Conditional Cash Transfers in Indonesia

Concept note

February 27, 2007

1.Key development objectives and rationale for Bank involvement

Indonesia has achieved remarkable progress in key human development indicators. Primary enrolment is close to universal. Child mortality has dropped faster than in most countries in the region, and basic health services are widely available. However, there are human development indicators in which Indonesia is seriously lagging and where special attention is required. These include the transition rate from primary to secondary school, child malnutrition, maternal mortality, and access to safe water andsanitation.For all these indicators, there is a strong correlation between poverty and outcomes. This suggests that a program that targets the poor, and provides them with the means to access basic health and education services could be an important component of a poverty strategy for Indonesia. Conditional cash transfers have proven to yield impressive results in this area in Latin American countries.

This year, the Government of Indonesia will launch a large pilot of two conditional cash transfer programs in six provinces. The programs promote healthy pregnancies, nutrition, and enrolment in primary and junior secondary school among the poor. These programs have a potential to make a substantial impact on improving human development outcomes among the poor and reducing poverty.

The World Bank has provided substantial support to help design these programs. This has been in the form of technical assistance, mostly supported by trust fund resources. The purpose of this concept note is to (1) document the developments with regards to the CCT programs in Indonesia to date, (2) document the Bank’s support to this initiative and (3) obtain agreement on the future direction of Bank support for the conditional cash transfer programs in Indonesia.

There a good reasons for the Bank to stay involved. First, it is in the CAS progress report(World Bank 2006)[1]. Second, the Bank has global experience in CCT programs, and provides financing and support for the two CDD programs that will form the basis for one of the CCT programs to be piloted. Third, the Government has requested continued Bank support. The support is a continuation of a close cooperation between the Bank and the Government to advice Government on the design and implementation of their poverty reduction and social protection strategy. The poverty assessment brings together a lot of the analytical work which has been shared with Government over the preceding years(World Bank 2006). For the UCT the Bank provided advice on the design, targeting and conducted, at the request of Government, several rapid assessments which improved the operation of the program mid-term. The community CCT is incorporated in the national CDD program (PNPM), which is one of the pillars of the Governments poverty reduction strategy, and currently implemented as an expansion of the World Bank financed KDP and UPP projects.

The note is structured as follows. Section 2 provides a brief historical background to the CCT program in Indonesia, which is important to understand the political and institutional context in which the CCTwill be launched. Section 3 describes the design, funding and institutional arrangements of the proposed CCT programs. Section 4 documents the Bank support to date and section 5 presents the key development objectives, components, and risks for future Bank engagement with the CCT programs.

2.Historical background to Conditional Cash Transfer program

Up until 2005Indonesia’s largest cash transfer program was in the form of fuel subsidies. These largely ended up with the better off, who are the owners of cars and consume the bulk of the gasoline. Kerosene was also subsidized, and even though this is used by the poor for cooking, the larger share of the subsidy ended up with the better off. Prices for these fuels were fixed by the Government, and increasing international oil prices thus translated automatically increased fuel subsidies. With the increase in world fuel prices in 2005, the subsidy level rose to 75 percent of total subsidies and transfers, 24.8 percent of total government expenditures and 5.1 percent of GDP.

In 2005 the Government of Indonesia took a bold move to reduce regressive fuel subsidies. The government raised fuel prices in March 2005 by a weighted average of 29 percent, followed by a more dramatic increase in October 2005, when prices rose by an additional 114 percent. While in March 2005 the price of kerosene was left unchanged, in the October price rise the kerosene price almost tripled. The annualized budgetary savings from the 2005 October fuel price increase (as they apply to the 2006 budget) are estimated to be equivalent to some US$10.1 billion. The budgeted amount for fuel subsidies in 2006 is down to some US$7.1 billion (in contrast to the US$15.7 billion it would have been in 2005 without the October price increase).

Before undertaking the major fuel price increase in October, in August 2005, the Government decided to put into place a targeted unconditional cash transfer (UCT) program. Each beneficiary family received about US$120 transferred in four tranches over the course of one year. Beneficiaries were selected by the Central Bureau of Statistics, with help from local administrators who provided initial lists of poor households to be considered for the program. The selected households were provided four coupons, which they could cash in at the nearest post office. The UCT program initially covered 15.5 million poor and near-poor households (some 28 percent of the population, in excess of the poverty rate of 16 percent) but after the initial round of payments, more people applied for the transfer and 3.7 million additional households were allowed to enter the program, resulting in a total of 19.2 million households becoming beneficiaries. The total annual budget for the program is estimated to be closeto US$2.4 billion. The last payment of the UCT was in September 2006.

The UCT was evaluated using rapid assessment fielded by two Indonesian research institutes(LP3ES 2005; SMERU 2005) In addition, a special module on the UCT was included in the February 2006 round of the panel Susenas household survey, a nationally representative household survey collecting data from 10,000 households.

Both indicate that the UCT performed rather well in term of transferring money to beneficiaries. 27 percent of the households in Susenas reported to have received a UCT. Of this group, only 2 percent reported not to have received the first payment. Of those who received the first payment, 94 percent reported to have received the full amount. For those who did not, the average cut was 20 percent. This was most frequently used to provide cash to households which were not selected for the UCT. The second most quoted reason is to pay for transport costs to the post office. These findings are confirmed by the qualitative assessments.

The targeting of the UCT program ended up to be similar to that of many other social assistance programs in Indonesia. The program was targeted towards the poor, but there is considerable scope for improvement. Mistargeting occurred both in the regional allocation of cards to regions, and the identification of individual beneficiaries. Mistargeting, in combination with poor socialization (providing information) of the program, were the basis of the most reported complaints in the qualitative assessments and were also frequently listed among complaints in the Susenas survey.

The UCT can be considered a success in the sense that the fuel price increases passed without any public upheaval. In the past smaller price increases triggered large scale protests. At the same time, the UCT programs did receive a lot of criticism in the media. The UCT program was criticized because it created dependency, viewing the poor as subjects, rather than investing in the poor to help themselves. The decision to launch a conditional cash transfer program can in part be viewed as a response to this criticism. Bappenas views the CCT as the embryo for a social safety net for Indonesia. The political backing for the CCT comes from the minister of Finance and Bappenas.

3.Design of the Conditional Cash Transfer Programs

Two CCT programs will be piloted, a household and a community based. Table 1summarizes the design of the two programs. The household CCT is modeled after similar programs in Latin America, building upon the experience gained in the UCT. Poor household receive a transfer conditional upon utilizing a basic set of health and education services. The community CCT builds upon the large experience in community driven development programs, in particular the World Bank funded Kecamatan Development Project and Urban Poverty Project. Communities receive a grant that they can use to increase utilization rates of the same basic services as targeted under household CCT. Facilitators guide the process of decision making and provide technical know-how on what works in improving health and education outcomes. Conditionality is introduced in the form of financial incentives, by making a portion of the block grant in second and subsequent years dependent on the performance in the previous year.

Our assumption is that the household CCT will be more effective in urban areas, where inequality is high and services are easily available, while the community CCT is better suited for rural areas, where inequality is low, collective decision making is easier and service supply is problematic. Good evaluation is needed to test this assumption. The household CCT has the advantage that because the conditionality is at the household level, it provides a strong incentive for those who receive the transfer. It also reduces poverty through the income transfer. It is a proven model, which has show to be successful in many other countries. Most of these countries, however, are Latin American countries, with much higher inequality than in Indonesia. It may be difficult in Indonesia to sustain a household targeted program if differences between poor and rich are not as pronounced. With the household CCT that is less of a concern, as the decision on how to share the community grant among households is taken by consensus. The community CCT also has the advantage that it can deal with small scale supply side problems, such a providing transportation for a service provider to come to the village. Since collective action making is part of the program, organizing these sorts of interventions is easier than in the household CCT. The disadvantage of the community CCT is that it provides a community level incentive to stimulate household demand. There could be pressure in the community to share the grant widely, limiting its impact.

The program will be piloted in six provincesof West Java, East Java, Jakarta, Nusa Tenggara Timur, Gorontalo and North Sulawesi[2]. In one district will either receive the household or the community CCT. The program is targeted towards the poor within these provinces, by excluding rich districts and sub-districts, and through a proxy means test in the household CCT. The household CCT is only implemented in sub-districts which are expected to be able to accommodate the additional demand for services. The community CCT is better equipped to handle small scale supply constraints as community could use the block grant to bring services to their village. The Community CCT is restricted to largely rural areas that have been exposed to KDP or UPP for at least 2 years, and thus have experience with village level planning.

The programs were initially envisioned as two pilots of the CCT, both to be funded from the regular Government’s budget. But the budget for the CCT was cut, and the Government also decided to scale up the existing CDD programs as a national poverty reduction program (PNPM). As a result the regular Government’s budget is used for the household CCT, while the community CCT is folded under the PNPM CDD expansion.

The poverty unit in Bappenas coordinates both programs, but the implementing agencies are different. For the household CCT it is the Ministry of Social Welfare (Depsos), chosen because the CCT is viewed as an embryo for a social security program. For the community CCT, Menkokesra is the agency proving overall coordination of the PNPM program. For implementation of the community CCT, it will rest with the Ministry of Home Affairs and the Ministry of Public Works, which are also the implementing agencies for the PNPM program.District governments are responsible for providing an office and staff for implementation of the household CCT program. This is formalized through a memorandum between the district and central government. At present, 31 out 49 districts have signed this memorandum. The rules allow for reduction of the size of the program in districts which consistently shirk on their responsibilities. Given the high visibility of the program, this threat will discourage shirking by district governments. For the community CCT, district support is less of a concern. The program is implemented through KDP and UPP project channels, which have very limited roles for districts.

The most recent Government planning documents indicate plans to scale up the household CCT gradually to 2010 and phase out in 2015. At its peak in 2010, the program will disburse about 12 trillion rupiah, a 10th of the fiscal space reported in the recent public expenditure review. The fiscal space is mostly the result of the reduction in fuel subsidies. By comparison, the cost of the UCT program was about 20 trillion Rupiah. The recent public expenditure review reports that the 2005 reduction in fuel subsidies freed up about 92 trillion rupiah. It would appear the program is fiscally sustainable.

Table 1 Design of household and community CCT programs

Household CCT
Program Keluarga Harapan / Community CCT
PNPM Generasi
Indicators targeted /
  • Pregnant women receive 4 pre-natal care visits
  • Pregnant mothers iron tables during pregnancy
  • Births are attended by a skilled health professional
  • 2 post natal care visit
  • Complete childhood immunizations
  • Monthly weight increases for infants
  • Vitamin A supplements for children under 5
  • Enrollment at primary school for children aged 6-12
  • 85 percent attendance at primary school for children aged 6-12
  • Enrollment at junior secondary school for children aged 13-15
  • 85 percent attendance at junior secondary school for children aged 13-15. HH CCT includes option for 16-18 year old to go to informal junior secondary equivalency training (packet B)

Size of transfer /
  • Minimum Rp 200,000, maximum Rp 2,200,000 per household per year depending on family composition
  • Average Rp 1,390,000
/ An average amount of 1.5 billion per sub-district
Conditionality /
  • Children under 6 are required to make required visits to public health clinic
  • in case of private provision, provider has to pass on utilization record to public clinic
  • Primary and junior secondary school confirm 85 percent attendance rate
  • Providers send monthly report to district on whether conditions have been met.
  • Next months payment is reduced if conditions are not met
/
  • Health service is collected through village level record keeping while school attendance is collected from regular classroom attendance records.
  • Two versions are piloted. One where second years’ block grant is independent of first years performance, one where it is. For the latter, second year the distribution of a portion of the block grant depends on village scoring of the 12 indicators. Budget fixed at sub-district level (inter-village competition)

Targeting /
  • Focused on income and nutrition poor sub-districts
  • Excluding sub-districts that lack supply of basic health and education services
  • Sub-district quota depending on nr of poor in sub-district
  • Ministry of Social Affairs identifies beneficiaries through proxy means test on UCT recipients and new candidates in areas where the UCT has too few recipients
  • Two richest districts in province excluded
/
  • Excluding Jakarta and urban areas
  • Districts needs to have had KDP
  • Two richest districts in province excluded

Implementing agencies / Leading agency: Ministry of Social affairs
Post office to transfer funds
Bureau of statistics make initial roster of beneficiaries /
  • Ministry of Home Affairs, for KDP
  • Ministry of Public Works, for UPP

Budget / $ 110 mln first year from central governments budget / $ 21 mln first year through KDP PNPM loan
$ 5 mln for first year through Dutch TF implemented through KDP
Expected coverage (nr of kecamatans) / 348 / 129
Expected coverage (nr of beneficiaries) / 500,000 / n.a.

4.Bank support to date

The development of the CCT programs has been supported byPREM,SDand HD in the World Bank Jakarta office. Most of the technical assistance has been supported by trust funds, in particular the DSF. The Bank has closely collaborated with ADB in defining the assistance. ADB has also provided additional financial support to hire technical assistance. The main recipient of the support has been Bappenas, and in particular the poverty division that is charged with preparing for the conditional cash transfer program.

The Bank and ADB jointly provided:

  • Two international experts on CCT, Tarsicio Castañeda and Luisa Fernandez, for a total period of 290 days.
  • Funding for 4 government delegates from Bappenas to participate in the international CCT conference, held in Turkey in June 2006.
  • A 2 week visit of Rita Combariza Cruz, the head of the Columbian CCT program.
  • A 1 week visit of C. Rafael Flores A., an international monitoring and evaluation expert from EmoryUniversity
  • Organized and provided funding for 6 government participants selected by Bappenas to visit the Oportunidades program in Mexico in August 2006.
  • Design and implementation of baseline survey (for both CCT programs) supported by a 1.5 million dollar grant provided by the Decentralization Support Facility.
  • The World Bank Jakarta office (PREM, SD) has provided substantial day to day support to Bappenas to support the design of the CCT. During the past year, an estimated two years of staff time may have been devoted to this. WBOJ staff provided assistance was provided on:
  • Conducting a review of the UCT, with a view of learning for the CCT.
  • Designing the targeting of the household CCT
  • Preparing the design and field manuals for the community and household CCT
  • Funding and mobilizing a local NGO to implement a small operational pilot for community CCT
  • Designing the evaluation framework and survey design and mobilizing
  • Recruiting local TA to prepare for CCT program
  • Support in operational planning for implementation

5.Key development objectives, components, and risks for future Bank support to the CCT

The objective is to provide technical assistance to Government required for the implementation of a cost-effective CCT program that is successful in reducing poverty and increasing human development outcomes among the poor.To this end, future Bank support for the CCT programs will focus on advising on the implementation of the pilots including proper monitoring and evaluation, and the expansion of the programs, if successful. Our expectation is that they will be, and that Government will want to expand the programs and integrate them in a wider strategy of social protection. The official manual for the household CCT contains a plan for expansion to 1,5 million households in 2008 ($330mln), 4 million in 2009 ($770mln) and 6,5 million in 2010 ($1,243 million). The KDP community CCT expects a budget of $56mln per year for 2008 and 2009.