives and nurses to provide outreach services Improveposyrnzrlt~(village health groups) organization and management to ensure that immunization, vitamin A and weighing services are efficiently carried out . monitoring Procuring weighing scales and height measures for growtl~ . Building infrastructure for health posts . Contracting private providers or NGOs to provide services in villages Contracting midwives or nurses to provide services in villages tlle The Central Statistics Office const~xcted initial roster of beneficiaries based on a health and education basic service survey (Szrrvej,Pelayaiiaii Dasal. Kesehataii da11 Penrliflikai~). Arranging health education in tenns of what preventative services people must receive Supporting community mobilization and social pressure towards current non- users Small access infrastructure such as roads and bridges which lead to education and health service Assisting with transportation costs and educatio~l materials for primary and junior secondary schooling. takes 12 - 14 months to complete. The cycle consists of four A cycle of PNPM Ge~zerasi main stages: socialization, village planning, village implementation and performance measurement. The village implementation stage takes approximately nine months. Communities are expected to use the block grants through a facilitated participatory planning process including social mapping, problem identification focus groups, hamlet level discussions, village level discussions, and sub-district level discussions with providers. Communities are also allowed to select their priority conditionalities (the same list of conditionalities as the PKH and will be described below). Although the block grants can be used for anything villagers collectively agree upon, the use of funds must be designed to improve at least one of the conditionalities. Service use and coverage of these conditionalities will be closely monitored and recorded (World Bank, 2007b). In 2007, 1625 villages are participating in PNPM Generasi. They are located in 129 sub- districts in 20 districts in five provinces.4 Target beneficiaries of the community-CCT are largely rural collllnuilities in areas with relatively inadequate supply of health and education service. Priority was put on areas that have been exposed to KDP for at least two years and thus have some experience with village level planning. 4 and The provinces of West Java, East Java, North Sulawesi, Gorol~talo East Nusa Tenggara. 2.2 Transfers and co~iditionalities .Uoz~sel~oId CCT The household CCT progranl aims to improve health and education outcomes, as well as increase consu~nption the poor by transferring cash directly to mothers on a quarterly of basis through the postal office. Internationally, household CCT programs are designed to provide cash amounts of approximately 15 to 20 percent of consumption of poor households. In the case of Indonesia, per household transfer amount depends on the number children and their ages. The support scenario and amount of transfers are presented in Table 1. If a mother is pregnant andlor has children aged 0-6 years, she will receive Rp 1,000,000 per year or Rp 250,000 per quarter, regardless of the number of under-five children. If a mother has two primary-school aged children (6-12 years) and one secondary-school aged child (13-15 years) and these children are attending school, she will receive Rp 1,800,000 per year or Rp 450,000 per quarter. A mother with children 0-6 years and three primary school-aged children will receive Rp 2,200,000 per year. A mother will continue receiving cash transfers in quarterly tranches so long as she meets the health and education conditions set by the project and goes to the post office to pick up the funds. If the mother fails to comply with the conditionalities, subsequent transfer payment is reduced. Cash transfers will be halted after a series of warnings. The conditionalities set by the project are as follows (World Bank, 2007b): Target households with children aged 0-6 years andlor pregnant or lactating mothers have to follow the protocols of preventive health care at public health facilities set by the Ministry of Health (MOH). These include5: For pregnant or lactating mothers . Four antenatal care visits and taking iron tablet during pregnancy Birth assisted by a trained professional . Two postnatal care visits for lactating mothers 2007. Government of It~donesia, For children 0-6 years: immunization and Vitamin A capsules twice a year Complete cl~ildl~ood . Monthly growth monitoring for infant 0-1 1 months and quarterly for children 1-6 years For target households with children aged 6-15 years (or children aged 15-18 years but not yet completed prinlary and secondary school), the education conditions are: Enrollment and attendance of a minimum of 85 percent of school-days for primary school children. . Enroll~llelltand attendance of a minimum of 85 percent of school-days for junior secondary school children. Poor household with children aged 15 -18 years who have not completed 9 years basic education can be eligible if the children are enrolled in an education program to complete 9 years equivalent. Cortzazrrrrity CCT The size of the bloclc grants provided to Community CCT sub-districts are pre- determined by the population size of the sub-district and poverty level. The average amount for the 2007 program is USD 8,400 (equivalent to Rp. 76,440,000 using exchange rate 1 USD = Rp. 9,100) per village. All participating villages also receive technical assistance in the form of facilitators and training. Two variants of Community-CCT are piloted. Half of the Community CCT groups will receive the perfo~l~lance-based incentives variant in year two, while the other half will receive the non-perfornlance based, non-incentives variant. The implementation of the two variants allows for assessment of the effectiveness of financial incentive and conditioiling of fiinds. The first-variant of Colnmunity CCT program makes the second years' block grant conditional on first year's performance. The conditionalities will be imposed through ' Source: World Bank, 2007a financial incentives to the communities. In the second year, villages will receive a bonus amount, on top of a fixed base amount. The size of the bonus depend on per-village performance of the use and coverage of services identified by the conditionalities. The budget for the village bonus grants is fixed at the sub-district level, hence creating inter- village competition for the reward grants. Health service indicators are collected through village level record keeping while school attendance is collected from regular classroom attendance records. In the second-variant Community CCT, the second years' block grant is independent of first years' performance. 3. BASELINE SURVEY METHODOLOGY 3.1 Survey desigli ' A series of three-wave evaluation surveys will be used to evaluate the impact on health and education indicators, consumption, and targeting of the two CCT schemes. The first of the three waves is the baseline CCT evaluation survey, carried out in 2007, prior to the implementation of the CCTS.' Two follow-up surveys are planned for 2008 and 2009, and if the program continues, additional impact surveys will be added for future years. The World Bank provides technical assistance to the Government for the CCT program, including the design and implementation of the baseline survey. The follow-up surveys will revisit the respondents of the baseline survey, creating a longitudinal panel. The advantage of this approach is that allows us to go beyond estimating average treatment effects, but also behavioral responses of individuals and households. The draw back is that panel data may suffer from attiition. If attrition is not random (i.e. attrition differs systematically for treated and controls, or poor and non-poor) then this could introduce a new source of bias to the impact analysis. The overall objective of the tlx-ee waves of evaluation surveys is to evaluate the impact and effectiveness of the two Conditional Cash Transfers, the Household CCT and the ' Source: World Bank, 2007b. The baseline survey was fielded between June and August 2007 by Gadjah Mada University Community CCT. The objective of the CCT baseline survey is to provide reference indicators to evaluate the impact of the two CCT approaches, specifically to collect baseline data on: Coverage data of targeted and non-targeted basic health services School enrolment and attendance Provision of health and education services Household coilsumption Household and community characteristics in target project locations The baseline survey col~ducts interviews with households, mothers, heads of villages or wards, health and education providers, and midwives, in order to collect information on the health and education status of children and their mothers, their use and experiences with health and education services, and the conditions of service providers. The questionnaires consist of five modules; each gathers information from a specific respondent. Module 1 comprises five books: Book 1A Household Core, Book 1B Married Women aged 15-49 years, Book 1C Children aged 6-15 years, Boolc ID Children aged 5 and under, and Book 1E Education test for reading and math of children aged 6-15 years. Module 2 to 5 each consists of one book; the respondents are villagelward heads (Book 2 Village Characteristics), heads of health center (Book 3 Community Health Facility), midwives (Boolc 4 Midwife) and headmaster (Book 5 School). 3.2 Sampling for the baseline survey Raizdonzizatio~z purpose, participation in the pilot programme has been ra~~domized For impact evaluatio~l over sub-districts. The CCT pilot was implemented in sub-districts that have been randomly assigned to the treatment group. The control sub-districts then reflect the counterfactual of the treatment group had they not been treated, and impact of the program can simply be derived by comparing the treatment and control groups (before and) after the pilot has been implemented. Due to the different nature of PISH and PNPM Gerzernsi, both followed a different process for selection of eligible sub-districts that would be subject to randomization. First, within each province, the 20 percent richest districts were excluded for both CCT programmes (based on school transition rates, malnutrition and poverty). Districts receiving KDP were eligible for PNPM Gelzerasi, from which 20 were selected, stratified by province. In NTT, East Java and West Java selection was random, in Gorontalo and North Sulawesi all eligible districts were selected. Within the selected districts, sub- districts were not eligible if they had participated in the Urban Poverty Program (UPP) or where less than 30 percent of the villages (desn) and urban precincts (kelurahan) are considered as r ~ ~ rbyl the national statistics office, BPS. This final screening yielded 300 a PNPM Generasi eligible sub-districts, which were randomly assigned to incentivized treatment (hereafter referred to as treatment I), non-incentivized treatment (hereafter treatment 11) and the control group. Randomization was stratified by district. The remaining districts were considered for PISH. The sub-districts that were deemed as "supply-side ready" were then randomly assigned to the PISH treatment and control groups. The supply side-readiness criteria were based on a statistical analysis of existing health and education facilities and providers in these sub-districts (World Bank, 2007a). The selection and rando~nizationprocedures are summarized in Figure 1 Sar~zplir zg The baseline survey was conducted in 660 sub-districts in six provinces where the PI(H household CCT and PNPM community CCT program are implemented. The household CCT baseline survey covers six out of the seven provinces that participate in the 2007 pilot project (excluding West Sumatra) while the community CCT baseline survey covers all five provinces of the pilot. The sampled PKH sub-districts are located in 44 districts (out of 48 districts in the pilot). The PNPM sub-districts cover all 20 participating districts. For the baseline survey, all 300 PNPM treatment and control sub-districts were included, while a subset of 180 sub-districts was randomly sampled from the PKH treatment and control group each, yielding a PKH sample of 360 sub-districts. The PKH sample draw was stratified by urban and rural classification.' Within all sub-districts, eight villages were sampled, within each village one ward (dusun) was randomly drawn, and &om each ward, five households were sampled. But the selection process for villages and l~ouseholds differs between the PKH and PNPM samples. For the PNPM villages and households, the selection process was straightforward. The villages were randomly selected from the full list of villages. Households in a ward were categorized into three groups: (i) households with pregnantllactating mothers or married women who were pregnant in the last two years, (ii) households with children age 6 to 15 years, and (iii) re~llaining households. From these groups, five households were randomly selected: two froill group (i), two from group (ii) and one from group (iii). With the PKH survey, there was a concern that random sampling of households could possibly yield a relatively small number of treated observations in the sample, if programme coverage would be small relative to the population size. Therefore a purposive sampliilg strategy was adopted, whereby the sampling procedure targeted UCT eligible households, as this is the subset of the population that would be considered as eligible for the PIW transfers. Since the UCT program itself suffered from leakage to the non-poor, a proxy means test was employed to select the poorer households on the UCT household list. One i~llplication this sampling strategy is that the PKH sample will be of perceived as relatively poor, compared to the actual population of the PKH localities. Villages in PKH sub-districts were first screened on UCT eligibility of all households. Only villages with at least five UCT eligible households per ward were considered for A sub-district is here classified as lural if the share of urban precincts (lieluruhu~t) less than 30 percent is of the total of usban precincts and mral villages, according to the 2005 PODES. 17 sampling. Thus, it is possible that less than eight villages were sampled for some sub- districts. In this case additional wards would be randomly selected from the remaining villages, as to balance the number of sampled wards in the sub-districts. Within a ward the UCT eligible households were classified similar to the PNPM households. Among the UCT eligible households five households were randomly selected from group (i) and (ii) only: two from group (i) and three from group (ii). The different steps in the sampling procedure are shown graphically in Figure 2 for the PNPM and Figure 3 for the PI- baseline survey. Sanzplirzg lzealtlz and erlrlcatiorz providers For each sub-district in the PIGI and PNPM samples, one community health center health centers operating (Puskesmas) was randomly selected from a list of all comn~unity in each sub-district. For sub-districts without community health centers, the sampling frame covered Puskces~rzas located in other sub-districts whose working area includes sub- districts without Puskcesrnas. Per sub-district three secondary schools were randomly selected from all secondary schools (public, private, regular/vocational, or other type of equivalent school) located in the sub-district. The sampling frame for midwives was constructed from two sources: a list of midwives working for health colll~nunitycenters in each sub-district but also running a private practice, and a list of private midwives. Information for the latter list was obtained directly from households. In each sub-district, two midwives are selected from the first list and two from the second list, yielding a sample of four midwives from each sub- district. Saiizple size arzd nort-resporzse The final saillple size is given in Table 2, broken down by type of respondent. The sample includes 14,326 (36,801 treated and 36,762 control) households from the PKH sub-districts and 12,000 households from the PNPM sub-districts (4,000 for treatment I, I1 and control group each). Besides households, the survey modules covered 10,899 children younger t11a113 years, 28,397 children age 6 to 15 years, 25,567 mamed women age 16 to 49, 658 coi~~munity health centers, 1861 midwives and 2564 schools. Non- response rates are low, well below one percent for individual household members, as shown in Table 3. However, we see very high non-response for the language and math test results, with 57 percent of children age 7 to 12 and 52 percent of children age 13 to 15 not completing the test modules. 3.3 Weights Weights were constructed so as to translate the survey data to a representative depiction of the reference population for the survey. In the case of the PKH survey, this is the group of households considered as poor among the UCT beneficiaries living in PKH eligible sub-districts. For PNPM Generasi, this is the entire population of the treatment and control sub-districts. Weights are also constructed for the schools, community health centers, and midwives. These facility weights should reflect the probability that a random individual living in the target area is exposed to the social service supplied by the specific school or health care provider. The weights are constructed in similar fashion except for the community health centers. Weights for households, villages, midwives, and schools reflect the probability of being sampled, while community health centers reflect the coverage area. The household weights are calculated as the inverse of the sampling probability Pr(house11old) = Pr(sub - district) x Pr(vil1age / sub - district) x Pr(ward / village) x Pr(househo1d / ward) That is, the probability that a particular household is sampled given that it resides in a treatment and control area (and is considered as UCT eligible, in case of the PKH survey). Below we elaborate on how each element in the sampling probability has been constructed. The sample probability of sub-districts, Pr(sub-district), is equal to 1 for the PNPM sample as all treatment and control areas are included in the survey. The sampling probability of PISH sub-districts is simply the ratio of the number of selected sub-districts in each of the 20 selected districts over the total number of eligible ("supply side ready") sub-districts in a district. The probabilities are stratified by urbanlrural and treatment/control status. The village sampling probabilities, Pr(vi1lage / sub-district), are based on the probability that a village is sainpled from a specific sub-district, conditional on being located in a sampled sub-district. For PNPM this is the ratio of the nuinber of selected villages over the total number of villages, according to the PODES 2005 data base. For PKH this is the ratio of the number of selected villages in a sub-district over the total number of villages that have at least 5 UCT eligible households per ward. In some PKH sub-districts less than 8 villages were sampled as results of the UCT eligibility sampling restriction, in which case Pr(vi1lage 1 sub-district) = 1. The construction of the conditional sampling probability for wards, Pr(ward I village), is similar for PKH and PNPM, since they were drawn randomly for both samples: the ratio of the number of selected wards over the total number of wards in a village. The probability of sanlpling households from the selected wards, Pr(househo1d I ward), is based on the classification of households into the three categories mentioned above in the n sampling section. L each ward, the probability of selecting a household is calculated as the ratio of the nuinber of sampled households £rom one of the three specific categories over the total nuinber of households in those categories in a ward. Note that the PNPM sample includes all of the categorized population groups, while the PI(H sample is restricted to households with pregnantllactating mothers, manied women who were pregnant in the last two years, or children age 6 to 15 years. Thus #type i 1 ii sampled households in PI(H ward Pr(househo1d /ward,PICH) = #type i / ii households in PKH ward #type i / ii / iii sampled households in PNPM ward Pr(househo1d /ward,PNPM) = #type i / ii / iii households in PNPM ward Similar to the housel~oldweights, weights for schools and midwives are constructed by multiplying the sub-district sampling probability with the probability that a school or midwife is sampled from the sub-district: Pr(educationihea1th facility) = Pr(sub - district) x Pr(faci1ity I sub - district) The probability contributioll Pr(faci1ity / sub-district) is based on the lists used as sampling frame. 11 case of community health facilities, we use the fraction of the sub- 1 district population falling within the service area of the health facility, reflecting the probability that a rand0111household in a sub-district is covered by this facility. 4. EVALUATION DESIGN AND BASELINE ANALYSIS 4.1 Evaluation design The nature of the CCT programmes is such that simply comparing CCT participants with non-participants will provide an inaccurate estimate of the programs' impact. Eligibility for PKH participation is based on household and regional indicators of relative deprivation, while targeting of PNPM Genevasi is based on relative conditions and perfolmance of public services in villages. Naive impact estimates would therefore suffer from classic selection bias: the treated sub-districts have worse outcome indicators to begin with; hence the irnpact of both CCT programmes will be underestimated. Another problem for identifying the impact of the CCT programs is that non-participants can be affected by the intervention. Such spill-over effects, or externalities, could occur within localities. For example, if local service provision is effected by changes in demand for these services as a result of the household CCT; if the village economy is affected by CCT the influx of I~ousel~old funds; or if I~ouseholdchange their behavior in order to be eligible for participation. But externalities can also come about between localities, for example if improved public services in a PNPM Generasi village are also available to individuals outside that area or if the program induces migration. In all these examples, non-participants can benefit or suffer from CCT participation by others. Bearing these potential sources of bias in mind, the empirical strategy for estimating the impact of the two CCT programmes builds on the hypothesis that random selection of sub-districts has removed any structural difference between treatment and control groups, such that, on average, they share the same characteristics and outcolne indicators in the absence of the program. On this basis, any differences observed after implementation of the programs can be attributed to the respective CCT interventions. Randomizing over larger geographical units such as sub-districts (and not households or villages) reduces the scope for spill-over effects between sub-districts. Since we will be comparing treatment and colltrol sub-districts, any bias from intra- and inter-village spillover effects is eliminated. In terms of iilte~ilal validity of the impact estimates, any selection bias will be removed due to randomization. For the PNPM survey the evaluation design will then be straightforward. With treated units being sub-districts and their populations, treatment effects are identified by the differences in average outcomes between the sampled individuals from the treated and non-treated sub-districts. Baseline data can be used to verify whether the treated and non-treated are indeed identical in absence of the CCT, and if necessary control for any pre-i~ltervention difference. For the PKH study the impact evaluation will b e less straightforward, as sub-district treatment status is randomized but individual selection is not. At the time of the baseline survey it was not yet ltnowu which households in the sample would be eligible for participation. In PKH sub-districts participation in the program was later determined by a selection process which iilvolved both BPS and the Ministry of Social Affairs. Part of this process entailed selectioil on statistical grounds, and part on interactio~lwith the communities thenlselves. Because this process was clearly not random, comparing PIW participants in treatlnent areas with all sampled households in control areas is likely to yield biased results. For a future impact evaluation we would therefore like to know which households in the PIC3 control sample would have been eligible if the same selection procedure would have been conducted in the control sub-districts. But since this is unknown, and due to the danger that not all elements of the selection procedure could be controlled for with the survey data, other techniques are called for. One approach could be to take estimate the intention to treat (ITT) effect, by adopting the same estimation method as for the PNPM experiment, and look at differences in average outcomes for the full salilples of treated and control sub-districts. The ITT basically reflects a weighted average of the direct effect of CCTs on participating households and the external effects for the whole population. The advantage of this approach is that it exploits the only true random variation in treatment assignment. The drawback is that it will not be possible to isolate the direct effect of the PIW. The difference between the ITT and the direct effect will become larger as the fraction of treated households in the sample decreases. There are some concerns regarding extemal validity of the impact study. The first relates to generalizing impact estimates., One should be cautious with drawing conclusions regarding expected effects of scaling up the CCT program, based on the estimated treatment effects fiom the inipact evaluation, because of the distinct nature of experimental areas. For example, the study areas have been selected based on specific criteria regarding poverty and supply side characteristics. As we find in section 5, some key health and education outcomes for the sub-districts in the survey differ from the national trends. Hence, we could expect that the CCT programs will have a different effect in the study areas compared to other areas of Indonesia. A related problem affects comparison of the PKH and PNPM schemes. Not only the different characteristics of PKH and PNPM sub-districts, but also the sampling strategy for PKH survey, makes it hard to directly compare impact estimates of both interventions, and interpret any differences. 4.2 Baseline analysis The goal of the baseline survey is therefore twofold: 1. To evaluate the balancing hypothesis: test for any pro-program differences in key outcome indicators and relevant characteristics. 2. To control for any observed pre-program differences in key outcome indicators, if these may occur in the impact evaluation. The baseline analysis presented in this report is mainly focused on the first point raised above: testing for differences between treatment and control areas. The results of this analysis will serve as input for future impact evaluations, and indicate whether any baseline controls are required. An initial aim of the baseline study was to analyze targeting of PI(H transfers. But this had to be abandoned since data on the PISH beneficiaries was not yet available. But even if that data had been available for this study, it would not have been possible to provide a complete picture on PICH targeting, since the PKH respondents are not representative of their respective sub-district populations. In the final section of this report we discuss how this problem could be addressed in the follow-up survey. Besides the balancing test, we will look at the main determinants driving the lcey outcome indicators, by regressing these on a number of individual and household characteristics, and supply side variables. We include district fixed effects as to control for differences in local health aid education policy. It has to be stressed that these regression results are merely correlations and cannot be interpreted as casual effects. In addition to this, these regressions can help shed some light on the scope for extemalities. Although the possibilities for this are limited due to methodological problems and the many confounding effects that hamper identification of the extent of extemalities, the data does allow us to investigate the relevance of potential conduits through which extemalities could occur. Such conduits include prices (as the CCT program may bring about behavioral responses by health care providers and fees charged for their services), migration, congestion and crowding out effects, peer effects, and fertility. Note that migration, fertility and peer effects through women's decision-making power are based on information in the household survey, and regional averages can only be calculated for the PNPM sample as this is representative for the sub-district population, while the PKH data is not. We can also investigate potential externalities by loolung at evidence fkom other health and education programs. For example, the survey provides information on individual participation in scholarship programs and a pro-poor targeted public health insurance (Askeskin), as well as aggregate coverage of these schemes (in schools and villages, respectively). Absence of spill-over effects would not rule out externalities due to the CCT programs, of course. But any negative correlation of health and education outcomes with aggregate participation in existing programs (while controlling for individual participation), would suggest that health and education services are sensitive to congestion and crowding out effects. In the empirical analysis presented in the following sections we will first sketch the context for health and education outcomes in the survey areas by comparing a selection of indicators to national outcomes, and discussing the results of the multivariate analysis for possible determinants of the key outcomes and scope for extemalities. The balancing properties of the sample will then be evaluated by mean comparison tests between treatment and control areas on - Target service coverage indicators in health and education - Key outcome indicators in health, education and child work - characteristics Individual and l~ousel~old - Health and education service provision - Village characteristics In the analysis standard errors have been adjusted for sampling design and clustering at sub-district level. 5. RESULTS OF THE BASELINE SURVEY This section presents the empirical results regarding the balancing properties of treatment and control groups. The first two subsections discuss the results for the target service coverage indicators and ltey education and health outcome indicators. Then we compare balance in other individual, household, village and service provider characteristics. 5.1 Education 5.1.1 Regional patterns We compare net enrohnent rates of primary and junior secondary school in the treatment areas of PKH and PNPM Ge1zerasi baseline survey to Susenas 2004 survey. Table 4 summarizes net enrolment rates by province in the baseline survey and for Susenas 2004. The figures for the treatment sample areas are presented separately by PKH and PNPM Generasi areas. Net primary enrolment in the PKH areas is 85.6 percent, which is lower than the average net enrolment rate according to Susenas 2004, reflecting the relatively poor population covered by the CCT pilot. Both the PKH sample and the Susenas show primary enrolment outside Java to be higher in rural areas. For Java enrolment is higher for urban PKH areas, while we see the opposite for the full population of West and East Java. Primary school enrolment in PNPM Genevasi is higher than in PKH areas, except for North Sulawesi and Gorontalo. Net junior secondary school enrolment in PKH areas is 51.9 percent, which is considerably lower than the national average of 65 percent. A pattern we also observe across provinces. In PNPM Gerzerasi areas, the enrolment rate is 62 percent (Table 5), closer to that of Susenas 2004. But in three provinces (North Sulawesi, Gorontalo and East Nusatengara) the PNPM Generasi baseline data generate higher net enrolment. Table 6 provides the transition rates to junior secondary schools by province obtained from the Ministry of National Education. It is calculated as the proportion of the number of new entrants to grade one in junior high schools within a year of primary school gaduation. The transition rates for a province can be higher than 100 percent partly because new entrants iin a province may come from primary school graduates in other provinces who moved into this province. The CCT baseline figures cannot be directly compared to that of the Ministry of National Education due to differences in data and formulae to derive transition rates. Table 6 also suggests that the transition rates may change substantially over time, with the province figures loolting particularly volatile. We could not calculate reliable transition rates from primary to junior secondary school with the data, as the timing of the survey (June to August) overlaps with the transition of academic years, which is the very moment that the decision tra~lsitioilsare (being) made. In itself this is not a problem for the impact evaluation, as long as randoinization balances pre-program primary school enrolment rates. Nevertheless, we do want to get an indication of the (balance in) current transition rates, as the transition from primary to juniors secondary is the main source of concern regarding the government's nine years basic education ambitions. We therefore include a very rough proxy of transition to secondary school: the junior secondary school enrolment rate conditio~lalon having completed primary school. The drawback of this measure is that it is not fully comparable with national statistics on transition rates. The advantage of this indicator is that it captures the transition from primary to secondary school and junior secondary drop-out, both of which are key bamers to achieving a universal nine year basic education target. Transition to junior secondary education is higher in urban areas, with large variation between provinces: low in West Java, high in Jakarta, Gorontalo and Nort Sulawesi. Particularly striking is the large disparity between urban and rural areas in East Nusatenggara. Overall, transition rates are higher for PNPM than for PKH areas. 5.1.2 Determinants Correlations of education and child work indicators with socio-economic, village, school and sub-district characteristics are given in Table 36 for children age 7 to 12, Table 37 for children age 13 to 15, and Table 38 for math and Bahasa test scores. Enrolment is among a specific age expressed as the gross pal-ticipation rate, which reflects e~uolment group, irrespective of the level of enrolment. Child work is defined by any work activity that involves earning illcome or being compensated in kind (referred to as economic work) and domestic work. The tables show regression results controlling for district fixed effects. It is i~nportantto reemphasize that these results serve a descriptive purpose and cannot be interpreted as causal effects due to potential confounding factors and endogenous variables. Nevertheless they can show us patterns in the main determinants of education outcomes and conduits of externalities. The latter could be captured by variables reflecting congestion (average class size in junior secondary school; spill-over effects from other interventions, by including scholarship coverage in schools, controlling for individual participation), peer effects (average school absence; average national school test scores in the sub-districts), and migration. School fees are not included as these are generally low under the BOS program. The results show that there is no gender gap in enrolment, but that girls are more diligent, do more domestic work than boys, and less economic worlc. They also perform better in math and language tests. In PNPM areas, young children from households that rely on agriculture for their main source of income do relatively more economic and domestic work than children of the same age from other households, while older children attend school less. Note that this does not reflect local affects as we already control for the rural status of the villages (which shows no significant correlation). Children in agricultural households perfonn less in all tests, for all age groups and samples. In contrast to this, we find that students score relatively high on Indonesian language (Bahasa Irzdolzesie) tests i n mral PISH villages. Surprisingly, for schooling and work we find little correlation between with expenditure quintiles or parental education, and where we do find significant results, the coefficients do not show a consistent pattern. However, both household expenditure and parental education are positively associated with test scores. Among children from 1~o~:seholds with relatively many small children, we see lower enrolment and higher worlc incidence. But in households with relatively older children, the prevalence of child worlc is lower. Children achieve lower test scores if they live in households with a high percentage of children overall. We see little correlation between the number of schools aud enrolment. but we do see lower work incidence. Regarding potential externalities, in-migration is associated with more work, less enrolment and lower tests scores. But the standard errors are large. If we look at externalities form other interventions, there seems no evidence that scholarship coverage in schools has affected education of other children. However, scholarship coverage is associated with lower outcomes, but again, standard errors are large here. There is some correlation with peer effect variables. High average student absence in junior secondary schools is associated with lower individual attendance and higher economic work incidence for younger childreti. For test scores we find that individual performance is better for math and Indonesia11 language if the average school performance in national exams is relatively high for math, but low for English. 5.1.3 Mean compariso~i tests Edztcntiort target inrlicntors The mean comparisoil tests for education target indicators are presented in Table 18 (PIM) and Table 20 (PNPM) for children age 6 to 12, and Table 19 (PKH) and Table 21 (PNPM) for children age 13 to 15. Note that the latter age group reflects the junior secondary school reference population, while the former is broader than the primary school age reference group, which in Indonesia typically is 7-12 years. However, in practice, school enrolme~lt11 Indonesia is not negligible among 6 year olds. This is also i reflected in the baseline survey data, with an average enrolment rate of 56 percent. We therefore include schooling of 6 year old children in the mean comparison tests, except for some age specific indicators (such as net enrolment rates). Net primary and junior secoiidary school enrolment rates are balanced across treatment and control groups. Net priillary enrolment lies below 90 percent for all groups (between 85 and 89 percent), which is low compared to the national average 93 percent (in 2004). Net junior secondary elirollile~lt also below the 2004 national average of 65 percent, lies but there is large varialion between PKH and PNPM groups: 52 to 54 percent, and 60 to 64 percent, respectively. For gross participation rates we see some statistically significant imbalances for children age 7 to 12 in the PNPM sample, and children age 13 to 15 in the PKH sample. Gross participation is close to universal for the youngest age group, ranging from 93 to 96 percent. For the older age group gross participation is lower, between 69 and 86 percent, but still well above the net junior secondary enrolment rate, indicating high repetition rates and delayed enrollnent in primary school. Table 18 and Table 20 show the absorption rate of new pupils in primary schools. This measure assesses the nunlber of newly enrolled primary school students as a fraction of the number of children aged 7 years old (the official primary school enrolment age). We calculate this at the sub-district level, taking weighted population means. We do not find statistically significant differences, but it has to be noted that standard errors are likely to be large given the relatively few obsemations (i.e. the number of sub-districts). Our proxy for junior secondary transition rates among 13 to 15 year olds is given in Table 19 (PIG!) and Table 21 (PNPM). Junior secondary school enrolment varies from 80 to 89 percent, being slightly higher for PNPM children, but with no statistically significant differences between treatment and control groups. School attendance rates are high, with 92 to 95 percent of enrolled children age 6 to 12 meeting the 85 percent attendance target (irrespective of a 1 or 2 week recall period). Full attendance rates are slightly lower. For the PKH samples there are no statistically significant differences in attendance, but for the PNPM we see some imbalance for the treatment I and control groups (87 against 92 percent). For the older age group we see significant differences in the school attendance target performance (2 week recall) in the PNPM sample: 94 percent in the treatment I group achieved the 85 percent attendance target, against 90 percent in the control group. Clzild tvork, costs of sclzoolirzg nrtd edricntiorz ouiconzes Mean comparisoil tests for child work indicators and costs of schooling are presented in Table 22 (PICH) and Table 25 (PNPM) for children age 6-12, and in Table 23 (PKH) and Table 26 (PNPM) for children age 13-15. We look at different levels of intensity in work: at least one hour of work in the last week and at least 20 hours of work. Overall, incidence of child work seems quite balanced, but we find statistically significant differences for the age group 13 to 15 in domestic work activities, for both PICH and PNPM. At this age 71 to 87 percent of children reports domestic work activities and 22 or 30 percent is engaged ill work activities that generate some kind of income, with 11 to 16 percent working at least 20 hours per week. For students below 13 there is no noticeable difference between treatment and control groups in distance (minutes) and costs (Rp.) of travel to school. But for older students differences in the PKH and PNPM treatment and control there are statistically significai~t groups. There appear to be no imbalances in the fraction of scholarship recipients. Education test scores are presented in Table 24 (PKH) and Table 27 (PNPM). The tables show the percentage of correctly answered questions for a set of home based tests, in math and language, for two separate age groups (7 to 12 and 13 to 15) irrespective of enrolment status. Children's test performance was similar in treatment and control groups. Children in PNPM areas performed slightly better than in PICH areas, in particular for children age 13 to 15. 5.2 Health indicators 5.2.1 Regiotlal patterns Ii~inizcnizationrnfes Immunizatioil rates vary greatly between provinces and samples. Table 7 shows the percentage of children age 0-36 months that received complete immunization at specific ages. Immunization is more prevalent in the PNPM Generasi than the PKH intervention areas. For the latter, coiuplete immunization rates are low on Java compared with the other provinces, with particular low rates on Java, where less than a quarter of children in the PKH sample have received the required immunization given their age. For children in the PNPM we see higher immunization rates and a contrasting pattern, with relative low rates in Gorontalo and East Nusatenggara. The IDHS figures apply a different reference group. It shows that more than half of children age 12-23 months have been h l l y vaccinated; that is, they received immunizations against tuberculosis, three doses against diphtheria, pertusis, and tetanus, and three doses against polio, and measles. Over 60 percent of children in East Java, North Sulawesi and East Nusatenggara have received all their vaccinations, compared to only 41 percent of childreil in West Java. This corresponds with findings in the PIW sample, that iminunizatioll in West Java lags behind other provinces Cltild r~torbiclity The baseline survey collected information on self reported illness an symptoms in the last month preceding the survey for children age 0-3 years. The prevalence of some illness (diarrhea, high fever, cough, and acute respiratory infection'') in different provinces is provided in Table 8. DKI Jakarta has quite high prevalence of diarrhea, fever and other provinces while prevalence of ARI is relatively high in coughing symptoms tha~l North Sulawesi and East Nusatenggara. Child morbidity is relatively low in East Java. Note that conlparisoll of these morbidity rates requires some caution, as self reported illness is biased upward when set against income. Self-reported illness typically depends on the affordability of care, as the rich reporting illness more often than the poor, which is surely not a result of the rich having a worse health status than the poor. On average, self reported morbidity rates are higher with PNPM respondents, but this is not consistent across provinces. Since the PKH sample is expected to be (on average) poorer than the 'O The symptoms of ARI are fever, couzl~ing, accompanied by short, rapid breathing. PNF'M sample, due to purposive sampling in PKH sub-districts, it is not clear whether difference in morbidity rates reflect differences in health status or reporting bias. Table 9 shows the percentage of children who suffered from ARI or diarrhea in the month before the baseline survey and who sought treatment at a health facility or with visiting health officers. Despite the high prevalence of child morbidity, 70 percent of cases with ARI synlptoins were treated in Jakarta. The treatment rate is also high in East Java and North Sulawesi, in contrast to Gorontalo (PKH) and East Nusatenggara (FNF'M). The 2002-2003 IDHS survey reports higher incidence and treatment rates for ARI and diarrhea compared to the CCT pilot (Table 10). In particular striking are the regional difference: high morbidity and low treatment rates in Gorontalo, against low incidence and high treatment rates in Jalcarta and East Java. CIzild iz14tritionalstatzrs Children's nutritional status is reflected by the incidence of child stunting, wasting and underweight. The indicators are anthropometric z-scores computed following the WHO Child Growth Standards. Stunting, based on a child's height and age, is a measure of chronic nutritional deficiency. Wasting, based on a child's weight and height, is a measure of acute nutritional deficiency while underweight, based on weight and age, is a composite measure of both acute and chronic malnutrition. Table 11 shows the percentage of children age 0-36 months that are malnutrition and severe malnutrition, according to WAZ, HAZ and WHZ scores.'' All indicators show that, overall, the extent of malnourishment is higher anlong children in PKH compared to PNPM. With some exceptions (notably severe underweight in North Sulawesi and East Java) this pattern is also observed in the different provinces. Stunting is more prevalent anlong the children in the survey than wasting. About half of the children aged 0-36 months are stunted, and around a third are categorized severe. There II Malnutritioil (undermeight, stunted, or wasted) is indicated by a z-score equal to or below -2, and severe malnutrition by a z-score equal to or below -3. 33 are strong regional disparities between provinces in Java, which have better children nutritional status, and ilon-Java. An exception is Jakarta which again has a higher prevalence of children underweight, stunting and wasting. birflz Assistance d111lri12g Information on birth deliveries by various types of assistance within the last two years preceding the survey is collected in the baseline survey from mamed women aged 16-49 years. A delivery is considered as assisted by a trained professional if it was attended by a doctor or midwife. Table 12 shows that the percentage of professioilally assisted deliveries is higher aillong women in PKH areas compared to PNPM Generasi, most likely because of the sub-district supply-side restriction for PKH participation and because PNPM Ge~zer-asi covers relatively rural sub-districts. Compared to the IDHS 2002-2003, the percentage of professionally assisted deliveries in both CCT pilot schemes is lower, which can be explained by the pro-poor design of the pilot. Professional assistance with deliveries is more prevalent in Jakarta, East Java and North Sulawesi, according to the CCT baseline survey and IDHS. Both data sources show that a large number of mothers in East Nusatenggara and West Java tend to choose a traditional birth. arzdyost~zntnl Al~terzatal enre According to the 2002-2003 IDHS, 92 percent of mothers who had a live birth in five years preceding the survey received at least one antenatal checkup at a health care provider. Further, 81 percent of mothers had four or more antenatal care visits, as is recommended by the Indonesian maternal health program (Central Bureau of Statistics Indonesia et al., 2003). The CCT baseline data reports that at least four antenatal care visits were made in 69 percent of pregnancies in PKH areas and 79 percent pregnancies in PNPM Ge~zerasi areas, during the 24 months preceding the survey (Table 13). On average, pregnant women in the PNPM Ge~le~asi intervention villages had 7.9 visits compared to only 6.6 visits in PIW villages. At province level we also see higher frequency of antenatal care visits for PNPM Gellerasi compared to PIG3 areas. Antenatal care is more frequent on Java than outside Java, with North Sulawesi having the lowest share (56 to 60 percent) of pregnant women meeting the CCT target of four antenatal visits. Distribution of iron supplements is an important component of antenatal care. According to the 2002-2003 DDHS, 78 percent of women who received antenatal care received iron tablets. Three in ten of these women took the recommended 90 or more tablets during pregnancy. According to the CCT baseline data 82 and 86 and percent of pregnant women in PICH and PNPM Generasi treatment areas, respectively, received iron pills during their pregnancy. These figures are higher than suggested in the 2002-2003 IDHS. However, only for 11.7 and 16.9 percent of pregnancies did women receive 90 or more iron pills during pregnancy, compared to 30 percent indicated in the IDHS data (Table 14). Incidence of having received the recommended 90 iron pills during pregnancy varies across province. For PICH areas it is particularly low in Gorontalo (3.6 percent) and West Java (8.7 percent), coillpared to East Java (14.3 percent). The percentages in PNPM Generasi areas are higher (ranging between 11 - 18 percent) than in PKH areas (3 - 14 percent). The CCT baseline survey also inquired about postnatal service from health care providers during the first 40 days after delivery. In PI- and PNPM Generasi areas, the average numbers of postnatal visits are comparable (3.5 and 3.6 visits) yet the distribution varies widely by province. On East Java, the average number is 4 postnatal visits, compared to less than 2 visits in East Nusatenggara. These frequencies translate to low average rates of meeting the CCT target of at least 2 postnatal care visits: 43 percent of live births in PKH and 52 in PNF'M Ger~err~si intervention areas, and go as low 25 percent in East Nusatenggara (PICH) and 35 percent in North Sulawesi (PNPM). Overall, the results show large scope for improvements in ante- and postnatal care in the CCT pilot areas, as the extent of care and iron supplements provided falls well short of recommended amount and the CCT targets (on average), with large variation between provinces. We would therefore expect a large potential impact of the CCT interventions. Mortality rates Some early childhood mortality rates calculated from the CCT baseline survey are presented in Table 16. They are neonatal mortality rates (children less than 30 days old) and infant mortality rates (younger than 12 months). The difference between the two is referred to as the post-neonatal mortality rate (age 1 to 11 months). Overall, the baseline data show ileonatal mortality rates of 41 deaths per 1000 live births in PICH intervention areas and 34 in PNPM Generasi areas. The infant mortality rates lies at 80 and 81 for the PKH and PNPM Generasi areas, respectively. This is considerably higher than the mortality rates de~ivedfrom the IDHS data, shown in Table 17. This shows neonatal nlortality rates ranging from 16 to 31 and infant mortality rates from 25 to 77 per 1000 live births, all well below the averages in the CCT intervention sub- districts. This difference could reflect the relative high degree of deprivation in the CCT pilot areas, due to the pro-poor design. It should also be noted that the reference period applied the IDHS 2002-2003 survey for calculatiiig mortality rates is 10 years before the survey date, while for the CCT baseline survey we use retrospective information for 5 years prior to the survey. According to the IDHS data, in general neonatal mortality is high East Nusatenggara and East Java, and relatively low in Jaltarta and North Sulawesi, while post-neonatal mortality is relatively high i11 Gorontalo and East Nusatenggara. The mortality rates calculated from the CCT pilot survey show a very volatile pattern across provinces, much stronger than is observed in the IDHS survey. The most likely explanation is that the number of observations in the CCT survey is not large enough to decompose indicators for such low frequency events by province. 5.2.2 Determinants The regressioll results are given in Table 39 and Table 40 for children's health target indicators and outcomes, respectively. Table 41 presents regressions for outpatient utilization and Table 42 for target indicators for married women age 16 to 49. We include potential externality variables that could capture effects from behavioral response of health service providers (prices) and migration. We also include individual and average village participation in the Aslceslcin program, to investigate experiences wit11 spillover programmes. In case of target indicators for manied women effects in other public l~ealtl~ we also include fertility (the crude birth rate per 1000 inhabitants over the last 24 months) and women's decision making power indices. Two indices are constructed based on as set of questions on whether women have a say on decisions regarding their children and household cons~~mption. indices range from 0 to 1, with a value of 1 indicating that a The woman has decision nlalcing power on all issues raised in the questionnaire. We include the individual indices and a sub-district average, to capture peer effects. Girls show better nutritional status (in both height and weight) than boys. Parental education shows a strong positive correlation with self reported morbidity and health targets for in~nlu~lizatio~l weighing. But health status and the programs' targets are and only weakly co~velated with agricultural activities of households or per capita household expenditure. A large share of children in the household is associated with a hlgher degree of malnutrition, lower i~lcidellce meeting immunization and weighing targets. Aslceskin of beneficiaries show higher morbidity and lower nutrition, which could reflect pro-poor targeting or indicate selectioll on needs. On the other hand, they are more lilcely to meet the target of monthly weighing. We see no evidence of spillover effects from the Aslcesln'n program at village level. Regarding other conduits for externalities, in-migration and crude birth rates show negative correlation with immunization and frequency of weiglling, suggesting potential spill-over effects through increased pressure on public services. The res~~lts supply side factors and prices are hard to interpret, wit11 both for positive and negative correlatio~l incidental statistical significance. and for Outpatient care is higher for females, and lower amongst l~ousel~olds whom agriculture is their main econo~nic activity. We also see a strong positive correlation with per capita household expeuditure and education of the head of household. This holds for health care from public, private aud traditional providers. There are also clear patterns for household composition. Utilization is lower for larger households, with a large share of children age 3 to 15 years, relative to adults (while controlling for individual age). Individuals from households with infants and toddlers, on the other hand, use more public and private care. Participating in Askeslcskin clearly increases pubic health care utilization. The size of the coefficients suggests a net effect on utilization and a substitution from private to public. The reported results are likely to be underestimates, given that they are not cleansed from possible selection effects. The supply side variables show some positive correlation with utilization of public and private outpatient care in PNPM sub- districts. We see potential for external effects on public care through in-migration. As with the earlier results for health outcomes, there appear to be no village level programme. externalities due to the Aslceski~z Women from agricultural households tend to make less use of professional assistance with births, and have less postnatal visits. Per capita expenditure and the level of education of the households head is positively correlated with target indicators for manied women. Pregnant wome~: in households with a large share of children younger than 2 years are more likely to have at least 4 antenatal visits and receive 90 iron pills. But young mothers in these households are less likely to meet the target for postnatal care. Women in households with relatively many older children are less likely to meet the target on postnatal care or have professionally assisted deliveries. Having Aslceskin insurance is positively associated with receiving the required iron supplements, professional assistwce at bii-th and postnatal care. But the results suggest the Aslceskin does not lead to externalities. \Vomen's influence on decisions regarding their children is positively correlated with meeting postnatal care targets. For women's decision-making power regardiug household coilsumption we find conflicting results between PKH and PNPM samples for receiving iron pills. However, the results do not give us a clue on how peer effects througl~increased women's decision-making power would play a role in meeting ante- and postnatal care, as we see both negative and positive coefficients for the sub-district averages. The crude birth rate is negatively correlated with assisted deliveries in PNPM areas, suggesting that if the community transfers induce fertility rates, this could increase pressure on services by midwives. Correlation with supply side factors and prices are again hard to interpret with both positive and negative correlation. 5.2.3 Mean comparison tests Healtlz targets i~zrlicrrtors The target indicators for preventive health care for pregnant women and mothers are given in Table 28 (PICH) Table 30 (PNPM). For both CCT programs the treatment and control samples are balanced. There are minor differences, but these are not statistically significant. But there are differences between the PKH and PNPM samples. Among pregnant women in the PKH sanlple about 70 percent have had at least 4 antenatal visits and 12 percent have received at least 90 iron pills. In the PNPM sample these numbers are higher, with almost 80 percent meeting the target for antenatal visits, and 17 percent for iron pills. The PNPM sanlple also performs better on targets for postnatal care, with around 52 percent having at least two visits, against 44 percent for PKH. There are no statistically significant differences between any of the samples for the percentage of deliveries assisted by a trained professional (ranging between 58 and 62 percent). Table 29 and Table 31 present the target health service coverage indicators for children aged 0 to 36 months in the PISH and PNPM samples, respectively. While the PNPM sample is balanced for all outcomes, the PKH is not balanced for the frequency of weight monitoring at health facilities and vitamin A receipt. The children in the PKH treatment sample have been weighed at a health facility slightly more often in the last two months than those in the control group. While the difference in the average frequency is not statistically significant (1.10 and 1.04, respectively), the difference in the percent of children that was weighed at least twice (34 and 28), is. Children in the PNPM treatment and control sample are weighed more often, ranging from 1.25 to 1.34 visits in the last two months. The PNPM sample also shows higher immunization rates and vitamin A uptake. Complete irnnlunization of children, given their age, is just below 40 percent in the PKH areas, and just above for PNPM. The higher immunization for PNPM as compared to PKH holds for all immunization types: BGC, Polio, DPT, Measles, and Hepatitis B. Vitamin A uptake for childre11 aged 6 to 36 months ranges from 36 and 38 percent receiving vitamin at least twice per year in the PKH treatment and control groups, respectively, to 37, 40 and 42 percent in the PNPM treatment I, treatment I1 and control groups, respectively. These pattenls of vitamin A uptake translate to approximately 50 percent of pote~ltialuptake, based on the biannual national distribution of vitamin A capsules. Healtlt outconzes Key health outcomes that we will discuss here are morbidity of childhood diseases, malnutrition and mortality alllong young children, and health care utilization rates for public, private and traditional curative outpatient care. We observe statistically significant differences in treated ARI for the PNPM treatment and control groups, while for the PI= groups there are differences in average height for age of children age 0 to 36 months, and overall outpatient contact rates. Morbidity, malnutrition and mortality rates are reported in Table 32 (PI(H) and Table 34 (PNPM). Malnourishment in terms of the Height-for-Age Z-score for children under 3 years old is slightly higher in the PIG3 control group compared to the treatment group. The other malnourislunent indicators (Weight-for-Age and Weight-for-height) seem balanced. Incidence of malnourishment is slightly higher in PKH sub-districts compared to PNPM sub-districts. Self reported incidence of diarrbea and acute respiratory illness amongst young children is balanced between treatineilt and control groups. These morbidity rates are slightly higher in the PNPM areas compared to PKH. Incidence of diarrhea in the last month ranges froin 25 to 30 percent alnong children under 36 months. ARI is reported for 18 to 22 percent of children. Neonatal inoliality rates are also statistically balanced between treatment and control groups, varying between 27 and 45 deaths per 1,000 live births. While the differences between the treatment and control groups seem sizeable, the standard errors are relatively large, such that null hypothesis of no difference can not be rejected. We do see a statistically significant difference for infant mortality in the PKH sample: the high mortality rate among childre11 younger than 1 year in the PKH treatment area (81 per 1000 live biiihs) is not matched by the control group (54). Outpatient utilization and contact rates are shown in Table 33 (PKH) and Table 35 (PNPM). These indicate the average number of visits to a health care provider in the last month, and the fraction of the population that visited a provider at least once. There are no statistically significant differences between treatment and control groups. On average, an individual in a PKH sub-district has 0.14 to 0.15 visits to a public or private health care provider (excluding traditional care), with 0.18 to 0.20 visits for a person in a PNPM sub-district. Roughly a 27 to 36 percent of modem care takes place at a private health care provider, and 30 to 33 percent at a public community health facility (Puskes7nas). We see similar patterns for contact rates, with the exception that there is a statistically significant difference for overall modem care (public and private) in the PKH treatment and control group. 5.3 Housellold and individual characteristics Household and individual characteristics are presented in Table 43 (PKH) and Table 44 (PNPM). proJiIe Age and de12tog~npltic The age profile shows the average age of respondents in the different modules of the baseline sulvey: the average age in months of children younger than 3 years (Book ID), the average age in years of children 6 to 15 years old (Book lC), married women 16 to 49 members (Book 1A). We see years old (Book 1B) and the average age of all housel~old no statistically significant differences. Average household size is balanced, with PKH households at 5.4 people and PNPM households around 4.4. Edrdcation attrrirznrent Education attairuuelit of the population age 10 years and older is not balanced: among the PNPM controls we see there is a higher share of the population with no degree and less with just prililary schooling education, compared to the treatment groups. Higher levels of education attainment do seem balanced. Education attainment is higher in PNPM areas compared to the PKH san~ple. Social prograr~~~tzes The survey also aslts about participation in social programmes and insurance schemes, such as Aslces social health insurance for civil servants, Aslceskin health insurance for the poor, SLTIBLT unconditional cash transfer schemes and Raskii~ rice subsidies. Consistent with PICH sanlpli~igstrategy, almost all PKH households receive SLTBLT assistance. Since the PNPM is a representative sample of the population, the SLTIBLT is much lower, around 34 percent. Houselzold cxperzditrrre Household expenditure is expressed as per capita monthly household spending, broken down by food and non-food spending. In addition to these aggregates, we specify per capita health and education spe~iding. a1 types of spending, the treatment and controls For are balanced. This is reflected in the mean comparison test and a graphical account of the spending. Figure 4, Figure 5 and Figure 6 graph the full distribution of l~ousel~old distribution of per capita total household spending, and education and health spending, respectively, for each treat~ilent and control group. For PKH households average expenditure is just below Rp. 200,000 per head per month, and between Rp. 325,000 and 340,000 for PNPM households. Head of lzouselzolrl We observe some statistically detectable differences for the characteristics of heads of households. Among PNPM ho~~seholds,percent of heads of household in the treatment 5 I group are female, compared to 8 percent for treatment I1 and 7 percent for the control group households. Educatiou attainmeut in the control group differs from both treatment groups, in particular with a higher incidence of non-educated heads of household (17, 18 and 22 perceut for treatme~~t I1 and control group, respectively) and a smaller share I, with only a primary school degree (52, 51 and 46 percent, respectively). Similar to overall education attainment, we see that, on average, heads of household in PNPM households enjoy a higher level education than those in PKH households. For all groups, cultivation of rice aud secondary crops is the main source of income for the head of household, ranging from 61 to 66 percent. Living co~zditiorzs The household survey collects information on an array of living conditions. These include characteristics of the house or dwelling (such as the construction material of the roof, walls aud floor), accessibility to clean drinking water and the source of water, sanitary facilities (such as type of toilet, waste disposal), cooking facilities. We find no statistical differences between treatment and control groups. On average living conditions are better for PBPM households than the sampled PKH households. Assets While we foulld expenditures to be balanced between treatment and control groups, we do see the occasioual differeilce in the accumulation of assets. The information on assets in the survey iilcludes type of land (such as imgated rice and non-irrigated rice land, dry land, housing land), land area (ha), household appliances (such as television, radio, refrigerator), means of trallsportatioll (such as bicycle, motor cycle, car, boat) and livestock. For the PNPM households, asset ownership balances between treatment and controls, but for the PISH groups, the treatment households own considerably less land (on average 0.22 and 0.39 ha, respectively) and own a car or motor boat less often compared to control households (0.2 and 0.5 percent, respectively). Corttruzrnityy nrticiprrtiort Participation in con~munityactivities and organizations does not differ significantly between control and treatment groups, for either CTT program. 5.4 Village characteristics and service provider characteristics Village characteristics The village characteristics are remarlcably balanced between treatment and control groups (PKH villages in Table 45 and PNPM villages in Table 46). While we do observe differences, they are never statistically significant because the standard errors are large. The only significant difference is found with the percentage of households working in agriculture. An important result for the CCT baseline is that village level variables that relate to the supply of education aud health services are balanced. The effects of the CCTs crucially depend on the supply of basic health and education services. Any difference in the supply of schooling and health services could cause heterogeneity in (latent) treatment effects, even when initial target and key outcome indicators are balanced ex ante. Other possible confounding effects could be caused by imbalances in health insurance coverage. Other relevant variables that are reported in Table 45 and Table 46, and where we find no statistically significant differences, include population size, religious composition, waste drainage systems used in the villages, availability to communication and media services, welfare and infrastructure indicators, and local wages for non-skilled worlcers. facility artd r~~idrvives Coitrrttuitity ltealtl~ Unlike the village characteristics, which show that treatment and control groups enjoy similar supply of health services in terms of the number of facilities, we find a number of significant differences when we loolc at the supply and quality of health care at individual health care providers. Table 47 to Table 50 show characteristics of community health centers and village midwives sampled for the survey. The questions in the survey provide very detailed info~-mationregarding physical aspects of the health care facilities (such as rooms, water supply, salitation, etc.), staffing, instruments and materials, stock of medicine, details on care recently provided, and prices. The tables list a large number of characteristics, but here we will o~ily point out the most important variables for which we found statistically sigllificant differences between treatment and controls. health center, we see similar patterns in imbalanced variables with For the cominu~~ity both the PKH and PNPM facilities. Most notable are the differences between treatment and control groups in the ~luiliber staff, water source, toilet facilities, number of rooms of (in particular treatment and inpatient rooms), frequency of vaccination of infants (DPT for PKH and BGC vacci~iatioiifor PNF'M), and frequency of weighing of children. Especially worrying are the ftequent the imbalance in availability in instruments and materials, and staffing, in particular for the PKH facilities. We also see some differences in the available stock of medicine and unit price of treatment for both CCT samples. However, these are minor and infrequent deviations. For midwives, the PNPM sample shows many more imbalances than the PKH sample. Discrepancies in electricity a~idwater source, and the number of assistants is seen for both PISH and PNPM midwives, as are significant differences in the availability in instruments, and the frequency of vaccinatio~ls for infants. There are also price differences between treatme~lta ~ coiltrols for public and private services offered. In d case of the PNPM public and private midwives seem more expensive in the control group, while for the PI- the midwives in the treatment group generally charge higher rates for their services. Sclzools An array of school characteristics is given in Table 51 (PKH) and Table 52 (PNPM). The variables presented show characteristics, qualification, and experience of the teaching staff, type of school, results for national and scl~oolexams, school class rooms and facilities, nu~nber students per class room, drop out rates, changes in enrolment, school of attendance, scholarships recipients, and budgets and revenues. The treatment and control groups are balanced in lllost va~iables.In this section we will only highlight the statistically significant imbala~~ces. The observed difference that at first seems most worrying is the discrepancy between PNPM dropout rates. However, this does not seem to be a systematic source of bias as the pattern in the discrepancies is not consistent: dropout rates for second grade students in junior secondary school are higher in the treatment I1 group than in the control group, while for the third grade the drop out rates are higher in the control group compared to both treatment gro~~ps. Also the percentage of students with scholarships is lower in PNPM control schools. Differences in the PKH schools that may reflect quality of scl~oolingare the higher number of first grade students per class room in the control group schools, while they have less funds available for infrastructure maintenance and study-teaching and extracurricular activities. There are some differences in characteristics of the school principal. For PKH schools the principals in control areas are more likely to be males compared to schools in treatment areas (91 and 87 percent, respectively), while in PNPM control area schools they are younger on average (23.8 years compared to 31.8 and 35.6 to the treatment I and I1 schools, respectively). The ilumber of computers and students toilet facilities is larger in PISH treatment groups and the level of school sanitation higher, compared to controls. In PNPM areas we see significant differences in the number of rooms, leaks in the ceiling and blackboard and chalk markers. 6. CONCLUSIONS AND RECOMMENDATIONS This analysis of the baseline survey for the PICH and PNPM Gerzerasi CCT pilot programs investigated whether randomized assignment of treatment and control status over sub-districts inanaged to balance the key outcome variables. The overall conclusion is that it does, and that it allows for an impact evaluation based on experimental methods. We divided the analysis by type and relevance of the variables with respect to the objectives of both CT programs: Target service indicators that will describe the conditions for receiving the CCT benefits * Key health and ed~icationoutconle indicators that reflect the key priority areas for social policy in Indonesia Q Individual and household indicators Q Village characteristics Q Education and health facility characteristics reflecting quality of social service delivery For each of these categories, we found no systematic imbalances between treatment and control groups. However, there are a few points of attention for forthcoming impact evaluations. There are differences in height for age z-score between the PKH treatment and control group. In addition, for both CCT pilots we find some discrepancies between treatment and control groups in gross participation rate, school travel costs and domestic work activities. On the other hand, incidental statistically significant differences can be expected in a randomized design. The observed discrepancies are scarce and are not unambiguously in favor of either treatrner~t control groups, while other key health and or education outcomes seem balances. This would suggest that there is no systematic bias in education or health status. Nevertheless, controlling for initial differences would be prudent. But we do find a higher frequency of statistically significant differences when we look at the school and health facility surveys. Although these differences seem, with no clear pattern in the imbalances, they do need to be taken into account for the impact evaluation. It concerns key variables on quality and supply of services (such as materials and supplies at health care facilities, the cost of midwife services, and national exam results) that could affect impact of the CCT programs, and drive impact heterogeneity. For the follow-up survey and subsequent impact evaluation, we can add to following notes and recomnn~endations: It would be advisable to conduct a follow-up survey in the period after the start of a new academic year. This would facilitate estimating the programs' effect on transition rates. Household decisions regarding the transition from primary to secondary school manifest themselves during the period around the start of the new academic year. To conduct a follow-up survey in the period June to August, as is the case with the baseline survey, would make it unnecessarily difficult (if not impossible) to translate answers from households to school transitions in a consistent manner. PIG3 impact evaluation poses some problems, given that we do not know latent eligibility status anlong households in control areas. Unlike CCT programs in other countries, program eligibility was not dete~lnined households in control groups. Moreover, the for process of establishing eligibility was based on statistical grounds with some degree of local influence. Therefore, while there exists an experimental counterfactual for the full sample in treatment sub-districts (i.e. eligible and non-eligible households), we do not have one for the eligible households alone. With the availability of high quality baseline data it should be possible to einploy non-experimental methods (such as combining double difference and matching methods) to resolve this problem. However, this would mean that we ignore the main feature of the survey, the randomized design. Randomization gives us an unbiased estimate of the average overall effect of the program on the sainpled population, or the intention to treat effect. This observed effect would then be a weighted average of the direct effect of CCTs on participating households and the external effects for the whole population. A key empirical challenge that remains is to isolate the direct effect fiom the overall effect (i.e. the treatment effect on the treated) while utilizing the rand0111variation in treatment assignment. Since the PI= respondents are not representative for the sub-district populations, but were selected by means of purposive sampling, it is not possible to evaluate targeting performance of the housel~oldCCT. In so far as this is of interest, it may be worthwhile to sample extra households in the follow up survey, to get a representative image of the sub-district population. This additional sample could, for example be women and children that do not appear on the UCT roster, but would otherwise adhere to the CTT eligibility criteria relating to age, pregnancy, lactating mothers, etc. Of course, caution would be required wit11 including this extra sample for impact evaluation. While in principle this would be possible since randomization has seemingly removed any systematic differences between the treated and non-treaded district populations, absence of baseline data would not malte it possible to detect or correct any remaining coincidental discrepancies. REFERENCES Central Bureau of Statistics Indonesia, National Family Planning Coordinating Board, Ministry of Health, and MEASURE/DHS+ (2003). Indo~zesia 2002-2003 Denzograplzic ar7d Heultlz Szwvey Key Filldings. Jakarta. U ~ Z PKH Government of Indonesia (2007). P e d o ~ ~ z aI~ ~ L I ? ~2007 Handa S, Davis B (2006). "The Experience of Conditional Cash Transfers in Latin Policy Review 24 (5): 515-536. America and the Caribbean", Developnze~~t Skoufias, Emmanuel and Susan W. Parker (2001). Conditional Cash Transfer and Their Impact On Child Work and Schooling: Evidence from the Progresa Program in Mexico, FCND Discussion Paper No. 123, Illtemational Food Policy Research Institute. UNICEF (2006). State of the World's Clzildvelz 2006: Excluded and Invisible World Bank (2006). Making the New Indonesia Work for the Poor. World Bank (2007a). Indonesia Co~nmunityConditional Cash Transfer Project Concept Note. World B a ~ k (2007b). TOR CCT (Phase I) Baseline Survey. Dasar Rumah Tangga CCT 2007 Metoclologi S~lrvei World Bank (2007~). PNPM Generasi Quarterly Updates November 2007 TABLES Table 1 Scenario of cash transfers for the household CCT program ........................................................................................................ 54 Table 2 Sample size .................................................................................................................................................................................. 55 Table 3 Nan-response rates for survey modules ....................................................................................................................................... 55 Table 4 Primary school net etrolment rates (percent) in CCT treatment areas and Susenas 2004 .......................................................... 56 Table 5 Junior secondary school net enrolment rates (percent) in CCT treatment areas and Susenas 200 57 Table 6 Transition to junior secondary school (percent) .........................57 Table 7 Percentage of children with complete inlnluniz . .................... 58.... Table 8 Self reported child morbidity. age 0 to 36 months m CCT treatment areas ........ 58 Table 9 Treatment sought for ARI and diarrhea in CCT treatment areas (percentage) ............................................................................ 59 Table 10 Prevalence and treatment of ARS. fever aud diarrhea. IDHS Survey 2002-2003 . ....................... .......................................... 59 Table 11 Nutritional status of children age 0 to 36 months in CCT treatment areas .............................................................................60 Table 12 Delivery assisted by doctor or midwife ilnnlunization in CCT treatment areas and SDHS Survey 2002-2003 (percent) ........61 Table 13 Antenatal care in CCT treatment areas ...................................................................................................................................... 61 Table 14 Percentage of mothers that received at least 90 iron pills during pregnancy. women age 16 to 49 in CCT treatment areas .... 62 ... Table 15 Postnatal vlslt m CCT treatment areas ....................................................................................................................................... 62 Table 16 Infant mortality rates in CCT treatment areas ........................................................................................................................... 63 Table 17 Infant mortality rates. D H S Survey 2003-2003 ........................................................................................................................ 63 Table 18 Target service coverage indicators for children age 6-12 years for household CCT treatment and control groups .................64 Table 19 Target service coverage indicators for children age 13-15 years for household CCT treatment and control groups ...............64 Table 20 Target service coverage indicators for children age 6-12 years for community CCT treatment and control groups................ 65 Table 21 Target service coverage indicators for children age 13-15 years for community CCT treatment and control groups ..............66 Table 22 Child work and cost of education for children age 6-12 years for household CCT treatment and control groups ...................67 Table 23 Child work and cost of education for children age 13-15 years for household CCT treatment and control groups ................. 67 Table 24 Language and math test scores for household CCT treatment and control groups (percentage of answers correct) ................67 Table 25 Child work and cost of education for children age 6-12 years for community CCT treatment and control groups ................. 68 Table 26 Child work and cost of education for children age 13-15 years for community CCT treatment and control groups ...............68 Table 27 Language and math test scores for community CCT treatment and control groups (percentage of answers correct) ..............69 Table 28 Target service coverage indicators for married women age 16 to 49 in household CCT treatment and control groups .......... 70 Table 29 Target service coverage indicators for children age 0 - 36 months for household CCT treatment and control groups ...........70 Table 30 Target service coverage indicators for married women age 16 to 49 in community CCT treatment and control groups ......... 72 Table 31 Target service coverage indicators for children age 0 . months for community CCT treatment and control groups..........72 36 Table 32 Health outcome indicators for children age 0-36 months for household CCT treatment and control groups ...........................74 Table 33 Health care utilization for household CCT treatment and control groups ................ 75 Table 34 Health outcome indicators for children age 0-36 months for co~~lnlunity treatment and control groups ......................... 76 CCT Table 35 Health care utilizatio~l community CCT treatment alld control groups ............................................................................. for 77 Table 36 Education and child work regressions. children 7 to 12 years ..................... Table 37 Education and child work regressions. childre11 13 to 15 years ................... Table 38 Test scores. children 7 to 15 years ....................... Table 39 Flealth target regressions. childr . ....................... .................. 84 Table 40 Health outcome regressions. children 0 to 36 months Table 41 Outpatient care regressions. all l~ouseholdmember Table 42 Target indicator regressions Tor married women ag .................................................................................................. 92 Table 43 Household characteristics for 1 1 Table 44 Household characteristics for community CCT treatment and contro Table 45 Village characteristics for household CCT treatment and control ............... Table 46 Village characteristics for con~munity CCT treatment and control ......................................................................................... 102 Table 47 Community health facility characteristics for household CCT treatment and control groups ................................................ 104 Table 48 Community health facility characteristics for community CCT treatment and control groups ........................................... 110 Table 49 Midwife characteristics for household CCT treatment and control groups ............................................................................. 119 Table 50 Midwife characteristics for community CCT treatment and control groups .......................................................................... 125 Table 51 School characteristics for household CCT treatment and control groups .............................................................................. 134 Table 52 School characteristics for community CCT treatment and control groups .............................................................................137 Table 1 Scenario of cash transfers for the household CCT program Support scenario Amount of transfer per household per year Fixed cash transfer Cash transfer for per household with a. Child age less than 6 years b. Pregnant or lactating mother c. Children of primary-school age d. Children of secondary-school age Average transfer per household Minimum transfer per l~ousel~old Maximum transfer per household Source: Republilc Indonesia, Pedoinan Umum PKH 2007 A.l Baseline survey sample Table 2 Sample size Household CCT Community CCT Treatment Control Treatment Treatment Control incentives no incent. Sub districts 180 180 100 100 100 Villages 1,369 1,354 768 768 777 Households 7,195 7,131 4,000 4,000 4,000 Individuals 36,801 36,762 16,446 16,375 16,739 3 Children ~nlder years 3,076 3,077 1,592 1,534 1,620 Children age 6-15 9,356 9,550 3,108 3,128 3,255 Mamed women age 16-49 7,408 7,365 3,617 3,580 3,597 Puskesmas 178 180 100 100 100 Junior high schools 507 507 277 283 287 Midwife 702 705 385 385 387 Table 3 Non-response rates for survey modules Survey module Non-response rate (%) Married woiuen age 16-50 years 0.22 Children age 6-15-~ears - Language test, children age 7-12 years Math test, children age 7-12 years Language test, children age 13-15 years Math test, children age 13-15 years Children age 0-36 months Community health facilities Midwives Schools A.2 Regional variation Table 4 Primary scl~ool enrolment rates (percent) in CCT treatment areas and net Susenas 2004 PICH PNPM Susenas 2004 Total 85.6 88.2 93.0 DKI Jalcarta 87.8 91.9 West Java 83.2 89.6 93.4 East Java 86.5 86.9 93.7 North Sulawesi 80.8 79.0 88.3 Gorontalo 91.0 87.1 88.9 East Nusatenggara 87.5 89.3 90.8 Urban 85.0 DKI Jakarta 87.8 West Java 84.2 East Java 87.9 North Sulawesi 72.3 Gorontalo 83.4 East Nusatetlggara 69.8 Rural 86.1 88.2 93.3 West Java 78.5 89.6 93.5 East Java 85.3 86.9 93.7 North Sulawesi 85.1 79.0 89.9 Gorontalo 97.9 87.1 89.3 East Nusatenggara 89.8 89.3 91.1 Table 5 Juuior secondary school net enrolment rates (percent) in CCT treatment areas and Susenas 2004 PISH PNPM Susenas 2004 TotaliNational 51.9 62.1 65.2 DKI Jakarta 34.7 76.1 West Java 41.4 72.2 61.7 East Java 60.2 64.4 67.1 North Sulawesi 50.2 75.0 67.9 Gorontalo 46.2 53.0 49.3 East Nusatenggara 39.9 48.4 43.3 urban 50.7 DKI Jakarta 34.7 West Java 41.7 East Java 58.9 North Sulawesi 49.4 Gorontalo 42.3 East Nusatenggara 61.3 Rural 53.3 62.1 60.1 West Java 40.1 72.2 54.4 East Java 61.4 64.4 62.1 North Sulawesi 50.6 75.0 67.4 Gorontalo 48.5 53.0 42.5 East Nusatenggara 36.5 48.4 36.4 Table 6 Transitiou to junior secondary school (percent) PIW PNPM National 2005/2006 a TotaliNational 80.1 87.8 79.7 DIG Jakarta 100.0 102.2 West Java 66.2 93.3 69.5 East Java 84.2 89.3 77.3 North Sulawesi 90.5 100.0 96.7 Gorontalo 97.5 90.4 95.6 East Nusatenggara 75.0 70.5 99.4 a) Source: Ministry of National Education Table 7 Percentage of children with complete immunization in CCT treatment areas and IDHS Survey 2002-2003 CCT Baseline Survey " IDHS Survey 2002-03 Complete Complete Complete immunization given immunization for immunization: age cl-~ildren 10 age BCG, Measles, months + 3 doses DPT & 3 doses Polio PIGI PNPM PKH PNPM DKI Jakarta 27.3 36.7 67.0 West Java 23.0 50.8 28.7 62.7 41.4 East Java 47.4 38.3 57.6 47.3 64.2 North Sulawesi 39.0 47.1 42.5 58.9 68.6 Gorontalo 56.6 27.0 64.2 34.4 56.6 East Nusatenggara 37.8 38.2 50.2 48.4 62.7 a) Percentage of children aged 0-36 months. b) Percentage of children aged 12-23 months. Table 8 Self reported child morbidity, age 0 to 36 months in CCT treatment areas Diarrhea Fever Cough Acute Respiratory DKI Jakarta West Java East Java North Sulawesi Gorontalo East Nusatenggara PNPM Generasi West Java East Java North Sulawesi Gorontalo East Nusatenggara Table 9 Treatment sought for ARI and diarrhea in CCT treatment areas (percentage) ARI treated Diarrhea treated PKH DIU Jakarta 70.2 48.6 West Java 40.9 38.0 East Java 79.4 65.0 North Sulawesi 78.3 53.9 Gorontalo 26.4 37.1 East Nusatenggara 67.9 56.1 PNPM Generasi West Java 65.2 54.9 East Java 69.1 65.7 North Sulawesi 91.7 55.5 Gorontalo 62.0 67.6 East Nusatenggara 44.0 50.4 Table 10 Prevalence and treatment of ARI, fever and diarrhea, IDHS Survey 2002- 2003 Percentage of children with Percentage of children with symptoms of ARI andlor ARI Fever Diarrhea fever got treatment National 7.6 25.9 11 56.8 DIU Jakarta 6.8 21.5 7.8 75.4 West Java 9.0 31.1 15.1 50.3 East Java 2.8 20.8 9.8 64.5 North Sulawesi 6.5 24.0 9.5 60.2 Gorontalo 13.8 32.6 12.2 41.0 East Nusatenggara 8.1 28.0 12.9 53.7 Table 11 Nutritional status of children age 0 to 36 months in CCT treatment areas Underweight (WAZ) Stunted (HAZ) Wasted (WHZ) Malnourished Severely Malnourished Severely Malnourished Severely malnourished malnourished mal~~ourished PKH DKI Jaltarta West Java East Java North Sulawesi Goro~italo East Nusatenggara PNPM Generasi West Java East Java North Sulawesi Gorontalo East Nusatenggara Table 12 Delivery assisted by doctor or midwife immunization in CCT treatment areas and IDHS Survey 2002-2003 (percent) PIG3 PNPM Generasi DHS Survey Total 61.8 59.5 66.3 DIU Jakarta 86.9 94.3 West Java 42.1 72.6 48.7 East Java 82.8 80.5 80.7 North Sulawesi 71.2 78.1 85.7 Gorontalo 60.7 51.6 48.8 East Nusatenggara 33.1 36.0 36.4 Table 13 Antenatal care in CCT treatment areas Average number of At least 4 antenatal visits antenatal visits (percent) PIG3 Total DIU Jakarta West Java East Java North Sulawesi Gorontalo East Nusatenggara PNPM Generasi Total 7.9 79.0 West Java 8.5 85.2 East Java 8.5 86.7 North Sulawesi 5.1 56.3 Gorontalo 8.4 71.7 East Nusatenggara 6.8 72.4 Table 14 Percentage of mothers that received at least 90 iron pills during pregnancy, women age 16 to 49 in CCT treatment areas Percent of mothers PICH PNPM Total 11.7 16.9 DKI Jakarta 10.6 West Java 8.7 17.3 East Java 14.3 15.1 North Sulawesi 10.8 11.2 Gorontalo 3.6 17.2 East Nusatenggara 10.2 18.1 Table 15 Postnatal visit in CCT treatment areas Average number of At least 2 postnatal visits postnatal visits (percent) PKH PNPM PKH PNPM Total 3.5 3.6 43.0 52.4 DKI Jakarta 3.0 73.1 West Java 3.2 4.3 56.3 67.3 East Java 4.7 4.7 45.9 53.5 North Sulawesi 3.2 1.8 40.4 34.5 GorontaIo 2.6 4.6 48.8 46.9 East Nusatenggara 1.6 1.9 24.8 40.9 Table 16 Infant mortality rates in CCT treatment areas Neonatal Post-neonatal Infant mortality mortality mortalitya PKH Total 41 40 81 DKI Jaltal-ta 0 0 0 West Java 47 35 82 East Java 35 46 82 North Sulawesi 57 9 66 Gorontalo 7 82 89 East Nusatenggara 43 45 88 PNPM Generasi Total 34 45 80 West Java 37 37 74 East Java 10 13 23 North Sulawesi 0 0 0 Gorontalo 40 106 145 East Nusatenggara 45 46 90 a) Computed as the difference between the infant and neonatal mortality rates. Table 17 Infant mortality rates, IDHS Survey 2003-2003 Neonatal Post-neonatal Infant mortality mortality mortality DKI Jakarta 18 17 35 West Java 25 19 44 East Java 28 14 43 North Sulawesi 16 9 25 Gorontalo 24 54 77 East Nusatenggara 31 28 59 A. 3 Mean comparison tests education and health Table 18 Target service coverage indicators for children age 6-12 years for household CCT treatment and control groups Variable Treatment Control Difference P NT NC Enrolled 0.8871 0.8694 0.0177 0.16 6715 6849 school Enrolled in prin~ary 0.8214 0.816 0.0054 0.65 6715 6849 Enrolled in secondary school 0.0653 0.0528 0.0125 0.14 6715 6849 Net enrolment primary school 0.8556 0.8623 -0.0067 0.57 5780 5901 at age 7 to 12 Gross participation rate 0.9315 0.9251 0.0065 0.52 5780 5901 At age 7 to 12 Primary school new pupil absorption at age 7 (in persons 1.3188 0.943 1 0.3757 0.24 180 180 avrg) Attend school 85% last 2 0.945 0.9415 0.0034 0.71 5094 5232 weeks Attend school 85% last week 0.9425 0.9467 -0.0042 0.73 4299 4154 Attend school 100% last 2 0.9104 0.9222 -0.0118 0.34 5094 5232 weeks Attend school 100% last week 0.9414 0.9447 -0.0033 0.79 4299 4154 Note: Results reflect fractions unless stated otherwise. Table 19 Target service coverage indicators for children age 13-15 years for household CCT treatment and control groups Variable Treatment Control Difference P NT NC Enrolled in primary school 0.1068 0.1504 -0.0436 0.01 2235 2236 Enrolled in secondary school 0.5724 0.5909 -0.0185 0.56 2235 2236 Net enrolment junior secondarv school Gross participation rate (age 13 to 15) Transition to'secondary school 0.8013 0.8352 -0.0338 0.21 1736 1729 Attend school 85% last 2 0.941 0.9175 0.0235 0.22 1316 1355 weeks Attend school 85% last week 0.9424 0.9377 0.0048 0.79 1133 1097 Attend school 100% last 2 0.9185 0.892 0.0265 0.25 1316 1355 weeks Attend school 100% last week 0.9423 0.937 0.0053 0.77 1133 1097 Note: Results reflect fractions unless stated otherwise. Table 20 Target service coverage indicators for children age 6-12 years for community CCT treatment and control groups Variable Treatment Treatment Control Difference IT Difference r, N N N I NI I-C NI-C I NI C Enrolled 0.8809 0.8797 0.881 1 -0.0002 0.99 -0.0014 0.94 2299 2326 2428 Ellrolled in priillary school 0.8145 0.8182 0.8347 -0.0202 0.27 -0.0165 0.3 2299 2326 2428 Enrolled in secondary school 0.0652 0.0615 0.0463 0.0189 0.11 0.0153 0.16 2299 2326 2428 Net enrolmei~t prilnary 0.8827 0.8817 0.888 -0.0053 0.74 -0.0063 0.68 1953 1976 2081 school Gross participation rate 0.9625 0.9547 0.9422 0.0203 0.08 0.0124 0.31 1953 1976 2081 (age 7 to 12) Primary school new pupil 1.1492 1.2545 1.1191 0.03 0.94 0.1354 0.74 97 95 96 absorption at age 7 (in persons avrg) Attend school 85% last 2 weeks 0.9171 0.9266 0.9385 -0.0214 0.34 -0.0119 0.48 1768 1751 1825 Attend school 85% last week 0.9288 0.9444 0.9372 -0.0085 0.58 0.0071 0.6 1545 1535 1485 Attend school 100% last 2 weeks 0.8747 0.8987 0.9189 -0.0441 0.06 -0.0202 0.27 1768 1751 1825 Attend school 100% last week 0.9261 0.9379 0.937 -0.0108 0.48 0.0009 0.95 1545 1535 1485 Note: Results reflect fractions unless stated otherwise. Table 21 Target service coverage indicators for children age 13-15 years for community CCT treatment and control groups Variable Treatment Treatment Control Difference p Difference p N N N I NI I -C NI - C I NI C Enrolled in urimarv school 0.1734 0.1621 0.1311 0.0423 0.2 0.0309 0.37 682 681 692 Ellrolled in secondary school 0.676 0.7021 0.6983 -0.0223 0.62 0.0038 0.94 682 681 692 Net enrol~llentjunior 0.5986 0.6424 0.6153 -0.0167 0.71 0.0271 0.56 682 681 692 secondary school Gross participation rate 0.851 1 0.8646 0.8315 0.0195 0.5 0.0331 0.28 682 681 692 (age 13 to 15) Transition to secolldary 0.8609 0.894 0.8814 -0.0205 0.59 0.0126 0.71 563 554 570 school Attend school 85% last 2 0.9445 0.9046 0.8985 0.0459 0.07 0.0061 0.85 460 485 504 weeks Attend school 85% last 0.8989 0.9224 0.8919 0.007 0.82 0.0305 0.29 400 443 421 week Attend school 100% last 2 0.923 1 0.8892 0.8801 0.043 0.12 0.0091 0.78 460 485 504 weeks Attend school 100% last 0.8982 0.9224 0.8919 0.0063 0.84 0.0305 0.29 400 443 421 week Note: Results reflect fractions unless stated otherwise. Table 22 Child worlc and cost of education for children age 6-12 years for household CCT treatment and control groups Variable Treatment Control Difference P NT NC Economic work activities 0.0679 0.0796 -0.0117 0.41 6715 6848 Domestic work activities 0.5443 0.5446 -0.0003 0.99 6675 6827 20 hrs work (non domestic) 0.0262 0.0283 -0.0021 0.76 6715 6848 20 hrs work (all) 0.1669 0.1794 -0.0124 0.55 6715 6849 Scholarship in last 2 years 0.0427 0.0514 -0.0087 0.31 5910 5981 Travel time to school (minutes) 15.61 15.20 0.41 0.67 5884 5968 Travel costs to school (Rp.) 129.56 96.62 32.93 0.35 5806 5804 Note: Results reflect fractions uiiless stated otherwise. Table 23 Child work and cost of education for children age 13-15 years for household CCT treatment and control groups Variable Treatment Control Difference P NT NC Economic work activities 0.2201 0.2227 -0.0026 0.93 2234 2235 Domestic worlc activities 0.7683 0.7146 0.0537 0.03 2219 2225 20 hrs wol-lc (11011 domestic) 0.1408 0.127 0.0138 0.51 2234 2235 20 hrs worlc (all) 0.4111 0.3866 0.0245 0.41 2235 2235 Scholarship in last 2 years 0.0841 0.0742 0.0099 0.59 1527 1565 Travel time to school (minutes) 20.59 21.38 -0.79 0.5 1515 1558 Travel costs to school (Rp.) 665.39 430.20 235.20 0.09 1492 1517 Note: Results reflect fractions unless stated otherwise. Table 24 Language and math test scores for household C C T treatment and control groups (percentage of answers correct) Variable Treatment Control Difference P NT NC Age7-12 Bahasa Math Total Age 13- 15 Bahasa 55.12 54.28 0.84 0.49 1061 1080 Math 52.58 52.38 0.2 0.89 1004 1037 Total 54.03 53.56 0.46 0.7 996 1024 Table 25 Child work and cost of education for children age 6-12 years for community CCT treatment and control groups Variable Treatment Treatment Control Difference p Difference p N N N I NI I-C NI-C I NI C work activities Ecoi~oillic 0.089 0.1022 0.104 -0.015 0.52 -0.0018 0.94 2299 2326 2428 Doil~esticwork activities 0.6066 0.6074 0.6499 -0.0433 0.18 -0.0425 0.14 2288 2323 2424 20 hrs work (non domestic) 0.0269 0.0401 0.0295 -0.0027 0.73 0.0105 0.29 2299 2326 2428 20 hrs work (all) 0.1669 0.1907 0.1684 -0.0015 0.94 0.0223 0.35 2299 2326 2428 Scholarsl~ip last 2 years in 0.0393 0.0354 0.0373 0.002 0.85 -0.0019 0.84 2042 2060 2155 Travel time to school (minutes) 15.49 16.49 16.26 -0.77 0.55 0.24 0.88 2039 2054 2150 Travel costs to school (Rp.) 164.28 156.38 139.56 24.72 0.55 16.82 0.63 2025 2033 2125 stated otheiwise. Note: Results reflect &actions ~~illess Table 26 Child work and cost of education for children age 13-15 years for community CCT treatment and control groups Variable Treatment Treatment Control Difference p Difference p N N N I NI I-C NI-C I NI C Ecoiiomic work activities 0.3032 0.2739 0.2552 0.048 0.31 0.0188 0.65 682 681 691 Domestic work activities 0.8556 0.8739 0.8079 0.0477 0.13 0.066 0.03 677 676 686 20 hrs work (non domestic) 0.1554 0.128 0.1097 0.0457 0.13 0.0183 0.48 682 681 691 20 hrs work (all) 0.4645 0.4178 0.4135 0.0509 0.26 0.0042 0.92 682 681 692 Scholarship in last 2 years 0.0713 0.0864 0.0982 -0.0269 0.33 -0.0118 0.68 562 566 577 Travel time to school (minutes) 25.16 22.67 21.01 4.15 0.09 1.67 0.33 560 563 575 Travel costs to school (Rp.) 978.49 635.86 670.22 308.27 0.07 -34.36 0.74 553 561 572 Note: Results reflect fractions unless stated otherwise. Table 27 Language and math test scores for community CCT treatment and control groups (percentage of answers correct) Variable Treatmellt Treatment Coiltrol Difference p Difference p N N N I NI I-C NI-C I NI C Age7-12 u Bahasa Math Total Age 13 - 15 Bahasa Math Total Table 28 Target service coverage indicators for married women age 16 to 49 in household CCT treatment and control groups Variable Treatment Control Difference P NT NC Nr. of antenatal visits 6.6033 6.6356 -0.0323 0.93 2842 2863 At least 4 antenatal visits 0.6907 0.685 1 0.0056 0.83 2842 2863 At least 90 iron pills given 0.1175 0.1242 -0.0067 0.75 2330 2308 Delivery assisted by doctor or midwife 0.6183 0.5738 0.0446 0.32 2323 2346 Nr. of postnatal visits 3.4825 3.6066 -0.1241 0.81 2323 2346 At least 2 postnatal visits 0.4301 0.4404 -0.0104 0.77 2323 2346 Note: Unit of analysis is each birth andor pregnancy per woman in the 24 month prior to the survey. Results reflect fractions unless stated otherwise. Table 29 Target service coverage indicators for children age 0 - 36 months for household CCT treatment and control groups Variable Treatment Control Difference P NT NC Complete iinmunization given age Complete inm~~nization for children age 10 months + Complete immunization BGC immunization Polio: 1 imnlunizatioll Polio: 2 ilnmu~lizations Polio: 3 i~nn~unizations Polio: 4 i~n~~lunizations DPT: 1 imlnullization DPT: 2 iinlnu~~izations DPT: 3 innnunizations Measles immunization Hepatitis B: 1 iin~nunization Hepatitis B: 2 immunizations Hepatitis B: 3 immunizations Not weighed in last two months 0.269 0.2769 -0.0078 0.78 2934 2930 Weighed once in last two month 0.4001 0.4388 -0.0387 0.13 2934 2930 Weighed at least twice in last two ~nonths 0.3309 0.2544 0.0465 0.08 2934 2930 Nr. of tillles weighed in last two months (in freq. unit) 1.0952 1.0421 0.0531 0.3 2934 2930 Receiving vitamin A of at 0.3581 0.384 -0.0259 0.36 2241 2198 least 2 per year during age 6 months - 5 years Nr. of times child received vitamin A (in freq. unit) 1.7361 1.8796 -0.1435 0.05 2877 2826 Nr. of opportunity to receive vitamin A (in &eq. unit) 3.6071 3.6616 -0.0544 0.48 3019 3013 Rate of uptake of vitaluin A 0.4795 0.52 -0.0405 0.11 2824 2765 from the official distribution Note: Results reflect &actions unless stated otherwise. Table 30 Target service coverage indicators for married women age 16 to 49 in community CCT treatment and control groups IJariab le Treatment Treatment Control Difference 11 Difference o N N N I NI I-C NI-C I NI C Nr. oraoteuatal visits 7.9407 7.7726 7.8945 0.0462 0.92 -0.1219 0.79 1682 1688 1702 At least 4 alltellatal visits 0.796 0.7852 0.7815 0.0145 0.69 0.0037 0.92 1682 1688 1702 At least90 iron pills given 0.1654 0.173 0.1934 -0.0281 0.34 -0.0205 0.49 1443 1465 1467 Delivcry assisted by doctor or midwife 0.5803 0.6095 0.6255 -0.0452 0.44 -0.0161 0.78 1346 1329 1357 Nr. of postilatal visits 3.6101 3.5374 3.1866 0.4235 0.31 0.3508 0.33 1346 1329 1357 At least 2 post~~atal visits 0.5105 0.5381 0.4834 0.027 0.48 0.0547 0.14 1346 1329 1357 Note: Unit of ailalysis is each birth andlor pregllailcy per woman in the 24 month prior to the survey. Rcsults reflect fractions unless stated otl~erwise. Table 31 Target service coverage indicators for children age 0 - 36 months for community CCT treatment and control groups Variable Treatment Treatment Control Difference p Difference p N N N I NI I -C NI-C I NI C Complete immunization 0.4302 0.3979 0.4289 0.0012 0.97 -0.031 0.46 1509 1422 1525 given age Complete immunization for 0.384 0.3424 0.3584 0.0256 0.44 -0.016 0.67 1535 1450 1558 children age 10 months + Complete inlmunization 0.5475 0.4846 0.53 11 0.0164 0.73 -0.0465 0.4 901 844 896 BGC immunization 0.832 0.8407 0.8472 -0.0152 0.56 -0.0064 0.83 1558 1480 1586 Polio: 1 imnlunization 0.8178 0.8436 0.8509 -0.033 0.22 -0.0073 0.79 1555 1473 1582 Polio: 2 immunizations 0.717 0.7173 0.74 -0.023 0.47 -0.0227 0.54 1540 1460 1568 Polio: 3 immunizations 0.6296 0.6102 0.6239 0.0057 0.88 -0.0137 0.75 1533 1452 1562 Polio: 4 immunizations 0.5238 0.4788 0.4856 0.0382 0.27 -0.0069 0.86 1531 1448 1558 DPT: 1 immunization 0.7466 0.7396 DPT: 2 immunizations 0.668 0.6488 DPT: 3 immunizations 0.5849 0.5563 Measles imm~u~ization 0.6075 0.5868 Hepatitis B: 1 im~nullization 0.7425 0.744 Hepatitis B: 2 in~~~~u~~izations 0.6261 0.6106 Hepatitis B: 3 immunizations 0.5475 0.5192 Not weighed in last two ~i~onths 0.2316 0.1959 Weighed once in last two month 0.3485 0.3226 Weighed at least twice in last two months 0.4199 0.4814 Nr. of times weighed in last 1.2451 1.3357 two months (in freq. unit) Receiving vitamin A of at 0.3978 0.3682 least 2 per year during age 6 months - 5 years Nr, of times child received 1.7218 1.6169 vitamin A (in freq. unit) Nr. of opportunity to receive 3.4763 3.4206 vitamin A (in freq. unit) Rate of uptake of vitamin A 0.4697 0.4466 from the bfficial distribution Note: Results reflect %actions unless stated otherwise. Table 32 Health outcome indicators for children age 0-36 months for household CCT treatment and control groups Variable Treatment Control Difference P NT NC Diarrhea last month 0.2885 0.2614 0.0271 0.25 3075 3075 Diarrhea treated High fever last month Cough last month Cough and rapid breath ARI last month ARI treated Illness last month Diarrhea or ARI Weight-for-age: not malnourished Weight-for-age: malnourished Weight-for-age: severely malnourished Height-for-age: not malnourished Height-for-age: malnourished Height-for-age: severely malnourished Weight-for-height: not malnourished Weight-for-height: malnourished Weight-for-height: severely malnourished Height (~111) Weight (kg) Mortality rate per 1000 live births Neonatal mortality 41 27 14 0.14 3210 3239 Infant mortality 81 54 27 0.04 3258 3283 Note: Results reflect fractions unless stated othe~wise. CCT Table 33 Health care utilization for l~ousel~old treatment and control groups Variable Treatment Control Difference P NT NC Nr. of outpatient visits Public 0.1106 0.1039 0.0068 0.36 36801 36762 Private 0.0388 0.0378 0.0009 0.84 36801 36762 Public and private 0.1494 0.1417 0.0077 0.39 36801 36762 Traditional 0.0044 0.0038 0.0007 0.55 36801 36762 Puskesnlas 0.0479 0.047 0.0009 0.85 36801 36762 Contact rate Public 0.0917 0.0842 0.0075 0.18 36801 36762 Private 0.0337 0.03 0.0037 0.3 36801 36762 Public and private 0.1214 0.1106 0.0108 0.09 36801 36762 Traditional 0.0034 0.003 1 0.0003 0.67 36801 36762 Puskesmas 0.0394 0.0384 0.001 0.79 36801 36762 Use of modern 0.6319 0.6302 0.0017 0.94 7516 7471 contraceptives amongst nlarried women Note: Results reflect fractions unless stated otherwise. Table 34 Health outcome indicators for children age 0-36 months for community CCT treatment and control groups Varial-rle Treatment Treati~leilt Coiltrol Difference p Differelice p N N N I-C NI-C Diarrhea last moiith 0.2506 0.047 0.031 0.3 1 Diarrliea treated I-Iigh fever last month Cough last month Cough and rapid breath ARI last month ARI treated Illness last month Diarrliea or ARI Weight-for-age: not nialtiourished Weight-for-age: n~alnourished Weight-for-age: severely malnourished Height-for-age: not malnourished Height-for-age: malnourished I-Ieight-forage: severely malnourished Weight-for-height: not malnourished Weight-for-height: malnourished Weight-for-height: severely malnourished Height (cin) Weight (kg) Mortality rate per 1000 live births Neonatal mortality 30 39 45 -15 0.26 -6 0.62 1633 1616 1698 Infant mortality 67 91 75 -8 0.61 16 0.37 1656 I634 1720 ls Note: R e s ~ ~ lreflect fractions unless stated otherwise. Table 35 Healtli care utilizatiot~for c o m m u ~ ~ iCCT treatment and control groups ty Variable Treatment Treatment Control Difference p Differeuce p N N N I NI I-C NI-C I NI C Nr. of outpatient visits Public 0.1223 0.1341 0.1302 -0.0079 0.49 0.0039 0.78 16446 16375 16739 Private 0.0606 0.0593 0.0715 -0.011 0.28 -0.0123 0.21 16446 16375 16739 Public and private 0.1828 0.1934 0.2017 -0.0189 0.24 -0.0083 0.62 16446 16375 16739 Traditional 0.008 0.0129 0.0123 -0.0044 0.43 0.0006 0.92 16446 16375 16739 Puskesmas 0.0557 0.0623 0.063 -0.0073 0.34 -0.0008 0.93 16446 16375 16739 Contact rate Public 0.0973 0.1056 0.1017 -0.0043 0.58 0.0039 0.69 16446 16375 16739 Private 0.0482 0.0463 0.0553 -0.0072 0.28 -0.0091 0.16 16446 16375 16739 Public and private 0.1387 0.1443 0.1483 -0.0097 0.33 -0.004 0.7 16446 16375 16739 Traditional 0.0063 0.0086 0.0078 -0.0015 0.57 0.0009 0.76 16446 16375 16739 Puskesmas 0.0461 0.0512 0.05 -0.0039 0.53 0.0013 0.86 16446 16375 16739 Use of modem contraceptives amongst married women 0.5744 0.563 1 0.5655 0.0088 0.74 -0.0025 0.93 3686 3656 3676 Note: Results reflect fractions unless stated otherwise. A.4 Regression results Table 36 Education and child work regressions, children 7 to 12 years (1) (2) (3) (4) (5) (6) (7) (8) PNPM: PRH: PNPM: PEW PNPM: PKI-I: PNPM: PKI-I: Gross Gross 85% att. 85% ail, Economic Economic Domestic Dotiiestic particip. particip. last week last week work work wol-k work Age -0.0001 -0.0001 -0.0048+ -0.0010 0.0225** 0.0234** 0.0654*" 0.0646'" [O.OOOl] [0.0001] [0.0025] [0.0016] [0.0036] [0.0027] Female 0.0004 -0.0002 -0.0133 -0.0303*' 0.2463** 0.2671"" [0.0004] [0.0002] [0.0084] [0.0055] [0.0120] [0.0089] Agriculture maill profession of liead of household -0.0000 0.0004 0.0228* 0.0030 0.0340* 0.0156 [0.0005] [0.0003] [0.0101] [0.0067] [0.0144] [O.Ol09] Roral village 0.0001 -0.0005 0.0108 -0.0059 0.0201 -0.0060 [0.0007] [0.0004] [0.0150] [O.OOSS] [0.0213] [0.0144] Quinlile 2 0.0008 0.0000 -0.0143 -0.0074 -0.0220 -0.0023 [0.0007] [0.0004] [0.0136] [0.0089] [0.0193] [0.0145] Quintile 3 0.0006 -0.0001 0.0150 -o.o1s1* 0.0240 0.0052 [0.0007] [0.0004] [0.0139] [0.0091] [0.0197] [0.0148] Quintile4 0.0006 -0.0002 0.0277+ 0.0055 0.0089 0.0260+ [0.0007] [0.0004] [0.0145] [0.0091] [0.0206] [0.0148] Quintile5 0.0005 -0.0007+ 0.0268+ 0.0156+ -0.0098 0.0121 [0.00071 [0.0004] [0.0150] [0.0095] [0.0214] [0.0154] Primary education, head of household -0.0005 0.0003 0.0243+ -0.0084 -0.0300 0.0387"* [0.0007] [0.0003] [0.0137] [0.0065] [0.0195] [0.0105] Junior secondary education, head of household -0.0001 0.0005 0.0018 -0.0101 -0.0230 0.0596** [0.0008] [0.0004] [0.0168] [0.0094] [0.0240] [0.0154] Senior secondary education, head of household 0.0001 0.0005 -0.0180 -0.0318* -0.0053 0.0323 [0.0008] [0.0005] [0.0170] [0.0123] [0.0241] [0.0200] Higher education, head of household 0.0000 0.0007 -0.0096 -0.03 11 -0.0850* 0.0264 [0.0012] [0.0021] [0.0246] [0.0509] [0.0349] [0.0825] Household size 0.0001 -0.0000 0.0064* 0.0015 -0.0102" -0.0104** [O.OOOl] [O.OOOl] [0.0028] [0.0017] [0.0040] [0.0027] Fraction of household age 0 to 2 years -0.0007 -0.0004 -0.0239 0.0260 0.0340 0.1 192" [0.0023] [0.0013] [0.0474] [0.0321] [0.0674] [0.0521] Fraction of household age 3 to 6 years -0.0477 0.2188"* [0.0484] [0.0475] Fraction of household age 7 to 15 years -0.0185 -0.1062** [0.0431] [0.0407] in Scl~olarsl~ip last 2 years -0.0025 0.0449X: [0.0214] [0.0201] I'ercenl BRKBN poor in village -0.01 14 -0.0443* [0.0209] 10.021I] Nunlber of primary schools -0.0036 -0.0066' [0.0033] [0.002S] Nunlber ofjunior secondary schools 0.0111+ 0.0013 [0.0061] [0.0052] Percent of shldents with scholarships in school -0.0071 -0.0248 [0.0182] [0.0187] Nunlber of students per classroom in grade 1 0.0002 0.0002 [0.0011] [0.0009] in Nunlber of students per classroo~ll grade 2 -0.0003 0.0008 [O.OOlO] [O.OOlO] per Number of sh~dents classroolll in grade 3 -0.0002 -0.0023*" [0.0009] [o.ooos] Average grade for UN: Indonesian -0.0118+ -0.00S0+ [O.OOGS] [0.0043] Average grade for UN: Math 0.0009 0.0040 [O.OOSG] [0.0068] Average grade for UN: English 0.0137 0.0056 [0.0098] [0.0064] School absence without permissiol~ -0.541 1** -0.2699" [0.1638] [O. 12841 In-migration rate in sub-district -0.0739 [O. 14091 Out-migration rate in sub-district -0.0970+ [0.0555] Constant 1.0265** 0.0028 - . . . . .. . . . . .. R-squared 0.01 0.00 0.01 0.01 0.04 0.03 0.15 0.15 Note: Includes district fixed effects. Standard errors in brackets. +significant at 10%; * significant at 5%; ** significant at 1% Table 37 Education and child work regressions, children 13 to 15 years (1) (2) (3) (4) (5) (6) (7) (8) PNPM: PKH: PNI'M: PU-I: PNPM: PKI-I: PNPM: PKI-I: Gross Gross 85% 85% Economic Economic Domestic Domestic particill. particip. attendance attendance work wol-k work wok -- week week Age -0.0014 -0.0000 -0.0446** 0.0120 0.0105 0.0247% 0.0082 0.0260" [0.0009] [O.OOll] [0.0126] [0.0078] [0.0138] [0.0096] [0.0121] [0.0101] Female -0.0011 0.0015 -0.0071 0.0152 -0.1 114** -0.0755*"; 0.2046"* 0.2649"" [0.0014] [0.0016] [0.0182] [0.0112] [0.0199] [0.0137] [0.0173] [0.0145] main profession of liead of household Agriculh~re -0.0022 0.0020 -0.0521' -0.0131 0.0183 0.0014 -0.0035 0.0028 [O.OOlG] [0.0018] [0.0214] [0.0132] [0.0235] [0.0161] [0.0204] [0.0171] Rural village -0.0017 0.0007 0.0186 -0.0135 0.0125 -0.0417+ 0.0160 -0.0168 [0.0022] [0.0025] [0.0290] [0.0189] [0.0329] [0.0216] [0.0286] [0.0228] Quintile 2 -0.0002 -0.0025 -0.0088 0.0045 -0.0099 -0.0157 -0.0224 -0.0071 [0.0023] [0.0027] [0.0310] [0.0195] [0.0337] [0.0236] [0.0295] [0.0249] Quintile 3 -0.0038 -0.001 1 0.0282 -0.0141 -0.0178 0.0120 -0.0339 0.0055 [0.0023] [0.0026] [0.0318] [0.0191] [0.0342] [0.0234] [0.0298] [0.0247] Qui1itile4 -0.0006 -0.0009 0.0233 -0.0145 0.0068 0.0275 0.0139 0.0378 [0.0024] [0.0027] [0.0318] [0.0190] [0.0344] [0.0234] [0.0300] [0.0247] QuintileS 0.0005 0.0009 0.0482 0.0128 -0.0130 0.0156 -0.0032 0.0014 [0.0024] [0.0027] [0.0323] [0.0196] [0.0351] [0.0240] [0.0306] I0.02531 Primary education, head of household -0.0000 -0.0015 0.0463 -0.0287* -0.0029 -0.0111 0.0354 -0.0099 [0.0022] [0.0018] [0.0298] [0.0131] [0.0316] [0.0161] [0.0276] [0.0170] secondary education, head of household Ju~iior -0.0007 -0.0040 0.0748* -0.0356+ -0.0233 0.0007 0.0570+ -0.0241 [0.0027] [0.0027] [0.0363] [0.0200] [0.0397] [0.0240] [0.0346] [0.0255] Senior secondary education, head of household -0.0003 -0.0001 0.0280 -0.0121 -0.0714+ 0.0364 -0.0169 0.0091 [0.0027] [0.0035] [0.0374] [0.0254] [0.0400] [0.0305] [0.0348] [0.0322] Higher education, head of household -0.0161** 0.0092 -0.0298 0.0702 -0.0855 -0.2654+ 0.0149 -0.1132 [0.0039] [0.0176] [0.0532] [0.1398] [0.0577] [0.1553] [0.0501] [0.1639] Household size -0.0008+ -0.0001 -0.0025 0.0018 0.0041 0.0022 0.0070 -0.0089* [0.0004] [O.OOOS] [0.0063] [0.0034] [0.0062] [0.0042] [0.0054] [0.0044] Fraction of household age 0 to 2 years 0.0051 -0.0000 -0.0658 0.0500 -0.0274 0.0430 0.0300 0.1948* [0.0081] [0.0102] [0.1073] [0.0741] [0.1185] [0.0897] [0.1033] [0.0949] Fraction of household age 3 to G years -0.0081 -0.0022 -0.0377 -0.0495 -0.1295 -0.0406 0.0764 -0.0083 [0.0739] Fraction of household age 7 to 15 years 0.0152 [0.0603] Scholarship in last 2 years 0.0373 [0.0245] Percent BKKBN poor in village 0.0479 [0.0330] Nuinber oiprimary schools -0.0032 [0.0043] Nutiiber ofjunior secondary schools -0.0030 [0.0077] Perceni of students with scholarships in school -0.0302 [0.0290] Nuniber of students per classrooln in grade 1 -0.0002 [0.0014] Nuriiber ofstudetlts per classrooni in grade 2 -0.0007 [0.0016] Nuliiber of students per classroom in grade 3 0.0019+ [0.0011] Average grade for UN: Indonesian 0.0083 [0.0075] Average grade for UN: Math -0.0161 [0.0107] Average grade for UN: English 0.01 15 [0.0101] School absence without permission -0.0670 [O. 18901 In-migration rate in sub-district Out-migration rate in sub-district Constant -0.1870 [0.0150] [0.0167] [0.1987] [0.1228] [0.21941 /0.1473] [0.1910] 10.15561 Observations 1490 2559 1106 1863 1496 2567 1489 2563 R-squared 0.03 0.01 0.05 0.01 0.04 0.03 0.10 0.13 Note: Includes district fixed effects. Standard errors in brackets. -1- significant at 10%; * significant at 5%; "* significant at 1% Fraction of household age 7 to 15 years -0.0679+ -0.0556* [0.0351] [0.0282] Percent BRKBN poor in village 0.0052 -0.0288+ [0.0178] [0.0148] Nuniber of primary schools 0.0027 0.0014 [0.0027] [0.0019] Stiiiibci. o ~ i i ~ ~ i sccoliil;~ry ioi. scliouls 0.0031 0.0025 [0.0048] [0.0035] of Percei~t students rvith scholarships in scl~ool -0.0227 0.0057 [0.0156] [0.0130] Number of students per classroom in grade 1 -0.0007 -0.0005 10.00091 10.00061 Number of students o e ~ - class~oom glade 2 in t0.001 i '0.000i [0.0009] Nu~liber stude~its classroon~ g a d e 3 of per in 0.0002 [0.0007] Average grade for UN: Indonesian 0.0048 [0.0056] Average grade for UN: Math 0.0135" [0.0066] Average grade for UN: English -0.0203** [0.0078] School absence without pernlission 0.2893+ [O. 15401 In-migration rate in sub-district 0.0508 [O. 11031 Out-migration rate in sub-district 0.1148" [0.0456] Constant 0.4945'* 0.4962** [0.0483] [0.0378] Obse~.vations 2892 4473 R-squared 0.09 0.06 Note: Includes district fixed effects. Standard errors in brackets. +significant at 10%; * significant at 5%; ** significant at 1% Table 39 Health target regressions, children 0 to 36 months (1) (2) (3) (4) (5) (6) PNPM: PKH: Con~plete PNPM: PKH: Weighed PNPM: PKH: Vitamin Complete immunization Weighed 2+ 2+ Vitanlin A A uptake rate iu~lnunizatio~~ for age uptake rate -- for age ;\gc 0,,,32>" o,,S'5:';:c .O,OqjS"" .0,"4Sq:::' o,o(j,, :*:a 0.0102+ [0.0113] [0.0096] [0.0112] [0.0097] [0.0110] [0.0104] Female 0.0030 0.0160 0.0169 0.0294' -0.0062 0.01 16 [0.0160] [0.0147] [0.0160] /0.0149] [0.0157] [0.0158] professio11of head oT11ousehold Agricultore tnai~i -0.0256 -0.0195 -0.0107 -0.0259 0.0247 0.01 13 [0.0192] [0.0180] [0.0191] [0.0183] [0.0188] [0.0193] Rural village -0.0303 -0.0002 -0.0403 0.0276 -0.1016** 0.0621* [0.0300] [0.0224] [0.0307] [0.0230] [0.0294] [0.0243] Quintile 2 0.0306 -0.0186 -0.0017 0.0174 -0.0382 -0.0410+ [0.02511 [0.0226] [0.025 I] [0.0228] [0.0247] [0.0244] Qiiintile 3 0.0022 0.0246 0.0226 0.0322 -0.0124 0.0151 [0.0260] [0.0233] [0.0259] [0.0236] [0.0256] [0.0249] Quintile4 0.04911- -0.0218 0.0040 -0.0080 0.0435 -0.0090 [0.0272] [0.0240] [0.0273] [0.0244] [O.O26S] [0.0260] Qtii11tile5 0.0450 -0.0104 -0.0199 -0.0278 -0.0150 -0.0120 [0.0291] [0.0253] [0.0291] [0.0254] [0.0286] [0.0271] Primary education, head oil~ouseliold 0.0309 0.0183 0.0455+ 0.0254 -0.0166 -0.0220 [0.0275] [0.0173] [0.0273] [0.0175] [0.0270] [0.0187] Junior secondary education, head of household 0.0789* 0.0429+ 0.1207*' 0.0377 -0.0171 -0.0225 [0.0319] [0.0253] [0.0318] [0.0256] [0.0314] [0.0273] Senior secoudary education, head of household 0.0971** 0.0128 0.1030** 0.0137 0.01 15 -0.0640+ [0.0334] [0.0316] [0.0332] [0.0322] [0.0328] [0.0341] Higher education, head of household 0.0559 0.0504 0.1166' -0.0432 -0.0565 -0.0612 [0.0492] [O. 14901 [0.0493] [0.1389] [0.0481] [O. 18031 Household size -0.0106+ -0.0005 -0.0097 -0.0125* -0.0108+ -0.0050 [0.0060] [0.0051] [0.0060] [O.OOSl] [0.0058] [O.OOSS] Fraction of household age 0 to 2 years -0.3923** 0.0829 -0.4347** -0.2677* -0.1387 -0.2315+ [0.1368] [0.1209] [0.1350] [O. 12251 [0.1332] [O. 13061 Fraction of household age 3 to 6 years -0.2625** -0.1 148 -0.3063** -0.2134"' 0.0295 0.1722* [0.0888] [0.0794] [0.0887] [0.0806] [0.0871] [0.0853] Fraction of household age 7 to 15 years Askeskin in Pelcent Askeskin l~ousel~olds a village Pexcenl RKKRN poor in village Nuinber of active integrated hcalth posts health C o l ~ u ~ ~ u n i t y center (or auxiliary) in village Nr of doctors providing health service in village Nr of (skilled) midwife providing health service Cost of visit to Puskesmas check by midwife at P uskesmas Cosl of ai~tenntal Cosl ofgcrlcral treatment by public lliidwife Cost of' antenatal check up by public midwife Cost of i~onnal delivery by public midwife Cost of general treatment by private midwife Cost of antenatal check up by private midwife delivery by private midwife Cost of l~ormal Average cost of child immunization by public midwife Average cost of child immunization by private midwife [0.0000] [0.0000] [0.0000] [0.0000] [0.0000] [0.0000] Crude birth rate in sub-district -4,710.89** -1,681.24 2,353.01 [1,807.40] [1,785.87] [1,769.29] 111-migrationrate in sub-district -0.2193 -0.4069+ -0.0389 [0.2364] [0.2389] [0.2352] Out-migration rate in sub-district -0.1024 0.3767*'U 0.0264 [0.10311 [O. 10341 jO.lOlZ] C:UI~SGIII~ U.369S - . ~ 0.24iU. .'~ 0.6152.- O.jjbu-. 0.4914.;.,. U.51Su.: [0.1046j [O.OS461 [0.1045] [0.08461 [O.l02S] [0.0909] Observations 3254 3692 3364 3796 3222 3612 R-squared 0.10 0.11 0.04 0.03 0.03 0.01 Note: Includes district fixed effects. Standard errors in brackets. +significant at 10%; * significant at 5%; ** sigllificallt at 1% Table 40 Health outcome regressions, children 0 to 36 months (1) (2) (3) (4) (5) (6) (7) (8) PNPM: PKH: PNPM: P ICH: PNPM: PIUI: PNPM: PKFI: Dial-rhea Diarrhea WAZ WAZ I-IAZ IlAZ WIIZ WIIZ ~ . -. ~ or ART or ART --. - . - - - . - . -. . .. . . .. - -. . . . 0,0280":; 0,0184~ , 3 9 9 : .0,.10j3+~ 0,~700:~2;: . : .0,[7j7:$: 0.0350 [O.OlOS] [0.0096] [0.0300] [0.0286] [0.0700] [0.0566] [0.0699] [0.0572] Female -0.0170 0.0093 0.1228** 0.1245** 0.2911** 0.3326** -0.0697 -0.0776 [0.0156] [0.0148] [0.0428] [0.0436] [0.1002] [0.0863] [0.0994] [0.0868] Agriculture maill profession of liead of liouseliold -0.0287 -0.0025 -0.0069 -0.0106 -0.0874 -0.1962i 0.0224 0.1637 [0.0187] [0.0182] [0.0512] [0.0534] [0.1197] [0.1059] [0.1188] [0.1064] Rural village -0.0244 -0.0131 -0.0553 -0.1043 -0.2690 -0.1281 0.0386 -0.2508+ [0.0297] [0.0228] [0.0810] [0.0667] [O. 19151 [0.1324] [O.l884] [0.1327] Quintile 2 -0.0242 -0.0223 -0.0135 0.1932** 0.3710" -0.0120 -0.3086* 0.1368 [0.0244] [0.0229] [0.0673] [0.0670] [0.1570] [0.1330] [0.1558] [0.1336] Quintile 3 -0.0290 -0.0159 0.0134 0.1164+ 0.0958 -0.1282 0.0058 0.1454 [0.0253] [0.0235] [0.0694] [0.0692] [0.1618] [0.1369] [0.1610] [0.1376] Quintilc4 -0.0237 0.0147 0.04 11 0.1595* 0.3789" 0.1509 -0.2186 0.0167 [0.0266] [0.0244] [0.0730] [0.0714] [0.1701] [0.1416] [0.1691] [0.1422] Quintile5 -0.0396 0.0310 0.0500 0.0875 0.2070 0.2455t -0.1805 -0.2288 [0.0284] [0.0255] [0.0780] [0.0748] [0.1819] [0.1478] [0.1816] [0.1489] Primary education, head of household -0.0430 -0.0055 0.0160 -0.0289 0.0527 -0.1065 0.0355 0.0222 [0.0265] [0.0175] [0.0731] [0.0512] [0.1713] [0.1018] [0.1698] [0.1019] Junior secondary education, head of household -0.0666* -0.0535% 0.1278 0.0099 0.0627 0.2150 0.1703 -0.2653t [0.0309] [0.0257] [0.0852] [0.0753] [0.1997] [0.1491] [0.1977] [0.1507] Senior secondary education, head of household -0.0933** -0.0250 0.1751" -0.0244 0.4056+ 0.2197 0.0406 -0.1101 [0.0323] [0.0325] [0.0890] [0.0955] [0.2085] [0.1886] [0.2067] [0.1902] Higher education, head of liousehold -0.1 168* -0.1739 0.2284+ 0.7059+ 0.5038 0.0878 -0.0609 0.3492 [0.0480] [0.1422] [0.1313] [0.4080] [0.3093] [0.8099] [0.3061] [0.8473] Household size -0.0132* -0.0068 0.0182 -0.01 12 0.0656+ 0.0262 -0.0303 -0.0332 [0.0058] [0.0051] [0.0160] [0.0152] [0.0373] [0.0300] [0.0370] [0.0303] Fraction of household age 0 to 2 years -0.0861 -0.2402* 0.1749 -0.6368+ -0.3331 -0.9412 0.5462 0.0828 [0.1315] [0.1214] [0.3616] [0.3600] [0.8498] [0.7131] [0.8369] [0.7224] Fraction of household age 3 to 6 years 0.1384 0.1126 -0.4621t -0.9122** -1.1488* -1.0810* 0.5963 -0.4086 [0.0865] [0.0802] [0.2362] [0.2359] [0.5521] [0.4663] [0.5482] [0.4688] Fraction of household age 7 to 15 years -0.5140 [0.4259] Askeskin -0.1004 [0.1271] Percent Askeskin households in a village 0.1574 [0.2691] I'erccnt l3KI . .. . II.IL.II. \ i\ii, 1\gc -U.OUUO -U.UUJ I -' 0.00 I S [0.0012] [0.0012] Agriculture main proression of head ~Eliot~sehold -0.0087 -0.0663";" [0.0168] [0.0178] Rmal village 0.0101 -0.0123 [0.0204] [0.0216] Quintile 2 0.0478' 0.0037 [0.0218] [0.0226] Quintile 3 0.0459" 0.0422+ [0.0224] [0.0233] Quintile4 0.0613** 0.0518* [0.0228] c0.02391 Quintile5 0.0497" 0.0307 [0.0239] [0.0251] Pri~iiarycclucation, head of liouseliold 0.0681"" 0.0449* [0.0196] [0.0207] Jill~ior secoodary education, head of household 0.1294** 0.1277** [0.0241] [0.0255] Seliior secondary education, head of houseliold 0.1660'" 0.1941** [0.0303] [0.0322] Higher education, head of household 0.1453 0.1811 [0.1420] [O. 15451 Household size 0.0106* -0.0014 [0.0041] [0.0050] Fraction of liousehold age 0 to 2 years 0.5817** -0.1796 [0.0743] [0.1155] Fraction of household age 3 to 6 years -0.2261** -0.2415*' [0.0707] [0.0783] Fraction of household age 7 to 15 years -0.0749 -0.1429* [0.0609] [0.0670] Women's decision making index: children Women's decision making index: household consuuiiption Askeskin ;\>kcsk111 l ~ ~ 1 l ~ ~ ~l l l ~\lll:lgc I'L.ILL~III l I l :I l l ~ l s Percent R R m N poor in village Numbcr of active integrated health posts health center (or auxiliary) in village Com~ilunity Nr of doctors providing health service in village Nuuiber of (skilled) midwife providing health service Cost of visit to Puskesmas Cost of antenatal check by midwife at Puskesnias Cost of general treatment by public midwife Cost of antenatal check up by public midwife Cost of normal delivery by public midwife Cost of general treatment by private midwife Cost of antenatal check up by private midwife Cost of normal delivery by private midwife Crude birth rate in sub-district - z i - i ,0 m* r r.,,,mwc- m e , i , ~ Q ,,,o.+oo, c 9 o u c - -"o:c;+c' e u - j? -2s A. 5 Mean comp:irison tests housel~olds, villages, health care providers and schools Table 43 flIoilseliiiitl characteristics for household CCT treatment and control Variable Treatment Control Difference Average age 0 - 36 nioiitllr (inoriil~s) 18.1036 18.4157 6 - 12 years iyenrsi 9.0286 9.0304 13 - 15 years (year\) 13.8186 13.7965 Full sanipie (!cars) 24.1488 23.4637 Female 0.4927 0.4926 Rural 0.528 0.5422 Education dcgrce obt.tiiicd (age>lO) None Primary Junior secondary Senior secoii\\iiedilia) Own radioltni~c recorder I O ~ Itelevisit>ii Ow11 parahol:~anteiiiin Own sho\vcosc'sidchoard Own refrigel-.itor Own bicycle >);iff Ow11 n~otorc!cleloii~hoard motor Own1 car/iiioia,r b o ; ~ ~ Own hand pli,~oe Own cliic1;eii iiiick Own pig Own aoal Own co\v;bii~ijlo Own horsc Commuilily pai:icipation Participaiioli III sociiil service group Participai~on prodiiction 111 0.052 0.0427 group Particip a t ioii :I>workci-s group 0.0341 0.0314 Participntiuii i i i nal. ~csource 0.0067 0.0062 managenlcni ;iou]~ Participaiioli i i i crciiii!finance 0.31 0.3458 group Participalioii i i i gavel-nmeiltal 0.0577 0.0625 group Participniioii iii religious1 0.6226 0.6395 traditional gri,np Participaiioli in recrenlional 0.0188 0.022 group Participaiioii i i i ~iiasspolitical 0.0082 0.01 14 organizaiioii Note: Resiilis I-eflccifractions u~lless stated otherwise. T a b l e 44 Household characteristics f o r community CCT t r e a t m e n t a n d control L7:!ri:!l>fc c ! ! i r i c ? Cociic-o! Dirfcrcricc IJ IDi fit-eiicc j i \ I' N I YI I i' I i' I Nl (~ Average age 0 - 36 ~iinntl~s(inoi?tlis) 17.2524 16.7113 17.1383 0.1441 0.81 -0.4241 0.53 1574 1516 1599 6 - 12 ycai-s (years) 8.95 11 S.9313 8.9649 -0.013s 0.86 -0.U236 U.73 2299 2326 242s 13 - 15 years (years) 13.8077 13.7165 13.8078 -0.0002 1 -0.0913 0.1 682 681 692 Full sample (years) 27.6592 27.6058 28.2724 -0.6132 0.42 -0.6667 0.39 14522 14470 14824 Fernale 0.5008 0.5045 0.5075 -0.0067 0.36 -0.003 0.68 16445 16375 16739 Education degree obtained (age> 10) None 0.2102 0.2238 0.2621 -0.0517 0.04 -0.0383 0.13 12162 12076 12336 Primaiy 0.47 0.4573 0.4237 0.0463 0.02 0.0336 0.07 12162 12076 12336 Junior secondary 0.1582 0.161 0.1479 0.0103 0.38 0.0131 0.29 12162 12076 12336 Senior secondary 0.1326 0.1292 0.138 -0.0054 0.76 -0.0088 0.61 12162 12076 12336 I-Iighcr 0.0289 0.0287 0.0283 0.0006 0.91 0.0003 0.95 12162 12076 12336 Per capita monthly expendit~lre Total (Rp.) 340273.16 325330.36 332882.64 7390.52 0.76 -7552.28 0.73 16446 16375 16739 Food (Rp.) 158009.45 136176.31 150632.58 7376.87 0.71 -14456.27 0.42 16446 16375 16739 Non food (Rp.) 18789.55 17853.85 15067.20 3722.35 0.26 2786.65 0.42 16446 16375 16739 Education (Rp.) 12355.84 10801.29 10389.24 1966.60 0.41 412.05 0.85 16446 16375 16739 Health (Rp.) 182263.71 189154.05 182250.06 13.64 1 6903.99 0.39 16446 16375 16739 Household size (persons) 4.42 4.4698 4.6028 -0.1828 0.11 -0.133 0.27 16446 16375 16739 Rural 0.1404 0.1348 0.1608 -0.0204 0.78 -0.026 0.7 16446 16375 I6739 Head of household Female 0.0527 0.0766 0.0716 -0.0188 0.01 0.005 0.56 16446 16375 16739 No education degree 0.1667 0.1796 0.224 -0.0571 0.04 -0.0444 0.1 16436 16356 16737 Primary 0.5166 0.5054 0.4614 0.0552 0.03 0.0439 0.07 16436 16356 16737 Junior secondary 0.1284 0.1392 0.1253 0.0031 0.84 0.0139 0.39 16436 16356 16737 Senior secondary 0.1493 0.1382 0.1487 0.0006 0.98 -0.0105 0.62 16436 16356 16737 Higher 0.0388 0.0377 0.0406 -0.0018 0.82 -0.0029 0.73 16436 16356 16737 Rice and secondary crops main profession Social progmmmeslinsurance .\..\l.~< .'.>l.~>l.lil SL'l'lBL'I Ioat 0.0377 0.0.:99 O\\ li,iiiiI i ~ l i i i i l i : 11 0.22-I: 0.3176 Own chickenlduck 0.5713 0.5742 O\vn pig 0.1395 0.15 Ow11 goal 0.203 0.1904 Ow11 cowv/buffalo 0.1484 0.1758 Owvi~ horse 0.0166 0.0203 Co111111ui1ity participalion in Participatioi~ social 0.3497 0.3313 service group Participation in productioil 0.1039 0.1089 group in Participatioi~ workers 0.0515 0.0557 group Participation in nat. resource 0.0153 0.0231 management group Participation in 0.297 0.3625 creditlfinance group Participation in 0.1225 0.1391 governmental group Participation in religious1 0.554 0.5826 traditional group Participation in recreational 0.0458 0.0397 group Participation in 0.0156 0.0203 masslpolitical organisation Note: Results reflect fractions unless stated otherwise. Table 45 \'illage characteristics for household CCT treatment and control Variable Treatment Control Difference P NT NC Population sire (persotis) 3829.5149 4022.7614 -193.2465 0.5 1366 1351 Number of sii!i-villages 4.1801 4.0791 0.1011 0.57 1369 1354 Number of lii~l~seliolds 1050.6421 1095.0403 -44.3982 0.58 1368 1351 Percentagc 01' liil~ilicsiii agricultural st .) 3 . Villagc suil;,~ Svoill 0.1761 0.1649 0.01 13 0.62 1369 1353 fueleartliqiia: ~Iotlier disasters in last 12 rnonil~ Village sul'k!. SI-0111 i~nsvest 0.4741 0.4646 0.0095 0.79 1369 1353 failure ill !as[ i 2 nioiitl~s Village sufli,! . lowcr iiico~iie due 0.3097 0.3039 0.0058 0.85 1369 1353 to croplseli j i : ,duct )price drop in last 12 iiionll; Average \\,a:. per day of 11011- 21553.7604 21486.409 67.351 1 0.94 1366 1350 skilled woi.l1l room Communily lie. '.tiilit) 113s: t ~ ~ e d i c i i ~ e room sewice Total budget l i b 2 , i l , ('1000 Rp.) . Number a f b a h .'i-! I ~iainths) given BCG vaccinativ Number of bab . given I -I I ~noiitiis) Polio vaccinattz.,~ - Number of baht . v-1 I i,?onths) given Hepatitis B vat t~ss, Numbel. o i biih: ' 1 - I I >n~w>ths) give,? DPT Hb Comh 3rc;iintioii Numbcr of bai~: 2-1 I inontiis) given DPTvuccinaiii~, . Number afbah: 3')-I i iiionths) givcii mcssle vaccin:i i Numbcr aipre: .; ! ;i~otllci.s givcn n ' vacciniltio~~ Numbcl- oiocn i ; 1hy pieg11301 imotlie~. (Kl) pure Number of ncn I .; by O I C S I Iiniothc~.~~~ (K4) Neonatal visit Baby 0-1 I ~iioib1. .\<,s$lied (babies) Child 12-35 mc. ! r <\cigl,ed (children) Child 36-59 m i I i\ci$l~ecl(children) % Baby 0-1 1 mm:'2 - -~nilci- dotted Iilie and above red line 8 , I ~.,i!!i moiiitoiiagca~~rl (babies) Child 12-35 ma I ' \ ilriclei dolled line and above red liiic c i .,hi, inonital-ing caid (children) . Child 36-59 mt. !: ~tt!de,-ilottcd linc and above red linc i I .ail11 inonitorinn caid u (childicn) Baby 0-1 I )not; 8 . ,;1~1cr line ofhc>lt$ md moniloi-ing car, I l,!,. I I a::~iiui.wdlinc aC heullii moiiitoit :..lid (ciiil~lren) Child 36-59 n w : i. 01ndcri-cd line of health monitoiv C . . : ~ I (cliiiiiiei~) Baby 0-1 I 11101: . ,, ::ii Ih~.n1111 monitoring card; matcmnl 3l.i ihi.;iItl, book (babics) Chili1 12-35 m, , B,, , , ~ > t lIhcallh > monitoring cat, !, st~.rm:dcl~ild Ihealtl, book (childici~ 1 Child 36-59 ni. ! I. ;, ttli iicnltli monitoi-iiig caw ,:..'rinal cliild health book (cliiliii.en C h iI : , ; x c n i i i f l i doseA vitamn (cl>ilclt Chili! 1-5 yeav. t, Ihhfil tiosc A vilazilin Motlcr in con1 : ; a : : give), IhigI~doscA --~ vita" ill (cliildi . Preet :tiit moiii. i , a 1:ililctr for blooii ~ ago (cl~ildicn) Nunihcr of l i b I., .~~;iil:iblr Numhciof fob, ' .tvlablu Numheiofva; I ,..ci!iiim availnblc Numl~eioitc~i. i t ' . ,:~i:iilablc Numl,ci of tile. \ W ' ~ ! Comiiablc Numlrci of g).! . I ':,?a1 table available Nurnl~cioisti.; I 1.1 cilwied clamps avail.oble Nuniher ofox! i , i:l!,tel- ;#\.ailable Nunibcrofinu. t., t\:iilabic Numbeiofni.: ' : : '.xi nvniiabic Number ofIBu~ .. .i.i.inccaciier available . Typc o f \ ~ a c c ~ n 8 , . -i. Ijciiity: Spccial coolingbox fii: Typ~ of\,accin :tcr'cald cliain . i . ;L. fnciiily: Frccrci Typr.i>fvaccitl .. .re ihcilily: Refit;.cmtol- Typc oivaccin ' , s . . :u taciiity: iVonc : Type i,fraccilt ' . ' . :c facility: Olhcl. Typt , i f s y r i i ~ ~, : . % t ivaccine injection: timi Nu, IW ofmi, ; , lull-tinic) , Nu, '?croin~ic' t~:,rt-~it>~c) ,;'L Nul. 1hcr ofvill s in,ln.ifc (Ii~ll-time) Nu,: berofvili. : i;:wiii. (pal-t-timc) Nu17 'her uf iio!, I 1. assi\i~nl (full-time) N u ~ :,er ofnut . , I . ! assisi;~ni(part-time) i Nun >crofpli;~ ii.: nssist:i~~t (Rd-time) Nu,? tncr of 1111:. :. ! . assi~t:int(pan-iit~~c) . Nur '.eiofall~ ,I.!, pcriiloncl (fiill- tim~ 0 Nus' ii ;.~i.oiliiel(tiill-time) Nut ,cr of ad,: i :.:r.at>ticl (pail-time) Nu ')er 0f01h ) ~..nliel (Tiill-time) Nu, bcr ofotl; I a-,iiocl ijial-I-time) Nut ber of do< . i l : l i priutc pmctice Nus ',ci ofdoc . \i:hout 1,i-ivatc placticc Nu,. oci of doi N s bci ofdei, u , 8111 psi) ;itc practice Nu! ',uiofdeii ' itliot~t i.rivna practice Nut ' k o f den Nui v r o l o u l ti8 piii.ctc piaclicc Nllt c r ofnol lli~iiitpi.~vole pl.aelice - : of no? ~- Nu - r Antalgin injccti 75O;ng ~ . : I - ? t n l il~ Number o f i v c i 'a51 ~ t , o : ~i11c Communily IHc I T:iciiit! tivlioingout: Pamcetaniol s) , IZ!ntng i:>iI-l,Onil Number o f w c i last lnint,ill ilic Community Hc i Faiilit! ri:i~i,ingout: Pancclamol la1 IO!,tng Number o f \\.ci. last inoiiili tile Communily Hc , Fa;ilit! rt:nningouf: Pancetamol tat : 501,bng i Number o f w c i iast i n o ~ i ! ltlic Communily I l c ', Facilit: ii$nt,ingaut: V i t A for childi i i o ~ l c5 i Numbcr ofwei, iast motiil, ilic Communily IIc > F:x:lct! rt$nx,ingout: BCG Nunibcr o f w c i Communily I l c DPT Nunlbcr o f wei last >,,0,1~1, t1,c Community I-lc 'iFa;ilil! riiiininqout: DPT Hepatitis I 'o,,,~>,, Number o f ~ v c i last 3,,<>1,'1, t1,c Communily HI. li Fa~,ilit)ilinningout: Polio Number o f wci Community Hc Hepatitis B Numbci of wcc Community Hc 1 Pa;$lhl! iiiniiiiig out: Mcaslcs Number o f wec ins1 :,iiii,'Il :i:c Communily Ik 'I Eiuilit! iiitiniiig out: Tetanus I b x a i ~ ' Service inside I 1boiI.l:o~ Svivvisit (ticket) Service inside I lboiliitn~ !Kcpsat visit (ticket) . Service itlsidc i bull.line l'iccnaiit mati~er cl~cck .,y a ~ ~ t l u i : ~ 18 Service insidc I buii,!in:. l'bi.gnnnt mother cl>cck 18 .,y diictnr Setvice inside I builillo:. 1)clirci.y service b y mid- c Service inside I buil,ixii; l>c/ii,ciy service by doc1 Service iiiside I lbuiltlnn: I!clivcry rooin Service inside ! boil,inng \ ' : i c i i o ~ extractioidfoict Servicc inside I buil~itn; iii iC imnlunizatiotl ~ Servicciniidc I b u i i ~ I ! i il)l'l' imm~~~>izalioii Service insidc I buil~.aiiy \,>ti jpolio immuoizatiao Serviccinsidc I Iheil~lli:; \Ii~:islc immunizatioli Servicc iiiside 1 buil~:tii. Ill' i' i i e p B Combo immun, :!ton Selvicc inridc i huil,i:ng ! li.j>:~titisB immuiiizntion I :l?il, tb8b,l~.*-5 Serviceinside i Roil l:!~:. i lk;utitis B imnirlnization I 1,alii !I,.,.I: 5 Service insidc : lhoiili!,; 1~t.iiius Tonoid [Tu i n >uni7.':ii>: :,)I lil.cglaiit molll~i Service iiiside i ' buil,I~n!: I .nvniIy planniiig pills Service issidc i buiI~,~ii:. 11 insertion ii - ~- Service iixidc ' , buil.':~!;: I! I! i-euactioil ~~~~~ z-io,n,c " 0 0 00 0 0 0000 0 000008888 2 00000 g 5 . 3 . - t o n 0 0 00 0 0 0000 0 _ _ _ _ _ _ + + + . C - CNln= 6 6 d = 2 P- a 5 2 $2 5 gs;? OD00 - - ;0000 %g d o 2 d 3 0 : : L"fq' = :..P-., N n 0 ,o N U *. n N ggzg wn- % r. a gEgXt 28% 2 8 ~ ~ tz+.30.. . i .- r 3 $ ~ ~ 2 .6 ! " = 2 b s r n u o ez qqq9 onmoo 9999°-00 4444 d d d 2 X C a 0 C - ,n,n , - N 2 = & 2 P- DU 2 N D*P-* - * Pn*uu 222 ,d d d g & 030; -7 00 d 2 n0 dddd 0 2 ooooo 0 0 gL=,4-_ Li LI zu:gg:;, .- C 37 3- -=if: 7 0 ?, g N ? 4 Y S x 0 3 g 4 NU 3s 2 on , 2, sss$-00 ddd$ g $arSg " " " z 6 2 24~ 04 ~ -gz 2 s " ~ s9 q 040 2 " _ , n * m g - nr. k - 2-gr z?gg g q o =I% % .. h % & r xz 2 5 D P- c.Z<% 0222 O%?zgo$om g K zag!,,m 3 4 -mr.-. 6; n c, ,9 m ;doc. 00-s o oo oo g 2 2 % ? f z I I - i ;zZL',? * m z ~ n2 2Zr. ;c , ? - ;. --,A,- % 2 2 3; - - z ~ 0 z ? m .% * 72 2 o 0 - ,omN " r n" n 0 " %gK ~ 0 0 dOo - 2 2 ~ ...~ ~ -. fl wnmr:.,; 5 ::W.: u D m " 000 - - %?2 2 n c > , ,, , n - 2 - 0 2 z z2 n c - g - - - m9y.1. " c , 0 , ,; '-P- q :-53:,, n 2 6; ;E -" 3 z * ?: 2 2 " -" 0 0 r. 0w dc. - s X BbaeoO~og dd P- d m d 0 2 N g OI /-In Community health facility has special 0.7818 0.8108 0.7301 0.0514 0.39 0.0804 0.17 100 100 I00 toilet for patients Commiiility lievlth facility has special 0.8091 0.5018 0.687 0.1221 0.04 0.1148 0.05 100 100 100 toilet for Puskesmas oficeis 'Toilet iisrd conlmonly by pnlienlc :~lltl 02455 0.1982 0287 -0.041 5 0.49 -00S8S 0 I4 100 100 100 1p~~skci.rill:is~ ~ V C I . O ~ ~ t n kacilily 1 ' ) ) ~ oil:xlt tn~cc u ~ ~ ~ z ~I~mlll,i l y 0.')909 I 0.991 5 -0.UU0-I O.'i7 0.UOS7 u.42 I00 IOU lo(! liar: o w l isirinc will> 01.willloill scptic I;,iii ~ ll l ')i'i , d i . * ;~ ~ i l ~ l l~i i l l i i !/11,1hI: I.l;l!l!j 0 'l'>O'I 11 ' i i ( i '1421, , ii O i i S i 0 (,(, 0 //t,'l/, il ,I //>/I 1011 //I,/ liar: o \ w lal~.iiicwith scptic tank i y p u o f lalritic cominoniiy h~31111 fhcilily 0 0.027 0.0087 -0.0087 0.57 0.0lS3 0.23 100 100 100 has: a u latrine without scptic task wn Community iiesltli facility bas: I 0.991 1 0 1 -0.009 0.24 100 100 100 Countedregistration tablc Community iiealtl~ facility has: waiting 0.9636 0.982 0.9913 -0.0277 0.17 -0.0093 0.64 100 100 100 room for patients Cammitnity iiealtl~ facility lhas: 1 I 1 0 I 0 1 100 100 100 linamination mom Community IienltB facility has: 0.8364 0.7658 0.9043 -0.068 0.19 -0.1386 0.01 100 100 100 h~jcclionhreatment roam Community I~cslth facility lhas: family 0.8 0.8018 0.8261 -0.0261 0.64 -0.0243 0.66 100 100 100 planning service room Cornmi~nity health facility has: delivery 0.6 0.5315 0.4957 0.1043 0.14 0.0359 0.61 100 100 100 moil, Cornmiioily iicollli facility llar: inp:ilicnl 1'00111 Community liealtb facility has: medicinc looill Commitnity licaltli facility has: labol-atoiy Bait cmclgcncy neonatal obsletiical service 'Total budget for 2006 ('1000 Rp ) Numbcr ofbabies 10-11 months) eiven. - BCG vaccination Number ofbvbics (0-1 1 months) given Polio vaccination Numbcl.ofbabics (0-1 I months) given llepatitis B vaccination Number ofbabies (0-1 I months) given 103.4762 136.0926 114.8532 -1 1.377 0.61 21.2394 0.34 95 97 96 DI'T Hb C ~ m b o vaccination Number of babies (2.1 1 months) . given 76.0377 55.8155 56.4486 19.5891 0.14 -0.6331 0.96 96 92 93 DPT vuceir~ation Number ofbsbics (9-1 1 months) given 59.0917 65.2569 54.0619 5.0298 0.59 11.1949 0.23 99 98 99 measle vaccination Number of pregnant mothersgiven T T 91.6606 94.789 75.3805 16.28 0.22 19.4085 0.14 99 98 99 vaccination Nambcr ofnew visit by pregnant mother (Kl)pure Namber orncw visit by pregnant mother K4) n i~ Xiiiiillbt ,,11>~gn.~:it i t i t l ~ cniil, con~plicatio~, Iliigli risk aiteodcd (puisoos) oI'prcgn:,nt N ~ ~ ! ~ ~ l ~ c r n~oll~cvs \vitI> !l: ,l, ~~l,,>,~l,~.,ti,>,, r i d ,ctcrrc,1 ((,C~S<,,,>) Nuinhcs ot~i~iotlicrs clitld blt.tii \%ili! 111 conlplicntiaii I lhigll risk nttendcd (persons) NunibcruI'iriuilicis ill chiid b i i t i ~ ail11 complication I high risk referred (persons) Delive~.y assisted by hcallh aificcr Neonatal visit Baby 0-1 1 months weighed (babies) Cliild 12-35 months weighed (children) Child 36-59 monllis weighed (childrcn) Baby 0-1 I months under dotted line and above red line orhealth monitoring card (babies) (children) Child 36-59 months ondcr doltcd line and nbovc rcd lint oI'heallli monitoring card (children) Baby 0-1 1 months under red line of Ihealih olonitoring card (babies) Cliild 12-35 months under red Iinc of health monitoring card (children) Cliild 36-59 months under rcd line of lieallll monitoring card (children) Baby 0-1 Imonths with health monitoring card imatemal child health book (babies) Child 12-35 months with health monitoring card lmalemal child health book (children) Child 36-59 months with health monitoring card / maternal c l ~ i l d health book (childrcn) Child 6-11 months given high dosc A viLlmil1 (children) Child 1-5 years givcn high dose A vitamin Mother i n confinement given high doos A Pregnant mother given tablets for blood 59.6019 58.9439 68.7358 -9.1339 0.36 -9.7919 0.32 94 96 93 regeneration (Fe3) (persons) Molher in confinement givcn tablets for 32.1068 40.7103 42.1321 -10.0253 0.14 -1.4218 0.83 94 96 93 blood regeneration (Fc3) (pcnons) Cliild giwn Ihi.11 ilose A iilnmin h momlis 21 17.1698 2040.5963 19S6.676 130401l 0.54 53.9?07 0S 97 08 97 I> ,: , 6nhi.n . -. 00 1 00 1 00 I SLO 9010'0- LE'O Z6ZO'O- 59560 6CPG'O ELZG'O :aUIJ3C,\/3U!>!l>JIII .?I(II?IIP.\V IOLUl?I1JCli?,[ IUZ-PU~~~~OSZ ua!13a!u! 00 1 001 00 l 550 EEEO'O- 68'0 6L00'0- 19Z8'0 82610 2818'0 o!%lcluv :ai!!3Jeh/ao!s!pau a[qc[!eAv s 3 ~ 0 0 salqes N P- - a I , c " z - - Z S h f m " " - 2 . . . . a m 1 0 ? 0 6 *' 2 2 d d d 0 c3 . CI 0 C 0 0 0 - - " .z q- X $ n $ 2 Lo - N N 2 - ~ m, 6 h 0 ? ~ < 2 x N a d ~i v dw x . o d o$ d z g . % : : S " s s X s s q z ? " 4 9 service bv midwife Service insidc the building: Delivery 0.4818 0.3784 0.4174 0.0644 0.36 -0.039 0.58 100 100 100 sci.vice hv doctor ~ ~ Suiv~cc illside tile bniiding: I>clivciy room -0.0055 09J 100 Sci-vicv insidi. tlie Ihuildin:: \':icnn, -0.0000 0.85 1 00 cnuactioll/iorceps Servicc insidc the building- BGC 0.0071 0.8 100 inin~~~nizatioc~ < < # . . : , ~ ~ . i : i ,.<,.,,lb,;, !:. .. ,,,%'I' 0 (in ! ,,,I, % ,P" ,,,,,,,:11,<>,, Y-; \,,ik 0 11252 0:i ion iiliiiiiiili~ltiuli Scl'vicc ijrside lllc buildi~lg: Measlc 0.0075 0.78 100 iil?nliiiii~atio~r Service inside thc building: 1)IyrIlcp B 0.0229 0.56 100 Combo immunization Scrvicc inside the buildins: I-lcpatitis B -0.013 0.74 100 for irnmur~izatian ehildl.cn undci. 5 Servicc insidc the building: Ilcpatilis B -0.0656 0.32 100 for immr~nizatian patient aver 5 Scrvicc insidc the building: Tetanus -0.019 0.41 100 Toxoid [TT] immunization for pregnant motller Service insidc thc building: Family -0.1115 ROI 100 planning pills - Scrvicc iilsidc thc buildine: IUD inscvlion -0.0403 0.36 100 Service insidc lilt building: IUD rdlactioii -0.0312 0.47 100 Service insidc tlie building: Implant -0.0166 0.76 100 i~lseiiinn Servicc iiisidc the building: Implant -0.0514 0.32 100 ietractioli Scrvice inside the building: Family -0.0198 0.49 100 planning iiijcclian Service inside the building: Side eCrects of -0.0763 0.1 100 use orcoiilraceptivei IUD control Service inside the building: Inpatient 0 I I00 lrcalment Cost (Rp) of. New visit (ticket) -255.3051 0.24 99 Cosl (Rp) of. Repeat visit (ticket) -232.6087 0.26 96 Cost (Rp) a t Prcsnant mother chcck up b) -148.0702 065 100 midwife Cosl (Rp) ot Pregnant mother check up by -271.3675 0.31 81 doctor Cost (Rp) of Delivery service by midwife -5831.1688 0.79 69 Cast (Rp) of: Dclive~y service by doctor 47294.0252 0.08 47 Coat (Rp) oE Delivery room 10579,3706 0.54 47 Cost (Rp) of: Vacum exlractionlforeeps 72157.8947 69833.3333 72238.1 -80.2005 1 -2404.7619 0.95 19 17 18 Cost (Rp) or RC'C immiinizotion 952.8302 797.1698 1040.909 -88.0789 0.68 -243.7393 0.25 97 95 95 l)l>'l' ~ $ ('mt (1tp) I > ? I ~ > > I ) A ; L I ~ < > ~ > 3225 747.4747 990.4762 2234.5238 0.29 -243.0014 0?)l 92 89 90 Cost (Rp) oC Ai,ti polio irninonizalio8i 86l.llll 7688679 945.4545 -84,3434 0.67 -1765S66 0.38 98 95 ' 95 (',,~\I vI' hlb~,t~lc (l;l>'h i~~~i~~ttt~i,:nli,~>) 029.Qn65 7S3.0ISO 95d.955 -2Z.O.iR-i 0.9 -171.9361 0.30 97 95 '16 Cost (Itp) 01: L)lrf l i c p I3 Combo 931.3725 811.8812 985.5769 -54.2044 0.79 -173.6957 0.4 92 90 91 inlmunizatio~l 3 Cost (Rp) of llcpalitis 1 immnnizatioii for 883.8384 776.699 1071.429 -187.5902 0.41 -294.7295 0.19 90 93 91 rc 1~ c l ~ i l ~ l tt11'~cr 5 ( ~,\: :p~ (; ,,f II~,,.$l!lh c t ~ , ~ ~ ~ ~ ~ , , , . ~ t ~ ~ s-.;t t 1; ~ IT?, ~ ss~sss~l 'i25.:s;-& -5~1.2~;; ii.bi -.?b..:'J>;5 ' I :('I ,o il, .i ;,~,li,,.;,>,v, < C'obt (Itl>i ut. lcta8tkb l<>hvtcar) 35.6341 36.2569 -0.6229 0.12 684 698 Midwilc's 1 . educ: St&\ 31 Midwii's i;i.t educ: l - y ~ . ~ !lpii,*,l.i pco: Mid\vilc's I:,-I educ: 2-yr ~ ; j > l tn:! pro: . Midwilc's 1;t.i educ: 3-yr ,.tce P~cilily Status: iiovciilment emplo cr I'NS Status: Covi.~iirncntemplo ee I > i T Status: Local govt colitv~ci Status: \'oiiixltccr No stntits since not \\,oil; 38 "ii: 1 li~nilli selvicc hciliiy Positio:,: llc.ici of facility Posilio!~: Co.ardinatii~g )mi,! ;it;. Posilit,!i: hii.iivifc Positicxi: \'ill.iee midwil'c owned lliciiii: Place olpii8:tte practice: Ii-ni~ Place oipii\.~te practice: o ICI Place oijrii! .rlcpractice: t i pl.iic Water sooicc piped ~ v a t ci'l\\l) i~ Water \olircc pumped nci Water iourci. well Water .t,i!rci. rain Water .uixici lake Watci .oitrc~ spring W ~ ~ C I . river .i>iiic. Watci- .t,irrci packed \v:ili. ~ir W a t c ~ ~ ~ uotherc ~ Water > o i i c i inside hoilsv Have I:.ti>nu Latrinc oa'it with septic l.l~,l< Latriiic o \ w ~.vithoul scplit tat:). Latrinc com:non or public Do not ii:nc ILitiine Havee!cctrii:ty Electi.ii:ty sit ,ice: PLN Electiiiity si,siire: puskcsii .is :!l.ce: puskcs~l.ii~.irnndo\\~l.ii genel:il,,i Electri, it? sc..nl-ce: solar I,\ Nmbl- i.llicii. in private pl:icti, 2 !place Nmbr s.l:~srt,lants: total Ninbi,i:ri\b..ianIs: midivi!~ Nmbi ivi:ksi! !mots: nurse Nmbi i d : , ~ m t a , ~ tatlici s: Keeps ;,;11:4:c bookskdni h~.t\\:rn prirnle &gov1 . i i - i i ~ . . S Ever st :iciii ,:d by P u s k c r ~ n ~ \ Last m.~~i!l? i , some: piisl;cin:ir I 0 0 0 Rp.) Last m.~iil! come: mimh::rsi lib11 ('I000 I: RP.) Last nl ,:!ili i i coma: pri\;ilr jib. ..lice Cl000 RP.) Last IT :P:IIixcomc: otbci- mcii I!.: p$nctice ~1000 ) :I> - Last ii,..nilli l$,Iaiinconic ('1'10;' !:p.) Last n! ,11!1i1c.laI spcnding lnnsji~il.n~eilicinee~q~i~~~~~~~i~. ':l>ei, fci. to puskc* il:!s ('1000 Rp.) Last n? :!iiiinct ilicome ('ll~00::I,.) Numb<, i,lchildren given liigli .'ose vit it last 6 ? > , ~ m t I ~ s Serve ; I $ 1pos):indu last monili Nuinbii ilfpilsyandu s e n d I;!.: m o i i ~ i i - Avg l i > : wspcilt atposyaniiii e;i. !I visit (minut, i N u m h i ! ~i1'b;~biesIchildsc~~ 5 visit iint:~,r posynii !ti last month (pcicciit) Num!,~ i > l ' l l b Metel. be1o:ig ti. Iliis private praclii'. Numil, a i l ' Riiceps belong io I !s pl.iv:ite this pi, ,.iir. p l n c t i c ~ Nomxbio t~iTinaculurn bcl,itig .i Illis private [,,:>ctii.e Numb< ,~TUlcrus Sonde l,i.lo! :!to this privuli, p i :~clii.e Nunib', ~>fGynccologicol 1:1bl belong to this pl-;::itc practice Nunili, t hr: (0-1 l montli,) 1 % ~ 'G VBCC~,!. iiii,, Nui~ib,: ~ > f.]served TOI.:a\< iiniuilt [. servic~ li:ib! ( O - l l m o n l l i ~A! li l'alio l vaccil>.'11111 Numb;. v > f [. .]sewed far goxi 't~rnent servic~. Iilb?. (O-llrnoetli~l11. .> B vaccik n ) i l N u m b . c > f [.]sewed for go\i.nrncnl aervici I kib! (0-1 l mon!li.) 1 8 ' T vaccin 'ii,i, Numb- , i f [. 1 served fa#. guxt :omci~t servici !!:ib) (0-1 l mooti:\) A ; :mlc vaccin8 t,>i, Nomli, ,,f [. .]seivcd for go,, orneo! sen ici i inb! (0-1 1 monllis) i i,"T-I ib . Con~i, 5 :i~.cination Numb, , i f [ . .]served for :oi< .ii~neiit servicr l'rcg~iaet motliei g l i o t '11' vacci~: '8.111 Numh : [ .]senred for gn! ..iio~ciit servic, I, i,incilt servic ' lotlrcr i n labour will, ;nmi,liliieh - . risk t i ,. nliules) Availablc ii~edicini, vaccine for pi-ivatc health scrvice: I':ii.~retamol s y a p 120mgl5ml-00ml ~ I ~ l t l e s ) Available maitcin, \.accine fol.pl.ivatc health service: I'ai.:,ctamol tablct 100mg (tablets) (tablets) Available incdicim vaccine for ~pi.i\~ate health scrvicc: \'it \ for children undci 5 (capsiili.) A v g Ibis per iiay \ji.,t on public services (lours) . Ave lils ocr cia\ il..:,t on orivatc scrviccs . , (110~1~) Govei-iimeiiiprii.i. icncml trcacixnt (Ilp.) Goveiiimciit price: :,i-e@ancy clieck up (RP.) Goveiii~~ieiirpiice :!ormat dcli\cry (Rp.) Gavcmmoii 1pi.ii.c .lcIiveiy witli complicatioi~ I(llp Goveinmoii 1pi.ii.i. IICG (babies) (Rp.) Govcii>meiitpiiic \)>tiPolio (babies) (Rp.1 Govcitimciii /price I )PT (babies] (Rp.) Goveiii~~ieiitpiiic \leasle (babics) (Ilp.) jprici, Goveiii~~~ciii I lcpatitis B (babies) (RP.) Gavci-imieiil pisic 1 (pl-epnaiit mom) 7 (Rp.1 Govcmmcni 1 p i . i ~ ~ l'iin~planning consullation (llp i Gavciiimciit l p i . i < ~ . ,! iin~plunning pills (Rp.) Govcriimmit priir. !?amplanning injection (RD.) G&ellillieiit pviii. IUD insertion ( ~ p . ) 5546.4761 4970.9035 piici. Gavciiit~~ei~t IUD ieliaciion (Rp.) 18104.8226 15435.3305 Goveiomciit jpriic 5ubcutoecoias cont~.acepti\e iii\i.r!on (Rp.) I 1585.159 10803.3773 Govcinment pitcc iubculancai:~ lion continccptiic ~ri.;::~~(Rp) 30337.4336 26755.8977 . Gavci-iimciit o r t i ; i'ontracenriie ridc effecls (Rp.) ~~~~ 17537.5667 15695.6522 Private piicc: ge~bc:.dlieatmcnt (Rp.) 2 754.4932 3584.3046 Private piicc: picg:t.~ocy cl~eck (Rp) 13109.1293 13031.7356 . Ole.) Private i~i'icv: n,.ll deliveiv . , . oo, 11868.3966 12251.5183 Privatc oiicc: dclt5~i.v with coint,licatinn (RP.) 316824.5805 3.18E+05 Privatc price: B('G ibokits) (Rp) 378367,5822 4.33E+05 Private price: Atiii l'olia @abicb) (Rp.) 4859.459 4652.3074 Private piicc: Dl' I il~abics) ( Priv:ire ipiici.: 1::ur-i ;lnlming eo~isiiltvliot~ (RP.) 4678.1761 5010.0915 Priv;itc oiicu: Fain :nlnnnin~~ills (Rn.1 1 139.3007 1131.363 (Rp 1 12634 1308 12931 3137 Noie: ~ c . s ~ i k reflect fractions unless stated othel-wise. z ~ z g g g 2 g & ' g 1 g % g g 8 1 o o o " 111%2%%1111Z?1Z?S211 o ,,,,,,,,.,,,,,, nn Z Z g C; o o o o o n n ~ o n n n n n o n o ! * P - P - P - r - C P - C h C C C C Z,,,C - cC C -,,, Pc . C Oonnrjnmnnnn",n CCC,". CP-Cc., nn00n 3 o f i 0inininin i f n nfC i Cf i f i f i mfnf i f i n f f f f f n ifi ififi 1nn fi i nn ~ I I a NeC-CeONWWP-'Oe P-NTICI* 00 CmmF\COP O O m Y ' C W m ic q...-P-nny,!,lc?cq ooododooooodo , q , o w, ooooo q - doow l z z 2 ~ 2 2 ~ ~ . - d c ? m o o ro- o a , i 0 I ,v ka a o r -NnOm x a - a g g s ~ r%:ss ~ ~ N = z ~ z g!g~ Z %~Y Z S ? ~ ~ g l 1 z & d d q qi o X ~ d q~ ~ XZ8S.i;q 8 2 y z i ~ ; g ; o q g $ o 9g 9 6 9 9 ~ 2 d+ 0 0 . o o o 9 d 94 q 0 0 q o o 49q94909900oq0q9 0 009 A - - , I - Y ) :' C a : 0 ic " z sd o o o% - zq,r.-nx o z 711i do amnms nmeoo " rn -nnV? ,, . , ,, , 0, P, *,~m ,N - ~ *v? W " , W ~ , - C ~ P ~ -i L. DOa i I od ddddd 0 0 0 0 0 0 0 ddddddddd * ': 0 a s 2 -vg%g. : : 2ZLZ2ZF4 g2gzgzs:fi a u ,b- N O - 3 o C 1 0 s F s 8 F 8 .. , m % oo+doo Oq9 090490 gzssz o , o o l 5 q 2,8°8=2680020Xgi;8s3 o??+e?O? Z+?+ddqaq c i ' cj U 5 E z-rn-c-, ,V?nmrn emn--+ -o qV?m.-hh*- -wV?OIuV? z o ~ " Enhoc. ,q9qr!q " " o z 9,.--$ ~ " ~ 2 ~ 2~ 2 $ 00-"L7 % 2 ~ z p ~ ~ % p g o g g z N " O I " u " NWoo.w,s , o s s ~ x ~ a ~ 2 * 3 ,onooo ooooc ddddd ddddddoo ododooooo w u g - E : ddddd L . : w nr-,o,,,,,m m - - 7 - m n?.w,,*,',n -,om-,omn mmmv-mmm mNv, m m o o m m r , n n r . ? nnr, n m o n m n n o mm nn n o mm n o P-nn e m s onn Vi ' n - x m N m " " " " 4 "7 N "7 N N nn ,-. N CINOriNNN N - N P - N W W n m m m mm N C I r l r l - i N m - a x s m mv,r,-,-.m mmmmmmm mmm m m m T ~ m m o m n o 2 m & &?Z n ? nn N N mm nn ZWZ x z C C C CZ C C n T , ~ - ~,. .-s m , P - . P-P-UP-P-P-P- P-P-VP-P-P-P- r n m c-~ n c n . 7 . KZK n n r, K n K nm P-P- P-P- o n m C m ViCN nnn m N n C N o - C n - - -- C " 7 P- n P-P- "7- n - , * - r e o m=P-~r,a+ z c r g" ? P . f zz xz: ? - - a s % d d c ' c S 6 o o o oo 6 d 66 oo ooo 6 66 " h - < mhC7.,P- - =- *,n 2agZ-g: nior.o-- g,g o U $ 2 X 3 5% ZZ ;X g$ gq q "NU g 3 - 3 $3 $?zooo? 5-2 3 - c . Z o r b - ++a;++ 0 C i = - o 9 0 oq, O q q q 7 E: y? 72 , , " - 0 0 p ? 2 2 b 6 N. 66 = c z , ; l s 2gg.z P - 5 W n N m C "" "Z N O - m z - 2 X 0 S S o d C c = 5 6 6d6.2V& "7' 0 0 0 2 ? o oo O d F - 1 ooo 6 Ei ; r z r ; s >~~~=$~ some-* o o c ? = + & d c ? " ? 6 0 0 0 ? -IO"-,r.* 99sqqq9 ;: % " 2" a xq-z z % ?;;2 O 2 ;2 rn N - gg ss $+ gsz : oq 7 T n" -44 3 "j % 5 -X 2 3 6 8 6 "U 5% 66 r n c r 7 > * cO c .,. O C r i q c 7 r = .,. z = 2.- 7ri c rn .- -Im.+m-," P- - I O h n N U n oo"loN-- oa lo m-z %%-: O E 3 -% 3 7-- 2 5 ~ 2 q ,2 s 2 2 23 55 - I,= P-U o " N o c ; = = d d d i j d d d dd" - N Kg 20- g 0 6 6 - " . . c " sg .- : z g x ': S Z Z $ = e -" e 9 - -: - 4 .- . z - .- /n g.- : g n " "5 - .. o i > .s B :. CO 2 - - ul C - Cs 0 - u O S i c -" oc . 2 - -8 2: E - b +~ . 1 1 ' 1 o > . = So ~ ,;,; >- E 8 - ; 9 o 5 - - 0 -x$ d e i :2 m > ." $ e 5 o * " + " S F ; o >:Y 2 s 0I e g g -- S ;s * g E 0 o m 5 z-0- S," - :'m 5 g -: ~ < g & . : 5 2 -, S - - m i ij ; 3 , . i 0 ° C -5 0 £ 2 -5 o " % - 3--.- -g - .z ~ z z $ E ;E! B ,, ,, s=5'5-,".? 6 c w ~ E ~ a x Z_Cf 5+m 5 2 E C .- i : ~, . ,;aX U hE ; < &;;;;;:;% , - - > E- C - O h > . " E c " " g g z 0 .s- : E - o " C " 8 " g .- . g : 2 2 - - a 2 o a s;s D . : .% g .5 :.8 .= :g 9 * 5. g FPE g.; ... s .5 : = " S c6.= - - 6 i r ... I ; .L. .'D .Y1 P 2E - lc?-f:f a > - .:s.~-. z 2 2" *'- t c = 3:: c . c, ;: ~ . ~ ~ . ~ o C 6 o o 6 o E ooE o ED $- ; j o.--,-kL.- s .=- C +:ti . ; + G 6 iY X g E ~ ~ -5: e L a - * . - o+ gr 5; " :, i ~ z, == ," . =D*SSS P O " ~ g g - b,cEh-"*tiu:.t:,: -.o- S" 2s.:m.9* -"-a 5 $ 2 g,gg,: g & E E z OW, i = ; , s 2 2 2I - ~ m & ~ z z z z ; ~ ~ & ~ 3 J - ~ J Z3 C A~ < , Z 1 r hz Y Z +~ 2 z E ? m d - - - - " - -w n ? 5 z E W t E , o m.5 25 2 2 = 2 ~ ~ Z r C Xz Z B . Z B > tnhle Number o f G '~iecoln~icnl belong to li fir- ,\:ti\ 2.1)-i ;. Nombcl- oCOayge~~ canister belong lo tliis private practice 0.2817 Niimbe~. ofleciib>tai- bcloiig to tliis pi-imtc psliclicc 0.1281 kit Nu,nbei of Wci~lling b c l o n ~ illis lo ,"'*a" p'"""C" I .4-1;9 Nuinbci afVscciiic currier belong lo this private pl.acticc 0.9491 Number of [...I scrvcd fo~~govenimcni service: Baby (0-1 1 months) BCG v~ccin~tion 13.1445 Nombel. of [...I servcd for gavcn,meni *eijicc. Dab) (0-1 Iinaiiiis) Anti l'olio vaccination 34.053 Numbcr o f [ ...I scrvcd for govemment service: Baby (0-1 lmonths) Hep B vaccination 20.526 Number of [...I served for govemment scrvice: Baby (0-1 1 months) DPT vaccin~tio~ 16.5603 Number o f [... servcd for govemmcnt ] scl.vice: Baby (0-1 1 months) Measlc v3ccindion 9.6019 Number of [...I servcd for govcmo>enl seivice: Baby (0-1 I months) DPl'llb Combo vaccination 15.6435 Number of [...I scrvcd for government sct.vice: Prcgi~ant mothcr givcn 'IT vaccinatio~l 15.7297 Number of [...I served fo~~govemmcnl K1 sc~vice: pmgnant mother visit 10.2245 Number of [...I served for govemment service: K4 pregnant mother visit 8.448 Number of [...I served for govemmcnt service: Pregnant moihcr with compllhigh risk treated 1.6008 Number of [...I served for govcmmcnt service: Pregnant motlier with complihigh risk referred 0.92 Number of [...I served far govemment service: Mother in labour with compliligh 0.5457 ,., ' " " r - " 2 m n o C : a z m r " " 3 ? z z q Y ) z ? . 2 , - 5 - Z!&Z ; ; 2 0 ' 1 N z hN i 2 Oz m- z - z 6 -. "? - : - - 7 -2 j *O 2 . 4 z 2 o - 2 2 - 2 " O . U ,n N N 3 9 2 2 - C 2 -, ( mX" $ _- , +Z *N 7z r W- 2 -.:. :' 5 e E - 4 2 o1 Number of [...I served for government and private srlvice: K1 pregnant mothcr visit 15.4744 16.988 12.6279 2.8465 0.27 4.3601 a09 374 381 382 >icill,~l coplvl i001,ig (c3pleti) 96.1 Sj4 7 1.6222 4659SS 49.5847 0.17 25.0235 0.4') 37J 381 3 82 ,\~;i,I.iblc ii,.iiiiiiii ~ J C C I I I L . Ibi govcmmciit lhcallh service: Ampicillin dry syrup 125rngl5ml @otllcs) 12.9477 22.4679 11.8035 1.1442 0.93 10.0644 0.45 374 381 382 Av:!ilablu mcdicindvaccioe for government liealth servicc: Anlalgia tablct 50Omg (tablets) 392.3185 297.8667 291.5915 100.7271 0.1 6.2753 0.92 374 381 382 Available mcdicindvaccine for govcmrncnt Bealtb service: Anlalgin injcctioii 250mgio1l-2ml (smpules) 15.2278 18.L523 12.9491 2.2787 0.63 5.2032 0.27 374 381 382 A\,ailable medicindvoccinc for government l~callll serviec: I'amcetamol syrup 120mgl5rnl-60ml (bottles) 13.6361 46.6814 60.8316 -47.1955 0.04 -14.1502 0.53 374 38 1 382 Available medicinelvaccine for government hcalth service: Pancelamol cablet 100n~g (iablcts) 107.9576 87.1072 97.6299 10.3277 0.72 -10.5227 0.71 374 381 382 Available mcdicineivaccine for go\~emment l~ealtli scrvicc: Paracetalnol tablet 5OOnlg (tablets) 451.2682 510.3868 404.0083 47.2599 0.57 106.3785 0.21 374 38 I 382 Available medicinelvaccii~e for govemmcnt healtli service: V i t A for childrcn undcr 5 (capsules) 151.5128 176.512 74.2391 77.2737 0.07 102.2729 0.02 374 381 382 Available mcdicincivaccine for private health scrvicc: Disposable syringe l m l (sets) 9.1252 18.1733 10.0977 -0.9725 0.79 8.0756 0.03 374 381 382 Available ~ncdicinelvaccine privale for llmlll? scrvice: Dislwsvblc syringe 2.5ml (sets) 39.2406 42.3627 39.1362 0.1045 0.99 3.2266 0.58 374 381 382 Availablc mcdicine/vaccine for private health service: Disposable syringe 5rnl (sets) 12.0592 14.9369 14.7048 -2.6456 0.42 0.2321 0.94 374 381 382 Available rnedicincivaccine for private health service: Amoxilline capsule 250mg (capsules) 24.1154 25.6463 24.8202 -0.7048 0.91 0.8261 0.89 374 381 382 Available rnedicinelvsccine for private hcaltl1 scrvice: Amoxillinc caplel 5OOrng (caplets) 104.43 107.7164 113.5104 -9.0804 0.46 -5.794 0.64 374 381 382 Available rncdicincivaccinc for private 7.8393 13.5882 7.2058 0.6334 0.85 6.3824 0.05 374 381 382 health service: A~norillinc syrup dry 125mgl5ml (bottles) Available mcdicinelvvccine for piivatc heal111 service: Ampicillin cvplct 500mg (caplets) 14.8097 23.9469 22.4127 -7.603 0.2 1.5342 0.8 374 381 382 Av:lilabIe medicinu'vnceine ibr privntc lic;~ltlis c ~ \ i i . u.\i>?],icilliiiilty,!iu], I25wg 5011 ~ ! ~ ~ ~ l l c ~ ] 1.7s; 7.(iiii j.8i'JiJ -2.0569 U.3 I j.i'j.33 U.22 374 381 js2 Availablc ~ncdicii~dvaccilrc pl-ivalc for l1cal11~ C ~ \ ~ C C S t .\iil.~lgin l ~ h l c500811g (tablcls) 118.1272 107.979 119.263 -1.1358 0.94 -11.284 0.49 374 381 382 Available mcdicinelvsccinc for privntc bcalth scrvicc: Aiitalgin i~ljcctioi, 250mglml-2ml (ampules) 1.7604 5.7846 1.5707 0.1897 0.94 4.2139 0.13 374 381 382 Available medieinelvaccine for private Iicaltb service: Paracclamol syrup 120md5ml-60ml (bottles) 10.8481 13.5912 9.6913 11569 0.72 3.8999 0.23 374 38 I 382 Availablc medicinelvaccine for private healtli scrvicc: Paracetn~nol tablet lOO~ng (lablca) 29.1617 30.4449 28.6403 0.5214 0.95 1.8046 0.82 374 381 382 Available medici~ieivuccinc privatc for hcalth scrvice: Pancelamol tablet 500mg (tablets) 241.7298 233.4309 239.9054 1.8244 0.94 -6.4745 0.8 374 381 382 Availablc mcdicineivaccine for private r~ndcr health selvicc: V i t A for c h i l d i o ~ 5 (ca~~sulc) 17.3846 17.0401 12.2474 5.1372 0.24 4.7927 0.28 374 381 382 Avg hrs pel- day spent on public services (hours) 5.8281 5.8028 5.4239 0.4043 0.09 0.3789 0.12 367 377 377 Avg hrs pciday spent on private setvices (hours) 3.6926 3.7118 3.7096 -0.017 0.96 0.0021 0.99 374 381 382 Government price: genenl treatment (Rp.) 11741.96 47551948 2753.4591 8988.4963 0.27 2001.7357 0.81 305 311 318 Government pi-ice: pregnancy check up (RII.) 2807.6923 2688.1 134 3171.7514 -364.0591 0.54 -483.638 0.41 343 357 359 Governmei,t "rice: normal dclivcrv . . . . (Ro.) 133602.17 139775.81 152922.2 -19320.03 0.11 -13146.3908 0.27 329 342 333 Govemment price: delivery with complication (Rp.) 180348.95 189521.90 172338.03 8010.918 0.73 17183.8696 0.47 I50 169 157 Government piicc: BCG (babies) (llp.) 894.8069 1853.4137 699688 195.1189 0.83 1153.7257 0.22 279 288 260 Government pricc: Anti Polio (babies) (RP.) 816.7989 754.6917 650.1524 166.6465 0.15 104.5392 0.37 281 288 265 Govemmcnt price: DPT (babies) (Rp.) 873,0053 872.14 720,155 152.8503 0.22 151.9849 0.22 280 287 260 Government pricc: Mcaslc (habies) (Rp.) 878.6667 941.689 710.1227 168.544 0.17 231.5663 0.06 279 288 263 Goventment pricc: Hepatitis B (habies) (RP.) 884.6154 906.0847 778.2546 106.3607 0.41 127.83 0.32 292 292 281 T Government price: I (pregnant mom) (RP.) 1168.4848 995.6085 877.8772 290.6076 0.03 117.7313 0.37 314 308 314 Government price: Fam planning 917.5947 553.0055 756.4252 161.1694 0.33 -203.4198 0.22 336 351 344 . .. Government price: Fam planning pills (Rp.1 1665.1376 1630.1908 1540.0718 125.0658 0.46 90.119 0.6 329 345 334 Government pricc: Fam planning injection 5254.4346 5070.5623 4761.0723 493.3623 0.26 309.49 0.48 340 353 347 IIII) I,,SL~~~~,~,, (I:], 1 1:n.10 57 121~207 11%l7~228 502 3373 o 86 -236: 1577 o .i? 180 213 I ~ I ! ' l l : . . .llil I .~..>. t2 I: . . !a,, j Iloi- ::!Y:II: I:(, ( 1 2 2 ~ ' [I :v . . . ,)-,,- IT> o: : Ir!! 2211 .: !;. liu\c!iin,iiil i,iiii. \iibciil.i~,cui#~ 11 C ~ > , , t , . , ~ C ~ ~ l , , ,C> L , t , < > , , \ : , ) , (ioicri,n?ci,t jli icc. Sobcul;incuor contraceptiue rclrlction (Ilp.) Governmentpricc: Contmceptive sidc clfccts (Rp.) l'rivalc viicc: rencral tccatrncnt (llu.1 (RII.) I'l-ivatc pl-icc: BCG (babies) (Rp.) I'l-ivaic price: Anti Polio (babies) (Rp.) Private price: DP"' (babies) (Rp.) Private orice: Measlc (babicsl (Rn.! Privnlc price: Fam planning consultation (R11.1 1336.1433 1125.7015 - Private price: Fan, planning .pills (KD.) . . 4774.3644 5161.9411 I . . I : I I : : I I , l ~ i l 17 j 10)77 73 IUI) L ~ , $ G I ~ I . , : (ltp 1 01 \:,tc p ~ ~ : c > 6Ji-I7 (2S113 77 IUO l't:<4!c ~ I I X r c l r d : ~(Itp ) ~~ 2511b7j 2792') 2 3 Private price: Subcutaneous contiaceptivc insertion (Rp.) 87081.88 8297 1 .SO Private price: Sobcutaiieous contiaccptivc cetmction (Rp.) 31407.29 31478 Private pricc: Contraceptive side efiefects (I~P.) 10782.78 11223.741 Note: Results reflect fractions unless stated otherwise. Table 51 ,6;c'i:!ol characteristics for household CCT treatment and control groups - - - Variable Treatment Control Difference P NT NC Male princip~~l 0.8719 0.9129 -0.04I 0.03 504 507 Education ol'ji'iilciliai SMAISMI,: ' 1 \ 0.0014 0.0027 -0.0013 0.64 504 507 Diploma 1 ? 0.0305 0.0486 -0.0181 0.14 504 507 Diploma 3 0.0652 0.0765 -0.0113 0.49 504 507 Diploma 4 0.1112 0.0808 0.0305 0.1 504 507 Post gradwit :(S2 53) 0.6157 0.55 71 0.0586 0.06 504 507 Experience of iiie psiiicipal as a 12.903 12.095 0.8081 0.18 487 492 teacher (yea] I Experience at"lic liuiiicipal at 9.6711 10.6085 -0.9375 0.11 500 505 other school f! Age of schoi~l'iiiiril3~1l 0.3595 0.3749 -0.0155 0.61 500 503 Main SMP 0.9244 0.9432 -0.0188 0.23 504 507 Public SMP 0.4037 0.3465 0.0572 0.06 504 507 Schools that 11. 1.c accseditation 0.1057 0.0986 0.007 0.71 504 507 Language use .ai sciiool Bahasa 0.9824 0.9857 -0.0033 0.67 504 507 School has scli 1i11 1in:il exam 0.8809 0.8747 0.0062 0.76 504 507 WAS) School has n.iiw>iiaifiiial exanl 0.9716 0.9663 0.0053 0.63 504 507 (UN) Percentage 0 1 ' . ~i:~Iciits passed 0.9969 0.9981 -0.0012 0.68 437 429 school final ch ,i:! Percentage o i i;!,ii.iiizpassed 0.9702 0.9703 -0.0002 0.98 432 423 national finai i ;::ill Facilities Nr. of clas, ' ioiiis 8.7776 9.1875 -0.4099 0.32 504 507 Nr. of lab(,i-:,-osics 0.9466 0.9041 0.0425 0.56 504 507 Nr. of libai i c 0.8647 0.8417 0.0231 0.31 504 507 Nr. of all liin jitis~,room 0.3098 0.3673 -0.0575 0.08 504 507 Nr. of sc11i.r Iii.:,lili affairs 0.5704 0.6175 -0.0471 0.14 504 507 room Nr. ofBP. 1%;: ! t ~ o ~ i i 0.6629 0.6597 0.0032 0.92 504 506 of NI--. schiw pricipal room 0.8943 0.9166 -0.0223 0.24 504 506 Nr. of teac!, st.niii i h 0.9637 0.9488 0.0149 0.35 504 506 Nr. of adixii .I iuiiiii 0.8593 0.8326 0.0267 0.27 503 506 Nr. of teat!!. I r i c 1.7003 1.6162 0.0842 0.17 504 507 Nr. of st~l(!.:i; ,i.i 3.7348 4.2371 -0.5023 0.04 504 507 Nr. of soc\c, i i - l ~ l a 1.0798 1.1235 -0.0437 0.43 504 507 Nr. of coi~,i> iwis 10.1471 7.8333 2.3138 0 504 507 Total Nunibi r ic,ic!icrin a 25.4566 25.8833 -0.4267 0.63 503 507 school pcr Number of sii~.~ciiis classroom 38.011 36.2254 1.7856 0.03 501 504 in 1st gradc Nuniber of Y: :.:!- pcs classroon~ 36.2077 35.1709 1.0367 0.18 499 498 in 2"dgradc Number of s.81 , ! i . lpcr classroom 34.2747 33.0477 1.227 0.13 488 487 in 3rd gradc Percentage i , . ' .rz,:t sh~dents in 0.0078 0.0073 0.0005 0.83 487 489 1st grade Percentage , 8 ti &idelits ni 0 0079 0 0087 -0.0008 0 72 475 479 2nd grade - - - . ~ -~ Percentage oi' \tudents in s'.:f 0.037 0.038 -0.001 0.9 472 475 3rd grade Percentage o;', !..,:it students in 0.0195 0.0239 -0.0044 0.24 488 489 1st grade Percentage o'. .I.:.t i i t students in 8 2st grade Percentage oi', , . q > . ~ i istudents in t 3st g a d e Percentage o f ' . !V!..iii~al studelits in 1st grade Percentage o i ' , I li:t,iial students in 2nd grade Percentage oi', i !!li,iial students in 3rd grade Average gi-ai!c :. r ' V : i d o i e s i a ~ I : ;.ilc 1 to 10) ( Average gmiie . : I ! S:iMat11 (grade scale 1 ! i t !. Average ginrli, . ' : 1 U:English (grade scale ! . . 1": Average ainiio I wr.i.ived per studelit in l ( ! O i .- (Rp.) Percent of >i!~i, -\it11 scliolarslup Average aliioii - . \ I ~ h o l a r s l ~ i p s per capita i I:,) School absci;: , i!l:t!iii permissioii Number oI'c1.1 branches . 11 school Nuniber ol'i!ii :I i l l school branches Number olc'a . ;I] open school Nuniber oii:ii ,I I I I ope11 school Tea~heragz~! , ~ t Teacher ehp$.s I sear) : $ Education,>:I. 1 1 I , SMtVShll.: ! ' ' Diploma ! ? Diplo~iia 3 Diplonxt -! Post gra~iii.~~ 53) . I ,: TotalbudyL,l2 , 15 I:!07 (1000' RP.) Total reveii\l<. )i, 2007 (1000' Rp.1 nos (1000' !:I Total educ.;ir~~s, :iditure spent by parents i i t I 6 f107(1000' ~ - esiracurric !'.I; (1000' <,ti ~!ies Rl).) Tntal exp 21299.8184 18201.37 3098.4469 0.22 471 470 m:iierials lCi00' Rp.) School sanii I c .!~ditionclean 0.81 81 0.761 5 0.0566 0.03 503 506 Tnble avai!:)'.! .:a k r s in class 0.99 0.9844 0.0057 0.42 503 506 Biackboanl .!I, ,ii ! ' 'iiiarkers in 11.9894 0.9747 0.0147 0.08 503 506 cI:1ss Floor made <;I 11.1 0.0553 0.0724 -0.0171 0.27 503 506 Ft~iictionil;g11 ' i s ' , ~ l n s s 0.7598 0.7413 0.0184 0.5 503 506 s L c ~ k inceil~! 0.2219 0.2098 0.0122 0.64 503 506 . - Arote:Res:ii1 I-c 'cct fractions uilless stated othelwise. ,nV)V)V)mCn w V)V)101010 10 z;: r N N N N N N . V) r. N * V ) V ) V ) V ) V ) r N N N N N N . V) V N m N V) m N Vr.l.r.r. N N N N N r N - - Z--cz O O O O O * cr.r.r.r.w 0 c m 0 0 0 0 0 W r . V r . F r . 0 r. w m W m 0 0 0 0 0 r.Cr.r.r. 0 C -8 N N N N N N N N N N N N N N N N N N N N N N % " r-mo*mm bWNOO\Cn 0 r T 2 2 2N Z?ZZZ m 2 m b m 1 0 N - IT. 3 C! C?P.'?":Ny O - P - O P . 0 oooooo 0 6000dd i w Nr. of school pricipal room Nr. of tcncliers r o o n ~ Nr. of adninin room Nr, ortcachcr IYC Nr. of student wc Yr. of soccrr field? v v ,>r ,,,,,, ,..,.. ,.,,. l,,i ,I -, : * . ! , I ., I,., >.,I.: ,, I school Nuiuber of students pcr classrooin in 1st grade Nuinbcr of students per classrooin in 2""rade Nuinber oCstndeiits per classroom in 3rd grade Percenlage of repeat sh~dents in 1st grade Percentage of repeat sh~dents in 2nd grade Percentage of repeat students in 3rd grade Percentage of dropout shtdents in 1st grade Percentage of dropout shtdents in 2st grade Percentage of dropout sh~deilts in 3st grade Pel-celita,xeof additional students in 1st grade Percentage of additional students in 2nd grade Percentage of additional students in 3rd grade Average grade for UN: Indonesian (grade scale 1 to 10) Average grade for UN. Math (grade scale 1 to 10) Average grade for U N English (grade scale 1 to 10) Average atnoilnt received per studenl in 200612007 (Rp.) Percent of students with scholarship Average amount of scholarships per capita (Rp.) School absence without permission Nltnlbev of classes in school branches Number of students in school branches Number of classes in ope11 school of Nun~ber students in open school Teacher age (year) Teacl~ex experjence (yeal-) Education of teachers SMA1SMRlMA Diploma 112 Diploma 3 Diploma 4 Post graduate (S2lS3) Total budget 200612007 (1000' RP.) Total revenue 2006/2007 (1000' RP.) BOS (1000' Rp.) Total education expenditure spent by parents in 200612007 (1000' RP.) Infrastructure maintenance (1000' RP.) Infrastructure maintenance and rehabilitation (I 000' Rp.) Study-teailhit~gand extracurricular activities (1000' RP.) Total expenditure for teaching illaterials (e.g. books and olhel-s, IOOO' Rp.) School sanitation condition clean Table available teachers in class Blackboard and chalWmarkers in class Floor made of earth Functioning lights in class Leaks in ceiling Nofe:Results reflect lractions ullless stated otherwise. Figure 2 Baseline sample selection PNPM 20 PNPM districts Treatment with rewards 100 sub-districts I Treatment without rewards 100 sub-districts Control 100 sub-districts I + Select all 100 sub-districts Randondy select 1 i Select all 100 sub-districts Randomly select Select all 100 sub-districts Randomly select S villages per sub-district 8 villages per sub-district S villages per sub-district Randomly select 1 ward per village Randomly select 1 ward per village I Randomly select 1 ward per village Randondy select Randomly select 1 Randomly select 5 households per ward per 5 l~ouseholds ward 5 households per ward 2 households with 2 households with 2 households with pregnantllactating prcgnantllactating pregnantllactating - wonlen 2 households with children age 6 to 15 women 2 housel~olds with children age 6 to 15 women 2 households with - 1 housel~oldfi-on1the remaining group - 1 housel~old from the remaining group children age 6 to 15 1 household from the remaining group Figure 3 Baseline sample selectioi~PICH 44 PIU3 districts 8 villages per sub-district 5 households per ward - + 2 households with pregnantllactating women 3 l~ouseholds 5 with children age Conditional on UCT eligibility 1 - 5 households per ward 2 housel~olds with pregnantllactating wolnen 3 housel~olds It015 with children age + Conditional on UCT eligibility 1 12 14 16 18 Ln per capita expenditure Household CCT: T Household CCT: C Community CCT: T (Inc) -- Community CCT: T (No lnc) Community CTT: C for Figure 4 Distribution of Ln per capita montlily expe~iditures treatment and control groups 5 10 15 Ln per capita education expenditure Household CCT: T Household CCT: C Community CCT: T (Inc) Community CCT: T (No lnc) Community CTT: C Figure 5 Distribution of Lu per capita moiithly education expenditures for treatment and control groups 10 15 Ln per capita health expenditure Household CCT: T Household CCT: C Community CCT: T (Inc) -- Community CCT: T (No lnc) Community CTT: C Figure 6 Distribution of Ln per capita monthly health expenditures for treatment groups and eo~itrol