Philippines - Devolution and health services : managing risks and opportunities

Report No. 1 2343-PH Philippines Devolution and Health Services: Managing Risks and Opportunities May 23, 1994 Country Department I Population and Human Resources Operation Division East Asia and Pacific Region Office FOR OFFICIAL USE ONLY MICROGRAPHICS Report No: 12343 PH Type: SEC Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization  CURRENCY EOUIVALENTS ANNUAL AVERAGE 1986-1993 1986 US$1.00 = P. 20.4 1987 US$1.00= P. 20.6 1988 US$1.00 = P. 21.1 1989 US$1.00 = P. 21.7 1990 US$1.00 - P. 24.3 1991 US$1.0C = P. 27.5 1992 US$1.00 = P. 25.5 1993 US$1.00 = P. 27.5 FEICAL YA GovernMent - Jamnary I to December 31. ACRONYMS ALOS - Average Length of Stay API - Annual Rate of Parasitic Incidence ARI - Acute Respiratory infecdon BCG - Tuberculosis Vaccine BF - Promotion of Breastfeeding BHS - Bareagay Health Stations CARI - Control of Acute Respiratory infections CDD - Control of Diarrheal Diseases COA - Commission on Audit CHSA -Comprehensive Health Service Agreement DA - Department of Agriculture DBM - Deparment of Budget & Management DENR - Departmnent of Environment& Natural Resources DOH - Department of Health DSWD - Department of Social Welfare & Development EPI - Expanded Program of Immunization FHSIS - Field Health Service Information System FIC - Fully Immnized Child FNRI - Food & Nutrition Research Institute GOP - Government of the Philippines IDA - Iron Deficiency Anemia IDD * Iodine Deficiency IMR - Infant Mortality Rate IPHO Integrated Provincial Health OffiMe IRA - Internal Revenue Allotment LGC - Local Government Code LGU - Local Government Unit MH - Maternal Health MCH - Maternal Child Health NCHS - US National Center for Health Stastics NCR - National Capital Region NEDA - National Economic Development Authority NGO - Non Govenmment Organization NTP - National Tuberculosis Control Program ORS - Oral Rehydration Solution PEM - Protein-Energy Malnutrition PNG - Papua New Guinea RHU - Rural {ealth Unit TF-IMR - Task Force on Infant Morality Rate VAD - Vitanin A Deficiency  FOR OFFICIAL USE ONLY PHILIPPINES DEVOLUTION AND HEALTH SERVICES: MANAGING RISKS AND OPPORTUNITIES Table of Contents PaQe No. EXECUTIVE SUMMARY ............................. I. PERFORMANCE AND STRUCTURE OF THE PHILIPPINES HEALTH SYSTEM - PRE-DEVOLUTION. A. Introduction . B. Health Status. 2 C. Health Delivery System: Structure and Evolution .11 D. Public Health Programs Performance 22 E. Health Expenditure and Finance 31 P. Implications for the Devolved System 39 II. DEVOLUTION OF HEALTH SERVICES IN COMPARATIVE PERSPECTIVE . .42 A. Health Provisions of the Local Government Code of 1991 .42 B. Implementation Progress .46 C. Organizational Types of Decentralization in Comparative Perspective .47 D. Potential Economic Effects of the LGC Health Provisions .so III. FISCAL IMPACT OF DEVOLUTION .63 A. Introduction .63 B. Local and Central Fiscal Context 64 C. Local Government Fiscal Behavior 1991 65 D. Estimating the Fiscal Gap .67 E. Conclusions and Implications for Policy 80 Annex 3A: Determinants of Local Health Expenditure Regressions 82 Annex 3B: Determinants of Locally Raised Revenues Regressions. 85 Annex 3C: Interviews with Local Executives .88 IV. ORGANIZATION AND MANAGEMENT ISSUES UNDER DEVOLUTION 89 A. Introductin .89 B. Scope of Devolution Induced Change.... 90 Ts douent has a restictcd distnbution and may be used by ecipiens only in the peonnan of their oflicisl dutis Ils contents may oot othewise be disc1oed thout Wold Bon& authrzto.  C. Managing Change: Organizational Structure ........................... 91 D. Managing Change: Management Systems ... 95 S. Ma.-.aging Change: Basic Health Programs ............................. 98 P. Policy Objectives and Implementation.. 103 Annex 4A: Retained Programs ........... - 110 Annex 4B: A Comparative Note on Managing Decentralization: Chile's Health Services ............. 111 V. MAKING DEVOLUTION WORK: ISSUES AND OPTIONS 115 A. Introduction ..15............ 1 B. Net Assessment of Health Services Devolution ..15............ 1 C. Actionable Policy Issues: Near Term... 118 D. Actionable Policy Issues: Longer Term. 124 B. The Future Role of the DOH ............ 127 Annex 5A: Actionable Issues and Ir.terventions ................. 129 Annex 5B: Improving Provincial Hospital Efficiency .................... 133 Annex 5C: DOH-LGU Assistance Mechanism Design and Implementation Issues ........................ 140 Annex SD: Expanding an Underdeveloped Public Health Program - The Case of Nutrition ......... 145 This report was prepared by a team composed of: Christopher Chamberlirr (mission leader)* Fadia Saadah (public health); Stanley Scheyer (public health policy); Richard Heaver (nutrition); Shaikh Hossain (health financing); mario Taguiwalo (consultant, institutions); Joseph Kutzin (hospitals); Charles Griffin (private sector); Larry Schroeder (consultant, comparative context); Benjamin Diokno (consultant, fiscal issues); James Herm (consultant, organization and management); Angela Murphy (formating and word processing). The coc eration and contributions of the Philippines Department of Health, the Department of Budget and Management, and local government governors, mayors and health staff are gratefully acknowledged.  List of Tables Table 1.1 -Infant Mortality Rates in Selected Asian Countries, 1970-90 .................................................. 5 Table 1.2 -Mortality Rates (per 100,000 population) for the Ten Leading Causes of Death, Philippines, 1985-9 .............8 Table 1.3 -Basic Nutrition Indicators for Selected Asian Countries (1987) ............... 9 Table 1.4 -Fertility Declines in Thailand, Indonesia, and the Philippines ... Table 1.5 -Total Health Care Expenditure (Mil Peso) .12 Table 1.6 -Department of Health Expenditure (Mil Peso) .32 Table 1.7 -Patterns of Public Health Expenditure in Selected Asia Countries (Average 1980-89) ............................. 32 Table 1.8 -The Department of Health Ex.penditures (Mil Peso, 1980 Price) .34 Table 1.9 -Uses of DOH Expenditures (Mil Peso, 1980 Prices) .35 Table 1.10 -Percent Share of The Sources of Financing Public Health Expenditure .36 Table 1.11 -Medicare Claims, Reimbursements, and Support Values 1987 .37 Table 2.1 -Cost of Devolved Functions .43 Table- 2.2 -Devolved DOH Assets and Personnel by Function and Destination .......... 44 Table 2.3 -organizational Types of Health Service Decentralization ................ 49 Table 2.4 -Possible Effects on Efficiency and Equity of Health Services ...... 61 Table 3.1 -Expenditure Program, Gross Net of Debt Purden, 1991-93 (in billion Pesos) .65 Table 3.2 -Ratio of LGU's with Fiscal Risk. 71. Table 3.3 -Internal Revenue Allotment, 1990-193 (in billion Pesos). 77 Table 4.1 -Role Assignment Matrix .93 Table 4.2 -Potential "Trade" Between Levels of Government .105 Table 5.1 -The Potential Benefits and Costs of Health Service Devolution in the Philippines .................. 116 Table 5.2 -Policy Objectives and Interventions for Devolution Management .120 Table 5.3 -Macro Options for DON Organizational Future .128  Figure 1.1 -Infant Mortality Estimates, Philippines, Selected Data Sources ................................................. 4 Figure 1.2 -Infant Mortality Rates and Per Capita GNP in 14 Asian Countries, 1989 ............... 6 Figure 1.3 -Percentage Malnourished Children under 6, Philippines, 1978-90 ................................................. 10 Figure 1.4 -Macro Health System Structure .13 Figure 1.5 -Population per Private and Public Hospital Bed with Provinces grouped by Household Income in 1988, for 197?, 1983 and 1S90 .17 Figure 1.6 -Populacion per Facility Ranked by Income, 1990 .18 Figure 1.7 -Household Utilization of Health Facilities by Income Decile, 1987 .19 Figure 1.8 -The Growth of Hospital Beds, 1972-1990 .21 Figure 1.9 -Immunization Coverage for All Antigens, Routine Report, 1980-1990, Philippines, DOH ............................. 24 Figure 1.10 -Degree of Prevalence of Schistosomiasis in Terms of Low, Medium, and High, 1986-91, Philippines, DOH ............. 28 Figure 1.11 -Expenditure on M:edical Care in Rural Areas, 1988 Expenditure Survey .................... ............ 37 Figure 3.1 -Municipal Fiscal Risk Against Real Property Assessment. 74 Figure 3.2 -Provincial Fiscal Risk Against Household Income .75 Figure 3.3 -City Fiscal Risk Against RPA .76 Figure 3.4 -Provincial Fiscal Risk Against Devolved Health Expenditure .79 Figure 3.5 -Municipal Fiscal Risk Against Devolved Health Expenditure .80 Figure 4.1 -Macro Health Systems Structure - Current and Future .92. Figure 4.2 -Existing Activities, Channels and Timings for the Distribution of National TB Program Logistics ........... 97 Figure 4.3 -Trade Potentials & Patterns of Resources and Functions .... 107 Figure 5.1 -Provincial Hospital Operational Efficiency Indicators, Philippines, 1990 ...................................... 137 A. Provincial Tertiary Hospitals - By Provincial Income B. Provincial Primary and Secondary Hospitals - By Provincial Income C. Provincial Tertiary Hospitals - By Region D. Provincial Primary and Secondary Hospitals - By Region List of Boxes Box 4.1 -Health Program Checklists for Devolution Management ....... 101  EXECUTIVE SUMMARY Tntroduction 1. Secretary Juan Flavier of the Philippines Department of Health (DOH) asked the World Bank to analyze the likely impact of the 1991 Local Government Code on the delivery of public health services. The DOH was concerned that the rapid and far-reaching decentralization mandated by the Code would impose a wide array of problems on the health system and result in sharply diminished performance. The Code has now been imnplemented, and local and central health officials are adapting to the new system. The major findings of this report are twofold: first, the balance between the risks and opportunities of the devolution of health services is a "close call"; second, several of the risks could seriously affect the performance of the national disease control programs unless they are immediately addressed by central and local goverrments. A. Background 2. The Philippines is currently implementing the 1991 Local Government Code (LGC). As a result, local governments will receive additional grants frou the national government and assume responsibility for delivering and financing selected services, including public health services. The resources are provided in the form of central internal revenue allotments that local governments are free to allocate among and within various sectors. 3. The public health system, perhaps the most affected sector within the national government, is undergoing a far-reaching structural and functional transformation. All the DOH facilities at the local level and over half of its staff (45,700) have been devol-ed to local governments at the barangay, municipal, provincial and city levels. Public health services and hospital operations at these levels are no longer subject to central financial or managerial control. In terms of the scope of health services, personnel and facilities involved, the number of local governments participating, and the high degree to which authority is being decentralized, the experience stands out as one of the most ambitious health decentralization initiatives ever undertaken in Asia. 4. There has been considerable debate in the Philippines concerning this initiative. Health services under the central Government, while improving in recent years, have not adequately addressed certain public health problems. Malaria is still endemic in some areas. Infant and childhood diseases are still among the leading causes of death. Acute respiratory infections and diarrhea, two leading killers, have not been controlled. Family planning and nutrition services are undeveloped. Thus, a substantial agenda is still unmet and must be tackled under a transformed public health system, now under the authority of 1,600 local governments.  ii 5. Some of the key public health programs still require central management and direction to assure national coverage and consistent technical approaches, a task made much more difficult by the devolution of health services. rhe incentives for local support of public health are incomplete because some benefits from their programs "spill over" to other jurisdictions. Also, localities may choose to support sophisticated clinical services and facilities and reduce their support for preventive programs which have only weak public demand. In addition, they may also direct services to their better-off constituents, leaving the poor underserved. 6. Yet, the same factors that threaten the public health system under devolution are also a source of optimism: local governments will be more accountable for the services they provide, and they will adapt services more closely to local needs and popular preferences. Further, inefficiencies arising from cumbersome central control will be %orrected more readily by autonomous local managers more conscious of cost; the public will be more willing to help pay for services if these are more responsive to local needs. 7. The risks and opportunities thus present the Government with urgent policy questions. Can the nLew decentralized system be made to work as well or better than its predecessor or will the initiative be overwhelmed with serious flaws? If it is worth pursuing, how should the potential problams be addressed and minimized? Finally, which issues must be resolved in the short term, and which can be postponed? 8. To answer these questions, the report first analyzes the performance of the pre-devolution health system (pre-1993). It idertifies its strengths and weaknesses, and for the key public health programs, such as TB, malaria, immunization and other disease control programs, it assesses how they may respond to the new decentralized system that is being rapidly introduced. The report then analyzes the risks and opportunities in three areas: (a) First, local governments face new incentives in budgeting for and managing health services under the decentralization law (the Local Government Code of 1991 or LGC). Theoretical and comparative data are used to explore the likely responses of local governments to the new arrangements and the impact of these responses on the efficiency and equity of the new system. Many of the expected benefits of decentralization are founded on the incentive structure for local decision makers, but potential problems arise from this autonomy, especially from the point of view of national health objectives, such as meeting the health needs of the poor and assuring adequate financial support for national public health programs; (b) Second, there is the issue of the adequacy of financial resources devolved to local governments and whether the latter will allocate sufficient resources for health services. Many observers expect that resources will not be adequate to support health and other devolved services. However, using actual fiscal data, the report estimates the underlying fiscal balance of each local government for 1993 (the first year of devolution) and the results do not  iii support earlier ?essimistic views on resource availability; (c) Third, organizational and management issues will surely materialize from the LGC. This Code was passed so as to empower local governments with new responsibilities and resources, not to improve the performance of the health system Rer se. Thus, there are a number of organizational asymmetries, gaps and unintended incentives in the decentralized health system that may seriously disrupt the management of national public health programs in the Philippines. 9. The approach taken in this report does not seek to identify a model decentralized system or structure to which the Philippines should aspire. This is the result of two factors: first, the LGC has already determined major features of the structure, and the Code is not likely to be repealed; second, the literature on dec'-ontralization provides little guidance on the question. Opinions range widely o'a the efficacy of decentralized vs. centralized organizational arrangements. Further, the literature is notably sparse in rigorous quantitative assessments of the impact of different decentralized structures on health service performance. 10. Thus, rather than seek to identify a model decentralized structure, the report focuses the analysis on actors, incentives and structures in the new system, identifies risks and opportunities (potential costs and benefits), and adopts a step by step approach to irncrease the new system's net benefits. The analysis will therefore lead to some implicit model of a 'better' decentralized structure, but it will not attempt to define a "best" system. B. Performance of the Pre-Devolution National Health System 11. The health system is characterized by a rough parity between private and public sectors in terms of health expenditure, number of hospital beds and manpower. thus creating a truly mixed system. The centralized public system was composed of a vast network of hospitals, clinics and health stations over the entire country, while the private system of hospitals, outpatient clinics and private practitioners also expanded to attain national coverage. The two sets of health pro-icers have competed for Uients in various health markets and have benefited from public subsidies and incentives. 12. The performance of this national system was less than exemplary from the late 1970s to the mid 1980s, when a protracted stall in the decline of the infant mortality rate (IMR) at around 60 per 1,000 live births let the Philippines fall behind other Asian countries for this key indicator of public well being. In the late 1980s, the DOH, with substantial inflows of foreign assistance, launched an effort to upgrade the performance of key public health programs (expanded program of immunizations [EPI], malaria, TB, maternal and child health, etc) and attempted to target the benefits of these programs to areas at highest health risk and of lowest socio-economic attainment.  iv 13. Results bave been impressive, both in the higher coverage rates for public health programs, and potentially, in terms of tneir impact on the INR (survey results should be available soon to determine the 1992 IMR and the child mortality rate). The EPI, malaria control, and schistosomiasis programs in particular have performed very well. Immunization rates for the six basic childhood diseases rose sharply in the second half of the 1980s with all except tetanus toxoid reaching over 70 percent of the target population, a result confirmed by household surveys. The annual rate of parasitic incidence or API for malaria declined by half between 1987-1991. EPI not only expanded coverage, but did so in an equitable fashion (with support of local governments), and coverage rates for poor and better off regions converged at higher levels. 14. However, the DOH priority disease control programs have not been equally successful, with some just starting to expand and others such as family planning/women's health and nutrition undeveloped, with little impact in the field; and, their level of development partly determines how they may fare under che new, decentralized structure. As meintioned above, EPI has done well, and local commitment, support and demand for child immunization services have followed program development. Thus, EPI is among the DOH programs least likely to suffer management disruptions or diminished funding from local governments. 15. The malaria program, on the other hand, has a long history of fluctuating effectiveness, probably due to lapses in coordination and uxnsustained effort to control the vector, to identify active cases, and to administer treatment. Although the program will be partly retained by the DOH, field operations depend on the active participation of local health staff. As a classic "public good" program in which much of its activities benefit society in general, local support for it may vary widely, which in turn, would affect its national scale and reduce its technical standards. 16. Programs to control acute respiratory infections, nutrition, and family planning have less to lose from decentralization, but there is an urgent need for these programs to be more effective, nationwide. Thus, expanding such undeveloped natiou..Al programs under a highly decentralized system will present the DOH with a difficult challenge. Not only will many local governments need to be persuaded to implement them in a consistent manner, but this will have to be done in a way that other programs do not suffer reductions in local support and effort. 17. Another important feature of the pre-devolution public health system is the pattern of public finance. Public funds have had a clear bias in favor of hospital construction and recurrent costs, despite recent efforts to increase spending on basic and preventive services. Thus, the pre- devolution system will pass on a large hospital infrastructure to provincial governments, the costs of which may prove as burdensome to them as to the DOH. However, the regioral distribution of DOH expenditure on facilities and services has been in most caszs modestly progressive, meaning that public health spending has benefited the poor disproportionately (a major achievement for a public system). The private system, while impaired by a proliferation of inefficient small hospitals, is nonetheless well distributed throughout the  V country and apparently is not being forced out of local markets, at least not yet, by the competition posed by public facilities and services. 18. Similarly, the overall health system has produced more than adequate manpower to meet the demand of the DOH and the private sector. In the late 1980s, the average graduating class of physicians was equal to 17 percent of the entire stock of practicing physicians in the country. Another strength of the former system was the successful DOH effort to devolve administrative authority to provincial health offices. Although well short-of the fully devolved structure now in place, this effort nonetheless built a substantial local capacity to manage health facilities and services at the provincial level. 19. As a prelude to devolution, the balance between strengths and weaknesses would seem to indicate that the health system achieved a wide- ranging rebound in performance compared to the mid 1980s, and that local governments have inherited a system more characterized by improving performance than systemic inefficiencies or inequities. However, decentralization could disrupt a number of public health disease control programs due to their vulnerability to divided responsibilities, loss of scale and coordination, and weak local demand. C. Risks and Opportunities in the Devolved System 20. Turning to the three areas of concern under the devolved system, this section assesses the net impact of devolution on the efficiency and equity of the public health system. First, the framework of analysis is described. Next, the new incentive structure for local government decision makers is assessed in terms of its likely impact on efficiency and equity of local health services. This is followed by sections on fiscal issues and organizational and management issues. Analytical Framework 21. Attempting to estimate the performance of a newly decentralized health system first roquires an analytical framework. The Secretary of Health articulated the importance of achieving increased efficiency and equity under the new decentralized system. The issues raised about devolution of health services, as noted above, deal with the adequacy of financial resources, mismanagement by local government officials, potential increases in costs, and the impact on the poor. The framework for the report should respond to these concerns and issues. 22. The framework chosen for this report is based on a cost-benefit approach. In moving from a centralized to a decentralized system, what are the likely costs and benefits (risks and opportunities) to the performance of the system and what will their impact be on the system's objectives? 23. To assess likely costs and benefits, three questions are addressed. Will resources to finance the system increase or decrease? Will these be put to the most effective use? And, will the production costs to deliver health  vi services rise or fall? The answers to those questions will affect the two critical objectives for the performance of the system: first, that the system operate efficiently, so that outputs (health services) are obtained for the least cost and that the level of output responds to both local and national priorities for health services; and second, that the new system deliver its benefits equitably. Local Incentive Structure 24. A major source of change is the new set of rules for local governments, as created by the LGC. These rules for raising, budgeting and managing resources establish new incentives for local governments; local officials can be exrected to respond to these incentives as they budget for and manage the health services devolved to them. Using theoretical perspectives and evidence from other countries, what will be the likely impact of the new incentive structure on the equity and efficiency of the new health system? 25. Efficienay. From the point of view of production or technical efficiency, decentralization on theoretical grounds could be expected to yield substantial gains in minimizing costs through more accountable local decision making, greater cost consciousness and innovation in how public services are produced, and more attention to cost recovery. Evidence from various countries is replete with examples of innovative, cost-reducing approaches to decentralized health delivery. For example, local governments in Nigeria turned to leasing as opposed to purchasing vehicles to deliver immunization services at considerable savings. 26. As for the overall economic efficiency of a decentralized system, the evidence suggests that localities will support health services at a level more in line with local requirements and may also integrate health services more efficiently with related services in sanitation, nutrition and education. Local governments can be expected to employ cost recovery more readily, which would help match health services more closely to local demand. Fiscal autonomy generally can be expected to increase the willingness of individuals to contribute towards public services, and decentralization will thereby lead to greater overall revenue effort. 27. But from a national point of view, local decisions may not be optimal. Risks are attached to over-spending on capital projects, such as duplicate hospital services or specialized clinics, and under-funding operations and maintenance. Such problems are not unique to decentralized arrangements, but they may be exacerbated by them. Poor administration of personnel, particularly political interference in personnel matters, can increase the costs of producing health services, and this has been found to occur more readily in locally controlled health delivery systems. Finally, production costs for national disease control programs may increase under decentralization due to inconsistent local participation and consequent loss of economies of scale, as could be the case in malaria or schistosomiasis control. Underspending on "spillover' services such as communicable disease  vii control may reduce economic efficiency, as would over-investment in purely curative services. 28. Based on the comparative and theoretical evidence, the efficiency effects of decentralization in the Philippines should be positive, but much of this potential benefit will need to be facilitated through the training of health managers and information campaigns to inform the people of their opportunities to shape local health services to their needs. Such health education efforts, as performed by the center, will be a crucial input to a successful decentralized system. 29. Equity. The equity impact of decentralization is a source of concern. On theoretical and comparative grounds. it is apparent that local government leaders are subject to strong incentives to direct services to the better off groups in their constituencies. Although there are exceptions, it is the potentially negative effect on low-income households or poor regions that is most commonly given as an argument against decentralized health service provision arrangements. Such disparities increased in Mexico and Papua New Guinea. A frequent outcome is that poorer local governments tend to allocate smaller proportions of resources to health services and are relatively less capable of processing requests for special assistance. In the Philippines, the central DOH had developed a moderately egalitarian distribution of facilities and services, an achievement unlikely to be improved by the local provisioning process set up by the LGC. 30. The comparative perspective, therefore, indicates a potential trade-off between efficiency gains and equity losses, if corrective interventions from the center are not taken. A second potential trade-off could occur between national and local health objectives, so that local needs for clinical and curative care are more accurately met, but key public interventions in communicable disease control and prevention are underfinanced at the local level. Local Government Fiscal Issues 31. The debate in the Philippines over health service devolution has often focused on the adequacy of local financial resources to support public health services and the potential misallocation of such resources. Concerning the adequacy of resources, it has been argued that the LGC would not provide sufficient additional revenues to cover the costs of supporting the devolved services, including healthi, and that the poorest local governments would be especially hard hit in fiscal terms. The second strand of the debate has addressed the allocative behavior of local governments and their allegedly low interest in adequately budgeting for health services. 32. Adeguacy of Local Resources. Expected total revenues and expenditures for each local government for 1993 were analyzed, including the central revenue grants (IRA) and local tax income on the revenue side, and the local expenditures on all services plus the annual recurrent costs of devolved functions on the expenditure side.  viii 33. The results show the following: 13 of 73 provinces (17.8 percent) will have insufficient revenue to cover total expenditures; 109 of 1,533 municipalities (7.1 percent) will have insufficient revenues (based on a sample of 533 municipalities); and none of the 60 cities will be so burdened. The remaining localities will all be in surplus, with revenues exceeding expected expenditures. These results indicate that the overall, first year, fiscal impact on local governments is favorable with small proportions facing a deficit as a result of the LGC. The total fiscal deficit faced by the 13 provinces is only P270 million or about three per cent of 1993 provincial IRA. Municipalities were burdened with about P420 million (also three per cent of 1993 municipal IRA). 34. This relatively favorable outcome on the fiscal effect of the LGC could be improved further with more local efforts to increase tax revenues. This would reduce, but not eliminate, the proportion of localities in deficit. This suggests that local governments may wish to intervene on the cost side as well as to balance budgets. It also suggests that the center, if it chooses to address this issue given its relatively modest dimensions, may wish to target short-term augmentation resources to only those poorer localities with large deficits and more urgent health problems. 35. It should be emphasized that these are estimates for 1993 using various assumptions. The resulting fiscal surplus or deficit shows an "underlying" fiscal impact, and does not take into account the budgeting decisions of local governments over the course of 1992 and 1993, as additional revenues and devolved services were transferred to them. These responses, as would be expected, have quickly found expenditures to consume the surpluses, while deficit localities have been forced to shrink expenditures to attain balanced local budgets, a legal requirement. Thus, to compute deficits and surpluses, the approach in this report assumes that local governments will finance all devolved functions fully, and that the central revenue grants would be fully distributed to local governments in 1993 for the first time. 36. Another such estimate of local fiscal gaps has been calculated by the Department of Budget and Management, using a different methodology and assumptions. Those results indicate a higher number of provinces in "deficit" (see Chapter III). 37. Two other findings are notable from this fiscal analysis. First, the underlying fiscal "surplus" for all cities totals nearly P4 billion, a substantial transfer from the center, and a source of concern given its budget constraints. Provinces and municipalities, as groups, also enjoy substantial net surpluses. The implications for health spending need to be kept in mind; if such localities allocate surpluses according to the share of health in the budget, the additional health spending above pre-devolution levels would be substantial. 38. The second finding is that the distribution of fiscal "deficit" across local governments is not concentrated in the poorest governments. In fact, for municipalities, the fiscal burden increases slightly with the tax  ix base; for provinces, fiscal deficits fall evenly on wealthy and poor governments. The fiscal balance also seems to follow the size of the devolved health costs, meaning that localities, especially provinces, with higher recurrent cost burdens of devolved health staff and facilities will tend to have lower surpluses/ higher deficits. The devolved health expenditures, therefore, will play a significant role in determining local fiscal status, and local governments in deficit will need to examine health expenditures carefully. In that regard, hospital recurrent costs will be of direct concern to provinces. There will be ample room for improving the operational and financial efficiency of devolved provincial hospitals, an area of considerable local interest and an opportunity for the DOH to lend technical assistance (see annex SB for a discussion of hospital issues). 39. Allocation of Resources. How local governments will allocate revenues toward their expenditure responsibilities, specifically in regard to health services, is difficult to predict. Analysis of 1991 local health expenditure can only provide some clues. The 1991 data show that health spending per capita is positively related to the wealth of the local government, meaning that local funding for health services in the pre- devolution period followed local government revenue. This is consistent with the comparative perspectives reviewed above, that equity objectives may not be well served by decentralization. However, as noted above, the large surpluses accruing to local governments, including poor local governments, afford an opportunity to raise local health spending over pre-devolution levels. 40. In sum, the fiscal dimension of the LGC, once thought to be a major cost/risk issue, is more likely to be a minor problem, with most local governments, including the poorer governments, enjoying substantial discretionary resources above what they will need to support devolved functions and their own programs. Thus, resource availability would appear to be adequate or even generous, and distributed fairly; however, the way in which local governments will spend these resources is a source of concern. Past behavior indicates that the poorer local governments will be less inclined to support health services than their wealthier counterparts. Consequently, spending disparities may widen between rich and poor localities despite "surplus" revenues. Organizational and Management Issues 41. The LGC contains specific provisions for health service management and therefore imposes arrangements that will have impacts on efficiency and, to a lesser extent, on equity. Thus, the risks and opportunities to the health system can also be approached from an organizational and management perspective. 42. On the plus side, the Code allows the DOH to retain a number of key functions that are best carried out at the central level, such as health information and education, foreign-assisted projects, licensing and accreditation. This array of retained functions presents the DOH with substantial opportunities to influence local health policy and implementation,  x without compromising the autonomy of local decision making, from which many of the benefits of a decentralized system would be expected to flow. 43. On the risk side, the LGC imposes change of control on nearly every management function and public health program under the DOH. Understandably, there will be a number of organizational asymmetries, gaps and unintended incentives which, taken together, pose a high risk of substantial costs to the national public health system. These issues are the most urgent of any single group of devolution-imposed risks and will need to be addressed soon by the national and local governments. 44. The organizational issues arising from the Code can be summarized as follows: (a) Potential loss of technical integrity. The LGC creates separate, autonomous health departments in 75 provinces and 1,533 municipalities, subject to loosely defined national standards of public health services. There is a risk that individual localities may politicize personnel selection and underfinance technical support so that technical quality suffers. Such issues as supervision, training, reliable medical supplies and equipment, and links with higher-level facilities are subject to local policy and priorities. Local health boards are mandated by the LGC to promote technical standards, but the health board role, essentially advisory in nature, lacks the authority to set health policy or enforce technical standards. Loss of technical integrity will lower the output of services, but without lowering costs. (b) Local capacity/authority asymmetries. The LGC effectively stratifies local health service delivery by assigning basic outpatient services and disease control services to municipalities, while assigning hospital services to provinces. This stratification does not correspond to previously developed management capacities, effectively burdening municipalities with responsibilities for which they are poorly prepared, and removing public health service responsibilities from provinces where capacities had been relatively well developed. This asymmetry in authority and capacity is particularly worrisome for the delivery of public health programs at the municipal level. Costs may rise at the municipal level, but yield no improvement in service. (c) Loss of provincial-level health service management. This asymmetry is further aggravated by the absence of provincial coordination and overall management of the basic public health programs. The LGC excuses provincial governments from such responsibilities and faced with their own budgetary pressures, provinces face an incentive to cut back on their support for public health program management. Without a provincial intermediary to funnel centrally provided program inputs to key public health programs, the performance of such flagship disease control programs as EPI, malaria control, TB  xi control and other maternal and child health programs can be expected to decline in the near term. (d) Possible breakdown of health management systems. Another issue arising from the above problems of stratification and loss of provincial level mane-gement is the expected deterioration in two key health management systems: the information system (and with it national health planning) and the logistics and procurement system. If the local delivery system is allowed to fragment into over 1,600 separate, uncoordinated health systems, then procurement and logistical arrangements to deliver central medical and other program inputs to the local service delivery points will be placed under considerable stress. In addition, the lack of incentives for local governments to send management information up to the DOH indicates a possible rapid breakdown of a nationally integrated health service management information system. This would seriously complicate development of provincial and national level health policy and planning functions as well as efforts to target central resources to areas of local need. Again there is the potential for program performance problems and cost increases as local governments attempt to create such management capacities for themselves. (e) Absence of a center/local assistance mechanism. Finally, the LGC makes no detailed provision for a new mechanism through which the DOH can direct assistance. It no longer has direct control over local service budgets and their targeting, a convenient central mechanism through which foreign assistance once flowed to local service delivery points. This affects both current projects and any future ones supported by donor assistance. Given the equity and efficiency rationale for selective DOH intervention at the local level, the development of a new assistance mechanism allowing the DOH to target project funds is receiving close attention. Net Assessment of Health Services Devolution 45. Reviewing these likely risks and opportunities, it is difficult to assert that the balance is either strongly negative or positive; rather, the net impact on efficiency and equity of the health system seems to be a "close call". There is a clear potential for some net efficiency gains and a roughly parallel potential for some net equity losses. A decisive benefit relates to the relatively positive effects on resource availability to local governments; if the analysis had demonstrated that a majority of localities would be unable to cover estimated expenditures as a result of the LGC, the net assessment might well have shifted into the high-cost column. D. Policy and Management ResRonses to the LGC 46. Because the LGC imposes potential costs and benefits so widely throughout the health system, the close call outcome creates a demanding  xii management challenge for the DOH and local governments. There is no basis for complacency; indeed, the provincial-municipal management problems contain large downside risks. Moreover, it will not be possible to target one or two single areas of high risk, but instead managerial attention will be spread across several categories of issues with high impact on efficiency and equity objectives. The DOH and its local government partners will need to agree on a policy framework and a corresponding set of near-term and longer-term priorities for action and implementation. Policy Actions - Near Term 47. From DOH policy statements and the analysis in this report, several policy objectives for devolution management can be defined for the near term, that is, over the transition period during which the new proprietors of the health system will seek to stabilize the management and performance of the system. These include: (a) Sustaining equitable and efficient performance of the leading disease control programs; (b) Reconstituting provincial public health program management; (c) Building local management capacity; (d) Strengthening local empowerment and self help. To attain those objectives, the management interventions listed below have emerged as high near-term priorities. 48. Negotiation of Provincial Intermediary Role. To address the stratification of the new system and the loss of provincial management of basic services, one approach would be to trade DOH resources for a restored provincial public health system. The restoration of provincial responsibility for the key public health programs (EPI, TB, malaria, schistosomiasis, maternal and child health, etc.) would be the objective of such 'trades.' 49. Restoring the provincial intermediary role is based on the clear "trading" opportunities between levels of government, each of which have specific needs for public health commodities and support services that can be supplied from within the national system, but do not have assured access as a result of the LGC. It is, therefore, possible that the central as well as provincial and municipal governments will have an active interest in negotiating such trades so that they can sustain public health services without having to establish, at considerable cost, individualized procurement and service support arrangements. 50. Thus, the DOH might trade medicines, vaccines and training services to municipalities, in return for a regular flow of management information and municipal agreement to participate in provincial planning and management of  xiii public health services. Provir.es might trade their resumption of managerial services over disease control programs in return for access to DOH financial and project assistance and hospital management assistance. These arrangements need not involve additional resources from the DOH; existing foreign-assisted projects and regular DOH budget support would cover the flow of resources to local governments. 51. Such trades or agreements between the DOH and provincial governments (and component municipalities) are the alternative to the many ad hoc arrangements that would likely emerge in patchwork fashion, locality by locality, program by program, system by system, and donor by donor. The LGC encourages such a patchwork approach because it provides for such a high degree of local autonomy over revenues, program choices and their management, 52. The DOH has proposed the negotiation of a "Comprehensive Health Services Agreement" with each province and its component municipalities. A Health Development Fund would account for the annual flow of DOH resources to the province and its municipalities. The Agreement would set out the local and central responsibilities for public health program management and delivery, with the province as the primary partner of the DOH. The Fund would make available the flow of program resources to local governments in return for signing the renewable agreements. 53. Municipal Health Management Canacity. The LGC now imposes responsibility for personnel management, logistics, training, budget preparation, overall planning, and expenditure management under the authority of municipal health staff who have limited experience in such functions. Provincial staff performed many of these functions in the centralized system. Thus, upgrading health management capacity and obtaining interim management support will be an urgent priority for many municipalities. This problem applies especially to larger municipalities with multiple facilities and large staff complements, and to municipalities with a large complement of new or inexperienced staff. Municipal facilities are the delivery points for most public health services, and serious management failures at this level will have a large impact on program performance. 54. The severity of this potential problem will be contained in the short term by the persistence of past linkages with provincial program coordinators and supervisors. These habits, however, will steadily diminish unless renewed, and relatively new management skills will need to be introduced so that the municipal health offices are able to present their budgetary needs convincingly and demonstrate efficient management of resources to local government executives and legislators. A provincial role in providing management support to municipalities would be a logical component of a restored "provincial public health system." 55. Health Education. Preparation of a national campaign to inform the public of its new responsibilities in the decentralized public health system would be a high impact intervention for the DOH in the near term. Vocal, persistent, popular demand for quality health services from local governments is a vital step toward local empowerment and the best guarantor that such services will be delivered.  xiv 56. Center-Local Assistance Mechanism. 1stablishing a new project assistance mechanism will be a fundamental prerequisite to building the DOH partnership role with local governments. It will require many steps and sub tasks, covering such issues as funds flow, definition of targeting criteria and cost-sharing formulas, and project development procedures and capacities. A workable, functioning mechanism would be a substantial contribution to the credibility of DOH offers of assistance to develop new public health services and help address equity concerns, while also attracting foreign donor finance. Local governments will participate in the Comprehensive Health Service Agreements partly based on expectations of future DOH assistance flows. 57. The most crucial elements in the design of such a mechanism are the choice of instrument (revenue grant vs. project assistance) and the targeting and cost-sharing criteria. The risks involved in poor design of the mechanism are substantial. For example, formula-driven revenue grants, even if earmarked for health spending, may have the effect of merely substituting for local resources in the financing of health services, so that no additional resources are made available to health. Another risk is that such assistance may depress local tax effort, thus perpetuating dependance on central resources. 58. If more discretionary instruments are used, such as project grants, a number of advantages ensue. The purposes of the grant can be controlled, the benefits can be well targeted to beneficiaries, and cost-sharing arrangements can be speci.fically tailored to local fiscal conditions. Moreover, donors will be more comfortable with such an arrangement. Even if clearly preferred as an assistance mechanism, the project grant (or loan) mechanism must be administered in a transparent fashion to assure accurate and equitable targeting and cost-sharing conditions, and poorer localities with low administrative capacity must be assisted in preparing project proposals. 59. Improving Provincial HosDital Efficiency. Although less urgent than other issues from the national point of view, the need to improve hospital efficiency, both operational and financial, will be a high priority for local governments and an important area for potential benefits from decentralization. Hospital recurrent expenditures will absorb a large part of provincial budgets and are an important factor in explaining provincial fiscal deficits. Therefore, governors may seek to raise efficiency and increase hospital revenues to ease overall budgetary pressures. 60. Analysis of operational efficiency data for all provincial hospitals reveals several interesting facts. First, financial efficiency (the proportion of revenues over total operating costs) is very low, averaging about 5 percent. This could be increased, especially in wealthier catchment areas. Second, operational efficiency occurs over a wide range, indicating that many hospitals fall ir:to the "less efficient" category and could benefit from quality improvements, consolidation, or even privatization. Some hospitals are overutilized and may require better demand management and capacity expansion. Moreover, both rich and poor provinces have similar efficiency problems.  xv 61. The DOR can play an important role in assisting provincial health offices to undertake effective hospital management improvements that will benefit both efficiency and equity objectives. Such assistance will also be a tradable service in the negotiation of a provincial role in managing public health disease control programs. Policy Actions - Longer Term 62. A number of issues are more appropriate to action in the longer term, not because they are any less important, but their impact on health program performance is slower. These issues are listed below. 63. Reform of the Management Information System. For the near term, the DOH and local governments may agree to continue the Field Health Service Information System as it was operating prior to devolution. However, that system will need to be reformed to respond more directly to local government information needs and be oriented to DOH assistance flows to local governments. 64. Health Plarming. Similar recommendations apply to the area-based health planning system introduced by the DOH in 1989. Its use should be sustained until the provincial intermediary role is well established and a more locally adapted planning system can be designed and installed. 65. Personnel Management Reform. Establishing an integrated personnel management system allowing for inter-governmental transfers, promotions and secondments would be a positive development for morale and efficiency. In addition, salary scales may need to be adjusted to allow local governments to recruit for service in less desirable locations. 66. Assessment of Procurement and Logistics System. The procurement system for the public health system has already beer. decentralized to a large extent and may be better adapted to a decentralized framework than other management systems. However, the newly devolved arrangements will create opportunities for efficiency gains, such as local pooling of procurement needs and concentration of procurement authority. An assessment of the procurement and logistics system should be undertaken after the devolved system has operated one or two years. 67. Modification of the IRA Formula. Due to the large fiscal surpluses accruing to cities and for other reasons, members of the national legislature have introduced bills to revise the revenue sharing (IRA) formula, as set down in the Local Government Code. Although the IRA formula appears to favor some local governments over others, the analysis in this report indicates that the formula is relatively benign in its fiscal impact on both poor and better off localities. Before a revision is legislated, several risks will need to be taken into account, such as the impact on local budget decision making, the distributional effects of a new formula and the relative expenditure needs of cities as opposed to provincial and municipal local governments. The surpluses generated by the current IRA formula for cities might prove to be less of a windfall than is apparent from the "surplus" estimations. Moreover, other policy interventions might be more appropriate to address the problem,  xvi such as further devolution of expenditure responsibilities to local governments. Additional analysis of this complex question could help determine whether an intervention is needed, and if so, how it should be designed to optimize the impact in both efficiency and equity terms. 68. Pooling of Resources and Other Efficiency Gains. Local governments will encounter opportunities to collaborate in the production of health services so as to avoid duplicated and unnecessary expenditures. However, the local response to these opportunities may benefit from encouragement and information on best practices from the DOH so that local governments more fully understand the benefits of such efficiency improvements. E. The Future Role of the DOH 69. This report assumes that the DOH will seek to establish itself as a full partner with local governments in the delivery of equitable and affordable health services for the entire public. This role has been eloquently articulated by the Secretary of Health, Dr. Juan Flavier. In particular, the DOH role would include responsibility for assuring that national health programs addressing communicable and preventable diseases are implemented consistently and equitably throughout the country. 70. Such a partnership role is not ordained by the Code, nor predicted by the decentralization experience of other developing countries. Central ministries of health in many cases have withdrawn into a regulatory and information-processing role, with minimal accountability for or involvement in the performance of the local health delivery system. This might be appropriate for more developed economies where differentials in health status and service access have been minimized, where the major communicable diseases have been controlled or eliminated, and local capacities to finance and manage public health services are advanced. But for the Philippines, inequities and service shortfalls are widespread and may well intensify under the devolved arrangements for public health services. Local capacities to deliver services are highly variable, and public health problems still require concerted, coordinated control efforts. 71. The newly decentralized system opens a vast potential for local variation in the level and purposes of health spending. In many cases, local decisions on health policy and financing may not be optimal from a national public health point of view. To pursue a full partnership role, the DOH will need to develop a broad range of instruments and capacities, including project assistance and incentives, regulatory pressures, public information campaigns and technical assistance. These can be deployed to influence governments, help develop their management capacity, and push their allocation of resources towards the most urgent public health problems affecting the most vulnerable, least able to pay groups. 72. How the Department will build the necessary capacities, obtain the necessary cooperation from LGUs, and secure sufficient external financial resources to fulfill its partnership role are its fundamental challenges for the rest of the decade.  Chapter I. PERFORMANCE AND STRUCTURE OF THE EHILIPPINES HEALTH SYSTEM PRE-DEVOLUTION A. Introduction 1.1 The Philippines is currently implementing the provisions of the 1991 Local Government Code (LGC). As a result of the LGC, local governments will receive additional grant resources from the national government and will assume responsibility for delivery and financing of selected public services, including public health services. The public health system, as arguably the most affected sector within the national government, will undergo a far- reaching structural and functional transformation. All the DOH facilities at the local level and over half of its staff will be devolved to local governments at the barangay, municipal, provincial and city levels. Public health services and hospital operations at these levels of local government will no longer be subject to central financial or managerial control. 1.2 Decentralization of health services will incur costs and benefits. Organizational change of this magnitude can be expected, at a minimum, to impose short-term declines in service output during the transfer of infrastructure and staff to local governments. After that "changeover" phase, the expected longer-term benefits of a decentralized system can begin to emerge, such as services better matched to local preferences and more innovative and efficient production arrangements. On the cost side, decentralization may result in a less effective targeting of services to the poor and the vulnerable. Thus the potential for trade-offs exists, between the attainment of higher efficiency at the expense of equity. 1.3 Many of the costs and benefits of decentralization arise from the specific provisions of the law itself. Unintended impacts on resource availability or on management functions may work through the system causing substantial additional losses or gains in efficiency and equity. Thus, the Local Government Code must be carefully assessed to determine the potential impact of the incentives it creates and the structure it imposes. The major questions for this study, therefore, can be framed as follows: (a) What are the potential costs and benefits of moving from a centralized to a highly devolved health system in the Philippines and what are the expected impacts of those costs and benefits in terms of equity and efficiency objectives? (b) Which cost and benefit issues merit policy attention in the near term to assure continuity of program performance and positive health impact? (c) Which cost and benefit issues could be deferred to the longer term to reform a maturing, devolved public health system? 1.4 Two assumptions underpin these questions. First, it is assumed that devolution as directed by the Local Government Code of 1991 (LGC)  2 is a political fact of life that cannot be substantially altered in the near term. The genesis of the LGC was fundamentally political, and the political forces in support of the Local Government Code of 1991 will continue to hold sway over those opposing it. 1.5 The second assumption is that the central Department of Health will seek to establish itself as a full partner with local governments in the delivery of equitable and affordable health services for all the people of the Philippines. If devolution is seen to impose high potential costs in specific areas, then it is assumed the DOH will seek to compensate or correct for those costs with the tools and resources at its command. 1.6 Because the devolution of services and infrastructure to local governments is currently in process, it is not yet possible to measure performance of tte new system and quantify the cost and benefit impacts of the central compared to the devolved system. That analysis will need to be undertaken a few years hence. With the objective of providing timely policy input, this report, instead, uses a combination of past, current, and comparative evidence combined with theoretical perspectives to construct a likely array of cost ar.d benefit impacts, from which policy implications are drawn. The evidence for cost/benefit findings can be grouped as follows, with each corresponding to a chapter in the report. (a) Pre-devolution performance of the national health system. (b) Comparative and theoretical perspectives on decentralization. (c) Local government fiscal issues. (d) Organizational and management issues arising from the LGC. 1.7 Chapter I begins the assessment of costs and benefits, focusing on the performance and structure of the former centralized system of public health and its interaction with the private sector. The strengths and weaknesses of the former centralized system will to varying degrees carry over to the r local government proprietors of the devolved structure, creating potential costs and benefits. Thus, Chapter I first examines health status in the Philippines, followed by the structure and evolution of the public and private system, the performance of public health programs, health expenditure and finance issues, and a final section summarizing the cost and benefit implications of the former system for the new devolved system of public health services. B. Health Status 1.8 This section of the chapter reviews the evidence on health status and its distribution regionally and by different population groups. The main sources of information on mortality, morbidity, and nutrition are the DOH reports, vital registration, census records, and a series of specialized surveys. Surveys are the most reliable source of data on the health indicators and their differentials.  3 Infant Mortality: Levels. Trends and Differentials 1.9 At the national level, widely differing estimates from different data sources hinder a clear-cut conclusion on the levels and trend of infant mortality in the Philippines. For trends, infant mortality estimates from vital registration, for example, illustrate a declining trend reaching as low as 30.1 in 1988 (Figure 1.1). The National Demographic Surveys (1983 and 1988) and the Contraceptive Prevalence Survey (1986) all present a flat pattern of infant mortality between the late 1970s until the mid-1980s. For the earlier period from the mid-1960s until the mid-1970s, the direct estimates also show a flat pattern, as does the 1978 Fertility Survey (not shown in Figure 1.1). 1.10 Figure 1.1 also illustrates the magnitude of difference between the different sources and methods of estimation of infant mortality levels in the PhilippinesL'. Although the various surveys show a flat pattern of the IMR, the estimates differ significantly on thse overall infant mortality level. The indirect estimate from the National Demographic Survey of 1988, for example, gives an IMR of 65 infant deaths per 1,000 live births for the reference period of 1986; the direct estimate from the same survey has been reported as 52 deaths per 1000 live births. 1.11 In view of these conflicting results, a Task Force for examining the issues on infant mortality estimates (TF-IMR) was established in July, 1991 in the Philippines. In its final report, the TF-IMR recommended the adoption of rates derived from indirect (higher than the direct) estimates based on the demographic surveys. The TF-IMR chose to discard rates derived from the vital registration system as a result of high levels of under- reporting. In choosing the indirect estimates, the TF-IMR cited the robustness of the indirect techniques, but no explanation for the unusually large difference between the direct and indirect measures was provided. Consequently, there remains some uncertainty concerning actual levels of the IMR over time. .1/ Indirect measures are based on the Brass method for estimating infant and child mortality measures using data on number of children ever born and children surviving by age of mother or duration of her marriage. These methods are robust to data errors, especially in dating of events. However, they are not very accurate in providing measures of infant mortality in the immediate past prior to the survey date since those are based on measures derived from a selective age group of mother i.e. 15-19 years old.  4 Figure 1.1 Infant Mortality Estimates. PhilipRines. Selected Data sources 80 \ --Vital Regist. C Aensus- 1980 \s tN1S-1983-D 60 + M O- 1983 -I -Ea- N0S-1988-D 7 --O N -1988-1 t G ^ Ps- 1986-D 40 - C GPS-1986-1 20 ,,, ,,1 ,,,l, 1960 1966 1972 1978 1984 1990 Year Source: "Final Report of the Task Force on Infant Mortality Rate," November 1991 1.12 Trends on the other hand, of infant mortality are more consistent across the different data sources, especially those derived from surveys. Despite reservations concerning the true level of the IMR, these sources clearly reflect a stalling in infant mortality decline between the late 1970s and mid-1980s. The 1990 census provides a less reliable estimate for the IMR at the end of the decade, and it points to a slight decline in infant mortality during the latter part of the 1980s. If such a decline is true, it will coincide with a period of rising economic indicators and serious efforts by the DOH in terms of providing preventive and curative care, especially for children and mothers. 1.13 Regional differentials in IMR are also significant. Rates are lowest among NCR and Region III (NDS indirect estimates for these two regions are 46 and 49 respectively) and highest among Regions XII and VIII (NDS indirect estimates are 104 and 89 respectively). In terms of time trends, the regional experience is mixed, reflecting flat, increasing, as well as  5 decreasing patterns of infant mortality. Unfortunately, provincial level data are not available for a more detailed analysis. 1.14 The apparently flat trend in IMR estimates is an unusual occurrence for East Asia. Table 1.1 places the Philippines IMR performance in comparative perspective using data for Asian countries from 1970-1990. Among the selected countries, the Philippines had rates comparable to those of Thailand and China in the period 1970-75. Over the ensuing 20 years, rates dropped to 32 and 31 for China and Thailand respectively, a decline that was not paralleled in the Philippines. Indeed, using either direct or indirect estimates of the IMR, the Philippines shows the lowest percentage decline in the region. Table 1.1: Infant Mortality Rates in Selected Asian Countries. 1970-90 % decline Country 1970-75 1975-80 1980-85 1985-90 1970-90 Bangladesh 140 137 128 110 21% China 61 40 39 32 488 India 135 126 110 99 27% Indonesia 114 105 95 82 28% Korea, Rep. 47 35 30 20 57% Lao PDR 145 135 122 110 24% Malaysia 42 34 28 18 57% Nepal 153 147 139 127 26% Philippines'- 64 52(60) 51(61) 52(61) 19%(6%) Thailand 65 53 38 31 52% Viet Nam 90 62 53 46 49% Source: World Bank Data File 1.15 If infant mortality and per capita income are plotted for Asian countries, as in Figure 1.2, it is apparent that the current Philippines IMR, while above the regional mean for its income level and far short of the achievements of Thailand and China, is not in the outlier category. This would suggest a linkage between national economic progress and the IMR. a well known relationship in inter-country comparisons. Estimates for the Philippines are derived as follows: the 1970-75 estimate was provided by the World Bank, Population and Human Resources; the 1975-90 estimates are derived from the direct estimates of mortality using the NDS-1988 data (as reported by the Philippines Task force on Infant Mortality (TF-IMR). The values in parenthesis represent the suggested values by the TF- IMR, based on indirect estimates, for an equivalent period.  6 Figure 1.2: Infant Mortality Rates and Per Capita GNP in 14 Asian Countries. 1989 140 120 0 Q 100 5 4 UU 42 60 - Phpinesl 40- 20 50 10 1500 2000 2500 3000 3500 4000 4500 Per Capita GNP Source: World Bank Data File Determinants of Infant Mortality 1.16 Several hypotheses can be advanced to explain the stall in the IMRa. Income growth seems to play a significant role in reducing infant mortality in the Philippines; over this period of mortality stall, the growth in household income as measured by household surveys also stalled, even while per capita GNP showed modest gains. In a cross-sectional analysis of household income and the IMR, however, income does not emerge as a signifip-ant variable, probably due to the single point estimate of income contained in the cross sectional survey data. Education of mothers, however, is indicated as an important determinant with cross-sectional analysis. The educational system in the Philippines has compared well with other developing countries, /3 Evidence for this section is drawn from World Bank, "Philippines: An Opening for Sustained Growth", 1993; a background paper analyzing Infant Mortality using the 1988 National Demographic Survey; and various secondary data sources.  7 but there are indications at the sectoral level of stalled improvement during the 1980s (parallel with the IMR) for cohort survival rates, an important indicator of educational efficiency and equity. Fertility rates also show a slow rate of decline since 1970 compared to other Asian countries. 1.17 Environmental variables (safe water, household sanitation) are indicated as significant correlates of the IMR outcome in a regional analysis regression, but it is not clear how access to such services may have contributed to the stalled IMR. Public health expenditure and a key public health output affecting infant and child health (DPT immunization levels), emerge as highly significant explanatory variables in the regional analysis. Those regions that managed reductions in infant mortality during the 1980s also were targeted with primary care and public health expenditure. 1.18 In the Philippines, determinants of IMR variation by household or by region are, therefore, likely to be similar to those reported for other countries. Explaining the IMR stall over time is a more difficult question. Some of the macro and sectoral indicators also show stalled performance, but these correlations amount to hypotheses, not findings. The duration of the IMR stall is an unusual occurrence in East Asia. If the IMR does not resume its expected decline in the late 1980s and early 90s, health policy assumptions will need to be reassessed and research on IMR determinants would be a high priority. Leading Causes of Mortality 1.19 Infectious and preventable diseases still represent the leading causes of death in the Philippines. Reports on infant mortality point to respiratory conditions and diarrheal diseases as the major killers among children under one year of age. In addition, nutritional deficiencies as well as septicemia and problems related to the birth event are classified among the ten leading causes of infant mortality. This mortality profile has lead the DOH to expand several public health programs addressing these main causes of death, namely respiratory infections and diarrhea. 1.20 Table 1.2 reports on the leading causes of death for the total population over a 5-year period. The Table shows that in 1989, the leading cause of death was pneumonia, followed by diseases of the heart and vascular system, tuberculosis, malignant neoplasms, diarrheal diseases, kidney disorders, and septicemia. Most of these causes are preventable and infectious in nature. An increase in the level of heart diseases and malignant neoplasms is also noted. This increase could be a reflection of better reporting or of a real increase in mortality risk. If mortality from chronic ailments is truly on the rise, such an increase could have important implications for public health services and the need to adapt them to a more demanding and expensive-to-treat range of conditions. The change in the disease pattern could vary regionally as well, thus affording local governments an opportunity to tailor local health services accordingly.  8 Table 1.2: Mortality Rates (Rer 100.000 goDulation) for the Ten Leading Causes of Death. Philippines. 1985-89 Year Disease 1985 1986 1987 1988 1989 Pneumonia 96.7 90.4 91.3 91.8 77.0 Diseases of the Heart 66.3 69.9 66.0 65.1 74.6 Diseases of the Vascular System 49.7 52.5 49.6 51.5 56.1 Tuberculosis (All Forms) 57.9 54.6 55.2 54.0 43.8 Malignant Neoplasms 33.2 32.8 33.4 33.9 36.5 Accidents 18.4 18.5 19.1 17.7 19.6 Diarrheal Diseases 21.1 19.4 23.2 21.8 13.5 Measles 14.7 11.2 13.8 16.1 11.2 Nephritis, Nephrotic Syndrome, 10.0 9.4 9.4 9.4 8.5 and Nephrosis Septicemia - - - 7.5 8.2 Source: DOH, "Philippines Health Statistics," 1985-1989 Nutritional Status 1.21 The nutritional problems in the Philippines are characterized by. protein-energy malnutrition (PEM) and micronutrient deficiencies. Tle available data show overall improvements in nutritional status during the second part of the decad, after a deterioration in the early 1980s. Continuing problems include protein-energy malnutrition (PEM), particularly among pre-school and school children, iron deficiency anemia, vitamin A deficiency, and goiter or Iodine deficiency. 1.22 Dietary Intake. The mean per capita caloric intake has remained inadequate over the last decade. Nationwide surveys conducted by the Food and Nutrition Research Institute (FNRI) in 1978 and 1982 showed that the caloric intake of Filipinos was about 89 percent adequate compared to the RDA. During the economic crisis in the 1980s, further declines in caloric intake are hypothesized and in 1987, it recovered slightly to 87 percent of the recommended allowance. Chronic dietary energy deficiency is particularly of concern among preschool children, and pregnant and lactating mothers. The mean energy intake of preschool children comprised only 65 percent of their RDA, while pregnant and lactating mothers barely reached 70 percent of their desirable caloric intake levels. In comparison, the results from the 1989-90 survey reflect improvements in the overall nutritional status. This mixed picture on dietary intake is reflected in both protein-energy malnutrition and micronutrient deficiencies, as discussed below. 1.23 Protein-Energy Malnutrition (PEMO. PEM refers to a range of clinical disorders due to a deficiency of protein or calories or both. It is indicated by the growth, weight and height of children and weight and height of adults relative to established standards. The internationally accepted  9 standard is the US National Center for Health Statistics (NCHS) Standard. Table 1.3 compares the malnutrition rates for the Philippines with three other Asian countries using minus two standard deviation of the NCHS standard as the cut-off point for malnutrition. Underweight, wasting and stunting rates among children under five in 1987, respectively, put the Philippines at an advantage relative to Indonesia but in a worse position relative to Thailand. The Philippines has the highest percentage (18 percent) of low birthweight in comparison to the other three countries. Table 1.3 Basic Nutrition Indicators for Selected Asian Countries (1987) % of Children Under Five % Low Country Birthweight Underweight Wasting Stunting China 6 21 8 41 Indonesia 14 51 11 46 Philippines 18 33 7 42 Thailand 12 26 10 28 Source: World Bank Data File 1.24 Trends in malnutrition rely on the Filipino standards for malnutrition. Figure 1.3 shows that underweight, a measure of current undernutrition among children under six, shows a decline from 21.9 percent in 1978 to 17.2 percent in 1982, increasing slightly to 17.7 percent in 1987, then decreasing to 14 percent in 1989-90. The prevalence of moderate and severe underweight has been declining, while milder degrees of underweight children (not shown) remain high at about 50 percent. This is partly attributed to shifts from severe to moderate and to mild malnutrition. Wasting among the under six, a measure of acute malnutrition, has declined slightly. 1.25 Micronutrient Deficiency. A series of micronutrient deficiencies affect health status in the Philippines. The three major health problems are: Iron Deficiency Anemia (IDA); Iodine Deficiency (IDD); and Vitamin A Deficiency (VAD). IDA is perhaps the most prevalent nutritional deficiency in the country. The 1987 National Nutrition Survey indicated that the prevalence of anemia has increased compared to results of the 1982 survey. Iodine deficiency is also common in the Philippines to the extent that the country is classified as the most goitrous country in Southeast Asia. Vitamin A Deficiency (VAD), as evidenced by night blindness and Bitot spots, occurs in clusters of high prevalence coinciding with low economic status.  10 Figure 1.3: Percentage Malnourished Children under 6. Philippines. 1978-90 25 Percentage 2S0 20 ..................... ................... ............................................................ 15 1 ........ ....... . .. 5 _ ... g 1.......-................. Underweight Wasted Stunted Wasted & Stunted Nt Inourished Children 0-6 Years 1978 M 1982 MM 1987 2 1989-90 Source: National Nutrition Surveys, 1978-1990 Fertility 1.26 Table 1.4 presents the Philippines' fertility levels and their rate of decline in comparative perspective. Similar to the pattern for the IMR, the Philippines performance lags behind two countries that began with similar demographic profiles and launched their family planning programs at roughly the same time. Analysis of this differential performance has isolated the level of family planning program effort and its consistency over time as the major factors affecting the lower Philippines' performance. Indeed, the Philippines falls behind all East Asian countries except China (which began the period with the highest CPR in East Asian developing countries) in terms of the annual percentage increase in contraceptive prevalencei4. As a determinant of health status, and with direct causal links to infant mortality risk, the TFR performance of the Philippines may be a contributing factor to the recently flat trend in the IMR. "4 World Bank, "New Directions in the Philippines' Family Planning Program", Population and Human Resources Division, Country Department II, Asia Region, October 1, 1991, pp. 44-50.  11 Table 1.4: Fertility Declines in Thailand. Indonesia. and the Philippines Country 1965\69 1988 Decline Thailand 6.30 3.0 3.30 Indonesia 5.57 3.4 2.17 Philippines 5.72 4.3 1.42 Source: World Bank, "New Directions in the Philippines' Family Planning Program", Population and Human Resources Division, Country Department II, Asia Region, October 1, 1991, pp. 45. Conclusions on Health Status 1.27 In reviewing the key health status indicators for the Philippines over time, there is no basis for complacency. Fertility has declined slowly, and key measures of malnutrition are comparatively high (low birthweight) or increasing (micronutrient deficiencies such as maternal anemia and iodine deficiency). On the leading causes of mortality, there is some evidence of increasing rates of chronic diseases affecting adults, but infectious and preventible diseases still represent the leading causes of death. The most striking phenomenon in measures of health status in the Philippines remains the apparent stalling of the IMR decline from the late 1970s to the mid 1980s. The performance of the IMR after 1986-87 may prove to show some improvement. overall, health status achievements are mixed; the new decentralized health system will need to accommodate initiatives to address relatively poor outcomes in fertility and nutrition, while sustaining public health programs that are targeted to at-risk children and mothers. C. Health Delivery System - Structure and Evolution 1.28 Introduction. The structure of the Philippines health system is dominated by a single characteristic, the roughly equal shares of the public and private sectors in terms of health expenditure (Table 1.5). The roughly equal size of the two sectors is also reflected in the public and private shares of personnel and facilities, both of which are widely and, in some cases, abundantly available throughout the country. 1.29 The well balanced presence of both public and private providers suggests they may either compete or complement one another in various regional and service markets and that they should enjoy comparative advantages in specific sub-sectors. In addition, private and public health infrastructure is distributed differently in response to 1^cal income levels and health risk, an equity issue of considerable interest given the Local Government Code. And finally, the evolution of the public system has deconcentrated some authority and responsibility from central to local components of the public delivery structure, another important precursor to the devolved system. These three issues and their implications are covered in this section.  12 Table 1.5: Total Health Care Expenditure (mil Peso) Percent Share National Public Private Total Public Private 1981 2736 5143 7879 34.7 65.3 1982 3309 6014 9323 35.5 65.5 1983 3921 7025 10946 35.8 64.2 1984 3596 8760 12356 29.1 70.9 1985 3779 6052 9831 38.4 61.6 1986 4870 6116 10986 44.3 55.7 1987 6582 6786 13368 49.2 50.8 1988 6874 6771 13645 50.4 49.6 1989 8587 7312 15899 54.0 46.0 1990 9085 7823 16908 53.7 46.3 Source: Except 1985 and 1988, private expenditures are based on NEDA and staff estimates. Public expenditures are from Orville Solon, Alejandro Herrin, et al, "Health Sector Financing in the Philippines," 1991. Structure of The Public System (Pre-Devolution) 1.30 The pre-devolution public system has had both central and local dimensions. At the central level, the Department of Health has presided over a national service delivery structure operating out of basic health and hospital facilities located at the Barangay (village) level, and up to the municipal, province, regional and national level (see Figure 1.4). The notable exception is found in the 60 chartered cities in the Philippines which have direct authority over their health delivery systems, subject only to supervision and regulation from the DOH. Thus, decentralization of health services has long been the case at the chartered city level. 1.31 The central DOH in Manila has been organized into five offices, each headed by an Undersecretary reporting to the Secretary of Health. The five offices have managed respectively: (a) Public health services(the major disease control programs, such as malaria, EPI, and TB); (b) Hospital services (management of national medical centers, regional hospitals, and the provincial level general hospitals);  13 Figure 1.4 Macro Health System Structure [urretil MatitrDa oa Kegs I. I Il F l l [ Offices ... .. .. .. . .... Provlnces Ice HospItalS IIospIialI * Districts | _ It#_ |Dist.H.Officea ~I -I) tIt urlion icipal I a ,',1 ' it::: osia s 1 1 111 RHU :::a 'r.'t:: AMIher 5 hS 1>S 516I ovr. a Oovn