H N P D I S C U S S I O N P A P E R Are Free Government Health Services The Best Way to Reach to Poor? Davidson R. Gwatkin September 2004 ARE FREE GOVERNMENT HEALTH SERVICES THE BEST WAY TO REACH THE POOR? Davidson R. Gwatkin September 2004 Health, Nutrition and Population (HNP) Discussion Paper This series is produced by the Health, Nutrition, and Population Family (HNP) of the World Bank's Human Development Network (HNP Discussion Paper). The papers in this series aim to provide a vehicle for publishing preliminary and unpolished results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For free copies of papers in this series please contact the individual authors whose name appears on the paper. Enquiries about the series and submissions should be made directly to the Editor in Chief. Submissions should have been previously reviewed and cleared by the sponsoring department which will bear the cost of publication. No additional reviews will be undertaken after submission. The sponsoring department and authors bear full responsibility for the quality of the technical contents and presentation of material in the series. Since the material will be published as presented, authors should submit an electronic copy in a predefined format as well as three camera-ready hard copies (copied front to back exactly as the author would like the final publication to appear). Rough drafts that do not meet minimum presentational standards may be returned to authors for more work before being accepted. The Editor in Chief of the series is Alexander S. Preker (apreker@worldbank.org); For information regarding this and other World Bank publications, please contact the HNP Advisory Services (healthpop@worldbank.org) at: Tel (202) 473-2256; and Fax (202) 522-3234. ` ISBN X-XXXXXX-XX-X © 2004 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 All rights reserved. ii Health, Nutrition and Population (HNP) Discussion Paper Are Free Government Health Services The Best Way To Reach the Poor? Davidson R. Gwatkina aPrincipal Health and Poverty Specialist, TheWorld Bank, Washington, D.C. USA Paper prepared for the World Bank's Resource Allocation and Purchasing Project Abstract: The paper asks whether the pursuit of universal coverage by a wide range of free government services constitutes the most promising approach to meeting the needs of disadvantaged population groups. The response given to that question is "probably not." The record to date points clearly to the danger that the benefits of subsidized government health services will flow primarily to the better-off, rather than to the poor for whom the services are intended. While there is no perfect approach to dealing with this issue, the record also points to several approaches that can significantly ameliorate the situation. Two of them, discussed in the paper, are the adoption of targeting measures to increase the proportion of benefits from government expenditures that flow to the poor; and the development of alternative, self- sustaining service financing and delivery mechanisms to serve the better-off. Successful implementation of approaches like these would allow governments to focus their efforts to achieve universal free coverage on a limited number of interventions that are particularly important for poor groups. Keywords: resource allocation and purchasing, health care financing, free government health services, universal coverage, targeting; social security. Disclaimer: The findings, interpretations and conclusions expressed in the paper are entirely those of the authors, and do not represent the views of the World Bank, its Executive Directors, or the countries they represent. Correspondence Details: Davidson R. Gwatkin; Room G7­005, Mail Station G7­702, The World Bank, 1818 H Street, NW, Washington, DC 20433 USA. Tel: (202) 473-3223; Fax: (202) 522-3234; Email: Dgwatkin@worldbank.org; iii iv Table of Contents FOREWORD............................................................................................................................VII ACKNOWLEDGEMENTS ..................................................................................................... IX INTRODUCTION........................................................................................................................1 THE BENEFICIARIES OF GOVERNMENT HEALTH SERVICE EXPENDITURES ....2 THE FIRST SOURCE: BENEFIT-INCIDENCE REPORTS....................................................................2 THE SECOND SOURCE: COMPARABLE COUNTRY STUDIES OF SOCIOECONOMIC INEQUALITIES IN HEALTH SERVICE USE.................................................................................................................2 FOCUSING GOVERNMENT SERVICES ON THE POOR..................................................3 TYPES OF TARGETING .................................................................................................................4 THE EFFECTIVENESS OF TARGETING ...........................................................................................5 ENCOURAGING THE BETTER-OFF TO PAY FOR THEIR OWN SERVICES .............7 PAYMENT FOR GOVERNMENT SERVICES .....................................................................................7 ESTABLISHMENT OF ALTERNATIVE MECHANISMS ......................................................................7 CONCLUSIONS ..........................................................................................................................9 REFERENCES...........................................................................................................................10 v vi FOREWORD Great progress has been made in recent years in securing better access and financial protection against the cost of illness through collective financing of health care. This publication ­ Are Free Government Health Services the Best Way to Reach the Poor by Davidson R. Gwatkin ­ is part of a series of Discussions Papers that review ways to make public spending on health care more efficient and equitable in developing countries through strategic purchasing and contracting services from nongovernmental providers. Promoting health and confronting disease challenges requires action across a range of activities in the health system. This includes improvements in the policymaking and stewardship role of governments, better access to human resources, drugs, medical equipment, and consumables, and a greater engagement of both public and private providers of services. Managing scarce resources and health care effectively and efficiently is an important part of this story. Experience has shown that, without strategic policies and focused spending mechanisms, the poor and other ordinary people are likely to get left out. The use of purchasing as a tool to enhance public sector performance is well documented in other sectors of the economy. Extension of this experience to the health sector is more recent and lessons learned are now being successfully applied to developing countries. The shift from hiring staff in the public sector and producing services "in house" from non governmental providers has been at the center of a lively debate on collective financing of health care during recent years. Its underlying premise is that it is necessary to separate the functions of financing health services from the production process of service delivery to improve public sector accountability and performance. In this Discussion Paper, Gwatkin makes a compelling case for including a pro poor focus under the strategic framework of resource allocation and purchasing. Without it the poor are often left out. Alexander S. Preker Lead Economist Editor of HNP Publications vii viii ACKNOWLEDGEMENTS The author acknowledges with appreciation the helpful advice and suggestions received from Cristian Baeza and Pablo Gottret. But neither should be held responsible for whatever errors may remain despite their efforts. The author of this Report is grateful to the World Bank for having published the Report as an HNP Discussion Paper. ix x INTRODUCTION Equity is a frequently stated justification for government involvement in the health care market. This is often taken to mean directly providing all segments of the population with a wide range of government-operated health services at no cost: free universal care. Yet a look at the record suggests that this goal all too often remains elusive, especially in poor countries; that governments in fact serve only a some of the population; and that the people served are disproportionately concentrated among the better-off. When this happens, government health services, far from promoting equity, work against it. What is to be done? One option is to persist: to increase the resources devoted to existing government-operated service programs in order to expand coverage beyond the high-income areas where they usually concentrate, out toward the poorer periphery. With enough time and resources, this strategy should eventually lead to inclusion of increasing numbers of poor people, and a diminution or elimination of inequalities in government coverage. To many observers, this would be the ideal situation, but the ideal can be a formidable enemy of the good. For example, it is difficult to envisage the availability of adequate government funds to permit significant progress toward a universal coverage goal in most, if not all, developing and transition economies. Where resources are scarce, a continued concentration on the expansion of government services as they currently exist would be much more likely to produce frustration than progress and limited, if any, increase in affordable services for the neediest. Such constraints need not be fatal, however, since the continued pursuit of universal coverage by widening the range of free government services is not the only, or necessarily the best, approach to benefit the poor. Even when equity-oriented governments do not wish to abandon altogether their universal free coverage aspirations, they have the option of limiting universal coverage efforts to a few interventions of particular epidemiological effectiveness, and of finding alternative ways to achieve equity for other services. Many alternative ways are available. For example, often the pattern of government expenditures can be modified to increase the proportion that benefits the disadvantaged. Or, to cite another possibility, one might try encouraging upper income people to cover a larger share of their health care costs, thereby freeing up government health service resources to serve previously excluded groups. The purpose of this chapter is to make and illustrate the basic point made above: that there are many ways for governments to pursue the goal of ensuring that the poor receive adequate, affordable services through alternative approaches to resource allocation and purchasing. This assertion is explored in the three sections that follow. The first section summarizes the information known about the distribution of benefits from government health services across social groups in order to document the regressive pattern that now frequently exists and the need for significant changes in approach if the poor are to benefit. The second and third sections 1 illustrate the kinds of changes that might be considered, discussing in further detail the two examples noted briefly above. THE BENEFICIARIES OF GOVERNMENT HEALTH SERVICE EXPENDITURES There are two principal sources of information about who benefits from government health service expenditures. One is a series of reports, using different variants of a methodology known as benefit incidence, which present information about overall government health service expenditures. The second is a set of 45 comparable country studies on socioeconomic inequalities in health status and service use, showing who uses a set of specific maternal and child health services. Both sources, while varied in nature, show that overall government health service programs usually--though not always--benefit the better-off more than the disadvantaged. This is true especially for secondary and tertiary care, which usually accounts for most government health care expenditures, less so for primary care and preventive services. THE FIRST SOURCE: BENEFIT-INCIDENCE REPORTS The many ins and outs of benefit-incidence methodology, used in some 25 to 30 country studies thus far, has been explained elsewhere (Castro-Leal, et al. 2000; Demery 2002; Gwatkin 2000). Briefly put, this methodology is the equity analogue to the better known cost-effectiveness analysis. That is, where cost-effectiveness analysis examines the amount of output produced per unit of resources invested in a health service program, benefit-incidence analysis is concerned with how that output is distributed across socioeconomic groups. It typically draws on two types of data: household surveys with information about socioeconomic status and utilization of health services, to estimate the use of different types of government services by each of several socioeconomic groups; and government statistics and financial records, which permit construction of an estimate for the unit cost of each type of government service covered. The volume of each service used by each socioeconomic group is multiplied by the unit cost of that service and then adjusted to account for any user fees paid by group members; and the results are summed for each socioeconomic group. The result is an estimate of the total amount of government health service resources used by the different groups. 2 Table 1. How Much Do the Poor Gain from Government Health Service Expenditures in Africa? (percent) Primary Care Total Care Percentage of benefit gained by: Percentage of benefit gained by: Poorest Richest Poorest Richest population population population population Country quintile quintile quintile quintile Côte d'Ivoire (1995) 14 22 11 32 Ghana (1992) 10 31 12 33 Guinea (1994) 10 36 4 48 Kenya (rural, 1992) 22 14 14 24 Madagascar (1993) 10 29 12 30 Tanzania (1992­93) 18 21 17 29 South Africa (1994) 18 10 16 17 Unweighted average 15 23 12 30 Note: The percentages refer to the total financial benefits from government health care expenditures accruing to the poorest and richest population quintiles Source: Castro-Leal, et al. (2000). Perhaps the best known set of cost-benefit estimates is a set of seven country studies for Africa, whose findings are summarized in table 1. As can be seen there, 30 percent of total government health care expenditures went to benefit the top 20 percent of the population--more than two- and-a-half times the 12 percent that benefited the poorest 20 percent. The record was better with respect to primary care; but even there, the top 20 percent of the population received on average one-and-a-half times as much gain as did the bottom 20 percent (23 percent of total financial benefit for the top quintile, vs. 15 percent of total financial benefit for the bottom quintile). 3 ni (7) (5) (2) care averages Richest quintile 30 21 27 )8(41 al health (7) (5) (2) (region Total Poorest quintile 12 19 13 )8(92 Population (5) (1) (2) -- the l Richest quintile 33 22 29 of ta left.'s To (5) (1) (2) Percent -- parenthese Poorest quintile 10 13 12 the 20 to Richest (2) (1) -- -- 34 41 and Care Richest quintile immediately Inpatient appears Poorest Hospital (2) -- -- to Poorest quintile 16 (1)5 that 1 ing average the Accru (2) (1) -- -- in Richest quintile 36 41 uded Services incl Outpatient (2) -- -- Health Poorest quintile 12 (1)7 countries of (7) (2) (2) number -- Government care Richest quintile 23 16 22 the from indicates Primary (7) (2) (2) -- Poorest quintile 15 21 16 Subsidy parentheses in Financial Europe acire (2001). figure 2. on Table percentage) Regi Africa Asia Eastern AmnitaL Gwatkin Each Note: Source: Table 2 presents summary figures from a collection of 22 benefit-incidence studies (including the 7 included in table 1), from four developing and transitional regions. The figures need to be interpreted with considerable caution because of possible interstudy methodological differences but are nonetheless of interest in showing that higher gains for the better-off are the norm in all regions, with the exception of Latin America (about which more later). All in all, the top quintile received more than the bottom quintile in 13 of the 22 countries--in 12 of the 14 countries outside of Latin America. This is similar to the findings from a more recent compilation showing a regressive pattern of overall expenditures in 17 of 25 developing societies covered and in 15 of the 17 societies outside of Latin America (Wagstaff 2002). THE SECOND SOURCE: COMPARABLE COUNTRY STUDIES OF SOCIOECONOMIC INEQUALITIES IN HEALTH SERVICE USE The 45 country studies (Gwatkin, et al. 2000)constituting the second source of information are based on the household data from the intercountry Demographic and Health Study (DHS) program. These data are from comparable large-scale household surveys that focus on maternal, child, and reproductive health. The data sets also include information about household assets or attributes (e.g., type of roof and flooring; availability of electricity; possession of bicycle, radio, watch; source of drinking water). This information was combined to form a single wealth index. The index was used to divide the study population into wealth quintiles; and the values were tabulated for about thirty health status and service use indicators (e.g., infant mortality, total fertility, immunization, use of modern contraception).1 Table 3. Distribution of Benefits of Three Government Maternal and Child Health Programs across Socioeconomic Classes A. Treatment rates by socioeconomic population quintile Percentage of cases receiving treatment in a government facility (unweighted country averages) Bottom Second Middle Fourth Top populatio population population population population Number of Intervention n quintile Quintile quintile quintile quintile countries Medical treatment, diarrhea 20.3 21.2 22.4 23.1 23.3 37 Medical treatment, acute respiratory infections 24.5 28.5 30.9 31.7 31.8 34 1 The service indicator data available do not indicate the amounts paid for public services. This introduces the possibility of distortion that would occur if, say, the better-off were to pay fully for their services, while the poor receive theirs for free. While this consideration deserves to be borne in mind while assessing the statistics, such anecdotal evidence as exists provides little basis for believing that it introduces a major distortion. 2 B. Intercountry differences in treatment inequality rates Col. 1 Col. 2 Col. 3 Col. 4 Number of countries Percentage of countries with higher treatment with higher treatment rate in top than in rate in top than in Intervention Number of countries bottom quintile bottom quintilea Medical treatment of diarrhea 37 24 64.9 Medical treatment of acute respiratory 34 26 76.4 infections Obstetric deliveries 45 43 95.6 a. Col.3/col.2 x 100. Source: Gwatkin, et al. (2000). Three of the indicators available for most, but not all, of the studies referred to use of public facilities: for treatment of children' diarrhea and of acute respiratory infections among children; and for obstetrical deliveries. The results are summarized in table 3. As can be seen in section A of that table, coverage rates increase steadily across quintiles for each of the three interventions, with the coverage in each quintile being higher than that in the quintiles below it. The rate of increase is modest for relatively simple interventions like medical treatment of diarrhea or acute respiratory infection. As shown in section B, in a substantial minority of countries, the coverage rate is somewhat higher in the bottom quintile than in the top quintile. For more complex treatments like attended deliveries, the picture is starker: the upper income groups are much more likely to receive care in public services in almost all countries. The coverage rate in the top quintile is on average nearly three times greater than in the lowest quintile. In brief, even the simplest interventions offered through government facilities usually reach the better-off at least somewhat more frequently than they do the neediest, for whom they are especially intended. For more complicated and expensive interventions to which the most government health service resources are dedicated, coverage inequalities favoring the better-off are much larger. FOCUSING GOVERNMENT SERVICES ON THE POOR For observers who believe that lessening disparities between the better-off and the disadvantaged is an important role of government health service provision, the figures presented in tables 1 through 3 are not encouraging. The picture, while by no means totally black, is certainly a much darker shade of gray than is often realized. 3 What, then, if anything, can be done to reorient government services more toward the disadvantaged groups who must be reached to achieve the equity objective? While no definitive answer is available, strong suggestions emerge from a long tradition of experience in the health and other social sectors in trying to reach the poor. TYPES OF TARGETING The measures used for this purpose, generally known by the infelicitous term "targeting," are of several types. Different authors employ different taxonomies and terminology in describing these types, but almost all differentiate between what might be called "individual," "categorical," and "self-" targeting.2 Individual or "direct targeting" applies to programs that, through some sort of means testing, seek to direct their benefits toward particular individuals who are poor. The means testing can be direct, through a careful, objective assessment of an individual's income, wealth, or both. Or it can be indirect, using "proxy means testing," under which individuals are selected on the basis of a few selected characteristics (e.g., type of roof on house, level of education) shown to be closely associated with poverty through careful statistical analysis of large household data sets. Categorical targeting also goes under a variety of other names--such as "indirect," "broad," "characteristic," or "indicator" targeting--each with a somewhat different connotation. The methods referred to by these different names share an emphasis on particular categories or groups of people or types of programs, rather than on individuals. Such emphasis is adopted in full recognition that it cannot produce the degree of precision theoretically available through an individual approach; but in the belief that the greater administrative feasibility and political benefits of indirect targeting will outweigh its theoretical shortcomings. Examples of indirect targeting include targeting particular population groups (say, landless agricultural workers) in which the prevalence of poverty is especially high; targeting geographic areas in which most people are poor; giving priority to services that are primarily relevant for the poor (such as clean water supplies in settings where most of the better-off already have access to clean water); and targeting through initiatives to deal with problems concentrated among the poor (e.g., communicable diseases). Self-targeting, or "self-selection," relies on programs that are made universally available but thought likely to be attractive only to the poor. It is used especially in public works programs. There, low-paying temporary manual labor jobs are made available to anyone who comes for them; but the low remuneration and short duration make them attract only people in severe need. A health example would be offering free hospitalization in large wards for anybody who wishes it, in anticipation that all but the poor will prefer to pay for a private room. These different types of targeting are by no means mutually exclusive. Programs that apply two or three simultaneously are common. 2 For a fuller, recent discussion of targeting types, on which this section draws heavily, see Coady, Grosh, and Hoddinott, 2002. 4 THE EFFECTIVENESS OF TARGETING There is no known instance of a targeting program's approaching perfection--that is, of reaching all poor people within a society, while excluding all those who are not poor. There are many cases where targeting appears to make no difference. Upon occasion, it even ends up subsidizing the better-off more than the disadvantaged. But, in many other instances, targeted programs have resulted in a significantly larger portion of benefits going to the poor than under the typical government service programs described in the preceding section. Many of these instances are from outside the health sector. The most comprehensive recent review, covering 67 programs from a wide range of sectors, found that the poor got more benefits in 70 to 75 percent of cases than they would have had the benefits been evenly distributed across the population (which would in itself be an improvement over the present situation with respect to government health service expenditure programs). On average, the poor got a bit over one third more. None of the numerous targeting methods used by the various programs seemed clearly superior to the others (Coady, Grosh, and Hoddinott 2002). A similar result emerged from another widely cited review that compared 18 carefully targeted Latin American government programs in a wide range of fields (including health) with 30 less well targeted programs. The principal finding was that: · In eight untargeted programs offering general food subsidies to entire populations, an average of 33 percent of the total benefit went to the poorest 40 percent of households. · In 22 government primary care and education programs that were "loosely targeted"--that is, provided benefits not specifically targeted but considerably more relevant for the poor because the services provided were already readily accessible to the better-off--58 percent of the total benefit accrued to the poorest 40 percent. · In the 18 more carefully targeted programs, the poorest 40 percent of the households got 72 percent of the total program benefits--over twice as high a percentage as the 33 percent that went to this poor group under the untargeted programs just mentioned (Grosh 1994).3 Within the health sector, by far the greatest amount of attention has gone to one particular type of individual targeting: that is, the identification of poor individuals qualified to receive exemptions from the user fees frequently introduced as a component of health sector reforms. Here, the record has been widely varied. 3The figures cited refer to medians rather than means and pertain to the poorest 40 percent of households rather than individuals. Because of higher fertility among the poor, the 40 percent of poorest households can be expected to contain more than 40 percent of individuals in the population. A report on benefits based on the poorest 40 percent of individuals would thus be likely to show at least somewhat lower figures for all three types of program described; but since each of the three types would be affected in a similar manner, their relative effectiveness in reaching the poor would not necessarily be changed. 5 At one end of the spectrum are the many countries, especially in Africa, whose governments simply issued decrees that the poor should be exempted and then largely forgot about them. Several reviews have shown that these have had little impact (Gilson, Russell, and Buse 1995; Nolan and Turbat 1995; Russell and Gilson 1997; Waters 1994). In other places sustained effort and resources have been applied, and there the record is much different. Examples include some of the Latin American initiatives included in the general review described above and a set of 9 successful efforts identified in another overall review of the exemption experience (Willis 1993). Perhaps the largest and best known of the more promising experience has been Thailand, which has been offering free medical care to low-income groups since 1975 through an initiative known as the "Low-Income Support Program."4 The program was modified numerous times as the government gained experience with it, and by the late 1990s was open to the about 25 percent of Thai families living below the country's nationally determined poverty line. Local officials of the Ministry of Home Affairs determined which families were qualified to participate and issued identification cards to the families that qualified. The cardholders presenting their cards at government health facilities were exempted from the usual fees. The cost was covered by a special allocation to the service-providing facilities from the Ministry of Public Health, which used around 8 percent of its budget for this purpose. By the late 1990s, some 11 million people--20 percent of Thailand's population--were participating in the initiative. Independent surveys suggest that about 80 percent of these people were indeed poor, and that about 65 percent of Thailand's poor were covered. While some observers would argue that a certain amount of "leakage"--coverage of nonpoor-- should be tolerated to win support from politically important groups,5 many Thai were seriously disturbed with the 20 percent of nonpoor who were inappropriately receiving benefits. Thus in 2001 the initiative was merged with a policy of universal care introduced by a new government. Though by no means perfect in its targeting accuracy, and thus controversial within Thailand, the Thai program's record nonetheless stands in stark contrast to the usual pattern of health service beneficiaries outlined in the preceding section and serves to indicate that exemption programs can be developed in which most benefits do flow to the poor. This and the other instances covered by the reviews also point to several guidelines for developing such programs. One is to have clear and easily verifiable eligibility criteria and of arranging for people other than clinical personnel to determine who is and is not eligible (e.g., by having responsible people thoroughly familiar with local financial conditions issue identification cards to the poor for presentation at health facilities). Another is to provide for financial support from an outside source to reimburse service providers for the revenue they forgo in providing free or subsidized services. Yet another is to be willing to experiment continually over the extended period of time that may be needed to get an exemption to work correctly. 4 The information in this section is drawn primarily from Khoman (1997). For an earlier report on the Thai experience, see Mills (1991). 5See, for example, Gelbach and Pritchett (1997a and 1997b). 6 In brief, effectively focusing government services on the poor is by no means easy. Effective targeting programs cannot be painlessly achieved simply by waving a wand or issuing a decree or two. Rather, they require determination, time, and money. But where these factors are present, they can make enough of a difference in the benefits across social classes to deserve serious consideration. ENCOURAGING THE BETTER-OFF TO PAY FOR THEIR OWN SERVICES Another option is to encourage the better-off to pick up a greater part of their own health bills. This can be done in many ways. One is to charge more for the government health services used by high-income people. Another is to use governmental authority to establish alternative, self- financed mechanisms through which the better-off can obtain services at their own expense. PAYMENT FOR GOVERNMENT SERVICES As long as government health services are free to and used by everyone, any service use by members of an upper income group represents a government subsidy to that group. However, the introduction of user fees presents an opportunity to employ differential financing that could alter the amount of subsidy going to different segments of the population. One way that differential financing can be used to promote equity is to reduce the amount that the poor must pay through the fee exemptions for the disadvantaged that have just been discussed. Another way of achieving the same objective is to get upper income groups to pay more. For example, the highest price possible could be charged for higher level medical care, which typically accounts for most government health care expenditures and which, as noted earlier, is almost always used disproportionately by the better-off. Given this pattern of use, any cost- recovery measures would be borne primarily by the upper income groups. Concerns about protecting the small minority of service users who are poor could be addressed through a fee waiver system. An approach of this sort would probably be considerably more feasible politically--and more justifiable on substantive grounds, as well--if accompanied by a health insurance program covering specified upper income groups. One obvious possibility is gradually introduced mandatory catastrophe insurance for all workers in the formal sector, financed through some combination of employee and employer contributions. This would protect the few but extremely unfortunate upper income households faced with medical bills that lie far beyond even their means. ESTABLISHMENT OF ALTERNATIVE MECHANISMS A more radical approach would be to get the better-off out of government facilities altogether-- that is, by getting the government out of the business of providing or financing services for the better-off. This might be done by, say, using government regulatory powers to foster the establishment of a fully self-financing private commercial health sector serving the better-off. 7 While few if any countries are known to have tried such an approach in the "pure" form just described, a rough approximation of it exists in the social security programs for formal sector workers in many places, especially in Latin America. Latin American countries have a long tradition of government social security programs that include health insurance for formally employed workers financed through payroll taxes and employee contributions. The programs often operate their own health facilities open only to enrolled workers, reimburse workers for services they obtain through the private sector, or both. As of the mid-1990s, such programs existed in 17 countries in Latin America and the Caribbean. The programs were generally large: in 9 of the 17 countries, the funds flowing through them represented one half or more of government health expenditures; in 5 of the 17, the programs covered more than half the population. In the 2 leading countries, Chile and Costa Rica, the social security health programs covered around 90 percent of the population and accounted for a comparable percentage of government health expenditures (Suárez-Berenguela 1998). In most of these programs, the middle- and upper income groups who constitute the majority of participants were directly covering most of the programs' costs, although some degree of government contribution to the programs was often involved. As of 1977 (the latest year with readily available comparative data), systems in four of the nine countries that provide figures were entirely self-financing, with no government subsidies; in the others, the amount of subsidy from government revenues ranged from 4 to 15 percent (World Bank 1987). Beyond this, evidence from one national initiative, in Chile, suggests that social security health programs can be used as a mechanism for channeling government subsidies to the poor. The Chilean program, known as FONASA, delivers no services itself, instead covering the costs of services provided by others, both government facilities and private practitioners. FONASA reaches about 70 percent of the country's population; around 40 percent of its beneficiaries are legally indigent and pay no premiums. Of the government subsidy that FONASA receives, more than 90 percent goes to services that reach these legally indigent. This suggests that governments of countries with effective social security health programs can think of grants to or contracts with those social security programs as a means of providing health services to the poor, as an alternative to direct service provision through government clinics, in much the same manner that nongovernmental and community organizations are used elsewhere (Bitran 1998). The social security programs as just described, however, are far from ideal. Leaving aside the questions often raised about their efficiency, there are equity issues as well. The most frequently mentioned one concerns the two-tier health system in most countries with social security programs--well equipped and staffed facilities for the better-off, who constitute the majority of social security program participants; much less adequate health ministry services for the neediest, who are largely excluded from social security systems. In addition, it is not possible to be sure that the volume of government health resources available for the poor is greater than would otherwise have been the case. While data inadequacies prevent any clear judgment, the possibility cannot be ruled out that without the political pressure created by upper income groups concerned about health services for themselves, government health service budgets are lower for each person served than they would have been had they been required to serve all income groups rather than just the needy. 8 All in all, as is often the case, the record is far from entirely clear or unambiguous. But at least part of it is striking: namely, just how much better government health care expenditures are oriented toward the poor in Latin America than in other regions. Particularly stark is the contrast with Africa, where there is rarely any significant alternative to government health services for upper income groups. In Africa, as shown earlier, the top 20 percent of the population gets two and a half times as much benefit from government health expenditures as does the bottom 20 percent. In Latin America, the figures from eight countries reproduced in table 2 show a situation that is almost exactly opposite: on average, the poorest 20 percent of the population gets more than twice as much benefit as the richest 20 percent (29 percent vs. 14 percent of total benefit). While this difference is no doubt attributable to many factors, it is difficult to resist the strong sense that Latin America's reliance on social security programs for the better-off has played a significant role. CONCLUSIONS Perhaps the best way to conclude is by returning to the question posed at the outset, "Does the pursuit of universal coverage by a wide range of free government services constitute the most promising approach to meeting the needs of disadvantaged population groups?" The response to that question presented in the preceding sections is "probably not." The record to date points clearly to the danger that the benefits of subsidized government health services will flow primarily to the better-off, rather than to the poor for whom the services are intended. While there is no perfect approach to dealing with this issue, the record also points to several approaches that can significantly ameliorate the situation. Two of them, discussed here, are the adoption of targeting measures to increase the proportion of benefits from government expenditures that flow to the poor; and the development of alternative, self-sustaining service financing and delivery mechanisms to serve the better-off. Successful implementation of approaches like these would allow governments to focus their efforts to achieve universal free coverage on a limited number of interventions that are particularly important for poor groups. 9 REFERENCES Bitran, R. 1998. "Equity in the Financing of Social Security for Health in Latin America." Major Applied Research Paper. Bethesda, Md.: Abt Associates Inc., Partnerships for Health Reform. Castro-Leal, F., J. Dayton, L. Demery, and K. Mehtra. 2000. "Public Spending on Health Care in Africa: Do The Poor Benefit?" Bulletin of the World Health Organization 78(1): 66­74. Coady, D., M. Grosh, and J. Hoddinott. 2002. "The Targeting of Transfers in Developing Countries: Review of Experience and Lessons." Draft paper for Safety Nets Primer Series. Washington, D.C.: World Bank, Social Protection Anchor Unit, April 16. Demery, L. 2002. "Benefit Incidence Analysis," Chapter 2 in Tool Kit: Techniques for Evaluating the Poverty Impact of Economic Policies. Washington, D.C.: World Bank, October. (http://poverty.worldbank.org/files/12995_TKWeb_Toolkit_all.pdf) Gelbach, J.B., and L.H. Pritchett. 1997a. "More for the Poor Is Less for the Poor: The Politics of Targeting." World Bank Policy Research Working Paper 1799. Washington, D.C.: World Bank, July. ------. 1997b. "Indicator Targeting in a Political Economy: Leakier Can Be Better." Manuscript. Washington, D.C.: World Bank, September. Unpublished. Gilson, L. S. Russell, and K. Buse. 1995. "The Political Economy of User Fees with Targeting: Developing Health Financing Policy." Journal of International Development 7(3): 369­401. Grosh, M. 1994. Administering Targeted Social Programs in Latin America: From Platitudes to Practice. Washington, D.C.: World Bank. Gwatkin, D.R. 2001. "Poverty and Inequalities in Health within Developing Countries: Filling the Information Gap." In Poverty and Inequality in Health: An International Perspective. D. Leon and G. Walt, eds. London: Oxford University Press. ------. 2000. "The Current State of Knowledge about Targeting Health Programs to Reach the Poor." Washington, D.C.: World Bank, February. (http://www.worldbank.org/wbi/healthflagship/dl_targeting.pdf) Gwatkin,D.R., S. Rutstein, K. Johnson, R.P. Pande, and A.Wagstaff. 2000. Socioeconomic Differences in Health, Nutrition, and Population. Washington, D.C.: World Bank, Health, Nutrition, and Population Department. Khoman, S. 1997. "Rural Health Care Financing in Thailand." In Innovations in Health Care Financing: Proceedings of a World Bank Conference, March 10-11, 1997, G.J. Schieber, ed. Washington, D.C.: World Bank. Mills, A. 1991. "Exempting the Poor: The Experience of Thailand." Social Science and Medicine 33(11): 1241­52.. Nolan, B., and V. Turbat. 1995. Cost Recovery in Public Health Services in Sub-Saharan Africa. Washington, D.C.: World Bank, Economic Development Institute. 10 Russell, S., and L. Gilson. 1997. "User Fee Policy to Promote Health Service Access for the Poor: A Wolf in Sheep's Clothing?" International Journal of Health Services 27(2): 359­79. Suárez-Berenguela, R. 1998. "El Gasto Nacional en Salud en las Américas, 1995" Consultancy Report ACS-0298-ASC9. Washington, D.C.:Pan American Health Organization, February. Waters, H. 1994. "Literature Review: Equity in the Health Sector in Developing Countries, with Lessons Learned for Sub-Saharan Africa." BASICS Project Document PN-ABW-162. Washington, D.C.: BASICS Project, AFR/SD, HHRAA Project. Wagstaff, A. 2002. "Inequalities in Health in Developing Countries: Swimming against the Tide?" Policy Research Working Paper 2795. Washington, D.C.: World Bank, March. Willis, C.Y. 1993. "Means Testing in Cost Recovery of Health Services in Developing Countries, Phase I: Review of Concepts and Literature and Preliminary Field Work Design." Major Applied Research Paper7. Bethesda, Md.: Abt Associates Inc., Health Financing and Sustainability Project, November. World Bank. 1987. Financing Health Services in Developing Countries: An Agenda for Reform. Washington, D.C.: World Bank. 11 About this series... This series is produced by the Health, Nutrition, and Population Family (HNP) of the World Bank's Human Development Network. The papers in this series aim to provide a vehicle for publishing preliminary and unpolished results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For free copies of papers in this series please contact the individual authors whose name appears on the paper. Enquiries about the series and submissions should be made directly to the Managing Editor Joy de Beyer (jdebeyer@worldbank.org) or HNP Advisory Service (healthpop@worldbank.org, tel 202 473-2256, fax 202 522-3234). For more information, see also www.worldbank.org/ hnppublications. THE WORLD BANK 1818 H Street, NW Washington, DC USA 20433 Telephone: 202 473 1000 Facsimile: 202 477 6391 Internet: www.worldbank.org E-mail: feedback@worldbank.org