65701 The Incidence of Social Spending in Côte d'Ivoire, 1986-95 Lionel Demery, Julia Dayton and Kalpanna Mehra, Poverty and Social Policy Department, The World Bank, Washington D.C. 1996 I. Introduction: the Context After years of successful growth in the 1960s and 1970s, Côte d’Ivoire has faced serious economic problems in more recent times. In part a result of terms of trade losses and an internal adjustment strategy, the country experienced negative growth for much of the 1980s and early 1990s (Demery, 1994). Living standards have fallen as a result. This period of economic decline also witnessed a significant surge in poverty, the headcount ratio increasing from 30 percent in 1985 to 46 percent in 1988. And by all accounts, it has continued to increase up to 1992 (Grootaert, 1993, 1994). These have been difficult times for the majority of the Ivoirian population, but especially for the poor. The deteriorating living standards of Côte d’Ivoire are not only reflected in GDP and income-based measures of welfare. The UNDP’s Human Development Index combines a range of indicators which measure welfare directly. According to the Human Development Index, Côte d’Ivoire ranked 136th out of 173 countries—23 places lower than its ranking on the basis of per capita GDP. Clearly, Côte d’Ivoire is lagging behind countries at similar income levels in promoting social development. Since January 1994, when the CFA franc was devalued, there has been a measure of economic recovery, and the decline in living standards has been arrested. The challenge is now to restore sustainable growth, and to ensure that such growth benefits the majority of the population, especially the poor. There are two basic requirements for broad based growth. First, growth must emphasize labor-intensive activities, such as agriculture and manufacturing. Second, Côte d’Ivoire must improve on its social development achievements. The population at large must have better access to human capital in order to benefit from the opportunities such growth presents. If the human resources of the poor fail to improve, their prospects of benefiting from growth will continue to be dim. Given this broader socioeconomic context, this analysis of public social spending seeks to answer two broad questions. First, to what extent has the reliance upon internal adjustment penalized social spending in Côte d’Ivoire, and led to a deterioration in the targeting of such spending to the poor? Second, does the evidence suggest that the poor have adequate access to human capital-enhancing services, so that any future recovery in economic growth will be broad based? The chapter is concerned only with the two critical human resources sectors, health and education. The following section deals with the approach that is taken in the chapter, this being ‘benefit’ incidence analysis. This method is then applied to spending on health (Section III) and on education (Section IV) for two years, 1986 and 1995. Section III describes the health system in Côte d’Ivoire, summarizes evidence from household surveys on health outcomes, and analyzes the incidence of government expenditures on health. Section IV then deals with education in like manner, beginning with the education system and outcomes at the household level, then analyzing the incidence of public spending on education. Finally, Section V makes comparisons with other countries, and Section VI draws some concluding observations. 2 II. Benefit Incidence Analysis Benefit incidence analysis has become an established approach since the path- breaking work on Malaysia by Meerman (1979) and on Colombia by Selowsky (1979). There has been a recent resurgence of interest in the approach, which was well reviewed in Van de Walle and Nead (1995). Until recently few applications have been attempted in Africa, though several studies are now becoming available (and are reported in section V below). The method seeks to measure how government subsidies on services such as health and education are distributed across groups in society. The distribution of these subsidies is determined by two broad factors. First, it depends on government spending itself, and how it is allocated within the sector. The lower the spending, and the greater the effective cost recovery, the lower will be the subsidy embodied in the service provided. Second, the distribution will depend on individual or household behavior—on who uses the service that the government provides. It is only by using the service (by sending a child to a primary school, or visiting the outpatient department at a hospital) that individuals and households can lay a claim to the in-kind transfer that is implicit in the subsidy. Benefit incidence analysis therefore brings together two sources of information: data on the government subsidy (estimated as the unit cost of providing the service, less any cost recovery back to the government) allocated to the different categories of service (primary schooling, in-patient hospital care, etc.); and information on the use of these services by individuals and households, which is usually obtained from household surveys. Data from these two sources are often difficult to match, usually because of the way in which the information is collected and disaggregated. This is a problem, for example, in matching public expenditures on health with the information provided in household survey data in Côte d’Ivoire (as discussed in section III). In general, government expenditures will be more equally distributed when the spending is concentrated on services that are used widely by the population, and used especially by poorer groups. If public expenditures are concentrated on primary education, or on primary-health facilities such as clinics or health centers—which are widely-used services benefiting poor and non-poor—public expenditures will tend to be more equally distributed. However, if governments spend more on high-cost services which are not generally used by poorer groups (such as university education or in-patient hospital care), the incidence of spending is likely to be more unequal. In sum, the benefit incidence of public spending depends both on the allocation of public expenditures within the sector, and on the behavior of households. An important limitation of the approach is that it does not necessarily measure the effect of government spending on the welfare of households. The real benefit to a household in using a public service lies not so much in the monetary value of that service (usually approximated by unit cost), as in the direct benefits it gives—better health, 3 literacy and better income-earning potential in the future. This should be borne in mind when interpreting the results of this chapter. III. The Public Provision of Health Services This section assesses the extent to which the poor in Côte d’Ivoire benefit from the public provision of health services. It is divided into three main sub-sections. The first documents the main characteristics of the system of health care in the country, providing estimates of public spending on health, and the subsidy that such spending implies. This is followed by a brief discussion of health outcomes: when household members become ill, how do they respond? The final section brings these two findings together in estimating the incidence of health spending in Côte d’Ivoire. A detailed analysis is provided for 1995 and then the change in the incidence of health expenditures in the past decade is examined by comparing the 1995 results with those for 1986. III. 1 The System of Health Care The public sector provides almost all health care in Côte d’Ivoire, as the private sector is very underdeveloped. The public sector is comprised of three levels of health care: primary, which includes preventive as well as basic curative services; secondary, which includes first-level referral hospitals; and tertiary, which encompasses referral care from the secondary-level and all specialty-care. The system can be characterized in two ways. First, the referral system does not work well, and the separate levels of health care tend to operate independently, instead of coordinating health care services (World Bank 1995d). One consequence of this compartmentalization is that the quality of the services varies substantial among the various types of facilities. The system has traditionally centered around tertiary-level health care services, which provide comparatively high standards of care and are well-funded. The primary level, however, has poor service delivery and facilities that are poorly staffed and equipped. In addition, little attention has been given to preventive and promotional activities such as mother and infant health care, family planning, immunization and nutrition (World Bank 1995d). The variation in funding among the levels is discussed in more detail in the following section. In the early 1990s, the Government began working to reorient the public health care delivery system. With the support of a Human Resources Sectoral Adjustment Loan (HRSECAL) from the World Bank, the program seeks to shift more resources to primary health care and first-level referral services. This new policy aims to integrate ongoing vertical health programs into a comprehensive system of primary health care, so that health care can be provided in the most cost effective manner. The new strategy is based on decentralization in the management of services and an increased participation by the community in both the management and financing of the primary level services. This approach represents a fundamental departure from past public health policy focusing on hospital-based care. To date, progress in all areas has been slow. 4 Public Expenditures on Health. Historically, Cote d’Ivoire has given high priority to public spending on human resources development. During the 1980s, the Government devoted over 50 percent of recurrent expenditures to the social sectors. Public expenditures on health accounted for about 6 percent of the total in the late 1980s and early 1990s, and the share has increased to about 8 percent in 1994. In real terms, spending on health has…. Within the health sector, the shift towards spending more at the primary level and less at the tertiary level did not occur as was intended in the health sector strategy. An annual increase had been planned in the share of recurrent expenditures for primary care, so that between 1991 and 1995 its share would increase from 35 to 42 percent of all recurrent expenditures. As shown in Table 1, recurrent expenditures for primary-level care have actually declined from 37 percent in 1991 to about 25 percent in 1994 and 1995. Within the recurrent expenditure category, about 75 percent is spent on wages for health care personnel. As in many other African countries, there is a preponderance of physicians in the capital, concentrated in tertiary-level care, and a dearth in rural areas. This is primarily a result of training in medical school, which focuses heavily on specialized services and not enough on basic health care and health promotion. A reorientation of the training of Ivoirian physicians would help to encourage future physician to practice primary health care (MSPAS 1995). The distribution of nurses and other health care personnel is well balanced among the regions of the country, a result of a decade-long campaign to promote an equitable distribution among the regions. Investment expenditures have varied considerably from year to year, showing no predominant pattern. Most importantly, actual disbursements in investment expenditures were consistently between one-third and one-half lower than the amount budgeted (not shown in table). This has made it hard to implement investment projects. 5 Table 1: Public Expenditures on Health, Recurrent and Investment, 1991-1995 (millions CFAF) 1991 (share) 1992 (share) 1993 (share) 1994 (share) 1995 (share) Recurrent Primary 15,223 36.8 10,737 25.4 10,037 26.0 11,348 25.1 13,478 26.1 Secondary 6,484 15.7 5,776 13.7 4,812 12.5 11,338 25.1 7,952 15.4 Tertiary 19,616 47.5 25,764 60.9 23,726 61.5 22,493 49.8 30,219 58.5 Total 41,323 100.0 42,277 100.0 38,575 100.0 45,179 100.0 51,648 100.0 Investment Primary 461 21.6 3,510 44.6 1,104 23.0 1,222 15.8 n.a. Secondary 649 30.5 2,949 37.4 1,924 40.0 1,557 20.1 n.a. Tertiary 1,020 47.9 1,416 18.0 1,779 37.0 4,975 64.2 n.a. Total 2,130 100.0 7,875 100.0 4,807 100.0 7,754 100.0 n.a. Grand Total 43,453 50,152 43,382 52,933 Memorandum Items Total as % of GDP GDP Deflator 91 91 91 (1987=100) Real Total 39,673 45,789 39,304 Percent change 13 -16 Note: All figures represent disbursed ('ordonnence') amounts. * 1995 annual figures are based on an extrapolation of the actual expenditures for the first eight months of the year. Sources: World Bank (1995b) for all except 1995. 1995 from preliminary tables supplied To what extent does the government subsidize the provision of health care in Côte d’Ivoire? Combining information on public-sector recurrent health disbursements for 1995 with information with data on health visits from the Social Dimensions of Structural Adjustment Priority Survey (PS) (1995),1 it is possible to estimate unit costs for health services. Public expenditure data are broken down simply by level (primary, secondary and tertiary), and by salary/non-salary. The survey obtained information on the type of health care facility visited, and these have been matched to the level of care they provide.2 Unfortunately, the questionnaire did not distinguish between hospitals at the secondary-level and those at the tertiary-level, and so it was necessary to combine these two categories when looking at unit costs. Hence, two levels of unit costs were established: one for consultations at health centers (including all primary-level facilities) 1 The survey was implemented in April 1995 by the Institut National de la Statistique. 2 Facilities considered at the primary level (called ‘health centers’ in the analysis) were: dispensaries and health posts (PMI). Hospitals and maternity hospitals were put into the seondary/tertiary level (called ‘hospitals’ in the analysis). Consultations that took place in the home were considered to be in the private sector. The consultations that fell into the ‘other’ category (for which nothing more is know), representing less than 5 percent of all visits, were not considered in the analysis. 6 and one for consultations to hospitals (including all secondary and tertiary-level facilities). For 1986, public recurrent expenditure data is only not available by level of care, so per unit cost had to be averaged across all types of facilities. The survey data are from the 1986 Living Standards Measurement Study. These survey data are roughly comparable to those collected in 1995, although they differ in several ways, as discussed in Box 1. In an ideal analysis, visits to health centers and other primary health care institutions should be distinguished from other visits (such as hospital visits) because they attract a lower public expenditure subsidy, and this should be taken into account in analyzing who benefits from health spending. For example, an in-patient visit to a hospital will require more spending and involve a larger public subsidy than a visit to a clinic or health center. And it is the poor who tend to use these low-cost facilities and the rich the high cost facilities. However, given the limitations of the 1986 public expenditure information, this was not possible. Comparisons between the two years, 1986 and 1995, is based on this more rough calculation, and an analysis for both heath centers and hospitals is also be presented for 1995. Table 2 reports government expenditures on health for 1995 by level of care. In 1995, the average cost of a visit to a primary-level health center (health post or dispensary) was CFAF 1,540, as compared with the average visit to a hospital of CFAF 1,760. The difference in the two unit costs is not as large as might be expected, since in other African countries the average cost of care at a hospital is significantly more expensive that at a health center. The reason for this small difference in unit costs is not clearly understood, but could be because so many people consult hospitals for minor outpatient services, causing the average cost of care to drop. For 1986, the average unit cost, an average across all levels of service, was CFAF 2,239, compared with CFAF 1,647 for 1995 (Annex Table 1). As these are an average of higher unit costs for tertiary care and lower costs for primary and secondary care, an analysis based on this unit cost is abstracting from a major source of inequality in health care delivery. If the poor use mainly lower levels of care, and the better off have a higher propensity to visit hospitals when sick, an analysis based on the mean subsidy for the public health sector as a whole would underestimate the true inequality in health service provision. 7 Table 2: Public Expenditures and Unit Subsidies for Health, 1995 (CFAF million unless stated otherwise) Salaries Non-Salary Total (less Cost Net Prorated Health Subsidy Level (Disbursed) (Disbursed) Expenditures Recovery)* Total Sub-Totals** visits per visit (Disbursed) (number) (CFAF) Primary: Primary: Primary: Primary 11,697 4,805 16,502 200 16,302 23,310 15,140,866 1,540 Secondary Secondary Secondary Secondary 5,378 2,655 8,033 925 7,108 and Tertiary: and Tertiary: and Tertiary: 25,277 14,361,100 1,760 Tertiary 9,840 5,441 15,281 4,711 10,570 Research 1,967 549 2,516 2,516 Administration 4,547 1,379 5,925 5,925 Social Affairs 3,393 2,775 6,168 6,168 Average: Total 36,821 17,603 54,423 5,836 48,587 48,587 29,501,966 1,647 *Cost Recovery revenues are estimated based on the 'Voted Budget' for 1995. **Spending research, administration and social Affairs has been pro-rated and added to the primary and secondary/tertiary health services, according to their respective shares in the budget, in order to account for all public health spending in the per-visit subsidy. Note: Annual spending was extrapolated based on actual spending through 8/31/95. Sources:Preliminary Tables from the Ministry of Health, 'Evolution des depenses publiques de Sante de 1992 a 1995, 10 Dec 1995. Cost Recovery data from 'Examen des Depenses Publiques dans le Secteur de la Sante', MSPAS, 1995. User charges. Cote d’Ivoire began the first phase of its cost recovery program in 1991, when fees were charged for consultations and drugs provided in tertiary-level facilities. In October 1994, user charges for pharmaceuticals and medical consultations were extended to all level of the public health care system. The basic principle is that 65 percent of the revenues remain at the local level to be used for the purchase of drugs, supplies and maintenance. The rest is returned to the Treasury and is used to help finance the recurrent costs of the health sector. The revenues retained at the local level are to be run by locally-selected committees call ‘Comites de gestion.’ The system of Comites de gestion has been set up, but is not yet functional in all local areas. In the spirit of the Bamako Initiative, there is no waiver or exemption policy. No analysis has been done to determine whether the fees deter anyone from using public health care facilities. Many suspect, however, that the use of public health care services declined following the devaluation, as a result of a sharp increase in the price of pharmaceuticals (which are imported). Overall, the system of cost recovery has been set up to provide an additional source of financing for the public health sector, with the goal that the funds generated from the program will substitute for allocations from the central government. Less attention has been given to other aspects of the cost recovery system, such as improving the quality of the services provided or extending access to those who do not currently use the services. III. 2 Ill-health and Household Responses Health Status in Côte d’Ivoire. Not only is Côte d’Ivoire one of the richest countries in Africa in per capita terms, but it also spends more than average on health as a share of GDP and in per capita terms. Nevertheless, health outcomes in Côte d’Ivoire are worse than those in comparable countries (Table 3). The infant mortality rate of 94 is higher than the average of 70 for less developed countries, and the maternal mortality rate of 1,000 per 100,000 live births is among the worst in Africa. Likewise, the fertility rate of 6.6 for 1992 is higher than for any other West African country and is higher than the average for all of Sub-Saharan Africa. These poor health indicators are mainly attributed to the deteriorating economic situation and to inappropriate human resources policies (World Bank, 1995d). Table 3: Health Status and Development Indicators, Cote d'Ivoire and Selected Countries Country Life Infant Maternal Total GNP Health Health Expectancy Mortality Mortality Fertility per Expend- Expend- itures itures at birth Rate Rate Rate capita as share per capita (M/F) (per 100,000 $US of GDP $US live births) (1992) (1992) (1988) (1992) (1991) (1990) (1990) Cote d'Ivoire 53/59 94 1,000 6.6 690 1.6 28 Other West African Countries: Burkina Faso 47/50 132 800 6.5 290 0.8 25 Ghana 53/57 81 1,000 6.1 400 1.2 14 Guinea 44/44 133 1,000 6.5 460 1.5 18 Senegal 47/50 80 950 6.1 720 1.7 30 All Africa 49/52 104 700 6.5 340 1.5 14 Less Developed Countries 62/65 70 450 3.6 900 2.1 41 Source: World Bank, 1994b. Illness and Household Response. There is little doubt, therefore, that much remains to be done to improve the health status of the Côte d’Ivoire population. What is the response of a household to an incidence of illness or injury? To what extent do households seek care for a sick or injured member? Both the 1986 and 1995 household surveys provides information about the patterns of health care use. Individuals were asked about whether or not they were ill during the preceding four weeks (1986) or two weeks (1995) and if they sought medical care. The incidence of public expenditures on health is determined fundamentally by this pattern of illness and treatment response. In 1995, 32 percent of the population reported an illness during the preceding 4 weeks,3 an small increase over 27 percent for 1986 (Annex Tables 2 and 3). This could reflect a worsening health situation, or it could be the reflection of different recall periods used in the two questionnaires, which could have caused the 1995 figures to overestimate the incidence of illness. People in the richer quintiles were more likely to report an illness than were those in the poorer quintiles.4 As with most surveys that rely on self reporting of illness, the poor (and uneducated) are less inclined to observe and therefore report an illness in the household. Women were slightly more likely than men to report an illness. There were no differences between rural and urban areas. 3 The number of people reporting an illness in the two week recall period was multiplied by 2, in order to be comparable with results from the 1986 survey. 4 Household members were allocated to quintiles according to the per capita total expenditures of the household to which they belong. Quintile 1 represents the poorest 20 percent of the population and quintile 5 the richest. 10 About 55 percent of those reporting an illness in 1995, or about 16 percent of the total population reported seeking some form of medical care. This was an overall increase from the 47 percent that reported seeking care in 1986. These averages are misleading, however, as the overall increase actually reflects an increase in health care consultations in the richest quintiles. There was a large increase in the share of people in the richest quintile who sought care (from 57 percent of those reporting an illness in 1986 to 67 in 1995), and there was no change in the poorest quintile (32 percent of those reporting an illness). The reasons for not seeking medical care are not fully understood, but two sources of information provide some explanations. First, the 1995 survey indicated that the main difficulties associated with seeking health care were the costs associated with the care. The high cost of prescription drugs and the consultation fee were cited as the two main reasons for not seeking care (these were selected over the cost of transportation, distance of the facility and the lack of qualified staff at the facility) (INS, 1995a). In fact, the price of prescription drugs has almost doubled since the devaluation. When asked in a qualitative survey about their coping mechanisms, the poor said that they often resorted to buying medicinal herbs at the market or pharmaceutical drugs from a ‘sidewalk pharmacist,’ instead of seeking care at a health care facility (INS, 1995b). Among those who went for a health care consultation, two clear patterns are evident (Figures 1 and 2). First, the poor predominantly sought care at primary-level health center and this tendency increased over time. In 1995, about 60 percent of all consultations in the poorest two quintiles were to health centers, an increase from 50 percent in 1986. Second, the richest two quintiles were more likely to seek care at a public hospital and this trend also increased between the two years. In 1995, 53 percent of all consultations in the richest quintile were to public hospitals (up from 44 percent in 1986). The use of private health care services remained relatively stable overall (8 percent in 1986 and 9 percent in 1995), but those in rural areas tended to seek care in the private sector more in 1995 (from 3 percent in 1986 to 8 percent in 1995) and those in urban areas relatively less (from 14 percent in 1986 to 10 percent in 1995). This might reflect the fact that poorest live in rural areas and are less able to afford services in the public sector (as discussed above). 11 Figure 1: Health Care Consultations: Distribution of Public and Private Facilities, 1986 100% 90% 80% share of annual consultations 70% 60% private 50% public hospital 40% public health center 30% 20% 10% 0% Rural 1 2 3 4 5 Urban Total Q uintiles Figure 2: Health Care Consultations: Distribution of Public and Private Facilities, 1995 100% 90% share of annual consultations 80% 70% 60% private 50% public hospital 40% public health center 30% 20% 10% 0% Rural 2 3 4 Urban richest poorest Total Q uintiles There are some gender differences in the pattern of health care use. Women were more inclined than males to seek care (especially in urban areas) and they were less likely to seek private care, although these differences were marginal. However, women’s health needs are different (and often greater) from those of males. This would lead us to expect that an even a higher proportion of women seeking care than what is reported here. It is possible that women find it harder to afford (either in time or money) the health services or that the services available do not meet their health needs. In addition, the fact that males are more likely to use private facilities suggests bias favoring men in decisions regarding the treatment of illness and injury (assuming that private care is of better quality). 12 III.3 The Incidence of Public Spending on Health What do the patterns of health care response by household members (as reported in the surveys) imply for the incidence of public spending? To translate this into public expenditure terms, incidence analysis simply allocates to those households which used publicly-provided health services, the subsidy (or unit cost) embodied in the care they received. Use of subsidized health services is therefore considered as an ‘in-kind transfer’ to those households that use the system. In effect, the analysis poses the question: What additional income would households need if they had to pay for services? The results of this exercise for 1995 are shown in Figure 3 and Annex Tables 4. The incidence of government spending on health is reported for each quintile, for females and males, and for rural and urban areas. On average, each member of the Ivoirian population gained CFAF 3,410 from health care consultations during 1995.5 But this is Figure 3: Per Capita Public Health Subsidy by Facility Type and Quintile, 1995 6000 5000 per capita public health subsidy 4000 Hospitals 3000 Health Centers 2000 1000 0 urban rural 2 3 4 average richest poorest Q uintiles distributed very unequally. The per capita health subsidy in rural Côte d’Ivoire is 34 percent lower than that in urban areas. And the subsidy going to the poorest quintile is 64 percent lower than the per capita subsidy obtained by the riches quintile. The per capita health subsidy for females is slightly higher than that for males. As discussed previously, this does not necessarily imply gender ‘bias,’ as the health needs of females and males are quite different. 5 The number of health visits for 1995 reported in the table is based on the visits reported by the PS. Since this was for the two week period prior to the interview, an annual estimate is obtained by increasing PS estimate by a factor of 26. This will lead to biases if the two-week period covered by the PS is not representative of health needs over the year. For example, if March/April are periods when fewer illnesses occur, the estimate obtained would be biased downward. 13 How well targeted are health subsidies to the poorest groups in Côte d’Ivoire in 1995? Overall, only 11 percent of the total health subsidy benefited the poorest 20 percent of the population. The poorest 40 percent (which approximates the poor population of the country) gained only 30 percent from government spending on health. The richest quintile, on the other hand, gained over 30 percent. Figure 4 shows the targeting of the public health subsidy compared with the distribution of income in the country (measured by per capita total household expenditures). The cumulative shares of individuals in the population, ranked by per capita expenditure, are measured on the horizontal axis. The vertical axis measures the cumulative shares of expenditures and public health subsidy. The subsidy for health centers is distributed the most equitably, as it lies the closest to the 90 degree diagonal. The hospital subsidy is distributed in the least equitable manner. Health expenditures are, however, more equitably distributed than the distribution of expenditures in the country. To some degree, this indicates that the public health subsidy is relatively pro-poor, in that it is distributed more equitably than are expenditures. Figure 4: Concentration Curves for the Public Health Subsidy, 1995 100 80 all health share of subsidy/expendiure 60 health centers 40 20 expenditures hospitals 0 0 20 40 60 80 100 share of population There are two sources of the inequality in the distribution of public health expenditures. First, the non-poor tend to use public services much more than the poor. Second, the poorest quintiles mostly use the services provided at primary care facilities (health centers), which are less expensive than those provided at hospitals. As shown in Figure 3, the largest contribution of the public subsidy for the poorest quintile is from consultations at health centers (although their average per capita subsidy for this level, CFAF 1,175 is lower than that for the richest quintile, 1,836). In contrast, the richest members of society are more heavily subsidized for their use of hospitals. 14 How has the targeting of public health services changed over time? This is especially interesting to assess given that the health sector has been going through a series of reforms during the last decade. When comparing the 1995 results with those for 1986, it is evident that the targeting of public health services to the poor has worsened (Figure 5 and Annex Tables 5 and 6). The poorest 20 percent of the Ivoirian population received 11 percent of the total public health subsidy in 1995, down from 14 percent in 1986. The share accrued to the richest 20 percent increased from 24 to 31 percent. As discussed above, this comparative analysis is the result of an averaging of the unit costs across all facilities (as disaggregated information was not available for 1986). These results, therefore, are likely to underestimate the degree of inequality in health spending in Côte d’Ivoire (because the analysis could not take into account the higher unit costs embodied in services used mainly by the non-poor, such as in-patient hospital care), there is clear evidence of marked inequality in government health spending. Figure 4: Distribution of the Public Health Subsidy, 1986 and 1995 70.0 60.0 share of public health subsidy 50.0 40.0 1986 30.0 1995 20.0 10.0 0.0 urban rural 2 3 4 richest poorest Q uintiles The results suggest that the public provision of health care gives too much emphasis to services used by the better-off urban populations, having relatively little to offer to the poorest rural communities. The richest quintile dominates the use of publicly-provided health care. 1n 1995, the richest quintile receives 31 percent of the public health subsidy and the second-richest receives 23 percent. Similarly, 52 percent goes to urban areas and 48 percent to rural areas, although most Ivoirians live in rural areas. Furthermore, a comparison with the incidence of public health spending in Cote d’Ivoire a decade ago indicates that the targeting to the poor is getting worse, not better. Ideally, however, the public system should be targeted to the poor, since better-off groups are able to obtain care in the private sector. This is not the case in Côte d’Ivoire; the public health system appears to be geared towards the needs of the better off, most of whom relied on the state system. Of those reported as ill among the top decile, only 15 percent sought care in the private sector. 15 IV The Public Provision of Education Services This section focuses on the education sector, examines use patterns of education services by different social groups, assesses the distribution of public subsidies on education and whether government spending has been targeted to the poor. It begins with a review of the education system of Côte d’Ivoire. It then reviews trends in education outcomes, such as literacy and enrollments rates. Finally it examines the incidence of the public education subsidy in 1995 and assesses progress in targeting to the poor over the past decade by comparing the 1995 results with those for 1986. IV.1 The Education System Formal education in Côte d’Ivoire is provided in the context of a long-standing traditional system, originating before the arrival of French colonists in the nineteenth century. This continues to be practiced in rural areas, where families entrust the education of their children to a respected family elder. The traditional system emphasizes learning by practice and experience (rather than reading and writing which is the key to formal education). Traditional schooling is also closely entwined with the encouragement and continuation of family values, equipping the child to participate in society. That children do not attend formal schools, therefore, does not imply that no education is acquired. This feature of Ivoirian culture should be kept in mind in interpreting the analysis that follows. The formal education system. The formal school system in Côte d’Ivoire closely follows the French system and is organized into a three-tier scheme. Primary education is a six-year program that aims to reach children 6-11 years of age. It is designed exclusively to provide preparation for secondary-level education, a privilege reserved for about 20 percent of those who complete primary school (few of whom come from poor households). Since such a small percent of students go on to the secondary level, there is a need to reform the primary-level curriculum so that it better addresses the needs of the majority of children, who will not receive any further formal education. This is one in education sector reform that is periodically debated in the Cote d’Ivoire (World Bank, 1995b).6 Entry to general secondary schooling is granted upon passing an examination given at the end of the sixth year of primary school. Everyone who passes the exam has the ‘right’ to a public secondary education, which is divided into two secondary cycles. Lower general secondary schooling has a four-year curriculum, leading to a final examination called Brevet d’enseignment du premier cycle (BEPC). Upper general secondary education students enter major areas of study, and specialize in the arts, social sciences, science, and so on. During this cycle, students prepare of the Baccalaureate examination. All students who pass this exam are entitled to a scholarship to the University. There is also a separate technical and vocational training system at the secondary level, which currently trains about 3,600 students. Higher or tertiary education 6 Fora for debate on this issue include: the Commission de Reforme de 1972, the Etats Generaux de l’Education in 1985 and the Plan d’Action National pour l’Education Pour Tous in 1992. 16 has a two-track curriculum: the university institutions or facultés (80 percent of tertiary- level students) and the professional schools or grandes écoles (12 percent of students).7 Private school accounts for about 10 percent of the enrollments at the primary level and about 30 percent at the secondary level. There is no private schooling at the tertiary level. Enrollments in private primary schools have declined over time, from a high of 32 percent of all enrollments in 1959/60. At the secondary level, private enrollments increased until the early 1980s, when they stabilized at about 30 percent. This sector serves two groups: (a) the richer families who opt to pay to send their children to private school, and (b) those students who are sponsored by the state in the way of a transfer to the private school (these students who benefit from this program are also generally from a privileged class). The later program of transfers was designed so that the private sector would absorb the overflow from public schools (when there was not enough spaces at public schools) and at the same time ensure that all eligible students receive an education. This is because, as stated above, students who pass the entry exams to both secondary and tertiary-level education are guaranteed a place in school by the government. As the number of students increases, the public system avails of the private system to accommodate the additional enrollments. Public expenditures on education. Between 1980 and 1985, public spending for education increased substantially and became increasingly focused on primary-level education (Table 4). Total recurrent expenditure on education grew by over 60 percent during the five year period (check figures!). At the same time, the share of recurrent expenditures allocated to primary education increased from 39 percent in 1980 to 45 percent in 1985 (Table 4). This increase for primary education came at the expense of spending at the secondary level (both general and technical), which declined from 36 percent in 1980 to 30 percent in 1985. The share allocated to tertiary-level care remained constant at 18-19 percent of the total. Since 1985, spending has remained relatively stable in both size (growing only modestly over the 15 year period) and composition. The share of spending for primary education inched up to about 47 percent in 1995. Spending for secondary and tertiary- level education has remained fairly constant over this period, with the exception of 1990, when the share for secondary increased to 34 percent and the share for tertiary decreased to 14 percent of the total. Table 4: Public Expenditures for Education (Recurrent), 1980-1995 1980 1985 1990 1995* Primary 52.9 75.2 96.6 93.3 Secondary General 48.4 49.8 72.6 58.7 Secondary Technical 10.3 12 12 9 7 The remaining students are enrolled in teacher training programs. 17 Tertiary, General and Technical 24 31.8 30.5 37.6 Total 135.6 168.8 211.7 198.6 Total as % of GDP GDP Deflator(1980=100) 100 130.6 119.2 n.a. Real Total 135.6 220.5 236.9 Percent change 62.6 7.5 As a share of total Primary 39% 45% 46% 47% Secondary General 36% 30% 34% 30% Secondary Technical 8% 7% 6% 5% Tertiary, General and Technical 18% 19% 14% 19% Total 100% 100% 100% 100% *1995 figures are based on an extrapolation of disbursements through 8-31-95. Sources: For 1980-90, World Bank 1995c and for 1995, preliminary tables from the Ministry of Education (1995). During the 1980s and early 1990s, Cote d’Ivoire experienced difficult economic times, which resulted in significant budgetary cutbacks. (Education alone accounted for about 45 percent of the government’s recurrent expenditures.) The continued budgetary problems revealed structural rigidities and inefficiencies in the delivery of education. The government was unable to reduce expenditures in two critical areas: (a) costs associated with staffing (teachers’ pay, in particular, which accounted for about 50 percent of the total public sector wage bill); and (b) subsidies for students at the secondary and university levels. As a result, funding for non-personnel operating costs suffered. Public Spending in 1985 and 1995. The government subsidizes both its own state-run schools and (to a lesser extent) officially recognized private schools. The subsidies to private education can largely be considered a subsidy to the rich.8 Both subsidies are taken into account in the analysis that follows. Table 5 combines these public expenditures with school enrollments to estimate unit costs in the education sector for 1986 and 1995 (Annex Tables 7 and 8 provides more detail on the unit costs). For public education, two sets of enrollment estimates were calculated (and are both reported in the Annex Tables): those derived from government sources, and estimates from the household surveys (Living Standards Measurement Study for 1986 and the Priory Survey for 1995). The two sources are reasonably close, except for secondary enrollments for 1995, which differ markedly for reasons not understood. Our preference is to use the survey enrollments for purposes of computing the per unit subsidy (these are shown in 8 This is because the government subsidy only covers the tuition costs of education for students who do not have a place in public school. The private schools (especially at the secondary level) are often not nearby the student’s home, which means that travel and room and board costs are substantial. These additional costs are not covered by the government subsidy, effectively excluding the poor from benefiting from this subsidy (Laurent, 1995). 18 Table 5), since these are the enrollments which determine which households gain from government education subsidies.9 Table 5: Education Recurrent Expenditures: Unit Subsidies by Level of Education, for Public and Private education, 1985 and 1995 (CFAF per student) Level 1986 1995 Public Private Public Private Primary 69,469 30,031 64,848 15,817 Secondary 84,196 82,801 97,385 96,608 Secondary Technical n.a. 566,064 72,915 Tertiary 587,667 399,935 Tertiary Technical n.a. 1,939,292 1,577,135 Source: Annex Tables 7 and 8. During both years, there is a steeply rising scale of unit costs. In 1995, this ranged from just CFAF 64,848 for each primary school enrollment to CFAF 1,939,292 for each enrollment in a tertiary technical institution. Government subsidies allocated through private schools are not trivial. In 1995, each student enrolled in a secondary school gained an average of CFAF 96,608 from government support, which about the same as the public school subsidy and one-third more than the public primary unit subsidy. Cost recovery. Cote d’Iviore currently employs a cost recovery system at all levels of education except for the University-level, where fee are paid and students receive stipends. At the primary level, the fee is 2,500 CFA per student (equivalent to about US$ x). The fee for secondary general is x and ….[to be completed after more information obtained from Catherine and Bettina]. IV.2 Literacy and Education Enrollment Literacy. In 1990, 54 percent of all adults over the age of 15 and 40 percent of women were literate (Table 6). These literacy rates are slightly higher that the average for all of Sub-Saharan Africa. They are much lower, however, than those for Ghana, a neighboring country with a per-capita income one-third lower than in Cote d’Ivoire. They are also lower than the average for all less developed countries of 63 for all adults and 52 for women. This is additional evidence that indicators of human development lag behind what one would expect for Cote d’Ivoire, given it relatively high per capita income. Table 6: Adult Literacy Rate for Cote d'Ivoire and Selected Countries, 1990 9 If the lower enrollment figures from official sources were used, and the resulting subsidies were allocated to households based on the enrollments recorded by the survey, public expenditures would be over-allocated, in the sense that the total education subsidy recorded as allocated to households would be computed to be greater than government education spending itself. 19 Country Adult Literacy Rate (percent of ages 15+) Total Female Cote d'Ivoire 54 40 Other West African Countries: Burkina Faso 18 9 Ghana 60 51 Guinea 24 13 Senegal 38 25 All Africa 50 38 Less Developed Countries 63 52 Source: World Bank, 1994b Trends in Enrollments. Enrollment rates are generally quite low, and this is a cause for policy concern. Until the mid-1980s, enrollment rates at all levels of education in Cote d’Ivoire steadily improved. Gross enrollment rates at the primary were estimated to be about 20 percent in 1948. They increased until the early 1980s, when gross enrollment rates peaked at about 77 percent. They dipped after the early 1980s, and are now slowly recovering. In 1986, survey estimates indicate that the gross enrollment rate was 63 percent, increasing to 75 in 1995 (Table 7 and Annex Tables 9 and 10). These averages mask even lower rates for some groups: the poor, girls and those living in rural areas. The gross enrollment rate for the poorest quintile in 1995 was 39 percent, less than half of the 88 percent for the richest quintile. The gross enrollment rate for girls was 69, compared with 79 for boys. The enrollment rate was 63 percent in rural areas and 92 in urban areas. Enrollment figures from the Ministry of Education indicate similar enrollment rates (Annex Table 11). These low enrollment rates indicate late entry and grade repetition—both symptoms of weaknesses in the educational system. The net primary enrollment rate in 1995 was just 51 percent for the country as a whole, but was lower for girls (at just 48 percent), for the poor (the net primary enrollment rate for the poorest quintile being just 34 percent—and only 30 percent for girls in the quintile), and in rural areas in general (40 percent). 20 Table 7: Gross Enrollment Rates, Primary and Secondary, 1986 and 1995 (percent) 1986 1995 Female Male Total Female Male Total Quintile Primary 1 28 49 39 45 65 56 2 54 64 59 60 77 69 3 50 73 61 69 75 73 4 74 81 77 84 88 87 5 87 89 88 103 104 103 All Cote d'Ivoire 57 69 63 69 79 75 Rural 46 62 55 55 68 63 Urban 70 81 75 87 97 92 Secondary 1 8 15 12 11 13 12 2 11 29 21 19 38 29 3 27 32 29 18 34 26 4 31 49 40 17 31 24 5 45 85 64 47 82 62 All Cote d'Ivoire 26 42 34 24 38 31 Rural 12 26 20 6 18 13 Urban 39 62 50 39 73 54 Source: Living Standards Measurement Study, 1986, and Social Dimensions of Structural Adjustment Priority Survey, 1995 In the early 1990s, the Government set a target of 90 percent gross enrollment for primary school the academic year 2000/01. At that time, however, the Government (and the Bank) severely underestimated the growth in the school-aged population, as a result, meeting this target could require an annual increase of 1,000 classrooms and 1,200 teachers. In addition, a sustainable school construction program has not yet been established, and the gap between programmed and actual expenditures has been high in recent years.10 The Government plans, nevertheless, to meet this ambitious enrollment goal with measures that include sensitizing the population, assisting financially areas where enrollment rates are under 50 percent, suppressing fees for the poorest groups, and conducting studies and research (World Bank, 1995a). 10 The main reasons being a high dependency on external financing of these activities and difficulties in mobilizing these resources. 21 At the secondary level, enrollments expanded at an annual growth rate of 16 percent during the 1960s and 1970s and at a 4 percent annual growth rate in the 1980s. In the late 80s and early 1990s, the rates stabilized or even slightly declined (the exact pattern varies by source of information). Overall, they remain, however, extremely low, and among the poor they are almost non-existent. The gross enrollment rate in 1995 was 29, a decline from 34 percent in 1986. These figures, however, differ from those reported by the Ministry of Education, whichh are sligtly higher for the 1990s. Again, there is evidence of bias against girls. In 1995, the gross enrollment rate was 21 for girls and 36 for boys. Regional differences were also marked (see poverty profile for more detailed breakdown). The gross secondary enrollment rate was 48 percent in urban areas but just 18 percent for the rural population. The difference was even more marked for girls (34 percent compared with 6 percent). Tertiary-level enrollments (including university and technical) started at about 300 students in 1960 and now include about 30,000 students. Between 1974, when the Universite de C⊥te d’Ivoire was established (formerly Universite d’Abijan), and 1991/92 University enrollments increased five-fold. This expansion in enrollments has been most dramatic in recent years – between 1985/86 there was a 100 percent increase in enrollments. This trend seems to be continuing, as enrollments between 1991/92 and 1992/93 increased by 13 percent. Additionally, the most rapid growth of entrants to secondary general education have occurred since 1985/86, suggesting that the enrollments at the tertiary level will continue to swell eight to twelve years later (e.g. during the mid-to-late 1990s) (World Bank, 1995b). Trends in Girls’ Enrollment. Girls’ enrollments at the primary level have been consistently growing at a faster rate than boys (Table 8). Despite these higher growth rates, girl’s enrollment rates have continued to lag behind boys’ rates, and this is problem is more severe problem among the poor. For the poorest quintile, girls’ enrollment rates in 1995 were only slightly more that half those of boys in the same group (Table 7). The participatory assessment, conducted in 1995, identified some of the reasons given for the differential in enrollment rates. The main reason cited was ‘lack of means’ to send all children to school. When this is the case, families opt to send the boys to school and to keep the girls at home to help with the housework. Contributing to this behavior is the perception by many Ivoirians that there are few employment prospects for school graduates (INS, 1995b). To speed up the closing of the gap between boys’ and girls’ enrollments, several additional measures have been suggested (World Bank 1995c). The goal is for these measures to increase parents’ propensity to send their daughters to school. These include: a sensitization campaign encouraging girls to go to school; recruitment of more female teachers; separation of boys from girls where there are double shifts; more flexible time schedules to allow girls to combine school with housework; lowering the primary school entrance age for girls so they could finish the first cycle before marriage; provide financial support or reduced costs (e.g. reducing fees or providing free or less- expensive textbooks). 22 Table 8: Average Yearly Growth in Primary School Enrollments by Gender, 1959- 1993 (percent) 1959/60-1969/70 1970/71-1981/82 1982/82-1992/93 Girls 13.6 8.3 3.4 Boys 7.1 7.2 2.8 Total 9.0 7.3 3.1 Note: Includes public and private enrollments Source: MEN/DPES in World Bank (1995c:4) Efficiency. These low enrollment rates, coupled with the relatively high levles of spending reported earlier suggest poor efficiency in the sector. Internal efficiency can be measured in terms of how often students repeat a grade of school and how long it takes a student to work through each level of education. Table 9 provides (very strong) evidence of the poor internal efficiency at the primary and secondary level. This is particularly the case at the primary level, where it takes on average almost nine years to complete the first cycle, three years longer than is intended. In addition, earning both the BEPC and the BAC degrees takes almost twice as long than planned. The reasons for such high repetition rates are not entirely clear, but are in part due to archaic standards. In addition, many children repeat the sixth year of the primary level mostly for non- pedagogical reasons, as access to secondary school is limited. The consequences of this low efficiency are that older students stay in school and impede the entry of younger students. A quasi-automatic promotion from grade to grade would sharply reduce repetition rates and draw the gross enrollment rate closer to the net rate. This would enable a significant reduction in total recurrent costs per primary-level student. If the average number of years to complete primary school were to decrease from nine to six, this would result in a savings of about one-third in the total costs of providing a primary school education (World Bank 1995c). Table 9: Internal Efficiency in Public Education: 1992-93 Levels Years Expected Average Number of Years to Complete Primary 6 8.9 Lower Secondary 4 5.6 BEPC Degree 4 8.0 Upper Secondary (complete) 3 4.1 BAC Degree 3 7.5 Source: World Bank, 1995b. IV. 3 The Incidence of Government Education Spending 23 What do the low enrollments discussed in the previous section imply for government expenditures and their distribution among the Ivoirian population? By bringing together information on expenditures per enrollment provided with household data on enrollments, it is possible to trace the incidence of public spending on education across the population. The results of this exercise for 1995 are reported in Figure 6 (with more information provided in Annex Table 12). Households were allocated the appropriate subsidy for each child enrolled in both public and private schools (at the primary, secondary, secondary technical, tertiary and tertiary technical levels). Figure 6: Per Capita Education Subsidy, by Level of Education and Quintile 30000 25000 20000 tertiary technical CFAF per capita tertiary 15000 secondary technical secondary 10000 primary 5000 0 urban rural 2 3 4 average richest poorest Q uintile s On average, in 1995 the Ivoirian population gained CFAF 15,560 per capita from subsidies paid through the education sector as a whole, CFAF 13,516 on average from public schooling and CFAF 2,043 from subsides to private schools. However, this education subsidy was very unequally distributed across the population. There were three main dimensions to this inequality. First, rural areas gained just CFAF 9,911 per capita during the year, less than half the CFAF 23,445 per capita that went to urban areas. Second, poorer groups gained less than the non-poor. The poorest quintile received a per capita subsidy (CFAF 10,995) equivalent to about one-third of the subsidy going to the richest (CFAF 27,396). About 14 percent of the total education subsidy benefited the poorest quintile, and less than 30 percent went to the poorest 40 percent of the population. The richest quintile, on the other hand received 35 percent of the education subsidy. This inequality applies to spending through both public and private schooling. Although spending through private schooling was extremely unequal (the poorest quintile gaining just 9 percent of the subsidy compared with 50 percent to richest quintile), spending on public schooling dominated (its pattern across the quintile being similar to the pattern of overall spending). The third source of inequality comes from gender differences. Overall, females gained only about a third of total education subsidies, with the inequality more marked among lower quintiles. In not sending girls to school, households are in effect allowing public education expenditures to benefit boys 24 disproportionately as compared with girls. On average the per capita education subsidy to boys is almost twice that to girls, and this appears to be similar in both urban and rural areas. Another way to examine the targeting of the public education subsidy is to compare the distribution of the education subsidy to the distribution of income in the country (measured by per capita total household expenditures). This is shown in Figure 7. Figure 7: Concentration Curves for the Education Public Subsidy, 1995 (Academic Education only) 100 80 all education share of subsidy/expendiure 60 primary 40 tertiary 20 expenditures secondary 0 0 20 40 60 80 100 share of population The cumulative shares of individuals in the population, ranked by per capita expenditure, are measured on the horizontal axis. The vertical axis measures the cumulative shares of expenditures and public education subsidy. The primary-level subsidy is distributed the most equitably, as it lies very close to the 90 degree diagonal. The concentration curve for all education combined lies below the diagonal, indicating that it favors the rich, but above the Lorenz distribution of expenditures. This means that the public education expenditures are more progressively distributed in the population than is wealth, and as such can be considered to have mild redistributive effects (although it is inequitably distributed in absolute terms). How has the targeting to the poor of the education subsidy changed over time? Comparing results from 1986 and 1995, we can see that the targeting of the education has improved during this decade (Figures 8 and 9, with further detail in Annex Tables 13 and 14 and Annex Figures 1 and 2). This comparison, however, only considers academic education as enrollment information for technical education was not available for 1986. Overall, the share of the public education subsidy going to the poorest quintile has increased more than two-fold during the past decade, from 7 percent in 1986 to 16 25 percent in 1995 (Figure 7). Likewise, the share for the richest quintile decreased from 46 percent in 1986 to 28 percent in 1995. The most important gains for the poor were at the primary level, where the share of the subsidy going to the poorest quintiles increased from 12 percent in 1986 to 20 percent in 1995. Nevertheless, as discussed above, Cote d’Ivoire still has a long ways to go before the public education subsidy is equitably distributed. Figure 8: Distribution Public Education Subsidy, 1986 and 1995 (All Education) 80 70 share of expenditures 60 50 1986 40 30 1995 20 10 0 urban rural 2 3 4 richest poorest quintile s Figure 9: Distribution Public Primary-level Education Subsidy, 1986 and 1995 60 50 share of expenditures 40 1986 30 1995 20 10 0 urban rural 2 3 4 richest poorest quintile s There are two main causes underlying the inequalities in the distribution of public education expenditures. First, household decisions on sending children to school have a direct effect on how government education expenditures are distributed across the population. The share of the poorest quintile in the total subsidy going to primary schooling is significantly less than that of the richest quintile, which is an uncommon finding in studies of this kind.11 Thus, although the poorest quintile is likely to have a much greater share of the primary school-aged population than the richest, their decisions 11 In Kenya, for example, Demery and Verghis (1994) found that the poorest quintile gained 22 percent from primary school expenditures, due mainly to the large numbers of primary-school aged children in the quintile. The richest quintile gained only 15 percent of the primary subsidy because it had few primary school-aged children. 26 not to enroll these children in school means that they gain little from government spending in the sector. Second, government expenditure allocations to the various levels of education also cause inequality in the incidence of spending. Government spending on primary education, for example, is far better targeted to the poor than all other sub- sectors. The poorest quintile gained almost 20 percent of primary school spending, but just 8 percent of secondary school spending, and none from secondary and tertiary technical education.12 Given that the government allocates almost 80 percent of its education budget to primary and secondary schooling (50 percent to primary schooling—see table 4), it would appear that the main source of inequality in education spending in Côte d’Ivoire comes from differences in household behavior—the decisions taken by households whether to send their children to school. Whether the low enrollment rates are due to supply-oriented factors (such as the access to schools or the quality of schooling) or to demand factors (low and declining income levels, and increasing opportunity costs of schooling) is critical for policy. It may be that although government spending on basic education is high, the composition of such spending may lead to poor services to the poor. If spending is targeted mainly to urban areas, there may be persistent problems of access to schooling by poorer rural communities. On the other hand, if the composition of spending (for example too little spending on non-salary components which appears to be the case from Annex Table 8) the quality of schooling may suffer, and households may find the service of little practical use. In either case, public expenditure reforms are called for. But if the main reason for low school enrollments among the poor is the poor economic climate in general, then other policies are called for (including those which restore economic growth and living standards, and possibly providing school scholarships for the poor). This is clearly a topic requiring further policy research. V. Comparisons with Other Countries Public spending on both education and health appear to be very poorly targeted to the poor in Côte d’Ivoire. How does the incidence of such spending compare with other countries? Table 10 compares the share of total government subsidies on education and health reaching the poorest and the richest quintiles in six other African countries and two countries from outside the region. 12 The poorest quintile is estimated to have gained 19 percent from tertiary, but this can be considered as an outlier in the data. The number of enrollments at this level picked up in the PS sample is very small. It so happened that some came from quintile 1. 27 Table 10: Incidence of Public Social Expenditures in Selected Countries (percent) Health Education Poorest Richest Poorest Richest 20% 20% 20% 20% Côte d’Ivoire (1995) 11 31 14 35 Africa Ghana (1992) 12 33 16 21 Guinea (1994) 4 48 5 44 Kenya (1992/3) 14 24 17 21 Madagascar (1993)* 18 24 9 44 South Africa (1993) 16 17 10 38 Tanzania (1993) 13 23 14 40 Other Countries Colombia (1992) 28 12 28 12 Malaysia (1989) 29 11 29 11 Sources: Demery and Dayton (1995). * preliminary results Inter-country comparisons must be handled with care, given the different methodologies and underlying data employed in each country. And the results also reflect the broader context of the education and health systems as a whole in the countries—the roles of the public and private sectors being an important element. The results do provide, however, the order of magnitude in inter-country differences. Compared with other African countries, the targeting of social spending in Côte d’Ivoire appears to be very weak. The respective shares of education spending in Côte d’Ivoire are the same as those in South Africa, a pattern which reflects years of apartheid policies in education in the latter. And health spending in Côte d’Ivoire seems to be among the poorest targeted in the region (only Guinea recording a lower relative share to the poorest quintile). By way of contrast, the two non-African countries reported in the table illustrate the potential for a much more targeted approach to social sector spending. In both Colombia (in 1992) and Malaysia (in 1989) social sector spending favored the poor rather than the better-off. These are the patterns that education and health sector reforms in Africa in general, and in Côte d’Ivoire in particular should aim to achieve. VI. Conclusions and Main Messages Social indicators show that social development is lagging in Côte d’Ivoire, compared to countries with similar income levels. The analysis presented here has also shown that public expenditures in the social sectors (health and education) are not distributed in a pro-poor way. The inequality is much more severe for health than for education. The poorest 20 percent of the population received only 11 percent of the public subsidy, and the richest 20 percent receive 31 percent. The is because the rich use public services much more so that the poor and because they predominantly avail 28 themselves of the more expensive public hospital-based services. At the same time, the rich use very few services in the private sector. This is unlike the pattern we would expect, where the poor use public services and the rich use mostly private services. The evidence also suggests that targeting of health services to the poor has actually worsened since 1986, when the poorest quintiles received about 14 percent of the public health subsidy and the richest quintile received 24 percent. This analysis provides strong evidence that the poor in Cote d’Ivoire do not benefit from public spending in education, although the inequality is not as pronounced as in the heath sector. The inequity is greatest at the higher levels of education, but it is of greatest concern at the primary level. Overall, only about 14 percent of the total education subsidy benefited the poorest quintile, and less than 30 percent went to the poorest 40 percent of the population. The richest quintile, on the other hand received 35 percent of the education subsidy. Yet, the poor stand to benefit the most from publicly- provided primary education, and they are the least likely to be able to buy it for themselves in the private sector. As discussed above, the reasons for this inequity are two-fold. First, they are the result of declining household demand for education. The private costs of schooling are high at a time when there is a poor employment outlook for graduates form all levels of education. At the same time, in the depressed economy opportunity costs for children’s time are high, especially for girls. Second, government resources could be increasingly allocated to primary and secondary education, and away from tertiary-level education, which subsidizes a limited number of wealthier student (who could afford to invest in their own education). In addition, a large proportion of recurrent costs at all levels of education are spent on teacher salaries, at the expense of cuts in much needed non-salary supplies. These results provide evidence of the continued need to reform the health and education sectors in Cote d’Ivoire. In health, there is a need to improve access by the poor to services. Spending needs to be re-oriented away from urban-based services to primary facilities serving rural areas. In education, the first step is more research into why enrollments especially of the poor especially at the primary level are so low despite significant budget allocations in sector. Some more specific policy recommendations follow: Health sector policy recommendations Shift resources to primary-level care. From this analysis it is clear that the poor are most likely to use public primary health care facilities. At the same time, the Ivoirian Government continue to favor funding for tertiary-level facilities over primary-level care, despite rhetoric to the contrary. The first priority needs to be to expand funding of primary-level services, if public health care is to have its intended impact on the lives of the most vulnerable Ivoirians. Strengthen the implementation of the cost recovery program. A well-run cost recovery program has the potential to provide additional funding at the local level, which when used correctly can help to enhance the provision of services. It can also help to 29 ensure the availability of essential drugs in primary-level health care facilities that might otherwise not be well-stocked. At the same time, it can help to improve the quality of the serviced received in tandem with the pharmaceuticals. Yet these are lofty goals that depend on well-run local management and supervision. The system in place in Cote d’Ivoire envision a large role for the locally-run ‘comites de gestion,’ but to date little effort has been made to facilitate the development of these local supervisory boards. Supporting and strengthening the development of the cost recovery system at the local level should, therefore, be a priority within the sector. At the same time, there is a need to investigate how the fees for consultations and drugs are affecting the utilization of health care. It is possible that the recent installation of a fee schedule has had adverse effects on the utilization of public health care by the poorest groups. After analyzing the situation, it is possible that adjustments on some fees could be made that would that promote more appropriate utilization of some services. Promote private sector development. It would also be appropriate for the Ivoirian government take steps to promote the development of the private health sector. The goal would be to encourage a private sector that could provide care for the richest Ivoirians. This would free resources in the public sector (that were previously monopolized by this group) that could be used to provide health care for those who cannot afford to provide for themselves. Education sector policy recommendations Expand primary-level enrollments of the poor. These results suggest the need for several policy changes. First, what can be done to increase enrollment, especially of the poor? The focus should be on providing all children with a basic education at the primary level. (It would be ideal if all children could stay in school to complete secondary education, but in light of difficult economic circumstances in Cote d’Ivoire and an economy with few prospects for employment after graduation, this is currently an unrealistic target.) As discussed above, more research is needed on the causes of low enrollments. If it due to ineffective spending, additional public expenditure reforms are called for. If it is due to deterioration in household circumstances, more deep-seated policies are called for. Several of the following policy changes would probably help to increase enrollments. Updating the primary-level curriculum to make it more appealing and relevant to the needs of children who are not necessarily preparing for advanced study is one way to increase the demand for primary education. A change in the repetition policy at the primary level would also make room for more children of school age to be enrolled and allow older children to move on in their education. One solution is a quasi-automatic promotion from grade to grade, which would sharply reduce repetition rates and would enable a reduction in total recurrent costs per primary-level student. Finally, it is obvious that an improved economic situation would help to provide discretionary income for families to spend on school fees and might improve the employment prospects of school graduates. 30 Increase public spending on primary-level education. On the supply side, difficult economic times and budget cutbacks are demanding a rigorous prioritization of resources in all sectors of the Ivoirian economy, including education. From an equity standpoint, the focus must be supporting primary-level education, even if this comes at the expense of funding for university-level study. The Ivoirian tradition of providing a university scholarship to all who pass the baccalaureate examination is in direct conflict with a potential policy to cut spending at the tertiary level, making this a very difficult political issue. In addition to equity considerations, there is evidence that the Ivoirian economy is not absorbing university graduates at the rate they are being produced, indicating that continued high public spending in this sub-sector to be inefficient as well as inequitable (World Bank, 1995c). Reform the wage bill. Finally, the dominance of the wage bill in recurrent budget for education has caused budget cuts to be made in non-wage spending. This is a continuing problem that has been addressed in other reports (World Bank, 1995c and 1995d), and it is imperative that the Ivoirians continue to search for a solution to the unnecessarily high wage bill. Expenditures that could be saved from this area could be re-allocated for needed items like school supplies and teacher training, which would increase the efficiency of the public education system. 31 References Demery, Lionel (1994) ‘Côte d’Ivoire: Fettered Adjustment’ in Ishrat Husain and Rashid Faruqee (eds) Adjustment in Africa: Lessons from Country Case Studies, World Bank Regional and Sectoral Studies, Washington D.C. Demery, Lionel, and Julia Dayton (1995) ‘International Comparisons of Public Expenditure Incidence’ (Unprocessed mimeo) Poverty and Social Policy Department, World Bank. Demery, Lionel and Mathew Verghis (1994) ‘The Incidence of Education Expenditure in Kenya’ (unprocessed mimeo) Education and Social Policy Department, World Bank. Grootaert, Christiaan (1993) The Evolution of Welfare and Poverty under Structural Change and Economic Recession in Côte d’Ivoire , 1985-88, Policy Research Working Paper 1078, World Bank, Washington D.C. Grootaert, Christiaan (1994) ‘Analyzing Poverty and Policy Reform: the experience of Côte d’Ivoire’ (Unprocessed mimeo), World Bank. Institut National de la Statistique (INS) (1995a) Analyse Descriptive des Problemes de Sante. Draft notes for the World Bank. Institut National de la Statistique (INS) (1995b) Enquete Qualitative sur l’Evolution du Niveau de Vie en Cote d’Ivoire. Preliminary Draft (not for circulation). Meerman, Jacob (1979) Public Expenditures in Malaysia: Who Benefits and Why? (New York: Oxford University Press). Ministere de la Sante et des Affaires Sociales (MSPAS) (1995) Examen des Depense Publiques dans le Secteur de la Sante. Republique de Cote d’Ivoire. Selowsky, Marcelo (1979) Who benefits from Government Expenditure? (New York: Oxford University Press). Van de Walle, Dominique, and K. Nead (1995) Public Spending and the Poor—Theory and Evidence Johns Hopkins University Press, Baltimore, Md. World Bank (1994a) ‘Human Resources Development Program, Education Budget Performance Matrix’ Africa Technical Department. _____. (1995a) Better Health in Africa: Experience and Lessons Learned. Washington, D.C. 32 _____. (1995b) Implementation Completion Report, Human Resources Development Program for the Republic of Cote d’Ivoire. Population and Human Resources Operations Division, West Central Africa Department, Africa Region. _____. (1995c) Cote d’Ivoire Education Sector Report. Population and Human Resources Operations Division, West Central Africa Department, Africa Region. _____. (1995d) Republic of Cote d’Ivoire Country Economic Memorandum. Country Operations Division, West Central Africa Department, Africa Region. Draft of March 20, 1995. 33