67312 A E SOCIAL ASSISTANCE PROGRAM AND PUBLIC EXPENDITURE REVIEW 4 THE WORLD BANK, JAKARTA OFFICE Indonesia Stock Exchange Building Tower II/12th Floor Jl. Jend. Sudirman Kav. 52-53 Jakarta 12910 Tel: (6221) 5299-3000 Fax: (6221) 5299-3111 Website: www.worldbank.org/id THE WORLD BANK 1818 H Street N.W. Washington, D.C. 20433, U.S.A. Tel: (202) 458-1876 Fax: (202) 522-1557/1560 Website: www.worldbank.org Printed in February 2012. Designed by Hasbi Akhir (hasbi@aisukenet.com) Cover photograph and photograph on page 11 provided by Anne Cecile Esteve/Matahati Productions/World Bank. Photograph on page 7 provided by Rythia Afkar. Copyright protection and all other rights reserved. The Social Assistance Program and Public Expenditure Review policy notes 1 through 8 together comprise Volume 2 of Protecting Poor and Vulnerable Households in Indonesia report. Both the report and the policy notes are products of the World Bank. The �ndings, interpretations, and conclusions expressed herein do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the Governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. For any questions regarding this report, please contact Vivi Alatas (valatas@worldbank.org) or Jon Jellema (jjellema@worldbank.org). JAMKESMAS HEALTH SERVICE FEE WAIVER SOCIAL ASSISTANCE PROGRAM AND PUBLIC EXPENDITURE REVIEW 4 JAMKESMAS Health Service Fee Waiver Table of Contents List of Abbreviations, Acronyms and Indonesian Terms 3 Executive Summary 6 1. Background 10 2. Objective, Program Size and Bene�t Adequacy 13 3. Targeting 18 4. Impact 22 5. Cost Effectiveness 27 6. Implementation 29 7. Public Financial Management and Sustainability 34 8. Summary and Recommendations 37 References 39 List of Figures Figure 1: Health Insurance Coverage, 2005 and 2010 12 Figure 2: Jamkesmas Expenditure and Coverage, 2006-2010 13 Figure 3: Jamkesmas Flow of Funds in Puskesmas 15 Figure 4: Jamkesmas Flow of Funds in Hospitals 16 Figure 5: Jamkesmas Flow of Funds, Overall 17 Figure 6: Jamkesmas Membership and Card Distribution 18 Figure 7: Jamkesmas Coverage by subgroup and decile, 2010 21 Figure 8: Jamkesmas Utilization by decile, 2010 24 Figure 9: Perception of Extent of Jamkesmas Bene�ts 30 Figure 10: Satisfaction with Jamkesmas Services 31 Figure 11: Media Perceptions of Jamkesmas, BLT, and Raskin 32 Figure 12: Number of Jamkesmas print media articles by topic 33 Figure 13a: Budget Execution Ratio by activity, 2008-2010 34 Figure 13b: Budget Disbursement Rate, 2008-2010 34 List of Tables Table 1: Jamkesmas at a Glance 14 Table 2: Jamkesmas Expenditure Summary 15 Table 3: Characteristics within Indonesian Populations, 2009 19 Table 4: Characteristics within Jamkesmas and non-Jamkesmas Populations, 2009 20 Table 5: Rates of Health Events, Health Facility Utilization, and OOP health expenditure shares 23 Table 6: Jamkesmas Targeting 25 Table 7: Spending Ef�ciency Indicators, 2006-2010 28 List of Boxes Box 1: Jamkesmas Financial Sustainability 36 2 List of Abbreviations, Acronyms and Indonesian Terms AFC Asian Financial Crisis APBN Anggaran Pendapatan dan Belanja Negara (Central Government Budget) APBD Anggaran Pendapatan dan Belanja Daerah (Regional budget, both Provincial and District budgets) ARI Acute Respiratory Infection Askes Asuransi Kesehatan (health insurance for government employees including military and pensioners) Askeskin Asuransi Kesehatan Masyarakat Miskin (Health insurance for the poor) Bidan Midwives Binkesmas Bina Kesehatan Masyarakat - Direktorat Jenderal (Community Health - Directorate General) BKKBN Badan Koordinasi Keluarga Berencana Nasional (Family Planning Coordination Agency) BLT Bantuan Langsung Tunai (Unconditional cash transfer) BLUD Badan Layanan Umum Daerah (Regional Public Service Agency) bn Billion BOK Bantuan Operational Kesehatan (Operational health assistance program) BPS Badan Pusat Statistik (Central Statistics Agency - Statistics indonesia) Camat Sub-district Head DG Direktorat Jenderal (Directorate General) Dinas kesehatan Local health services implementing agency DPA-RSUD Dokumen Pelaksanaan Anggaran (Budget Implementation Document) DRG Diagnostic Related Groups GOI Government of Indonesia IFLS Indonesian Family Life Survey Jamkesda Jaminan Kesehatan Daerah (Local level health insurance scheme for the poor) Jamkesmas Jaminan Kesehatan Masyarakat (Health insurance scheme for the poor) Jamsostek Jaminan Sosial Tenaga Kerja (Workforce social security) JPKM Jaminan Pemeliharaan Kesehatan Masyarakat (Community health insurance scheme) JPKMM Jaringan Pengaman Kesehatan Masyarakat Miskin (Health safety nets for the poor) JPK-Gakin Jaminan Pemeliharaan Kesehatan Keluarga Miskin (Health Insurance for the Poor) JPS-BK Jaring Pengaman Sosial Bidang Kesehatan (Health Safety Net) JSLU Jaminan Sosial Lanjut Usia (Social cash transfer for the elderly) JSPACA Jaminan Sosial Penyandang Cacat Berat (Social cash transfer for the disabled) Kab/Kota Kabupaten/Kota (District) Kartu Miskin Card for the Poor Kartu Sehat Health cards (for the poor) Kemenkes Kementerian Kesehatan (Ministry of Health, MOH) Kemenkeu Kementerian Keuangan (Ministry of Finance, MOF) KemenkomInfo Kementrian Komunikasi dan Informatika (Ministry of Communications and Information Technology) KPPN Kantor Pelayanan Perbendaharaan Negara (State treasury service of�ce) 3 JAMKESMAS Health Service Fee Waiver LAKIP Laporan Akuntabilitas Keuangan Pemerintah (annual �nancial accountability reports) Lurah Village Head LHS Left hand side (of graph) M&E Monitoring and Evaluation Non-BLUD Bukan Badan Layanan Umum Daerah (non-Regional Public Service Agency) OECD Organization for Economic Co-operation and Development OOP Out of pocket costs (for health care) PKSA Program Kesejahteraan Sosial Anak (Social cash transfer for disadvantaged children) PODES Potensi Desa (Survey of village potential) PPJK Pusat Pembiayaan Jaminan Kesehatan (Center for Health Security Financing) PT Pos Kantor Pos (National post of�ce) Puskesmas Pusat Kesehatan Masyarakat (Community health center) SKPD Satuan Kerja Perangkat Daerah (Sub-National Government Task Force) SG Secretary General SSN Social Safety Net Surat Miskin Letter Certifying Poverty Susenas Survei Sosio-Ekonomi Nasional (National Socio-Economic Survey) Rp Indonesian Rupiah RSU Rumah Sakit Umum (Public Hospital) RT/RW Rukun Tetangga/Rukun Warga (Neighborhood) RHS Right hand side (of graph) TK-LHK Tenaga Kerja di Luar Hubungan Kerja (Workers with Employment of out of Formal Work Relations/ informal sector) UCT Unconditional Cash Transfer US$ United States Dollars 4 5 Executive Summary Macroeconomic growth and incomes have been on the rise since the Asian Financial Crisis (AFC), but health service utilization and health outcomes in Indonesia have been slower to improve. Poor households – who typically have less experience with the healthcare system – have been especially disadvantaged. Shocks from adverse health events are a persistent risk for poor Indonesian households, negatively affecting household incomes, productivity and overall well-being. Health insurance coverage remains low overall and especially low for poor households. Jamkesmas could provide valuable bene�ts by allowing cardholders to acquire preventative, curative, and catastrophic health care services without fees. When it promotes healthy households, keeps students active, alert, and participating in their education, returns adults to work sooner, and saves households from the high costs of healthcare, Jamkesmas’ sizeable individual bene�ts should be matched by increased social bene�ts resulting from a healthy and productive population. Over 75 million Jamkesmas cards have been distributed, allowing access to nearly unlimited health services at both public and private inpatient and outpatient care facilities. In 2005, the Government of Indonesia (GOI) introduced the health service fee waiver that would eventually become Jamkesmas. Over 2008 to 2009, coverage was expanded while poor household targeting and prioritization were re�ned, the Jamkesmas name was introduced, and authority for program implementation shifted government agencies. The recent expansion to near-poor households means that Jamkesmas cards could cover more bene�ciaries than all other public and private health insurances combined. Jamkesmas has been provided to poor households, but many non-poor have also received Jamkesmas bene�ts due to dual central and local targeting processes which have led to frequent mismatches and errors in coverage. Jamkesmas cards were printed for poor households on registered lists kept by the Badan Pusat Statistik (Statistics Indonesia, BPS), subject to veri�cation by a combination of local authorities in consultation with Jamkesmas and BPS 6 of�cials. In practice the veri�cation and consultation phases were implemented with revisions to lists and sometimes were not implemented at all; in addition, Jamkesmas card distribution was sometimes controlled by village- and neighborhood- level authorities. Furthermore, many alternative cards and were initially accepted as proof of Jamkesmas coverage. As a consequence, many non-poor ended up with access to the Jamkesmas bene�ts that are ultimately provided by public and private health service providers. Jamkesmas has had positive impacts on health service utilization… Households with Jamkesmas access increased their health service utilization rates over previous levels by signi�cant amounts. This is true for inpatient and outpatient care and services at both private and public providers and at both primary health care centers and secondary service providers like hospitals. However, when introduced in 2005, the program did not lower households’ out-of-pocket spending on health and later in 2009 when Jamkesmas was revised and delegated to the Ministry of Health it again did not appear to reduce total out-of-pocket healthcare costs (see below). …but Jamkesmas utilization speci�cally by poorer and more vulnerable households has lagged. The weaknesses in targeting that allowed non-poor households to access Jamkesmas bene�ts combined with greater prior exposure to the healthcare system led to greater impacts for non-poor households. Larger increases in utilization rates for non-poor households occurs whether visits are to primary or tertiary service providers and whether the services are publicly or privately provided, but is especially pronounced for inpatient services in private care providers. Jamkesmas impacts, especially for poor households, are limited partially because of underprovided support operations like socialization and outreach… Jamkesmas does not have a bene�ciary induction or outreach process beyond the initial socialization process that takes place during the consultation and veri�cation stage (mentioned above) which itself was done haphazardly. Bene�ciaries are not provided enough advance information to encourage take-up 7 JAMKESMAS Health Service Fee Waiver or allow effective program use; what information they do receive usually comes from the service providers themselves. Likewise, an appeals and grievances process is provided for on paper, but bene�ciaries are generally unaware of how to use it and those complaints that are lodged are not resolved. The implementing agency spends an inordinate amount of time collecting and processing claims, leaving it little time for effective monitoring of other program processes. …the proliferation of competing regional health schemes… A signi�cant number of districts (335 out of 498 in 2010) are providing Jamkesmas-like plans at no additional cost to households. In many cases, these Jaminan Kesehatan Daerah (Jamkesda) provide equivalent or even enhanced bene�ts for anyone not covered by Jamkesmas, encouraging those who are covered by Jamkesmas to switch to Jamkesda-provided services. …and the indirect costs of receiving health care services that are not addressed by Jamkesmas. Jamkesmas has had very little effect on out-of-pocket healthcare spending; poor Jamkesmas users have seen the largest utilization increases at the providers which are nearest and cheapest: public primary health care centers, or Pusat Kesehatan Masyarakat (Puskesmas). Jamkesmas was not developed to encourage supply-side improvements in either quantity or quality and for many participants it remains just as costly as before to get to the door of a health service provider. Diagnosis and doctor-recommended treatment plans are not affected by Jamkesmas and poor households are often urged to use brand-name instead of generic drugs and undergo more-expensive-than-necessary diagnostic testing. Health service providers �nd Jamkesmas dif�cult and costly to implement resulting in fewer services provided, and funds spent, on Jamkesmas bene�ciaries. Local regulations regarding public health center management often conflict with Jamkesmas mandates, leaving health service providers confused and unwilling to use Jamkesmas funds to provide Jamkesmas bene�ciaries with planned services. Both public health centers and hospitals report dif�culty in generating the plans of action (health centers) or claims for services provided (hospitals) that must be submitted before Jamkesmas funds can be disbursed by the central level. Both providers also indicate that veri�cation of those submissions is delayed, resulting in underfunding for most of the year combined with a large disbursement of Jamkesmas funds at the end of the year. The reports on usage and management of Jamkesmas (made by service providers to Jamkesmas implementation teams) on which the Jamkesmas monitoring and evaluation process depends are often not made because of a lack of funds or quali�ed personnel or both. The future costs of an improved Jamkesmas program have not been adequately publicized and Jamkesmas’ �nancial, �scal, and political sustainability is uncertain. The Jamkesmas fee waiver is paid for with current government revenues; there is not yet a self-sustaining, risk-pooled Jamkesmas fund �nanced through contributions or co-payments. Current budget formulation for Jamkesmas does not incorporate the �nancial burdens that an adequately- provided or non-supply-constrained program would impose. Currently, the per-member cost of providing Jamkesmas services is not based on actuarially-projected rates of demand for healthcare services nor on projected increases in the cost of such services. Current estimates for a program operating with fewer supply-side constraints imply a per-member Jamkesmas cost approximately twice as large as the government is currently providing. More so than other household social assistance programs, the �nancial burdens and therefore the �scal and political sustainability of the Jamkesmas program are not well understood. 8 9 1. Background Improving the health status of its citizens has long been a goal of the Government of Indonesia, but poor households lack the same low-cost access that non-poor households enjoy. Continuous improvements in health access and outcomes for all citizens have long been a focus of Government of Indonesia social policy… Health insurance has been available in Indonesia since at least 19601; in the 1990s the GOI began experimenting with large-scale programs providing health care services targeted speci�cally to poor households. The �rst Indonesian health card (kartu sehat, circa 1994) provided poor families with free curative health care at community health centers, (Puskesmas) and referral care in 3rd class inpatient wards at district hospitals.2 The next kartu sehat iteration, Jaring Pengaman Sosial Bidang Kesehatan (JPS-BK) (Health Safety Net), was introduced during the AFC in 1997-1998. It provided capitation grants directly to the public health providers (village-level providers including midwives, Puskesmas, and hospitals with 3rd class beds) who were meant to make available free curative, preventative, outpatient, inpatient, and mother-and-child care to cardholders.3 During the early 2000s, a set of compensatory programs 1 See “Social Assistance Program and Public Expenditure Review 8: History and Evolution of Social Assistance in Indonesia� in this collection. 2 Kartu Sehat did not provide incentives for Puskesmas staff to identify bene�ciaries or to accomplish card distribution; no formal means testing was involved; and the program was not well-marketed to the target population group; all of these factors contributed to low program coverage and utilization. 3 Private providers were not included in the scheme. Quantitative work suggests that while targeting was pro-poor, there was also considerable leakage to the non-poor (see Sparrow et al, 2008). Also, even though the program did appear to result in a net increase in the use of outpatient care by poor households with JPS-BK cards, because many non-poor cardholders switched from private to public providers, this crowding out meant most of the bene�ts accrued to the non-poor (see Pradhan, Saadah and Sparrow, 2007). 10 came packaged with a GOI decision to reduce fuel subsidies; the health services portion of these programs provided free inpatient care for the poor at public district hospitals, free hepatitis B vaccines for poor children, and free generic drugs. During a further reduction in subsidies in 2003 funds were added to existing initiatives to cover both referral inpatient care and basic health services at health centers.4 …but take-up and coverage of health protection schemes have been stubbornly low. Estimates from recent household survey data indicate that less than half of the population is covered by any form of health insurance.5 The social health insurance scheme for civil servants, Asuransi Kesehatan (Askes), and the program for formal private sector employees,6 Jaminan Sosial Tenaga Kerja (Jamsostek) are the second and third largest health insurance programs (after Jamkesmas) but together cover no more than 10 percent of Indonesians. There are also private insurance schemes (covering less than 3 percent of the population) and some small community health insurance programs (around 1 percent or less); see Figure 1. Jamkesmas coverage has doubled in the past 5 years to over 15 million households (2010), which gives Jamkesmas bene�ts to approximately 29 percent of all Indonesians (according to nationally representative household survey data). Nevertheless, approximately 60 percent of all Indonesian households are still without any formal health insurance. 4 See � Social Assistance Program and Public Expenditure Review 8: History and Evolution of Social Assistance in Indonesia� in this collection for more information on the history of social assistance initiatives in Indonesia. 5 Rokx et al. (forthcoming). 6 Since 2006, Jamsostek has been required by law to cover the informal sector as well and the Ministry of Labor’s Jamsostek TK-LHK (Informal Workers Social Protection Scheme) is meant to cover informal workers. Program growth has been slow, however, and as of March 2010 it covered only 240,000 people (including members and dependents). Jamsostek TK-LHK bene�ciaries are included in the overall Jamsostek number above. 11 JAMKESMAS Health Service Fee Waiver Figure 1: Health 70 Insurance Millions of Households 60 Coverage7, 2005 50 and 2010 36 40 39 30 20 1 15 7 10 2 3 3 5 5 5 0 Source: Susenas 2005, 2010 Note: In 2005, a household is considered “covered� if at least one of the household members reports having (a particular type of) insurance. In 2010, a household is considered “covered� by Jamkesmas if the respondent reports that the household can access Jamkesmas. Poor households �nd it dif�cult to cope with health-related setbacks and are less well-insured against these risks than the general population. Poor and near-poor households generally have less pre-natal healthcare, higher rates of unattended birth, and higher rates of child malnutrition and report more frequent household welfare losses “due to health factors�.8 These health events are a more frequent source of loss for poor households than even unemployment, which is itself often a consequence of consecutive or cumulative negative health events. Jamkesmas is a health service fee waiver developed to improve utilization of available healthcare services while reducing healthcare costs for targeted poor and vulnerable households. Jamkesmas gives cardholders the ability to acquire preventative, curative, and catastrophic health care services without fees. Jamkesmas cards are targeted to poor and vulnerable households and cover virtually unlimited use of all available healthcare services and facilities.9 Jamkesmas’ bene�t package can encourage cardholders to acquire any and all services their diagnosis requires without additional cost. Jamkesmas, though targeted to poor and vulnerable households, brings social bene�ts as well: when it promotes healthy households, keeps students active, alert, and participating in their education, and returns adults to work sooner, the resulting increase in a healthy and productive population bene�ts all Indonesians. This report assesses the effectiveness of Jamkesmas to determine where and how reforms could lead to better outcomes for poor households. Quantitative analysis can determine if Jamkesmas is reaching poor households and how successful Jamkesmas has been in increasing utilization and lowering out-of-pocket (OOP) costs. Qualitative information on program delivery and program operations, gathered from a large set of stakeholders, will shed light on areas which would bene�t from reform. As Jamkesmas expansion is being considered as one pathway to universal health insurance coverage in Indonesia,10 the report also discusses the �scal implications of an expanded Jamkesmas and the �nancial sustainability of the current program. 7 In Figure 1, the Jamkesmas category includes households covered by older health insurance schemes oriented towards the poor, such as the kartu sehat or JPK-Gakin (Jaminan Pemeliharaan Kesehatan Keluarga Miskin), that existed prior to the introduction of Jamkesmas (as Askeskin) in 2005 and which have continued to exist in parallel to Jamkesmas. The Jamkesmas category may also contain households covered by the local Jamkesmas top-ups known as Jamkesda. Though Jamkesda coverage is rapidly increasing (see below) and coverage of most of these other schemes is estimated to be low (see Rokx et al.), the total coverage of these alternatives cannot be reliably disaggregated from Jamkesmas coverage. Therefore, the absolute numbers of households with certain Jamkesmas coverage is actually somewhat lower than the numbers presented in Figure 1. 8 See World Bank (2006) or Rand (2011). 9 On paper (in regulations and technical manuals), Jamkesmas even covers treatments not available in the cardholder’s area by paying for referrals and transfers. In practice, such service is not likely available. 10 In 2004 the GOI passed into law a universal health care framework (along with four other areas of social insurance) and in 2009 re-iterated that expanded Jamkesmas coverage would be part of the drive to cover all citizens with a health care scheme. 12 2. Objective, Program Size and Bene�t Adequacy Jaminan Kesehatan Masyarakat is a tax-�nanced fee waiver entitling cardholders to free in- or out-patient care at hospitals and primary care centers. Jamkesmas is a tax-�nanced fee waiver11 entitling members to free in- or out-patient care at hospitals and primary health centers. Jamkesmas has been available since 2006 (as Asuransi Kesehatan Masyarakat Miskin or Askeskin), providing health insurance for the poor and near-poor, targeted through the application of a means test, with a comprehensive package of free health services and bene�ts. The implicit value of bene�ts received depends on a bene�ciary’s or household’s actual utilization of covered health services. The program is �nanced by the central government (from current revenues) and does not require any insurance contributions or cost-sharing on the part of bene�ciaries or local governments. Figure 2: 6,000 100 Target Number of Beneficiaries Jamkesmas (Millions of Cardholders) Expenditure 4,500 75 and Rp billion Coverage, 3,000 50 2006-2010 1,500 25 0 0 Sources and Notes: Kemenkeu and Kemenkes data. With over 70 million cards distributed, Jamkesmas is most likely the largest initiative delivering free, subsidized, or insured healthcare in Indonesia. According to of�cial data 76.4 million poor and near-poor Indonesians – about a third of the population – are covered by the Jamkesmas program making it the largest Indonesian health services scheme. The national socioeconomic survey of households (Susenas) in 2010 con�rms that approximately 15 million households – and therefore about 60 million individuals assuming all household members can equally access the Jamkesmas card for health care utilization – report having access to Jamkesmas in 2010.12 Coverage of schemes similar to Jamkesmas but delivered by regional governments – the Jamkesda mentioned above – is estimated at an additional 27.5 million individuals.13 Jamkesmas consumes a signi�cant amount of the central government’s budgets for both health and household-targeted social assistance. The bulk of the Jamkesmas budget is derived by multiplying the Jamkesmas “premium�14 by the number of cards distributed. After increasing signi�cantly in 2007, mainly due to an expansion 11 While Jamkesmas’ objectives include providing �nancial protection from health costs, it was not established, and does not operate, on an insurance basis, with pooled contributions re-distributed according to veri�able outcomes. As a tax-�nanced transfer, its operation is similar to the rest of the Cluster 1 programs. Presidential Decree number 15 (2010) gives to the Of�ce of the Vice President the mandate to coordinate the acceleration of poverty reduction programs and Jamkesmas is explicitly mentioned as one such program. 12 Household coverage (as measured in the Susenas survey) may differ from of�cial estimates of the number of cards distributed for at least 3 reasons: (1) Susenas survey weights may not reflect the correct probability of contacting a Jamkesmas household; (2) households themselves may be Jamkesmas cardholders but mistakenly report coverage by Jamkesda (or any other similar-to-Jamkesmas scheme) or may think they are covered even though they are not cardholders; and (3) not all distributed cards have actually reached bene�ciary households (see Section 6 below). The same discrepancy between of�cial bene�ciary numbers and household survey (Susenas) bene�ciary numbers also occurred in BLT: of�cial bene�ciary numbers of BLT recipients in 2005 were 19.1 million households while the number of bene�ciaries according to the Susenas surveys from that year was 15 million households. 13 Soewondo, P., 2010. “Universal Health Coverage: Cost Estimation.� TNP2K Presentation �les. 14 Currently set at Rp 6250 per person per month, or approximately US$ 0.70 (at November 2011 nominal exchange rates). 13 JAMKESMAS Health Service Fee Waiver in bene�ciary numbers15, central government spending on Jamkesmas has largely stabilized in nominal terms and actually declined in real terms. In 2010, Rp 4,763 billion was spent on the program, equivalent to around one �fth of total Kementerian Kesehatan (Ministry of Health, Kemenkes) spending. As a household-targeted social assistance (SA) scheme, it is also a large program: Jamkesmas expenditures are second only to Raskin and account for almost 20 percent of all central government resources devoted to SA programs.16 Nearly 88 percent of Jamkesmas expenditure is allocated to hospital-based care and one tenth is allocated to primary care provided through local health centers (Table 2). The remaining expenditure goes towards operational management and to PT Askes, which updates and manages the distribution of membership cards.17 Table 1: Of�cial name: Jaminan Kesehatan Masyrakat (Jamkesmas); prior to 2008: Jamkesmas Asuransi kesehatan untuk keluarga miskin (Askeskin) at a Glance Program type: Health Service Fee Waiver Program Type and inaugural year Permanent, tax-�nanced, 2006 (start/usage year) Coverage National (100% provinces, 100% districts) Of�cial Number of bene�ciaries 76.4 million individuals (2010) Of�cial value of bene�t Potentially unlimited coverage of all healthcare services consumed; GOI budgets costs of Rp 6,250 per cardholder Public expenditure (2010) Rp 4,763 billion (US$ 524.62 million) Administrative cost per recipient Rp 9,362 (US$ 0.9) Percent of poor households covered 41% (2010)* Key policy and executing agency Kemenkes (Ministry of Health ) Key implementation agencies (role) Kemenkes (processing of fee claims, veri�cation of bene�ciary lists), Hospitals and Local Health Centers (service providers, fee claims), Support operations partners (role) BPS (targeting and eligibility); PT Askes (printing and card distribution to local governments); Kemenkeu (Ministry of Finance) (silent fund disbursement) Local Government participation Socialization, card distribution, monitoring and evaluation *Susenas 2010 Kemenkes is the key policy and executing agency for Jamkesmas. Jamkesmas is recorded as two separate activities in �nancial reporting documents. Each activity has its own budget and is implemented by a separate Directorate General (DG) within Kemenkes.18 The DG for Community Health is responsible for executing the budget, verifying bene�ciary lists and processing fee claims from and authorizing payments to health centers. The DG for Medical Services performs the same functions for hospitals and is also responsible for transferring funds to PT Askes for its management of membership card distribution. In addition, an ad hoc team in the Center for Health Financing and Risk Protection under the Secretary General (SG) is responsible for overall management, coordination and safeguarding activities related to Jamkesmas.19 The service providers included in the Jamkesmas program network are mainly government-owned facilities20 and are 15 The increase in spending in 2007 is also partly explained by the MoH decision to temporarily allow poor households to access Jamkesmas services without displaying a Jamkesmas card (open membership), which led to a spike in hospitalization. This decision followed problems with PT Askes’ distribution of Jamkesmas cards. 16 Excluding expenditures on the GOI’s emergency and temporary unconditional cash transfer, BLT. BLT is not a continuous program and has been distributed twice, in 2005 to 2006 and again from 2008 to 2009. 17 The budget distributed to PT Askes was recorded under transfers to hospitals and attached as a professionals / services fee. 18 Starting in 2011, the budgets for the two activities/interventions were uni�ed under the DG for Medical Services, although the functions remain separated under the two respective DGs. However, the management and safeguarding budget remains under the Secretary General. 19 The safeguarding responsibilities consist of activities such as administration, targeting, follow up on complaints, socialization, training and monitoring and evaluation. 20 Private providers are allowed and currently account for roughly 30 percent of all facilities , but a smaller share of total hospital beds, in the Jamkesmas network. 14 paid through different schemes: local health centers are paid by capitation,21 while hospitals are reimbursed for inpatient services on a fee-for-service basis and, from 2009, these payments use a Diagnostic Related Group (DRG) provider payment mechanism. Table 2: 2006 2007 2008 2009 2010 Jamkesmas Total Jamkesmas (Nominal, Rp billion) 3,074 4,567 4,448 4,620 4,763 Expenditure - Hospitals 1,696 3,402 3,600 3,535 4,168 Summary - Health Centers 1,350 1,073 647 888 496 - Operational management 27 89 125 134 99 - Bene�ciary cards (PT ASKES) 3 77 63 63 Analytical series: Total Jamkesmas (Constant 2009 prices, Rp bn) 4,375 5,843 4,817 4,620 4,409 Total Jamkesmas (US$, Rp million) 336 498 456 446 525 Share of Kemenkes budget (%) 25 29 28 26 20 Share of central government SA spending (%) 10 32 13 17 18 Memo items: Target number of bene�ciaries (Million) 60.0 76.4 76.4 76.4 76.4 Monthly premium per bene�ciary (Rp) 5,000 5,000 5,000 5,000 6,250 Sources and notes: Kemenkes, Kemenkeu, BPS and World Bank staff calculations. Local health centers are allotted Jamkesmas funds based on the size of the populations they serve and general activities (Figure 3). As mentioned above, local health centers receive Jamkesmas funds on a capitation basis. Jamkesmas budgets are determined by Kemenkes, in collaboration with the Dinas Kesehatan, based on the number of poor in a Puskesmas area, the type of Puskesmas (with or without inpatient care services), and the geographical characteristics of the region in which the Puskesmas operates. Before receiving Jamkesmas allotments, each Puskesmas must submit a schedule of activities that includes promotive, preventive, curative, and rehabilitation services. After the plan is approved by the Dinas Kesehatan, Jamkesmas funds can be withdrawn by the Puskemas from the local branch of PT Pos (National Post Of�ce system). Figure 3: APBN Jamkesmas Flow of Funds in Puskesmas KPPN Kemenkes: Bina Kesehatan JPKMM Management Team Masyarakat (Binkesmas) Bank KPPN Funds accountability Dinkes Kab/Kota PT. Pos POA Proposal Jamkesmas Fields of work Puskesmas Management Team Recommendation letter POA approval for PT. Pos Funds utilization Funds accountability & POA proposal for next year Source: CHR-UI – 2010 21 Starting in 2011, health centers must instead make claims for fees to the local district Health Of�ce (Dinas Kesehatan). See Section 6 formore on the burden of reporting in Jamkesmas-network health care providers. 15 JAMKESMAS Health Service Fee Waiver Hospitals are allotted Jamkesmas funds based on the number of poor in the catchment area as well as the type of hospital, which is based on the availability of medical services. The flow of funds from the central level to hospitals also varies based on hospital ownership status (see Figure 4). For the �rst type of public hospital – a non-Badan Layanan Umum Daerah, non-BLUD, or a “non-regional public service agency� – Jamkesmas funds (for claims submitted) proceed from the Kemenkes budget to the hospital’s bank account. However, non-BLUD agencies receive operating funds through local budgets, and budget management rules are also partially determined by local government (according to principles developed by the Satuan Kerja Perangkat Daerah, SKPD, or Sub-national Government Task Force), so a non-BLUD public hospital is required to retribute its Jamkesmas funds to the local government’s treasury.22 In the second type of public hospital – a Badan Layanan Umum Daerah, BLUD – there is no obligation for Jamkesmas funds to be retributed to local government treasuries, so BLUD public hospitals have more (but not complete) autonomy over the use of Jamkesmas funds. Finally, private hospitals also receive Jamkesmas funds through deposits to their bank accounts, and they retain complete autonomy over Jamkesmas fund use. Figure 4: APBN Jamkesmas Flow of Funds in Health Safety Net Management Hospitals KPPN Team DPA RSUD Bank KPPN Feedback Hospital Bank Independent Account APBD 1. Public Hospital 2.Private Hospital 1a.RSU Non-BLUD: 1b.RSU BLUD: The funds are managed Funds utilization The funds are managed entirely by the hospital and Funds Local Treasury for payment for entirely by the hospital. No can be used for operational accountability medicines, blood obligation to deposit the funds as account stock, ambulance, funds into local treasury receivables. laboratorial, etc. Remaining funds shall be deposited into the local treasury. Source: CHR-UI – 2010 Hospitals claims are reviewed and paid on the basis of a Diagnostic Related Group (DRG) provider payment mechanism. Hospitals and staff are asked to compile claims using an Indonesia-speci�c DRG software. Once compiled, an independent veri�er – each hospital, public or private is assigned one (or more) independent veri�er, who works as a government contractor – reviews the claims to verify that the correct claiming procedures have been followed and that the correct data and documents have been submitted. Claims are then forwarded to the Center for Health Financing and Risk Protection at Kemenkes, where claims are reviewed, payments are approved, and feedback to hospitals (on the appropriateness of procedures prescribed to Jamkesmas patients) is provided. 22 Before retributing Jamkesmas funds, some public non-BLUD hospitals spend Jamkesmas funds received on general operating costs (medicines and other medical supplies, laboratory materials, ambulance services) before sending on remaining funds to local treasuries – see Section 6 below. 16 Implementation and support operations are shared among a host of agencies. Funds are channeled to providers directly from the Treasury via PT Pos or commercial bank accounts once authorization letters are received from Kemenkes. Other central government agencies involved include BPS (targeting) and Kementrian Komunikasi dan Informatika (Ministry of Communications and Information Technology, Kemenkominfo) (socialization). In addition, local governments also get involved in socialization, card distribution, and monitoring and evaluation activities and receive some funds from the SG via the deconcentration fund mechanism (see also Sections 3 and 6 below for more details on all these implementation and support processes).23 Figure 5 below summarizes the Jamkesmas institutional arrangements and flow of funds mechanisms. Figure 5: Jamkesmas Kemenkes Flow of Funds, Overall DG of Community DG of Medical Services Secretariat General Health Secretary of DG - Secretary of DG Fees for PT Askes Center for Health Planning Unit -Planning Unit (membership cards) Financiing and Risk Protection (P2JK) Centrally executed Administrative PT. POS Bank BRI Transfer to PT Askes administrative cost costs for local governments (through deconcentration fund) Direct transfer to Direct transfer to Puskesmas Hospital Free health services to Free health services to Sources and Notes: World Bank staff based on DG Manuals. 23 Deconcentration funds were mainly used for monitoring and evaluation (in selected districts and provinces), socialization and coordination, and also program and activity formulation for that year. These expenditures are not consistently made in all regions in all years. 17 JAMKESMAS Health Service Fee Waiver 3. Targeting Jamkesmas card distribution is pro-poor, but large numbers of poor households are excluded from the program and many non-poor are included. Jamkesmas reaches households everywhere in Indonesia. The precursor to Jamkesmas – Askeskin – printed enough cards to distribute to approximately 10 million households in 2005 and 2006.24 When responsibility for implementation was transferred from PT Askes (a state-owned insurance provider) to Kemenkes and renamed Jamkesmas (2008), additional cards for new bene�ciaries were printed and distributed to regions, increasing the of�cial number of bene�ciaries (by 2010) to about 19 million households (and 76.4 million individuals) in all provinces. Such broad coverage means Jamkesmas is the primary health insurance for over 70 percent of all Indonesians with any health insurance and nearly 30 percent of the entire Indonesian population. Targeting and card distribution are decentralized operations that rely on lists of poor households from 2005 (Figure 6). Individuals hold cards, but targeting is done at the household level. Kemenkes determined district-level quotas based on 2005 list of poor households constructed by BPS for BLT,25 with the number of near-poor and below households (19.1 million) on the list being multiplied by an average size of four members to estimate the number of individuals to be covered by the Jamkesmas program (76.4 million).26 Kemenkes then partnered with local governments to do veri�cation checks of the selected households to get more complete information on all individuals in a household. Standard operating procedure was for district of�cials (including from the Jamkesmas program), in consultation with village and lower-level government of�cials, authorities, and health service facility administrators to verify, update, or revise the list of households selected to receive cards. Figure 6: BPS National Quota Kemenkes Jamkesmas Membership and Card PT. Askes PSE 05 District Quota (Central) Distribution BPS District Regent/Mayor Data validation by Camat, Lurah, RT/RW, Midwives, Cadres Decree of Regent/Mayor Askes Regional Card Printed Card distribution by RT/ Cardholder Midwives/Cadres Source: CHR-UI, 2010 24 Jamkesmas coverage recorded in the 2005 Susenas (7 million households; see Figure 1) could differ from the of�cial number of bene�ciaries (based on the number of cards printed and distributed to regions) for a variety of reasons: see Section 2 and footnote 12 above. 25 A temporary unconditional cash transfer program; please see “Social Assistance Program and Public Expenditure Review 2: BLT� in this collection. 26 However, Sparrow et al. (2008) state that the districts themselves were not only responsible for identifying eligible participants, but also for counting them; that is, determining the district quota. Decentralized targeting procedures, local revision to of�cial methods, and a lack of monitoring of the bene�ciary selection are common across all the social safety net initiatives covered in this series, but the latter two characteristics are more noticeable in Jamkesmas, Raskin, and the cash transfers for vulnerable groups JSLU, JSPACA, and PKSA. See “Social Assistance Program and Public Expenditure Review 3: Raskin� and “Social Assistance Program and Public Expenditure Review 7: JSLU, JSPACA, and PKSA “in this collection as well as a companion piece, “Targeting Poor and Vulnerable Households in Indonesia�. 18 Decentralized targeting procedures were prone to revision and resulted in targeting errors. The consultative process described above was not tightly controlled. In some areas, of�cials and interested parties consulted the 2005 BPS list of the poor and others the Badan Koordinasi Keluarga Berencana Nasional (Family Planning Coordination Agency, BKKBN) criteria in order to identify those in their communities who should receive Jamkesmas cards. A number of communities followed neither BPS nor BKKBN lists of the poor while village midwives and Puskemas cadres often selected bene�ciaries based on their own criteria, searching for instance for females with infants (regardless of economic status) or households where the primary wage earner faced large health care costs or an extended visit to a healthcare provider.27 The bene�ciary veri�cation procedure was supposed to involve household representatives (so household rosters with names, ages, and addresses for all members could be correctly generated) but in many cases veri�cation teams neither conducted home visits nor invited household heads to village meetings. This resulted in frequent distribution errors later (see Section 6 below). Poor socialization hampered targeting as well, with village leaders often not knowing who the intended targets were.28 Delays in card distribution meant that individuals could use older versions of kartu sehat cards or a letter certifying poverty (surat miskin) from their village head to access Jamkesmas services,29 although this was prohibited in 2008.30 Individuals without local residence cards, often the poor and local migrants, were not eligible.31 To date there has been no major update to Jamkesmas bene�ciary data, which is based on lists compiled in 2005 containing socio-demographic characteristics within households in that year, and which are known to have excluded many then-poor and vulnerable households.32 The resulting distribution of cards was modestly pro-poor. Table 3 demonstrates that households with a Jamkesmas card are sometimes better off than an average poor Indonesian household. For example, access to safe water and sanitation is higher, and the number of household members lower, for an average Jamkesmas recipient. Other categories such as remoteness – captured by the presence of schools as well as the “rural� indicator – are higher in Jamkesmas households than average poor households. Table 3: % of all % of poor % of Characteristics Indonesians population Jamkesmas within Indonesian who: who: recipients who: Populations, 2009 Do not have access to bottled, tap, or well water 19 30 25 Do not have access to private sanitation 35 58 51 Live in villages without a primary school 1 2 3 Live in villages without a junior secondary school 36 42 43 Live in rural areas 52 63 65 Live with more than 5 household members 27 43 28 Have less than primary education 19 32 34 Are illiterate 9 15 13 Work in agriculture sector 38 59 48 Sources: Susenas 2009, Podes 2008, World Bank calculations. Note: “Work in …� refers to shares of working individuals, not all Indonesians. 27 Neither was the consulting group tightly controlled. In most cases, the composition of the consulting group depended on who was available, who had been assigned from by higher-level agencies or political bodies, and who was interested. Furthermore, not all districts actually managed a consultation and selection process but simply tried to �nd the same households that BPS had used to calculate district-level quotas. See Sparrow et al. (2008), Son and Sparrow (2009), SMERU (2010), or ICW (2008)not in ref. Both lists mentioned (BPS and BKKBN) use proxy consumption indicators to identify the poor, although the BPS list uses a broader and more appropriate set of indicators. See this collection’s companion report “Targeting Poor and Vulnerable Households in Indonesia�. 28 Sparrow, Suryahadi and Widyanti (2008). 29 Sparrow, Suryahadi and Widyanti (2008). 30 Son and Sparrow (2009). 31 Son and Sparrow (2009). 32 See Targeting Poor and Vulnerable Households in Indonesia (World Bank, 2012a) 19 JAMKESMAS Health Service Fee Waiver Table 4 demonstrates that poor households not receiving Jamkesmas (incorrect exclusion) are more likely to be urban, less likely to be working in the agricultural or informal sectors, and have higher levels of education. Non-poor households receiving Jamkesmas (incorrect inclusion) have similar levels of education and work in the formal sector as often as excluded poor households, but they are more frequently urban and less often in agriculture. These same non-poor households who are included in Jamkesmas are less well educated, less formal, less urban, and more agricultural than non-poor households who did not get Jamkesmas. In other words, non-poor households receiving Jamkesmas are not so different from most excluded poor households, except they are more urban and less often in agriculture, while they are worse off than most non-poor households without Jamkesmas. These patterns, which indicate that expenditure poverty alone does not identify recipients of social assistance bene�ts, especially in urban areas, is common across the national household-targeted programs. Table 4: Poor households Non-poor households Characteristics non- non- within Jamkesmas Jamkesmas Jamkesmas Jamkesmas Jamkesmas and non- Household or Household head is/has: (% of households) Jamkesmas Primary school or less 80 71 69 44 Populations, Agricultural Sector 60 56 48 30 2009 Formal Employee 17 22 23 38 Urban 30 33 39 58 (average among households) Household members 4.6 4.6 3.5 3.6 Child dependency ratio 61 58 39 38 Source: Susenas & World Bank calculations. Jamkesmas targeting shows room for improvement (see Section 6 below for dif�culties encountered in Jamkesmas implementation including targeting). Figure 7 demonstrates that while total Jamkesmas recipients are similar to the number of poor and near-poor, coverage among the targeted group of households (the poorest three deciles according to household expenditure) is 39 to 50 percent while coverage rates among the non-targeted groups declines slowly from 36 percent in the 4th decile to 7 percent in the 10th decile (the richest 10 percent of Indonesian households by expenditure), resulting in 23 percent coverage in non-target households. As with most Indonesian social assistance programs with a negotiated community targeting procedure, female-led households are more likely to be bene�ciaries of the program than male-headed households with the same economic status. Likewise, rural households more often end up with Jamkesmas cards than urban households with similar (adjusted) expenditure levels. In a comparison of targeting outcomes, and with 100 percent representing perfect targeting according to program design, Jamkesmas was at 16 percent better than random (while the other major programs, BLT and Raskin, were at 24 and 13 percent, respectively).33 Usage of Jamkesmas is relatively constant across deciles, with about one in three cardholders (from any decile) reporting using free health services in the last six months (see Section 4 and Figure 8 below). 33 That is, targeting outcomes under Jamkesmas (BLT, Raskin) are 16 (24, 13) percent better than if the same number of bene�ts had been distributed randomly. See Targeting Poor and Vulnerable Households in Indonesia (World Bank, 2012a) for more detail. 20 Figure 7: 100 Jamkesmas National 90 Coverage by Urban Percentage of Household Covered subgroup and 80 Rural decile, 2010 Female headed household 70 60 50 40 30 20 10 0 Source: Susenas 2010 21 JAMKESMAS Health Service Fee Waiver 4. Impact Jamkesmas does encourage health service utilization, but positive impacts are larger for non-poor households. Jamkesmas cardholders do not make copayments and there is no explicit cost-sharing required of local governments. These bene�ts are more generous than even Askes for civil servants, which requires copayments for inpatient care. For services provided to Jamkemas bene�ciaries, hospitals are reimbursed on a fee-for-service basis with payments based on a DRG mechanism34 and primary-care providers are paid a per-capita grant based on forecast use by the number of bene�ciaries in a catchment area.35 Local governments are not explicitly delegated any cost-sharing36 or provision of human or physical capital. The Jamkesmas provider network is mostly limited to public health facilities but private providers and facilities can be accredited for inclusion in the Jamkesmas provider network.37 Poor households frequently face negative health-related events but spend relatively little on health care services, so a generous fee-waiver system is expected to increase cardholders’ utilization... Poorer households in Indonesia report slightly higher rates of acute respiratory infection (ARI) and diarrhea as well as higher rates of work and school disruptions from illness (Table 5).38 If Jamkesmas were not available, poor households would be expected to have the highest out-of-pocket costs as a share of total expenditure given the demographic characteristics in their households (Table 5).39 Despite this, health service utilization – either outpatient visits or number of days of inpatient care – is generally lower for poor households than non-poor households40 and poor households generally spend less (both absolutely and as a share of total expenditure) than rich households on health care (Table 5). All this suggests that a health care fee waiver like Jamkesmas could provide a substantial bene�t to poor households. …Though major supply-side constraints may temper any expectations regarding Jamkemas’ ability to effectively deliver all desired healthcare services to poor households. The very generous de jure bene�ts promised are signi�cantly different from the services actually available on the ground within the Jamkesmas network of providers. Healthcare provider shortages – of hospital beds, doctors, specialists, diagnostic tools, medicines, biomedical equipment and devices, and other services – are a serious issue in Indonesia.41 Not only does this mean that cumulative Jamkesmas claim costs are well below what they would be if all promised services were available and if the program were actively promoted and bene�t packages were explained to members,42 it also means the de facto bene�t provided by Jamkesmas is signi�cantly lower than the de jure bene�t; see Section 6 below for more detail.43 34 In a “Diagnostic Related Groups� system, payments are made based on the average cost of treating a given diagnosis. If the provided treatment costs less than the DRG-set payment, the provider usually keeps the excess payment. If the provided treatment costs more, the provider receives only the DRG-set payment. 35 See Section 2 above for more on Jamkesmas expenditure flows. 36 As mentioned above, non-BLUD hospitals retribute Jamkesmas funds to regional general treasury accounts, but this is not a cost-sharing arrangement per se and is a result of regional government regulations (not Jamkesmas operating procedures). 37 Private hospitals currently account for 30 percent of hospitals in the network, but this 30 percent comprises a much smaller share of the total beds in the network and they are found mostly in urban areas in Java. 38 Table 5 is based on information from Sparrow et al. (2008). See also World Bank (2006), which notes that health shocks are a frequent cause of welfare losses in poorer households. 39 Expected OOP health expenditures are estimated by statistically predicting what households with certain demographic, health/illness, and location characteristics would spend if they had the incomes, knowledge, and health-seeking behavior of Indonesians who are not constrained by income (those in the 90th percentile or above of the Indonesia-wide expenditure distribution). See Sparrow et al. (2008) for details. 40 A slight quali�cation is necessary: rates of outpatient care in public health centers (Puskesmas) for poor households are double that for rich households (in 2006 (2009), 0.08 versus 0.04 visits last month (0.1 versus 0.056 visits), respectively, for quintile 1 and quintile 5 households). The overall higher outpatient rate in rich households comes from public hospital visits and private hospital and/or private doctor visits, which are several times higher in rich households. 41 See Box 1 in this note as well as Guerard et al. (2011). 42 This in turn may mean there is very little incentive to improve quantity or quality of, or access to, services available as total Jamkesmas costs would be much higher in the absence of existing supply constraints. 43 Access costs – transportation, food, lodging, chaperones, child care, lost wages – are not addressed by Jamkesmas, so any implicit reduction in actual bene�ts drives up relative costs of using the Jamkesmas card for acquiring healthcare services. 22 Table 5: Per capita ARI Influenza Diarrhea Any health Illness Outpatient Inpatient OOP health share (%) Rates of expenditure complaint disrupted visits per days per of total household quintile work or month year expenditure Health school Events, incidence (%) household per capita actual predicted Health panel Cross section panel Facility 2005 2005 2006 2007 2005 2005 2005 2008 Utilization, 1 17.0 12.4 2.5 26.6 27.4 29.7 16.3 0.12 0.04 1.7 11.8 15.2 and OOP 2 17.0 12.2 2.1 26.8 27.9 30.8 16.2 0.14 0.05 1.8 7.2 9.8 health 3 16.5 12.0 1.9 27.4 28.5 31.6 16.4 0.16 0.07 1.9 5.2 7.0 expenditure 4 16.2 11.7 1.5 27.6 28.9 32.1 16.2 0.18 0.10 2.2 3.6 4.6 shares 5 15.4 13.7 1.4 26.6 28.1 30.4 14.2 0.19 0.17 2.8 1.9 1.9 Sources and Notes: Susenas, various years, and World Bank staff calculations. Predicted OOP health expenditures are based on Son and Sparrow (2009). Utilization rates for outpatient and inpatient services have increased after fee waivers were made available…. In 2005 when the Jamkesmas predecessor Askeskin was introduced, the new program lead to increases in rates of outpatient utilization of approximately .05 to .06 visits per household member (per month).44 Also in 2005, inpatient utilization went up overall by approximately 0.1 days per household member (per year) as a result of household access to fee waivers.45 For both outpatient and inpatient care, the bulk of the increases were due to increased visits to public facilities. Between 2007 and 2009, when the number of Jamkesmas cards distributed increased by over 16 million, increases in Jamkesmas coverage increased district-level inpatient utilization rates by approximately 0.1 visits per person per month on average.46 …and at both primary and secondary facilities… Following the introduction of Askeskin in 2005, all households with cards increased the number of visits to primary care centers (Puskesmas, local health clinics, and local hospitals) by 0.09 visits per person per month (an increase in primary care utilization of nearly 50 percent). For the poorest 20 percent of Indonesian households, the likelihood of a primary care visit increased by an even larger amount (relative to their pre- Askeskin rates). At secondary facilities (hospitals), Askeskin increased the likelihood of household visits by 0.8 visits, which is a nearly 50 percent increase over pre-Askeskin rates.47,48 …but poor households bene�tted least. All of the positive increases in utilization due to the introduction of Askeskin in 2005 mentioned above – outpatient, inpatient, primary, and secondary care – are larger for households in the upper deciles of the expenditure distribution than for poorer households in the lower deciles. For example, Askeskin in 2006 encouraged an extra 0.04 to 0.06 public facility outpatient visits per person per month for the poorest 20 percent of households, but encouraged an extra 0.08 to 0.9 of these same visits in the richest 20 percent of households (and an extra 0.09 to 0.12 of these visits in the 2nd-richest 20 percent of households). In inpatient care larger bene�ts accruing to non-poor households are even more noticeable: in the richest quintile households acquired an extra 0.9 days per year of inpatient care (public or private) because of Askeskin while Askeskin households in the 2nd poorest quintile consumed 44 Except where otherwise noted, discussion in this section of Askeskin/Jamkesmas impacts during and after the year 2005 refers primarily to results from Sparrow et al. (2008) and Son and Sparrow (2009). 45 During the 2005 introduction of Askeskin/Jamkesmas there was a simultaneous introduction of the temporary direct cash transfer to households, Bantuan Langsung Tunai (BLT). As mentioned above, Askeskin targeting was based to some extent on BLT targeting procedures and lists of bene�ciaries. BLT cash transfers also led to statistically signi�cant increased outpatient utilization rates – see Sparrow et al. (2008). Increases in health care utilization due to BLT in 2008 were more pronounced for uninsured households, where uninsured includes those not receiving a Jamkesmas card, and for outpatient services – see �Social Assistance Program and Public Expenditure Review 2: BLT� in this series. 46 Outpatient utilization rates also increased on average where Jamkesmas coverage increased, but estimated impacts from increased Jamkesmas coverage are not signi�cantly different from zero. The difference-in-differences (DD) regressions of district-average changes in utilization rates on changes in district Jamkesmas coverage rates also control for baseline levels of coverage, exposure to epidemics and natural disasters, changes in rate of reporting of illnesses, and province �xed effects. Data for the regressions comes from a district-level panel created from the Susenas household survey in years 2007, 2008, and 2009. 47 Son and Sparrow (2009). Over all deciles from poor to rich, there is a signi�cant reduction in the number of secondary care visits by Askeskin/ Jamkesmas cardholders over time. 48 This pattern is con�rmed in cross-section correlations using 2009 data. In particular, in cross-section, outpatient care utilization, inpatient care utilization, and total number of days of inpatient care consumed are larger in poor households with Jamkesmas than poor households with no insurance. These same poor Jamkesmas households have a much higher public to private facility-visit ratio than either insured poor households or uninsured poor households. In cross-section in 2009, this is especially true for inpatient care visits. See Rokx et al. (forthcoming). 23 JAMKESMAS Health Service Fee Waiver only an extra 0.08 days per year.49 Increased inpatient days are split roughly equally between public and private providers for non-poor Jamkesmas households while poor Jamkesmas households favor public providers.50 Over the period 2007 to 2009, which includes the large increase in Jamkesmas coverage mentioned earlier, the poorest 30 percent of Indonesian households did not appreciably increase either inpatient or outpatient utilization rates in districts with increased Jamkesmas coverage.51 Additional impact analysis using the IFLS panel survey – which is representative of approximately 80 to 85 percent of the Indonesian population and two waves of which span the period during which Askeskin was introduced – �nd approximately zero increase in the probability (over prior-to-Askeskin probabilities) that a household with Askeskin would visit either a public or private healthcare provider for outpatient or inpatient care.52 Unexpectedly, rates of utilization of Jamkesmas cards are approximately equal between covered poor and non-poor households… As discussed above, Jamkesmas targeting is pro-poor, but only mildly so, so it is no surprise that some non-poor households are able to leverage Jamkesmas fee waivers to increase utilization of healthcare service. Figure 8 demonstrates that while Jamkesmas cards cover nearly 30 percent of the population, the overall utilization rate for those with Jamkesmas is approximately 11 percent. In the poorest 30 percent, middle 40 percent, and richest 30 percent of households, about 39, 38, and 34 percent (respectively) of those with coverage used Jamkesmas for healthcare services in the past 6 months. Figure 8: Jamkesmas percent Utilization by decile, 100 2010 90 Jamkesmas Coverage 80 Utilization (conditional on membership) 70 60 50 40 30 20 10 0 1 2 3 4 5 6 7 8 9 10 Household Per Capita Expenditure Decile Source: Susenas 2009 …which may indicate that Jamkesmas is not effectively serving poor households. Table 6 below shows that the targeting of Jamkesmas is progressive in expected OOP expenditure (see above), meaning those households predicted to have higher healthcare costs based on demographics, location, and health status alone are covered by Jamkesmas at 49 Son and Sparrow (2009). The estimated coef�cient for the Jamkesmas impact on household inpatient care in the poorest 20 percent of households, 0.04 days per year, is positive but statistically indistinguishable from zero. This means that the poorest Jamkesmas households are consuming close to zero extra days of inpatient care, and the increase for the richest Jamkesmas households is an entire order of magnitude greater than that for poorer households. 50 Son and Sparrow (2009). 51 See footnote 46 for details. Here, changes in household utilization rates (and the district averages created from them) are calculated only for households in the 3rd decile or below of the Indonesia-wide expenditure distribution. 52 Giles and Satriawan (2012). IFLS-recorded rates of inpatient and outpatient utilization by poor and non-poor households prior to Askeskin agree broadly with Susenas records: there is a higher probability that non-poor households were using the healthcare system. Estimated impacts were slightly larger, but still not signi�cantly different from zero, for inpatient care in either poor households (from the bottom 40 percent of the IFLS consumption distribution) or all households. Increases in the probability of poor household inpatient utilization were also not statistically distinguishable from increases for all households. 24 higher rates.53 Given their higher expected OOP (see Table 5 above), a relatively larger share of covered poor households would be expected to utilize Jamkesmas to acquire free healthcare. Instead, conditional utilization rates indicate that the propensity to use Jamkesmas (when covered) is approximately equal across poor and non-poor households, and therefore across all levels (high and low) of predicted OOP health expenditure. Table 6: OOP health expenditure share % of households with Jamkesmas Jamkesmas quintile Actual OOP Predicted OOP Targeting 1 (low OOP share) 7.2 1.8 2 6.9 4.9 3 7.4 7.6 4 7.8 10.0 5 (high OOP share) 10.1 12.7 Sources and Notes: Susenas 2008 and World Bank staff calculations. Predicted OOP health expenditures are based on Son and Sparrow (2009). Jamkesmas does not seem to signi�cantly reduce poor households’ direct OOP costs… In 2005 households who received the new Askeskin program allocated 0.2 to 0.4 percentage points more of the household budget to health care. Within the poorest 20 percent of Indonesian households, those receiving cards increased the healthcare expenditure share by 0.3 to 0.5 percentage points over 2005 to 2006.54 Impact analysis using the IFLS panel survey indicates that over a slightly longer time period – 2000 to 2007, including the introduction of Askeskin in 2005 – Askeskin signi�cantly lowered the level of expenditure on health-related items for any household using Askeskin, but the expenditure-reducing effect was weaker (by approximately half) for poor households.55 More recently in a 2009 cross-section of Indonesian households, the poor and near-poor56 show approximately equal expenditure shares on healthcare regardless of health insurance coverage: shares are approximately 1.7, 1.8, and 1.8 percent for those without insurance, with Jamkesmas, or with other health insurance respectively.57 At the district level between 2007 and 2009 (including the signi�cant increase in Jamkesmas bene�ciaries during early 2009), increases in Jamkesmas coverage led to no change in average expenditure shares on OOP health care costs; among poor households only, there was also no change in OOP health care cost shares attributable to increased Jamkesmas coverage.58 …while observers note the supplementary costs of access are not reduced by Jamkesmas … From conversations with bene�ciaries in four Javanese districts, cardholders stated that having Jamkesmas did not noticeably reduce overall out-of-pocket costs incurred by using the healthcare system. They also noted that the costs of registration and transportation as well as fees for medicines or additional check-ups were signi�cant and not addressed by Jamkesmas. In addition, they reported that healthcare system utilization for serious or major health events often required the assistance of other family members or caretakers and that transportation and the cost of food and lodging for these additional participants was signi�cant and not covered by Jamkesmas.59 53 This may in part be due to the participation of local authorities, including health authorities, in the selection of Jamkesmas bene�ciaries (described above in Section 2 and 3 and below in Section 6): observers noted that individuals and households with existing or emerging health issues were often allocated Jamkesmas cards. 54 Son and Sparrow (2009) attribute rising healthcare expenditures to the combination of “lumpiness and indivisibility of care� together with reduced barriers to healthcare access, and therefore increasing utilization, for the poor. They also state that “Indirect barriers remain large for inpatient care, presumably due to high opportunity costs such as foregone earnings and travel distance to hospitals�. 55 Giles and Satriawan (2012). 56 Approximately the poorest 3 deciles of the Indonesian expenditure distribution. 57 For all poor households with at least one inpatient visit during the past year, healthcare expenditure shares with Jamkesmas (6.2 percent) are slightly lower than for households with no insurance (7.5 percent) or households with other insurance (6.7 percent). However, as mentioned above, the Jamkesmas impacts on inpatient utilization are far greater for nonpoor households than poor, suggesting that the slight reduction in OOP costs from Jamkesmas does not contribute to large cumulative decreases in OOP spending for poor households. 58 For all households, existing levels of Jamkesmas coverage in 2007 were associated with statistically signi�cant reductions (over 2007 to 2009) in OOP costs of roughly 0.3 percentage points. For poor and near poor households in the �rst three deciles of the expenditure distribution, 2007 coverage levels were associated with increases in OOP cost shares (from 2007 to 2009) of roughly 0.1 percentage points, though these estimated increases are not statistically signi�cantly different from zero. 59 CHR-UI, 2010. 25 JAMKESMAS Health Service Fee Waiver …and a lack of effective socialization and outreach has left many bene�ciaries unable to use the Jamkesmas program. Sections 5 and 6 below discuss the available funds for, and the extent and quality of, Jamkesmas socialization. Section 6 notes that expectations and knowledge among the Jamkesmas-card-holding population as well as lower-level government of�cials and health facility staff varies widely. Equally true is that local-level standard operating procedures and informal practices regarding Jamkesmas vary widely. This suggests that knowledge of, and expectations regarding, the Jamkesmas program are not wide-spread or consistent, which further suggests that previous experience with health care services is a pre-requisite for maximizing Jamkesmas bene�ts. Altogether, Jamkesmas appears to reinforce good behaviors among experienced poor and non-poor households but is less effective at encouraging use in poor households.60 The impact estimates discussed above – taken over the period when Askeskin was �rst introduced and also when Jamkesmas completed an increase in nationwide coverage – suggest that Jamkesmas has not yet obviously increased utilization or reduced the total burden of direct plus indirect spending on healthcare speci�cally for poorer households. However, IFLS data indicates that regardless of expenditure level, households with prior exposure to the healthcare system were more likely to use Askeskin, so poor households with �rst-hand knowledge and previous experience may have used the program to reduce the total cost of healthcare services they were anyway consuming. Jamkesmas may therefore encourage knowledgeable poorer households to spend on healthcare an amount closer to that which would be expected given their demographic characteristics, but if a main Jamkesmas objective is to bring new patients from disadvantaged households into the health care system, more must be done to facilitate this effort (see Section 6 below). 60 Compare this to results in Uganda, where total bene�ts from the re-abolition of health service user fees was greatest for the least well-off households; see Deininger and Mpuga (2004). 26 5. Cost Effectiveness Assessing the ef�ciency of Jamkesmas spending is challenging because of the exclusion of some administrative expenditure from the Jamkesmas budget and �nancial reports Assessing the ef�ciency of Jamkesmas spending is challenging because of the exclusion of some administrative expenditure from the Jamkesmas budget and �nancial reports. Unlike most other social assistance programs, Jamkesmas is unique in having a clear economic classi�cation de�nition in the budget, allowing a more straightforward examination of budget implementation. Nonetheless, it is dif�cult to fully quantify administrative costs for the Jamkesmas program as some important administrative spending is excluded. For example, high-level budget data for Jamkesmas by economic classi�cation de�nes almost all spending on Jamkesmas as social assistance expenditure, mainly dominated by transfers to hospitals and Puskesmas. Expenditure classi�ed as goods, services, and capital is very low, suggesting very minimal administration costs of approximately 4 to 5 percent of program expenditure. However, this system of classi�cation excludes central government civil servant salary expenditures which are recorded under each Secretariat Unit within each DG involved in implementing the program rather than under the Jamkesmas budget. Complicating matters, �nancial reporting covers centrally-executed spending only and excludes spending on administration by local governments (funded from deconcentration funds or own budget) or heath service providers (hospitals and Puskesmas). The lack of a single comprehensive report covering all spending complicates ef�ciency analysis, and calculation of ef�ciency indicators based on existing data only could be misleading as it may underestimate total administrative costs. Estimates indicate that administrative costs for the Jamkesmas program are low, although this excludes the costs incurred by health service providers and is therefore likely an underestimate. Including estimated salary expenditure and deconcentration funds,61 Jamkesmas administrative costs are low and have declined after coverage was increased in 2007. For example, the average annual administrative cost per bene�ciary was approximately 20 US cents in 2010 (down from nearly 1 US dollar in 2006), while the overall administrative overhead ratio was just 3 percent (down from 18 percent previously). International experience indicates this is a relatively low administrative overhead ratio: a study on health programs in eleven countries concluded that a minimum administrative cost to administer such programs was between 9 and 10 percent.62 However, the data compiled here does not include real administrative costs borne by hospitals and local health centers in support of the program, so it likely underestimates the actual comprehensive administrative cost of Jamkesmas. 61 Due to lack of information in the LAKIP on salary expenditure and transfers to PT Askes, Kemenkeu budget realization data is used to calculate these costs. Salary expenditure is estimated by dividing the total salary expenditure within one DG evenly by the total number of Satker/Directorate in that DG. In addition, the membership management fee paid to PT Askes was assumed to be a professional and/or services cost underneath goods and services expenditure. 62 See Gwatkin (2000). 27 JAMKESMAS Health Service Fee Waiver Table 7: 2006 2007 2008 2009 2010 Spending Unit Cost (Total spending/no. 51,226 59,783 58,223 60,474 62,338 Ef�ciency bene�ciaries, Rp) Indicators, Administrative costs per bene�- 9,117 1,556 3,006 2,685 2,229 2006-2010 ciariy (Non-bene�ts/No. bene�- ciaries, Rp) in US$ 0.90 0.20 0.30 0.30 0.20 Administrative overhead ratio 18% 2% 5% 4% 3% (Non-bene�ts/Total spending) Memo items: Of�cial bene�ciaries 60,000,000 76,400,000 76,400,000 76,400,000 76,400,000 Number of civil servants n.a. n.a. n.a. n.a. n.a. Value of annual Health Services 42,108 58,226 55,216 57,788 60,109 (Rp) Total spending (Rp bn) 3074 4567 4448 4620 4763 o/w Bene�ts 3046 4451 4243 4478 4592 o/w Non-bene�ts* n.a. 116 203 142 170 o/w Civil servant salaries n.a. 4 6 8 8 o/w Admin/other n.a. 98 141 86 105 o/w Targeting n.a. 1 15 7 8 o/w Follow up n.a. 0 0 0 0 o/w Socialization n.a. 3 9 6 7 o/w Training n.a. 3 20 14 16 o/w Training n.a. 3 20 14 16 o/w Evaluation (M&E) n.a. 6 12 22 26 Sources and Notes: World Bank staff calculations based on Financial Accountability Reports (Laporan Akuntabilitas Keuangan Pemerintah or LAKIPs) and Directorate General Data. *While original budget data allocates personnel expenses to the DG as a whole, expenses for each individual Activity/Directorate have been divided evenly. The administration budget is highly fragmented and additional resources for administration could enhance the effectiveness of the Jamkesmas program. The small budget for administration is fragmented among different units and functions, raising concerns about overlapping responsibilities, coordination and economies of scale. Of the Rp 170 billion total in non-bene�t spending in 2010, 60 percent was accounted for by the central government, leaving 40 percent for local governments (via deconcentration fund), which in some regions was used to execute activities similar to those being done by Kemenkes at the central level. From the same total, the Jamkesmas program consumed funds for the following activities: civil servant salaries (Rp 8 billion), administration and other costs (Rp 105 billion), targeting (Rp 8 billion), socialization (Rp 7 billion), training (Rp 16 billion) and monitoring and evaluation (Rp 26 billion). Of particular note is the amount spent on socialization, which represents just 4 percent of all non-bene�t costs or less than 1 percent of total program costs, and on follow-up, which is not usually funded. These outreach activities are noticeably ineffective (see Section 4 above and Section 6 below); part of the reason may be the lack of budgetary attention these activities receive. 28 6. Implementation Program support operations – initial targeting and veri�cation, socialization and complaint and grievance systems, and a monitoring and evaluation cycle – are not yet provided regularly or with suf�cient quality. Veri�cation and monitoring of bene�ciary selection is not done systematically. As mentioned above, the selection of bene�ciaries is actually done at the local level and practices and personnel involved vary from region to region. Final results are not generally monitored, evaluated, or addressed by the Jamkesmas program administrators at the central level. Field researchers have noted that some local Jamkesmas selection teams did not update or verify either the 2005 BPS lists used to assign households to the Jamkesmas program nor collect appropriate data on households selected for Jamkesmas but outside the BPS lists. In addition to a lack of updating of demographic or economic changes to bene�ciary households and a system to accommodate such updates, neither has there been an update at the national level of the Jamkesmas registry. So, national data does not capture, for example, the new addresses of long-time Jamkesmas members. The number of cardholders registered by PT Askes is at 72 million, or about 4.5 million fewer bene�ciaries than cards printed. Potentially eligible households were not able to provide a bottom-up check on data validation and card distribution because of a lack of socialization of both the Jamkesmas program and eligibility rules. The �rst round of socialization usually occurs (informally) with the �rst distribution of Jamkesmas cards or even later when a cardholder visits a healthcare service center and is informed by healthcare service staff of the parameters of the program. Card distribution varies from village to village also, so possibilities for informal socialization and interaction with knowledgeable stakeholders are not the same everywhere. The delegation of bene�ciary selection and the lack of follow-up or veri�cation of the selections have led to abuse and malfeasance in some areas. For example, nepotism, close relationships between village of�cials and non-poor Jamkesmas recipients, and political considerations have all led to exclusion and inclusion error in the Jamkesmas program.63 Furthermore, the agency responsible for printing the cards (PT Askes) is responsible for checking lists of bene�ciaries produced by local-level selection against the national list of poor households produced by BPS and if there are unexplained disagreements, PT Askes will occasionally refuse to print cards for the locally-suggested list. This also leads to exclusion and inclusion error. Socialization of program goals and rights and responsibilities to both providers and bene�ciaries has been insuf�cient. Because of the variation in the local-level selection process for Jamkesmas cardholders mentioned above, socialization of these procedures required considerable advance planning. Field research found that the objectives of the data collection and validation activities were not widely socialized to enumerators or potential bene�ciaries.64 Socialization of program content and included services as well as goals and the rights and responsibilities of Jamkesmas cardholders should accompany card distribution (at the latest) and ideally should be re-iterated frequently. Field research indicates that bene�ciaries most frequently received such information from health center and hospital staff only after visiting the provider. Figure 9 below indicates that Jamkesmas socialization activities have overall not effectively produced a knowledgeable cardmember base: the majority of cardholders do not truly understand the bene�ts to which they now have access. For both inpatient and outpatient service types, the overwhelming majority (usually over 80 percent) of bene�ciaries either did not know or were misinformed about coverage of select services or types of care. Figure 9 also demonstrates that information regarding fees and charges for medicines are not well-publicized. 63 See CHR-UI (2010) for more on this issue in Jamkesmas and either the report from World Bank (2012a) Targeting Poor and Vulnerable Households in Indonesia or RAND Urban Poverty study for more on perceptions of, and reasons for, mistargeting in the large Cluster 1 social assistance initiatives. 64 See CHR-UI,2010 29 JAMKESMAS Health Service Fee Waiver Figure 9: At Puskesmas, Inpatient Services At Hospital, Inpatient Services Perception Laboratory 3rd class of Inpatient services of Extent of Medicines Medium and Major surgery Jamkesmas Inpatient ICU/ICCU Bene�ts Birth delivery Birth delivery Emergency unit response Emergency unit response 100% At Puskesmas, Outpatient Services At Hospital, Outpatient Services Medical services Laboratory Radiology & electromedic Dental treatment Dental treatment Medicine Pre - natal cares Blood services Family planning services Pre - natal cares Sources and Notes: adapted from Indonesia Corruption Watch (2008). *All treatments, services, diagnostics, and medicines listed above are of�cially covered according to Jamkesmas regulations. Weakly provided socialization and recruitment activities likely explains a large portion of the pro-rich distribution of the Jamkesmas impacts on health service utilization (see Section 4 above). With very little experience with many of the providers now accessible under Jamkesmas, and with very little knowledge about the Jamkesmas program goals and their own eligibility, as well as about the services covered and the service providers’ rights and responsibilities and a Jamkesmas card holder’s rights and responsibilities, poorer households likely use Jamkesmas less frequently than richer households who have more experience with health care consumption. Underprovided socialization may have led to only small changes in health care access, which is a precondition of utilization. Monitoring and evaluation of the bene�ciary experience or service provider treatment and behavior towards bene�ciaries has not happened. Field researchers have noted that Jamkesmas cardholders have experienced longer- than-normal queues or delays in service because the service provider needed to verify Jamkesmas status or took too long to provide a referral; discrimination in favor of regularly-paying patients; and charges for covered medicines. Bene�ciaries have complaints regarding quality of service generally. The same research team found that service providers have experienced delays in funding; have too few resources to complete Jamkesmas-mandated reports; and too few additional staff and supplies to accommodate increased demand from Jamkesmas cardholders which leads to a lowering in the quality of care. The Jamkesmas implementing agency is not always aware of the frequency and depth of these particular issues partly because monitoring activities for bene�ciaries and service providers are not consistently funded. There is also no system for the “red flags� mentioned above to be incorporated into a follow-up system that can suggest solutions and provide remedies. Bene�ciaries and service providers have complaints and grievances but they are not addressed systematically. Potentially eligible bene�ciaries without cards frequently desire explanations regarding their status while those with Jamkesmas cards most often ask why other household members have not been allowed to access Jamkesmas services. As veri�cation and updating procedures are not automated, Jamkesmas cards frequently contain small errors in a bene�ciary’s name, address, or identity card number (among other things). These errors often cause the service provider to ask for supporting documentation like a reference letter from a lower-level health provider or a government-issued identi�cation card. This delays service and increases costs for bene�ciaries, which they note when making complaints. As mentioned previously, service providers are sometimes not able to keep up with increased demand and the consequent decrease in quality and late payment or delays in claims processing and funding. Finally, the regional of�ces of the implementing agency note that they too have not been given any extra resources to handle all of the processes that support Jamkesmas (claims payment and auditing; bene�ciary support and monitoring; health facility monitoring and auditing; and others). Complaints that are received are not incorporated into a framework for addressing them in real time and incorporating the accumulated complaints into revisions in operating procedures or policy. Notwithstanding the sources of complaints mentioned above, bene�ciaries appreciate the care provided to them through the Jamkesmas program. Figure 10 below shows Jamkesmas cardholders’ levels of satisfaction with different personnel available at most all healthcare service providers. However, it should be remembered that not 30 all personnel are available everywhere and patients are occasionally rejected because of a service provider’s inability to perform the relevant procedures. Earlier it was pointed out that Jamkesmas bene�ciaries are not consistently aware of what rights, responsibilities, and levels of quality of care they should receive. In that sense, high levels of satisfaction with unanticipated free medical care should be expected. Figure 10: Patient's level of satisfaction of health care services Satisfaction with Jamkesmas Services nurse Satisfied doctor Not Satisfied medical staff Inddifferent/Not know 0% 50% 100% Source: adapted from ICW (2008) The accumulation of unused Jamkesmas funds indicates that basic health service providers are having trouble with the conflicting local and national regulations regarding Jamkesmas and health provision generally.65 Often there are local regulations indicating that basic health service provision must be free for all residents. With such regulations in place, service providers, local governments, and regional health of�ces are hesitant to utilize all available Jamkesmas funds for providing services for Jamkesmas bene�ciaries only, and as a result Jamkesmas fund absorption (by providers) can be as low as 10 percent. Other local regulations might stipulate maximum wage rates for basic health service staff (who are treating patients under a universal free basic health service) which are then applied when a Jamkesmas bene�ciary receives treatment; the locally-regulated rate is often below the Jamkesmas-mandated maximum; this also limits Jamkesmas fund utilization by local governments and service providers. In another example of the effects of disharmonized regulations on the accumulation of unused funds, local regulations may stipulate that, for example, extension and outreach services (beyond the physical Puskesmas location) must be funded from a particular source such as Bantuan Operasional Kesehatan (BOK, or Health Service Operational Assistance) and therefore Jamkesmas funds may not be used for those same activities. When such activities account for a large part of Puskesmas costs or are a major part of Puskesmas-proposed Jamkesmas activities (see above), the Puskesmas is forced to let Jamkesmas funds accumulate unused. Jamkesmas network hospitals report the opposite problem: claim review and disbursement procedures take too long, leaving them underfunded for much of the year. Some hospitals indicate that the claim generation procedure is a burden and they are not well-equipped to complete claiming procedures rapidly enough. In some cases, hospitals are not organized or administered well enough to generate, track, and archive the claims necessary for re- imbursement. Hospitals have also reported that they have neither the requisite number of DRG software operators, nor a high degree of con�dence in the staff delegated to the DRG and claims-generation processes; this can lead to either insuf�cient remuneration for services provided (because the claims submitted incorrectly reflect the actual procedures performed) or “up coding�, where claims submitted reflect higher-value procedures than those that were actually provided.66 In addition, most hospitals feel the claim veri�cation procedure, which involves both an independent veri�er and the central-level Center for Health Security Financing (PPJK), takes too long and disrupts the gradual and timely disbursement of operational funds, leaving the hospital underfunded for most of the year; this impression is con�rmed in budget disbursement rates discussed below in Section 7. 65 See the report from Center of Health Research at University of Indonesia (CHR UI, 2010) for more details. The qualitative study was conducted in 2 provinces (South Sumatra and West Java) and includes focus groups discussions, in-depth interviews and direct observations involving many informants from various levels: bene�ciarieis, communities, local implementation partners, hospital and health center staff and authorities, PT Askes staff, and staff from Provincial and District Health Of�ces (among others) . The study involved approximately 120 informants, seven hospitals, and four health centers. 66 Private hospitals have also indicated that the INA-DRG re-imbursements (from PPJK) are lower than actual costs of services provided, leading to the same hospitals asking for “cost sharing� from Jamkesmas patients to cover the gap. 31 JAMKESMAS Health Service Fee Waiver Monitoring and evaluation of claims and service-provider payment processes is underprovided. When there are long delays in claims processing from slow veri�cation or claims payments at the central level or when advance capitation allocations are unused, under-used, or mis-used, an emergency monitoring team is put into action to resolve the delay or utilization problem. However, a regular monitoring schedule of all claims, service-provider payment procedures, or service provider or local-level funds utilization has not been put in place and required reports (from hospitals and Puskesmas to Jamkesmas teams at the district level) are often not submitted. Both the service providers themselves and the of�ces of the implementing agency at the local level (Dinas Kesehatan) have claimed that they lack the funds, personnel, and expertise to complete reporting activities. The independent claims veri�ers assigned to monitor claims and the claims process in Jamkesmas hospitals are often asked (by either local governments of Dinas Kesehatan) to switch to Jamkesda claims veri�cation instead; this contributes to delays in the Jamkesmas claiming, veri�cation, and funds disbursement procedure as well as to increased opportunities for “up coding�, prescribing more expensive medications, and other costly practices. Despite accumulating dif�culties, media opinion regarding has continued to be favorable, which is mirrored in bene�ciary sentiment. The average trend in media evaluation of Jamkesmas – determined by a comparison of the number of positively-toned to negatively-toned Jamkesmas articles appearing in print media – improved slightly over 2008 to 2009 (Figure 11). In 2008 sentiment was already quite high and positive articles with a focus on Jamkesmas (instead of just mere mention) represented 78 percent of all Jamkesmas-focused articles; by 2009 the same share had improved to 79 percent.67 Figure 11: Jamkesmas Raskin 1,500 1.00 Media Perceptions of 0.80 Average Media Sentiment* 1,000 Jamkesmas, 0.60 BLT, and 0.40 Number of Articles Raskin 500 0.20 0.00 0 -0.20 -500 -0.40 -0.60 -1,000 -0.80 Positive Mention (LHS) Positive Focus (LHS) Average Mention (RHS) Negative Mention (LHS) Negative Focus (LHS) Average Focus (RHS) Sources: World Bank, media analysis Notes: Articles are divided into “mention� (mention the program in passing) and “focus� (in-depth treatment). Articles are evaluated as positive (including neutral) or negative in sentiment. Positive sentiment takes a +1 and negative sentiment a -1, making average sentiment between -1 (all negative) to +1 (all positive). The majority of media attention in Jamkesmas has been on program implementation (Figure 12). Jamkesmas has avoided most of the political controversy and fund allocation questions surrounding other large programs like BLT and Raskin (for example). Most media coverage has instead focused on implementation issues, in particular socialization and targeting (which are also major BLT issues) as well as supply-side issues (a major Raskin issue) and the distribution of Jamkesmas funds. Given the dif�culties mentioned above, continuing media coverage in these areas may mean that the Jamkesmas program “honeymoon period� may be drawing to a close and program administrators may need to work harder to solve current Jamkesmas constraints in order to satisfy bene�ciaries and the public at large. 67 From a range of -1 to 1, with -1 meaning all articles are unfavorable, and 1 meaning all articles are favorable or neutral. 32 Figure 12: Program effectiveness 73 91 Number of Program extending 8 Jamkesmas Policy print media Program expansion 9 articles by Fund allocation 22 24 topic Dependency and exit strategy 4 8 Program socialization 57 91 Targeting 80 64 Implementation Complaint handling 9 Distribution process 64 44 Supply issue 56 37 Fraud and corruption 3 20 M&E 18 15 2008 2009 Interplay Political Campaign related 14 22 Political economic biases 05 0 100 150 200 Sources: World Bank, media analysis 33 JAMKESMAS Health Service Fee Waiver 7. Public Financial Management and Sustainability Several different management issues threaten the program’s longevity. Budget execution rates in Jamkesmas have declined signi�cantly in recent years despite the increases in Jamkesmas coverage and allocations – see Figure 13a. In 2009, the government revised its membership system and limited fee waivers to those poor households with Jamkesmas cards only; previously-accepted alternate forms of identi�cation no longer guarantee free access to healthcare services. Also, the number of regional healthcare schemes (Jamkesda) has been on the rise - there were 60 such schemes in 2008 and over 300 by 2010 - and they are collectively covering more and more recipients. Jamkesda schemes are both substitutes for and competitors of Jamkesmas-provided care, and bene�ciaries with a choice often choose the locally-provided Jamkesda. Finally, when DRG-based remuneration was instituted for hospitals, this limited spending by those providers; at the same time, the Puskesmas facilities often faced contradictory regulation (circa 2009) that limited the ability of the facilities to recoup their capitation-based Jamkesmas subsidies.68 Jamkesmas 2010 disbursement improved over 2009, but this was due primarily to reduced planned budget allocations; in other words, Jamkemas 2010 realized spending was relative to a noticeably lower planned level of expenditure. Jamkesmas budget disbursement has been skewed towards the end of the year in the past but has shown recent signs of improving – see Figure13b.69 In 2008, just 26 percent of the budget was disbursed in the �rst half of the year; by 2010, fully 50 percent of the budget was disbursed in the �rst half of the year. In 2008 and 2009, around 30 percent of the total budget was disbursed in the last quarter, whereas in 2010, 40 percent of the budget was disbursed in the same period. This disbursement pro�le reflects a number of factors. First, the lack of guidelines regulating provider claims submission has led providers to economize on administrative and processing costs by delaying reimbursement requests until they are able to make a one-time, cumulative request late in the budget year. Second, delays occur when Kemenkes veri�es submitted claims. Finally, after claim veri�cation, there is often yet another delay before the Kemenkeu Treasury Of�ce (Kantor Pelayanan Perbendaharaan Negara, KPPN) actually delivers payment. Regulations in 2009 obliged KPPN to deliver payments within one week of �nal claim veri�cation and this contributed to the recent improvements in the �rst-half disbursement pro�le. Figure 13a: Budget Execution Ratio by activity, Figure 13b: Budget Disbursement Rate, 2008-2010 2008-2010 Sources and Notes: World Bank staff calculations based on Directorate General Data 68 For example, some districts, which have authority over most social service spending, prohibited Puskesmas from accessing central-level Jamkesmas funds (provided on a per-capita basis). These districts argued that Puskesmas should not be allowed to charge fees (to either the government or to households directly) for services performed for poor households. Also see Section 6 above. 69 Here, disbursement is equivalent to the transfer of funds from Kemenkeu to services providers. 34 A proliferation of regional health insurance and free healthcare schemes may conflict with Jamkesmas and has increased demand-side pressures on public healthcare providers. These schemes, generically known as “Jamkesda�, for Jamkesmas Daerah, are in content modeled after the Jamkesmas scheme and provide free healthcare or healthcare insurance to local citizens.70 A review of select Jamkesda schemes71 noted that in “most of these schemes it is not clear how the local government subsidy is coordinated with [Jamkesmas].� The same study noted that when a city program started providing free care at public primary healthcare centers, those same public facilities experienced heavy increases in traf�c and funds spent on subsidized drugs, a reduction in staff time and resources available for public outreach (traditionally a large component of the services offered), and an increase in demand for unnecessary procedures or higher levels of service.72 Interviews with stakeholders in local governments managing Jamkesmas funds as well as funds for Jamkesda note that overlapping initiatives result in an accumulation of unused funds in the regional health account. Likewise, Jamkesmas inspectors sent from the central Kemenkes are sometimes obliged to perform claims veri�cation and other audit functions for the local Jamkesda scheme. Financial and budget planning for current and future Jamkesmas costs is inadequate and Jamkesmas’ legal status is also uncertain – see Box 1. Currently, the GOI allocates a Jamkesmas budget to Kemenkes that is equal to the number of bene�ciaries times a per-bene�ciary “premium�73 of Rp 6,250 (or approximately US$ 0.70) per month. The current per-capita funding scheme takes the low utilization rates and low demand for most of the current menu of medical procedures as given, even though impact analysis has shown that the Jamkesmas program itself leads to increases in both of those rates especially among previous system users. Without revisions to these costs and the method and frequency with which these costs are reviewed, demographic change as well as program maturity and increasing demand will not be reflected in program costs. The �nancial sustainability, and therefore the political life, of the program will be threatened. The government’s plans for universal insurance coverage will also matter for Jamkesmas. Currently, the social security law of 2004 provides for universal coverage for all Indonesians in a health insurance scheme as well as a pension scheme, and old age savings scheme, a workplace accident scheme, and life insurance. As the health fee scheme currently covering the single largest share of Indonesian households, Jamkesmas’ mandate is clearly affected by progress in the GOI’s universal insurance goals.74 Kemenkes currently acts as purchaser of services, program administrator, claims administrator (and auditor), regulator of services, and in most cases the health care provider. In addition to imposing a signi�cant burden on Kemenkes staff – who still ful�ll as many non-Jamkesmas functions as before – the integration of all functions in one agency does not produce incentives to achieve ef�ciency or reduce costs, monitor and promote quality or accountability, or advocate for bene�ciaries. For example, interviewed program staff at the regional level reported spending most of their available time processing claims and making sure claims payment proceeds on schedule. In addition, it creates dif�culties in decomposing the total amount of resources Jamkesmas requires. For example, the supply-side subsidies from Kemenkes to public hospitals do not show up as Jamkesmas costs and therefore Jamkesmas “premiums� do not currently account for the less expensive, GOI-subsidized price of care. 70 In 2008, there were approximately 36 district-level Jamkesda insurance and 65 free-healthcare-for-all schemes operating across Indonesia. There are also an unknown number of province-level and below-district-level schemes that are again similar in content to Jamkesmas or the Jamkesda schemes. Not all Jamkesda schemes prioritize poor households; some promote universal coverage or free health care for all citizens. All Jamkesda were together estimated to cover 2.3 million Indonesians in 2008. 71 Ascobat et al. (2009). 72 Several additional questions arise when two or more programs cover the same household with partially overlapping menus of service such as: (1) Are there household risks (in both health outcomes and the �nancial cost of health events) that remain unaddressed by both programs?; (2) How are the bene�t packages sustainable at the local level?; (3) How will local-level schemes be prevented from consuming (either directly or indirectly) resources, subsidies, and the civil servants and healthcare personnel meant to be dedicated to the national scheme?; (4) How is bene�t coordination between schemes organized?; (5) How is claims payment coordination between schemes organized?; and many others. In short, if managing one health insurance scheme is complicated, managing to coordinate two or more schemes and still produce good outcomes for bene�ciaries is at least twice as complex. 73 As mentioned previously, Jamkesmas is a tax-�nanced government expenditure without any of the traditional insurance elements like actual premiums, risk-pooling, or cost-containment mechanisms. 74 Also see Guerard et al. (2011) (no ref )for more on overlaps between universal health insurance plans and Jamkesmas. 35 JAMKESMAS Health Service Fee Waiver Box 1: Four factors may cause actual expenditures to signi�cantly exceed current budgeted expenditures: Jamkesmas 1. Coverage: Of�cial Jamkesmas coverage is now 76.4 million individuals but signi�cant inclusion and Financial exclusion errors remain. Given the frequent household entry into and exit out of poverty in Indonesia, Sustainability further coverage increases are likely as the newly poor enter the program while the non-poor or newly non-poor remain covered. 2. Bene�t package: On paper, the Jamkesmas bene�t package is the largest in the country, even better than the Askes or Jamsostek insurances. Its bene�t package is more generous than even those available under national health insurance programs in OECD countries. Virtually all primary, secondary and tertiary care procedures are covered, along with medications, lab tests, radiology, and other expensive diagnostic procedures. 3. Utilization: Many households eligible for Jamkesmas are not aware that they can join; some households with Jamkesmas cards are unaware of how to access the free healthcare system; others in remote areas cannot afford the cost of transportation to the nearest public health center. Slow rates of exit from traditional medicine also reduce demand for public health services. If Jamkesmas were actively facilitated, costs would increase substantially. 4. Supply side constraints: Current Jamkesmas costs from medical care provided are limited by inadequate service provision and perceived low service quality. The medical services available are not suf�cient to meet expected demand or the mix of services demanded (given the virtually unlimited bene�t package). Shortages of doctors, specialists, hospital beds and other services make paper bene�ts a promise in name only. The GOI likely has very little incentive to improve quantity or quality of services available (or access to that care) as total Jamkesmas costs would be much higher in the absence of existing supply constraints. Jamkesmas costs are thus likely to rise sharply in the absence of signi�cant reforms. Targeting should be improved and reviewed regularly so only eligible households remain covered. The bene�t package should be reduced both because the government should not promise services it cannot deliver reliably everywhere and the cost of actually delivering promised services would be signi�cantly higher than current costs. The government currently sets the cost of Jamkesmas at about Rp 6,500 per member per month, but estimates (which assume that current supply-side constraints remain) put the cost at Rp 12,000 per month or greater (see “Actuarial Costing of Universal Health Insurance Coverage in Indonesia�, Guerard et al (2011)). Proper data for more precise estimates of Jamkesmas costs in the absence of supply side constraints is not available. The political sustainability of the Jamkesmas program is also uncertain. Within Kemenkes and elsewhere in the government, there is strong support for expanding the program to the entire population to achieve universal health coverage. The tax-�nanced health coverage concept may be viable, but these plans contradict Law #40 of 2004 mandating the establishment of national health care via social health insurance. If Law #40 were implemented, those covered by Jamkesmas would become members of the social insurance system; the required contributions for many Jamkesmas members would be paid by the government; and Jamkesmas would be phased out. The current administration has not demonstrated a �rm commitment to developing the overall strategy, speci�c program design, �nancing mechanism, or the governance structure of the various bodies, funds, and insurance instruments that Law #40 implies. Consequently, social insurance, universal health care, and Jamkesmas all have uncertain political futures. However, as long as Jamkesmas intends to provide health service access to poor households, all aspects of the Jamkesmas program (bene�t package, administration and health care delivery, and �scal sustainability) need sustained improvement. 36 8. Summary and Recommendations Jamkesmas could serve an important risk-mitigation function for poor households if it performed outreach and encouraged healthy behaviours Jamkesmas could serve an important risk-mitigation function for poor households if it performed outreach and encouraged healthy behaviours. Poor households �nd health shocks to be one of the most serious risks to welfare, and coverage of either government- or privately-provided health insurance is low overall and especially low for poor households. There is evidence that poor households do not consume enough health services given their demographic characteristics. Jamkesmas utilization impacts are greatest for households who have greater familiarity with and lower cost of accessing health care systems and services while poor households are often confused regarding their own eligibility; covered bene�ts and procedures; the obligations of care providers; and their own rights when receiving care and other program details. In order for Jamkesmas to be effective for poor households, more effort and funds should be spent on facilitation and socialization and recruitment into the Jamkesmas network of providers. New bene�ciaries and old Jamkemas registries should be revised using the 2011 poor household registry and uni�ed database when it becomes available. Once household status has been veri�ed, the distribution of Jamkesmas cards should remain free from local-level revision and influence and should follow the 2011 registry. Jamkesmas inclusion and exclusion error should both improve once these revisions are completed. Jamkesmas needs to address the supplemental costs of health service access that constrain poorer households. Jamkesmas has not reduced out-of-pocket expenditures on health care-related costs overall, but there is evidence that it will as long as bene�ciaries are already using the healthcare system. However, bene�ciaries mention that �nding funds for travel, for hosting a companion or caretaker during a longer stay, and fees for name-brand medicines or more expensive procedures is dif�cult and discourages them from using Jamkesmas or any health care services. If Jamkesmas were proactive and strategized with community-level providers of goods and services like public transport and child care, it could better serve its client households and reduce overall costs of access. Jamkesmas must develop a strategy for ensuring that bene�ciaries get the quality care they need. This will require (1) monitoring of service providers and in particular their treatment plans and charged services, (2) establishment of a complaints and grievance system where bene�ciaries can report unsatisfactory services and have them remedied, and (3) better socialization of Jamkesmas bene�ts, goals, and rights and responsibilities so that informed bene�ciaries can provide grass-roots monitoring and appraisal of services provided under the Jamkesmas name. If monitoring from the top-down can be supplemented by reports from informed system users in a bottom-up approach, it is more likely that overcharging and unnecessary procedures will decrease. The Jamkesmas program should contain explicit formulations of both patient rights and minimum service standards for providers in the network. Currently, Jamkesmas does not address the supply of health care services nor the quality of care provided. Remote bene�ciaries cannot take advantage of Jamkesmas bene�ts because there is no nearby health care provider. Bene�ciaries mention that they prefer private providers (some of which are in the Jamkesmas network) to public because of the difference in the quality and standards of care. Jamkesmas bene�ciaries have mentioned being turned away in favour of non-Jamkesmas patients and of being made to wait longer for service. The Jamkesmas program should standardize services by formulating explicit patient rights and minimum service standards for providers in the network. Areas without facilities currently meeting those standards should develop, together with Jamkesmas of�cials, a plan for improvements that will allow them to meet Jamkesmas standards. Jamkesmas needs to invest in a much more comprehensive monitoring, evaluation, and remediation program. When local regulations regarding public health center management conflict with Jamkesmas mandates, Jamkesmas implementers should be made aware as soon as possible. Either new Jamkesmas regulations or revised local regulations should be formulated to allow the Jamkesmas program to serve as many bene�ciaries as possible without constraint. Reporting requirements for public health centers and hospitals should either be eased or Jamkesmas should include 37 JAMKESMAS Health Service Fee Waiver more funds for service providers to hire the personnel necessary to manage the current reporting requirements. Likewise, the Ministry of Health needs to address delays in the claims or service plan veri�cation procedure so that Jamkesmas funds can be disbursed smoothly and dependably. Finally, the veri�cation team should be enlarged and the monitoring of Jamkesmas fund usage and management should rely both on reports from hospitals as well as frequent (and unannounced) visits from Jamkesmas monitoring teams. The focus should be on discovering and describing dif�culties early, and working with service providers on feasible short and medium-term solutions. Jamkesmas needs to develop medium and long-term scenarios that are scienti�cally costed to ensure the program’s longevity. Households support the Jamkesmas program, but the political and �scal future of Jamkesmas is uncertain. If Jamkesmas continues to deliver bene�ts to poor households, clearer cost and implementation scenarios for the next two to �ve years should be developed, including a possible merger into a universal health insurance scheme. 38 References Bredenkamp, C., Tandon, A., Harimurti, P., Pambudi, E., Rokx, C. forthcoming. “Aren’t Good Intentions Enough? A New Perspectives on Indonesia’s Jamkesmas Program.� Jakarta, Indonesia: World Bank. Centre for Health Research. 2010. “Health Insurance in Indonesia: A Stakeholder Accountability Review of the Jamkesmas program.� Jakarta, Indonesia: University of Indonesia Davidson, R., Gwatkin, D., Wagstaff, A., and Yazbeck, A.S. 2005. “Reaching the Poor with Health, Nutrition, and Population Services: What Works, What Doesn’t, and Why.� Washington, D.C.: World Bank Deininger, Klaus and Paul Mgupa. 2004.. “Economic and Welfare Effects of the Abolition of Health User Fees: Evidence from Uganda,� World Bank Policy Research Working Paper 3276. Washington, DC: World Bank. Gani, A., H. Thabrany., Pujianto, F., Yanuar, T., Tachman, A. Sirergar, H., Wahyu, S., Soerachmad, Widyastuti, Yulherin, and Nurbaiti 2009. “Good Practices of Local Health Financing Schemes in Indonesia: Their Contribution toward Universal Coverage of Health Insurance.� AusAID. Jakarta, Indonesia. Giles, J. and Satriawan, E. 2012. “Evaluating Health Insurance for the Poor: Evidence from Indonesia Family Life Survey�, forthcoming. Washington, DC: World Bank. Grosh, M., del Ninno, C., Tesliuc, E., & Ouerghi, A. 2008. For Protection and Promotion: The Design and Implementation of Effective Safety Nets. Washington, DC: World Bank. Guerard, Y.,Wiener, M., Rokx, C., Schieber, G., Harimurti, P., Pambudi, E., Tandon, A. 2011. “Actuarial Costing of Universal Health Insurance Coverage in Indonesia: Options and Preliminary Results�. Health, Nutrition and Population HNP. Discussion Paper. Gwatkin, D. 2000.. The Current State of Knowledge about Targeting Health Programs to Reach The Poor. The World Bank Publications on Poverty and Health www.worldbank.org/povertyandhealth. Indonesia Corruption Watch ICW. 2008. “Study Policy Warranty Health Community Assurance.� Jakarta, Indonesia Corruption Watch. Pradhan, M., Saadah, F., Sparrow, R. 2007. “Did the Health Card Program Ensure Access to Medical Care for the Poor during Indonesia’s Economic Crisis?� The World Bank Economic Review. RAND. 2011. “Indonesia Urban Poverty Analysis and Program Review.� Draft report for upcoming World Bank study : Urban Poverty in Indonesia. Soewondo, P. 2010. “Universal Health Coverage: Cost Estimation.� TNP2K Presentation �les Son, H., Sparrow, R. 2009. “Impact Evaluation of Social Health Insurance for the Poor in Indonesia: A Panel Data Analysis.� Sparrow, R., Suryahadi, A. and Widyanti, W. 2008. “Public Health Insurance for the Poor: Targeting and Impact of Indonesia’s Askeskin Programme.� Working Paper. Institute of Social Studies. Sparrow, R., Suryahadi, A., and Widyanti, W. 2010. “Social Health Insurance for the Poor: Targeting and Impact of Indonesia’s Askeskin Programme.� SMERU Working Paper. The SMERU Institute World Bank. 2006.. Making the New Indonesia Work for the Poor. Jakarta, Indonesia:World Bank. World Bank. 2010. Indonesia Jobs Report: Towards Better Jobs and Security for All. Jakarta, Indonesia:World Bank. World Bank. 2011. “Indonesia Health Sector Review: Does Jamkesmas Protect the Population from Health Expenditure Shocks?� Jakarta, Indonesia:World Bank. World Bank. 2012a. Targeting Poor and Vulnerable Households in Indonesia. Jakarta, Indonesia: World Bank. World Bank. 2012b. Protecting Poor and Vulnerable Households in Indonesia. Jakarta, Indonesia: World Bank. World Bank. 2012c. “Social Assistance Program and Public Expenditure Review 1: Public Expenditure Review Summary.� Jakarta, Indonesia: World Bank. World Bank. 2012d. “Social Assistance Program and Public Expenditure Review 2: BLT.� Jakarta, Indonesia: World Bank. 39 JAMKESMAS Health Service Fee Waiver World Bank. 2012e. “Social Assistance Program and Public Expenditure Review 3: Raskin.� Jakarta, Indonesia: World Bank. World Bank. 2012f. “Social Assistance Program and Public Expenditure Review 4: Jamkesmas.� Jakarta, Indonesia: World Bank. World Bank. 2012g. “Social Assistance Program and Public Expenditure Review 5: BSM.� Jakarta, Indonesia: World Bank World Bank. 2012h. “Social Assistance Program and Public Expenditure Review 6: PKH.� Jakarta, Indonesia: World Bank. World Bank. 2012i. “Social Assistance Program and Public Expenditure Review 7: JSLU, JSPACA, and PKSA.� Jakarta, Indonesia: World Bank. World Bank. 2012j. “Social Assistance Program and Public Expenditure Review 8: History of Social Assistance in Indonesia.� Jakarta, Indonesia: World Bank. 40 Indonesia Social Assistance Program and Public Expenditure Reviews are policy notes summarizing current issues in the Government of Indonesia’s major household-targeted social assistance programs. Design, implementation, coverage, impacts, budgetary demands, cost effectiveness, stakeholder perceptions, and program history will be examined through qualitative and quantitative analysis for each program: conditional and unconditional cash transfers, subsidized rice distribution, a health fee waiver, cash scholarships, and cash transfers for the neglected elderly, disabled, and at-risk youth. Evidence-based appraisal can assist policymakers in evaluating programs and deciding whether and how they are achieving Indonesia’s social protection goals. Reviews 1 through 8 together comprise a companion volume to the “Protecting Poor and Vulnerable Households in Indonesia� report. “Protecting Poor and Vulnerable Households in Indonesia� provides a public expenditure review of the wider social assistance sector, summarizes results from Reviews 1 through 8, explores gaps in both population coverage and risk coverage of current social assistance programming, and offers recommendations and suggested reforms for achieving an integrated and coordinated household-based social assistance system. THE WORLD BANK, JAKARTA OFFICE Indonesia Stock Exchange Building Tower II/12th Floor Jl. Jend. Sudirman Kav. 52-53 Jakarta 12910 Tel: (6221) 5299-3000 Fax: (6221) 5299-3111 Website: www.worldbank.org/id